Indian

Epistemology and Metaphysics

 Also available from Bloomsbury

The Bloomsbury Research Handbook of Indian Aesthetics and the Philosophy of Art,
Arindam Chakrabarti
The Bloomsbury Research Handbook of Indian Ethics,
Shyam Ranganathan
The Collected Writings of Jaysankar Lal Shaw: Indian Analytic and Anglophone Philosophy,
edited by Jaysankar Lal Shaw
An Introduction to Indian Philosophy,
Christopher Bartley
Indian Epistemology and Metaphysics

Edited by
Joerg Tuske

Bloomsbury Academic
An imprint of Bloomsbury Publishing Plc
 Contents

Contributors

Introduction: Indian Epistemology and Metaphysics Joerg Tuske

Part I Knowledge, Language, and Logic

1 Epistemology, Metaphysics, and Logic in Prācīna Nyāya and Buddhist
Philosophy Claus Oetke
2 Fallacies and Defeaters in Early Navya Nyāya  Stephen H. Phillips
3 Jayarāśi and the Skeptical Tradition  Eli Franco
4 Jainism: Disambiguate the Ambiguous Piotr Balcerowicz

Part II Consciousness and the External World

5 Proofs of Idealism in Buddhist Epistemology: Dharmakīrti’s Refutation of External
Objects Birgit Kellner
6 Materialism in Indian Philosophy: The Doctrine and Arguments Pradeep P. Gokhale
7 Sāṁkhya: Dualism without Substances Ferenc Ruzsa
8 Śaiva Nondualism Raffaele Torella
9 An Indian Debate on Optical Reflections and Its Metaphysical Implications: Śaiva
Nondualism and the Mirror of Consciousness Isabelle Ratié

Part III Universals and Momentary Existence

10 A Road Not Taken in Indian Epistemology: Kumārila’s Defense of the Perceptibility of
Universals John Taber
11 The Role of Causality in Ratnakīrti’s Argument for Momentariness Joel Feldman

Part IV Self, No-Self, and Self-Knowledge

12 Self or No-Self? The Ātman Debate in Classical Indian Philosophy Alex Watson
13 Where the Self and Other Meet: Early Indian Yogācāra Buddhist Approaches to
Intersubjectivity Roy Tzohar
14 Luminous Mind: Self-Luminosity versus Other-Luminosity in Indian Philosophy of
Mind Matthew MacKenzie
15 Solving Kant’s Problem: K. C. Bhattacharyya on Self-Knowledge Jay L. Garfield

Part V Concepts and Cognitions

16 Nyāya Theory of Concepts Keya Maitra
17 Vasubandhu’s Arguments for the Cognition of Nonexistent Objects Zhihua Yao

Glossary of Frequently Used Sanskrit Terms

Chronological Table of Main Indian Thinkers and Texts Mentioned in This Volume
Index
 Contributors

Piotr Balcerowicz, of no nationality (which he emphasizes), is professor of philosophy

and Indian studies, currently based in Warsaw, Poland. He specializes in the Indian
philosophical tradition, with emphasis on epistemological thought and Jainism. He teaches
Indian philosophy and Indian religion, as well as intercultural relations, conflict
resolution, and contemporary history of South Asia, Central Asia, and the Middle East. He
has published extensively on Indian philosophy and religion, especially Jainism, but also
on the Middle East and Central Asia and Afghanistan. Since 2002, with his NGO
Education for Peace, he has been involved in various development cooperation projects in
Afghanistan, Pakistan, Burma/Myanmar, and Africa.
Joel Feldman was born in New York City in 1968. He received a BA from Hampshire
College in 1994 and a PhD in philosophy from the University of Texas at Austin in 2003.
He is currently associate professor of philosophy at Rider University, where he has been
teaching since 2004. He is coauthor of Ratnakīrti’s Proof of Momentariness by Positive
Correlation: Transliteration, Translation, and Philosophic Commentary (2011).
Eli Franco is the director of the Institute for Indology and Central Asian Studies, Leipzig
University, and a member of the Saxon Academy of Sciences. He has published extensively
on various traditions of Indian philosophy, especially on Lokāyata and the Buddhist
epistemological tradition. He is the author of Perception, Knowledge and Disbelief: A
Study of Jayarāśi’s Scepticism (1987 and 1994); Dharmakīrti on Compassion and Rebirth
(1997); The Spitzer Manuscript: The Oldest Philosophical Manuscript in Sanskrit (2004);
Dharmakīrti on the Duality of the Object (2014; with Miyako Notake). His editorial work
includes Beyond Orientalism: The Work of Wilhelm Halbfass and Its Impact on Indian and
Cross-Cultural Studies (1997; with Karin Preisendanz); Yogic Perception, Meditation and
Altered States of Consciousness (2009); From Turfan to Ajanta. Festschrift for Dieter
Schlingloff on the Occasion of his Eightieth Birthday (2010; with Monika Zin); Religion
and Logic in Buddhist Philosophical Analysis, Proceedings of the 4th International
Dharmakīrti Conference (2011; with Helmut Krasser et al.); and Historiography and
Periodization of Indian Philosophy (2013).
Jay L. Garfield is Kwan Im Thong Hood Cho Temple Professor of humanities and head of
studies in philosophy at Yale-NUS College, professor of philosophy at the National
University of Singapore, recurrent visiting professor of philosophy at Yale University,
Doris Silbert Professor in the humanities and professor of philosophy at Smith College,
professor of philosophy at Melbourne University, and adjunct professor of philosophy at
the Central University of Tibetan Studies. Garfield’s most recent books include Engaging
Buddhism: Why It Matters to Philosophy (2014); Madhyamaka and Yogācāra: Allies or
Rivals? (2015; edited, with Jan Westerhoff); The Moon Points Back: Buddhism, Logic and
Analytic Philosophy (2015; edited, with Yasuo Deguchi, Graham Priest and Koji Tanaka);
Indian Philosophy in English from Renaissance to Independence (2011, with Nalini
Bhushan); and Contrary Thinking: Selected Papers of Daya Krishna (2011, edited with
Nalini Bhushan and Daniel Raveh).
Pradeep P. Gokhale retired as professor from the Department of Philosophy, University
of Pune, in 2012. Since then he has been working as Dr. B. R. Ambedkar Research
Professor in the Central University of Tibetan Studies, Sarnath, Varanasi, India. He is the
author of Inference and Fallacies Discussed in Ancient Indian Logic (With Special
Reference to Nyāya and Buddhism); Vādanyāya of Dharmakīrti (The Logic of Debate);
Hetubindu of Dharmakīrti (A Point on Probans); coauthor of Recollection, Recognition and
Reasoning (A Study in the Jaina Theory of Parokṣa-pramāṇa); the editor of: The Philosophy
of Dr. B. R. Ambedkar; and coeditor of Studies in Jainism and Indian Moral Philosophy:
Problems, Concepts and Perspectives. He is also the author of many books and articles in
Marathi. These books and articles pertain to themes in Classical Indian philosophy, social
and moral philosophy and philosophy of religion.
Birgit Kellner has been director of the Institute for Cultural and Intellectual History of
Asia at the Austrian Academy of Sciences in Vienna since 2015. Her research focuses on
Buddhist epistemology and logic in India in its interaction with non-Buddhist philosophies,
and on Tibetan adaptations and interpretations of Indian Buddhist thought. She has
worked especially on the knowledge of nonexistence (the pivotal concept being “non-
apprehension,” anupalabdhi), and on the nature and structure of consciousness and
cognition, as epitomized in controversies surrounding the status of “aspects” (ākāra) and
the notion of “reflexive awareness” or svasaṃvedana. Kellner earned her degrees from
the Universities of Vienna (Tibetan and Buddhist Studies, Mag. phil. 1994) and Hiroshima
(Indian Philosophy, PhD 1999), and subsequently pursued her research on Buddhist
epistemology, logic, and the philosophy of mind in the framework of various prestigious
research fellowships and projects in Vienna and Hamburg. After an appointment as
visiting assistant professor at UC Berkeley she became professor for Buddhist studies at
the University of Heidelberg in 2010 before taking up her present position.
Matthew MacKenzie is associate professor of philosophy at Colorado State University.
He works in the areas of Buddhist philosophy, Indian philosophy, and philosophy of mind.
Keya Maitra is professor and chair of the Department of Philosophy at University of
North Carolina, Asheville. Her current research interests include Indian philosophy,
philosophy of mind, Third World feminism, and feminist philosophy of mind. She has
published articles in Asian Philosophy, Hypatia, Philosophy in the Contemporary World,
Southwest Philosophy Review, and International Journal of Philosophical Studies along
with edited volumes. She contributed On Putnam (2002) to the Wadsworth Philosophers
Series. Her recent publications include “Mindfulness, Anātman and the Possibility of a
Feminist Self-Consciousness” and “The Questions of Identity and Agency in Feminism with
Borders: A Mindful Response.” Her The Philosophy of the Bhagavad-Gītā containing a
fresh translation of the Hindu text Bhagavad-Gītā prepared for philosophers is
forthcoming. She is coediting the volume Feminist Philosophy of Mind. She also received a
Fulbright-Nehru Academic and Professional Excellence Senior Research Award to India to
work on her project on the epistemology of mindfulness.
Claus Oetke was born in Germany and trained at the University of Hamburg (Germany)
in Indology, philosophy, and Sinology. He was awarded a doctorate in 1973, following
which he was a lecturer (“wissenschaftlicher Assistent”) at the Institute for Culture and
History of India and Tibet in Hamburg until 1983. He received a habilitation in 1983 and
professorship at the same institute. He served as a short-term visiting professor in Vienna
(Austria) and Kiel (Germany). From 1987 until 1992 he held the “Heisenberg Professur” of
the “Deutsche Forschungsgemeinschaft” (DFG), which led to visiting professorships at the
University of Pennsylvania (Philadelphia, USA), The Australian National University in
Canberra (Australia), and the University of Texas at Austin (USA). From 1993 until
January 2014, he was the head of the Department of Indian and Tibetan Studies at the
University of Stockholm. Since February 2014 he has been professor emeritus of the
University of Stockholm. His main topics of research include Tibetan translations from
Chinese Buddhist texts, problems of person and self in the Indian philosophical tradition,
studies in early Madhyamaka-Buddhism, Indian theories of inference and proof (“Indian
Logic”), theoretical issues of interpretation, and linguistic semantics.
Stephen H. Phillips is professor of philosophy and Asian studies at the University of
Texas at Austin (appointed 1982), and has been visiting professor of philosophy at the
University of Hawaii, Manoa (1995), and Jadavpur University, Kolkata (2008). He received
a PhD from Harvard University (1982) after having attended Harvard College (A.B. 1975)
and a yoga ashram school in India. Author of seven books, including Yoga, Karma, and
Rebirth named by Choice an “Outstanding Academic Title” for 2010, he has recently
published Classical Indian Epistemology: The Knowledge Sources of the Nyāya School
(2012), which presents classical Indian views about knowledge in terminology suited for
the Western philosophy professional. Phillips has teamed with N. S. Ramanuja Tatacharya
to translate about a third of the massive and monumental fourteenth-century Tattva-cintā-
maṇi by Gaṅgeśa. He is editor or coeditor of several anthologies and has published more
than sixty papers in scholarly journals.
Isabelle Ratié is professor of Sanskrit language and literatures at the Sorbonne Nouvelle
University (Paris). She has published several monographs on Śaiva and Buddhist
philosophies (Le Soi et l’Autre. Identité, différence et altérité dans la philosophie de la
Pratyabhijñā, 2011, Weller Prize 2012; Une Critique bouddhique du Soi selon la
Mīmāṃsā, 2014; and Self, No-Self, and Salvation. Dharmakīrti’s Critique of the Notions of
Self and Person, 2013, with Vincent Eltschinger). She has also edited with Eli Franco the
collective volume Around Abhinavagupta. Aspects of the Intellectual History of Kashmir
from the Ninth to the Eleventh Century (2016). She is currently working on the edition
and annotated translation of several Śaiva and Buddhist works, including some hitherto
unknown fragments of Utpaladeva’s lost Vivṛti on the Īśvarapratyabhijñā treatise, and, in
collaboration with Vincent Eltschinger, the second chapter of Śaṅkaranandana’s
Dharmālaṅkāra.
Ferenc Ruzsa was born in Budapest, Hungary, in 1957 and got his MA in Indology and
philosophy at ELTE University, Budapest. After ten years of being a computer programmer
he started to teach at the Institute of Philosophy, ELTE, in 1993; since 1998 he is
associate professor at the Department of Ancient and Medieval Philosophy. He also taught
for thirteen years at the Buddhist College, Budapest (in 2007–2009, he was rector of the
college). He received his PhD from the Hungarian Academy of Sciences in 1995 for a
thesis on Īśvarakṛṣṇa’s philosophy. His researches focus on early Indian philosophy, with
emphasis on its origins, Buddhism, and Sāṁkhya. He is the author of two books and more
than fifty papers, all of which can be read at https://elte.academia.edu/FerencRuzsa.
John Taber is professor of philosophy at the University of New Mexico. He has published
two books: Transformative Philosophy: A Study of Śaṅkara, Fichte, and Heidegger (1983)
and A Hindu Critique of Buddhist Epistemology: Kumārila on Perception (2005) and
numerous articles on Mīmāṃsā, Advaita Vedānta, Buddhist epistemology, and Indian logic.
Raffaele Torella is professor of Sanskrit at the University of Rome “Sapienza,” where he
has also taught Indian philosophy, Indian religion, and Indology. He is the coordinator of
the South Asia Section in the Sapienza Doctoral Course in “Civilizations and Cultures of
Asia and Africa.” He is the editor-in-chief of Rivista degli Studi Orientali and Rivista di
Studi Sudasiatici. His main fields of research are Kashmiri Śaivism, linguistic speculation,
Buddhist epistemology, and manuscriptology. His recent book publications include
Utpaladeva, Philosopher of Recognition (2016; coeditor with B. Bäumer); The
Īśvarapratyabhijñā-kārikā of Utpaladeva’s with the Author’s Vṛtti. Critical Edition and
Annotated Translation (2013); Śivasūtra with Ksemarāja’s Vimarśinī (2013; in Italian); The
Philosophical Traditions of India: An Appraisal (2011); Eros and Emotions in India and
Tibet (2007; with G. Boccali; in Italian). He has been the scientific editor and coauthor of
the section “Science in India” in the multivolume work History of Science (2002; in
Italian). Along with Bettina Bäumer, he has organized the first International Workshop on
Utpaladeva (Indian Institute for Advanced Study, Shimla, 2010) and the conference on
“The Human Person and Nature in Classical and Modern India” (Rome, 2013). His
forthcoming book is The Saṃvitprakāśa of Vāmanadatta, a critical edition and English
translation.
Joerg Tuske is professor and chair of the Philosophy Department at Salisbury University,
Maryland. He graduated from the Universities of London and Cambridge and spent a year
at the University of Pune in India. He is the author of several articles on Indian
epistemology, logic, and philosophy of mind.
Roy Tzohar specializes in the history of philosophy with a focus on Buddhist and
Brahmanical philosophical traditions in India. He is currently a tenured faculty in the East
Asian Studies Department at Tel Aviv University. He holds a PhD (with distinction) from
the Religion Department at Columbia University (New York, 2011), and an MA in
philosophy (summa cum laude) from Tel Aviv University’s Interdisciplinary Program for
Outstanding Students (2004). His current research, for which he was awarded the Marie
Curie IRG fellowship of the EU, concerns intersubjectivity and language in the Indian
Buddhist Yogācāra thought. He is also working on a book manuscript on the works of
Buddhist poet and philosopher Aśvaghoṣa. His monograph Meaning in the World and in
Texts: A Buddhist Theory of Metaphor is forthcoming.
Alex Watson is professor of Indian philosophy at Ashoka University, Delhi, prior to which
he was preceptor in Sanskrit at Harvard University. He is the author of The Self’s
Awareness of Itself (2006) and An Enquiry into the Nature of Liberation (2013; with
Dominic Goodall and Anjaneya Sarma), as well as numerous articles in such journals as
the Journal of Indian Philosophy. After completing his DPhil at the University of Oxford, he
held research fellowships at Wolfson College, Oxford, the French School of Asian Studies,
Pondicherry, and Kyushu University and was a guest lecturer at the University of Vienna.
Zhihua Yao is associate professor of philosophy at the Chinese University of Hong Kong.
His research interests cover Buddhist philosophy, Indian philosophy, and philosophy of
religion. His publications include The Buddhist Theory of Self-Cognition (2005) and
various articles in the Journal of Indian Philosophy, Philosophy East and West, the Journal
of Indian Council of Philosophical Research, the Journal of Chinese Philosophy, the
Journal of Buddhist Studies, and Comparative Philosophy.
 Introduction: Indian Epistemology and Metaphysics
Joerg Tuske

The present volume provides an overview of some of the recent developments in Indian
epistemology and metaphysics. However, with the exception of the author of Chapter 15,
all of the contributors to this anthology discuss philosophers who lived before 1700 CE.
This means that the term “recent developments” does not refer to the philosophical ideas
which are discussed but rather to the historical scholarship that helps us to understand
these ideas and put them into context. The terms “epistemology,” “metaphysics,” and even
“philosophy” are of course Western concepts with Greek roots. There has been
considerable debate about whether these terms can be applied to the Indian context and
the history of the study of Indian texts demonstrates this. For example, some Western
philosophers, such as Edmund Husserl, have claimed that Indian texts are religious and
mystical and lack the argumentative rigor of the philosophical texts of the Western canon.
In particular, they contend that the Enlightenment idea of a separation of philosophy and
religion and the resulting development of philosophy as a separate academic discipline is
a uniquely Western phenomenon which has shaped the nature and guiding questions of
the discipline. According to this line of thought, the absence of an Enlightenment in other
parts of the world, including India, makes it impossible to use categories such as
“philosophy” in the Indian context. I am pleased to say that this view has become very
much a minority view due in large part to the efforts of philosophers, such as Bimal
Matilal, who have pointed out the similarities between the questions addressed by
historical Indian writers and Western philosophers, particularly analytical philosophers.
However, within this process interpreters have sometimes overemphasized these
similarities and not paid enough attention to the specific cultural and intellectual
background of these Indian writers.
The present volume is a compilation of chapters by contributors who take the cultural
and intellectual background of the Indian texts under discussion seriously and aim to
interpret these texts on their own terms. As is obvious from the chapters, this requires a
significant level of scholarship, in particular linguistic expertise. This expertise has
traditionally been the domain of Indology. While many of the contributors have their
academic homes in Indology departments, some are housed in philosophy departments.
Regardless of their institutional affiliations each contributor brings to this volume their
knowledge of the historical context of the original arguments as well as an appreciation of
their philosophical significance. This then raises the question of the way in which the
terms “philosophy,” “epistemology,” or “metaphysics” apply to the Indian context. There is
no one word in Sanskrit or other Indian languages that would translate exactly as
“philosophy.” However, when reading certain Indian texts it is clear that they address
questions such as “What is knowledge?,” What is good?,” or “What is the self?,” all of
which are questions that we would classify as “philosophical” if they were raised in the
context of Western texts. In fact, Sanskrit has terms such ānvīkṣikī (logic or rational
inquiry), darśana (view or doctrine; the so-called Indian philosophical schools mentioned
below are each referred to as a darśana), pramāṇa-śāstra (systematic treatise on the
sources of knowledge, i.e., epistemology), or padārtha-śāstra (systematic treatise of the
fundamental categories, i.e., metaphysics). This is the reason why we can talk about
“Indian philosophy” and even about “Indian epistemology” or “Indian metaphysics.”
For readers of Western philosophy many of the concerns of the Indian authors will be
familiar and some will seem alien. The contributors to this volume aim to draw
connections between their sources and the concerns of Western philosophers where
possible. However, the main aim of these chapters is not to draw a connection between
Indian texts and Western philosophy. Rather, it is first and foremost to provide readers
with an idea of the concerns of Indian authors. Hopefully, this brings out an appreciation
for the intellectual depth of the Indian traditions and shows that, while there is an overlap
between the topics of these texts and those of the Western philosophical tradition, the
Indian texts have their own context and have to be read within that context. This means
that if this volume is classified as “comparative philosophy,” as volumes of this kind often
are, then it is because it first and foremost aims to provide an overview of certain Indian
intellectual traditions and only then aims to point out possible similarities as well as
differences between these traditions and Western philosophy. So the aim is not to start
with a question from the Western philosophical context and to try to answer it by applying
Indian material. Neither is the aim to make the Indian material “relevant” to Western
philosophy. One underlying assumption of all of the contributors to this volume is that the
Indian materials are interesting in their own right and form intellectual traditions that
merit their own interpretations. Once the work on these interpretations has progressed (it
is work that cannot ever be finished), it makes sense to assess the possible intellectual
connections to Western philosophy. This volume is a contribution toward interpreting the
Indian texts and making responsible connections to discussions on Western philosophy.
For the reader who does not have a background in Indian philosophical texts, the
following overview of Indian philosophical traditions might be useful.

Indian philosophical schools

As mentioned above there is no formal distinction between what would count as a
religious text or a philosophical text in the Indian traditions. Many of the early sources,
for example, the Vedas and Upaniṣads, are regarded as sacred and also raise
philosophical issues. By the start of the first millennium ce we begin to see texts that are
written in a more technical language addressing issues that would be quite familiar to
anybody with a background in Western philosophy. However, most of these Indian texts
are of a soteriological nature. They are supposed to lead to liberation from the cycle of
rebirths, a belief which is shared by almost all Indian intellectual traditions. This means
that even a text on a very technical subject often contains at least a note explaining how
knowledge of this particular subject matter aids in the achievement of liberation.
Indian philosophical thinking is divided into a number of different schools or textual
traditions (darśanas). In late classical doxographies, these traditions are broadly divided
into those that accept the authority of the Vedas (āstika—affirmers) and those that do not
(nāstika—deniers). Among the latter, the most prominent for the purposes of this volume
are the Buddhists, Jains, and materialists (Cārvāka/Lokāyata). Among the former, we have
six traditions that are often referred to in sets of two, indicating strong doctrinal
affiliations between the members of each pair. They are the Sāṃkhya-Yoga, Mīmā ṃsa-
Vedānta, and Nyāya-Vaiśeṣika traditions. In addition to these traditions there are a few
others that are counted among the āstika, for example, the Grammarians (Vyākaraṇa) and
Śaivism aspects of which are discussed in Chapters 9 and 10 of this volume.
These textual traditions developed over a period of more than one thousand years with
many starting around the beginning of the first century CE, although some texts can be
dated hundreds of years before that. In many cases the development of these texts takes
the form of a short foundational text (sūtra) with later texts as commentaries on the
foundational text or on previous commentaries, written by subsequent writers. In some
instances, especially in the case of early texts, we do not know who the authors were,
even though there might be a particular name which is associated with the authorship.
For example, traditionally, the foundational text of the Nyāya school has been ascribed to
Gautama Akṣapāda. However, it is much more likely that this text was written by a
number of authors after a long period of oral transmission. In other cases we know the
names but the dates might be uncertain. In most cases, however, the authorship and dates
are well established.
Knowledge of a number of languages is required in order to study the Indian material.
Most of the texts were written in classical Sanskrit, a development of the vedic Sanskrit of
the Vedas and Upaniṣads. Some of the early Buddhist texts were written in Pāli. As
Buddhism spread throughout Asia, Sanskrit and Pāli texts would be translated into other
Asian languages, for example, Tibetan, classical Chinese, and Japanese. Over the
centuries some of the Sanskrit and Pāli originals were lost and continued to exist only in
Tibetan or Chinese translations. In some cases they were then retranslated into Sanskrit
or Pāli. Some contributors to this volume use retranslations into Sanskrit. Whenever such
texts are cited, the titles are prefixed by an asterisk. Many of the Indian texts have been
translated at least partially into English, but there are still many that have not been
translated into any European language. Even if they have, the translations are often quite
old and outdated. In some cases new manuscripts are discovered which require a
reinterpretation of the text. This means that a large part of the required work for
interpreting these texts is philological, which adds a certain amount of difficulty. Several
attempts to compile definitive editions and translations (in so far as such a thing is
possible) are under way but it is a task that requires more than one lifetime.
Rather than providing a general overview of each of the schools, which can be found in
many introductory works to Indian philosophy, I would like to provide a brief introduction
to the chapters of this volume.

Outline of this volume

This volume provides readers with chapters by some of the most important scholars in the
field. Some of these scholars are very well established and some of them are only at the
beginning of their careers. All of the contributors were given free choice in the selection
of their topics. This means that there is not a “natural” classification for these chapters.
However, as is evident from the table of contents, they are organized into several themes.
The first theme has the broad title “Knowledge, Language, and Logic.” Chapter 1 by
Claus Oetke provides an extensive overview of the similarities and differences between
the views of the Prācina Nyāya (old Nyāya) school and the seventh-century Buddhist
philosophers Candrakīrti and Dharmakīrti. The term “Prācina Nyāya” refers to a division
within the Nyāya school between the old and new (Navya) Nyāya schools. While this is a
temporal division that occurred probably in the early fourteenth century beginning with
the Nyāya thinker Gaṅgeśa1 more importantly it is a methodological division in that
Gaṅgeśa began to use a very technical vocabulary that is sometimes compared to modern
technical philosophical jargon. Chapter 1 provides us with a very clear overview of some
of the epistemological and metaphysical issues discussed in India and thereby shows that
the application of these Greek terms is indeed applicable to the Indian material. In
addition, Oetke’s chapter highlights a very important aspect of Indian thinking: rational
inquiry. He provides us with an account of disagreements between different traditions of
Indian thinking (in this case particular Nyāya and Buddhist authors). This is one of the key
features of Indian thought and a major point of comparison with Western philosophy.
Indian thought is based on rigorous debate in the same way that Western philosophy is.
Most of these debates occurred between philosophers of different philosophical schools
but, in a more subtle way, also between philosophers of the same school.
In Chapter 2, Stephen Phillips discusses the notion of inference (anumāna) in early
Navya Nyāya sources. According to the Nyāya school, inference is one of four sources or
instruments of knowledge (pramāṇa). The others are perception (pratyakṣa), comparison
(upamāna), and verbal testimony (śabda). In discussions of inference as a source of
knowledge it becomes important to distinguish between reliable and unreliable
inferences. Only reliable inferences can lead to knowledge and therefore count as
instruments of knowledge. So Nyāya texts accept a certain model of a reliable inference
and they also provide examples of unreliable inferences or fallacies. The aim of this is
twofold: first, there is an epistemological argument about the production of knowledge
and second discussions of inference serve to provide a manual on how to win debates.
The third chapter by Eli Franco gives a fascinating account of a largely under-
researched topic in Indian philosophy, namely, skepticism. In particular, Franco focuses on
the ninth-century philosopher Jayarāśi who was affiliated with the Lokāyata materialist
school. Franco explains and evaluates Jayarāśi’s arguments against the idea, defended by
other philosophical schools, for example, the Nyāya and various Buddhist authors, that
there are instruments of knowledge (pramāṇas). One of the main features of Franco’s
chapter is that he interprets Jayarāśi first and foremost as a skeptic and not as a
materialist who also holds skeptical views.
In the final chapter of the first section, Piotr Balcerowicz shows how the Jains argued
for their “theory of viewpoints” which is the idea that any true description of a state of
affairs is uttered from a particular viewpoint. He makes the important connection
between this logical theory and the idea that our language is always imprecise because it
relies on assumptions that speakers of a language gather from the context and which are
not explicitly stated. Language then is always ambiguous and the Jain theory of the “four
standpoints” is an attempt to disambiguate this language.
The chapters in the second section, entitled “Consciousness and the External World,”
all deal with different Indian responses to the problem of the relationship between the
world and what Western philosophers call “mind.” Even though it is possible that the
etymology for the term “mind” is derived from the Sanskrit word manas, it is very difficult
to find an equivalent of the term “mind” in Sanskrit or other Indian languages. One of the
most important differences between “mind” and “manas” is that, according to many
Indian traditions, the latter is physical and acts as a switch between the different external
sense organs and it also acts as the “inner sense organ” by which we perceive our own
cognitions. So, the term cannot be used to describe, for example, the notion of a Cartesian
mind which accounts for the immaterial, first-person core of an individual.
In the first chapter of this section, Chapter 5, Birgit Kellner discusses the view of the
Buddhist philosopher Dharmakīrti that external objects do not exist and in fact are mere
cognitions. She explores some of the parallels between this view and the Western view of
idealism in its various forms. Most importantly she draws connections between
Dharmakīrti’s views on cognition and inference. The aim is to gain an understanding of
how the arguments for the view that external objects do not exist fit into Dharmakīrti’s
larger view of inference.
Pradeep Gokhale discusses Indian materialism (Lokāyata), a school of thought which is
often neglected in the study of the history of Indian philosophy, in Chapter 6. This is a
particularly important task because most of the original sources of this school have been
lost and what we know is often gleaned from quotations and summaries by philosophical
opponents of materialism. Naturally, we do not know how accurate these second-hand
accounts are. Gokhale provides a careful reconstruction of some of the main materialist
arguments and counters some of the most common objections to these arguments.
In Chapter 7 Ferenc Ruzsa provides a fascinating interpretation of the dualism of the
Sāṃkhya school of philosophy. According to most interpretations, dualism is the view that
there are two irreducible substances, such as the mental and the physical. In Sāṃkhya,
dualism is expressed as the division between puruṣa (person, soul) and prakṛti (nature).
As humans then we are a mix of these two and one of the main aims of Sāṃkhya is their
complete separation. Ruzsa argues that while Sāṃkhya philosophers were certainly
dualists, puruṣa and prakṛti do not qualify as substances in the traditional sense. In fact,
what we think of as a substance is nothing over and above its changing qualities,
according to Ruzsa’s interpretation of Sāṃkhya sources. Substances then reduce to their
qualities, which is a view that we also find in Buddhist teachings.
The last two chapters of this section, by Raffaele Torella and Isabelle Ratié,
respectively, provide an insight into a branch of Śaiva philosophy which is nondualistic. It
stands in contrast with the Tantric Śaiva tradition of Kashmir which is dualist in nature.
The term “Śaiva” refers to the worship of the god Śiva. The nondualistic element of this
philosophy is expressed as the identity of the individual soul and Śiva. Torella’s discussion
focuses on the relation between cognitions and the objects of our cognitions. He relates
the nondualistic Śaiva view to epistemological discussions in Buddhist philosophy about
the nature of cognition and its relationship to the external world.
Ratié’s chapter is a wonderful discussion of the ontological roles of mirror images in
Śaiva thought. These deserve special attention because consciousness is often compared
to a mirror in Indian philosophy and so the question is what the ontological status of these
reflections is. In particular, Ratié provides a fascinating insight into the different uses of
the analogy of consciousness, not just in Śaiva nondualism but in Indian philosophy in
general.
The short section entitled “Universals and Momentary Existence” features discussions
of metaphysics in Mīmāṃsā philosophy and the thought of the Buddhist philosopher
Ratnakīrti. In the first chapter of this section, John Taber discusses the interesting
question why those Indian philosophers who argued for the existence of universals also
thought that these universals are perceptible. This discussion provides an interesting
contrast with philosophical arguments in the Western tradition where those who argued
for the existence of universals usually claimed that they are imperceptible except through
their instantiations.
The second chapter in this section by Joel Feldman provides an overview of some very
influential arguments of the Buddhist philosopher Ratnakīrti in favor of the doctrine of
momentariness (kṣaṇika-vāda) of all existence. According to this view, there are no
temporally persisting entities. So not only are there no universals, but not even
particulars that endure through time.
 The next section entitled “Self, No-Self, and Self-Knowledge” covers one of the central
philosophical topics in Indian philosophy: the nature of the self. In Chapter 12, Alex
Watson discusses the main arguments of several schools of Indian philosophy and he
clarifies their positions on the existence and nature of the self. The positions range from
the idea that the individual self is an unchanging substance to the view, held by some
Buddhist thinkers, that the self is nothing but a series of momentary existences.
In Chapter 13 Roy Tzohar introduces an argument from the writings of the Buddhist
philosopher Vasubandhu that relates closely to some of the discussions in the second
section of this volume about idealism or the “mind-only” view. Vasubandhu is an idealist
and he also argues that the self is reducible to a series of momentary states. One difficulty
for any idealist is to account for the phenomenon that different people seem to have the
same, or at least similar, experiences (“intersubjective agreement”). A realist, that is,
someone who believes in the existence of a mind-independent external world which we
perceive, could make the argument that the fact that we often agree with others on the
nature of our experiences is evidence for the existence of an external world which causes
us to have these uniform experiences. So the reason why you and I agree that there is a
book of a certain size and shape in front of us is that there really is a book that has this
size and shape. According to Vasubandhu, intersubjective agreement is not an argument
for a mind-independent external world. In order to make this claim, however, we have to
question our idea of what it means to be a self and, especially, what it means to be “other.”
The next chapter by Matthew MacKenzie addresses the concept of luminosity
(prakāśatā) in Indian philosophy. This concept describes the subjective nature of our
experiences. Indian thinkers disagreed on whether consciousness has self-illumination
(svaprakāśa) or other-illumination (paraprakāśa). Self-illumination refers to the idea that
consciousness is reflexive, which means that it presents itself as subjectivity without an
object. Other-illumination, by contrast, is the idea that consciousness is presented
through the presentation of the objects of consciousness. For those familiar with the
history of modern Western philosophy, the example of the contrast between Descartes’
and Hume’s views might be helpful. Descartes argues that the self can be experienced as
self without any particular thought, whereas Hume argued famously that he can never
detect such a self. Instead, introspection can only ever reveal different thoughts
(“perceptions” in Hume’s terminology). So according to Descartes consciousness would
have self-illumination whereas according to Hume it would have other-illumination. Of
course, neither Descartes nor Hume use these terms and my comparison neglects a whole
range of differences between these thinkers. Nevertheless, as MacKenzie mentions in the
conclusion to his chapter, the Indian debate has many points of contact with Western
philosophical debates about the nature of subjectivity, self-consciousness, and self-
representation in the phenomenological and analytical traditions.
In the final chapter of this section Jay Garfield discusses the thought of the twentieth-
century Indian philosopher Krishnachandra Bhattacharyya on the subject of self-
knowledge. Bhattacharyya was influenced by Western philosophy and especially by the
thought of Immanuel Kant. Garfield highlights how Bhattacharyya grapples with the
problem of self-knowledge or the knowledge of the subject of itself. He inherits this
problem from Kant but his treatment of the problem is also heavily influenced by his
training in Indian philosophical traditions and therefore makes for an interesting and
fruitful episode in cross-cultural philosophy.
The final section entitled “Concepts and Cognitions” contains two chapters. The first of
these by Keya Maitra provides an account of the views of the Nyāya school on concepts or
“distinguishing features.” In particular she relates it to the important distinction between
perception without concepts (nirvikalpaka) and perception with concepts (savikalpaka).
Maitra connects the arguments with a Western account of concepts by Ruth Millikan.
In the second chapter in this section, Zhihua Yao discusses Vasubandhu’s arguments for
the cognition of nonexistent objects. Of particular importance here is the problem of
having cognitions about the past, that is, cognitions of things which do not exist anymore,
and the future, that is, cognitions of things that do not yet exist. Yao highlights
Vasubandhu’s views on the ontological status of the past and the future and thereby
provides an insight into an important debate among thinkers of different Buddhist
schools.
As mentioned at the beginning of this introduction, the chapters in this volume provide
an overview of some of the most exciting recent scholarship in Indian philosophy. The
following bibliography lists a few important background works as well as some selected
suggestions for anybody interested in deepening their acquaintance with many of the
issues discussed in this volume.

Note
1. There is some disagreement about his dates. They range from the twelfth to the early fourteenth
century.

Bibliography
Dasgupta, S. (1922), A History of Indian Philosophy, 5 vols. Delhi: Motilal Banarsidass. (A slightly
outdated but still useful overview of Indian philosophical schools.)
Ganeri, J. (2001), Philosophy in Classical India: The Proper Work of Reason. London: Routledge. (A
more recent introduction connecting Indian philosophy and analytical philosophy.)
Halbfass, W. (1992), On Being and What There Is. New York: SUNY Press. (A classic study of the
ontology of the Vaiśeṣika school of Indian philosophy.)
Matilal, B. K. (1986), Perception: An Essay on Classical Indian Theories of Knowledge. Oxford:
Clarendon. (An influential book connecting Indian philosophy and Western analytical philosophy.)
Mohanty, J. N. (1992), Reason and Tradition in Indian Thought. Oxford: Oxford University Press. (A
discussion of different aspects of Indian philosophy by an author who is also trained in
phenomenology and analytical philosophy.)
Patil, P. (2009), Against a Hindu God: Buddhist Philosophy of Religion in India. New York: Columbia
University Press. (This book contains a splendid discussion of the different methodologies of studying
Indian philosophy.)
Phillips, S. H. (1996), Classical Indian Metaphysics. Chicago: Open Court. (A sophisticated discussion
of selected themes in Indian metaphysics with some translations of key passages.)
Phillips, S. H. (2012), Epistemology in Classical India: The Knowledge Sources of the Nyāya School.
New York: Routledge. (A comprehensive introduction to the epistemology of the Nyāya school of
Indian philosophy.)
Potter, K. H. (ed.) (1970–), Encyclopedia of Indian Philosophies. Delhi: Motilal Banarsidass. (This
multivolume project is an excellent resource for the study of all areas of Indian philosophy. The
volumes are organized by school and in chronological order. Information about this source can be
found at: http://faculty.washington.edu/kpotter/.)
Radhakrishnan, S., and Moore, C. A. (eds) (1957), A Sourcebook in Indian Philosophy. Princeton:
Princeton University Press. (A useful and easily available collection of translations of Indian texts.
These translations should only be used as a means of first acquaintance.)
Taber, J. (2005), A Hindu Critique of Buddhist Epistemology. New York: Routledge. (This volume
contains an excellent translation and discussion of the chapter on perception in the Ślokavārttika by
Kumārila Bhaṭṭa.)
 Part One

Knowledge, Language, and Logic

 1
Epistemology, Metaphysics, and Logic in Prācīna Nyāya and
Buddhist Philosophy
Claus Oetke

I
With regard to the tradition of philosophical thought demarcated by the expressions
“Prācīna Nyāya” and “Buddhist philosophy” the words “epistemology,” “metaphysics,” and
“logic” designate spheres exhibiting specific and significant affinities and interrelations.
Those areas are themselves connected in particular ways with subject matters which
could be characterized by the terms “theory of argumentation” as well as “methodology of
cognition.” Some of the connections are neither self-evident nor trivial. Thus a more exact
account of the relations, explanations for the existence of the described situation, and an
elucidation of the extent to which the terms “logic” and “epistemology” are applicable in
the pertinent field constitute major topics of this chapter.
A remarkable phenomenon lies in the fact that with respect to the traditions of “Prācīna
Nyāya” and “Buddhist philosophy” it is quite often difficult to differentiate between
“logic” and “epistemology” due to the circumstance that individual statements or
doctrines do not permit unambiguous assignments to these designations. Insofar as those
difficulties cannot be exclusively reduced to uncertainties of interpretation they pose the
question of the legitimateness of employing the terms “logic” or “epistemology” in this
area. However, from a theoretical point of view it is possible to discern a basis for the idea
of a subject matter exhibiting affinities both to epistemology and to logic in some
legitimate sense of the words. One can easily recognize that a significant connection
exists between the category of linguistic items and that of mental or psychological entities
in view of the fact that the latter one encompasses specimens pertaining to objects which
are simultaneously possible contents expressed by linguistic units. In particular
expressions such as “NN knows/believes/recognizes/ ... that p” patently suggest that items
to which certain cognitive acts are directed can also occur as something that is expressed
by linguistic objects, specifically by expressions of the form “that p” (in English). To be
sure, as far as one can see, the idea of any type of entities which can function both as
possible contents of cognitive (or other mental) acts and as possible contents of linguistic
units seems to be completely absent in Prācīna Nyāya or in Buddhist philosophy. This,
however, does not refute the hypothesis that those traditions were concerned with
phenomena which according to our understanding relate to such entities or that they
dealt with properties which, as a matter of fact, belong to objects of the above specified
kind. But evidently this theoretical possibility alone cannot justify the assumption that
anything deserving to be called “logic” or “epistemology” existed in Prācīna Nyāya,
Buddhist philosophy, or anywhere else in the domain of Indian philosophical thought.
Generally the idea that logical or epistemological properties are a matter of concern in
some philosophical tradition provides a possibility of enhancing the plausibility of
speaking of epistemology and logic regarding that tradition. Against this background the
supposition that Nyāya or Buddhist philosophies might implicitly have addressed issues
pertaining both to (propositional) contents of cognitive acts and to contents of linguistic
objects opens a prima facie promising perspective. For contents expressible by linguistic
items can be reasonably viewed as possible bearers of the property of being true and
accordingly as entities capable of exhibiting the quality of being logically true. Hence,
given that an intelligible concept of logical truth can be explicated, the occurrence of
teachings exhibiting affinities both to the fields of logic and to epistemology licensing the
acknowledgment of the existence of some variety of logic as well as of epistemology is at
least in principle conceivable. The crux is, nevertheless, that apparently this possibility
has not been realized in Nyāya or Buddhist philosophies, and it is doubtful that this
happened anywhere in the tradition of Indian thought. There is no need to contemplate
the question whether textual sources exhibit some term which could be rendered by
“logically true” or “logical truth.” Even in the absence of an expression representing a
concept of logical truth one might reasonably attribute a(n implicit) grasp of that concept
under certain circumstances. This would be the case if in certain texts one encountered,
for example, a statement which reads as follows: “If something or something else is the
case, then if the first (alternative) is not the case then the second is the case, and if the
second is not the case then the first is the case.” It would not be capricious to infer from
the occurrence of such an assertion an awareness of the principle that if a disjunction is
true then, if a certain member is not true, then (at least one of) the remaining member(s)
of the disjunction is true and that accordingly the person who made the statement
recognized a fact which could also enable him to recognize that any sentence in which “A”
and “B” in the expression “If A or B and not-A, then B” are replaced by sentences
(expressing propositions) is bound to express a truth. In this manner the way is paved for
acknowledging that there is a special variety of truth, or more precisely that there is a
special basis for the truth of sentences or their stated contents which not every linguistic
unit expressing a truth possesses—and that the same holds good for the complementary
concept of untruth or being untrue. Thus the decisive fact lies in the circumstance that, as
far as one can see, no manifestation of a recognition of some special way of being true (or
untrue) similar to the idea of logical truth occurs in the (early) tradition of Indian thought
and that, specifically, one cannot detect an awareness of the possibility that certain
contents differ from others with respect to the necessity of its truth or lack of truth.
At this point it might be pointed out that many textual sources suggest an
acknowledgment of familiar “logical” principles of reasoning, for example, by arguing that
since B must hold true if A holds true, and since B cannot be true A cannot be true either.
Although this form of reasoning is reminiscent of the so-called rule of modus tollendo
tollens the phenomenon cannot be exploited for diagnosing the possession of a concept of
logical truth. The decisive reason for the inappropriateness of this reasoning is not the
circumstance that merely following some familiar logical principle does not license the
attribution of a corresponding logical concept or even the existence of logic in some
tradition of thought. Even an explicit acknowledgment to the effect that generally “A”
cannot be the case, given (i) that “B” is not the case and (ii) that “if A then B” holds good,
does not entitle us to attribute to the concerned subject the view that any complex
sentence of the form “If A then B and not-B, then not-A” or the content expressed by it
represents a logical truth and that for certain contents truth exhibits some distinguishing
feature which is absent with respect to other contents. However, it would not be
capricious to credit a person in the envisaged situation with the view that certain
sentences or their contents exhibit a particular relational property, such that the truth of
certain pieces of content provide support or evidence or even necessitate the truth of
some other piece(s) of content.
We contend that at most the ascertainment of certain relational properties between
pieces of content could provide a basis for supposing the existence of some variety of
logic in the tradition of (earlier) Nyāya or Buddhist philosophy. This raises, however, the
problem whether the relational properties that are concerned can be legitimately viewed
as logical properties or are at least related to logical properties in a manner which
renders well-founded a classification of certain tenets or teachings as logical theorems or
doctrines. In this connection it is most crucial to sharply distinguish between the
character of the relations as they are assessed by us and the views which representatives
of the traditions of Nyāya or Buddhist thought might have held about them.

II
There is a presumably very old definition of inference (anumāna) which highlights the
significance of relation or relational properties and which says that inference is the
establishment of a rest from something perceptible on account of some connection
(sambandhād ekasmāt pratyakṣāc cheṣasiddhir anumānam)—or the establishment of a
rest from something perceptible which is connected, if the variant reading sambaddhād
instead of sambandhād were hypothesized. A textual source known under the designation
Ṣaṣṭitantra (commonly regarded as a work of a Sāṃkhya teacher called Vārṣagaṇya or
Vṛṣagaṇa) provides even a specification of the pertinent type of relation by enumerating
seven varieties, which read: (1) connection of master and (his) belongings
(svasvāmibhāva), (2) connection of original substance and (its) transformation
(prakṛtivikārabhāva), (3) connection of cause and effect (kāryakāraṇabhāva), (4)
connection of producer and produced (nimittanimittakabhāva), (5) connection of whole
and part (mātrāmātrikabhāva), (6) connection of common occurrence (sahacāribhāva),
and (7) connection of annihilated/obstructed and annihilator/obstructor
(vadhyaghātakabhāva). It should be pretty obvious that none of the terms employed for
specifying the kinds of relationship exhibits any affinity to the expression “logical,” and
the hypothesis that the mentioned varieties represent various types of logical relation
does not appear probable either. Various examples provided in the text, such as the
connection between master and servant or soul and primordial matter illustrating the first
variety, the connection between potter and pot or soul and activity of primordial matter
illustrating the fourth type, and various other illustrations, support the diagnosis that a
specification of diversities of logical relations is not intended. Moreover, the formulations
of the text insinuate that the envisaged items to be related are not possible contents of
sentences or objects of cognition, anything which is capable of being true or false, but
rather objects or sorts of objects which are part of an external (physical) world. It needs
to be duly acknowledged, however, that this circumstance does by itself not refute the
supposition that relations are involved which in fact pertain to contents capable of being
true or false. This becomes manifest if one transforms the idea of, for example, inferring a
master from a servant, to the idea of inferring the existence of a master from the
existence of a servant, and this in its turn to the idea of inferring (the proposition) that a
master exists (somewhere) from the proposition that a servant exists (somewhere), and
similarly in other cases. Nevertheless, the supposition that (exclusively) logical
connections are at stake appears improbable both in view of the presented illustrations
and on account of the terms by which the relations are characterized. This diagnosis is
corroborated by other examples found in textual sources, such as inferring (prior) clouds
from the swelling of water in a river, inferring prior rain from an increase of the amount of
water and the swiftness of the current in a river, inferring that there is an imperceptible
movement of the sun on account of deviations of its location at different places at
different times, or even inferring future rain from the existence of clouds. In all such
cases it would be most natural to suppose that the link between that which is inferred and
its basis is a relation of probability and not more. However, it has been explicitly pointed
out that certain alleged cases of inferences are not compelling, and according to a remark
in the second book of the Nyāyasūtra-s (Sūtra 2.1.36 in the edition or W. Ruben [1928]
and 2.1.39 in the edition of G. Jhā [1939]) pertinent examples ought to be modified by
making the basis of inference more explicit. It remains doubtful whether the supposition
that the pertinent examples merely exhibit a lack of explicitness is really correct and
faithfully reflects the intentions of previous theoreticians. At any rate, even with respect
to the revised versions, such as inferring previous rain from the fact that the increase of
the amount of water is not confined to a particular section of a river, which excludes the
possibility that water has been dammed up at a certain place, one can ascertain that they
do not exhibit the same sort and the same degree of conclusiveness which is commonly
associated with the notion of logical implication or entailment.
Nevertheless, by clearly differentiating between our own assessment of the subject
matter and that which had been given by writers of the past, one has to concede the
possibility that the exhibition of a maximal degree of conclusiveness was always regarded
as a requirement of acceptable inferences. This would not rule out the stance of denying
doctrines of inferences the status of logical theories on account of the fact that the
conception of maximal conclusiveness entertained by subjects of a foreign tradition of
thought deviates from our conception and specifically from the concept of conclusiveness
which we regard as essential for the idea of logic. However, it needs to be admitted that
even under those circumstances the position could be maintained that significant
affinities exist between doctrines of inferential reasoning, as they have been developed in
the traditions of Nyāya or Buddhist philosophy, and theories of logic in a more genuine
sense. In this connection it is appropriate to draw the attention to another aspect: In the
previously mentioned list of varieties of “connection” specified in the Ṣaṣṭitantra each
type is illustrated first by an example taken from the world of everyday experience and
after this by an example taken from metaphysics, exemplifying the establishment of
metaphysical tenets advocated in the Sāṃkhya school. It follows from this that the
envisaged connections must have been regarded as exhibiting a certain degree of “topic
neutrality” by being applicable to subject matters of common sense as well as to issues of
metaphysical or “scientific” investigation. This feature permeates theories of inference as
a whole and is also endorsed by the circumstance that the derivation of the occurrence of
fire from the occurrence of smoke on the one hand and the establishment of the
impermanence of sound from its being produced on the other hand constitute the most
prominent examples cited in textual sources. It is not eccentric to hypothesize as an
underlying idea that metaphysical tenets can be established by principles whose
soundness is vindicated by common experience and to suppose that the aspiration to
demonstrate the truth of metaphysical tenets was a decisive factor for the elaboration of
doctrines of inference, in the sense of the technical term “anumāna.”
Leaving open for the time being whether teachings of anumāna can be properly
classified as “logical,” one could at least suppose the existence of affinities on account of
the following three characteristics: (1) There are doctrinal tenets of inference which can
be interpreted as pertaining to relations between (propositional) contents expressible by
linguistic units. (2) Identifications of maximal degrees of conclusiveness constitute a
relevant aim. (3) Pertinent connections are specified in fairly “topic neutral” terms. In
spite of this the above presented account on the basis of earliest sources reveals
theoretical problems. First it is by no means certain—and in fact not even plausible—that
the connections, as they have been specified in the Ṣaṣṭitantra, always exhibit the feature
that the occurrence of one connected item guarantees the occurrence of the other, or,
against the backdrop of relations between contents, that the truth of one content follows
from the truth of another with a maximal degree of conclusiveness. Second, the list of
varieties of connection is haphazard; one cannot recognize that it represents a systematic
and exhaustive selection. Third, and this is most important, the inventory does not
disclose any unifying principle; it does not make plain whether there is some feature
which binds the varieties together and, if that were the case, that the pertinent principle
of grouping is also relevant. With respect to theoretical developments in the historical
period which is at stake here, it is almost always impossible to encounter explicit
motivations for revisions, and this holds good in particular for the origination of the
doctrine, which is widely known under the designation “trairūpya,” or the teaching of the
three characteristic of a valid reason. Nevertheless, that the above specified deficiencies
have in fact motivated the development of this later doctrine and that they were
responsible for the circumstance that caused the earlier teaching to fall into oblivion is by
no means improbable.
It seems that the existence of the canon of the “marks” or characteristics of valid
inferences associated with the term “trairūpya” has played a most decisive role for the
tenet that in the tradition of Nyāya and Buddhist philosophy a variety of logic has been
developed, which is sometimes called “Indian Logic.” The crux is, however, that the view
that the form of reasoning accounted for by the doctrine of trairūpya represents a
specimen of deductive reasoning and that accordingly the correlated doctrine deserves to
be called a logical theory rests on debatable premises. To be sure, one could reach such a
result on the basis of the assumption that the three pertinent conditions should read as
follows:

1. The object o about which an inference is concerned, exhibits the property of being H;
or alternatively: o is H (corresponding to the first condition).
2. There is something which exhibits the property of being H, which also exhibits some
property of being S; or: something is both H and S (corresponding to the second
condition).
3. Everything which does not exhibit the property of being S does not exhibit the
property of being H; or equivalently: everything which is H is (also) S (corresponding
to the third condition).

If one supposes that the above specified propositions function as premises in an

argument, one may deduce the following conclusion:

4. The object o about which an inference is concerned exhibits the property of being S;
or: o is S.

Obviously 4 cannot be false if 1, 2, and 3 are all true. But there are various reasons which
render the suggested account questionable. First, for conclusively deriving the
consequence specified by 4 the second requirement is obviously superfluous. This would
even hold good if the second condition would be taken as saying that everything that
exhibits H also exhibits S, as it has been alternatively suggested. Second, pertinent
textual sources do not depict the matter as if a conclusion should be drawn from a
multiplicity of premises in inferences which are complying with trairūpya. The
formulations intimate rather a view according to which the exhibition of a quality on the
part of a particular object or group of objects follows from the exhibition of some other
quality on the part of the concerned object or group of objects. If the matter were
transformed to a relation of consequence between propositional contents, one should thus
suppose that a consequence is drawn from a single premise. Nevertheless, even in view of
those facts it could still be maintained that doctrines of trairūpya refer in an indirect
manner to a relation of “deductive derivation” by stipulating that a group of premises
implies a consequence if the latter can be deductively derived from the former, that is, if it
is logically entailed by it, and if, possibly, additional conditions are fulfilled. Thus one
might think that the elaboration of the doctrine of trairūpya is based on some imperfect
grasp of some variety of deductive inference or logical consequence. It is because of
highlighting the doubtfulness of this diagnosis that the following circumstance possesses
primary importance: The material adequacy of the above presented account of the content
of the conditions of trairūpya is highly questionable. This holds true in particular for the
second and third conditions. According to an alternative explication the second
requirement would be adequately represented if the expression “something” were
replaced by “something which differs from o” or “something else” and if in the
formulation of the third condition “everything which” were replaced by “everything which
differs from o and which.” Although the deviances of formulation appear minor they
embody a most significant difference: According to the modified version it could not be
maintained that satisfaction of the three conditions, taken together, guarantees or
“logically entails” the pertinent conclusion. Instead one might regard the canon of the
three characteristics of a valid inference as a “methodological principle” specifying a
condition under which it is (or should be) legitimate to extrapolate a regularity that holds
good in some realm to a different domain, specifically the domain of objects which are
under consideration in some specific situation of inference or argumentation. The
questionable nature of the initially presented explication is additionally enhanced by the
fact that it mismatches certain formulations to be found in textual sources. In the
Praśastapādabhāṣya (also known under the designation Padārthadharma-saṃgraha), a
“classical” treatise of the Vaiśeṣika school, a passage occurs in which the conditions two
and three are presented by saying that it should be known regarding (the quality
functioning as) a valid indicator that it occurs together with the quality that is to be
established in some or all instances different from the object(s) which is (are) to be
investigated in a particular situation of inference (or argument) (anumeyadharmānvite
cānyatra sarvasminn ekadeśe vā prasiddham) and that in accordance with means of valid
knowledge a valid indicator must never occur in the domain which is complementary to
(that of the occurrence of) the quality to be established (anumeyaviparīte ca sarvasmin
pramāṇato ‘sad eva).1 The crucial feature lies in the occurrence of the expressions
“known” as well as “in accordance with means of valid knowledge.” Despite the fact that
various textual sources exhibit variations and lack of consistency as far as formulations of
the three conditions are concerned one cannot dismiss the occurrence of “epistemic
terms” as irrelevant. If one looks a bit closer in the matter, one can recognize that the
import of the wording is by no means definite and that the formulations permit various
nonequivalent explications which cannot be considered here. The important point is that
one cannot simply suppose that the conditions of trairūpya exclusively relate to relations
of truth between different pieces of content and that presumably some authors were
disposed to consider as a requirement for soundness of inferences or acceptability of
arguments matters pertaining to states of knowledge of persons—though the significance
of this difference appears to be disregarded not only by authors of the past but also in
modern research.
If the epistemic versions of trairūpya can be correctly explicated in terms of knowledge,
one could still maintain that the canon of trairūpya entails a specification of relations
which pertain to matters of truth. The doctrine of trairūpya is not the only instance which
intimates that validity of inferences or acceptability of arguments depends on epistemic
requirements encompassing truth-relations as ingredients. A fairly ancient definition of
the term “anumāna” is presented in a treatise ascribed to the famous Buddhist dogmatist
Vasubandhu, known under the title Vādavidhi. According to this explanation inference—or,
presumably, more correctly: the basis of inference—is the observation of an object not
occurring without (the item which is to be inferred) for someone who knows that
(nāntarīyakārthadarśanam tadvido ‘numānam).2 Probably the thought which the writer
intended to communicate is that a sound inference is distinguished by the circumstance
that (1) the inferring person knows by perception that a particular object exemplifies a
particular quality H; (2) all objects, apart from the pertinent one, which exhibit H, also
exhibit a quality S; (3) the inferring subject knows that (2) is the case; (4) the person
draws the conclusion that the pertinent object exhibits S. (Here and elsewhere the letter
“H” and “S” should be reminiscent of the technical terms hetu and sādhya respectively,
which could be alternatively rendered by the words probans and probandum.) To be sure,
if one disregards the component “apart from the pertinent one” in (2), one might distil
from this elucidation a relation between a universal proposition and its instantiation and a
corresponding connection of “logical entailment,” even if this does not exhaust the
stipulated requirements. The crux is that, notwithstanding the circumstance that the
disregard harmonizes with the linguistic sense of the formulation, it remains—in view of
the occurrence of the component “who knows that” (tadvid-)—doubtful whether it
complies with the communicative intentions envisaged by the writer of the text. Given
that logical relations pertaining to possible contents of linguistic units relate to matters of
their truth or lack of truth, one is entitled to assert that classifying teachings on
inferential reasoning of the sort considered here as logical theories captures at best a
partial aspect of the matter.
The supposition that reflections on inferential reasoning corresponding to the technical
term “anumāna” were essentially motivated by metaphysical concerns is suited to explain
why in the traditions of Nyāya or Buddhist philosophies issues of truth-relations
themselves did not lie in the focus of interest and came into the picture at most in an
indirect way. This circumstance can also explain why from a logical point of view the
results offered by those traditions of thought could appear rather meager. This holds good
both under the aspects of types of sentences or sentence-contents and under the aspect of
forms of reasoning which are taken into consideration. The doctrine(s) of anumāna which
were current during the times when Buddhist philosophy flourished in India are tailored
to the establishment of propositions expressible by subject-predicate sentences of the
form “a is F” and to certain varieties of existential propositions inasmuch as they can be
intuitively transformed to contents expressible by subject-predicate sentences. This
imposes a considerable restriction of scope as far as possible conclusions are concerned.
Thus complex propositions and sentence-forms like “There is either something which is F
or there is something which is G” as well as sentences exhibiting embedding, in particular
modal or deontic statements of the type “It is possible that p,” “It ought be the case that
p” fall beyond the range. However it is true that in connection with discussions of the
“means of cognition” (pramāṇa) called arthāpatti examples are presented which refer to a
form of reasoning where a conclusion is derived from a multiplicity of premises, for
example, in the derivation of the proposition that a certain person, say Devadatta, exists
at a certain time outside his own house, given the information that (1) the person is living
at the pertinent time and (2) he cannot be found in his house.3 If one makes the implicit
premises explicit one can distil a reasoning of the form:
If A, then B or C
A
Not B
Ergo: C
However, as far as one can see, authors of pertinent textual sources have not only omitted
to explicate the formal structure of the reasoning but also missed the opportunity of
identifying in a systematic manner other structures of inference which exhibit similar
properties of conclusiveness. The hypothesis that metaphysical aspirations were an
overriding concern is apt to (at least partially) explain the diagnosed lack of interest.

III
A wish to acquire knowledge pertaining to doubtful metaphysical propositions intimates
the idea of employing uncontroversial knowledge as a means of establishing their truth by
using methods of ascertainment whose reliability is vindicated by everyday experience.
Accordingly a desire to demonstrate the truth of controversial metaphysical tenets
suggests the thought of showing that knowledge concerning the content of the pertinent
tenets can be obtained by employing a method of acquiring this knowledge on the basis of
undisputable facts which exemplifies the specified quality. A doctrine which is known
under the name “five-membered syllogism” can be regarded as a reflection of this idea. It
is true that the interpretation of the canon of the five members is a much debated issue
both as far as questions of content and questions of function are concerned. It must be
also admitted that formulations encountered in textual sources are usually too inexplicit
for making definite decisions in matters of interpretation and that the possibility exists
that different writers at different times associated deviating imports with individual
elements of the scheme of five members and envisaged divergent functions. Thus the
following presentation can merely represent one alternative reading which possesses
enough plausibility to be seriously taken into consideration.
The character of the individual members is approximately indicated by technical terms,
which are commonly rendered by: (1) Assertion (or: thesis, or: proposition), (2) reason, (3)
example (or: exemplification), (4) application, (5) conclusion. In original sources different
terms are encountered for signifying those elements, namely, pratijñā for the first, hetu or
apadeśa for the second, dṛṣṭānta or udāharaṇa for the third, upanaya or anusaṃdhāna for
the fourth, and nigamana or pratyāmnāya for the fifth member. Those members could be
connected with the following functions:

1. Recognition of the fact that the truth of some content—for example, that expressed by
“Sound is impermanent”—is not beyond doubt.
2. Identifying some content which is beyond doubt and furnishes a possible evidence for
the truth of the content in question— for example, the content expressed by “Sound is
produced.”
3. Ascertaining that in unquestionable cases the same type of evidence provides support
for the truth of a similar content in question— for example, ascertaining that with
respect to examples, such as pots, something’s being produced constitutes evidence for
its being impermanent.
4. Adopting the supposition that the situation with respect to the reliability of evidence
prevailing in the unquestionable cases holds also true with respect to the problematic
case—for example, hypothesizing that since in unquestionable cases something’s being
produced provides evidence for its being impermanent it deserves to be supposed that
the same type of evidence is reliable also in the present case.
5. Recognizing that the problematic content deserves to be regarded as true—for
example, recognizing that the proposition that sound is impermanent deserves to be
considered as true.

To be sure, various textual sources intimate that the five-membered syllogism comes into
play in the context of interpersonal argumentations. However, the description presented
above indicates that a neutral specification of the function can be given which is
applicable both to processes of obtaining knowledge by individuals and to acts of
convincing other subjects of the truth of a particular controversial content. It might be
noted in passing that the question of whether certain elements ought be eliminated
depends on particular viewpoints, but is not mandatory on every viewpoint. An elimination
of the first member could rely on the view that the occurrence of the last member
implicitly reveals the fact that the concerned content is questionable, if one assumes the
stance that presentation of evidence is only appropriate if a relevant content is open to
doubt or controversy. The fourth step could be considered as dispensable if one
presupposes that extrapolation is an essential ingredient of every act of establishing the
truth of something which could be questioned. Anyhow, the underlying idea presented
above exhibits a high degree of generality. Thus there is no compelling reason to confine
it to cases of establishing contents expressible by atomic subject-predicate statements on
the basis of contents where some attribute is ascribed to the same subject or to relate the
inherent principle of extrapolation exclusively to regularities of co-occurrence of qualities
in quality-possessors. Furthermore, in principle, it would have been possible to apply the
notion of extension of knowledge from familiar to less familiar cases or from undoubted
facts to less indubitable propositions to various cases differing with respect to the type or
the degree of persuasiveness by which conclusions are linked to pieces of evidence and to
draw explicit distinctions in this regard. The circumstance that this possibility was
obviously not realized in the traditions of earlier Nyāya and Buddhist philosophies should
not be dismissed as insignificant, particularly against the background of the fact that
different examples which are encountered in textual sources call for a more
differentiating account. Again the predominance of metaphysical concerns could furnish
an explanation: The high importance attributed to tenets, such as the existence—or
nonexistence—of a soul, the permanence—or impermanence—of sound, the existence—or
nonexistence—of a world of objects of experience existing independently of their
experience, the existence—or nonexistence—of items persisting in time, in combination
with the need to defend metaphysical claims against opponents made it rather
unattractive to acknowledge that the conclusions rest on a support that is less than
maximally compelling. This constitutes a major reason why, as far as the type and
conclusiveness of support is concerned, we must neatly distinguish between the
assessments made by advocates of the tenets themselves and our own assessment of the
matter.

IV
In one respect at least the idea which underlies the teachings usually associated with the
denomination “Indian Logic” is much broader and far more encompassing than that of
“logic” in its most usual sense. In principle the idea is apt to constitute a starting point for
developments in the areas of epistemology, theory of argument and deductive reasoning
at the same time. However, many of the ways which were paved thereby have not been
entered and potential possibilities were apparently not realized in the traditions of Nyāya
and Buddhist philosophies. To be sure, an intuition of the generality of the underlying
conception might have contributed to the contention advocated in Buddhist schools and in
Vaiśeṣika that with the exception of perception (pratyakṣa) all other alleged means of
knowledge or cognition (pramāṇa) which had been approved by others, in particular
“word” (śabda), “gesture” (ceṣṭā), “comparison” (upamāna), “presumption” or
“implication” (arthāpatti), “inclusion” (saṃbhava), “absence” (abhāva), and “tradition”
(aitihya), are, inasmuch as they possess validity, merely variants of inference (anumāna).
But the idea of deriving something more controversial from something less controversial
is definitely not inherent in the general concept of derivation or drawing a conclusion
itself. In this respect the previously delineated notion implies a narrowing of scope which
does not pertain to current concepts of deduction or logical consequence. Hence the fact
needs to be emphasized that in spite of this it would be unfair to assert sweepingly that
the traditions of Nyāya and Buddhist philosophies as a whole tacitly accepted a limited
outlook in this regard.
It seems that the merit of widening the perspective is in the first place due to two
Buddhist philosophers who flourished in the seventh century, namely, Candrakīrti and
Dharmakīrti. The revision introduced by the first one is more fundamental and more
important than the reforms initiated by the latter. But presumably in both cases
metaphysical concerns played a crucial role.
It must be admitted that with respect to Candrakīrti the proposed assessment partly
presupposes readings of crucial textual passages in his work Prasannapadā which are
debated. As a detailed investigation pertaining to questions of interpretation of individual
sentences is not apposite here, it will be taken for granted that the account is not crucially
affected by reliance on unwarranted philological premises. A decisive point is that the
following portrayal exhibits a high degree of theoretical plausibility.
It has been implicitly communicated above that a central ingredient of inferential
pieces of reasoning of the “traditional” type is dependence on a basis which ought to be
true. This requirement can be satisfied if knowledge pertaining to an ordinary world of
experience is exploited for recognizing and establishing debatable propositions,
particularly propositions pertaining to metaphysical matters. Hence the attitude of
somebody who acknowledges or advocates the truth of metaphysical claims relying on
such methodological principles would be threatened by inconsistency if he did himself not
believe that the basis on which he grounds his views or contentions possesses the quality
of being true. As a proponent of tenets of Madhyamaka teaching Candrakīrti had to
establish propositions which question the reality of the phenomenal world of experience.
Thus given that he did not intend to advocate contentions which were not considered as
true by himself, he could hardly derive relevant tenets from (allegedly) indubitable truths
in the ordinary world and avail himself of a method which appeals to the circumstance
that the reliability of the method is vindicated by commonly acknowledged experiences. It
appears after all that the writer of the Prasannapadā explicitly avows in the introductory
section of his work that proving a proposition requires (at least among others) a recourse
to a method by which the proponent could himself be convinced of its truth. In such a
situation three options are available, namely, (1) adopt an inconsistent way of behavior, (2)
refrain from the ambition of validly demonstrating the considered tenets, and (3) employ a
different method of recognizing and proving relevant metaphysical tenets. I contend that
the author, who is considered as the principal representative of the so-called Prāsaṅgika
school in Madhyamaka Buddhism, chose the third alternative.
In order to overcome the dilemma the writer of the Prasannapadā exploited the
possibility of indirect proofs. This means in a nutshell that the truth of a proposition is
established by hypothetically supposing its contradictory, deriving from this supposition
an absurdity or contradiction, deducing the untenability of the supposition itself and
concluding that only the contradictory of its content can be true. This method appeals to
principles which are reminiscent of familiar logical ideas, in particular the concepts of
reductio ad absurdum and of double negation. As a matter of fact, the establishment of
the most relevant metaphysical tenets and of the central doctrine of Madhyamaka relies
on the exploitation of additional principles, among which is the theorem that if both “A”
and “B” are not true, a disjunction of “A” and “B” cannot be true either and that if a
disjunction (of “A” and “B”) is entailed by a proposition (“C”), then this proposition (“C”) is
not true if the disjunction is not true (i.e., if both “A” and “B” are not true). In this manner
it can be derived, that in the final analysis no entity originates due to the fact that if
something originated it would (1) either exclusively originate from itself or (2) exclusively
from some different entity or entities or (3) partially from itself and partially from
something different or (4) without any cause, in combination with the circumstance that
all the alternatives (1)–(4) are untenable. The method of indirect proof, associated with
the technical term “prasaṅga,” comes specifically into play in the context of refuting the
individual alternatives. It is a remarkable fact that, as far as one can see, the underlying
principles indicated above have not been made explicit in the pertinent textual passages.
Thus there remains room for additional critical reflections on different levels, particularly
(i) with respect to the relationship between the envisaged suppositions and the alleged
absurdities or contradictions and (ii) with respect to the logical principles which are
tacitly employed. These matters are by no means irrelevant. It is true that the intuitively
assailable points concern mainly special claims to the effect that particular assumptions
entail specific incongruous consequences. But neither the presumption that consequences
which appear absurd are in fact indefensible nor the supposition that logical principles
which seem uncontroversial must possess absolute validity deserve to be taken for
granted. To be sure, toward the end of the introductory section of the Prasannapadā,
Candrakīrti articulates the claim that his method of prasaṅga-reasoning complies with
actual ordinary practice. But from a philosophical viewpoint this consideration, even if it
should be true, is definitely unsatisfactory. Obviously the author was ready to accept
intuitions about entailment, absurdity, and validity of logical principles without seeing a
reason to put them into question. This attitude is criticizable on at least two accounts:
first it is worthwhile to examine the soundness of intuitions because one could obtain
novel insights by recognizing why they are sound, if they are. Second, even commonly
shared intuitions need not be trustworthy in the final analysis. Specifically with respect to
established logical principles we are nowadays aware of the possibility that they must be
regarded as problematic or even be rejected in the light of logical reflection or
philosophical analysis. Such ideas were presumably foreign to the writer of the
Prasannapadā. Again the concern for metaphysics furnishes a partial explanation of this
indifference. In spite of the shortcomings highlighted above the founder of the
Prāsaṅgika-school deserves to be considered as a theoretical philosopher of highest
importance in the Indian tradition. His significance does not merely stem from the fact
that he was a most influential commentator of a central text of his predecessor Nāgārjuna
and that he presented a convincing account of the nature of the philosophy of
Madhyamaka. Presumably far more relevant is the circumstance that Candrakīrti clearly
discerned that metaphysical undertakings call for methods of acquisition of knowledge as
well as argumentative reasoning which fundamentally differ from the one envisaged in
classical expositions of anumāna. A central distinguishing feature of prasaṅga in contrast
to anumāna is that in the former no atomic proposition which needs to be true functions
as a basis of deriving a consequence. Only a proposition which is hypothetically assumed
for the purpose of refutation and a complex proposition representing a relation of
entailment serve as starting points in inferences and proofs of the prasaṅga-type. Thereby
a radical departure from the narrow framework of anumāna-reasoning is achieved—it
appears, though, that its importance is underestimated.
We are entitled to presume that also in Dharmakīrti’s teachings on inference and proof
metaphysical aspirations played a determining role even if this is less obvious than in the
case of Candrakīrti. A major reason for being dissatisfied with earlier accounts of
inference lies in problems of extrapolation. As a matter of fact, a difficulty connected with
inferring propositions whose truth cannot be verified by perception in principle and thus
on inferring the truth of metaphysical claims on the basis of observable facts appears to
have been discerned quite early in history. The relevant predicament is the following:
Whereas in ordinary cases, such as in the case of inferring the existence of fire from the
observation of smoke in a specific situation, a connection between the concerned items,
such as the occurrence of smoke and the occurrence of fire, can be verified by perception
in certain instances, this is impossible if pertinent “metaphysical” issues are to be settled
by inference, such as the existence of imperceptible substances like the soul or the air—
which is according to Vaiśeṣika doctrine perceptible by touch but not by vision—or other
metaphysical tenets which cannot be vindicated by ordinary experience. It was, however,
widely supposed that this circumstance does not constitute an insurmountable problem.
In certain sources a distinction is drawn between two types of inferences designated by
terms such as dṛṣṭaliṅga, pratyakṣatodṛṣṭasaṃbandha (inferences where the connection is
observed by perception) on the one hand and expressions such as adṛṣṭaliṅga,
sāmānyatodṛṣṭaliṅga, or sāmānyatodṛṣṭa on the other hand.4 The idea is that in the latter
type of cases a connection between probans and probandum, although not perceptible in
its specific nature, can be observed with respect to its general nature. Thus the (alleged)
movement of the sun being imperceptible in principle is to be inferred on the basis of
different locations at different times due to the fact that differences of positions at various
times can be seen to be connected with (visible) movement in other cases. More precisely,
the proposition that there is an invisible movement of the sun is supposed to be inferable
from its difference of position at different times on account of the fact that existence of
movement in general can be seen to be connected with differences of position. A similar
strategy is suggested for other cases concerning facts transcending ordinary experience.
However, the implicit extrapolation poses also other problems. First it raises the
question whether it is legitimate to extrapolate from observed regularities in a number of
cases to a regularity in some unknown case. This is a general problem affecting ordinary
inferences and the establishment of metaphysical tenets alike. Even with respect to the
traditional example of the conjecture that fire occurs since smoke is to be seen it is a
relevant question what should warrant the assumption that smoke can exclusively arise
from fire. The involved problem of extrapolation is still more acute if a transference of
some ordinary domain to an altogether different domain of metaphysics is at stake, but is
not confined to such cases. There are indications that Dharmakīrti was fully aware of that
problem. To be sure, one could attempt to circumvent some problems simply by
conceptual stipulations, for example, by laying down that causal history determines
identity, so that it might be declared that only items which originate from fire qualify as
smoke in a genuine sense of the term. But apart from the fact that this looks like an ad
hoc solution and merely shifts the difficulty to the issue of vindicating that something
qualifies as smoke in the genuine sense of the word, it appears pretty implausible with
respect to examples like inferring impermanence from the property of being produced.
Should one really assume that something which does not perish cannot be produced?
Furthermore, and in this regard it is difficult to ascertain whether Dharmakīrti was aware
of the pertinent fact, it does not hold true that numerical increase of confirmation
correlates with increase of probability. Obviously the probability of living for another year
does not increase in proportion with the increase of previous years of survival. In addition
to this, examples have been presented before Dharmakīrti which call into question the
reliance of the satisfaction of the demands implied in the above expounded doctrine of the
“three characteristics” (trairūpya). This holds true in particular for the phenomenon
which is associated with the term “sadvitīyaprayoga.”5 Here the crucial point is that a
probans for a disjunctive probandum meets all the criteria of certain current versions of
trairūpya but permits the derivation of arbitrary propositions with the help of intuitively
undisputable logical principles. Apart from “statistical confirmation” altogether different
factors come into play in connection with extrapolating from known to unknown cases,
which have not been accounted for in the above expounded doctrine of trairūpya or by
other theoretical ingredients of earlier teachings.
But even a deeper problem can be ascertained. Granted that a certain regularity holds
in fact true without any exception in an unlimited number of cases, one might wonder why
this is possible after all. Again it is not definitely clear whether Dharmakīrti was aware of
the pertinent question. It can be ascertained, nevertheless, that certain teachings are
objectively suited to provide an answer to this problem. If one raises the question what
could be responsible for the circumstance that smoke is always connected with fire, if one
took for granted for the sake of argument that this is really the case, then the following
answer presents itself: the connection exists because of the existence of a causal
relationship between the items. In this context it is important to realize that this
consideration cannot eliminate the fact that inferences from the occurrence of causal
effects to their causes are bound to be unsafe and that different reasons are responsible
for this. First it is always difficult, if not impossible, to eliminate the possibility that items
which seem to be connected by a causal link are in fact not connected in this manner.
Second, the notion of a causal connection does not entail absolute invariance. It is in the
final analysis not inconsistent to suppose that causal relations are subject to changes in
space and time, that, in other words, nature and its laws are not uniform and unalterable.
It seems rather improbable that Dharmakīrti’s philosophy offers satisfactory answers to
these problems. However, the question of whether the Buddhist philosopher at least
addressed those issues can be left open here.
In principle it would have been possible to frankly acknowledge that inferences make
conclusions only probable without necessitating them. Obviously this was not the stance
which Dharmakīrti adopted. Instead he solved the problem of the inadequacy of statistical
support for extrapolations by bringing into play two varieties of “essence-relations,” called
svabhāvapratibandha, connected with the expressions kāryahetu as well as tadutpatti on
the one hand and svabhāvahetu as well as tādātmya on the other. Whereas the first two
terms might be rendered by “reason/evidence [which is an] effect” and “origination from
this”—or even “causality”—respectively, the import of the third and fourth terms could be
represented by the renderings “reason which is an essence” and “being (or having) this as
self”—or “identity in reality”—respectively. It has been observed by previous scholars that
thereby Dharmakīrti’s account in some respect reverts to the one distinguished by the
enumeration of various kinds of connections as it occurs in the above considered passage
of the Ṣaṣṭitantra. Evidently the first variety of “essential connection” should account for
inferences from effects to their causes, which are beset by the uncertainties highlighted
above. Remarkably a standard example of the second type occurs which could be
considered as a prototype of an “analytic” inference, namely, deriving that something is a
tree from the fact that it is an Aśoka-tree (śiṃśapā or the tree Dalbergia Sissoo). Thus it
could be expected that Dharmakīrti satisfies the desideratum of distinguishing between
different types and degrees of rigor regarding the connection between premises and
conclusions in inferences, while presenting with respect to svabhāvahetu-s a rather
unfamiliar explanation in terms of “essences”—and the idea of an item’s being present
due to the mere existence of an item. This presumption is, however, not true. The most
important application of the idea of svabhāvahetu in the realm of philosophy is made by
Dharmakīrti himself in the context of the argument that everything must possess
momentary existence due to the (alleged) fact that only items which possess causal
efficiency can be acknowledged as existent.6 There is no need to explore the details of the
argument because the following contention is indubitable: The reasoning does not exhibit
the property that the relevant conclusion follows from the truth of acknowledged
premises with the same degree of rigor as in inferences where one derives that something
must be a tree because it is an oak and so on. As maximum of rigor is not the
distinguishing characteristic of svabhāva-inferences in general it is hard to recognize any
plausible motivation for introducing the category of svabhāvahetu-s as a distinct type at
all, except on the supposition that a main concern of the Buddhist philosopher was to
open a way for the establishment of particular metaphysical tenets, most prominently the
thesis of the momentary nature of all entities (of a physical world). Against the
background of the fact that not only the realm of all inferences licensed in Dharmakīrti’s
framework but even the special category of svabhāvahetu-s exhibits a heterogeneous
character the suspicion arises that the account of inference provided by the philosopher
might be beset by fundamental defects. Presumably those shortcomings are more serious
than the ones emphasized in certain textual sources of Brahmanical schools like Nyāya
and Vaiśeṣika, where it is claimed that the Buddhist doctrines of the tradition initiated by
Dharmakīrti are excessively restrictive disallowing certain valid inferences.7 Nonetheless,
the theory presented by Dharmakīrti possesses importance not merely due to the fact that
it copes—or attempts to cope—with certain problems which seem to have been neglected
before. It exhibits also significance on account of the circumstance that it transcends the
framework of establishing novel truths on the basis of commonly acknowledged facts,
even if the innovations are less radical than those which have been introduced by
Candrakīrti. It is obvious that certain specimens relegated to the domain of svabhāvahetu-
s exemplify a relative order with respect to the certainty of the basis and the conclusion
which runs counter to the relative order in traditional examples like the inference of fire
from smoke. Moreover, Dharmakīrti has advanced clear examples which can be credited
with the quality of exemplifying maximal conclusiveness. It is not improbable that a
profounder assessment of the nature of those examples and their distinguishing features
on the part of Dharmakīrti and his tradition was impeded by the predominance of
metaphysical aspirations.

V
Neither in ancient Nyāya, if this term is taken as encompassing the teachings of the
system of Nyāya in the narrower sense as well as related Brahmanical schools like
Vaiśeṣika and Mīmāṃsā, nor in Buddhist philosophies flourishing in India a systematic
investigation was undertaken regarding the ways in which the truth of contents
expressible by linguistic units can be related to other truths or to contingent matters. It is,
however, not wrong to contend that the manner in which knowledge is related to other
knowledge constituted an object of interest in those traditions. Since metaphysical
concerns naturally induce the question of what can be known and how one can know
something this is not surprising. Confronted with metaphysical issues it is natural to
acknowledge as relevant the question on what conditions metaphysical belief can be
counted as knowledge. Hence it deserves to be noted that tenets which do not obviously
exhibit an epistemological character can be interpreted as providing an implicit answer to
the pertinent question. This holds good with respect to the previously cited statement of
the Vādavidhi which intimates an acknowledgment of some instantiation of the following
schema: “If somebody believes (that) B (is true) and knows (by C) (that) A (is true), then
he knows (that) B (is true).” With respect to inferential knowledge it means that if
somebody (NN) knows by perception that some object exhibits a quality H and knows that
H does not occur without S, then he (NN) knows that the object exhibits the quality S, if
he (NN) believes that this is the case. Obviously from a formal viewpoint the above
specified schema could be also implemented in other ways, such as by substituting ‘by
reliable verbal communication” for “by perception,” but more radical alterations would be
equally possible. In any case, however, the pertinent theorem can be brought in
correlation with the “traditional” explication of knowledge as a true belief which rests on
appropriate grounds. It is true that, as far as one can see, neither the formal schema nor
any possible instantiation of this schema has been explicitly formulated in textual sources
of Nyāya or Buddhist philosophy. For its acknowledgment one can merely rely on indirect
evidence. But possibly precisely the disinclination of making such epistemological
theorems fully explicit has prevented a more critical reflection on the subject matter.
Since in various doctrines (before Dharmakīrti) the idea of “(invariable) co-occurrence”
between probans and probandum has been related to uncontroversial instances different
from the case in question so that inference relies on extrapolation it is questionable that
principles complying with the specified schema are objectively adequate. As, in addition
to this, numerous examples are to be found in textual sources where conclusions are
presented which follow from pertinent premises in a manner which is far from compelling,
it could be regarded as more suitable to hypothesize a variant version of the principle
which reads: “If somebody believes (that) B (is true) and knows (by C) (that) A is true and
if B is in fact true, then he knows (that) B (is true).” This reflects the stance that belief
qualifies as knowledge if (1) the believed content is in fact true and (2) certain other facts
are known. What those facts which have to be known are could be specified with respect
to individual types of knowledge in an epistemological theory. This would permit to assign
to knowledge of co-occurrence relations a relevant function of establishing knowledge
even for cases in which instances of probability-reasoning are at stake. A theory of
“means of (acquiring) knowledge” in the sense of the term “pramāṇa” could be employed
for specifying possible instances of the letter “C” in the schema. But this is in fact a
secondary issue. Although the modified principle, which is applicable in a more general
manner, appears to be more suitable from an objective point of view it would be a mistake
to attribute its acknowledgment to representatives of Nyāya or to Buddhist philosophies.
The decisive reason is not that such an epistemological maxim was not explicitly
recognized. Decisive is rather is the circumstance that it is far from certain that
representatives of those traditions were willing to admit that inferential knowledge and
knowledge in general might merely rely on probability. If it had been clearly
acknowledged by some writer that means of (acquisition of) knowledge are expedients for
increasing the chances of recognizing truths in the long run without guaranteeing the
attainment of truth in every single instance, it could be appropriate to attribute
acknowledgment of the modified principle as a stance which is most compatible with the
outlook. As long as such statements are not found it must be deemed probable that
representatives of ancient Nyāya and Buddhist philosophies would be disposed to regard
the first version as adequate. To be sure, we who interpret and analyse the theories which
can be reconstructed from the textual sources might deem that the “means of knowledge”
specified in Indian traditions of thought can at best perform the more modest function
described above. But this underscores the potential relevance of deviances between our
assessments of epistemological matters and those prevailing in the investigated
traditions.

Notes
1. Cf Nenninger (1992), p. 31.
2. “Die Schlußfolgerung ist das Sehen eines notwendig verbundenen Gegenstandes für jemand, der
das weiß” (Frauwallner, 1957, p. 120).
3. See, for example, the passage in the Vyomavatī which runs: “āgamapūrvikā tu jīvati devadatte
gṛhe nāstīti vākyād gṛhe ‘sattvaṃ pratipadya tadanyathānupapattyā bahirbhāvapratipattiḥ”
(Sastri, 1984, p. 178).
4. Cf., among others, Randle (1930/1976, pp. 148 ff.).
5. The standard example is an argument saying that a pot must be such that it differs from some
(existing) entity which is either the pot itself or some man who is nothing but a body possessing
characteristics of consciousness because it is not a lotus, like, for example, a wall.
6. The assumption is that mere existence of an item entails its momentariness without dependence
on any additional contingent factors.
7. Cf., for example, Nyāyakandalī, p. 498 (Śarmā & Chattopadhyaya, 1963) and Vyomavatī, p. 157
(Sastri, 1984).

Bibliography
Dunne, J. D. (2004), Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom Publications.
Frauwallner, E. (1957), “Vasubandhu’s Vādavidhiḥ.” Wiener Zeitschrift für die Kunde Süd- und
Ostasiens, 1, 104–146.
Jhā, G. (ed.) (1939), Nyāyasūtra of Gautama: A System of Indian Logic. Poona: Poona Oriental Series
No 58.
Nenninger, C. (1992), Aus gutem Grund: Praśastapādas anumāna-Lehre und die drei Bedingungen des
logischen Grundes. Reinbek: Dr. Inge Wezler Verlag für Orientalistische Fachpublikationen.
Oetke, C. (1996), “Ancient Indian Logic as a Theory of Non-monotonic Reasoning.” Journal of Indian
Philosophy, 24, 447–539.
Randle, H. N. (1930/1976), Indian Logic in the Early Schools. A Study of the Nyāyadarśana in Its
Relation to the Early Logic of Other Schools. Oxford/New Delhi: Oxford University Press.
Ruben, W. (1928), Die Nyāyasūtras. Text, Übersetzung, Erläuterung und Glossar. Leipzig:
Abhandlungen für die Kunde des Morgenlandes, XVIII. Band, No 2.
Śarmā, D., and Chattopadhyaya, K. (eds.) (1963), Praśastapādabhāṣya (Padārthadharmasaṅgraha) of
Praśastapāda with the Commentary Nyāyakandalī of Dharabaṭṭa. Gaṅgānātha-Jhā-Granthamālā, vol.
1. Varanasi: Varanasya Sanskrit Vishvavidyalya.
Sastri, G. (ed.) (1984), Vyomavatī of Vyomaśivācārya. Varanasi: Sampurnanand Sanskrit
Vishvavidyalaya.
Stcherbatsky, Th. (1962), Buddhist Logic. New York: Dover Publications.
 2
Fallacies and Defeaters in Early Navya Nyāya
Stephen H. Phillips

The Nyāya school of classical Indian philosophy boasts a literature that is continuous in
the sense of building on itself from an Ur-text, Gautama Akṣapāda’s Nyāya-sūtra (c. 200
CE), on until practically the modern period. A series of commentaries each elucidating not
only the original sūtras but also earlier commentaries—by Vātsyāyana (c. 400),
Uddyotakara (c. 600), Vācaspati (c. 950), and Udayana (c. 1000)—dominates the school’s
early period. But also other authors—a few before Udayana and many afterward—
continue to defend and elaborate a siddhānta, a “settled” worldview, into the so-called
Navya (“New”) era. Gaṅgeśa (c. 1300), for example, draws on both NyS commentaries
and non-commentarial literature while speaking for the school. Nyāya authors are also
generally aware of developments in other schools, claiming, for example, to correct the
reasoning of opponents and reestablish positions that have come under attack (or
counterattack) from Mīmāṃsakas and Buddhists in particular.1 There are innovations, to
be sure, and a few changes in outlook although more commonly only elaboration or
expansion of argument. The overall flavor of the school is conservative in the sense of a
felt call to defend the views of the NyS and not to jettison them except rarely.
Furthermore, although there are tremendous doctrinal divides between Buddhist
logicians and Naiyāyikas, there is not much difference when it comes to inference as a
knowledge source and fallacies.2
So what of the division between Old and New Nyāya, prācīna and navya? This
intriguing distinction does not mean quite the same as employed by classical authors as it
does among modern scholars. Let me explain and thereby set the focus for the
philosophically substantive part of this contribution that follows.
Wada (2000, pp. 442–450) argues that Navya Nyāya begins with Udayana in that
Udayana formulates the positions and techniques of analysis most distinctive of later
Nyāya philosophy. The judgments of more than a dozen scholars are cited either in
support of this view or against the opinion attributed to me, among others, that it is rather
Gaṅgeśa who should be awarded the distinction of being the original Navya Naiyāyika.
What Wada says is almost all correct, but he overlooks the indexical use of “navya” among
later authors that is much like how “new” is used in English. For Gaṅgeśa, Udayana
belongs to Old Nyāya because, relative to himself and contemporaries and near
predecessors such as Maṇikaṇṭha, Udayana was recognized as a comparatively distant
figure, not that he belonged to an ancien régime, that he was prerevolutionary in some
sort of comprehensive way as opposed to details of analysis. For later authors such as
Mathurānātha, sometimes Gaṅgeśa himself is not counted as navya but only Raghunātha
and his sixteenth- and seventeenth-century followers (Ingalls, 1951, pp. 126–127).
Of course, to be new in the sense that a philosophy would be new is not only to be later
than something similar but also to constitute a significant change. Here the division
between Old and New Nyāya becomes problematic, since, to pick Udayana’s most
immediate big-name predecessor, Vācaspati, the tenth-century Naiyāyika, is responsible
for more than thirty significant advances relative to Uddyotakara (Ranganath, 1999, pp.
7–78) and Uddyotakara even more relative to Vātsyāyana (I dare say). Udayana’s
recognition of Vaiśeṣika ontology (and drawing out the importance of the category of
absence) along with his ontological treatises may well be the most dramatic development
in Nyāya’s long and rich history. But Uddyotakara also recognizes a Vaiśeṣika position as
correct on several matters, doubt and universals (NyV 1.1.23 and 2.2.66), to name just
two.
Now a most telling change is Gaṅgeśa’s Tattva-cintā-maṇi “(Wish-Fulfilling) Jewel of
Reflection on the Truth (about Epistemology)” (henceforth Jewel) spawning its own
commentarial literature which comes to dwarf the core quartet of NyS commentaries.
According to the judgment of Dinesh Bhattacharya (1958, p. 109): “The work of Gaṅgeśa
became highly popular very soon and was studied and commented upon in various centres
of culture of India. It not only cast the works of the old school of logic into oblivion but the
neo-logical works of his predecessors also faded into insignificance and gradually were
forgotten due to its overwhelming popularity and all-embracing character.” Nevertheless,
for sheer originality, it may well be Raghunātha (c. 1500) who takes the prize within
Nyāya’s long run. He does so definitely in the opinion of Naiyāyikas in Bengal, that is,
Navadvīpa (Ganeri, 2011, p. 45 and elsewhere). Indeed, the works of Raghunātha are of
such complexity and length—while presupposing familiarity with Gaṅgeśa’s Jewel in
particular—that I do not propose to include them in this study. Hence the qualifier early in
my title. This chapter is concerned with the views of Gaṅgeśa, his contemporaries, and
near predecessors including the handling of NyS positions. Raghunātha writes at length
about the fallacies discussed below, but his views will not be aired mainly because he has
so much to say (and much of importance) but also because Raghunātha scholarship is
comparatively little advanced on our topic of fallacies.3 My aim is to present an overview
of fallacies and defeaters in Nyāya through the early period of Navya Nyāya, focusing in
particular on Gaṅgeśa’s Jewel.

Inference as a knowledge source

The concept of the inferential fallacy is derivative from that of inference, anumāna, which
produces new knowledge, new inferential knowledge, anumiti, according to Naiyāyikas
including Gaṅgeśa, from other things known. In everyday life, we need knowledge in
order to get what we want and avoid what we want to avoid (Vātsyāyana, avatāra to NyS
1.1.1). Continuously we make inferences for these ends, for example, from hearing
thunder to the presence of nearby clouds, automatically, as Gaṅgeśa points out (Tirupati I:
339; Calcutta: 427–428), without wishing to infer (contrary to Uddyotakara, NyV 1.1.5).
Often we do wish to infer, however, to find out something important. In philosophy, just
about everything may be taken to be in dispute, even things like a self that seem to be
(ap)perceptually given. Thus we make inferences in order to establish right positions. In
philosophy (vāda), we try to construct arguments that match the natural deductions that
lead to useful knowledge in everyday life.4 Hence we produce formal inferences designed
to coerce a neutral judge (madhyastha, the “one in the middle,” a usage dating all the way
back to Vātsyāyana) and indeed an opponent, to accept a conclusion, to see its truth. The
conclusion would be, to consider “inference for oneself,” sva-artha-anumāna, a new bit of
knowledge, and in “inference for others,” para-artha-anumāna, perhaps a new bit for them
but also a proposition that joins others as siddhānta, the right and certified way to
believe.5
Staal (1973), Matilal (1998), and others present the classical pattern with a variation on
the predicate-logic notation of modern logic,6 but the latter works pretty well since an
individual picked out by a Fregean name is about the same as a classical Indian dharmin,
“property-bearer,” also thought of as a location or substratum, adhikaraṇa (very broadly
interpreted). The relation binding properties and their bearers in the Indian systems—
Nyāya, Buddhist, Mīmāṃsaka, and so on—is occurrence, occurrence of a property at a
location or site. This is also figured as possession and captured in Sanskrit by a possessive
suffix or by the genitive case. A site has a property. For example, a mountain has
smokiness—which is the same as saying that the mountain is smoky. Later Naiyāyikas talk
about qualification: the mountain is qualified by smokiness. To represent an inferential
subject (pakṣa, e.g., a mountain, a) as a bare particular qualified by a prover property
(hetu, e.g., smokiness, H) entailing that it is qualified by a probandum property (sādhya,
e.g., fieriness, S)—(1) a is H, (2) all things H are things S, therefore (3) a is S—is generally
okay but fails to be sensitive to the Naiyāyika contention that we invariably experience
locations as clothed, as propertied or qualified (viśiṣṭa), and so there must be one or more
properties that specify the inferential subject (the pakṣatā, subjecthood, as specified by
being-a-mountain, for instance). The specifier (avacchedaka) of the subjecthood is usually
irrelevant: it does not matter that it is a mountain that is smoky in the stock example;
whatever it is that is smoky is fiery and we could point and say, “That is fiery, since it is
smoky.” However, for some fallacies, such as “The lake is fiery, since it is smoky,” the
specifier of the inferential subject is relevant since, for example, no lake is fiery.
Nevertheless, a simple predicate logic of names and properties and quantification is
usually adequate, and it is misleading to designate the subject as an Aristotelian “minor
term” because its character apart from its possessing the prover property (H, often
mislabeled the “middle term”) is, as I say, usually irrelevant. But a proposition such as
“The mountain is smoky,” Ha, is considered, in the fashion of Aristotle, to represent a fact
or state of affairs comprised of items, not ideas, a fact that entails another fact, “The
mountain is fiery,” Sa, in virtue of a pervasion or inclusion relationship between two
properties in extension, vyāpti: (x) (Hx → Sx). Anything that possesses, or is qualified by,
smokiness (H) possesses, or is qualified by, fieriness (S). In overview, Nyāya’s inference is
not simply a deductive pattern because how a vyāpti or generalization is known is thought
of as part of the inferential process.
It is important to appreciate that deductively valid arguments are viewed as fallacies in
Nyāya if in fact there is no pervasion or a is not H (this is called asiddhi, failure to know
the prover as a property of the inferential subject or failure to know the vyāpti), even
though it is recognized that if (counterfactually) a were H and all Hs were Ss, then a
would be S. The premises have to be known by the inferring subject in a good inference,
and only those inferences that produce knowledge (implying truth of belief) are good or
genuine as opposed to anumāna-ābhāsa, pseudo-inferences, close imitators of the real
McCoys, imitators that however slight the deviation from the genuine do not produce new
knowledge. Thus proper reconstruction of the inferential form would have each of three
lines (two premises and the conclusion) prefixed by the epistemic operator “(K),” to be
read, “A subject P knows that . . .”

1. (K) Ha (P knows that something a, e.g., a mountain, is H, e.g., smoky).

2. (K) (x) (Hx → Sx) (P knows that everything H is S; e.g., everything smoky is fiery).
3. (K) Sa, provided that parāmarśa has occurred. (P, who puts the two premises together,
knows that that thing a is S; e.g., the mountain is fiery).
The tacked-on third condition, parāmarśa, is no separate operation. “Inferential
reflection,” parāmarśa, is a matter of putting the information together from the two
premises such that, psychologically speaking, the inferential conclusion becomes the
content of a bit of occurrent knowledge. The nature of this “reflection” is a hot topic, both
within Nyāya and without. Uddyotakara makes it the central pillar of his theory, and
Gaṅgeśa devotes a long section to defending its necessity (contra Mīmāṃsakas in
particular). From the first-person point of view, the psychological process of “reflection,”
parāmarśa, is not immune to error both in the sense that it may be fed wrong information
and that one could reason incorrectly. But through that process, inferential knowledge is
produced, and epistemic warrant—or “certainty” (niścayatā)—passes from premises to
conclusion, and we act unhesitatingly, for example, to put a fire on yonder mountain out.
Things are yet more complicated, as we come now to our central topic. Inferential
knowledge is defeasible, or, more precisely stated, what a subject takes to be inferential
knowledge may turn out to be pseudo, non-genuine, a false cognition imitating a true one,
or even in Gettier-style cases an accidentally true cognition masquerading as one
genuinely inference-born. And not only are individual reasoners fallible, knowledge has a
social dimension and defeaters may come from outside. That is to say, not only would, for
instance, awareness of a counterexample be a defeater, but also if someone were to
present a counterinference to a conclusion opposed to ours, no longer would we have
inferential knowledge. Awareness of any of several kinds of “blocker” of “reflection” can
undermine the generalization or the existential premise on which reflection depends.
There are potential preventers of inferential awareness, “defeaters,” bādhaka, leading to
belief relinquishment by someone who has hitherto not noticed a counterexample or the
like or who has drawn a conclusion erroneously or who has been successfully challenged
even though his or her belief is true and arrived at properly before acquiring the new
information.
It is a platitude that inductive generalizations are defeasible, and that the conclusion of
an inductive argument is only probable, not guaranteed by the truth of the premises. A
peculiarity of the classical Indian approach is that although induction is crucial to know
the pervasion or general rule employed in an inference,7 the result is looked upon as
certain (niścita), as a bit of knowledge (pramā and niścaya), as Matilal (1998, p. 15) has
pointed out. The key bridge idea is that of innocence of belief—including inductive
correlation—not so much as “until proven guilty” but “until reasonably challenged or
checked.”8 Part of Nyāya’s response to skepticism is to insist that certainty shows itself in
a person’s acting unhesitatingly while admitting that anyone could be wrong with just
about anything. We act confidently on the basis of bits of knowledge, although just about
any cognition/belief so regarded could turn out to be false in fact.
Nyāya’s view of presumptive justification should be understood together with another
epistemological oddity, that of the infallibilism of the knowledge sources. Although we are
fallible as knowing subjects, the sources of knowledge themselves—the pramāṇa of which
inference is one—invariably generate knowledge, according to the Nyāya conception. The
sources are factive (Dasti and Phillips, 2010). We know that p if in fact our belief that p
has been produced by a genuine source. Success in action and our desire to continue to
succeed lead us to trust what we take to be pramāṇa, although, again, in just about any
given case it is possible that we be fooled, that what is operative with us in the given
instance is not a genuine knowledge-producer but a close imitator, a pramāṇa-ābhāsa.
Identification of defeaters is how we or another shows that what looks to be a genuine
pramāṇa is not.
With regard to inference, we find several paradigmatic cases of error concerning a
putative prover property, such as apparent smokiness belonging to a mountain, “smoke”
which could be found, for example, to be dust, such that the existential premise is false:
there is no smoke on the mountain (and we do not know it is fiery). Sometimes a
generalization is found to be hasty, such as that everything earthen is scratchable by iron,
a putative rule or pervasion defeated by the counterexample of a diamond (a diamond
being considered earthen but not scratchable by iron). And sometimes an opponent
presents a counterinference based on a different correlation (everything F is not-S, and a
is F, countering a conclusion that Sa). But so long as there is no defeater (bādhakam vinā)
that we are aware of, we do have knowledge so long as an inferential process (or
perceptual, etc.) has in fact been operative in generating our belief and our inferential
conclusion is true.9 Certification is a different matter (Phillips, 2012, pp. 17–32; cf.,
Ganeri, 2011, pp. 134–135), where pramāṇa become methods of knowledge self-
consciously employed as best we are able. By certification we have reflective knowledge
as opposed to first-level or unreflective knowledge where the pramāṇa work
automatically. Certification brings with it a higher barrier to doubt or challenge, as will be
elaborated below. Fallacy identification is the obverse side of certification, of the coin of
showing either that we do not know what we thought we knew or confirming that it is
indeed true.
Much attention is devoted to how pervasion or the general rule is discerned. The
answer is broadly inductive: we depend on putatively similar instances, sapakṣa, which
are known to exhibit the probandum and thus should also, at least in one case (as first
insisted upon by Diṅnāga), exhibit the prover (e.g., Hb and Sb)—and putatively dissimilar
instances, vipakṣa, which are known not to exhibit the probandum and should lack the
prover, too (e.g., ¬Sc and ¬Hc: to be read “S is absent at c” and “H is absent at c”). The
example, dṛṣṭānta, is normally a single instance of sapakṣa where the prover also occurs
(Hb and Sb), its citing thus indicating that there is inductive support. In other words, by
reference to cases that are sapakṣa—that is, known instances of the probandum—where
the prover is also found, we marshal positive evidence supporting an inference-
underpinning pervasion (anvaya, “positive correlation”). And by reference to cases that
are vipakṣa—that is, known instances of the absence of the probandum—we garner
further support by not finding the prover (vyatireka, “negative correlation”).10 Citing an
example is shorthand to indicate that such a procedure supports assertion of a vyāpti,
“pervasion” or general rule.
The inductive procedure is significant for the wider theory of justification more
generally and for the logic of an important type of defeater called the upādhi in particular.
As pointed out, each line of our representation of the embedded inferential pattern should
be prefixed by an epistemic operator, and, provided that a subject P comes “reflectively to
grasp the connection” between Ha and (x) (Hx → Sx), P has knowledge and thus
warranted cognition that Sa. But not if there is a blocker of the reflection. Being-
warranted and thus knowledge are not closed under deduction considered in abstraction
from the psychological process of “reflection,” parāmarśa. And it is of course with respect
to the requirement of pervasion-knowledge, vyāpti-jñāna, that reflection is especially
liable to be checked. For how can we be sure that the limited correlations that we have
observed or heard about are not due to some other factor, something accounting for the
prover’s and the probandum’s observed togetherness without which a putative prover
would not appear the sign of the probandum it seems? Pervasion indeed comes to be
defined as the lawful connection of two properties H and S, a connection free of any such
“additional conditions,” upādhi, such that something’s being an H guarantees its being an
S, no matter what else is true about it.

Defeat by an upādhi
The upādhi is a rather complex defeater, although to point one out in order to refute a
putative inference is a common exercise in late philosophic texts. Conversely, to defend a
bit of inferential knowledge from a charge that the putative pervasion it relies on is not in
fact upādhi-free is also a common strategy. The upādhi is a characteristic that secures the
pervasion of the prover by the probandum such that without it the prover would
“deviate”—vyabhicāra, deviation, that is, the counterexample, being the most prominent
and obvious of the defeaters, and the first on the Nyāya-sūtra’s list of hetv-ābhāsa,
“misleading prover.” The upādhi is not, strictly speaking, itself a hetv-ābhāsa but rather
inferentially undercutting counterevidence, evidence of a counterexample. Thus
knowledge and justification do not pass from premises to conclusion if we become aware
of an upādhi, which would be a blocker of “reflection” through undermining the
generalization on which such reflection depends. Thus awareness of an upādhi is a
preventer of inferential knowledge, a “defeater,” bādhaka, leading to belief
relinquishment by someone who has hitherto not noticed the undercutting condition and
who has erroneously arrived at a conclusion on the basis of a falsely induced pervasion as
preserved in memory.
According to a traditional definition which although refined by Gaṅgeśa suits most
purposes, an upādhi is a property U such that

1. U pervades the probandum, the sādhya, S (i.e., anything that is S is U, sādhya-

vyāpaka), and
2. U does not pervade the prover, the hetu, H (i.e., there is something that is H but not U,
sādhana-avyāpaka)
1. (x) (Sx → Ux)
2. (∃x) (Hx & ¬Ux)
It follows then that

3. (∃x) (Hx & ¬Sx). (There is something that has the prover without having the
probandum.)

Thus, there is “deviation” and no relation of pervasion:

4. ~(x) (Hx → Sx) (Thus it is false, for example, that fire (H) entails smoke (S), since
smoke entails wet fuel (U) and there is something, e.g., a hot ball of iron, where there
is no wet fuel but there is fire—or so it is thought, to use the stock example.)
Typically, demonstrating an upādhi does even more than block a putative inference; it
suggests why the effort may have looked good although it is flawed. The upādhi is
typically an “additional condition”: Ha would not entail Sa unless there were in addition
Ua.

5. (x) ((Hx & Ux) → Sx)

The combination of the prover (H) with the additional condition (U) would correctly
predict the probandum (S), as in the standard case of wet fuel (U) added to the otherwise
faulty prover fire (H) in what otherwise would be a faulty inference to smoke (S). Not
everything fiery is smoky; for example, there is fire (it is thought) in a hot ball of iron but
no smoke. However, it is true that wherever there is fire and wet fuel there is smoke.
Gaṅgeśa identifies upādhis that are somewhat exotic in that they are not additional
conditions that would restore a conclusion that was undercut such that (5) need not be
true while the U property is nevertheless a defeater (trans. Phillips and Tatacharya, 2002,
pp. 127–132). However, this is a minor point. In broad overview, the upādhi explains why
an original inference looked good: the upādhi was overlooked. Only with the additional
condition U occurring along with the original putative prover H is there generation of
knowledge of something as S.
 The Nyāya-sūtra’s hetv-ābhāsa among other fallacies
Three lists from the Nyāya-sūtra—of types of (1) jāti, “(fallacious) rejoinders,” (2) nigraha-
sthāna, “occasions for censure,” and (3) hetv-ābhāsa, “pseudo-provers”—along with a
fourth not in the sūtras, types of (4) tarka, “(suppositional) reasoning,” need to be
surveyed given our topic.
The first, jāti, may be dealt with briefly, since jāti are dismissed by Gaṅgeśa, saying that
they are not good responses in a debate (i.e., in vāda which is concerned with uncovering
the truth as opposed to merely winning) presumably for the reason that we know they are
tricks, philosophic conundra meant to distract attention and keep us from appreciating a
perfectly good inference (TCM Tirupati II 165; Calcutta 983). Prets (2001) argues that
many seem to be proposed as legitimate tests for the similarity on which a generalization
is based, and finds a much more positive understanding of them in NyS, chapter one, as
opposed to chapter five whose “First Daily Lesson” (of two) is devoted to their catalogue
in twenty-four varieties (and may well be an interpolation). Prets may be right about very
early Nyāya including portions of “Vātsyāyana’s” bhāṣya, but long before Navya Nyāya
(clearly with Uddyotakara) jāti come to be understood as misleading analogies and
classified as tricks of debate. For example, against an inference to Sa on the basis of Ha
and familiar things exhibiting both the probandum and the prover (Hb, Sb, etc.), the
tricky opponent argues that since Sa is dubious (Sa is the conclusion to be arrived at and
so is admittedly dubious in context), the examples putatively supporting the
generalization are also dubious for the reason—asserted by the proponent of Sa—that the
examples are like the case at issue. Philosophic work has to be done in practically every
case of a jāti such that by the end we have a better grip on the notion of relevant
similarity. Matilal (1998, pp. 73–80) discusses distinct jāti listed in a Buddhist text, the
Upāya-hṛdaya, which he shows similarly are exposed as groundless analogies or
disanalogies but whose philosophic lessons flow from Buddhist metaphysics and attitudes
about testimony which are in sharp contrast with those of Nyāya. A jāti is then a fallacy in
the very broad sense of a misleading analogy, and there is less agreement about them
across school than about the nature and members of the three other lists.
“Pseudo-provers,” hetv-ābhāsa, in contrast, are—Gaṅgeśa says in the same passage
where he disparages jāti—good responses, since they are genuine blockers of inferential
knowledge or are considerations leading to belief-relinquishment. Apparently there is a
difference between honest and dishonest error. Now the hetv-ābhāsa are not quite formal
fallacies. None of them is strictly speaking a deductive fallacy. They are rather of a
peculiarly epistemic nature, and as such are included in the Nyāya-sūtra’s list of nigraha-
sthāna, broadly “defeaters,” bādhaka, or more specifically within a debate context, “minus
points,” or “occasions for censure,” as the compound is sometimes translated. So let us
take up nigraha-sthāna and then the hetv-ābhāsa.
Now Dharmakīrti devotes his Vāda-nyāya to simplification of the Nyāya-sūtra’s list of
nigraha-sthāna. The Nyāya-sūtra puts hetv-ābhāsa within a set of twenty-two nigraha-
sthāna which include “switching theses” (pratijñā-antara), “skipping steps” (nyūna), tu
quoque (“two wrongs don’t make a right,” mata-anujñā), “missing the opportunity to
censure” (paryanuyojya-upekṣaṇa), and “contradicting an accepted thesis”
(apasiddhānta). Dharmakīrti reclassifies according to whether the fault belongs to an
original proponent or opponent of a thesis and further according to whether a debater
should or should not have stated something or other—a schema which Matilal (1998, p.
87) says is “a more systematic and sophisticated formulation of the types of clincher.”
However, to my mind the later Naiyāyika classification of the defeaters as faults
pertaining to a person (puruṣa-doṣa) as opposed to logical flaws (hetu-doṣa) is still more
sophisticated.11 The hetv-ābhāsa are logical flaws; at least four of the five are and the one
that is not, sat-pratipakṣa, “counterinference,” while clearly epistemic, is also not simply
an “occasional error” (puruṣa-doṣa), as will be explained.12 The remaining nigraha-sthāna
are, we might say, procedural or presentational flaws, agential flaws, such as equivocation
and semantic or grammatical incoherence, that have to do with a debater’s mastery of the
procedures of an organized debate including language. And there are a few that may be
viewed as violations of (Gricean) conversational maxims (such as “irrelevance,” artha-
antara, which runs afoul of the Gricean maxim of relation: Be relevant). All the nigraha-
sthāna are, to be sure, defeaters in a debate context, but few are utilized very often as
arguments in philosophic texts and do not command much attention among later
Naiyāyikas despite their elaboration throughout the “Second Daily Lesson” of the Nyāya-
sūtra’s chapter five.
Gaṅgeśa and cohorts give quite a lot of attention, however, to the five “pseudo-provers”
discussed in the Nyāya-sūtra and its commentarial literature. Most of the remaining
defeaters used and discussed by Naiyāyikas fall within the category of tarka, or, we might
say, within the extended tool set of philosophic reasoning. Now tarka, which is more
precisely “suppositional reasoning,” includes many logically interesting items, and we
shall note ten forms after reviewing the popular “pseudo-provers.” The hetv-ābhāsa when
identified are thought directly to reveal an inferential error; the others, including the
upādhi, are indirectly vitiators. It is important to appreciate that the “fallacies of the
prover” are not the only recognized fallacies in the sense of epistemic “defeater,”
bādhaka, but it is also convenient to separate the lists and consider the forms of tarka
separately.
There are also a few other rules surrounding inference and debate whose violations
have not made any of the lists that I know of, although they are used argumentationally to
refute a mistaken position. For example, there is the upādhi, which is recognized as an
inferential undercutter but does not make the lists. “Proving what is already accepted,”
siddha-sādhana, is also a complaint often heard but not classified. And there is a rule
about absences that the absentee (pratiyogin, “countercorrelate,” e.g., my glasses that
are not on the table) must be something real and known previously by a knower of the
absence. The idea of an absence of rabbits’ horns on the table breaks the rule and so such
a pseudo-absence could not figure in an inference. (“Absence” is a legitimate metaphysical
category according to Nyāya and some of the other schools, a category with a certain
logic or “grammar,” namely, absentee-Absence-location, that is to say, any absence has an
absentee at a location, e.g., my glasses, the absentee, on the table, the location. The
absence of my glasses on the table is an object which according to Nyāya I directly
perceive.) The rule about the absentee having to be something real and known previously
(the mind of the subject provides the absentee) can be subsumed under a wider rule,
namely, that all predicates be “familiar,” prasiddha, not in the strongest sense of “previous
direct acquaintance” with the properties they represent but in the sense of “locatable”
(Matilal, 1998, p. 147). An unlocatable property cannot figure in inferential or any kind of
knowledge. The absential rule may also be seen as falling within a set of rules flowing
from a schema of categories, substances, qualities, actions, and so on, à la Aristotle and
debate rules he identifies in, especially, the first chapter of his Topics.

The five hetv-ābhāsa

“Deviation,” savyabhicāra, the first “pseudo-prover” on the Nyāya-sūtra’s list, comes in
three varieties. In its simplest form, it is (1) the “common,” sādhāraṇa, which is just
knowledge of a counterexample (Hb, ¬Sb). However, Gaṅgeśa (TCM Tirupati II, 28;
Calcutta 816) formulates the error as an indecisiveness between two alternatives, Sa and
¬Sa, both being evidenced by known correlations with the putative prover H (some Hs
correlating with Ss and some with not-Ss). The Nyāya-sūtra and other texts use the term
“anaikāntika,” which means “indecisive,” as a synonym. Then there is (2) the “unique,”
asādhāraṇa, which is a special instance of such indecisiveness defined as a putative
prover’s not being known to occur in either the sapakṣa or vipakṣa, places where the
probandum is known to occur or not to occur, respectively. In other words, while there
may be negative correlations (¬Hb, ¬Sb, etc.), there are no positive correlations and
some of the putative prover’s absences (¬Hs) correlate with presences of the probandum
(Ss) as well as with some of its absences (¬Ss) so that there is nothing to go on. There is
no evidence of either an S or a not-S. For example, “Sound (śabda) is impermanent (S),
since it locates (śabdatva) being-sound (H).” The universal, being-sound or soundhood, is
located only in sound, which is the inferential subject, and so it cuts no ice as a prover. Of
course, when everything in the universe is an S, the case is different. Gaṅgeśa and
company spend much effort separating out good inferences called “positive-only,” kevala-
anvayin (e.g., “The pot is nameable, since it is knowable,” everything being nameable:
nameability is a universally locatable property) as distinct from this fallacy. A third variety
of deviation is (3) the anupasaṃhārin, the “inconclusive,” where the property specifying
the inferential subject is universally locatable such that there can be no evidence of things
outside it (nothing is outside it) without begging the question. For example, “Everything is
impermanent (S), since it is knowable (H).” Here the challenge is to separate out good
inferences where the subject is everything, such as “Everything is nameable, since
everything is knowable.”13
The second hetv-ābhāsa according to the traditional list is viruddha, the
“contradictory,” which Gaṅgeśa defines contrapositively: “The contradictory (putative
prover) is to be the countercorrelate (absentee) of an absence that pervades the
probandum (S → ¬H).”14 In other words, Ha proves the very opposite of a putative
probandum, that is, ¬Sa. Here problems arise with regard to double absences, double
negations, which are not always equivalent to presences, for example, the absence of a
mutual absence of a pot is not the pot but potness, according to Raghunātha (Matilal,
1968, p. 54). The Vaiśeṣika theory of absence types, which runs through Udayana and
Gaṅgeśa as well, is much criticized in the post-Gaṅgeśa period (Bandyopadhyay, 1977, pp.
114–116). Discussion of the difference between relational and mutual absences and of the
varieties of relational absence in the fashion of the classical commentaries would take us
too far afield. In essence, the contradictory is the straightforward fallacy of a putative
prover establishing a probandum’s absence, albeit when the probandum is itself an
absence there are certain difficulties and a need for further refinement.
The third hetv-ābhāsa is sat-pratipakṣa, “counterinference” (the prakaraṇa-sama of the
Nyāya-sūtra), which is, as has been explained, a putative prover that would establish a
conclusion Sa except that its evidence base consisting of correlations, positive and
negative, between things H and S, is matched by an opponent’s inference to ¬Sa complete
with correlations between a counterprover J and ¬S. About this fallacy Bandyopadhyay
(1977, p. 127) says, “The very possibility of this check implies that the premises of the
countersyllogism cannot be proved as false by the first disputant who is too confused to
question them.” The point, however, is not that the proponent of an original inference is
confused about the evidence for an opposed conclusion but rather that unless that
evidence can be shown to be flawed the first disputant loses his or her warrant for Sa
even if the proposition is true and there is nothing wrong with the evidence or
correlations between things H and S. This fallacy shows that there is a social dimension to
knowledge, as already pointed out.
Gaṅgeśa views the fourth hetv-ābhāsa, called asiddhi, the “unestablished” (the NyS’s
sādhya-sama), as three separate fallacies (TCM Tirupati II: 115; Calcutta 919): (1) a
putative prover as “unestablished” with respect to location, āśraya-asiddhi, (2) its being
“unestablished” with respect to essential nature, sva-rūpa-asiddhi, and (3) asiddhi by way
of not being pervaded by the probandum, vyāpyatva-asiddhi. This last is said not to be a
separate fallacy but rather reduces to deviation (vyabhicāra), the contradictory
(viruddha), or the defeated (bādha). Still, it is recognized that the word is used for a
person’s failure to be warranted in the case of either the existential or the universal
premise. Since ways the required pervasion-knowledge can be blocked are covered by the
three other fallacies, the “unestablished” as sui generis is a prover’s not being known to
qualify the inferential subject (no pakṣa-dharmatā), and this can occur in a couple of
fashions, (1) and (2) above. This is also the meaning of Nyāya-sūtra 1.2.8 which lists
sādhya-sama as the type of fallacy where the putative prover is “equal to the probandum
(in being unknown as veritably qualifying the inferential subject).”15 Finally, (1) āśraya-
asiddhi occurs in different ways, one, when the inferential subject fails to exist, for
example, “The sky-lotus is fragrant, since it is a lotus” (being-a-lotus cannot be situated in
the putative inferential subject because there is none, because there is no sky-lotus), and,
two, when a subject mistakes something such as dust for smoke—a perceptual error in
essence but one that vitiates a putative inference nonetheless. And (2) sva-rūpa-asiddhi
occurs when the putative prover is contradicted by the specifier of the pakṣatā or
subjecthood, for example, being-a-lake with smokiness.
The fifth and final hetv-ābhāsa is bādha or bādhita, the “defeated,” more precisely, the
“defeated in advance.” Nyāya-sūtra 1.2.4 and 1.2.9 mention a fallacy that is reinterpreted
by Vācaspati and others as the bādha addressed by Gaṅgeśa and company: kāla-atīta and
atīta-kāla, the “mistimed.” As brought out by Bandyopadhyay (1977, pp. 157–158), there
are two ways that an argument can fail to be timely, one, by trying to establish something
that is already well-known (siddha-sādhana), and, two, by trying to establish something
the opposite of which is already known to be true, bādha, the “defeated,” the “defeated in
advance,” as it were, a kind of “patent falsehood.”
The patent truth that all of the hetv-ābhāsa in Gaṅgeśa’s view have an epistemic nature
can be seen from his opening statement in the siddhānta section on the definition of this
fallacy:

We answer. (The “defeated,” bādha, amounts to) certainty (niścaya) about the probandum’s
absence on the inferential subject, a certainty that derives from certainty about the knowledge
status of the cognition of the probandum’s absence. If certainty about an object were to come from
its cognition alone, determinations of veridicality (i.e., certification) would be useless and there
would be the unhappy consequence that there would be certainty about an object from the
knowledge that its cognition is erroneous.
And there is (knowledge of) deviation when there is certainty about the knowledge status of the
cognition of the probandum’s absence from there being such supervenience or because there is
superior evidence (that the inferential subject lacks the probandum as opposed to having it), not
otherwise (i.e., bādha begins with certified knowledge of the probandum’s absence and only then
proceeds to knowledge of deviation). Thus before there is knowledge of deviation depending on
certainty about the probandum’s absence (and afterwards as well), i.e., in both cases the
knowledge status of the cognition of the probandum’s absence has already been known. This is
then the (epistemic) flaw, since there is the supervenience.16

In other words, bādha depends on there being not only first-order, unreflective knowledge
that ¬Sa but also self-conscious certification. The probandum’s absence (¬S) entails that
there is deviation, and so one could say that the underlying fact that accounts for the
inference failure is the lack of pervasion which is equivalent to deviation. But although a
prover as “defeated” is underpinned by pervasion failure and indeed deviation—the fallacy
is in this way “supervenient”—it is also, from the first-person point of view, that which a
subject or debater notices, namely, that it is already known that the inferential subject
lacks the probandum. To identify this fact, which is a fact about what we are certain of, is
the fallacy and vitiator.
To insist that bādha demands certified knowledge, which presents a higher barrier to
doubt than with the pramāṇa, the “knowledge sources,” working automatically (as with
knowledge of nearby clouds from hearing thunder), Gaṅgeśa avoids certain difficulties
with regard to dialectical situations, such as whether an opponent’s inference to the
conclusion ¬Sa should count as a counterinference, successfully blocking an original
inference to Sa, or as an instance of the “defeated” in that we already know from our own
inference that ¬Sa is false. To repeat, certification raises our level of confidence above
that of unreflective knowledge. No inference to the conclusion that fire is not hot will
shake us from our conviction, reinforced by much experience and by reflection, that fire is
hot.
Still, no Naiyāyika that I know of presents clear formula to determine degrees of
justification in cases of pramāṇa conflict and the like, whether perception trumps
inference or vice versa and under what conditions. So just how disagreement impacts
justification is not clear. An epistemology of conflicting reasons and evidence is not
worked out, and much seems to depend on circumstances of topic or the prestige of
authority (as with people who are considered nipuṇa, expert, sharp-eyed, in certain kinds
of perception). However, Nyāya has developed rich resources for dispute resolution in its
theory of the ways the scales of justification may be tipped to the one side or the other by
unfavorable or favorable tarka, “(suppositional) reasoning,”17 the drawing out of untoward
consequences, considerations of simplicity, and so on. To this we turn now in the final
section of this chapter.

Tarka: the forms of philosophic argumentation

Although tarka is treated in an early section of the inference chapter of Gaṅgeśa’s Jewel
and no list of types of “(suppositional) reasoning” is forthcoming, he and other Naiyāyikas
recognize several types: (1) ātma-āśraya, “self-dependence” (or begging the question),18
(2) anyonya-āśraya, “mutual dependence” (mutual presupposition), (3) cakraka,
“circularity” (reasoning in a circle), (4) anavasthā, “infinite regress,” and (5) aniṣṭa-
prasaṅga, “unwanted consequence” (a list present in Udayana, ATV 863), plus (6)
prathama-upasthitatva, “being presupposed by the other,” the “first established” (a form
of anukūla, “favorable,” tarka), (7) utsarga, “(hasty) generalization,” (8) vinigamanā-
viraha, “differentiation failure,” (9) lāghava, “theoretic lightness” (another form of
favorable tarka), and (10) gaurava, “theoretic heaviness.”19 Let us take an overview.
Sometimes our beliefs/cognitions have been generated by processes that would be
counted pramāṇa if they did not face counterconsiderations. But in facing
counterconsiderations—in being reasonably challenged—they are not trustworthy and do
not guide unhesitating effort and action. There is a social dimension to knowledge where
reasoning reigns, resolving controversy. These are the ways of tarka, “hypothetical” or
“suppositional reasoning.” Paradigmatically, tarka is called for in order to reestablish a
presumption of truth in favor of one thesis that has putative source support against a rival
thesis that also has putative source support, a thesis and a counterthesis both backed up
by, for example, apparently genuine inferences (co-counterinference, the most common
situation) or by competing perceptual or testimonial evidence. By supposing the truth of
the rival thesis and (in Socratic style) showing how it leads to unacceptable consequences
or breaks an intellectual norm, one repossesses a presumption of truth, provided one’s
own thesis does indeed have at least the appearance of a pramāṇa in its corner. Gaṅgeśa
joins a consensus across the classical schools that such arguments are not in themselves
knowledge-generators although they can swing the balance concerning what it is rational
to believe.
Furthermore, “reasoning,” tarka, is what a philosopher is good at, the drawing out of
implications of opposed views and testing them against mutually accepted positions
(siddhānta), according to, broadly speaking, criteria of coherence but also of simplicity.
The standard variety shows that an opponent’s hypothesis (or an opposite thesis, ∼p)
violates an intellectual norm. There are a few forms of “favorable” tarka, for example,
having one’s own thesis presupposed by the opponent’s while the reverse does not hold.
Nevertheless, usually tarka is thought of as “unfavourable,” only indirectly supporting a
thesis p by establishing a counterconsideration against a rival ∼p, or, sometimes, against
a rival hypothesis q which would explain a set of commonly recognized facts in different
terms.
Gaṅgeśa in the tarka section of his Jewel’s inference chapter sees suppositional
reasoning as called for primarily in the special circumstance of doubt about a pervasion
for which there is other evidence. In the preceding section (TCM Tirupati I: 171; Calcutta
175), he pointed out that repeated observation (“even hundreds of times”) is insufficient
to guarantee knowledge of a pervasion: not all things made out of earth are scratchable
by iron (diamonds are a counterexample). He acknowledges that the possibility of a
counterexample cannot be eliminated. Nevertheless, we cognize pervasions, and reliably
under certain circumstances, and meaningful doubt can be eliminated—this is the upshot
of his view. Reasoning can show “untoward ramifications” of not accepting a conclusion
that is evidenced. Doubt or challenging need not be paralyzing, for there are the forms of
tarka to eliminate wrong views and thus indirectly support right views, even in some
cases support them directly, as with (6) above, prathama-upasthitatva, “being
presupposed by the other.”
In sum, formulating inferences and detecting inferential fallacies as well as use of tarka
constitute ways to proceed in philosophic debate and inquiry. Broadly, these are forms of
argumentation, not the only forms, for there are also the pramāṇa of perception and
testimony (and even analogy, upamāna, if the question is a word’s meaning) whose
identification counts, along with inferences self-consciously carried out, as certificational
according to Nyāya. Inferences, for example, are, however, the most prominently
employed forms of argumentation in Nyāya texts, albeit on a few topics the pramāṇa is
testimony.

Notes
1. For example, Uddyotakara says in his opening verse that his purpose is to remove the “ignorance
imparted by bad reasoners,” or, better, “misleading reasoners,” kutarkin, namely, as Vācaspati tells
us explicitly, the arguments of Diṅnāga and followers (NyV and NySVTatp, avatāra to 1.1.1;
Nyāyadarśanam: 1).Vācaspati uses the word “ācchādita”: the Nyāya teaching or śāstra has become
“covered over.” The point is to dispel the fog, not to innovate or correct core positions
(Nyāyadarśanam: 2), although innovate one does of course in certain instances.
2. The development of logic and epistemology in classical India is a long story to the writing of which
many scholars have now contributed, with some disagreeing with my contention. But the opposed
parties learned from one another and it’s the convergences that are striking from a distance. A
comparison of two late treatises both called Tarka-bhāṣā bears this out: by the Buddhist
Mokṣākaragupta (trans. Kajiyama [1998]) and by the Naiyāyika Keśava Miśra (trans. Iyer [1979]).
For example, three types of the vyabhicāra fallacy are identified and analysed almost identically
(Kajiyama, 1998, pp. 113–114; Iyer, 1979, pp. 104–107). There are some important divergences, for
 example, the Buddhist concept of antar-vyāpti, “internal pervasion” (Mishra, 2002), whereas good
inferences so analyzed are in the Nyāya view “negative-only” (Phillips, 2012, p. 63). Gokhale
(1992) attends carefully to several differences in the Buddhist and Naiyāyika approaches (the
classifications of Diṅnāga’s Nyāya-mukha, which seem to depend on Vātsyāyana but are also more
sophisticated, similarly those of the Nyāya-praveśa of Śaṅkarasvāmin—Diṅnāga’s disciple or,
according to the Chinese tradition, Diṅnāga himself—do not entirely match up with Nyāya’s), but,
given the sharp doctrinal divides, the convergences are striking even in the early period and they
become tighter over the centuries. Of course, had the two sides not agreed at least largely on
forms of argumentation, how would they have been able to engage one another so vigorously?
3. For example, there is no discussion of fallacies in either Vol. 6 or Vol. 13 of Encyclopedia of Indian
Philosophies (1993 and 2011, ed. by K. Potter and S. Bhattacharyya), within the long introductions
to the periods covered of Nyāya philosophy (Vol. 6 including Raghunātha and Vol. 13 authors
immediately following him) by Sibajiban Bhattacharyya. Ganeri (2011), although providing a lucid
overview of Raghunātha’s positions in metaphysics and his innovations in technical language, has
nothing on fallacies. A notable exception to the general lack of scholarship here is Bandyopadhyay
(1977), which discusses advances and clarifications by Raghunātha. Nevertheless,
Bandyopadhyay’s excellent book will be mined by me almost only for its bits on Gaṅgeśa.
4. Taber (2004) argues successfully against Oetke (1996) that the paradigm logical form embedded
in a good inference is monotonic, not non-monotonic, as Oetke had contended. New information is
irrelevant to the validity of the pattern itself, although it may well be relevant to a subject’s
justification for acceptance of the premises. Examples of inference presented in classical texts
often seem non-monotonic, but this is precisely because it is realized that care is to be taken with
the premises, whether each is warranted, and with putting them together consciously in
parāmarśa, “reflection.” Of course, the one pattern focused on and used in philosophy is not
capable of handling all the kinds of reasoning that we use in everyday life. An adequate theory of
everyday reasoning is not the Nyāya project (which is, rather, examination of a particular pattern
and causal nexus proved to generate new knowledge that is neither perceptual nor testimonial,
“following”—“anu-” in “anumiti”—from other things known). Again, the reason the Nyāya pattern
(of universal instantiation and modus ponens) seems non-monotonic is that there may well be
questions about whether a subject knows the premises. And the fallibilism that attaches to the
premises of course attaches to the conclusion, too. (Cf. Israel [1980] who similarly voices an
epistemological complaint against the very idea of non-monotonic logic, according to Koons
[2013].)
5. In an “inference for others,” there are five “members” (avayava) to be construed as a single
statement governed by grammatical and semantic rules and designed to provoke inferential
knowledge in another. The idea—from Gautama through Navya Nyāya—is that the five-part form is
an ideal ordering corresponding to requirements of syntactic binding or “expectation” and
semantic “fittingness” (ākāṅkṣā and yogyatā) crucial to testimonial knowledge of a sort that would
prompt the hearer to draw a desired conclusion and have inferential knowledge for herself. In
other words, there is to be a provoking of inferential knowledge by a statement that provides the
inferential terms in the proper relationship. Thus violations of semantic fittingness, for example, or
some other sort of failure properly to communicate in the context of a debate, are sometimes
identified as defeaters of inference although the faults have nothing to do with logic. As far as
logic is concerned, it is recognized that an “inference for others” depends on the cogency of a
corresponding “inference for oneself.”
6. For example, Staal (1973, p. 158): “(x) (A(h,x) → A(s,x)) ‘For all x, if h occurs in x, then s occurs in
x.’” Here the same idea will be represented: (x) (Hx → Sx) “For all x, if x locates H, then x locates
S,” or, more simply, “For all x, if x is H, then x is S,” with “H” as a variable representing a prover
property, “S” a probandum property, and “x” an individual (e.g., a, sometimes not a concrete but
an abstract entity) or a set (even everything: see below). Concerning why in some cases this
representation is inadequate from the Buddhist perspective—although not in the same ways I think
for Nyāya—see the web publication by Glashoff, “The formal content of the Trairūpya doctrine,
Dignāga’s Hetucakra, and Uddyotarakara’s extension.”
7. The Buddhist position from Dharmakīrti on is a little different in recognizing in addition to
inductive generalization in effect an a priori, antar-vyāpti (see in particular Mishra [2002]). Nyāya
does not, or does only through what it calls the extraordinary sensory connection (alaukika-indriya-
saṃnikarśa) with a universal such as cowhood such that a future cow, a cow not yet existing, will
be known, for example, to have a dewlap. And universals are themselves said to be known from
recurrent experience (anugata-anubhava).
8. The Nyāya solution to the problem of induction would seem to be bootstrapping: the methods of
positive and negative correlation themselves correlate with the production of knowledge and
success in action without, so far, a counterexample.
9. According to Nyāya, which is an externalism (to use the contemporary term of art), to know that p
is not necessarily to know that you know that p. All that is necessary for knowledge is that p be
true and the subject’s belief that p be generated by a veritable pramāṇa.
10. Tuske (1998) discusses the “raven paradox” that attaches to inference by negative correlation:
sight of another black raven (Hb, Sb) provides support for the generalization that all ravens are
 black (all H are S), but sight of a white dove, which is neither a raven nor black (¬Hc, ¬Sc), does
not. Gaṅgeśa’s solution to the problem (Phillips, 2009), if a solution it be, is different from
Diṅnāga’s as explained by Tuske. In brief, there is a relevance condition as brought about by the
requirement that a probandum be “familiar,” prasiddha.
11. A puruṣa-doṣa is also called an anitya-doṣa, an “occasional flaw” (e.g., TCM Tirupati II, 22;
Calcutta 798) as opposed to a “constant flaw,” nitya-doṣa, which is equivalent to a hetu-doṣa: see,
for example, Mathurānātha’s commentary on the TCM section on deviation (Calcutta 799–800).
Later, Gaṅgeśa contends that sometimes even “failure to grasp a pervasion is (not a logical flaw
but) a flaw belonging to a person or subject”: vyāpti-ajñānasya puruṣa-doṣatvāt (TCM Tirupati II,
49 and Calcutta 826). Cf. Bagchi (1953, p. 207 n.3). Nevertheless, the reason for separating out
the hetv-ābhāsa from other nigraha-sthāna is that all the hetv-ābhāsa sport a regularity
considering evidence such that what they are most like are inductive errors highlighting inductive
rules.
12. Siderits (2003) is overfussy in insisting that anumāna is not an argument in that it is not, like an
argument, composed of propositions. (Yes—but the epistemic events making up the inferential
process that generates inferential knowledge and the knowledge itself embed propositions. How
else could debate proceed?) However, Siderits correctly sees that inference and the fallacies of the
prover, hetv-ābhāsa, are fundamentally epistemic and neither strictly deductive nor inductive. And
they are also not somehow both in his view. Still, it seems to me obvious that there are both
inductive and deductive considerations present in anumāna, which may be abstracted by us while
being mindful that in the final analysis we are talking about epistemic success and failure.
13. Matilal (1998, p. 128) overlooks this part of the effort when he writes, “According to the later
Nyāya school, any argument that has a conclusion (a thesis) of the form ‛Everything is F’ is
fallacious, because it would be inconclusive.” Indeed, Gaṅgeśa says explicitly: “‘Everything is
nameable since it is knowable,’ which is an inference with a good and genuine prover (and
inferential knowledge is generated thereby), is not an inference where a universally present
property specifies the inferential subject, because the prover here occurs on things about which it
is known with certainty that they have the probandum”: sarvam abhidheyam prameyatvād iti sad-
dhetau na kevala-anvayī pakṣatā-avacchedakaḥ niścita-sādhyavad-vṛttitvāt (TCM Tirupati II: 68;
Calcutta 839). In Gaṅgeśa’s view, things are meant severally by the pronoun, there being no class
character of “everything-hood.” Thus Matilal is right in general in that an inferential subject
specified by a universally locatable property would involve an instance of the fallacy, since there
would be no correlations on which knowledge of pervasion could be based. Nevertheless, the
inference above is a good one.
14. sādhya-vyāpaka-abhāva-pratiyogitvaṃ viruddham (TCM Tirupati II: 77; Calcutta 855).
15. Vātsyāyana’s example (NySBh 1.2.8): “Shadow is a substance, since it has motion (and whatever
has motion is a substance).” The fallacy does not lie in the purported pervasion, “Whatever has
motion is a substance,” but with the putative prover’s being a property of the inferential subject. In
fact, shadows do not move in that a shadow is to be analyzed as an absence or obstruction of light
and the apparent movement is to be accounted for by the movement of the obstruction relative to a
source of light or to the light-source’s own movement. So the prover (having-motion) is
unestablished as qualifying the inferential subject (shadow); it is “equal to the probandum
(substancehood)” in being in this way unproved.
16. TCM Tirupati II: 136; Calcutta 960. ucyate | pakṣe sādhya-abhāva-niścayaḥ sādhya-abhāva-jñāna-
pramātva-niścayāt | jñāna-mātrād artha-niścaye prāmāṇya-vaiyarthyam bhramatvena jñānāt artha-
niścaya-prasaṅgaś ca | tathā ca upajīvyatvena adhika-balatvena vā pakṣe sādhya-abhāva-jñānasya
pramātva-niścaye savyabhicāro na anyathā ity ubhayathā api sādhya-abhāva-niścaya-adhīna-
vyabhicāra-jñānāt pūrvaṃ jñātaṃ sādhya-abhāva-jñāna-pramātvam eva doṣaḥ upajīvyatvāt |
17. Ganeri (2001, p. 151ff) pioneers the translation, “suppositional reasoning.” Technically, tarka is
classified as a kind of false cognition, one that we realize is false while it is entertained.
Vātsyāyana’s commentary on the tarka sūtra, NyS 1.1.40 (Nyāyadarśanam: 321–322), mentions
belief in rebirth, common to both Buddhism and Hinduism, as in tension with the (Buddhist) thesis
that there is no enduring self (ātman). Both sides present apparently good inferences. This draws
the Nyāya inferential knowledge of a self into question. But here tarka reestablishes a presumption
in favor of an enduring self: “If there were no enduring self, rebirth would be impossible.” Kang
(2010, p. 1) renders tarka as “reflective analysis without requiring further factual information on
the object of investigation.” This is not to say that such reasoning would not test a hypothesis
against facts but that these would be known and accepted facts, not new ones to be uncovered.
18. The important subrule that the inferential subject may not be cited as an example of correlation
for risk of begging the question is denied by Buddhist logicians such as Ratnakaraśāntipāda with
the notion of antar-vyāpti, “internal pervasion.”
19. Bagchi (1953, p. 151) presents the list following a note by Viśvanātha on NyS 1.1.40
(Nyāyadarśanam, pp. 325–327), where the seventeenth-century follower of Raghunātha provides
examples of each of the first five in three varieties (utpatti-sthiti-jnāpti-dvārā). Viśvanātha
acknowledges in the Vṛtti that the last five are also viewed as forms of tarka by some philosophers,
but, so he contends, they do not draw out an unwanted consequence and so are not strictly
speaking tarka but rather pramāṇa-sahakārin, “aids to the workings of the knowledge sources”
 (Nyāyadarśanam, p. 327). But of course so are the genuine forms of tarka.

Bibliography
Aristotle (1997), Topics, Books I and VIII. Trans. R. Smith. Oxford: Clarendon.
Bagchi, S. (1953), Inductive Reasoning: A Study of tarka and Its Role in Indian Logic. Calcutta:
Munishchandra Sinha.
Bandyopadhyay, N. (1977), The Concept of Logical Fallacies. Calcutta: Sanskrit Pustak Bhandar.
Bhattacharya, D. C. (1958), History of Navya-Nyāya in Mithilā. Darbhanga: Mithila Institute.
Bhattacharyya, S., and Potter, K. H. (eds.) (2011), Encyclopedia of Indian Philosophies, vol. 13. Nyāya-
Vaiśeṣika Philosophy from 1515 to 1660. Delhi: Motilal Banarsidass.
Dasti, M., and Phillips, S. (2010), “Pramāṇa are factive—a response to Jonardon Ganeri.” Philosophy
Easy and West, 60, 535–540.
Dharmakīrti, Vāda-nyāya (ed.) and (trans.) P. P. Gokhale (1993), Vādanyāya of Dharmakīrti: The Logic
of Debate. Delhi: Sri Satguru Publications.
Diṅnāga, Nyāya-mukha. trans. G. Tucci (1930), The Nyāya-mukha of Dignāga. Heidelberg: Materialien
zur Kunde des Buddhismus.
Ganeri, J. (2001), Philosophy in Classical India: The Proper Work of Reason. London: Routledge.
Ganeri, J. (2011), The Lost Age of Reason: Philosophy in Early Modern India 1450–1700. Oxford:
Oxford University Press.
Gaṅgeśa, Tattva-cintā-maṇi (TCM), inference chapter. (1) N. S. Ramanuja Tatacharya (ed.) (1982 and
1999), with the Prakāśa commentary by Rucidatta Miśra and a subcommentary by
Dharmarājādhvarin, 2 vols. Tirupati: Kendriya Sanskrit Vidyapeetha Series 33. (2) K. Tarkavagish
(ed.) (1884–1901, reprint 1991), with the commentary of Mathurānātha. 2 vols. Calcutta: Asiatic
Society. The upādhi section translated: S. H. Phillips and N. S. Ramanuja Tatacharya (2002), Gaṅgeśa
on the Upādhi, the “Inferential Undercutting Condition.” New Delhi: Indian Council of Philosophical
Research.
Gautama, Nyāya-sūtra (NyS). A. M. Tarkatirtha, T. Nyayatarkatirtha, and H. K. Tarkatirtha (eds.)
(1936–1944, reprint 1985), Nyāyadarśanam (Nyāya-sūtra with four commentaries, the Nyāya-sūtra-
bhāṣya (NySBh) of Vātsyāyana, the Nyāya-vārttika (NyV) of Uddyotakara, the Nyāya-vārttika-
tātpārya-ṭīkā of Vācaspati Miśra, and the Vṛtti of Viśvanātha). Calcutta: Calcutta Sanskrit Series 18.
Glashoff, K., “The formal content of the trairūpya doctrine, Dignāga’s hetucakra, and Uddyotakara’s
extension.” http://webapp5.rrz.uni-hamburg.de/syllogism/ks_dignaga.html.
Gokhale, P. P. (1992), Inference and Fallacies Discussed in Ancient Indian Logic, with Special Reference
to Nyāya and Buddhism. Delhi: Sri Satguru Publications.
Grice, H. P. (1975), “Logic and conversation,” in P. Cole and J. L. Morgan (eds.), Syntax and Semantics,
vol. 3. New York: Academic Press, pp. 41–58.
Ingalls, D. H. H. (1951), Materials for the Study of Navya-Nyāya Logic. Cambridge: Harvard University
Press.
Israel, D. (1980), “What’s wrong with non-monotonic logic,” in Proceedings of the First National
Conference on Artificial Intelligence. Palo Alto, CA: American Academy of Artificial Intelligence, pp.
99–101.
Kang, S. Y. (2010), “An inquiry into the definition of tarka in Nyāya tradition and its connotation of
negative speculation.” Journal of Indian Philosophy, 38, 1–23.
Keśava Miśra, Tarka-bhāṣā. D. R. Bhandarkar and Pandit Kedarnath (eds.) (1937), Tarkabhāṣā of
Keśava Miśra. Bombay: Bombay Sanskrit and Prakrit Series 84. Trans. S. R. Iyer (1979), Tarkabhāṣā
of Keśava Miśra. Varanasi: Chaukhambha Orientalia.
Koons, R. (2013), “Defeasible reasoning,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy.
http://plato.stanford.edu/archives/spr2013/entries/reasoning-defeasible/.
Maṇikaṇṭha Miśra, Nyāya-ratna. V. S. Sastri and V. Krishnamacharya (eds.) (1953), Nyāyaratna of
Maṇikaṇṭha Miśra. Madras: Madras Government Oriental Series.
Mathurānātha, Māthurī (commentary on Gaṅgeśa’s TCM). See Gaṅgeśa (2).
Matilal, B. K. (1968), The Navya-Nyāya Doctrine of Negation. Cambridge: Harvard University Press.
Matilal, B. K. (1998), The Character of Logic in India. Ed. J. Ganeri and H. Tiwari. Albany: State
University of New York Press.
Mishra, A. (2002), Antarvyāpti. New Delhi: Indian Council of Philosophical Research.
Mokṣākaragupta, Tarka-bhāṣā. Embar Krishnamacharya (ed.) (1942). Gaekwad’s Oriental Series 94.
Baroda: Oriental Institute. Trans. Y. Kajiyama (1998), An Introduction to Buddhist Philosophy: An
Annotated Translation of the Tarkabhāsā of Mokṣākaragupta. Vienna: Arbeitskreis für Tibetische und
Buddhistische Studien, Universität Wien.
Oetke, C. (1996), “Ancient Indian logic as a theory of non-monotonic reasoning.” Journal of Indian
Philosophy, 24, 447–539.
Phillips, S. H. (2009), “From the Tattva-cintā-maṇi by Gaṅgeśa, the kevala-vyatireki-prakaraṇam,
‘negative-only inference,’ annotated translation and commentary,” in P. Balcerowicz (ed.), Logic and
Belief in Indian Philosophy, Warsaw Indological Studies, vol. 3. Delhi: Motilal Banarsidass, pp. 435–
505.
Phillips, S. H. (2012), Epistemology in Classical India: The Knowledge Sources of the Nyāya School.
New York and London: Routledge.
Potter, K. H., and Bhattacharyya, S. (eds.) (1993), Encyclopedia of Indian Philosophies, vol. 6. Indian
Philosophical Analysis: Nyāya-Vaiśeṣika from Gaṅgeśa to Raghunātha Śiromaṇi. Delhi: Motilal
Banarsidass.
Prets, E. (2001), “Futile and false rejoinders, sophistical arguments and early Indian logic.” Journal of
Indian Philosophy, 29, 545–558.
Ranganath, S. (1999), Contribution of Vācaspati Miśra to Indian Philosophy. Delhi: Pratibha Prakasan.
Siderits, M. (2003), “Deductive, inductive, both or neither.” Journal of Indian Philosophy, 31, 303–321.
Staal, J. F. (1973), “The concept of pakṣa in Indian logic.” Journal of Indian Philosophy, 2, 156–166.
Taber, J. (2004), “Is Indian logic nonmonotonic?” Philosophy East and West, 54, 143–170.
Tuske, J. (1998), “Diṅnāga and the raven paradox.” Journal of Indian Philosophy, 26, 387–403.
Udayana, Ātma-tattva-viveka (ATV). V. Dvivedin and L. S. Dravida (eds.) (1986 reprint), Bibliotheca
Indica 170. Calcutta: The Asiatic Society.
Uddyotakara, Nyāya-vārttika (NyV). See Gautama.
Vācaspati Miśra, Nyāya-vārttika-tātparya-ṭīkā (NyVTatp) (“Notes on (Uddyotakara’s) Intention in his
Nyāya-vārttika Commentary”). See Gautama.
Vātsyāyana, Nyāya-sūtra-bhāṣya (NySBh). See Gautama.
Viśvanātha, Vṛtti on the Nyāya-sūtra. See Gautama.
Wada, T. (2000), “The origin of Navya-Nyāya and its place within the history of Indian logic,” in S. Hing
and T. Wada (eds.), Three Mountains and Seven Rivers: Prof. Musashi Tachikawa’s Felicitation
Volume. Delhi: Motilal Banarsidass, pp. 439–462.
 3
Jayarāśi and the Skeptical Tradition*
Eli Franco

“Skepticism” may be defined as a philosophical standpoint or attitude which consists in

doubting, distrusting, or disbelieving all knowledge claims. As such, it was often
considered the bankruptcy and end of philosophy. Indeed, if the skeptics are right and
dogmatists are wrong, none of our cherished opinions and beliefs, none of the facts or
data we hold to be true, none of the sciences like physics or biology can be properly
established to be true.
It is one of the merits of Richard Popkin (1923–2005)—not only the most renowned
historian of Skepticism, but also a very distinguished skeptic himself—to have changed
our perception of the historical development of Western philosophy, especially in the
modern period from the renaissance to the twentieth century. Thanks to his vast and
fascinating body of work, Skepticism is no longer considered just an anti-intellectual
negative attitude that aims at discrediting all human endeavors toward knowledge, but as
one of the major forces in the development of modern philosophy, a development that to a
great extent may be described in terms of crises pyrrhoniennes and the responses they
provoked. Furthermore, the enduring influence of Hume has put Skepticism at the basis
of all modern schools of thought, including pragmatism, positivism, analytical philosophy,
existentialism, and philosophy of science.
The study of Skepticism in the Indian philosophical tradition is still in its infancy. To my
knowledge, none of the histories of Indian philosophy attempts to follow its development
through the centuries, assess its impact on rival schools of thought, or even identify it as a
current of thought. This is partly due, no doubt, to the fact that the history of Indian
thought is still a relatively neglected area of studies, but the main reason for this neglect
is that Skepticism in South Asia, unlike its counterpart of the European tradition, never
formed an independent school of thought (such as Pyrrhonism or Academic Skepticism).
We know of skeptics affiliated with the Madhyamaka tradition, with Lokāyata Materialism,
or with Advaita Vedānta, but not of just skeptics. Skeptics in South Asia have usually
(though not always) not seen themselves as a part of a single tradition, and perhaps justly
so.
This leads us to the question of how to define or characterize Skepticism in South Asia.
To the question “How can one know anything?” all Indian philosophers answer
unanimously: By having a means of knowledge (pramāṇa). This answer may sound almost
tautological and no two significant philosophers would understand the term in the same
manner—opinions vary widely about the nature, the number, the definition, and the scope
of pramāṇas—nevertheless the notion of pramāṇa played a crucial role in structuring
South Asian epistemologies: It is around this concept, its definitions, varieties, and
applications to metaphysical issues that Indian philosophy developed in its most dynamic
period (roughly from the fifth to the twelfth century). Skepticism in South Asia may thus
be defined as the philosophical stance that rejects all pramāṇas, all means of knowledge.

Self-reference and circularity of means of knowledge

The first sweeping rejection of all means of knowledge that has come down to us is
contained in a short treatise1 attributed to Nāgārjuna, the founder of the Madhyamaka
school.2 According to Nāgārjuna’s opponent, everything is established by four means of
knowledge: perception, inference, comparison, and verbal testimony.3 But if everything
has to be established by a means of knowledge, how are the means of knowledge
themselves established? If they are established by other means of knowledge, these other
means of knowledge also have to be established by yet other means of knowledge and
thus an infinite regress arises. If, on the other hand, the means of knowledge are
established without means of knowledge, the tenet that everything has to be established
by a means of knowledge has been abandoned. And why should certain things be
established by a means of knowledge, but others (the means of knowledge themselves)
not? The opponent tries to solve the problem by claiming that the means of knowledge
establish their objects as well as themselves, just as fire illuminates other things as well
as itself. Nāgārjuna explains at some length that the analogy is flawed because actually
fire does not illuminate itself, nor does it, when burning an object, burn itself. Further, the
means of knowledge cannot be established in relation to their objects, because in that
case the objects would be established independently of the means of knowledge.
Alternatively, interdependence between the means and the objects of knowledge would
arise. Thus, it is not possible to establish the means of knowledge by themselves, nor by
one another, nor by other means of knowledge, nor by their objects, nor accidentally.

The true characteristic of means of knowledge—truthfulness

The above is undoubtedly a powerful critique of the means of knowledge.4 However, the
most detailed and devastating critique of the pramāṇas was penned by Jayarāśi Bhaṭṭa, a
skeptic philosopher affiliated with the materialist Lokāyata school who probably lived in
the ninth century CE. His work, “The Lion Which is a Calamity for (All) Principles”
(tattvopaplavasiṃha), is unique. It is the only work that survived from the Lokāyata
school; it is also the only work in which full-fledged Skepticism is propounded without an
affiliation to a religious tradition such as the Madhyamaka or Advaita Vedānta. Unlike
Nāgārjuna, Jayarāśi does not concentrate on the circularity and interdependence of the
means of knowledge, but on their true characteristics (sal-lakṣaṇa). The entire work
revolves around the different characteristics, or definitions, of the means of knowledge
according to major philosophical schools of his time, most prominently, Nyāya, the
Buddhism of the Dharmakīrti school, and Bhāṭṭa Mīmāṃsā.
The most important characteristic of all means of knowledge is their truthfulness or
reliability. In what follows I shall summarize in some detail Jayarāśi’s arguments against
the two most important means of knowledge, namely, sense perception (pratyakṣa) and
inference (anumāna). They illustrate nicely, I believe, the general flavor of the book.

Four criteria of truth

Suppose one sees at some distance a certain object, say a body of water. How is it possible
to know that the cognition5 is true? Considering the various epistemological theories that
were current in Classical India one can enumerate four criteria of truth: (1) the fact of
being produced by undisturbed/non-defective causal factors (aduṣṭa-kāraka), (2) the
absence of sublation/contradiction (bādhā), (3) the efficiency of the activity toward
obtaining the perceived object (pravṛtti-sāmarthya), and (4) the intrinsic validity of the
cognition (svataḥprāmāṇya).

The production by non-defective causal factors

The first criterion would not do, for if the cognition is said to be true because the process
of its production has not been disturbed in any way, how is precisely this fact, that the
process has not been disturbed,6 known? It cannot be known through perception because
the good function of the sense of sight,7 and so on, is beyond the senses themselves. It can
also not be known by inference, because there is no proper inferential sign (or reason) for
such inference. If one says that the true cognition itself is the sign for the fact that the
causal process has been undisturbed, the fallacy of interdependence arises: one knows
that the cognition is true because it arises from a non-defective causal process, and one
knows that it arises from a non-defective causal process because it is true (TUS 2.8–14).

Absence of sublation
The second criterion to be considered is the absence of sublation or contradiction. For
instance, one sees a piece of silver from afar, but when approaching it, finds that it is
mother-of-pearl. The cognition of silver is sublated by the cognition of the mother-of-pearl.
One has a cognition of water, but when approaching the place where the water should be,
one realizes that the supposed water was only a reflection of sunrays. The cognition of
water is sublated by the cognition of sunrays.8 The problem with this criterion of truth is
the interval of time that separates the sublating and sublated cognitions. Consequently,
one cannot be sure at any given moment which of one’s cognitions is going to be proved
false in the future. Further, in certain cases a sublation may never arise. A deceptive
sense faculty may remain so, a disturbed mind may never be cured, certain objects are too
subtle to ever be apprehended correctly, or one may simply die before the sublation has
had a chance to arise. This does not mean that non-sublated cognitions are true. In other
words, a sublation may tell one that a certain cognition was false, but the absence of
sublation does not tell one that a certain cognition is true (TUS 2.15–24).9
The opponent may attempt to defend his position by claiming that although the absence
of a sublating factor for a certain individual person may be due to chance (one may not go
toward the water and never realize that it was only sun rays), in the case of true
cognitions, sublation never arises for anyone. This seems to amount to a commonsense
argument: How do I know that there is only one moon in the sky? Because everybody else
says it is so. However, Jayarāśi comments that this manner of interpreting the absence of
sublation would be impractical. Short of being omniscient, one does not know what all
people think about a particular subject matter (TUS 2.23f.).

Efficiency of activity
The third criterion examined by Jayarāśi is the efficiency of the activity. Basically it is a
pragmatic principle of confirmation: one sees a certain object, one undertakes an activity
in order to obtain10 that object, the activity is successful and the object is obtained, the
cognition of the object is therefore true.11 The trouble with this criterion, Jayarāśi says, is
that the efficiency itself has to be perceived, but one cannot be certain about the
truthfulness of this perception: is the efficiency of the activity perceived or not? If it is not
perceived, how do you know that it exists? If it is perceived, how do you know that its
perception is true? By another efficiency of activity? But that other efficiency of activity
would have to be perceived by another perception, which would have to be confirmed by
another efficiency of activity, and so on ad infinitum (TUS 3.9–15).12
Jayarāśi’s argument raises the same objection that is often used to criticize
pragmatism. To repeat William James’s example,13 if one believes that there are tigers in
India, one is prepared to find tigers there and would be surprised if none are found.
However, as Chisholm and others have pointed out, it is not always the case that a
cognition will give satisfaction if and only if it is true. One may go to Syria, meet some
tigers there, and think one is in India, or one may go to India, not find any tigers there,
but mistake a cat for a tiger.14
There are some additional problems regarding the obtainment of the object. Does it
mean that one obtains the same object that has been seen before? Ideally, this should be
the case. However, certain objects last for just a moment. If one sees water in a river, one
cannot obtain the same water each time one perceives it. According to Nyāya Vaiśeṣika
metaphysics, every watery substance is immediately destroyed when its parts move and a
new substance is formed from the new arrangement of the parts. Thus, one cannot usually
obtain the same water that was seen before. The Naiyāyikas try to bridge the gap
between objects that are seen and those that are obtained by saying that they belong to
the same kind (tajjātīyodakaprāpti). This criterion, however, is both too narrow and too
wide. In some cases, a true cognition may not lead to the obtainment of an object of the
same kind (e.g., a cognition of the stars and of the moon or of other objects beyond one’s
reach), in other cases one may obtain a similar object even when acting upon a false
cognition (e.g., after an erroneous vision of water, obtaining some water accidentally).
Through his work, Jayarāśi discusses still other attempts to account for the obtainment of
an object and equally rejects them as unsatisfactory (TUS 3.19–4.3, 4.22–24, 9.1–4).

Intrinsic validity
The final theory considered by Jayarāśi claims that the validity or truthfulness of a
cognition does not depend on any external criterion, but rather is intrinsic or immanent to
the cognition. This theory is associated above all with two philosophical schools:
Sāṃkhya, which considered both validity and falsity to be intrinsic, and Mīmāṃsā, which
considered validity to be intrinsic, but falsity extrinsic.15 Jayarāśi is probably arguing
against the Mīmāṃsakas,16 who designed their theory so that it would guarantee the
validity of the Veda, which was considered to be eternal and without an author
(apauruṣeyatva).17 Had the Veda been composed by an author, the Mīmāṃsakas argue,
one could suspect it to contain faults, as faults are found in any human composition.
However, being without an author, the Veda is beyond the very possibility of having faults.
Jayarāśi’s analysis of this position is quite elaborate. He raises the questions whether
truthfulness is a property (dharma) of cognition or its own nature (svabhāva), in the
former case, whether it arises before, together, or after the cognition, in the latter case,
whether it is determined by the existence of the cognition alone or by other factors, and if
so, what these other factors may be, whether they participate in the production of
cognition, whether they are perceived, whether the cognition itself is perceived, and if so,
whether by itself or by another cognition. Needless to say, none of the above alternatives
proves satisfactory (TUS 9.18–10.26).
The above could hardly be considered to exhaust what Jayarāśi has to say on the issue
of the truthfulness of perceptions, which also involves metaphysical issues such as the
existence of universals, the plurality of individuals, the relation of inherence, and the
problem of nonexisting objects. However, the above, I believe, suffices to show that the
problem of the criterion of truth arose in the Indian philosophical context as well, and that
Indian skeptics, like their European counterparts, clearly saw that the problem simply has
no solution. In the rest of this chapter, I will describe Jayarāśi’s arguments against the
possibility of a valid inference.

Jayarāśi’s critique of inference

Critique of the Nyāya theory of inference
When perception has been shown to be unreliable, the dismissal of inference follows
easily. Inference is characterized (or defined) in the foundational work of the Nyāya
tradition, Nyāyasūtra (NS 1.1.5), as being preceded by perception (tatpūrvakam
anumānam), which clearly implies that without perception inference cannot take place.
For instance, one observes in places like kitchens that there is an invariable relation or
concomitance between smoke and fire: Wherever there is smoke there is fire.18 This
determination of the relation gives rise to a mnemonic trace (saṃskāra) which enables the
later recollection of the relation. After a while one sees again the so-called inferential sign
(liṅga), in our case smoke, and recollects its invariable relation to fire. One combines the
observation of the sign (i.e., smoke) with the recollection of the relation between smoke
and fire, and infers fire (parāmarśajñānam anumānam). This account of the inferential
process in the Tattvopaplavasiṃha (TUS 64.20–25) specifically follows the Nyāya
tradition, but such a description is generally accepted by all philosophical traditions.
Obviously, if there is no perception of the relation between smoke and fire, the
inference of fire from smoke is impossible. Sense perception is thus the cause (or more
precisely, one of the causes) of inference, and if the cause is absent, the effect does not
arise. In other words, without perception one cannot grasp the invariable relation
between smoke and fire, and without this grasping, inference cannot arise (TUS 65.1–5).19

The impossibility of grasping the invariable relation

As is clear from the above, the core of a valid inference is the true determination of an
invariable relation or concomitance (vyāpti, avinābhāva) between two types of things, and
such a relation should hold good, of course, for all places and all times. It is, therefore, the
very possibility of knowing such a relation that is the main target of Jayarāśi’s arguments.
He further raises the question of the relata of such a relation: are they individuals,20
universals, or both? First, a relation cannot hold between two universals, such as smoke-
ness and fire-ness, because universals do not exist.21 For the same reason, a relation
cannot hold between a universal and an individual. Finally, an invariable relation cannot
be perceived between individuals, because individuals are infinite in number and a
perception extending to all individual occurrences of smoke and fire is impossible.
Perception is confined to the here and now; it cannot grasp individuals that are remote in
place, time, or nature.
However, is it necessary to perceive all individual smokes and fires to determine the
relation between smoke and fire? Couldn’t one grasp it by observing only a few
individuals (katipayavyakti; TUS 65.22f.)? In that case, Jayarāśi replies, the relation holds
good only for the observed individuals, not for others.22

Can causation establish the invariable relation?

The opponent23 attempts to secure the validity of inference by claiming that smoke is
produced by fire. The presupposition of this argument seems to be the universality of the
causal relation, that is, if one establishes that in some cases smoke is produced by fire,
one can assume that this is always the case.
Jayarāśi responds that a causal connection cannot be established and asks whether one
perceives causation by means of sense perception. The beginning of the following
discussion is again fragmentary24 because the manuscript is damaged, but Jayarāśi
probably argued that fire and smoke may arise by chance, in which case there is no causal
relation between them. As long as a universal relation is not apprehended (and, according
to Jayarāśi, it cannot be apprehended), this possibility cannot be excluded. The fact that
smoke arises connected to a specific place and time (TUS 66.15:
niyatadeśakālasambandhitva,25 i.e., not anywhere and at any time) does not preclude the
possibility that it arises accidentally or without a cause (ākasmikatva, kāraṇam vinā).
Another reason why this causal relation cannot be established is that one cannot
perceive smoke as an effect. This is the case because one does not perceive smoke
ceasing to exist; for if something does not cease to exist, it is also not produced (TUS
66.21f.). Discontinuation of existence cannot be perceived by perception, says Jayarāśi.
What would be the perception of such discontinuation? Would it consist in the perception
of smoke, the perception of another object, or a perception without an object? The
perception of smoke affirms the existence of smoke, not its cessation. The perception of
another object affirms the existence of another object, not the cessation of smoke. And a
perception without an object neither affirms nor negates anything because it is similar to
the perception (or non-perception) of a mute, blind, or deaf person (mūkā-
ndhabadhiratūlyatva).

Destruction and discontinuation of existence

The opponent claims that the discontinuation of existence (sattāviccheda) is nothing but
destruction (pradhvaṃsa),26 for discontinuation is brought about by destruction. Thus, by
means of the perception of destruction one perceives that an object has ceased to exist.
Jayarāśi retorts that destruction is an altogether different object of cognition
(viṣayāntara). The cognition of destruction affirms destruction and does not negate
anything (not even the existence of a particular object). If the opponent claims that
because smoke does not appear in the cognition of the destruction (pradhvaṃsa) of
smoke, smoke is annihilated; thus the whole world would be annihilated because it also
does not appear in the cognition of the destruction of smoke (dhūmapradhvaṃsa-jñāna)
(TUS 67.15–24).
The distinction between discontinuation of existence (sattāviccheda) and destruction
(pradhvaṃsa) may seem at first sight to be a gratuitous play on words, but it becomes
meaningful when one considers the historical background of these arguments. The nature
of destruction or perishing has been debated for centuries in the Indian philosophical
tradition, mainly between two groups of protagonists, the Nyāya-Vaiśeṣikas and the
Buddhists.27 The Buddhists argued that all entities are of a perishable nature, and since
this is their nature, their perishing does not depend on any external causes; rather it is
immanent and arises spontaneously from the entities themselves, just like heat in fire, or
the fact of being made of metal for a metallic object. Conversely, whatever depends on
other causes is not necessary, for example, the color in a certain piece of cloth. Thus, if
perishing were not necessary, it would be possible for something to be produced and not
to perish. Consequently, things arise from their causes as perishable by nature. Thus, all
things only exist for a moment. Further, if something does not have a perishable nature,
nothing could destroy it, for example, one cannot destroy the heat in fire (or if it is
destroyed, fire itself is also destroyed). Perishing is thus not different from the thing itself
and does not arise from other causes than the causes of the thing itself. If it were
different, it would arise from other causes, and the thing would be perceived even when
the perishing occurs, for something is not destroyed because something else (namely,
perishing) arises. Further, one says “the pot is destroyed”; if, however, perishing were
different from the thing itself, one would have to say “perishing has arisen.” In that case
perishing would not be relevant to the pot itself. Therefore, perishing (vināśa) is nothing
but the own nature of things (bhāvasvabhāva). The Naiyāyikas retort that this theory
implies that perishing is identical with existence, which is absurd. Either the thing exists
for more than a single moment, in which case it is not momentary, or it exists for only one
moment, in which case the perishing occurring while the thing exists has to happen just
as the thing arises. Thus, the thing would arise and perish at the same time. If, however,
the thing exists for one moment and the perishing occurs in the next, existence and
perishing would not be the same because of their difference in time. The Buddhists may
object that this is not a fair representation of their doctrine. They do not claim that
existence/arising is its own nonexistence (i.e., existence and perishing are not identical),
but that the second moment is the nonexistence of the first moment. The Vaiśeṣika reply is
that this is without substance because although the previous and posterior moments are
different individuals, they are not contradictory in nature. Thus, it would be possible that
the two exist side-by-side, just like two pots:28 the existence of one pot does not imply that
the other has perished.29
Jayarāśi, for his part, seems to turn the tables on the Naiyāyikas. Can they really claim
that there is a contradiction between smoke and its destruction? Could one not say that
the alleged cognition of the destruction of smoke is simply the cognition of another object,
which cannot be related to a negation of the existence of smoke? Technically speaking,
the destruction of a substance, say of a pot by a hammer, is explained by Nyāya-Vaiśeṣikas
as follows: the conjunction between the hammer and the pot brings about movement in
the parts of the pot, which causes a disjunction among the parts of the pot, which brings
about the destruction of the pot. The fact that a substance (smoke, a pot, etc.) is
considered to be a continuous object does not solve the difficulty: no matter how long a
substance exists, when it is destroyed it is no longer there, and the moment of “its”
absence/destruction cannot be related to any moment of its existence. If it were related,
the same substance would exist and not exist. Thus, the argument that the Naiyāyikas
raise against the Buddhist tenet actually applies to their own position.

Negation and contradiction of existence

The Naiyāyika whose arguments are presented in the NK explains the absence (abhāva) of
a thing as having the nature of the negation (pratiṣedha) of that thing because the
relevant cognition arises, for example, a cognition with the content “the pot does not
exist.” When this absence arises, the pot ceases to exist, its continuation is the non-
continuation of the pot, and its perception is the non-perception of the pot. There is a
mutual contradiction/opposition between the two (paraspara-virodha).30
Just as the proponent in the NK, the opponent in the TUS suggests that there is a
contradiction or opposition (virodha)31 between smoke and the destruction of smoke.
Jayarāśi proceeds to analyze what this opposition might consist in. What is meant by this
alleged opposition? Is it the fact of having different forms (atadākāratā), the fact of not
belonging to the same time (asamānakālīnatā), the fact of one not being perceived when
the other is present, the fact of being produced by the other (tajjanyatā), the fact of
producing the other (tajjanakatva), being the agent of a different action
(bhinnakriyākartṛtva),32 the fact of arising from different causes (bhinnahetūtpādyatva), or
the fact of being supported by a certain substratum and not being supported by it
(āśritānāśritatva). In the following passage, all of these alternatives are considered one by
one and none proves to be satisfactory (TUS 67.25–70.5). I will summarize here only
Jayarāśi’s arguments concerning the third of these alternatives, a passage that is typical
for Jayarāśi’s bantering and teasing manner of arguing.
The opponent claims that an opposition between two things means that the one is not
perceived when the other is present. However, non-perception of something may be due
to many reasons, such as distance, lack of light, and so on, and not to its destruction. Non-
perception of something does not imply that it has been destroyed. The opponent objects
that the cases pointed out by Jayarāśi are different from what he has in mind; when
something is not perceived due to lack of light, and so on, it may be perceived again when
the conditions have changed, but when something has been destroyed, it is never
perceived again. But by whom, Jayarāśi asks, is it never perceived again? By a particular
perceiver or by everybody? If it is never perceived by a particular perceiver, this is not
correct. There are many reasons why a particular perceiver may not perceive an object.
And it happens that certain perceivers do not perceive an object, whereas others do. This
does not mean that the object does not exist. If the opponent claims that the object is not
perceived again by any perceiver, this is also not correct, because it is impossible to know
the minds of all people and ascertain that they have not perceived the relevant object.
Furthermore, when something has been destroyed, is it supposed to be not perceived in a
specific place and a specific time or at all places and all times? The former alternative
does not imply that the object has been destroyed, and the latter cannot be known (TUS
68.10–69.2).

Further reasons for non-ascertainment of causal relationships

For the following reason too, says Jayarāśi, a causal relation cannot be determined: One
would have to perceive the effect immediately after perceiving the cause. However, things
are not produced from a single cause, but from a causal complex (kāraṇa-sāmagrī), and a
causal complex (i.e., the collection of all factors that are involved in the production of
something) in its entirety is beyond the realm of the senses (atīndriyatva) (TUS 70.7–13).
Furthermore, when one says that a piece of cloth is produced from the threads, is it
because the cloth and the threads arise one after the other or because they are perceived
one after the other? The mere fact that two things arise one after the other does not
establish a causal connection between them. For instance, one could not claim that a
certain piece of cloth (ambara)33 is the effect of certain threads and not of many other
things that arise at the same time and are not considered to be produced by the threads.
The opponent makes his position more precise: The effect has as its cause something that
arises prior to it and is connected with positive and negative concomitance to it (TUS
70.19: anvayavyatirekayuktasya yasya pūrvakālabhāvitvaṃ tatkāraṇam tad asti). Positive
and negative concomitance (anvaya and vyatireka) between two things, X and Y, means:
Where X is present, Y is present, and (by modus tollens) where Y is absent, X is absent. In
the same manner, when the effect is present, the cause is present, and the cause is absent
when the effect is absent. Here too we find a close parallel in the discussion in the NK34:
“However, accepting only the things of which we perceive the capacity [to produce certain
effects] by positive and negative concomitance with that [effect] as being a causal
complex in respect to it, we do not deserve [your] criticism.”
Next, Jayarāśi questions the meaning of the expression “(connected with) positive and
negative concomitance.” Does it convey a cause or something else? If it conveys a cause,
this is not objectionable, but we have a tautology (TUS 70.22: na kiṃcid bādhyate):
something is a cause because it has the nature of a cause. If it expresses something else,
we do not know what it is (na jānīmaḥ kiṃ tad iti).35 Further, without the corresponding
cognition, one cannot know whether two things arise at the same time, one after the
other, or not at all. One might consider establishing that two things are produced one
after the other because one perceives them in succession. However, one cannot conclude
that two things arise one after the other just because they are perceived one after the
other, for one cannot establish that something is missing (or has not yet arisen) by the fact
that it is not apprehended or by an apprehension of something else. Further, even two
things that arise at the same time may be perceived one after the other, like the two horns
of a bull.36 Similarly, two eternal things (which do not arise at all), such as the universals
“bovinity” and “existence” (gotvāstitva), may also be perceived one after the other.
Furthermore, one can perceive the cause after the effect, as in the case when one
perceives first the cloth and then the threads. It is also possible that two things that arise
successively be perceived by means of a single act of apprehension, that is, at the same
time. Jayarāśi raises a number of other arguments against the determination of causality
that we must skip here for reasons of space.

The object of inference

A famous argument that is often used by materialists and skeptics against the validity of
inference is this: If inference has a universal for its object, it establishes what has already
been established and is thus superfluous; if it has the particular as its object, there is no
“following,” that is, no invariable concomitance between the proving property and the
property to be proved (sāmānye siddhasādhyatā, viśeṣe ’nugamābhāvaḥ).37 For instance, if
one infers from smoke fire in general or the universal “fire-ness,” the inference does not
convey anything which was not known before, because the connection of smoke with the
universal “fire-ness” (or “being fire”) was already established when one perceived the
invariable connection between smoke and fire. If, however, a particular fire is inferred
when one sees smoke somewhere, the inference is inconclusive because there is no
concomitance between smoke and a particular fire.
This argument is known to Jayarāśi, and he modifies and develops it (TUS 72.12f.).38 It
is said that the property to be proved by the inference based on causation, such as
between smoke and fire, is the fact that the effect is preceded by its cause (tatkāraṇa-
pūrvakatvaṃ kāryasyānumīyate). But what does it mean for the effect to be preceded by
its cause? Is it the connection to the existence of its own cause
(svakāraṇasattāsambandha), (its own)39 existence (sattā), or another property
(dharmāntara)? If it is the connection, this is not correct because this connection is
perceived by means of perception, and it is not appropriate to infer what has already been
perceived because the reason (which is perceived at the time of inference) would also
have to be inferred. In this connection Jayarāśi interprets the phrase “if inference has a
universal for its object, it establishes what has already been established” as follows. He
understands the word sāmānya, etymologically meaning “commonness,” quite literally as
“being common” (samānatā), and “being common” he explains as a thing (artha) being a
common object of both perception and inference (not, as the word “sāmānya” referring to
a universal is usually understood, namely, as being common to all individuals of its class,
e.g., as bovinity is common to all cows). Similarly, if existence of the effect is to be
inferred, it too is perceived by means of perception, and thus inference is superfluous. In
this case it can also be said that inference establishes what has already been established.
Alternatively, the phrase can be said to mean that something existing is established (TUS
72.21: siddha = vidyamāna). That is, if the object were common to perception, inference
would establish something existing. But neither the connection to the existence of its own
cause nor the existence (of the effect as effect) is known to exist.40
If the opponent claims that “the effect being preceded by its cause” is a different
property (different from the already discussed connection and existence), and this
property is inferred, then—inasmuch as this property resides in a perceptible locus
(upalabdhilakṣaṇaprāptādhikaraṇastha) (such as smoke)41—it has already been
apprehended by means of perception, and thus inference is superfluous. Nor is there a
cause for the non-perception of the property, such as obstruction, distance, and so on,
because the inferential sign, such as smoke, is clearly visible. In this sense it has been
said: If inference has the particular as its object (i.e., if its object has not already been
perceived by means of perception), there is no apprehension42 (viśeṣe ‘nugamābhāvaḥ).
Here “particular” is interpreted by Jayarāśi as meaning something special/particular to
inference, namely, something that is perceived only by means of inference (TUS 73.1:
viśeṣo hy anumānaikagrāhyo ‘rthaḥ). With such a thing a relation cannot be perceived,
and when a relation is not perceived, one cannot infer anything on the basis of the
inferential sign (liṅga). And if being preceded by a cause is a different quality, and it is
this quality which is inferred from the inferential sign, the cause itself, like space, God,
and so on, would not be perceived.
If one claims that the effect itself, as being preceded by a cause, is inferred, this is only
fumbling due to the weakness of the opponent’s mind (prajñāmāndyaviceṣṭita). The effect
as such has already been perceived and thus its inference is superfluous. Again, the same
criticism can be voiced, but with a significant variant (73.9): sāmānye siddhasādhanam.43
The word sāmānya is interpreted again as being a common object to perception and
inference. When such an object is being established, its establishing/proof (sādhana) is
superfluous. Besides, if the effect is also apprehended from inference, the cause would
not be apprehended by any means (neither by perception nor by inference), and in this
case one could not think or speak about time (because without relationships involving
priority or posteriority, as prominently exemplified in the causal relationship, time is
impossible?).

Inference from cause to effect

In the same manner, the inference from cause to effect is impossible (TUS 73.14–23).
According to the Naiyāyikas, the cause properly speaking is the complex of all causal
factors (kāraka-sāmagrī), and since it cannot be apprehended by means of sensory
perception, it cannot be apprehended by any other means of knowledge (inasmuch as they
all presuppose perception). Further, the causal complex changes at every moment. Thus,
by the time the inferential process has been completed, the causal complex no longer
exists. If one claims that one need not apprehend the entire causal complex in order to
infer its effect, that is, that it suffices to apprehend a part of the causal complex, that is,
some of the causal factors, in order to infer its effect, this is not correct because only a
part of the causal complex does not yet produce the effect, nor does it make known the
effect, because this is absurd.

The inference of impermanence (anityatva)

Another famous inference that is often used as an example in Indian discussions of
inference derives impermanence (anityatva) from the fact of being produced (kṛtakatva).
This inference is admitted by virtually all philosophical traditions because the assumption
that whatever is produced must, at some point, be destroyed or perish may be considered
one of the fundamental presuppositions of the Indian philosophical tradition.44 The above
inference is also disputed by Jayarāśi as being invalid. Again he raises the three
alternatives concerning the property to be proved: Impermanence of a thing is either the
effect’s connection to the existence of its own cause, its own existence, or another
property. To these he adds two more alternatives: The property to be proved might be the
destruction of sound45 or the own nature of sound. The first three alternatives are rejected
for the same reasons as above. If, however, the destruction of sound is to be inferred in
the inference of impermanence, this is not correct, because a connection of the sound to
its destruction is not apprehended. When a sound exists, its destruction does not (yet)
exist. And if it does not exist, it cannot be related to the sound. For a relation is not
possible when there is only one relatum. Furthermore, absence (of sound) cannot be an
object of cognition. Even if it were, it would not be a sign of sound.46 And even if it were a
sign of sound, it would be an absence of cognition, not an object of cognition, because it
does not exist. In the same manner, the inference of sunset from sunrise
(tapanodayāstamiti) is rejected (TUS 73.25–74.9).
I do not recall having encountered the above-mentioned inference somewhere else;
presumably it would use the sunrise as a reason to infer the sunset.47 It is not clear,
however, whether it was claimed that these two properties of the sun stand in causal
relation to each other. It is remarkable that the example (though not the implied inference
itself) also appears in the NK: “As for what was said, namely, that because absence (i.e.,
the destruction of a thing) is certain, it does not depend on other causes, this [statement]
too [is rendered] inconclusive by sunrise and sunset.”48 The example is used to show not
only that things which are certain are dependent on causes, but also that they do not
happen at once, for although sunrise and sunset are certain, they occur one after the
other.49

Inference of the soul (ātman)

So far Jayarāśi has dealt with mainly theoretical problems concerning the validity and
reliability of inference as such. He now moves on to a famous case where inference was
put to practice by the Naiyāyikas. One of the most heated controversies between
“Hindus”50 and Buddhists centers on the existence of an ātman, that is, the existence of a
permanent self or soul that appropriates for itself an infinite series of bodies, one life after
the other. The Buddhists generally denied the existence of such a soul, and in the classical
period of Indian philosophy the doctrine of self-lessness or soul-lessness (anātmavāda)
was considered, for instance, by the influential Buddhist philosopher Dharmakīrti, to be
the central doctrine taught by the Buddha. The Naiyāyikas and other philosophical
traditions attempted to prove the existence of a permanent soul by means of various
inferences, which Jayarāśi proceeds to criticize.
One of the proofs for the soul is based on the assumption that phenomena such as
pleasure, pain, awareness/consciousness, which are considered to be qualities by the
Naiyāyikas, cannot exist by themselves, but need a support or substratum to reside in.
This substratum, they say, is the soul. Thus, the relationship between the soul and
feelings, knowledge, and so on is understood as the relationship between a substance and
its qualities, the same relationship that is found in the material world between substances
such as earth, water, fire, and so forth, and certain qualities such as color, smell, flavor,
and so on.
Jayarāśi rejects this inference on the simple grounds that the soul cannot be inferred
from pleasure, pain, and so on, because their relation or connection to the soul is not
apprehended (TUS 74.11f.). What exactly is the property to be proved in the above
inference? Is it that pleasure, knowledge, and so on have a substratum, or is it that the
soul itself is recognized as a substratum, or is it the nature of awareness and so on? If it is
that they have a substratum, the soul is not recognized thereby. If, however, the soul itself
is claimed to be proved, then the inference amounts to the form: There is a soul because
of awareness (asty ātmā vijñānāt). However, this inference is not valid because the formal
requirement that the reason and the property to be proved must be properties of the same
subject51 is not fulfilled: “being awareness/cognition” and “being soul” are not properties
of the same thing. Thus the fallacy of having different loci (vyadhikaraṇa) occurs.
The third option is that the nature of awareness is inferred as something that is
supported by something else (āśrita). However, this is presumably known by means of
sensory perception. Thus, one should name another object for the inference. Jayarāśi now
turns to another critical issue. How are awareness, pleasure, and so on said to be
connected to the soul? Is it by their mere existence, by the fact that they are produced by
the soul, by the fact that they produce the soul, by the fact that they inhere in the soul, or
by being identical with the soul?
The first alternative is clearly false. If awareness, pleasure, and so on are said to belong
to the soul because of their mere existence, by the same undifferentiated reason they
could be said to belong to anything and everything, and thus everything would be
conscious or happy. If awareness is said to belong to the soul because it is produced by
the soul, then the sense of vision, light, the objects, and so on would all be conscious like
the soul, because they all produce awareness. The next alternative, namely, that
awareness produces the soul, seems gratuitous because the Naiyāyikas themselves do not
consider this to be the case. The subsequent alternative, that awareness belongs to the
soul because it inheres in the soul, is rejected because the relation of inherence does not
exist.52 Even if it were to exist, any substance would be the soul, since inherence is
common to all substances. It is indeed one of the most peculiar features of Nyāya-
Vaiśeṣika metaphysics that the relation of inherence (samavāya) is considered to be one,
not only qualitatively but also numerically.
Finally, if it is said that awareness, pleasure, and so on belong to the soul because they
arise as something that has the same nature as the soul (ātmatādātmya), the soul would
undergo a transformation, that is, it would assume different states, just as awareness and
so on do, and thus the soul would lack continuity; in this case, however, recollection,
inference, recognition, and so forth would not be possible (TUS 74.11–75.23).
Next, Jayarāśi proceeds to criticize the inference of the soul according to the Jainas
(TUS 76.24ff.), the Sāṃkhya (TUS 79.20ff.), Vedānta (TUS 81.15f.), and Mīmāṃsā
traditions (TUS 82.7ff.), before he resumes the criticism of inference with the refutation of
the Buddhist theory of inference.
The phrase sāmānye siddhasādhyatā is used again in Jayarāśi’s discussion of the
Mīmāṃsā inference of the soul, but from a slightly different perspective. According to the
Mīmāṃsā, a means of knowledge must have a new object, that is, it must be an agent of
apprehension of a hitherto unapprehended object (anadhigatārthagantṛ and similar
expressions).53 This is because the means of knowledge have their respectively
determined objects (niyataviṣayāṇi pramāṇāni), and their specific objects thus mutually
exclude each other (itaretaravyāvṛttiviśeṣaviṣayāṇi); inference does not “act
upon”/“involve” (pravartate) an object of perception and vice versa. In the light of this
theory, Jayarāśi states (82.12): viśeṣe ‘nugamābhāvaḥ. “If [inference has] a particular [as
its object], there is no [invariable] relation.” In this case he interprets the determined
particular as referring to an object that can be grasped only by means of its respectively
determined means of knowledge (niyatapramāṇagrāhyo ‘rthaḥ). In the case of such an
object the invariable relation could not be apprehended. If, however, inference also acts
on an object determined by perception, then the two means of knowledge would have a
common object (sādharaṇa, samāna), and, in that case, “if [inference has an object that] is
common, something that has already been established is established (sāmānye
siddhasādhyatā), because this object has already been apprehended by perception.”54
Jayarāśi then suggests another understanding of sāmānye siddhasādhyatā. He begins
with the phrase sāmānye siddhe sādhanam. This reading may be translated: “If the
universal [were] established (i.e., existed), there is proof (i.e., the inference).”55 But the
universal is not established, and inference is therefore impossible. Under this
interpretation Jayarāśi also interprets siddha in the phrase siddhasādhanam as something
existing (82.22: siddhasya sādhanam—vidyamānasya sādhanam). Putting the two
interpretations together, one may translate the whole phrase as “If the universal [were
established, inference would] establish/prove something existing.” However, the universal
(such as “fire-ness”) does not exist and therefore the inference does not make anything
known.
Jayarāśi suggests yet another possible understanding of the phrase: siddhasādhanam
could be understood as a karmadhāraya compound (siddhaṃ sādhanam), which has to be
translated as “an established establishment/proof,” which means an existing
establishment/proof (vidyamānaṃ sādhanam). However, the universal “smoke-ness,”
which is used in an inference to establish fire, does not exist. Therefore, a universal
cannot be a proof.56 In this sense one could translate the phrase: If [there were] a
universal, [there would have been] an establishment/proof.
Finally, Jayarāśi suggest another interpretation of the same phrase by explaining (TUS
83.1) siddhasādhanam as jñātam anumānaṃ sādhanaṃ bhavati. This interpretation may
be translated as: “a known inference (i.e., inferential sign) becomes a proof.” But a
universal such as “smoke-ness” cannot be known because it does not exist. Or even if it
existed, it could not be apprehended because there is no means to apprehend it: the
nature of a universal is to be pervasive (anusyūtaṃ rūpam), and as such it can neither be
apprehended in itself alone nor in any single individual, but only in many individuals.
However, at the time of inference one does not apprehend many individuals, but only a
single one. Thus the universal cannot be apprehended and therefore it cannot function as
a proof.

Critique of the Buddhist theory of inference

The last target of Jayarāśi’s arguments against inference are the Buddhists, and the
chapter concerned with them is by far the longest (TUS 83–109). However, most of the
chapter does not deal with inference directly, but with its specific applications to
controversial metaphysical topics; for instance, the Buddhist criticism of the whole
(avayavin), the discussion of the object of non-perception (anupalabdhi), the rejection of
an external object (bāhyārtha), and the proof of momentariness (kṣaṇikatva). These
topics, however, are clearly beyond the scope of this chapter.
Jayarāśi begins again by stating that the relation (i.e., the invariable relation between
the reason and the property to be proved) is not apprehended, and raises the same
alternatives as in his criticism of inference according to the Naiyāyikas: Does the relation
obtain between two universals, a universal and a particular (svalakṣaṇa), or two
particulars? The first two alternatives are rejected because universals do not exist. If it
obtains between two particulars, is it a relation between two particular objects or two
particular cognitions, or a particular cognition and a particular object? If it is between two
particular objects, like smoke and fire, how does the cognition of their causal relation
arise? Is it by their mere existence, by their being connected to the respective form, by
the smoke’s being produced by fire, or by a combination of these alternatives (sāmastya)?
Most of the arguments used by Jayarāśi here have already been mentioned above and
need not be repeated. Among the points of criticism that are specific to Buddhist
epistemology, however, one may mention the doctrine that every awareness bears the
form of its object. In fact, the object of awareness is defined already in Abhidharma
sources as the cause of awareness that bestows its own form upon it (ākārārpakatva or
similar expressions).57 In later periods, this assumption has become somewhat
problematic because an element of being or an entity (and awareness is an entity for the
Buddhists) can normally have only one form, whereas an awareness seems to have two
forms, its own and the form of its object. In this connection Jayarāśi quotes from an
unidentified Buddhist source: “A mirror appropriates the form58 of a face, but does not
abandon the form of a mirror; in the same manner the awareness too, when appropriating
the form of an object, does not abandon the form of an awareness.”59 Jayarāśi proceeds to
show that in fact the mirror does not appropriate the form of a face. What does
appropriation of a form mean in this case? Does it mean having the nature of a face, or
the arising at the contiguous place (aviraladeśotpāda, i.e., the arising of the mirror
immediately next to the face)? In the first case, there would be no difference between the
face and the mirror. In the second case, the form of the object would arise in proximity to
the cognition, but not in it, and thus the respective determination of cognitions according
to their objects (pratikarmavyavasthā) would not be possible.60
Many of Jayarāśi’s arguments center on specific Buddhists tenets such as
momentariness, the nature of cause, own nature/own being (svabhāva), form (ākāra/rūpa),
the nature of awareness (jñāna and equivalent expressions), self-apprehension
(svasaṃvedana), and so on. It is, therefore, impossible to summarize his arguments here
without first explaining these Buddhist doctrines themselves. Consequently I can treat
here only one example that involves, again, the dictum discussed above. According to the
Buddhists, awareness or cognition is a momentary independent entity (not related to any
substratum such as a soul) that apprehends an object as well as itself (i.e., an awareness
is always also a self-awareness; it is not apprehended by another awareness).
Presupposing this theory Jayarāśi claims that one cannot determine fire and smoke to be
cause and effect (TUS 88.24ff.). Cause and effect exist at successive moments (for the
cause must precede its effect and both are momentary); are they apprehended by a single
awareness, or by their respective awarenesses? If they are both apprehended by a single
awareness, this awareness cannot be momentary; alternatively, if the cause of smoke
continued to exist up to the moment of the effect, the cause would not be momentary. If,
however, fire and smoke are apprehended by their respective awarenesses, the awareness
of fire affirms the existence of fire, and that of smoke, the existence of smoke. In this case
the phrase sāmānye siddhasādhyatā means: With regard to a generality, that is, if the
mere existence of smoke [is apprehended], something established is being established,
that is, the awareness or apprehension of smoke would be valid but useless. However,
with regard to particularity, that is, if something different is assumed to be apprehended,
namely, that smoke and fire arise one after the other, there is no “following,” that is, there
is no activity, that is, one cannot apprehend smoke in its temporal relation to fire (viśeṣe
bhede ... ‘nugamābhāvo vyāpārābhāvaḥ).
As could be expected, the Buddhists being criticized by Jayarāśi are Dharmakīrti and
his followers. According to Dharmakīrti, only two types of relation are truly invariable and
can be used in a valid inference: the relation of the effect to its cause (but not of the cause
to its effect, which is uncertain) or the “relation”61 of identity or sameness of nature
(tādātmya). Up to this point, Jayarāśi’s arguments are primarily concerned with the
former type of relation and the corresponding type of inference. The last section in the
chapter on inference in the TUS is dedicated to the so-called svabhāva-anumāna,
inference from own nature. The standard example for this type of inference is: “This is a
tree because it is an Aśoka tree.” In this case one essential property, being a tree, is
inferred from another essential property, being Aśoka-tree.62 Both properties belong to
one and the same thing. With his first argument Jayarāśi objects precisely to this (TUS
108.5f.): A relation should occur between two distinct things; no relation is possible for a
single thing. Further, when impermanence (anityatva) is inferred from being produced
(kṛtakatva), does the former appear (in the awareness) when the latter is apprehended? If
it does not appear, how could there be identity of what appears and what does not
appear? If, however, impermanence does appear, inference becomes superfluous.
In the above I have attempted to summarize and explain as briefly as possible Jayarāśi’s
main arguments against perception and inference, which are the most important means of
knowledge. Other means of knowledge are equally criticized, notably, implication or
circumstantial evidence (arthāpatti), analogy (upamāna), absence of apprehension
(abhāva), inclusion (sambhava), tradition (aitihya), and verbal communication (śabda).63
However, with the exception of verbal communication, these means of knowledge are of
minor importance in the Indian philosophical tradition and are therefore discussed by
Jayarāśi only briefly. They are mainly rejected on the grounds that they are either not
valid or are included in inference (which has already been rejected).64

Notes
* As always, I am greatly indebted to Karin Preisendanz for her thoughtful and pertinent comments.
1. See VV 31ff.
2. The originality of this discussion cannot be ascertained. Fragments of a similar discussion are
present in the Spitzer Manuscript, which may predate Nāgārjuna. The circularity of the pramāṇas
seems to have been a standard topic in Abhidharma literature (see Franco 2004a).
3. The four means of knowledge are accepted by the Nyāya tradition and thus it is widely assumed
that here Nāgārjuna is criticizing the Naiyāyikas. This assumption is unwarranted. On the basis of
 the Spitzer Manuscript one can ascertain that certain Buddhists, probably from the Sarvāstivāda
tradition, also accepted these four means of knowledge. The same can be seen in the
*Upāyahṛdaya, an early Buddhist text on dialectics that has survived only in Chinese translation.
4. The similarity to Sextus Empiricus’s argument about the criterion of truth is remarkable and
evident. See Bury (1935, p. 179): “Those who claim for themselves to judge the truth are bound to
possess a criterion of truth. This criterion, then, either is without a judge’s approval or has been
approved. But if it is without approval, whence comes it that it is trustworthy? For no matter of
dispute is to be trusted without judging. And, if it has been approved, that which approves it, in
turn, either has been approved or has not been approved, and so on ad infinitum.”
5. “Cognition” in this chapter is used to translate terms such as “jñāna,” “vijñāna,” “dhī,” “citta,” and
similar words. It refers to a brief moment of awareness of a certain object. According to the
“Hindu” schools, the moment of awareness is a quality of the soul; according to the Buddhists, who
deny the existence of a soul, it is an independent momentary entity.
6. Causes of “disturbance” could include the senses being impaired by disease—for instance, the
timira sickness that causes one to see a double moon, and so on—the mind being disturbed by
hunger, or the object being too subtle to be perceived correctly.
7. The word “indriya” (sense) is often misunderstood and mistranslated as “sense organ”; however, it
always refers to the sense faculties, or sense capacities, not to the sense organs (therefore “the
sense of sight” and not “the eye”). When the Naiyāyikas say that senses are not perceptible by the
senses themselves, they do not mean, of course, that the eyes and ears are imperceptible, but that
the senses of sight or of hearing are so.
8. Although it may seem that the absence of contradiction presupposes a coherence theory of truth,
such an assumption would be wrong. To my knowledge, Indian philosophers always presuppose a
correspondence theory of truth.
9. In spite of the different context, the similarity between Jayarāśi’s statement and Popper’s principle
that scientific hypotheses or theories can be refuted but not verified is quite clear.
10. The criterion also functions in a negative way, that is, one may be successful in avoiding or getting
rid of a perceived object.
11. This criterion seems to presuppose a hierarchy in the reliability of the senses; while seeing water
may be deceptive, tasting it seems to be considered a much more reliable experience. In general,
sensual errors are discussed only with respect to the sense of sight, vision being almost always
taken as the model for the senses. Examples of errors concerning hearing, tasting, and so on are
extremely rare.
12. Another interpretation of this passage was proposed by Walter Ruben: How does one know that
the efficiency of an activity is a criterion of truth? One needs a criterion in order to choose a
criterion; the second criterion needs a third one, and so on.
13. See James (1909, chapter 2: “The Tigers of India”).
14. See Chisholm (1977, p. 97).
15. That is, validity is immanent to a cognition, while falsity comes from external factors.
16. Sāṃkhya in Jayarāśi’s time was no longer a force to be reckoned with in Indian philosophy.
17. An excellent discussion of this theory can be found in Taber (1992).
18. The invariable relation does not have to be symmetric. It is generally acknowledged in Indian
philosophical works that there is no smoke without fire, but that there is fire without smoke.
19. In this connection the question whether in the Indian philosophical tradition inference is
deductive or inductive has to be addressed. Unfortunately, a great deal of what has been written
on the subject is useless, because there is no such thing as the “Indian inference.” Indeed, it would
be surprising if all Indian philosophers were of a single opinion. Conceptions of inference differ
even within the same tradition. Thus, for the Buddhist Dignāga inferences can be said to be
inductive, whereas for Dharmakīrti, who developed Dignāga’s ideas further, they are deductive. In
emic terms the question was phrased whether the subject of inference (pakṣa) is part of the group
of things that possess the property to be proved (sapakṣa) or whether it is excluded from it; see
Tillemans (1999) and Franco (2004).
20. The term used by Jayarāśi, “svalakṣaṇa” (own characteristic), is originally a specifically Buddhist
term. It is interesting that for Jayarāśi it has lost its Buddhist connotations and is used as a term
for “individual” in general; in this sense it is interchangeable with the term “vyakti” used in the
Mīmāṃsā and Nyāya-Vaiśeṣika traditions.
21. Universals according to Nyāya-Vaiśeṣika are independent eternal entities that are inherent to all
individuals of a certain class and the cause for the recurrent cognition of different individuals as
belonging to this same class, for example, the universal “bovinity” (gotva) is present in all cows
and the cause for our recognition of all individual cows as also being cows; the universal “pot-
ness” is present in all pots and causes us to recognize individuals of various shapes, sizes, and
colors as pots, and so on. The existence of universals is refuted in some detail in TUS 4.5–7.11.
22. Unfortunately the palm-leaf manuscript is broken at this point and the further development of the
argument is thus missing. Note that Jayarāśi’s important work survives only in a single manuscript
dating from the end of the thirteenth century.
23. Philosophical works in Classical India were often written in the form of a dialogue between
opponent and proponent. The role of the opponent is often reduced to raising objections against
 the positions and arguments of the proponent. In some cases the opponent is fictitious, in others a
historical person whose arguments are quoted or paraphrased; sometimes it is difficult to know
whether the objections raised by an opponent in a given work correspond to actual arguments
adduced in other works, or whether the author is only mentioning possible objections as he
develops his theory. The TUS uses a large number of arguments of real opponents, a number of
whom have been identified in Franco (1987), which deals with the section on sense perception.
Comparable research for the section dealing with inference remains to be done. Note, however,
that there are close parallels between the discussion here and the one in the Vaiśeṣika work
Nyāyakandalī (NK). Although Śrīdhara, the author of the NK, lived two centuries later than
Jayarāśi, the materials he summarizes are clearly of a much earlier date. See the discussion below.
24. See n. 23 above.
25. This expression evokes the beginning of the Viṃśatikā, where the Buddhist philosopher
Vasubandhu points out that objects in a dream, although unreal, also arise related to a specific
place and time.
26. TUS 67.15: atha sattāvicchedo hi pradhvaṃsaḥ, so ‘nenātmasāt kriyate (The discontinuation of
existence is indeed destruction, for it absorbs it).
27. A lively summary of this debate can be found in Shastri (1964, chapter VI “Defence of Substance,”
pp. 154–234, esp. pp. 203ff.). My presentation here follows the exposition in NK 188ff., which
seems to have close affinities with the position of Jayarāśi’s Nyāya opponent.
28. More precisely, two pots that belong to different series. As all things are momentary, the
Buddhists interpret the continuation of objects like pots as a series of quick successions of similar
momentary atoms.
29. NK 198.3–7 (Dvivedin’s ed. pp. 78–79): [bhāvasya] ekakṣaṇāvasthāyitve tu tasya kṣaṇāntare
sthityabhāva iti na bhāvābhāvayor ekatvam kālabhedāt. atha matam: na brūmo bhāvaḥ
svasyaivābhāvaḥ kintu dvitīyakṣaṇaḥ pūrvakṣaṇasyābhāva iti tad apy asāraṃ pūrvāparakṣaṇayor
vyaktibhede ’pi svarūpasya virodhābhāvāt. yathā ghaṭo bhinnasantativartinā ghaṭāntareṇa saha
tiṣṭhaty evam ekasantānavartināpi saha tiṣṭhet, dvitīyakṣaṇagrāhipramāṇāntarasya
tatsvarūpavidheś caritārthasya prathamakṣaṇaniṣedhe pramāṇatvābhāvāt.
30. See NK 198.7–199.3: abhāvas tu bhāvapratiṣedhātmaiva, ghaṭo nāstīti pratītyudayāt. tatas
tasyotpattir bhāvasya nivṛttiḥ, tasyāvasthānaṃ bhāvasyānavasthitiḥ, tasyopalambho
bhāvasyānupalambha iti yuktaṃ parasparavirodhāt.
31. This term is often translated as “contradiction,” but here it refers to an ontological opposition.
32. This compound is ambiguous and also allows other interpretations, namely, “having different
actions and agents” or “being different actions and agents.”
33. Radhakrishnan (1971, p. 242), where this passage is translated, takes this word in its well-known
meaning of “ether,” but in this context the equally attested meaning “cloth” seems more probable
to me.
34. NK 195.6–7: vayaṃ tu yatra yeṣām anvayavyatirekābhyāṃ sāmarthyam avagacchāmas, tatra
teṣām eva sāmagrībhāvam abhyupagacchanto nopālambham arhāmāḥ.
35. That is, the opponent fails to account what it is.
36. A similar argument is raised by Dharmakīrti’s commentator Prajñākaragupta; see PVA 69.4:
yasyopalabdhiḥ prathamaṃ tat tasya yadi kāraṇam |
na khalāntargataṃ bījaṃ hetuḥ syād aṅkurodaye || 439
37. For various quotations of this verse in Jaina, Mīmāṃsā, Nyāya, and Buddhist sources, see Franco
(1987, p. 398).
38. In doing so he plays on the various meanings of the word “sāmānya,” which means a universal,
something common, a common generic property, something general, and so on. He also interprets
the locative case in various ways, sometimes as locative absolute (expressing both real and
counter-real conditions), sometimes as locative of reference (“in respect to,” etc.).
39. Radhakrishnan (1971, p. 245), however: “the existence [of the causal priority].”
40. Radhakrishnan (1971, p. 245, n. 2), however: “neither what is already known nor what already
exists.”
41. Since the property in question is the effect’s “being preceded by its cause,” it cannot reside in the
subject of inference, for example, the mountain, but in the inferential sign itself (e.g., smoke).
42. Here Jayarāśi seems to interpret anugama as avagama; Radhakrishnan (1971, p. 245) interprets
anugama here as perception: “In the case of unique particularity, there is no perception.”
43. I am not sure whether Jayarāśi knew the above-quoted phrase in two variants (-sādhananam and -
sādhyatā, both are metrically possible) or whether he simply replaces the latter with the former to
make his understanding of it clear.
44. The Judeo-Christian notion of a soul, the world, or matter as having been created at some point in
time and continuing to exist forever would seem very odd to Indian philosophers. Exception is
made only for negative entities. Absence/inexistence of something because it is destroyed begins
at the moment of destruction and continues forever.
45. Sound is the subject of a frequently discussed inference which takes the form: Sound is
impermanent, because it is produced, like a pot.
46. The inferential sign in the above-mentioned inference is “being produced” (kṛtakatva). It seems
 therefore that the word “liṅga” is not used here in the meaning of “inferential sign,” but as an
equivalent to “lakṣaṇa” (characteristic mark).
47. Or the other way round, but this seems less likely in this context, for the destruction of sound has
to be inferred.
48. NK 202.1–2: yac ca dhruvabhāvitvād abhāvasya hetvantarānapekṣeti uktam, tad api savitur
udayāstamayābhyām anaikāntikam.
49. See the Pañjikā on the passage quoted in the previous note: savitur udayāstamayau
dhruvabhāvinau krameṇa ca bhavataḥ.
50. The use of the term “Hinduism” when applied to the classical period before the tenth century is
problematic inasmuch as it suggests a continuity and uniformity to the present day.
51. For instance, the property possessing smoke and the property possessing fire must belong the
same locus (or the same property possessor), otherwise one could infer that this mountain has fire
because that (other) mountain has smoke.
52. This relation is examined and rejected in a separate section of the TUS (7.12ff.).
53. TUS 82.7–8: pramāṇāntarānavadhāritārthaviṣayatvāt pramāṇānām “[The inference of the soul is
not valid] because means of knowledge have as their objects things that have not been [previously]
determined by another means of knowledge.”
54. For a discussion of petitio principii in the Mīmāṃsā theory of inference, see Bhatt (1989, pp. 251–
254); for the Mīmāṃsā inference of the soul, id., pp. 382–404.
55. For according to the Mīmāṃsakas the relation between reason and the property to be proved
obtains between universals.
56. Under this interpretation, it would be possible to read: sāmānye ‘siddhasādhanam. “If/when a
universal [is supposed to be that by means of which one infers], the proof does not exist.”
57. On this definition of object, see Franco (1987, p. 426f.), and further references therein.
58. Literally, “makes own having the form.”
59. TUS 85.18–20: ādarśamaṇḍalaṃ mukharūpatāṃ svīkaroti na cādarśarūpatāṃ parityajati, tathā
jñānam api viṣayarūpatāṃ svīkurvan na vijñānarūpatāṃ tyajati.
60. I thank Phyllis Granoff for the interpretation of this argument.
61. Whereas the causal relation obtains between two things, such as smoke and fire, identity or
sameness of nature refers to two properties of the same thing; it is therefore not an ontological
relation, but only a conceptual one.
62. Although it may appear so at first sight, this type of inference is not simply based on inclusion;
inferences such as “sound is impermanent because it is produced,” or “sound is momentary
because it exists,” are further examples for this type of inference. Of course, in a certain sense, all
valid inferences can be said to involve inclusion in the sense that fire can be said to “include”
smoke.
63. See Mohanta 2009.
64. This chapter was written before Esther Solomon’s translation, Jayarāśi Bhaṭṭa’s
Tattvopaplavasimha (An Introduction, Sanskrit Text, English Translation & Notes), was
posthumously published in 2010. However, I do not consider that my summary and interpretation
of Jayarāśi’s arguments need to be revised in view of this publication. It is unfortunate that
Solomon’s translation, which was obviously not considered by the author herself to be ready for
publication, was printed without any editorial revisions. A more recent translation by V. N. Jha
(Ernakulam 2013) is not up to the usual scholarly standards.

Abbreviations and bibliography

Bhatt, G. P. (1989), The Basic Ways of Knowing. Delhi: Motilal Banarsidass.
Chisholm, R. (1977, 2nd ed.), A Theory of Knowledge. Englewood Cliffs, NJ: Prentice Hall.
Franco, E. (1987), Perception, Knowledge, and Disbelief. Stuttgart: Franz Steiner (repr. 1994, Delhi:
Motilal Banarsidass).
Franco, E. (2004), “Xuanzang’s proof of idealism (vijñaptimātratā).” Horin, 11, 199–212.
Franco, E. (2004a), “A Note on Nāgārjuna and the Naiyāyikas,” in S. Hino, T. Wada, Three Mountains
and Seven Rivers: Prof. Musashi Tachikawa’s Felicitation Volume, Delhi 2004, 203–208.
James, W. (1909, repr. 1979), The Meaning of Truth. Cambridge, MA and London: Harvard University
Press.
Mohanta, D. (2009), Studies in Jayarāśi Bhaṭṭa’s Critique of Knowing from Words. Kolkota: The Asiatic
Society.
NK: Nyāyakandalī. Being a commentary on Praśastapādabhāṣya, with three sub-commentaries
(Ṭippaṇa of Naracandrasūri, Pañjikā of Rājaśekharasūri, Kusumodgama of Śiḍila). Ed. J. S. Jetly and
V. G. Parikh (1991), Baroda: Oriental Institute. (Another edition of the Praśastapādabhāṣya and
Nyāyakandalī, edited by V. P. Dvivedin (1984, 2nd ed.), Delhi: Sri Satguru.)
NS: Ruben, W. (1928), Die Nyāyasūtras. Text, Übersetzung, Erläuterung und Glossar. Leipzig:
Abhandlungen für die Kunde des Morgenlandes, XVIII. Band, No 2.
PVA: Pramāṇavārtikabhāshyam or Vārtikālaṅkāraḥof Prajñākaragupta. Ed. R. Sāṅkṛtyāyana (1953),
Patna: Kashi Prasad Jayaswal Research Institute.
Radhakrishnan, S., and Moore, C. A. (1957), A Source Book in Indian Philosophy. Princeton: Princeton
University Press.
Sextus Empiricus, Against the Logicians. Trans. R. G. Bury (1935). London: W. Heinemann.
Shastri, D. N. (1964, repr. 1976), The Philosophy of Nyāya-Vaiśeṣika and Its Conflict with the Buddhist
Dignāga School (Critique of Indian Realism). Delhi/Varanasi: Bharatiya Vidya Prakashan.
Taber, J. (1992), “What did Kumārila Bhaṭṭa mean by svataḥprāmāṇya?” Journal of the American
Oriental Society, 112, 204–221.
Tillemans, T. J. F. (1999), “On sapakṣa,” in his Scripture, Logic, Language: Essays on Dharmakīrti and
His Tibetan Successors. Boston: Wisdom Publications, pp. 89–116.
TUS: Tattvopaplavasiṃha of Jayarāśi Bhaṭṭa, edited by S. Sanghavi and R. C. Parikh (1940), Baroda:
Gaekwad’s Oriental Series. (TUS 1–64 are translated and commented upon in Franco [1987].)
VV: The Dialectical Method of Nāgārjuna: Vigrahavyāvartanī, translated by K. Bhattacharya, edited by
E. H. Johnston and A. Kunst, repr. (1978, repr. 2002), Motilal Banarsidass, Delhi. (See also Yonezawa,
Y. [2008], “Vigrahavyāvartanī Sanskrit transliteration and Tibetan translation.” Journal of the
Naritasan Institute for Buddhist Studies, 31, 209–333.)
 4
Jainism: Disambiguate the Ambiguous*
Piotr Balcerowicz

Ceci n’est pas une pipe, c’est une devise:

Half a bee, philosophically,

Must ipso facto half not be.
But half a bee, has got to be,
Vis a vis its entity.
—D’you see?
But can a bee be said to be
Or not to be an entire bee,
When half the bee is not a bee,
Due to some ancient injury.1

How can we communicate in a language which is by nature imprecise and ambiguous? For
instance, what color is the honey bee, in one word? Black? Yellow? Imagine a simple quiz,
or crossword, when one is required to fill in the color and has a limited space for it. If I
write “black,” do I lie, while neglecting the yellow stripes? If I answer “yellow,” is my
response false which ignores the black parts? Would it not be unnatural, one could argue,
to impose such limitations on our language? Why should one do it in the first place? But
isn’t this precisely what we always do with any natural language? Suppose we have a
series of items on display: honey bees, charcoal, white storks, and a flag of Afghanistan,
and we are asked: “Which of these is yellow”? Should the correct reply be “the honey
bees” or “none”? Or was the question formulated incorrectly, being imprecise? Which
conditions should the question meet in order to be precise so that one could provide a
correct reply? In ordinary language, however, both the question “Which is yellow?” and
the reply “The honey bees are yellow” would be well understood, whereas a precise
question, for instance, “Which of these items which combine black parts of their
exoskeleton, substance, plumage, and fabric with some other parts featuring other colors,
except for charcoal which does not combine black with any other color, contains yellow
parts?,” would be considered awkward, to say the least. Similarly, a precise reply of the
sort “No item is yellow, except for honey bees, which display some yellow color in some of
their stripes, though the fact is that the black is found primarily in the exoskeleton, with
yellow pigments dominating in the hair, color patterns depending on the combination 7
pairs of genes which can be responsible for 2,181 different genotypes,” would similarly
sound outlandish. Despite being seemingly precise, the reply would violate most of the
Gricean conversational maxims, such as the maxim of quantity (speaker’s contribution
should be as informative as required), relevance (speaker’s contribution should relate to
the purpose of exchange and avoid unnecessary digressions), and manner (speaker’s
contribution should avoid obscurity and ambiguity).
What is the honey bee? It is defined as a specimen of a subset of the bees of the genus
Apis and family Apidae which are identified by the intense production of honey, so a honey
bee is a bee which produces honey. Suppose a honey bee does not produce honey at all
but otherwise looks and behaves like all honey bees; is it a honey bee or not, or a partial
honey bee, that is, half a bee? Now suppose we see a honey bee flying with the collected
nectar and pollen back to the hive. What would be the truth value of an explanation given
by a parent to the child: “Look, the bee is making honey”? Obviously, the bee is not
making honey at this precise moment but is performing an action which is in a series of
actions which jointly constitute the process of making honey.
All the above, which may sound at first either trivial or silly, points to something
extremely important, namely, that in our everyday communication, but in fact also when it
comes to the usage of strictly formalized language, we heavily rely on the context of the
speech act, which by definition provides just a fraction of information, whereas most of it
is supplied by the context. It is neither possible nor necessary to provide all information in
a speech act, a sentence, which is by definition ambiguous. It is ambiguous because it is
necessarily incomplete for, in a natural language, all utterances stand in need of
additional analysis which has to take into account other factors such as context, as in the
exchange: “Could you tell me what time it is?” (which is formally a yes/no question!)
—“The supermarket is closed.”—“But my brother has his birthday tonight!”—“There’s a
petrol station on the corner.” These seemingly unrelated sentences shouldn’t make any
sense, but they apparently do. Why? How do we manage to convey the meanings, ideas,
thoughts, and description, if we do it with incomplete and ambiguous sentences? How is it
possible that ambiguous does not mean indefinite? These questions, including the process
of decoding from utterance → proposition → meaning, concern what has been a matter of
interest for pragmatic logic in the West over a few decades. But precisely the same
questions, addressed in India centuries before Western philosophers did that, led to the
development of the doctrine of multiplexity of reality (anekānta-vāda) by the followers of
Jainism which matured in the period between fifth and seventh centuries CE, but which
had its roots two and a half millennia back.
The term itself, anekānta-vāda, came to be as closely associated with Jainism as, for
instance, kṣaṇika-vāda, or the doctrine of momentariness, with Buddhism, a religious and
philosophical rival of Jainism of roughly similar antiquity. It literally means that the world,
and every entity in it, is by nature “of not one (an-eka) aspect (anta) or facet.” In other
words, the “cut-gem like” world of infinite facets, and everything that exists, is a complex
structure, both internally (being a whole consisting of infinite facets amenable to infinite
modes of expression) and externally (being related to all other entities through infinite
relations). For this reason it is not possible to provide a full, that is, exhaustive, account of
anything, and what we do with every speech act and proposition, also reflected in
thought, is to provide just a tiny slice of the item we refer to. We select only those aspects
of it which we consider relevant to our purposes. Like ourselves, the recipients of our
communications have at their disposal an effective method or strategy to relate the partial
message to reality in a meaningful manner so that they understand the meaning. The
Jaina doctrine of multiplexity of reality has therefore two basic aspects, ontological and
semantic-logical.
Both structurally, in the sense how the theory is structurally and logically constructed,
and historically, how and why it really began, ontology has priority over epistemology and
semantics. Historically speaking, the Jaina theory seems to be a dialectical interplay
between two basic standpoints concerning the nature of the world and all phenomena.
One classical extreme worldview in ancient India was eternalism (śāśvata-vāda), which
postulates that all that really exists has to be eternal, permanent, unchanging, such as the
views professed by the Sāṁkhya school. All phenomenal world is potentially contained, or
dormant, in the unmanifest primordial factor (prakṛti), which endures and emanates itself
through a quasi-evolution, a view which would resemble that of Parmenides of Elea: what
really exists cannot cease to exists, and nothing which does not exist cannot come into
being. The other extreme worldview was that of annihilationism (uccheda-vāda) with its
claim that all that exists necessarily exists just for a short while, which for the Buddhists
and Jainas was the perspective adopted by the materialists. However, for the Jainas, an
instantiation of annihilationism would also be Buddhism with its doctrine of
momentariness (kṣaṇika-vāda), or of transitoriness, according to which all that exists
appears just for a moment in order to disappear the very next moment and produce its
semblance. This position would resemble that of Heraclitus of Ephesus, who posited that
the whole world and its phenomena are in a perpetual flux (πάντα ῥεῖ, “everything
flows”).
For the Jainas it was originally a problem of change and identity. If everything were
permanent, how could we account for the change and process we perceive all around us?
How could a living being be a living entity, if to live entails to undergo change and
progress through a process? The living being per se, either called “soul” (jīva) or “agent”
(ātman), consists in thought, cognition, and experience, which by definition are processes,
and a process cannot exist without change and transformation. An absolutely permanent,
in the sense of eternal and changeless, living being would be dead, a contradiction in
terms. However, how can one preserve its identity, if it undergoes a constant process of
change? Identity seems to contradict change. What does it mean to say that Theodore
(Devadatta) we know now is the same person we met fifteen years ago, if he has been in
the constant process of change? The present Theodore must therefore be a different
entity than the erstwhile Theodore, and hence he cannot share his identity with his earlier
alter ego. What is it that remains unchanging in any entity by the force of which we can
identify it as the same thing? The system of Vaiśeṣika, or the school of the philosophy of
nature, postulated a special kind of additional entity, or a category, known as individuator
(viśeṣa) which would attach to every individual thing to make it singular, identifiable, and
self-same. For the Jainas that was a device which would multiply entities beyond necessity,
and in addition was not attested to empirically.
The Jainas realistically considered that, without any need to take recourse to such
imperceptible and unaccountable entities as individuators, the reality combines both
natures, the permanent and transitory. In fact, they distinguished four interrelated layers,
or aspects, in every individual entity. The aspect which was responsible for the
persistence in time and identifiability of an individual entity was called substance
(dravya). Substances could not exist alone: they are necessarily accompanied by, and
distinguishable from each other through, a set of qualities (guṇa) they possess. These
were succinctly defined as follows: “The attributes through which a substance, such as
the soul (jīva) and the non-soul (ajīva), can be recognized are known as sensorily
perceptible and imperceptible qualities (guṇa), which are characterized by essence
different from the substance” (PSā 2.28). Qualities accompany substances in a quasi-
permanent manner: for instance, a particular substance of earth is characterized and
individuated by a particular set of qualities such as color, taste, smell, and touch. These
may gradually transform, but will never disappear: earth will always be possessed of such
a set of qualities. While substances were to explain the preservation of identity of entities,
qualities were conceived of as their distinguishing properties for, theoretically speaking,
one thing is to identify an entity as itself, but another to distinguish it from other entities.
This theory was clearly an extension of the Vaiśeṣika categorization of entities, the first
two (out of classical six) being substances (dravya) and qualities (guṇa). The Jainas added
another ontic layer which was to explain the change. These were modes (paryāya), or
discernible modifications, which qualified both substances and qualities. A change in color
or shade was expressed through modes, for example, one mode (darker shade of blue)
replacing another. In fact, both substances and their qualities were thought of to undergo
incessant process of transformation and change, and precisely that aspect was verbalized
as modes. Both qualities and modes could not be suspended in the air: they had to be
located in or related to a particular substratum, that is, in their substance. While modes
were that layer of continuously changing aspect of a thing (or event, phenomenon, etc.)
possessed of its two other aspects, that is, substance and qualities, which was observable
and nameable, things also possessed a changing aspect which was sometimes palpable,
but often hardly discernible in a moment-to-moment observation. Only a longer period of
time would allow one to state that a minor change has taken place, but even then the
transformations were so microscopic and almost undetectable that there was no adequate
vocabulary to name them. These were called ineffable, transient occurrences (vivarta,
vartanā), which were frequently neglected in Jaina expositions of the theory, and even
more so by modern researchers. In other words, modifications and transformations, be
they rapid and insignificant, as long as we could consciously observe them and find a
linguistic description for them, would be classified as modes. Those which would not enter
the threshold of such discernibility and nameability (hence their actual existence, even
though amenable to nonconceptual perception, would be either beyond our conceptual
judgment or a matter of a post-factum inference) would be termed “transient
occurrences.” Every phenomenon was therefore considered a complex whole consisting of
four such ontic layers, two of which would account for permanence and preservation of its
identity (substances and qualities), and in these aspects a thing was immutable and self-
same, and two would explain continuous change and transformation (modes and transient
occurrences). This is how the Jainas took their own “middle path” vis-à-vis the ontic
character of every entity, which—unlike the Buddhist middle way of avoiding two
extremes—consisted in a combination of such seemingly incompatible views of eternalism
and nihilism. However, to say that one and the same thing is both permanent, hence
immutable, and changing, ergo mutable, would seem to lead to obvious contradictions. A
method to reconcile such standpoints in a seamless way was therefore called for.
On top of that, the Jainas maintained that all “that is existent is furnished with
origination, annihilation and permanence” (TS 5.29), in other words, “origination,
continuity and destruction take place in the world, consisting of souls and matter, by way
of transformation due to combination and separation” (PSā 2.37). The continuous
transformation was an inalienable feature of everything that existed, for “there is no
object without transformation” (PSā 1.10), but it also implied that through it entities
would conjoin to produce larger structures or disjoin, an element of relation (involved in
both processes of conjoining and disjoining) being most crucial, for the Jainas believed
that all things are directly or indirectly related to each other. These three coexisting and
simultaneous aspects of existence, such as origination (utpāda, udaya), continued
existence (sthiti, dhrauvya) and cessation, or disintegration (bhaṅga, vyaya, apavarga),
were considered necessary to explain the process of change. Again, also this assumption
of three aspects of existence seems to have been a response to the Buddhist theories
developed primarily by the schools of Abhidharma and Sarvāsti-vāda which postulated
either three or four, respectively, conditioned factors, known as “markers” (saṁskṛta-
lakṣaṇa), namely, origination (utpāda), continuity (sthiti), deterioration (jarā, vyaya), and
extinction (bhaṅga, nirodha), or second-order elementary constituents of reality (dharma)
that were believed to attach themselves to every first-order elementary constituent of
reality “marked” (lakṣya) by them and thereby become determined in its momentary
existence (kṣaṇika). For the Jainas, all these parallel processes were real, but concerned
different aspects of one and the same entity.
As Jaina thinkers would sometimes state it, to know one thing means to know
everything, inasmuch as everything is interrelated. To know one singular entity, one
should be required to know both all its modes, including past and future, and its complex
interrelatedness, that is, the relations in which it entered, enters, and will enter with
other entities, but also relations which are precluded. Otherwise, our knowledge of the
singular thing would be partial. This would lead to a paradoxical conclusion that “the one
who does not know simultaneously all objects in all three times and in all three worlds,
cannot know even a single substance with all its modes” (PSā 1.48). A corollary of this line
of thinking was the need to admit omniscience of salvific import, which was considered
the ultimate spiritual goal, tantamount to absolute perfection and liberation (mokṣa).
Even without postulating such complex ontic structures of modes and relations
enveloping each and every singular entity and determining its character, considered both
self-same and changing, an accurate description of every little thing, such as
communicating the actual color of the honey bee, may pose a problem and we quickly
realize that the verbal and eidetic apparatus at our disposal is extremely limited.
A handy maxim (nyāya) which the Jainas would frequently use was: “Every sentence
functions with a restriction,” which expressed the limitations naturally imposed on our
thoughts about reality and communication capability. In fact, our language is always a
shorthand for more complex ideas and descriptions. Neither can we think of all aspects of
one and the same thing simultaneously nor can we express them all in one statement. The
language functions within the confines of the impossibility to express the essence of a
singular thing, contextualized through all its modes of existence and relations with the
rest of the world, all at once. Both our thought and the language which reflects the
thought cut the reality into tiny slices, and the way the reality is portioned depends on our
pragmatic needs of either action or communication. An obvious conclusion would be that
every sentence and all communication is ambiguous in lacking precision and inability to
provide a full account of things we would like to describe or express. How is it at all
possible that despite the vagueness and equivocation ingrained in the language people do
manage to communicate ideas? For certain reasons, less related to epistemology and
semantics but more connected with ethics and code of monastic discipline, the Jainas
proposed three different strategies to explain this paradox. Furthermore, they claimed
that the process of communication based on necessarily ambiguous messages can be both
disambiguated and formalized in a system. In fact, we could claim that the doctrine of
multiplexity of reality was precisely such an attempt to provide formalization of the
semantic process of communication and interpretation, accomplished without the use of
any symbols (surprising as it may be, the Indians abstained from the usage of symbols in
epistemology, logic, and semantics; unlike in traditional grammar). One of prime tasks of
an ordinary language user and a philosopher alike was to precisely disambiguate the
ambiguous web of names and propositions within a language. The former would do it on a
daily basis whereas the latter was supposed, in addition, to provide a formalized
description of the disambiguation methodology.

Layers within the name

Three such disambiguation strategies were proposed by the Jainas. Historically the oldest
of the three is most probably the theory of the four standpoints (nikṣepa, nyāsa).2 It
postulated four crucial theoretical determinants as tools of analysis of the semantic field
of a term or singular expression: name (nāma), material representation (sthāpanā),
substance (dravya), and actual condition (bhāva) of an entity analyzed (TS 1.5). In this
sense, four basic semantic layers should formally be distinguished. Let us take as an
example the term “honey bee” and its usage. Suppose an entrepreneur has established a
transport company bearing the name Honey Bee. In this particular context, whenever one
uses the term “honey bee” one does not refer to a specimen of the insect class but to a
company bearing such a name. The statement “The honey bee delivered five tonnes of
steel ball bearings last week” would not make our eyebrows rise. This is the usage of the
term from the standpoint of the name (nāma-nikṣepa). Suppose, now, I point to a
photograph of a honey bee and say: “This is a honey bee.” One could, of course, object
and say: “No, it’s not. It is a photograph of the honey bee,” and of course would be right.
However, while referring to a photograph with “a honey bee” we do it from the standpoint
of material representation (sthāpanā-nikṣepa). In ordinary parlance, we do not confuse
the levels of the discourse. René Magritte employed precisely this implied ambiguity in his
series of paintings La trahison des images, depicting a pipe with the subscript “Ceci n’est
pas une pipe” (This is not a pipe). Despite the general need for precision and accuracy, we
deliberately use ambiguous expressions and are understood, whereas the precision would
lead to misunderstandings and even serious trouble in such cases. Imagine, for instance, a
dialogue at the passport control at an airport and the officer pointing to a photograph of
yours in your passport and asking: “Is that you?,” and you consistently insisting: “No, of
course not!” The next semantic layer of every term is that of substance (dravya-nikṣepa).
From this standpoint we can refer to any specimen of a subset of the bees of the genus
Apis, whether dead or alive, working or asleep. However, to use the term “honey bee”
with most precision would demand that we apply it to an Apis specimen which is presently
regurgitating, that is, engaged in the process of making honey, which gives it its proper
name. This is the usage of words from the standpoint of actual condition (bhāva-nikṣepa).
From this standpoint, a honey bee which is not engaged in the process of regurgitation
and not producing honey would not fulfil the conditions of description of “the honey bee”
and accordingly could not be labeled as honey bee in the true sense; it would be, at most,
a ... half a bee! What we do with words is precisely this: we oscillate between the four
different basic layers of meanings which every term is prone to express, depending on our
pragmatic needs. What allows us to select the proper meaning layer of a term is the
context, the supplier of additional information. A confusion of the layers may be quite
innocuous, may be a humoresque basis for jokes, but also may lead to serious problems
such as cases of, say, blasphemy, when the offended side confuses the layer of the name or
material representation with that of substance and interprets alleged sacrilege done to a
material object as an act against a prophet, god, or some other idea.

In the search of context

A term is related to its designatum, the designated thing, and to assign the correct
relation necessitated by the context is the scope of the doctrine of standpoints. In
contradistinction to it, a sentence relates to an event, process, or fact, or to a thing or
things in an action, state, or condition. In the same way as terms, the usage and
interpretation of sentences are also context-sensitive.
One of the ways we can use sentences is to refer to particular events. The second
disambiguation method developed within the doctrine of multiplexity of reality is the
sevenfold method of conditionally valid predications, known as the theory of viewpoints
(naya-vāda).3 The starting point for this method is the sentence, not just a singular term.
It helps one to allocate the proper context out of a range of possible meanings a sentence
may connote. Take, for instance, “The honey bee makes honey” (madhu-makṣikā madhu
karoti).4 This particular sentence can apply to a range of possible situations, or “worlds.”
For instance, it can refer to the whole subset of the bees of the genus Apis in general, and
assume a form of a definition, the subject of the definition being the universal “honey
bee.” It would still hold valid even if there were particular bees which would not fulfil the
function of making honey. But the sentence may equally refer to a particular honey bee we
have been watching over some time. In some other languages which do not grammatically
distinguish between the continuous and simple tenses as English does, for example, in
Sanskrit or German, the same sentence can describe a current situation when a particular
bee is just collecting nectar. How does the language accomplish such multiple tasks with
one and the same sentence? How do the recipients of the sentence pick up the correct
situation the actual utterance is intended to refer to? This issue is embedded in a more
general question concerning the actual meaning of an utterance, or what is the truth
value of a particular utterance which purports to describe a certain event. In this, the
Jainas followed an approach, generally adopted in India, of the correspondence theory of
truth: a sentence “p” (“it is raining”) is true when indeed p (when it is raining). An
utterance is the actual realization of an idea or thought in a speech act, writing, or
otherwise through a sequence of symbols, such as phonemes or graphic signs, of a
sentence (or “ideal sentence”) we have in our mind. The process of communication
requires that our thoughts as sentences, which have a particular meaning assigned by us,
are translated into actual series of phonemes or graphic signs as utterances. How does
the meaning creep back into the utterances when the recipients hear them or read them
and are required to assign the proper meaning to them? This process, of course, involves
the issue of assigning a truth value to such utterances, too. We are used to a
straightforward relation between an utterance and its truth value as in Figure 4.1.

4.1 Simplified relation “utterance–truth value.”

But clearly, such an approach, based on the simple relation “utterance-truth value,” is
rather an inaccurate description of the whole mechanism, inasmuch as the interpretative
process is intermediated by a range of additional aspects. Obviously, we have to take into
consideration a range of factors to interpret a given utterance correctly, such as the
speaker’s intention and linguistic conventions governing the usage of words, as in Figure
4.2.

4.2 Expanded, but incomplete relation “utterance–truth value.”

However, as the Jainas indicate, these two additional factors prove insufficient, and
other aspects x1, x2, ... xn, such as time factor, universal or nominal reference, synonymy,
and so on, have to be taken into account in order to help us allocate the proper context, as
in Figure 4.3.
4.3 Complex relation “utterance–truth value.”

The Jainas traditionally, since around the fifth century CE, distinguished seven such
major parameters, or interpretative factors, which they called viewpoints (naya). They
claimed these would take into account basic conditions necessary in a proper allocation of
an utterance to its proper context, namely, (1) comprehensive (naigama), (2) collective
(saṅgraha), (3) empirical (vyavahāra), (4) direct (ṛju-sūtra), (5) verbal (śabda), (6)
etymological (samabhirūḍha), and (7) factual (evaṁ-bhūta) viewpoints, as in Figure 4.4.5
This septuplet would not exhaust all possible viewpoints or interpretation strategies, as
these could theoretically be infinite, however the seven were considered basic and most
frequently applied in daily discourse. Some thinkers (NAV 29.13) claimed, however, that
the model should exhaust all possible conceivable perspectives.

4.4 Complex relation “utterance–truth value” systematized through basic seven

viewpoints.

They would further group these seven viewpoints in two major headings, either
substance-expressive (dravyârthika) (1–3) and mode-expressive (paryāyârthika) (4–7)
viewpoints, or object-bound viewpoints, operating by means of object (artha-dvāreṇa
[pravṛtta]) (1–4) and speech-bound viewpoints, operating by means of speech (śabda-
dvāreṇa [pravṛtta]) (5–7). This typology was supposed to comprise such aspects as the
speaker’s intention (who allocates a particular meaning to the sequence of verbal signs he
chooses to convey the idea he has in mind), the universal-particular distinction, temporal
aspect, and a range of linguistic conventions, synonymy, and so on.
In this approach, every utterance has to undergo a process of interpretation in which a
particular context is assigned in which it holds true, the inapplicable contexts or those in
which it is false being excluded. This method was believed to consistently yield one and
only one context in which a particular utterance under particular conditions is both
applicable and true. What were the actual meanings, or spheres of reference, of these
seven viewpoints, one by one?
The first of these, the comprehensive viewpoint (naigama), a most general, open-ended
interpretation of an utterance, sometimes excluded from classifications, takes into
account a possibly extensive, all-inclusive context consisting of a complex of meanings and
connotations evoked by the utterance. It is the least interpreted and analyzed approach,
when it does not even play a role whether we speak of a whole class of things or
individuals of the class. In some cases it may indeed be quite irrelevant whether the
sentence “the honey bee makes honey” (madhu-makṣikā madhu karoti) refers to a class of
things (“the honey bee” and “honey”), a universal, or a particular item of the class, an
individual. The speaker, if asked about this ambiguity, would say: “Whatever. Does it
matter?,” because distinctive features of individuals or constitutive characteristics
representative of a given class are not important at this level of communication. Very
often this comprehensive viewpoint coincides with a colloquial, unreflected usage of an
unspecified reference, which indiscriminately comprises both the particular and the
universal.
The second viewpoint is collective (saṅgraha). It cumulatively refers to a whole class of
individuals, or to the universal, which constitutes the denotation of a given utterance. In
this, it forms a basis for any taxonomy, classification, and definition. As one thinker
expressed it, “The collective viewpoint is the process of synthesis of one facet out of all
possible facets of things” (TBh 1.35, p. 32). From this perspective, “the honey bee makes
honey” (madhu-makṣikā madhu karoti) takes all the three elements of the utterance—“the
honey bee,” “the production,” and “honey”—in a general sense and takes the utterance to
refer to universals or classes of things. It does not matter which honey bee, which honey,
and what activity is involved.
The empirical viewpoint (vyavahāra), which comes third, refers to a concrete individual
selected from a class. It is called “empirical” because we normally do things with concrete
things, namely, in most cases, we handle macroscopic objects and, while describing
processes and actions, we generally tend to refer to such concrete individual objects
which enter our practice (vyavahāra). Superficially the same utterance “the honey bee
makes honey” (madhu-makṣikā madhu karoti) singles out a concrete individual honey bee,
selected out of a whole class, and describes what it generally does or is supposed, or
expected to do.
The next step in the concretization of the context is accomplished by means of the
direct viewpoint (ṛju-sūtra), which narrows the point of reference down to the temporal
condition or manifestation of an individual. The point of reference is such that it is
concurrent with the instant characterized by the action or by the condition in which the
individual thing finds itself in and which is being expressed by the utterance. As a rule, it
is the present slice of time continuum through which a referent passes to which the direct
viewpoint applies, though points in time other than the present moment are theoretically
conceivable. This is the transient, momentary aspect of the thing to which an utterance
refers, and in English this would be usually expressed with the present continuous tense,
but in other languages which do not have this grammatical form, an ordinary present
tense will still be used, as in the Sanskrit sentence madhu-makṣikā madhu karoti (“the
honey bee makes honey” in the sense of “this honey bee is producing honey”). The
corresponding point of reference of this sentence is an event when a particular honey bee
is currently engaged in the process of honey making, for instance, by regurgitating, or
even bringing the pollen and nectar back to the hive, but not any other situation. As the
Jainas would express it, the direct viewpoint emphasizes one particular, that is, present,
mode (paryāya) of a thing, and the substantial and non-momentary character of the entity
is intentionally ignored. Since all the remaining viewpoints, like the direct viewpoint,
emphasize a selected mode of an object in action, for this reason, they all are classified as
mode-expressive (paryāyârthika), unlike the first three, known as substance-expressive
(dravyârthika), which disregard modes and primarily relate to the substance and qualities
of the thing.
In contradistinction to all four above viewpoints denoting an object in action as such,
hence occasionally known as object-bound viewpoints (artha-dvāreṇa), the remaining
three are speech-bound (śabda-dvāreṇa), namely, they introduce additional semantic
distinctions which no longer relate to the thing itself but to specific linguistic devices
employed to describe it.
The first of these speech-bound viewpoints is sometimes (TBh) called present (or
accurate) verbal viewpoint (sāmprata-śabda-naya). By introducing an additional aspect of
linguistic forms of expressing one and the same thing, it limits the scope of application to
the present time and highlights the verbal reference, including grammatical distinctions.
At this level of communication, it is still irrelevant which verbal means we use, provided
they will all be properly understood. Take, for instance, a few sentences which in certain
circumstances will express one and the same phenomenon we are observing: “The honey
bee is making honey” (madhu-makṣikā madhu karoti),6 “The apis is producing golden
ambrosia,” “The drone is participating in the honey making.” This viewpoint allows one to
use any of these, or similar sentences, to describe an event to the same effect. One
intentionally neglects possible differentiation between the shades of meanings and treats
them as fully interchangeable. Similarly, we may refer to Venus seen above the horizon
with expressions “Hesperus is there” or “Phosphorus is shining.” Here, one fully enjoys
the freedom to pick any one from a range of expressions to denote one and the same
event due to linguistic flexibility, when certain semantic distinctions are consciously
overridden, all however within bonds of a prevalent linguistic convention in consonance
with which users of the language agree upon a selection of verbal expressions that denote
a particular individual, event, or phenomenon.
The sixth viewpoint, which further limits the context of an utterance, is called
etymological (samabhirūḍha), not without a reason. It draws a distinction among
seemingly synonymous utterances which is based on the divergent derivation of their
elements. The presupposition is that ideal synonymy is nonexistent, for even apparently
close synonyms possess their distinct shades due to which one is not a full substitute for
the other in all contexts and situations. No synonyms can be considered equivalent. Out of
the three utterances provided as examples for the verbal viewpoint, “The drone is
participating in the honey making” will mean something quite different than “The honey
bee is making honey,” when we consider that the drones are males, not workers, and
instead of gathering pollen and nectar they fly to mate with the queen (which could still
be ultimately considered a part of a complex process of the production of honey). In the
case of example two, one could argue that not every apis, being the genus, is the honey
bee, which is just a subset. Further, “to produce,” “to make,” and “to participate” have
quite different meanings, whereas “golden ambrosia” may be, say, maple syrup. Words
and expressions evoke different images in mind and can be associated with different
situations which make a seeming synonym inapplicable. A synonym may even become a
misnomer in particular circumstances. For instance, to say in the evening that
“Phosphorus is there” is as mistaken as to claim in the morning that “Hesperus is
shining,” even though both would refer to one and the same planet Venus. Hence, from
this particular viewpoint, the language user has to be quite selective when it comes to
choosing an adequate sentence to describe what he has in mind. Out of the possibilities
still open at the level of the verbal viewpoint, one has to choose that particular utterance
which precisely, that is, all etymological and grammatical nuances considered,
corresponds to what one intends to express. As some Jaina thinkers (SSi 1.33) add, this
viewpoint can be quite a useful tool while dealing with homonyms. We can still use quasi-
synonymous phrases but with caution, remembering that they express quite different
things.
The seventh and last viewpoint is the factual viewpoint, which has the narrowest sphere
of application. It stipulates that we use a particular utterance with awareness of
etymological and grammatical nuances but we also apply it to such a context in which the
thing finds itself in a condition precisely corresponding to such nuances. The utterance
“The honey bee is making honey” can be considered true only when it is pronounced at a
particular moment and refers to the situation when the honey bee is actually engaged in
the process of regurgitation, but not when it is cleaning the hive, feeding the larvae,
“dancing” to exchange information, or depositing the collected nectar and pollen in the
hive. Similarly, while pointing to the sky in the morning in the direction where
Phosphorus, or “light-bringing” star, can normally be seen, we are permitted to say
“Phosphorus is shining” provided the planet is actually visible here and now, and is not
obscured by clouds. To say that “Phosphorus is shining,” in a situation when there are
clouds in the sky obscuring Venus, would be still correct at the sixth level, that of the
etymological viewpoint. With all last three viewpoints, the default time reference is the
present moment, but clearly they could also be applied to the past or future.
As we can see the viewpoints are nested in the sense that every subsequent viewpoint
delineates a class of its referents which is a subclass of the preceding viewpoint. We start
with the most comprehensive set of all kinds of elements, indiscriminately including both
individuals and classes, or universals, and gradually narrow the description down to an
individual at a particular time, usually the present moment, which can be referred to with
an utterance that closely, that is, etymologically and grammatically, corresponds to its
current condition.
In fact, we could group all these seven viewpoints under index (i), as in Figure 4.5,
which would then be an appropriate description of this method without any loss of
contents and meaning.

4.5 Complex relation “utterance–truth value” with the index comprising basic
seven viewpoints.

What this theory demonstrates is that no utterance is simply either true or false. Every
sequence of phonemes of graphic signs has to be interpreted and the process of
interpretation necessarily involves the assignment of the proper context. Only then we
can decide whether a sentence is true or false. In other words, since the interpretation of
an utterance and assignment of truth value is not a binary function “utterance → truth
value,” in order to determine its truth value, we have to ascribe it to its specific viewpoint
type that supplies the contextual information which is lacking in the utterance as such,
inasmuch as it “functions with a restriction.” The viewpoints function as such context-
indicators, or intermediary parameters which help us assign the correct context and truth
value. All of the most important context-indicators are indices, comprised under index i.
An utterance will yield truth or falsehood depending on the adequate interpretation of its
context which is determined by means of indexation. Without such indexed determination
of the context of an utterance (through the set of viewpoint indices), the utterance
remains meaningless and its truth value cannot be assigned. As long as we do not know to
what situation the utterance “The honey bee makes honey” is supposed to refer and we do
not know its referent, its meaning remains indeterminate and not liable to truth/falsehood
evaluation.
Accordingly, we can have the following simple model of the context-based
interpretation I of the utterances α, β, γ ... that belong to a class F of formulas (possible
utterances):

I = < D, I, A >

In this model of interpretation I, D is the domain of admissible interpretations (a class of

conceivable situations (phenomena or individuals in action, process, condition, state, etc.)
denotable by the utterances α, β, γ ... ; I is a class of indices i, or context-indicators
(viewpoints); A comprises i-indexed classes of actual denotata.
In other words, Ai is a particular class indexed with a given index i∊I (the i-interpreted
class), which groups actual situations that correspond to those in circumstances described
by index i. The truth value of the i-interpreted utterance α—namely, either ||α||i = 1 for
true or ||α||i = 0 for false—therefore depends on the actual context represented by the
circumstances delimited by elements of the class I (indices) in the interpretation I. Thus,
the truth value of any i-interpreted utterance α depends on the actual context represented
by an index i (viewpoint) of the class I in the interpretation I, and the archetypal index is
circumscribed by the following coordinates:

i = < c, a, t, l, e, s >

In view of some bovine creatures so revered in India, we may call this the CATLES model.
In the formula, the variable c designates the general class C∊D of possible denotata of
(situations refereed to with) utterances α, β, γ ... The variable a is an element of the class
C (viz., a particular individual of the class C). The variable t is the temporal reference (the
point of time is usually the present moment of “now,” which steadily progresses on the
time axis, so in this sense it can also be treated a set of variables, inasmuch as it is not a
constant). The variable l comprises linguistic conventions in accordance with which
utterances α, β, γ ... are pronounced. The variable e indicates etymology and other
grammatical categories in accordance with which one distinguishes between the
meanings of apparent synonyms α, β, γ ... In other words, e describes an equivalence
relation between etymology and grammatical structure of a particular phrase and its
meaning, in view of which—for the range of expressions α, β, γ ...—we will have three
different coordinates eα, eβ, eγ, ... The variable s represent the current status of the
individual or event that is the denotatum of α, β, γ , ..., and the fact that the referent
currently manifests the quality by which it is being referred to by α, β, γ , ..., in
accordance with etymological and grammatical requirements. In this hierarchical, nested
model (Figure 4.6) every subsequent viewpoint introduces a new indexical coordinate,
except for the first viewpoint in the case of which the context-defining parameters remain
indeterminate.

4.6 The hierarchical model of the seven viewpoints.

What the model of viewpoints is certainly not about are different “languages of
metaphysics,” that is, different levels of description ordered hierarchically into a layered
structure in which every layer describes a slice or reality (e.g., in terms of macroscopic
things down to molecular entities and subatomic particles). It is rather about how we
determine the relevant context for utterances and thereby assign truth values to them.
This theory, however, is not without problems and inconsistencies, one of them being how
the three speech-bound viewpoints can consistently be integrated into the whole structure
and what their actual role in the communication is. The fact that different thinkers offered
sometimes quite divergent interpretations of each viewpoint points to the lack of
unanimity among the Jainas but also reveals that specific solutions proposed were found
inadequate and unsatisfactory by the Jainas themselves. Nevertheless, a theory which
matured around twelve hundred to fourteen hundred years ago seems to provide valuable
feedback to modern philosophy of language and semantics and makes one aware how
much our utterances are context-sensitive.

A thing of many facets

Since, as the theory of viewpoints claims, it allows for only one aspect of a real thing
(vastu) (or one context in which the thing is seen) to be verbally expressed through an
utterance interpreted with the help of it, it is sometimes called an incomplete account
(vikalâdeśa), inasmuch as the utterance is applicable only to a particular object under a
particular viewpoint. We could say that, in this theory, an utterance is in search for its
context. The Jainas distinguished that incomplete account from what they claimed
provided a complete account (sakalâdeśa) of a really existing thing. It is a method the
starting point of which is the thing in itself and an attempt to provide a full description of
it through a range of formalized sentences, by way of the search for all possible
statements that can be made about the real thing. This is the third disambiguation
strategy proposed by the Jainas to convert statements expressed in natural languages,
which are by nature ambiguous, into propositions whose meaning is unequivocal. It is
called the sevenfold modal description (syād-vāda).7 This theory provoked the highest
number of objections from other systems which claimed it would violate some basic rules
of logic, especially the law of contradiction. The main reason for the criticism—but also
for the (misapplied) attribution of many-valued logic to it by modern researchers—was
that what Jainas expressed in natural language (and Indians never used symbols for
logical analysis) may seem like ascribing contradictory properties (P and non-P) to one
and the same substratum. However, as we will see, such criticism was based on
misreadings of the theory, and it does provide a valuable insight into how the language
works, with all basic logical rules preserved intact. The theory (vāda) derives its name
from the term syāt8 which is a sentential functor meaning “somehow,” “in a certain
sense,” or “from a certain perspective,” and is explained as a particle “expressive of
multiplexity of reality.” An alternative name for the theory is sapta-bhaṅgī, that is, “a
theory of seven (sapta) basic figures (bhaṅga),” or “the seven-figured model.”
The theory claims that about each and every entity, phenomenon, or event, consisting of
a substantial layer (dravya), of innumerable properties (guṇa) and of infinite modes
(paryāya) and entangled in infinite relations with other entities, one can express an
infinite number of true sentences, which can be formally standardized into seven basic
figures due to their structural properties. Every such formal sentence contains at least the
following conspicuous elements: (1) a sentential functor syāt, which means “in a certain
sense”; (2) the verbal element “is” or “exists” (asti), which is consistently the third person,
present indicative, and has a copulative meaning, being a connective with the third
element; (3) this third element is a predicate, symbolized here as P, Q, and so on, which
however is hardly ever expressly mentioned, but is always implied. In addition, some
sentences have negation which formally is a sentential negation the way it is
grammatically expressed (e.g., “x is not P”), but what is meant by the brief style is
actually a predicative negation (e.g., “x is non-P”). In addition, some of the seven
standardized figures contain also an element “inexpressible” (avaktavya) which is a kind
of complex description (being a conjunction of predicates). All the seven statements, or
figures (bhaṅga), all of them considered true,9 are always represented as follows:

1. “In a certain sense, x [indeed] is P”—syād asty [eva].10

2. “In a certain sense, x [indeed] is non-Q”—syān nâsty [eva].
3. “In a certain sense, x [indeed] is P and, in a certain sense, [indeed] is non-Q”—syād
asty [eva] syān nâsty [eva].
4. “In a certain sense, x [indeed] is inexpressible”—syād avaktavyam [eva].
5. “In a certain sense, x [indeed] is P and, in a certain sense, [indeed] is inexpressible”—
syād asty [eva] syād avaktavyam [eva].
6. “In a certain sense, x [indeed] is non-Q and, in a certain sense, [indeed] is
inexpressible”—syān nâsty [eva] syād avaktavyam [eva].
7. “In a certain sense, x [indeed] is P, in a certain sense, [indeed] is non-Q and, in a
certain sense, [indeed] is inexpressible”—syād asty [eva] syān nâsty [eva] syād
avaktavyam [eva].
The Jainas themselves emphasized, first, that the above formalized seven figures do not
involve any contradiction, and, second, they understand negation in the classical sense:
“x, y are contradictory iff x = ¬y.” The seven-figured model was meant to disambiguate
statements which serve a kind of shorthand for more complex assertions. The phrase of
the first figure (syād asti) would literally mean “x, in a certain sense, exists,” or “x, in a
certain sense, is,” but it is in fact a truncated statement, which should be understood as
“in a certain sense, it (some object under discussion) indeed exists as ...” or “in a certain
sense, some object x indeed has a property P.” As we can see, what its critics considered
to be contradictory properties attributed to one and the same thing, for example, P and
non-P, was never the case in Jaina theory. What may seem like contradictory properties
are always two different properties (e.g., P and non-Q) in the sense that they are
predicated of one and the same thing from different perspectives, in different context,
time frame, and so on, distinctions which an unanalyzed natural language conceals. The
key term was a sentential function which acts as a complex index in which a range of
contextual parameters are hidden.
The sentential functor turns an assertion or negation into a modal sentence, in which
the indicative mood is transformed into “perspective” mood. In other words, the functor
syāt (“in a certain sense”) introduces a particular perspective into every sentence, and the
perspective is expressive of a particular aspect (deśa) or facet (aṁśa) of the thing, which
we can call “parameter.” Four such basic parameters that qualify the way we predicate of
a thing are traditionally distinguished: substance (dravya), place (kṣetra), time (kāla),
condition (bhāva), but these could be extended infinitely to incorporate any other
perspective, or parameter, from which we can predicate a certain property of a thing or
deny it. These perspectives were considered implicit in any statement we make in a
natural language: while describing any object, we usually do it from a certain perspective.
For instance, when we say that a pot is made of clay, we predicate the property “made of
clay” of a pot from a perspective of its substance, not its mode. To predicate the same
property of the pot from the perspective of its mode, that is, transient occurrence, would
mean that the pot could lose its clay as its material substratum and still remain a pot, an
utter impossibility (a clay pot without the property of “being made of clay”!). Thus, the
parameters and aspects, contained in the sentential functor, are an integral element of the
disambiguation strategy of an apparent assertion or negation, for these are considered to
be principally expressed through modal sentences, where modality means perspectivism.
Accordingly, ordinary sentences should always be interpreted through the parameters,
embedded in the modal functor syāt.
In addition, predicates, such as “black,” “made of clay,” and so on, are again never
straightforward. In fact, they always point to a range of other predicates which are
relevant in the case analyzed, but are not explicitly verbalized. An explicit predicate P in
the assertoric sentence Px (“x is P”) carries a range of hidden predicates, both positive,
such as predicates R {A, C, E, G, ...}, and negative, that is, their hidden implied
counterparts Q {B, D, F, H, ...}, which are ultimately negated: ¬Qx. Let us take an original
example11 to see how the predicates work in a seemingly simple sentence “this pot is
made of clay.” The singular predicate, however, implies a range of other predicates, which
can be implicitly asserted or denied:

x is A = “x is made of clay”
x is ¬B = “x is not made of water etc.”
x is C = “x is related to the city of Pāṭaliputra”
x is ¬D = “x is not related to the city of Kānyakubja etc.”
x is E = “x is existing in autumn”
x is ¬F = “x is not existing in spring etc.”
x is G = “x is something black”
x is ¬H = “x is not something red etc.”

While predicating its substance of a material object, we also imply that it has a
particular color, occupies particular space and time coordinates, and so on, but we
implicitly deny possessing a different substance, different color different space and time
coordinates, and so on. A message may be simple and straightforward, but only
superficially, inasmuch potentially it conveys more information which is there to be
expressed and decoded through additional modal assertions and denials.
Thus, all negative predicates ¬Q {¬B, ¬D, ¬F, ¬H, ...} are merely implied by the
affirmative predicates R {A, C, E, G, ...} of the sentence “In a certain sense, x indeed is
made of clay,” but not expressly stated in the first figure. They come to the fore in the
second figure “this pot is not made of clay,” implied by the first figure.
At this stage, we can explain how two orders or parameters are actually introduced into
the theory. What we call parameters above are in fact only first-order parameters, such as
substance S, place or occurrence O, time T, condition C, and the list can be extended to
include mode, aspect, relation, distinction, material substratum, relation, serviceability,
verbal designation, and so on.
Accordingly, the predicates R {A, C, E, G, ...} and ¬Q {¬B, ¬D, ¬F, ¬H, ...} are treated
as one predicate P indexed with the set of the four basic parameters {PSx, POx, PTx, PCx,
...}, for instance, as follows:
“With respect to substance, x is P”: PSx,
“With respect to place, x is P”: POx,
“With respect to time, x is P”: PTx,
“With respect to condition, x is P”: PCx, etc.
However, this would still be a simplification because we also have second-order
parameters, with which first-order parameters are indexed in their turn: RS1x, ¬QS2x,
RO1x, ¬QO2x, RT1x, ¬QT2x, RC1x, ¬QC2x, ..., for instance:
“with respect to substance S1, x is ...”: RS1x,
“with respect to substance S2, x is not ...”: ¬QS2x,
“with respect to place O1, x is ...”: RO1x,
“with respect to place O2, x is not ...”: ¬QO2x,
“with respect to time T1, x is ...”: RT1x,
“with respect to time T2, x is not ...”: ¬QT2x,
“with respect to condition C1, x is ...”: RC1x,
“with respect to condition C2, x is not ...”: ¬QC2x, etc.
Every sentence is taken to embed a set of hidden parameters that delineate the context,
and a simple predicate, say, P of any statement about a real thing x (Px) is in fact a
compounded predicate, the proper understanding of which necessitates an analysis with
the help of additional parameters. On the basis of the above, we can formulate a general
rule:

∀x . ∃σ σ: Pπx,

For every thing x, there is a particular perspective σ such that it can be interpreted as parameter π
with respect to which x is P,

where π = {S, O, T, C} is the set of the first-order parameters of substance = S, place

(occurrence) = O, time = T, and condition = C, and each first-order parameter carries
second-order parameters which specify which of two (or more) parameters of one set (say,
substance S) is meant (S1 or S2).
The third and fourth figure (bhaṅga) may jointly present some interpretative problems,
for they may either seem contradictory or redundant, though the Jainas would maintain
that neither is the case.
The third figure (“In a certain sense, x [indeed] is P and, in a certain sense, [indeed] is
non-Q”) seems to be a conjunction of figures 1 and 2. However, it is not, the Jainas
explain, and the third figure (bhaṅga) carries some additional meaning over and above a
mere conjunction. It will soon become clear what kind of meaning it is. It is grounded in
the claim that every sentence carries two semantic layers: the primary meaning and the
secondary (implied) meaning, and these two layers become relevant in figures 3 and 4. To
understand how it happens, we will have to deal with these two figures jointly.
But what does it mean, in the first place, to predicate the property of inexplicability
(avaktavyatva), that is, impredicability, inassertability, or inexpressibility, of a thing, as in
the fourth figure? Would it not be a contradiction to express anything about something
which is not expressible, including the property “inexpressibility”? Clearly not in this
case, the Jainas would maintain, because what the fourth figure means is not some kind of
absolute inassertability or inexpressibility, but it is a description of the fact that it is a
virtual impossibility to predicate two seemingly contradictory properties of one and the
same thing at the same time. Simultaneity should be interpreted as the case when
statements predicate two incompatible sets of properties of a numerically one object from
exactly the same reference point. Consecutivity automatically changes the points of
reference and a contradiction is dissolved. For instance, as we have seen, the list of
second-order parameters includes also second-order time-parameters, which are exact
temporal reference points that index the basic parameter of time as distinguished from
basic parameters of substance, place and condition, and so on. Accordingly, we can both
assert and deny particular properties of a given object with reference to time (as
distinguished from predication with reference to, e.g., place, substance, and condition),
but we can apply different points of time.
In addition, figure 4 may be taken to refer to the linguistic incapacity, or human
incapability, to express an affirmation of certain properties and the negation of some
others in one breath, which is not the case with figure 3.
However, to understand what both figures 3 and 4 express and how they are different,
another idea is introduced into the theory, emphasis (arpaṇa). In figures which are
conjuncts, every predicate carries an emphasized property (arpita), sometimes called
primary (mukhya), but also a not-emphasized property (anarpita), known as “secondary”
(gauṇa) or subordinate (upasarjanīta), which is usually the opposite of the emphasized
property.
Accordingly, emphasis is an explicit verbal pronouncement of a property, whereas non-
emphasis (anarpaṇa) means that a property is not explicitly mentioned in a sentence
although it is logically implied or entailed. An illustration sometimes adduced is the
relationship in the inference (anumāna; e.g., “sound is impermanent because it is
produced”) based on the negative concomitance (vyatireka; e.g., “whatever is permanent
is not produced”), which is necessarily implied by the formal pronouncement of a positive
proof formula based on positive concomitance (anvaya; e.g., “whatever is produced is
impermanent”), albeit it is not expressed. Both positive and negative kinds of
concomitance are considered logically equivalent, but they may convey different shades of
meaning depending on practical, rhetorical, or didactic needs.
Figure 4 expresses the simple fact that it is not possible to simultaneously predicate
two properties, one positive (P) and one negative (¬Q), both emphasized as primary, of
one and the same thing. Simultaneity should be interpreted as the case when statements
predicate two incompatible sets of properties of a numerically one object from exactly the
same reference point. Consecutivity automatically changes the reference point and in this
way a contradiction is avoided. For instance, as we have seen, the list of second-order
parameters includes also second-order time parameters, which are exact temporal
reference points that index a basic parameter of time as distinguished from basic
parameters of substance, place, and condition. Thus, we can both assert and deny
particular properties of a given object with reference to time (as distinguished from
predication with reference to, e.g., place, substance, and condition), and we can apply
different time points.
In other words, two properties cannot be emphasized simultaneously, but only
consecutively in two separate sentences. In compounded, complex figures (nos. 3–7), out
of two properties P and non-Q, only one (or none) can be emphasized, that is, explicitly
verbalized, whereas the other one is logically implied or entailed. This means that the
emphasis applies only when we want to express two properties, but it has no function to
perform when only one property is expressed or denied, as in figures 1 and 2.
Following such understanding, two different strategies are suggested to differentiate
between the third figure (“In a certain sense, x [indeed] is P, in a certain sense, and
[indeed] is non-Q”) and the fourth one (“In a certain sense, x [indeed] is inexpressible”).
The first interpretation, which is the most prevalent (e.g., Akalaṅka, RVār 4.42;
Abhayadeva-sūri, TBV 1.36), rules that both figures can be expressed with “equal
expressive force,” namely, either both emphasized or both not emphasized, only
consecutively, and that is the third figure. When one intends to express P and non-Q with
“equal expressive force” simultaneously, this is the case of inexpressibility. This is the
more traditional way to differentiate between figures 3 and 4: to assume that the third
statement (“in a certain sense, x [indeed] is P, in a certain sense, and [indeed] is non-Q”)
consecutively expresses two distinct properties, which are not contradictory,12 because
they refer to two different contexts, or they have two different sets of parameters,
whereas the fourth statement (“in a certain sense, x [indeed] is inexpressible”) expresses
two distinct parameterized properties, which do not stand in contradiction, but there is no
linguistic tool at our disposal to express them simultaneously. An example sometimes
given (SVM1 28.12–13) for the fourth figure (“inexpressible”) is the case of determining
the sex of the fetus in the womb without prior knowledge before birth (“this is a boy/girl”):
one may resort to “the third sex” (napuṁsaka), namely, neither male nor female.
According to the second interpretation, in the third figure, one property is emphasized,
whereas the other is not. In other words, if both properties, P and Q, are expressed with
“equal expressive force” simultaneously, this is the case of inexpressibility (fourth figure),
because either both are emphasized or both are not (RVār 4.42). However, as long as they
are expressed with different expressive force, even at the same time, P is expressed and Q
is not or vice versa, this is the case of the third figure.
Jaina thinkers unanimously emphasize that neither figures 3 and 4 nor any of the last
three figures involves contradiction, because no two contradictory predicates (P and non-
P) are ever predicated of one and the same substance; one property is asserted in the
case of, say, substance, whereas the other one is denied in the case of quality or mode,
and so on (ĀMī 16, YA 48), namely, it is the “P and non-Q” situation. For instance, to say
that “in a certain sense, a honey bee is indeed black with respect to its odd stripes and is
indeed yellow with respect to its even stripes” (figure 3, with both predicates expressed
with equal expressive force, but consecutively) involves no less contradiction then to say
“In a certain sense, a honey bee is indeed YbElLaLcOkW wWiItThH rReEsSpPeEcCtT tToO
iItTsS oEdVdE NsStTrRiIpPeEsS”13 (figure 4, with both predicates expressed with equal
expressive force, but simultaneously).
Emphasis, as an integral part of the sevenfold modal description, should be
symbolically integrated in the formalized model; let’s take symbol ε to represent it, and
symbol π for parameter:

∀x . ∃σ σ: Pπσx

For every real thing x, there is always a particular perspective σ such that it can be interpreted as
parameter π with respect to which x is P and the property P is emphasised under condition ε.

In addition, symbol ε1 will stand for “property under emphasis” (“emphasized property”)
and ε0 for “property under no emphasis” (“property not emphasized”). To provide an
illustration for the third figure with the four basic parameters of substance (S), place or
occurrence (O), time (T), and condition (C), the situation may be represented in a more
detailed manner as follows:

1. AS1ε1x & ¬BS2ε0x.

“In a certain sense, that is, with respect to substance S, a given pot x exists as being made
of clay” (AS1x) and “with respect to substance S, a given pot x does not exist as something
made of water” (¬BS2x).

2. CO1ε1x & ¬DO1ε0x.

“In a certain sense, that is, with respect to place O, a given pot x exists in the city of
Pāṭaliputra” (CO1x) and “with respect to place O, a given pot x does not exist in the city of
Kānyakubja” (¬DO2x).

3. ET1ε1x & ¬FT2ε0x.

“In a certain sense, that is, with respect to time T, a given pot x exists in the autumn”
(ET1x) and “with respect to time T, a given pot x does not exist in the spring” (¬FT2x).

4. GC1ε1x & ¬HC2ε0x.

“In a certain sense, that is, with respect to condition C, a given pot x exists as
something black” (GC1x) and “with respect to condition C, a given pot x does not exist as
something red” (¬HC2x).
The remaining figures 5–7 in the Jaina model will be permutations of the three basic
ones: 1, 2, and 4. As we can see, things have become quite complex at this stage. I will
refrain here from providing two possible attempts of formalization with detailed
description, and will restrict myself just to a brief exposition of one such formalization14:

1. Pπ1ε1x.

“In a certain sense, x [indeed] is P.”

2. ¬Pπ2ε1x.

“In a certain sense, x [indeed] is non-Q.”

3. Pπ1ε0x & ¬Pπ2ε0x.

“In a certain sense, x [indeed] is P and, in a certain sense, [indeed] is non-Q.”

4. Pπ1ε1x & ¬Pπ2ε1x.

“In a certain sense, x [indeed] is inexpressible.”

5. Pπ1ε1x & Pp1ε0x & ¬Pπ2ε0x.

“In a certain sense, x [indeed] is P and, in a certain sense, [indeed] is inexpressible.”

6. ¬Pπ2ε1x & Pp1ε0x & ¬Pπ2ε0x.

“In a certain sense, x [indeed] is non-Q and [indeed] is inexpressible.”

7. Pπ1ε1x & ¬Pπ2ε1x & Pπ 1ε0x & ¬Pπ2ε0x.

“In a certain sense, x [indeed] is P, in a certain sense, [indeed] is non-Q and, in a certain
sense, [indeed] is inexpressible.”
How to read this? P is a predicate variable which comprises a range of positive
predicates R {A, C, E, G, ...} and a range of negative predicates Q {B, D, F, H, ...} (denied
of the object); π is a set of the first-order parameters {S, O, T, C} of substance, place
(occurrence), time and condition, and so on (can be extended indefinitely) which
determine the parameter with respect to which predicate P is considered; ε is emphasis
which indicates that a given property is either expressed (ε1) or suppressed (ε0). For
instance, Pπ1ε1x states that an assertion “object x is P” is to be explicated through a certain
first-order parameter π (π1), for example, “in view of its substance, a jar is made of clay,”
and such property is verbally emphasized (ε1), that is, the predicate is expressly stated in
language. In the case of the second figure ¬Pπ2ε1x, the proposition “the same object x is
non-P” means that the predicate should be understood with respect to some other first-
order parameter π (π2), for example, “in view of its substance, a jar is not made of water,”
and likewise it is verbally emphasized (ε1).
The above does not aspire to be the ultimate formalization of the sevenfold modal
description (syād-vāda), but it shows the way it can be done. The Jainas themselves were
often not quite unanimous as to how to interpret certain figures, which means one would
have to provide a formalized model for specific interpretations offered by certain Jaina
thinkers separately. Such formalizations as the one above will probably reveal minor
problems with this theory, such as redundancies. Despite certain deficiencies, what this
model—all done by Jaina philosophers in natural language without any recourse to
symbols—shows is how many different true statements can be made about one and the
same object, event, phenomenon, or situation without running into a contradiction.

Conclusion
The theory of modal description as well as the two ones previously described, the theory
of the four standpoints and the theory of viewpoints, highlight an extremely important
element of communication and thought, which has usually been neglected by philosophers
in the past: our communication is necessarily limited and the users of language supply the
missing information from the context in order to complete “shorthands” which we are
bound to use. One of the prime tasks of the philosopher is to propose adequate tools that
should disambiguate the language and develop them to fit the requirements of efficient
communication and description. The principle that language, communication, and
description are by nature ambiguous stipulates that suitable procedures of reasoning
should likewise be developed which would seriously take into account the context-
dependency of every term, expression, proposition, and description. And this is what
pragmatic logicians do now, several centuries after this kind of semantics started.

Notes
* Work on this chapter has been generously supported by the National Science Centre of Poland
(Research Project: History of Classical Indian Philosophy: non-Brahmanic Schools, National
Science Centre, 2011/01/B/HS1/04014).
1. “Eric the half a bee,” an excerpt, Monty Python Sings CD booklet, 1989 Virgin Records, p. 10.
2. This doctrine most probably goes back to an ancient teacher Pārśva, mythologized and considered
the twenty-third tīrthaṁkāra (“ford maker”), who most probably lived in the sixth century bce,
perhaps one hundred years before Vardhamāna Mahāvīra (died c. 400 BCE), mythologized as the
twenty-fourth tīrthaṁkāra; see Balcerowicz (2016, pp. 174–185).
3. The exposition of the naya-vāda in this chapter, its formalization in particular, is based on
Balcerowicz (2001; 2003).
4. Bees (śyāma-bhramara) are mentioned in the context of viewpoints by JTBh 2 § 8.
5. This theory was probably a joint product of two traditions, Jainism and Ājīvikism; see Balcerowicz
(2016, pp. 186–204).
6. Remember that in many languages, especially in Sanskrit and Prakrit, there is no distinction
between the simple present and the present continuous tenses, so madhu-makṣikā madhu karoti
means both “the honey bee makes honey” and “the honey bee is making honey,” so throughout our
analysis in this chapter we effectively deal with nominally one and the same sentence seen through
 seven viewpoints.
7. It is discussed in much detail in Balcerowicz (2015).
8. Formally it is the third person optative of the verbal root “to be” (“could be, may be”).
9. There were numerous attempts to formalize this theory in a way which would assign every figure
a distinct truth value, which would ultimately lead to a model of many-valued logic, but such
attempts do not accurately depict what the Indian themselves said in their texts. Examples are
numerous, for example, Barlingay (1965, pp. 6, 65), Mukerji (1977, pp. 230–233), Matilal (1981,
pp. 54–56; 1991, pp. 12–16), Pandey (1984, p. 163), Bharucha–Kamat (1984), Gokhale (1991, pp.
83–84), Ganeri (2002), Priest (2008), Schang (2008a, b; 2010). Such formalization attempts were
rather projections of modern concepts onto Jaina ideas and resulted from misreading the texts.
This Jaina theory leaves no room for many truth values. See: Balcerowicz (2015, pp. 184–195).
10. The particle eva in the sense of “indeed” or “exclusively” was introduced around 600 CEin order
to restrict the applicability of the property predicated of the real thing, being a semantic method to
restrict the range of the term that denotes the property.
11. SVM1 23.113–119, p. 143.12–18 = SVM2, p. 210.7–12.
12. The criticism commonly wielded against the syād-vāda in India, but also the assumption of some
modern researchers who impose many-valuedness on the theory, was that it implied the
contradiction of “P and non-P” (instead of what should be adequately interpreted as: R and non-Q).
13. This is the way I try, imperfectly as it were, to graphically reproduce the simultaneity of
expressing two predicates of equal expressive force.
14. For the details, see Balcerowicz (2015, pp. 221–224).

Abbreviations and bibliography

ĀMī: Samantabhadra, Āpta-mīmāṁsā. (1) Ed. Jain, P. (1914), Ācārya-śrī-samantabhadra-svāmi-viracitā
Āpta-mīmāṁsā syād-vāda-vidyāpati-śrī-vidyānanda-svāmi-viracitā pramāṇa-parīkṣā ca. Kāśī:
Sanātana Jaina Granthamālā 7, 8. (2) Ed. and trans. Shah, N. (1999), Samantabhadra’s
Āptamīmāṁsā. Critique of an Authority [Along with English Translation, Introduction, Notes and
Akalaṅka’s Sanskrit Commentary Aṣṭaśatī]. Ahmedabad: Sanskrit-Sanskriti Grantha-mālā 7, Dr.
Jagruti Dilip Sheth.
Balcerowicz, P. (2001), “The logical structure of the naya method of the Jainas.” Journal of Indian
Philosophy, 29, 379–403.
Balcerowicz, P. (2003), “Some remarks on the naya method,” in P. Balcerowicz (ed.), Essays in Jaina
Philosophy and Religion. Delhi: Motilal Banarsidass, pp. 37–67.
Balcerowicz, P. (2015), “Do attempts to formalise the syād-vāda make sense?,” in P. Flügel and O.
Qvarnström (eds.), Jaina Scriptures and Philosophy. London and New York: Routledge, pp. 181–248.
Balcerowicz, P. (2016), Early Asceticism in India: Ājīvikism and Jainism. London and New York:
Routledge.
Barlingay, S. S. (1965), A Modern Introduction to Indian Logic. Delhi: National Publishing House.
Bharucha, F., and Kamat, R. V. (1984), “Syādvāda theory of Jainism in terms of devial logic.” Indian
Philosophical Quarterly, 9, 181–187.
Ganeri, J. (2002), “Jaina logic as the philosophical basis of pluralism.” History and Philosophy of Logic,
23, 267–281.
Gokhale, P. P. (1991), “The logical structure of syādvāda.” Journal of Indian Council of Philosophical
Research, 8, 73–81. (Rpt in Shah [2000, pp. 75–86].)
JTBh: Yaśovijaya, Jaina-tarka-bhāṣā. Ed. and trans. Bhargava, D. (1973), Mahopādhyāya Yaśovijaya’s
Jaina Tarka Bhāṣā with Translation and Critical Notes. Delhi, Varanasi and Patna: Motilal
Banarsidass.
Matilal, B. K. (1981), The Central Philosophy of Jainism (Anekānta-vāda). Ahmedabad: L.D. Series 79,
L. D. Institute of Indology.
Matilal, B. K. (1991), “Anekānta: both yes and no?” Journal of Indian Council of Philosophical Research,
8, 1–12. (Rpt in Shah [2000, pp. 1–16].)
Mukerji, R. N. (1977), “The Jaina logic of seven-fold predication,” in A. N. Upadhye, N. Tatia, D.
Malvania, M. Mehta, N. Shastri, and K. Shastri (eds.), Mahāvīra and His Teachings. Bombay:
Bhagavān Mahāvīra 2,500 Nirvāṇa Mahotsava Samiti, pp. 225–233.
NAV: Siddharṣi-gaṇin, Nyāyâvatāra-vivṛti. Ed. and trans. Balcerowicz, P. (2001), Jaina Epistemology in
Historical and Comparative Perspective: A Critical Edition and an Annotated Translation of
Siddhasena Mahāmati’s Nyāyâvatāra, Siddharṣigaṇin’s Nyāyâvatāra-vivṛti and Devabhadrasūri’s
Nyāyâvatāra-ṭippaṇa. Stuttgart: Volumes I and II. Alt- und Neu-Indische Studien 53,1 and 53,2. Franz
Steiner Verlag. (Second rev. ed. 2009, Delhi: Motilal Banarsidass.)
Priest, G. (2008), “Jaina logic: a contemporary perspective.” History and Philosophy of Logic, 29, 263–
278.
PSā: Kundakunda, Pavayaṇa-sāra [Pravacana-sāra]. Ed. Upadhye, A. N. (1984), Śrī Kundakundācārya’s
Pravacanasāra (Pavayaṇasāra), a Pro-Canonical Text of the Jainas, the Prakṛit Text Critically Edited
with the Sanskrit Commentaries of Amṛtacandra and Jayasena. Agās (Gujarat): Śrī Paramaśruta-
Prabhāvaka-Maṇḍala, Śrīmad Rājacandra Āśrama. (First ed., Bombay 1935.)
RVār: Akalaṅka, Tattvârtha-vārttika (Rāja-vārttika). Ed. Jain, M. K. (1953–57), Tattvārtha-vārttika [Rāja-
vārttika] of Śrī Akalaṅkadeva. Edited with Hindi Translation, Introduction, Appendices, Variant
Readings, Comparative Notes etc. Parts I–II. Delhi: Jñānapīṭha Mūrtidevī Jaina Grantha-mālā,
Sanskrit Grantha 10, 20, Bhāratīya Jñānapīṭha Prakāśana. (Second ed., Delhi 1982.)
Schang, F. (2008a), “Truth-values are not the whole story (A semi-paranormal logic of acceptance and
rejection within n-valuation),” paper presented to 4ème Congrès Mondial sur la Paraconsistance
(WCP4)—Ormond College, University of Melbourne, July 13–18, 2008, manuscript,
http://poincare.univ-nancy2.fr/digitalAssets/28005_SCHANG_WCP4__Truth-
values_are_not_the_whole_story.pdf (accessed on March 4, 2009).
Schang, F. (2008b), “Jaina logic in the light of n-valuation,” typescript 2008.
Schang, F. (2010), “Two Indian dialectical logics: saptabhaṅgī and catuṣkoṭi.” Journal of Indian Council
of Philosophical Research, 27, 47–76.
Shah, J. N. (ed.) (2000), Jaina Theory of Multiple Facets of Reality and Truth. Delhi: Bhogilal
Leherchand Institute of Indology.
SSi: Pūjyapāda Devanandin, Sarvârtha-siddhi. Ed and trans. (into Hindi) Shastri, P. (1934), Ācārya
Pūjyapāda’s Sarvārthasiddhi (The Commentary on Ācārya Griddhapiccha’s Tattvārtha-sūtra). Ed. and
trans. (into Hindi). Varanasi. (Rpt, Delhi: Jñānapīṭha Mūrtidevī Jaina Grantha-mālā 13, Bhāratīya
Jñānapīṭha Prakāśana, 2000.)
SVM: Malliṣeṇa, Syād-vāda-mañjarī. (1) Ed. Dhruva, A. B. (1933), Syādvādamañjarī of Malliṣeṇa with
the Anyayoga-vyavaccheda-dvātriṁśikā of Hemacandra. Bombay: Bombay Sanskrit and Prakrit
Series No. 83. (2) Ed. and trans. (into Hindi) Jagadīśacandra Jain (1992), Kali-kāla-sarvajña-śrī-
hemacandrâcārya-viracitā-anya-yoga-vyavaccheda-dvātriṁśikā-stavana-ṭīkā śrī-malliṣeṇa-sūri-praṇītā
syād-vāda-mañjarī. Agās: Śrī Paramaśruta Prabhāvak Maṇdal, Śrīmad Rājacandra Āśram.
TBh: Umāsvāti, Tattvārthâdhigama-bhāṣya. See TS1.
TBV: Abhayadeva-sūri, Tattva-bodha-vidhāyinī. Ed. Sukhlāl Saṅghavi and Becardās Dośi (1984),
Saṁmati-tarka-prakaraṇa by Siddhasena Divākara with Abhayadevasūri’s Commentary, Tattva-
bodha-vidhāyinī. Vols. I and II. Kyoto: Rinsen Buddhist Text Series VI–1, 2. (Rpt from the original
edition published in 5 vols. Amdāvād: Gujarāt-purā-tattva-mandir-granthāvalī 10, 16, 18, 19, 21,
Gujarāt-purā-tattva-mandir, 1924–31.)
TS: Umāsvāmin, Tattvârtha-sūtra. (1) Śvetāmbara Recension: Mody, M. K. (ed.) (1903–1905),
Tattvārthādhigama by Umāsvāti Being in the Original Sanskrit with the Bhāṣya by the Author
Himself. Calcutta: Bibliotheca Indica/Bibliotheca Indica New Series 1044, 1079, 1118, Asiatic Society
of Bengal. (2) Digāmbara Recension: See SSi.
YA: Samantabhadra, Yukty-anuśāsana. Ed. Lāl, I. (1921, Saṁ.: 1977), Yukty-anuśāsana of
Samantabhadra. Edited with the Ṭīkā of Vidyānandācārya. Bombay: Māṇikacandra-Digambara-Jaina-
Grantha-mālā, Nāthūrām Premī Maṁtrī/Māṇikacandra-Digambara-Jaina-Grantha-mālā-samiti.
 Part Two

Consciousness and the External World

 5
Proofs of Idealism in Buddhist Epistemology: Dharmakīrti’s
Refutation of External Objects
Birgit Kellner

This chapter is dedicated to Tosaki Hiromasa and Tilmann Vetter (†), without whose
pathbreaking contributions to the study of Dharmakīrti’s theory of perception in the
Pramāṇavārttika and the Pramāṇaviniścaya it could never have been written.1

Medieval Indian Buddhist philosophers developed a variety of proofs of “mere-cognition”

(vijñaptimātra(tā)), the signature concept of the Yogācāra school which stands for the
principle that when cognition is aware of objects, there is just (mātra) that awareness, and no
external object that would correspond to its content. Mere-cognition is not found in the
earliest Yogācāra literature, and it only gradually rises to the prominent position that it
occupies in pertinent works of Vasubandhu (probably between 350 and 430 CE),2 and in later
philosophical literature predominantly of the Buddhist tradition of logic and epistemology
(pramāṇa), initiated by Dignāga (c. 480–540 CEor slightly earlier) and Dharmakīrti (active
between mid-sixth and mid-seventh century CE).3
There has been much debate in recent times whether mere-cognition philosophically
represents a form of idealism.4 More specifically, objections were raised to earlier
interpretations that consider mere-cognition to correspond to what in the history of Western
philosophy has been called “subjective idealism”: the view that objects cannot exist without
being cognized, or, as it is sometimes put, that there is no “mind-independent world.” But an
important methodological consideration seems to be largely missing in this debate. The
positions that scholars take in the idealism debate are typically supported through studies of
the semantics of key terms, of textual context, of doctrinal and intellectual background, and of
individual arguments that individual Buddhist thinkers advance in support of mere-cognition.
But when all that has been, rightly, taken into consideration, the question remains whether
the various individual proofs form part of a whole. Do they indicate a strategy, a method?
More generally, is the attribution of a philosophical “position” to a historical author complete
without an enquiry into a method, strategy, or style of reasoning that she pursues? Such
questions are all the more pertinent for Dignāga and Dharmakīrti, who engaged in the
construction of rigorous theories of inference and proof. In addition to pursuing certain
practices of argumentation that might follow more or less established patterns and
conventions, they explicated standards of validity that were to serve as yardsticks for all
argument. This, then, raises the question how individual areas of investigation (such as mere-
cognition) are related to explicit logical theorizing.
What first springs to eye in the different proofs of mere-cognition found in the works of
Vasubandhu, Dignāga, and Dharmakīrti is that they advance through a primarily critical
method that refutes external objects. This is not surprising, for arguing that external objects
do not exist (or cannot be known) is an evident choice of method for establishing that
whenever cognition is aware of something, there is just that awareness, and not an external
object. Vasubandhu argues along these lines in his Viṃśikā Vijñaptimātratāsiddhiḥ, the “Proof
of Mere-Cognition in Twenty Stanzas,” that comes with a prose auto-commentary. Dignāga
disproves external objects in his “Investigation of the Object-Support” (Ālambanaparīkṣā,
short ĀP), a treatise in verse furnished with a prose auto-commentary, the
Ālambanaparīkṣāvṛtti (ĀPV). But it is Dharmakīrti who presents by far the broadest array of
arguments against external objects, in the chapters on perception (pratyakṣa) of his two
comprehensive epistemological treatises, the “Commentary on the Means of Valid Cognition”
(Pramāṇavārttika, short PV) in verse, and the “Ascertainment of the Means of Valid Cognition”
(Pramāṇaviniścaya, short PVin) in verse and prose. To bring Dharmakīrti’s arguments into
relief, it will be helpful to first review the arguments advanced by Vasubandhu and Dignāga.
Vasubandhu’s position has been a matter of some debate, and some have argued that it differs
strongly from that of Dharmakīrti. Dharmakīrti, for his part, continues to some degree along
lines of argument established by Dignāga.
 The refutation of external objects in Vasubandhu’s Viṃśikā and Dignāga’s
Ālambanaparīkṣā
Idealist readings of Vasubandhu’s Viṃśikā (short Vś) tend to focus on a set of famous and
widely quoted arguments offered in Vś 11–15. Vasubandhu here argues that objects of
perception neither exist as continuous wholes (avayavin), nor as a multitude of minute and
indivisible particles, nor as a single such particle, that is, an atom. Like Dignāga and
Dharmakīrti, Vasubandhu focuses his discussion on the perceptual awareness of objects; the
term “cognition” in what follows should therefore be understood as referring primarily to
sense perception. For Vasubandhu, the analysis of perceived objects into physical parts is
necessary—for otherwise a host of anomalies would occur—but the very idea of a single,
indivisible atom turns out to be logically impossible. For one, atoms would have to be
connected with each other to form macroscopic objects. But if an atom were simultaneously
connected with other atoms on each of its six sides (the cardinal directions, above and below),
it would have to have six parts and could no longer be single.5 Non-idealist interpreters of the
Viṃśikā construe this argument as showing that “reason” cannot postulate external objects as
causes of experience, since they are logically incoherent,6 or as pursuing the more limited,
phenomenalist claim that the objects we are experiencing are not physical objects.7
While these and other arguments against idealist readings of the Viṃśikā cannot be easily
dismissed, we have recently argued8 that one can still make better sense of the text as a
whole as pursuing subjective idealism. Our argument is extensive and cannot be reproduced
here in full. But the main points are as follows. In the ninth chapter of his
Abhidharmakośabhāṣya, Vasubandhu argues for a negative thesis—that a self as substance
over and above cognitions does not exist—through what effectively amounts to an extended
argumentum ad ignorantiam: an ātman or pudgala does not exist because neither of the
accredited means of valid cognition (pramāṇa) perception, inference, or scripture offer any
evidence for its existence. A closer look at the Viṃśikā led us to find the same strategy at
work: external objects do not exist because there is no evidence for them through any
pramāṇa. In this reading, the mereological arguments in Vś 11–15 simply support the larger
argument that there is no scriptural evidence for external objects, for it clarifies how the
Buddha’s teaching that the constituent factors are without essence (dharmanairātmya),
epitomized in mere-cognition, is to be correctly interpreted: it does not proclaim the existence
of objects of perception. It is true that Vasubandhu does not explicitly state the conclusion to
the argument from ignorance—that objects do not exist—but this can be accounted for in two
mutually nonexclusive ways. On the one hand, it can be argued that he did not have to—what
other conclusion is there to draw from the complete absence of evidence for external objects
than their nonexistence? It is not far-fetched to believe that the nonexistence of something is
the best explanation for the utter lack of evidence for its existence.9 On the other hand,
Vasubandhu’s reticence at stating his conclusion might have soteriological motivations, for in
the closing stanza (Vś 22) of the Viṃśikā he stresses that mere-cognition in all its aspects is
the domain of the Buddha—how dharmas truly are can only be realized in a higher meditative
state, in transconceptual meditation or nirvikalpasamādhi.
The point of departure for Dignāga’s argumentation in the ĀP(V) is that the “object-
support” (*ālambana) of a perception has to fulfil two conditions: it must be a cause of
perception and perception must have its appearance (*ābhāsa).10 Perception, in other words,
is related to its object through causation and resemblance. In Indian epistemological
literature after Dignāga this ultimately representationalist theory of perception is taken to be
characteristic for the Sautrāntika school of thought.11 With this definition as a yardstick,
Dignāga develops an argument along the following lines. Some claim that individual atoms
are the object-support because they cause perception, while others hold that a collection (Tib.
’dus pa, Skt. *samudāya) of atoms is the object-support because it appears in perception.
What appears are, after all, spatially extended shapes and colors. Dignāga argues that the
first proposal is wrong because perception does not carry the appearance of individual atoms,
just like it does not carry the appearance of the sense-faculty (ĀP 1). But while perception
bears the appearance of the collection, the collection cannot be its cause, for it does not exist
substantially, just like the second moon in the sky that is seen by persons suffering from
floaters (*timira) (ĀP 2). Only atoms are causally efficacious, not collections. Even if
Dignāga’s opponent were to modify his proposal, as he attempts to do, the fundamental
problem remains that there is an incongruence between what is seen and what, in virtue of a
given ontological theory, is believed to exist. What is seen is not real, and whatever might be
real cannot be seen. Dignāga concludes that the object-support of perception is not outside,
but inside consciousness (ĀPV 178,2f.). The object-support is the “cognizable form inside”
(antar jñeyarūpam, ĀP 6) that merely appears as if it were external. As this appearance exists
simultaneously with its perception, it seems to violate the requirement that cause and effect
must exist in temporal succession. Dignāga consequently justifies that something
simultaneous can be a cause, based on the principle of “logicians” (Tib. gtan tshigs pa) that
“when the one is present or absent, the other is also present or absent”
(*bhāvābhāvayoḥtadvattvam, ĀP 7a with ĀPV).12 To preserve the notion of successive
causality, one may alternatively consider that the form of an object produces a capacity
(*śakti)—a seed (*bīja)—within the series of mental episodes that constitutes consciousness.
This seed later gives rise to a perception with the same object-appearance (ĀP 7cd). Dignāga
does not explicitly identify the object-support in this model; he does not say whether it is the
earlier object-form or the seed. But the basic conclusion remains that the object of perception
is internal (ĀP 8bcd).
On the whole, Dignāga proceeds in the ĀP(V) by evaluating different candidates from a
given ontology in terms of whether they can fulfil a given definition of the object of
perception. This atomistic ontology is not subjected to any criticism comparable to that of
Vasubandhu. It is simply determined as an unsuitable foundation of perceptual experience.
Vasubandhu’s mereological arguments effectively demolish the notion of atoms—although this
might not be his only and not even his main goal of proof—whereas Dignāga’s arguments are
logically consistent with the existence of a physical world made up of indivisible and minute
material particles—a world, that, however, could not in any way causally impinge on human
cognitive faculties. Both according to our novel reading of the Viṃśikā and according to
earlier idealist interpretations, Vasubandhu argues that external objects do not exist (but in
different ways). Dignāga’s arguments, by contrast, only support the logically weaker,
phenomenalist claim that external objects cannot be perceived, or that the objects of
experience do not exist outside consciousness. Dan Arnold recently characterized these
positions, respectively, as an “epistemic idealism,” expressed with the claim “that what we are
immediately aware of must be understood in terms of the intrinsic properties of cognition,”13
and an ontologically committed “metaphysical idealism” that denies the existence of a
physical world. At first sight, this distinction seems well-suited to capturing a significant
philosophical difference between Vasubandhu and Dignāga. But as so often happens, bringing
Dharmakīrti into the picture makes matters more complicated and offers opportunities for
rethinking what initially appears to be obvious.

Dharmakīrti’s refutation of external objects

The framework: realism and idealism as different levels of analysis
There is evidence to support that Dharmakīrti was generally committed to idealism. He wrote
a treatise to prove the existence of other mental continua, the Santānāntarasiddhi, and there
devoted considerable effort to averting the danger of solipsism. Who but a committed idealist
would feel the need to counter solipsism?14 But a philosopher’s commitment to a position
need not mean that their arguments in favor of it are strong enough to support it; one cannot
exclude that there are internal tensions, and it remains therefore necessary to attend to
Dharmakīrti’s actual refutation of external objects. Moreover, Dharmakīrti’s attitude to a
Yogācāra position is nuanced. His epistemological theory encompasses a theoretical hierarchy
between a realist theory of cognition and its idealist counterpart—“idealist” at least in the
sense of a theory that denies perceptual content to derive from, or correspond to, external
objects that exist independently of consciousness.15 In the hierarchy of idealism and realism,
idealism is superior because it provides the more accurate analysis of cognition, yet realism
remains the default level of analysis in most areas of philosophy in which Dharmakīrti
engages, notably in his theory of inference. This theoretical hierarchy is at the same time a
soteriological one in that the idealist theory represents a level of analysis that corresponds
more closely to how beings that are further advanced on the Buddhist path to liberation are to
experience reality. If the perceptual process is scrutinized, the view held “in the world” (loke)
that perception is of external objects turns out to be misguided. As a matter of fact,
perception only experiences—is aware of—an object’s form within itself; it does not
experience an external object that differs from it in nature (parātman). This internal form is
superimposed, projected outward, as it were, due to a “disturbance” (viplava, 431),16 that is,
due to a systematic cognitive distortion issuing, no doubt, from ignorance (avidyā) as a
fundamental conditio humana that on the Buddha’s diagnosis keeps unawakened human
beings entrapped in saṃsāra.17 The appearance of cognition, by nature undifferentiated, as
different in terms of subject and object, is also a “disturbance” (upaplava, 212, 214; viplava,
331). It is out of an “inner disturbance” (antarupaplava) that cognition, one in nature, shows a
multiplicity of false appearances, subject, and object (361). Philosophically, distortions issuing
from ignorance act as an error theory that allows to explain how ordinary beings come to
think and experience their cognitive life in a way that does not represent how things really
are.18 Scriptural teachings of the five psychophysical aggregates of living beings (skandha) or
of the twelve sense-spheres (āyatana) imply the existence of external objects, but such
teachings are spread by the Buddha merely in accordance with worldly understanding. In
doing so they set aside nonduality (advaya) as that which is ultimately true (tattvārtha). In
this the Buddha resembles an elephant who shuts one of its laterally set eyes and then only
sees to one side (219; read with the commentary M1 184,10–14).

The transition from realism to idealism: defending Sautrāntika realism as a

preparation for the proof of idealism
The theoretical hierarchy of idealism and realism entails that one cannot speak of
Dharmakīrti either as realist or as idealist. By default, he operates on the presupposition of
external reality, suspending arguments that realism is flawed, but in certain contexts he
explicitly argues for idealism and occasionally also addresses the relationship between these
two levels of analysis. Of particular interest in this regard, and for the refutation of external
objects as such, is an extended argument that makes a transition of realism to idealism.19
Here Dharmakīrti first defends a Sautrāntika-style theory of perception based on causation
and resemblance against Brahmanical interlocutors (301–319).20 He then turns to questioning
and ultimately rejecting the supposition, which underlies this theory, that perception is of
external objects (320–337). Although the initial defense of the Sautrāntika model may seem to
be of little relevance to proving idealism, it can be seen as the first step in Dharmakīrti’s
proof of idealism, for it establishes a key ingredient of it and also prepares the ground for
questioning realism in the first place.
The point of departure for Dharmakīrti’s discussion is the connection of cognition with a
particular object—or, in other words, cognition’s object-specificity. It is obvious (pratyakṣa) to
everyone from their experience that awareness is restricted (prativedana), first and foremost
insofar as it is limited to particular objects.21 Awareness is always awareness of something;
awareness does not occur to us pure and simple. But what accounts for cognition’s object-
specificity? Dharmakīrti argues that whatever differentiates cognition according to its object
must belong to cognition itself, or to cognition’s nature (ātman). Cognition is differentiated
according to its object by having the object’s form (ākāra, rūpa) within itself; this constitutes
cognition’s “resemblance” (sārūpya) to the external object.22 Nothing extraneous to cognition,
like one of its causes, can account for its object-specificity. One criterion for determining the
factor that accounts for it is to examine whether cognition’s object changes when that factor
changes. This, Dharmakīrti argues, is not true for the sense-faculty, one of the causes of
perception. Whether the sense-faculty is sharp (paṭu) or dull (manda) certainly affects and in
some way differentiates perception, but not in the relevant way of making it about a different
object, for the “condition for the close connection [between perception and its object] is
absent there” (pratyāsattinibandhanābhāvāt). But if one wished to argue that some specific
property of experience, caused by the external object, but not tantamount to resemblance,
accounted for object-specificity, then one would have to say what precisely this spurious
feature might be. As Dharmakīrti ridicules his interlocutor in the PVin, if some indeterminate
feature should determine of which object cognition is, things would be truly well
determined!23 It is true that others have specified properties of cognition that they might
claim to be caused by the object. But these are reducible to cognition’s having the object’s
form, to its form-possession. What is the “(mere) seeing of an object” (arthālocana), used to
define perception in a number of other schools,24 other than cognition’s having that object’s
form (310; PVin 1 32,15–33,1)? Similarly, the cognition of a qualifier (viśeṣaṇajñāna), such as
the color white, can only account for the cognition of what is qualified by it (viśeṣyajñāna), as,
for example, a white cow, if both are distinguished by having the form of their respective
object (313bcd; PVin 1 33,1–3). The sense-faculty, moreover, is common to every sense
perception (312; PVin 1 32,15). Every visual perception is caused by the visual sense-faculty,
hence the sense is not a suitable basis for distinguishing the seeing of blue from that of
yellow. The “sense-object contact” (indriyārthasannikarṣa) that is according to Nyāya,
Vaiśeṣika, or Mīmāṃsā among the causes of perception applies to an object in its entirety. But
one only sees objects with some of their qualities, like color, whereas others, like their being a
mass of atoms, remain unseen (316; PVin 1 33,8–10).25 In sum, if one holds that perception is
of an external object, one has to grant that perception itself has that object’s form. No other
factor involved in the causation of perception can account for its object-specificity. Whatever
feature of perception itself might be invoked turns out to be reducible to “resemblance,” as
specified in terms of form-possession, or is unsuitable, as the sense-object-contact.
In order to appreciate Dharmakīrti’s point here as well as his further strategy, we should
bear in mind that realists active prior to (or more or less simultaneous with) Dharmakīrti
appear to have considered the object-specificity of cognition as evidence for realism.26 The
Mīmāṃsaka Kumārila argued that cognition’s object-specificity was only possible on the
premise of external reality. His remarks from the beginning of the nirālambanavāda chapter
of the Ślokavārttika even suggest an argument to this effect: if cognitions were devoid of
external objects (nirālambana), scholastic distinctions like that between prior position and
established position (pūrvapakṣa/siddhānta), and philosophical distinctions such as that
between valid and invalid cognition would collapse—as would categories that form the
foundation of ritual science.27 Much like Dharmakīrti, Kumārila elsewhere states that it is
established for all living beings that something is apprehended having the form of blue or
yellow, long or short.28 But for Kumārila, if one recognizes the object-specificity of cognition,
one also has to recognize external reality as its foundation. Vasubandhu counters a similar
realist challenge at the beginning of the Viṃśikā by first pointing to dream experience: a
spatial or temporal restriction (niyama) of cognitions is also known to occur in dreams, and
dreams surely occur without external objects. Why should the same not also apply to waking
cognitions?29
Considering a realist argument for external reality from cognition’s restricted nature as the
backdrop of Dharmakīrti’s argumentation offers new perspectives. At this stage, Dharmakīrti
believes to have brought his imaginary opponent to agree that a perception has its object’s
form. The opponent would thereby be driven to concede the general point that cognition is a
type of entity that can have “form.” As is well known, this is one of the main points of
disagreement between Buddhist epistemologists and their Naiyāyika and Mīmāṃsaka
opponents, as these opponents attribute “form” to external objects, and deny that cognition
can have form; they are “advocates of formless cognition” (nirākārajñānavādin). By having
established that cognition has form, Dharmakīrti already sets the stage for his refutation of
external objects. If restricted awareness is nothing but cognition’s carrying the object’s form
within itself, why should cognition be of an external object in the first place? What is the
evidence that we cognize something external?30

Arguments against external objects

Dharmakīrti’s actual arguments against external objects can be divided into two groups,
although this division is not made explicit in the text. The first group of arguments points out
defects of the Sautrāntika theory of perception, while the second group argues on the ground
of the nature of cognition. It is striking that Kumārila distinguishes two types of arguments
that others advanced against external objects along quite similar lines: those based on an
examination of the object, and those based on (an examination of) the means of valid
cognition, that is, cognition as offering evidence for objects.31

Defects of the Sautrāntika theory

Dharmakīrti’s focus on pointing out defects specifically of the Sautrāntika model of
perception, and not of realism in general, might be construed as a weakness in argument, but
it becomes more plausible if the preceding defense of Sautrāntika realism is regarded as the
first step in an extended argument to prove idealism. For, if read in this manner, Dharmakīrti
first strove to commit his Brahmanical interlocutors to a causation-cum-resemblance account
of perception’s object-specificity, and now believes to have established that perception has the
form of its object. If the Sautrāntika theory is the most rational realist account of perception,
which Dharmakīrti believes to have established, then it is sufficient to point out the defects of
this theory in order to refute realism as a whole.
Dharmakīrti first (321cd) picks up the argument from incongruence sketched in Dignāga’s
Ālambanaparīkṣā and -vṛtti discussed above. While a number of minute and subtle atoms are
supposedly the cause of perceptual awareness, what appears in it is a single, “coarse” (sthūla)
form. Yet, this coarse, spatially extended form is not a feature of what causes perception, the
individual atom.32 The next stanza 322 continues by saying, “Therefore (i.e. given the
argument from incongruence), cognition does not have the form of the object; or, if it had,
that [form-possession] would be deviating (vyabhicārin).”33 The nature of the “deviation” is an
illustrative case to show how commentators read their own agendas into the text. According
to the commentator Manorathanandin, probably active in the second half of the eleventh
century, the brief mention of a “deviation” hints at a new argument that appeals to perceptual
illusion. If the Sautrāntika accounts for restricted awareness by stipulating a resemblance
with an external object, then this is fallacious because there are cases of restricted awareness
where there is no resemblance. In perceptual illusions like that of a mesh of hair or a double
moon experienced by someone suffering from the timira-disease (i.e., floaters), one is aware
of a specific object, yet there is no external object to which the appearance in cognition would
resemble.34 The Sautrāntika, wishing to resist such a move, might point to causation as an
additional condition: whatever is the object also serves as the cause for the cognition that
bears its form. But there is no real double moon that causes said illusion; the illusion would
turn out to be not a case that makes the independence of cognitive content from external
reality plausible, but simply an aberration, caused by some defect in the sense-faculty or some
other cause of cognition.35 The following stanza 323 then points out a further problem that
arises from the Sautrāntika account. Unlike Manorathanandin, the earlier commentator
Devendrabuddhi (ca. 630–690 CE) does not take Dharmakīrti’s mention of a “deviation” as a
reference to perceptual illusion, but rather directly connects it with the argument that follows
in 323.36 Dharmakīrti himself also does not speak of any additional “deviation” in the section
of the Pramāṇaviniścaya that otherwise closely parallels these stanzas in PV, nor does he
invoke an analogy with illusion there. This is worth emphasizing because earlier Yogācāra
sources use dream experiences to make the doctrine of mere-cognition plausible.37 As already
noted, Vasubandhu also appeals to the appearance of unreal objects in dreams in the very
beginning of the Viṃśikā. The Buddhist position is even presented with a formal “dreaming
argument” (Taber, 1992) by Kumārila, and Prajñākaragupta (c. 750–810 CE) and
Jñānaśrīmitra (c. 980–1030 CE), two significant and original thinkers in Dharmakīrti’s
tradition, also advanced formal dreaming arguments.38 Yet, Dharmakīrti nowhere makes such
an argument, and apparently did not place great weight on appeals to perceptual illusion in
his refutation of external objects.
The argument in 323 questions that causation-cum-resemblance unequivocally establishes
that perception has external objects. Assume a situation where a person has two perceptions
with identical object-appearances, for example, blue, in immediate sequence. This situation is
less contrived than it might initially seem. Given that the Sautrāntika assumes objects as well
as mental events to be of only momentary existence, any seemingly continuous perception
would in fact just be a succession of perceptual events with identical appearances. And many
of our perceptions, if not all, are seemingly continuous. In this situation, the earlier
perception is a cause of the later one; in the technical terminology that Abhidharmic analysis
developed to classify the causes of perception, it is the “immediately preceding homologous
condition” (samanantarapratyaya); hence we can dub this argument the
samanantarapratyaya-argument. Both perceptions have the same form of blue. The preceding
perception therefore fulfils both conditions for being an object—causation and resemblance—
and it could therefore just as well be considered the object of the later one! The Sautrāntika
believes that his definition of the object of perception by causation and resemblance limits the
role of the “object” to an external object, but this is inconclusive.39 The Sautrāntika responds
by pointing to a subsequent determinative cognition (adhyavasāya): When a determinative
cognition with the content “this was seen” or “this was heard” arises after a perception, it
must have been preceded by an experience of that which was seen or heard, that is, of the
object. But such a determination simply does not occur with respect to an immediately
preceding cognition, hence that cognition is not the object. We do not determine “this
preceding cognition was seen.” Yet, Dharmakīrti insists, it is precisely the close connection
(pratyāsatti) between perception and its object that is under scrutiny: only when such a
connection exists can a subsequent determination arise. That connection remains to be
accounted for.40 And, to complete Dharmakīrti’s argument, if it were to be accounted for by
causation and resemblance, then there would be no reason why the determination should not
just as well refer to the preceding and homologous condition; the initially raised problem
remains. Although the argument is premised on an ontology of exclusively momentary events,
it does not logically depend on it. All that is needed is a realist view that considers mental
events to have other mental events among their causes. When this is granted, a sequence of
two cognitions with the same mental image would trigger the problem that both the external
object and the preceding cognition fulfil the definition of an object of perception, that is, to
cause a subsequent cognition that has the same form.

Problems with resemblance

In a later section of the Pramāṇavārttika, Dharmakīrti points out additional problems of the
notion of resemblance that underlies Sautrāntika realism. If cognition resembled its putative
external object in all respects, it would no longer be cognition; it would be an external object.
But if it were only partially similar to the object, then “everything would be aware of
everything else.”41 The cognition of a pot, Devendrabuddhi explains, would be that of
potsherds, as the cognition does after all resemble potsherds in some respects because both
potsherds and cognitions have the property of being cognizable (*jñeyatva). If one were to
object that the cognition of the pot is not one of potsherds because it simply has the form
“pot,” then the next problem arises: the cognition of one pot would be a cognition of all pots.42
Cognition’s possession of an object’s form should make it indexically linked to one specific
object, to the one particular pot in front of my eyes. Alas, form-possession cannot do this
work, for it is a generic notion.43

Arguments from the nature of cognition

Dharmakīrti concludes from the argument from incongruence and from the
samanantarapratyaya-argument that there is nothing other—no external object—to be
experienced by cognition; cognition is only aware of itself. This is one of several claims that
come to be condensely expressed in the notion of svasaṃvedana, “reflexive awareness.” Now
that objects are obviously brought to awareness in a restricted fashion
(pratyakṣaprativedyatva), the phenomenon to be accounted for, is also just cognition’s own
nature. Cognition shines forth by itself (svayaṃprakāśate); it is not of anything other, of an
external object, and it is also, qua cognition, not brought to awareness by any further
cognitive act.44 The second group of arguments against external objects that can be detected
in Dharmakīrti’s works offers support for the first part of this two-part theorem: cognition is
not of anything else. To the extent that it is aware of objects, it is only aware of an internal
object-form contained within itself. Unpacking the two arguments in this group is not an easy
task, for they are less explicit than the arguments in the first group. They have also, in
traditional interpretations, been taken to serve additional purposes, over and above the
refutation of external objects—and for one of them, the famous sahopalambhaniyama-
inference, Dharmakīrti himself states that it can be used to establish that cognition has its
own form and that of its object on the basis of realist presuppositions (397); the inference
does not refute external objects on all interpretations he himself considers legitimate. As a
result of these complexities, a full discussion would lead us too far; a limited outline with a
focus on just how these arguments refute external objects if so understood will have to
suffice.
The first argument, referred to as the saṃvedana-argument by Takashi Iwata who provided
a meticulous analysis of it,45 is presented in such cryptic form in the Pramāṇaviniścaya that a
slightly elaborated paraphrase is more helpful than a literal translation:

What is called “awareness” (saṃvedana) is just appearing-in-a-certain-way (tathāprathana) because it

has that nature (tādātmyāt). This awareness is not of anything else, just like the awareness of cognition
itself is not of anything else. For this reason, too, it is not possible that awareness applies to a thing
other than cognition itself.46

Cognition cannot be of some other external thing; whatever it apprehends must be within
itself. Object-awareness is like cognition’s reflexive awareness of itself (ātmasaṃvedana), for
cognition is by nature just an intransitive “appearing-in-a-certain-way,” as opposed to a
transitive apprehension of something else. Elsewhere, Dignāga and Dharmakīrti deny that
cognition might perform any kind of activity (vyāpāra) directed at an external object.47 “To
appear” (prathate)48 is an intransitive verb, for it makes no sense to ask the “what” question
for an object: “What did you appear?” is a nonsensical question, as opposed to “what did you
read?” or “what did you dream about?”49 To liken cognition with transitive activities such as
cutting down a tree with an axe, as Dharmakīrti’s Brahmanical interlocutors as well as
Sarvāstivāda Buddhists are prone to do, may seem in line with ordinary language use, but it is
not analytically accurate. As Manorathanandin points out, when Dharmakīrti uses expressions
like “cognition cognizes itself” (dhīr ātmavedinī, 329), this is only metaphorical parlance. In
reality, cognition simply arises with awareness for its nature, comparable to light that has
luminosity for its nature.50
The saṃvedana-argument professes that both dimensions of awareness that belong to
cognition are ultimately one, and intransitive: its awareness of itself qua cognition, and its
awareness of an object’s form that it contains within itself. Readers familiar with the more
recent history of Western philosophy may feel tempted to link this argument with so-called
one-level accounts of consciousness as are characteristically advanced by philosophers in the
modern phenomenological tradition, such as Jean-Paul Sartre.51 Conscious mental states are
intrinsically and pre-reflectively conscious of themselves while being conscious of objects.
This self-consciousness is not subsequently produced, or otherwise separate from object-
consciousness. But the conclusion that Dharmakīrti draws here is one that phenomenologists
shun: given a one-level account of consciousness, cognition cannot be of an external object.
Dharmakīrti does not spell out the further implications himself, but the direction of his
argument seems to be that if cognition cannot apply to anything outside itself because it has
the nature of intransitive awareness, by implication whatever might be “out there” and does
not partake in cognition’s luminous nature (which is the foundation of its intransitivity) cannot
become manifest at all. As later authors in Dharmakīrti’s tradition put it, “insentient” (jaḍa)
objects simply cannot appear because to appear means to have the nature of “illumination” or
“shining forth” (prakāśa), which is exclusive to consciousness.52 Even if insentient objects
might be causes of cognitions, which the saṃvedana-argument when taken on its own does
not strictly speaking rule out, they still cannot possibly appear within cognition. External
reality is not perceptually accessible. This conclusion is one of the main obstacles for a
wholesale assimilation of Buddhist epistemology to contemporary phenomenology, however
instructive certain parallels might be.53
The second argument in this group is the famous sahopalambhaniyama-inference, the
inference from the necessary joint cognition of an object like blue color and its perception to
their “non-difference” (abheda). This inference has a particularly rich tradition of
interpretation. It is widely discussed among non-Buddhist philosophers and awaits full
philosophical assessment.54 At its locus classicus in the Pramāṇaviniścaya, the focus of the
inference is placed on the object’s non-difference from perception. Even though the form
(rūpa) of blue appears as different from experience, it is not different from it because the two
are necessarily cognized together, like the two moons seen by someone suffering from
floaters. Blue and its perception are cognized together because “there is no cognition of one
of these two without a [simultaneous] cognition of the form of the other.”55
For good reasons, commentators have rejected an interpretation of saha (“together,”
“joint”) as just indicating simultaneity, for there are after all many cases where ontologically
different objects are cognized simultaneously. Simultaneous perception of objects simply does
not equal identity.56 Rather, it must be necessary for perception to be cognized when an object
is perceived, and, conversely, that an object is perceived when its perception is cognized. The
second claim seems less controversial. It seems reasonable that when I am aware of a
perception that shows blue color, blue color is perceived. The first claim is more
controversial, for it is not obvious that a perception of an object is not possible without a
simultaneous awareness of that perception. In the Pramāṇaviniścaya, Dharmakīrti sets out to
justify this claim by arguing that “for someone who does not perceive perception, the
perception of the object is not established either.”57 Perception cannot be cognized after it
grasps its object, for this would lead to an infinite regress since perception itself would have
to be grasped by a further perception, and so on. It therefore has to be cognized
simultaneously with its object. But whether Dharmakīrti anywhere establishes the premise of
this argument, that all cognitions have to be cognized, remains to be studied; as presented in
PVin, the argument seems question-begging.58
Commentators identify two further passages in the Pramāṇavārttika as stating the same
sahopalambhaniyama-inference. Of these, the passage 387–390ab is closer to the
Pramāṇaviniścaya version and contains the main ingredients of the inference: the conclusion
is that there is no “separation” or “difference” (vivekitā, 389) of object and cognition, or, more
focused, that the object (artha) is not different from cognition, that it “does not extend
beyond” (avyatirekitva, 390) cognition. The reason is that the object is necessarily brought to
awareness simultaneously with cognition (sakṛt saṃvedyamānasya niyamena dhiyā saha
viṣayasya ..., 387; cf. also saṃvittiniyama in 388). In stanza 389, the necessity of the joint
perception is explicated in two claims that anticipate the corresponding two-part explication
in PVin: “no object is [observed being experienced] without awareness; nor is awareness
observed being experienced without an object” (nārtho ’saṃvedanaḥkaścid anarthaṃvāpi
vedanam ... saṃvedyamānaṃdṛṣtaṃ); recall that the corresponding part in the version in PVin
is “there is no cognition of one of these two without a [simultaneous] cognition of the form of
the other.”
In the passage 333–335, the realist first asks what would be wrong in assuming that an
external object is experienced. Dharmakīrti answers: “nothing at all! Only this [remains to be
asked:] Why would it be said that an external object is perceived?” At this point in the
discussion it is established for both parties that cognition has form. The question remains
whether this form originates from an external object or from imprints left by earlier
experience in the mental series. Dharmakīrti answers in stanza 335:

Because [something blue] is not apprehended without the additional qualifier (upādhi) of perception,
[and] because [blue] is apprehended when this [qualifier of perception] is apprehended, perception has
the appearance of blue. There is no external object by itself.59

The argument is very close to a sahopalambhaniyama-inference, if not fully identical with it:
the conclusion is that there is no external object “by itself” (kevalaḥ), a conclusion that can
plausibly be understood to mean that there is no external object that would be different from
cognition, that is., separate or independent from cognition. The reasoning to support this
conclusion consists in a joint apprehension, expressed in two claims that structurally
correspond to the ones from stanza 388. But there may be some significance to the
characterization of perception as an “additional qualifier” (upādhi) of the apprehended object.
It is one thing to say that when blue is apprehended, it is always apprehended as qualified by
its perception, but it is another thing to say that when blue is apprehended, its perception is
also apprehended. Whenever I perceive blue, I am aware of blue perceptually, but this does
not have to mean I am aware of the perception of blue (or of perceiving blue). The argument
presented in 335 may therefore be a weaker form of the sahopalambhaniyama-argument that
does not yet involve the innate reflexive awareness of perception, svasaṃvedana, in quite the
same way as the inference from PVin. But the conclusion, that there is no external object by
itself—independent from cognition—seems to be the same in all versions of this intriguing
argument.
Dharmakīrti continues, in 336, by stressing that cognition’s restricted nature does not
depend on external objects; a mechanism of awakened imprints manages to account for it.60
Earlier experience leaves traces in the mental series that give rise to object-specific
cognitions. Dignāga, as shown above, invoked a similar model at the end of the ĀP(V).
Dharmakīrti thereby takes up the realist’s challenge that Vasubandhu’s opponent and
Kumārila both raise: that restricted awareness demands external reality. Imprints in the
mental series also help to explain the difference between valid and invalid cognition that
according to Kumārila collapses without external objects: An invalid cognition is one arising
from imprints left behind in the mental series by a disturbed cognition, and that therefore
does not lead one to attain a desired goal. By contrast, a valid cognition is one arising from
strong imprints (dṛḍhavāsanā) that has an uninterrupted connection with the desired goal and
is reliable in action.61 Even the determination of the causal relation between seed and sprout
or fire and smoke, and drawing inferences on its basis, is possible without assuming external
objects (392–396), just on the basis of mental appearances; this is the doctrine of the wise
(viduṣāṃvādaḥ, 397). External objects are, in short, not required to explain any of the
phenomena that realists explain by them. Dharmakīrti suggests an inference that
nevertheless might be used to prove the existence of external objects: external objects could
be proven “from absence” (vyatirekāt). When all other causes for a perception are assembled,
and perception still does not arise, its absence logically implies that an additional cause is
needed—and that further cause might well be the external object, unless the idealist were to
claim that that additional cause is a special material cause of the cognition, that is, a
preceding mental episode in the same mental series.62 It seems then that this inference is a
theoretical possibility, to be used if the (superior) idealist account is not adhered to, if it is
suspended for the purpose of explaining things as they are in the world.

Conclusions: Dharmakīrti’s idealist position and aspects of logical method

Arguments from Dharmakīrti’s rich and extensive, and still largely unexplored analysis of
svasaṃvedana (425–539) might well have a further bearing on the refutation of external
objects. As the discussion of the second group of arguments has shown, Dharmakīrti’s
arguments are terse to the point of occasionally being cryptic, and only a close reading—
invariably guided by pointers from commentaries—can advance our understanding of them.
But, for now, let us suppose that the arguments discussed above are Dharmakīrti’s main
ones against external objects, and indicative of his strategy, assuming that he has one. First of
all, it is evident from them that Dharmakīrti nowhere produces straightforward arguments
that the external world does not exist. Like Dignāga, he does not advance the kind of
mereological arguments against external objects that Vasubandhu put forward in Viṃśikā 11–
15. Nor does he advance an argument from ignorance that, as we argued, can be uncovered
from the Viṃśikā. Dharmakīrti’s first group of arguments is quite involved, for it points out
defects in one specific realist theory, the Sautrāntika’s, that is in his view the most rational
one. These arguments lead Dharmakīrti to conclude that cognition is not of anything else—
and this conclusion is then supported by the second group of arguments that argue from the
nature of cognition. Nothing clearly indicates how Dharmakīrti conceived of the relationship
between the two groups of arguments. But when compared to the first group, the second
group, while far more difficult to interpret, presents stronger and more independent evidence
for the absolute imperceptibility of external objects. It is not just that the most rational realist
theory of perception has defects, but even more: there is something to the very nature of
cognition that makes it impossible for it to be of external objects.
 Given that Dharmakīrti in all these arguments reveals himself as being chiefly concerned
with what we perceive, it seems prima facie reasonable to read him as pursuing the point that
the objects we experience are not, and indeed cannot be, outside consciousness. This may
initially reinforce the plausibility of Dan Arnold’s distinction between an “epistemic idealism”
(Dignāga/Dharmakīrti) and an ontologically committed “metaphysical idealism”
(Vasubandhu).63 But one challenge to this distinction was recently raised by Isabelle Ratié. If
Dharmakīrti indeed argues for the complete and utter imperceptibility of the external world,
is this not tantamount to arguing for its nonexistence in a different way? Is not epistemic
idealism already ontologically committed because what is absolutely imperceptible can only
be nonexistent? After all, in Dharmakīrti’s ontology “existence” is defined as causal capacity,
through the notion of arthakriyā. If an object is fully imperceptible and can never produce
even the minimal effect of a perceptual awareness, is it not thereby nonexistent?64
At this point, considerations of logical method and theory become pertinent. Dharmakīrti
was active after Vasubandhu, and the relationship between their views and arguments has
therefore historical dimensions that abstract comparison cannot properly grasp. Dharmakīrti
introduced significant innovations in logical theory, notably the restriction of acceptable
logical reasons to three, and only three, types. We should not only expect that his various
proofs are in accordance with his own theory or at least do not openly violate it (however
difficult the interpretation of that theory may turn out to be), but we should also consider that
features of logical theory determine the nature of claims that can be proven with its help.
Might the refutation of external objects be such a case? And is there anything in
Dharmakīrti’s work that signals how he might have, or could have, conceived of Vasubandhu’s
arguments against external objects from the vantage point of his own position?
Dharmakīrti himself does not explicitly address Vasubandhu’s arguments. But as Arnold
(2008) and Ratié (2014) have noted, a rare passage in Manorathanandin’s
Pramāṇavārttikavṛtti on 335 speaks to this difference; it deserves to be quoted in full:

[Sautrāntika opponent:] No! Even so, [i.e.] even if there is no argument proving an external object,
which is [on your view?]65 beyond the reach of the senses, the non-existence of external objects is not
[thereby] established. [Manorathanandin:] Because [we] establish what [we] intend only as far as
[saying] that cognition appears, whereas the external object does not appear at all, we do not have any
regard for negating the external object, which behaves like an [imperceptible] demon [and] is without a
means of valid cognition that proves it.66 But if the opponent were to strongly insist on negating the
[external object], he should be made to examine the master [Vasubandhu’s] negation of atoms
according to whether one supposes that [the external object] has parts or is partless.67

The Sautrāntika opponent points out that Dharmakīrti’s argument does not prove the
nonexistence of external objects, even if he might have managed to establish that there is no
evidence for their existence. This, now, is a classic objection to arguments from ignorance:
Absence of evidence does not constitute evidence of absence. Manorathanandin does not
dispute the objector’s point: indeed, Dharmakīrti did not prove the nonexistence of external
objects by showing that there is no evidence for them. Manorathanandin does not say this
explicitly, but a look at Dharmakīrti’s theory of inference supports an even stronger point:
Dharmakīrti could not have proven the nonexistence of external objects in this way.
In Dharmakīrti’s logical framework, negation is proven through a special type of reason
called “non-apprehension” (anupalabdhi), but the scope of this reason is very limited. One can
only infer that objects which would be necessarily perceived in a certain situation if they
existed, but which are not perceptually apprehended, are suitable for cognitive, linguistic,
and physical treatment as nonexistent.68 This type of reason does not permit universal
ontological denial, for it presupposes that the negated object is of a kind that will necessarily
be perceived (given the presence of all additional causes for its perception) if it exists. For
that to be established requires that the object is one that can be perceived in principle. Thus
restricted, inferences based on the non-apprehension of a perceptible object can only prove
the occasional (situational) absence of objects, not the nonexistence of an entire class of
objects. Throughout his works, Dharmakīrti offers different reasonings for why general
arguments from ignorance are flawed.69 The most straightforward of these is that remote
objects (viprakṛṣṭa)—objects remote in time or place, or by their very nature, such as piśāca-
demons70—cannot be proven as absent on the ground that they are not apprehended through
perception, inference, and scripture because they lack the causal capacity to produce a
cognition of themselves. They might therefore exist without giving rise to a pramāṇa, and the
absence of a pramāṇa for them cannot prove their nonexistence.
Dharmakīrti, in short, effectively eliminates arguments from ignorance from his theory of
inference. It is, then, in keeping with his logical theory that the possibility for proving the
nonexistence of external objects through arguments from ignorance is not even considered in
Dharmakīrti’s works (as far as I can tell), and that he does not conclude from the absence of
evidence for external objects that they do not exist. Dharmakīrti’s proofs that external objects
are absolutely imperceptible may entail their nonexistence, along the lines suggested by
Ratié. The difference between epistemic and metaphysical idealism may thereby become
minimized and lose its significance as a heuristic device for detecting different varieties of
mere-cognition in Indian Buddhist thought. But his logical theory prevents Dharmakīrti from
proving the nonexistence of external objects through arguments that merely demonstrate
their imperceptibility. This theory opens up an evidential gap, in a manner of speaking,
between imperceptibility and nonexistence. A Dharmakīrtian might claim that totally
imperceptible objects are as good as nonexistent, but when further pushed to prove it would
have to resort to other arguments than those used by Dharmakīrti to refute external objects,
one where absence of evidence does not become evidence for absence.
Such arguments could, of course, be precisely those that Vasubandhu put forward in
Viṃśikā 11–15. According to Manorathanandin, Vasubandhu’s mereological arguments merit
consideration only if the opponent were to stubbornly insist71 on negating the external object
(proving its nonexistence) after Dharmakīrti’s proofs of its imperceptibility have been pointed
out to him. Vasubandhu’s arguments are not declared to be invalid; there is nothing formally
wrong with them. It is only that they are irrelevant because they prove something that is not
of value, since the epistemic inaccessibility of the external world is all there really needs to be
shown. Metaphysical idealism, to use Arnold’s term, is for those who insist on pursuing
irrelevant questions.72

Notes
1. Research for this chapter was undertaken within the research group “Practices of Argumentation in
Transcultural Perspective” (directed by Joachim Kurtz and Birgit Kellner) of the Cluster of Excellence
“Asia and Europe in a Global Context—the Dynamics of Transculturality” at the University of
Heidelberg. I am, as always, grateful to John Taber, whose readings of Buddhist philosophy have
greatly enriched my approach (though he may not agree with my conclusions).
2. This date is conjectured by Deleanu (2006, vol. 2, pp. 186–194).
3. Cf. Schmithausen (2001) for a succinct historical overview on vijñaptimātra(tā). Cf. Krasser (2012) for
a recent proposal to move Dharmakīrti up from Frauwallner’s 600–660 CEand place his time of activity
in the middle of the sixth century. Krasser argues that Bhāviveka (490/500–570 CE) referred to
Dharmakīrtian ideas and arguments. The implications of this proposal, as well as its methodology,
remain to be assessed in consideration of a wider context.
4. Some positions in this debate are reviewed in Kellner and Taber (2014).
5. Vś 12. See Kapstein (1988) for an insightful philosophical appreciation of this and other arguments
from Vś 11–15, and Oetke (1992) for a rigorously argued different reading.
6. Hayes (1988, p. 100).
7. Oetke (1992).
8. Kellner and Taber (2014).
9. In Kellner and Taber (2014), we discuss arguments from ignorance in Indian philosophy more
generally, and also address their logical limitations and the parameters that make some of them
plausible abductive evidence (though they are not deductively valid).
10. The ĀP(V) are not extant in Sanskrit (save for a few quotations), but only preserved in Chinese and
Tibetan translation. Asterisks indicate that Sanskrit terms of the ĀP(V) are tentatively reconstructed
from the Tibetan. My analysis is based on the Tibetan translation, and in its general contours seems
consistent with the Dharmakīrtian tradition in India. Interpretations in Tang China, and even the
translations by Paramārtha and Xuanzang, seem to differ to some degree from my analysis. Cf. Lin
(2007 and 2014), and Lusthaus (2014). For a recent philosophical appraisal of the ĀP(V), see Siderits
(forthcoming a).
11. The Chinese exegete Xuanzang, to whom we owe much insight into the intricate doctrinal fabrics of
medieval Indian Buddhist thought, attributes this definition to “Hīnayāna masters” in general, with the
exception of Saṃmitīyas (De La Vallée Poussin [1928, p. 42f.]).
12. Literally: “when [x]​ is present [or] absent, [y] has that (i.e., presence or absence).” For a different
construal, cf. Frauwallner (1930, p. 183). My construal is based on Uddyotakara’s explication of the
same principle to explain the relationship between pramāṇa and pramāṇaphala at NV 6,11f. By
“logicians,” Dignāga likely referred to Naiyāyikas (though not to Uddyotakara himself, who was active
after him).
13. Arnold (2008, p. 15). Arnold here compares Dharmakīrti, not Dignāga, with Vasubandhu, but basically
considers Dharmakīrti’s position to be the same that I have reconstructed for the ĀP(V).
14. I am grateful to Robert Sharf for bringing the relevance of this point to my attention.
15. See Dreyfus (1997), McClintock (2003), and Dunne (2004) for analyses of this hierarchy in terms of a
“sliding” or “ascending” scale of analysis, and Kellner (2011a) for a critical assessment of Dunne’s
version of it on the basis of some of the material covered in the following. Aspects of the following
section were also treated in Kellner (2009) (in German). In Kellner (2011a and b) I suggested to
characterize the different positions as “externalist” and “internalist” (rather than “realist” and
 “idealist”), but in the present context the latter, more established distinction is without harm, since the
nature of the positions themselves is the topic of enquiry.
16. In the following three-digit numbers without any further qualification refer to stanzas in the PV’s
chapter on perception. When variants are not explicitly discussed, the Sanskrit text corresponds to that
given in the edition of Tosaki (1979 for stanzas 1–319; 1985 for stanzas 320–539). The verse numbering
also corresponds to Tosaki’s. Tosaki did not have access to manuscript materials that became
accessible more recently and permit more comprehensive text-critical work on the PV and its
commentaries. Cf. Kellner (2010a) for a comprehensive review of the PV’s editorial history and extant
witnesses, and for first results of text-critical analysis, resulting in some emendations of Tosaki’s text
and a better understanding of the mechanisms resulting in textual change.
17. For in-depth studies of ignorance in Dharmakīrti’s thought, cf. Eltschinger (2009; 2010).
18. Cf. also Kellner (2009, 70f.).
19. This argument occurs in the first part of his treatment of the means of valid cognition and its result
(pramāṇa/pramāṇaphala), at PV 3.301–337 and the close parallel PVin 1 30,9–36,9, which takes
Dignāga’s pithy remarks in PS(V) 1.8cd-10 for its point of departure. Dignāga employs an idealist
analysis of perception in this context, but does not offer a refutation of external objects in the PS(V),
which presents itself as the culmination of his life-work and was therefore most likely composed after
the ĀP(V).
20. For this section, cf. also the helpful brief summary in Taber (2005, p. 197f., n. 98), where
correspondences in the pratyakṣapariccheda of Kumārila’s ŚV are pointed out.
21. The peculiar term “prativedana,” “restricted awareness,” occurs only at 320 (corresponding to PVin 1
33,11), but the object-specificity of perception constitutes the main explanandum throughout the entire
section 301–337. The commentator Manorathanandin expands the compound prativedanam as
pratiniyataṃvedanam, and interprets it to also include cognition’s restriction to a particular mental
series, not only to a particular object, cf. M1 215,9: nīlādyākāreṇa
pratiniyataṃvedanaṃpratisantānaniyataṃvā.
22. 302–305, corresponding to PVin 1 31,4–32,3.
23. PVin 1 31,11–32,2: na hi paṭumandādibhiḥsvabhedair bhedakam apīndriyārthenaitad ghaṭayati, tatra
pratyāsattinibandhanābhāvat. asty anubhavaviśeṣo ’rthakṛtaḥyata iyaṃpratītiḥ, na sārūpyād iti cet.
atha katham idānīṃsato rūpaṃna nirdiśyate? nedam idantayā śakyaṃnirdeṣṭum. anirūpitena nāmāyam
ātmanā bhāvān vyavasthāpayatīdam asyedaṃneti suvyavasthitā bhāvāḥ. Cf. also 305cd. The term
“pratyāsatti” for the close connection between a cognition and its object also occurs in 324, cf. p. 111.
24. Cf. Sāṅkhyakārikā 28, Praśastapādabhāṣya (Bronkhorst and Ramseier, 1944: 44,11 and 45,11f.), and
ŚV pratyakṣapariccheda 71–72a, 112–113 (Taber, 2005).
25. The example is taken from M1 213,22. Dharmakīrti’s basic argument goes back to PSV 1 10,8–12 (cf.
esp. sarvātmanā sannikarṣāt). The commentator Jinendrabuddhi traces this sannikarṣapramāṇavāda
back to a Vaiśeṣika author named Śrāyasaka (PSṬ1 118, 15). This author is not known from any other
sources.
26. The relative chronology of Dharmakīrti and Kumārila is a complicated subject, but suffice it to say that
they were most likely near-contemporaries, perhaps even contemporaries, so that interaction—each
responds to the other—cannot be ruled out.
27. ŚV nirālambanavāda 1–3; cf. Taber (2010, p. 279f.).
28. ŚV śūnyavāda 5 (cf. Taber, 2010, p. 283): tatra tāvad idaṃsiddhaṃsarvaprāṇabhṛtām api //
grāhyataṃnīlapītādidīrghādyākāravastunaḥ//
29. On our new reading of the Viṃśikā, Vasubandhu does not put forward a veritable “dreaming
argument” for idealism, but rather uses this argument to dispute that there is inferential evidence for
external objects (cf. Kellner and Taber, 2014).
30. 320c: tad (sc. prativedanam) arthavedanaṃkena, corresponding to PVin 1 33,12.
31. ŚV nirālambanavāda 17: arthasya parīkṣaṇāt ... pramāṇam āśritaḥ... (Taber, 2010, p. 280).
32. At 195–196, Dharmakīrti presents a refined version of the Sautrāntika theory: many “conglomerated”
(sañcita) atoms are said to be the external object of cognition, for the single atoms have the distinctive
feature (viśeṣa)—of being capable to producing a perception—only when they are together, not
individually. This explains why many things are the object of (nonconceptual) perception, yet without
being a universal (sāmānya), which according to Dignāga and Dharmakīrti has to be grasped by
conceptual cognition and cannot be apprehended by sense perception. But even in this refined theory,
the atoms lack a coarse appearance (211). Cf. Dunne (2004), for extended discussion of this section of
the PV. In the Pramāṇaviniścaya, Dharmakīrti complements Dignāga’s argument by adding a refutation
of the “whole” (avayavin) of the Vaiśeṣika, as an example for a singular, spatially extended object; cf.
PVin 1 34,10–35,6, with parallels in PV 2.84–85; cf. Funayama (1990), based on the Tibetan translation
of PVin.
33. 322ab: tan nārtharūpatā tasya satyāṃvā vyabhicāriṇī. Cf. also 320: tad arthavedanaṃkena tādrūpyād
vyabhicāri tat / “By what is that (i.e. restricted awareness) an awareness of an external object? On
account of [cognition’s] having the form of that [object]? That is deviating.”
34. M1 216,9 on 322 (dvicandrajñānādiṣu), and also M1 215,14f on 320:
dvicandrakeśoṇḍūkajñānādyākārasyārtham antareṇāpi bhāvāt.
35. M1 216,10–12, introducing 323: na kevalād arthasārūpyād arthasaṃvedanatvaṃyena vyabhicāraḥsyāt,
kiṃtarhi sārūpyatadutpattibhyām. te ca dvicandrajñānādīnāṃna staḥ, candradvayasyābhāvāt
tadutpatter ayogāt.
36. Det D219b4=P257b3. Devendrabuddhi makes no reference to the double moon or other illusory
objects in his commentary on 320 and 322.
37. MSg II.14, ASBh 42,6f.
38. ŚV nirālambanavāda 23. Cf. Taber (1992, p. 219; 1994), Kobayashi (2011) (arguing that Vś 1 does
present a formal inference), and Kellner and Taber (2014, p. 736, n. 94) (arguing that this is a less
charitable reading of the stanza). In our new reading of the Vś, we take the first part of the Viṃśikā
(Vś. 1–7) to be directed against inferential evidence for external objects: there are no reasons to
postulate external objects because a number of facts that are usually explained by them can also be
accounted for through the doctrine of mere-cognition.
39. 323: tatsārūpyatadutpattī yadi saṃvedyalakṣaṇam / saṃvedyaṃsyāt
samānārthaṃvijñānaṃsamanantaram // PVin 1 33,12–34,1: anantaraṃtarhi
vijñānaṃtulyaviṣayaṃviṣayaḥprāpnoti. Cf. also Kellner (2011a, p. 295) (following an interpretation first
given in Tosaki [1985, p. 7]), and Arnold (2008, p. 10 (advocating a different interpretation).)
40. 324–325: idaṃdṛṣṭaṃśrutaṃvedam iti yatrāvasāyadhīḥ/ sa tasyānubhavaḥsaiva pratyāsattir vicāryate
// dṛśyadarśanayor yena tasya tad darśanaṃmatam / tayoḥsambandham āśritya draṣṭur eṣa viniścayaḥ//
(For vedam in 324, PrA’ reads cedam; for darśanaṃ in 325, attested in MA, PVZh, Rt, sādhanam is
attested by PVt, PrA’, PrB.) Cf. also PVin 1 34,2–5; Kellner (2011a, p. 295). For a different interpretation
of 324, see Arnold (2008, p. 10f.).
41. 434: sarvātmanā hi sārūpye jñānaṃajñānataṃvrajet / sāmye kenacid aṃśena syāt
sarvaṃsarvavedanam // The stanza is quoted in commentaries on ŚV śūnyavāda 20 with reversed
halves, and with tu for hi (Kellner, 2010a, p. 180, n. 54). The version cited here is preserved in PrA’, PrB,
and PVt. For the first half, Det, MA, Rt, and PVZh read na ca sarvātmanā sāmyaṃajñānatvaprasaṅgataḥ,
“And [cognition] is not completely the same [as its object], for then it would absurdly follow that it is
not cognition.” For further text-critical discussion, cf. Kellner (2010a, p. 200, esp. n. 111).
42. Det D244b5f = P289b4–7. The second problem is also mentioned in M1 248,11.
43. Cf. King (2005) for similar arguments against likeness-based theories of perception in medieval
European philosophy.
44. 326–327: ātmā sa tasyānubhavaḥsa ca nānyasya kasyacit / pratyakṣaprativedyatvam api tasya
tadātmatā // nānyo ’nubhāvyas tenāsti tasya nānubhavo ’paraḥ/ tasyāpi tulyacodyatvāt tat svayaṃtat
prakāśate // Cf. Kellner (2010a, pp. 196–199) for a discussion of variants in 327. Most significantly, for
tat svayaṃtat prakāśate (“Therefore, that [cognition] shines forth by itself”), which is attested in MA,
PVZh, Rt, and probably also Det, the witnesses associated with Prajñākaragupta’s commentary
(including PVt) attest svayaṃsaiva prakāśate, which was also adopted by Tosaki and corresponds to the
parallel stanza PVin 1.38 (“It is this very [cognition, sc. buddhiḥ] that shines forth by itself”). As
outlined in Kellner (loc. cit.), the PVin stanza selectively influenced the transmission of its counterpart
in the PV. In place of tasyāpi tulyacodyatvāt in PV, PVin 1.38 has grāhyagrāhakavaidhuryāt.
45. Cf. Iwata (1991, pp. 9–15).
46. PVin 1 42,3–6: saṃvedanam ity api tasya tādātmyāt tathāprathanam, na tad anyasya kasyacid
ātmasaṃvedanavat. tato ’pi na tad arthāntare yuktam. Cf. Iwata (1991, p. 9) (based on the Tibetan
translation; the Sanskrit was not available at the time).
47. PS 1.8cd (Kellner [2010b, p. 219], with further discussion of background in Abhidharma and Yogācāra
literature) and PSV ad PS 1.9d (op. cit., p. 223); for Dharmakīrti, cf. 308=PVin 1.37.
48. The verb form prathate is used in 349.
49. Cf. Legrand (2009) for a helpful clarification of the transitivity/intransitivity of consciousness.
50. M1 218,5–7. Cf. Watson (2014) for an illuminating (!) overview of different ways in which the light
analogy is used in controversies between idealists in Dharmakīrti’s tradition and Naiyāyikas.
51. Zahavi (2005, 20ff.).
52. Cf. Watson (2014, p. 415, especially n. 37). One of the clearest articulations of this trope is TS 2000:
vijñānaṃjaḍarūpebhyo vyāvṛttam upajāyate / iyam evātmasaṃvittir asya yājaḍarūpatā //
53. Cf. Coseru (2012) for an illustrative attempt to work toward such an assimilation.
54. Taber (2010, 292f.) gives some clues as to which direction such an assessment might take.
55. PVin 1.54cd: sahopalambhaniyamād abhedo nīlataddhiyoḥ. PVin 1 40,1–3: na hi bhinnāvabhāsitve ’py
arthāntaram eva rūpaṃnīlasyānubhāvāt tayoḥsahopalambhaniyamād dvicandrādivat. na hy anayor
ekākārānupalambhe ’nyopalambho ’sti.
56. Cf. Iwata (1991, 66–103) for a detailed overview of commentarial interpretations of saha in
sahopalambhaniyama, including rejections of simple simultaneity. Dharmakīrti uses the word
“simultaneous” (sakṛt) in 387, but also makes it clear in the immediately following stanza 388 that
there is no necessity of a (joint) awareness for different objects like blue or yellow (saṃvittiniyamo
nāsti bhinnayor nīlapītayoḥ).
57. PVin 1.55ab: nāpratyakṣopalambhasyārthadṛṣṭiḥprasidhyati. Cf. Kellner (2011b, p. 420, n. 28) for a
justification of this translation from the context, against one that takes the genitive
apratyakṣopalambhasya as an absolute genitive, following the Tibetan translation dmigs na: “if
perception is not perceived . . .”
58. In Kellner (2011b), I argued that Dharmakīrti does not establish this claim in the Pramāṇaviniścaya
passage, but it cannot be ruled out that some of the arguments presented in the second half of the
Pramāṇavārttika’s chapter on perception, still largely unexplored, fulfil this task. For further
philosophical reflection on the infinite regress, cf. Siderits (forthcoming b).
59. 333–335: yadi bāhyo ’nubhūyeta ko doṣo naiva kaścana / idam eva kim uktaṃsyāt sa bāhyo ’rtho
’nubhūyate // yadi buddhis tadākārā sāsty ākāraviśeṣiṇī / sā bāhyād anyato veti vicāram idam arhati //
darśanopādhirahitasyāgrahāt tadgrahe grahāt / darśanaṃnīlanirbhāsaṃnārtho bāhyo ’sti kevalaḥ//
(333d: for ’nubhūyate PrA’ reads ’nubhūyeta; 334: for -viśeṣiṇī PrA’ reads -niveśinī; 335: kevalaḥ PrB,
and perhaps also PrA’ (difficult to read), PVZh, against kevalam MA. The Tibetan translation yan gar
 found in Det, PVt, and Rt is inconclusive. Cf. Taber (2010, p. 291) for translation and discussion of these
stanzas.
60. 336: kasyacit kiñcid evāntarvāsanāyāḥprabodhakam / tato dhiyāṃviniyamo na bāhyārthavyāpekṣayā //
61. PVin 1 43,14–44,6; cf. Krasser (2004, p. 143f.) for text and translation.
62. PVin 1 58d: bāhyasiddhiḥsyād vyatirekataḥ, elaborated in PVin 1 43,10–12: satsu samartheṣu anyeṣu
hetuṣu jñānakāryāniṣpattiḥkāraṇāntaravaikalyaṃsūcayati. sa bāhyo ’rthaḥsyāt, yady atra kaścid
upādānaviśeṣābhāvakṛtaṃkāryavyatirekaṃna brūyāt. The basic argument is also offered in 390d–
391ab (Krasser, 2004, p. 142f.). Some traditional interpreters seem to take this inference as a response
by the Sautrāntika to criticism by the Yogācāra, while others construe it as an expression of the
Sautrāntika’s view that the external object is only inferred, and not perceived (see Kyūma [2011, p.
314, n. 28] for textual references).
63. Cf. above p. 106.
64. Ratié (2014, p. 361).
65. It is not evident from the text whether the qualifier that the object is beyond the reach of the senses
expresses the Sautrāntika’s own position or that which he attributes to the Yogācāra. The former
entails that the Sautrāntika here claims the external object can only be inferred, and not perceived, but
this is not such a fixed position that it can be unproblematically supplied to Manorathanandin’s
argument (cf. above n. 62). But this point of uncertainty does not affect the main argument.
66. Here my translation follows Ratié (2014, p. 359, n. 22), against Arnold (2008, p. 16).
67. M1 220, 16–20 (MA 43a4–6): na, tathāpi parokṣasya bāhyasya sādhakasyābhāve ’pi nābhāvasthitir iti
cet, pratibhāsamānaṃjñānaṃbāhyaṃtu na pratibhāsata eveti tāvataivābhimatasiddheḥ,
sādhakapramāṇarahitapiśācāyamānabahirarthaniṣedhe nāsmākam* ādaraḥ. yadi tu
tanniṣedhanirbandho garīyān sāṃśatvānāṃśatvakalpanayā paramāṇupratiṣedha**
ācāryīyaḥparyeṣitavyaḥ. (* For niṣedhe nāsmākam, the edition prints niṣedhenāsmākam. The
manuscript MA reads niṣedhesmākam, with -nā- added in the bottom margin with a correction sign in
the actual text. ** For paramāṇupratiṣedha MA, the edition M1 prints paramāṇupratiṣedhe.) Cf. also
Ratié (2014, p. 359, n. 23). My translation follows Ratié (2014, p. 358) (with minor and largely stylistic
differences), against Arnold (2008, p. 16) (who only translates Manorathanandin’s response).
68. Cf. esp. Kellner (1999; 2003) for further elaboration of Dharmakīrti’s complicated theory of non-
apprehension.
69. For a brief overview, cf. Kellner and Taber (2014).
70. To be precise, piśācas are thought of as imperceptible by their nature (svabhāva) for human beings,
that is, in relation to a particular type of cognizing subject. Cf. Kellner (1999) and now especially the
pertinent remarks in Ratié (2014).
71. Interpreting Manorathanandin’s nirbandha as “insistence” (Ratié: “obstinacy”) is one significant
departure from Arnold’s interpretation, where the word is taken (oddly) to refer to a neutral desire.
72. Ratié (2014) entertains both possibilities in her interpretation of Manorathanandin’s passage: that the
ontological question becomes irrelevant or that epistemic idealism already entails an ontological
position. Considering Dharmakīrti’s elimination of arguments from ignorance, the second possibility
seems a less likely account of the systematic implications within Manorathanandin’s rich passage—
even though, as Ratié then expounds, later Śaiva authors exploited it.

References
Primary sources in Sanskrit and canonical Tibetan translations

ĀP Dignāga’s Ālambanaparīkṣā. See Frauwallner (1930).

ĀPV Dignāga’s Ālambanaparīkṣāvṛtti. See Frauwallner (1930).
ASBh Abhidharmasamuccaya-bhāṣyam. Deciphered and edited by Nathamal Tatia.
2nd edn. Patna: K.P. Jayaswal Research Institute 2005.
Det Devendrabuddhi’s Pramāṇavārttikapañjikā. Tibetan translation. Tshad ma
rnam ’grel gyi dka’ ’grel, translated by Subhūtiśrī(śānti) and (Rma) Dge ba’i
blo gros. D 4217 Che 1–326b4, P 5717b Che 1–390a8.
M1 Dharmakīrti’s Pramāṇavārttika with a Commentary by Manorathanandin.
Edited by Rāhula Sāṅkṛityāyana. Appendix to Journal of the Bihar and Orissa
Research Society 24–26 (1938–40).
MA Sanskrit manuscript (on paper) of Manorathanandin’s Pramāṇavārttikavṛtti,
written by Vibhūticandra. In: Shigeaki Watanabe (ed.), A Sanskrit
Manuscript of Manorathanandin’s Pramāṇavārttikavṛttiḥ. Facsimile Edition.
Patna/Narita: Bihar Research Society/Naritasan Institute for Buddhist
Studies, 1998.
MSg Asaṅga’s Mahāyānasaṅgraha. La somme du Grand Véhicule d’Asaṅga
(Mahāyānasaṃgraha) [édité et traduit] par Étienne Lamotte. Louvain-la-
Neuve, Université de Louvain, Institut Orientaliste, 1973.
NV Nyāyabhāṣyavārttika of Bhāradvāja Uddyotakara. Edited by Anantalal
Thakur. Delhi: Indian Council of Philosophical Research, 1997.
PrA’ A modern transcript of an incomplete paper ms. of Prajñākaragupta’s
Pramāṇavārttikālaṅkārabhāṣya, written by Vibhūticandra, extending from
 the commentary on PV 3.302 to the end of the work. Microfilm of the Nepal-
German Manuscript Preservation Project in Kathmandu (reel no. A1219/26)
(cf. Kellner 2010a, pp. 167 and 168).
PrB Palm-leaf manuscript of Prajñākaragupta’s Pramāṇavārttikālaṅkārabhāṣya.
In: Shigeaki Watanabe (ed.), Sanskrit Manuscripts of Prajñākaragupta’s
Pramāṇavārttikabhāṣyam. Facsimile Edition. Patna/Narita: Bihar Research
Society/Naritasan Institute for Buddhist Studies, 1998.
PS Dignāga’s Pramāṇasamuccaya. Steinkellner, Ernst (ed.). 2005. Dignāga’s
Pramāṇasamuccaya, Chapter 1: A Hypothetical Reconstruction of the
Sanskrit Text with the Help of the Two Tibetan Translations on the Basis of
the Hitherto Known Sanskrit Fragments and the Linguistic Materials Gained
from Jinendrabuddhi’s Ṭīkā.
http://www.ikga.oeaw.ac.at/Mat/dignaga_PS_1.pdf, with revisions of 2014:
http://www.ikga.oeaw.ac.at/mediawiki/images/f/f3/Dignaga_PS_1_revision.pdf
(last accessed June 25, 2015).
PSV Dignāga’s Pramāṇasamuccayavṛtti. See PS.
PSṬ1 Jinendrabuddhi’s Pramāṇasamuccayaṭīkā. Ernst Steinkellner, Helmut
Krasser, Horst Lasic (eds.): Jinendrabuddhi’s Viśālāmalavatī
Pramāṇasamuccayaṭīkā, Chapter 1. Part I: Critical Edition, Part II:
Diplomatic Edition. [Sanskrit Texts from the Tibetan Autonomous Region 1].
Beijing; Vienna: China Tibetology Publishing House; Austrian Academy of
Sciences Press, 2005.
PV 3 Dharmakīrti’s Pramāṇavārttika, chapter 3 (pratyakṣa). See Tosaki (1979,
stanzas 1–319; 1985, stanzas 320–539).
PV 2 Dharmakīrti’s Pramāṇavārttika, chapter 2 (pramāṇasiddhi). See Miyasaka
(1971–72).
PVt Tibetan translation of Dharmakīrti’s Pramāṇavārttika. Tshad ma rnam ’grel
gyi tshig le’ur byas pa, translated by Subhūtiśrīśānti and (Rma) Dge ba’i blo
gros, revised by *Bhavyarāja (Skal ldan rgyal po) and (Rṅog) Blo ldan śes
rab, retranslated or revised by Śākyaśrībhadra and Sa skya paṇḍita. D 4210
Ce 94b1–151a7, P 5709 Ce 190a4–250b6.
PVZh Zha lu ri phug manuscript of PV, readings reported in: Pramāṇavārttikam by
Ācārya Dharmakīrti. Edited by Rāhula Saṅkṛtyāyana. Appendix to Journal of
the Bihar and Orissa Research Society 24 (1938).
PVin 1 Dharmakīrti’s Pramāṇaviniścaya. Chapters 1 and 2. Edited by Ernst
Steinkellner. [Sanskrit Texts from the Tibetan Autonomous Region 2].
Beijing; Vienna: China Tibetology Publishing House; Austrian Academy of
Sciences Press, 2007. For Corrigenda, cf. Wiener Zeitschrift für die Kunde
Südasiens 51 (2007–8), 207–208, as well as
http://ikga.oeaw.ac.at/Mat/steinkellner07_ corrigenda.pdf (last visited June
26, 2015).
PVSV Dharmakīrti’s Pramāṇavārttikasvavṛtti. The Pramāṇavārttikam of
Dharmakīrti: the First Chapter with the Autocommentary. Edited by Raniero
Gnoli. Serie Orientale Roma 23. Rome: Istituto Italiano per il Medio ed
Estremo Oriente, 1960.
Rt Tibetan translation of Ravigupta’s Pramāṇavārttikavṛtti. Tshad ma rnam
’grel gyi ’grel pa las le’u gsum pa. D 4225, Phe 1–174a7; P 5722, Phe 1–
208a7. Translators unknown.
Sāṅkhyakārikā Sāṅkhyakārikā. Yuktidīpikā. The Most Significant Commentary on the
Sāṅkhyakārikā. Critically edited by Albrecht Wezler and Shujun Motegi.
Stuttgart: Franz Steiner Verlag, 1998.
ŚV Svāmī Dvārikadāsa Śāstrī, Ślokavārttika of Śrī Kumārila Bhaṭṭa, with the
commentary Nyāyaratnākara of Pārthasārathi Miśra. Varanasi: Tara
Publications, 1978.
TS Dvarikadas Shastri: Tattvasaṅgraha of Ācārya
Shāntarakṣita with the Commentary ‘Pañjikā’ of Shri Kamalashīla. 2 Vols.
Varanasi: Bauddha Bharati, 1968; reprinted 1981.
TSP Kamalaśīla’s Tattvasaṅgrahapañjikā. See TS.

Vś Vasubandhu’s Viṃśikā. Vijñaptimātratāsiddhi: Deux Traités de Vasubandhu:

Viṃśatikā et Triṃśikā. Edited by Sylvain Lévi. Paris: Libraire Ancienne
Honoré Champion, 1925.

Research literature
Arnold, D. (2008), “Buddhist idealism, epistemic or otherwise: thoughts on the alternating perspectives of
 Dharmakīrti.” Sophia, 47, 3–28.
Bronkhorst, J., and Ramseier, Y. (1994), Word Index to the Praśastapādabhāṣya: A Complete Word Index to
the Printed Editions of the Praśastapādabhāṣya. Delhi: Motilal Banarsidass.
Coseru, C. (2012), Perceiving Reality: Consciousness, Intentionality, and Cognition in Buddhist Philosophy.
New York: Oxford University Press.
De La Vallée Poussin, L. (1928), Vijñaptimātratāsiddhi. La Siddhi de Hiuan-Tsang, traduite et annotée par
Louis de la Vallée Poussin (Vol. I). Paris: Librairie Orientaliste Paul Geuthner.
Deleanu, F. (2006), The Chapter on the Mundane Path (Laukikamārga) in the Śrāvakabhūmi. A Trilingual
Edition (Sanskrit, Tibetan, Chinese), Annotated Translation, and Introductory Study. 2 vols. Tokyo: The
International Institute for Buddhist Studies.
Dreyfus, G. (1997), Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpretations. Albany:
State University of New York Press.
Dunne, J. (2004), Foundations of Dharmakīrti’s Philosophy. Somerville: Wisdom Publications.
Eltschinger, V. (2010 [2009]), “Ignorance, epistemology and soteriology—Part I.” Journal of the
International Association of Buddhist Studies, 32/1–2, 39–83.
Eltschinger, V. (2011 [2010]), “Ignorance, epistemology and soteriology—Part II.” Journal of the
International Association of Buddhist Studies, 33/1–2, 27–74.
Frauwallner, E. (1930), “Dignāgas Ālambanaparīkṣā: Text, Übersetzung und Erläuterungen.” Wiener
Zeitschrift für die Kunde des Morgenlandes, 37, 174–194.
Funayama, T. (1990), “Bubun to zentai.” Tohō Gakuhō, 62, 607–635.
Hayes, R. P. (1988), Dignāga on the Interpretation of Signs. Dordrecht: Kluwer Academic Publishers.
Iwata, T. (1991), Sahopalambhaniyama: Struktur und Entwicklung des Schlusses von der Tatsache, dass
Erkenntnis und Gegenstand ausschliesslich zusammen wahrgenommen werden, auf deren
Nichtverschiedenheit. Vol. 1. Stuttgart: F. Steiner.
Kapstein, M. (1988), “Mereological considerations in Vasubandhu’s ‘proof of idealism’
(Vijñaptimātratāsiddhiḥ).” Idealistic Studies, 18/1, 32–54.
Kellner, B. (1999), “Levels of (im)perceptibility. Dharmottara on the dṛśya in dṛśyānupalabdhi,” in Shoryu
Katsura (ed.), Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of
the Third International Dharmakīrti Conference, Hiroshima, November 4–6, 1997. Wien: Verlag der
Österreichischen Akademie der Wissenschaften, pp. 193–208.
Kellner, B. (2003), “Integrating negative knowledge into Pramāṇa theory: the development of the
dṛśyānupalabdhi in Dharmakīrti’s earlier works.” Journal of Indian Philosophy, 31/1–3, 121–159.
Kellner, B. (2009), “Buddhistische Theorien des Geistes—Intentionalität und Selbstbewusstsein,” in Birgit
Kellner and Susanne Weigelin-Schwiedrzik (eds.), Denkt Asien anders? Göttingen: V & R unipress, pp.
55–75.
Kellner, B. (2010a), “Towards a critical edition of Dharmakīrti’s Pramāṇavārttika,” in Jürgen Hanneder and
Philipp A. Maas (eds.), Text Genealogy, Textual Criticism and Editorial Technique = Wiener Zeitschrift für
die Kunde Südasiens 52–53 (2009–2010), 161–211.
Kellner, B. (2010b), “Self-awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti: a close
reading.” Journal of Indian Philosophy, 38/3; special issue on “Buddhist Theories of Self-Awareness
(svasaṃvedana): Reception and Critique,” 203–231.
Kellner, B. (2011a), “Dharmakīrti’s criticism of external realism and the sliding scale of analysis,” in
Krasser et al. (2011), pp. 291–298.
Kellner, B. (2011b), “Self-awareness (svasaṃvedana) and infinite regresses: a comparison of arguments by
Dignāga and Dharmakīrti.” Journal of Indian Philosophy, 39/4–5; special issue on 14th World Sanskrit
Conference (eds. Shōryū Katsura, Mark Siderits, and Kiyotaka Yoshimizu), 411–426.
Kellner, B., and Taber, J. (2014), “Studies in Yogācāra-Vijñānavāda idealism I: the interpretation of
Vasubandhu’s Viṃśikā.” Asiatische Studien/Études Asiatiques, 68/3, 709–756.
King, P. (2005), “Rethinking representation in the Middle Ages,” in Henrik Lagerlund (ed.), Representation
and Objects of Thought in Medieval Philosophy. Aldershot: Ashgate, pp. 83–102.
Kobayashi, H. (2011), “On the development of the argument to prove vijñaptimātratā,” in Krasser et al.
(2011), pp. 299–308.
Krasser, H. (2004), “Are Buddhist Pramāṇavādins non-Buddhistic? Dignāga and Dharmakīrti on the impact
of logic and epistemology on emancipation.” Hōrin, 11, 129–146.
Krasser, H. (2012), “Bhāviveka, Dharmakīrti and Kumārila,” in François Voegeli, Vincent Eltschinger,
Danielle Feller, Maria Piera Candotti, Bogdan Diaconescu, and Malhar Kulkarni (eds.), Devadattīyam:
Johannes Bronkhorst Felicitation Volume. Bern, Berlin, Bruxelles, Frankfurt am Main, New York, Oxford,
Wien: Peter Lang Verlag, pp. 535–594.
Krasser, H., et al. (2011), Religion and Logic in Buddhist Philosophical Analysis: Proceedings of the Fourth
International Dharmakīrti Conference, Vienna, August 23–27, 2005. Edited by Helmut Krasser, Horst
Lasic, Eli Franco, and Birgit Kellner. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Kyūma, T. (2011), “On the (im)perceptibility of external objects in Dharmakīrti’s epistemology,” in Krasser
et al., pp. 309–318.
Legrand, D. (2009), “Two senses for ‘givenness of consciousness.’” Phenomenology and Cognitive Sciences,
8, 89–94.
Lin, C.-K. (2009 [2007]), “Object of cognition in Dignāga’s Ālambanaparīkṣāvṛtti: on the controversial
passages in Paramārtha’s and Xuanzang’s translations.” Journal of the International Association of
Buddhist Studies, 30/1–2, 117–138.
Lin, C.-K. (2014), “The uncompromising quest for genuine Buddhism: Lü Cheng’s critique of original
enlightenment,” in John Makeham (ed.), Transforming Consciousness: Yogācāra Thought in Modern
 China. New York: Oxford University Press, pp. 343–374.
Lusthaus, D. (2014), “Lü Cheng, epistemology, and genuine Buddhism,” in John Makeham (ed.),
Transforming Consciousness: Yogācāra Thought in Modern China. New York: Oxford University Press, pp.
317–342.
McClintock, S. (2003), “The role of the ‘given’ in the classification of Śāntarakṣita and Kamalaśīla as
Svātantrika-Mādhyamikas,” in Georges Dreyfus and Sara McClintock (eds.), The Svātantrika-Prāsaṅgika
Distinction: What Difference Does a Difference Make? Somerville: Wisdom Publications, pp. 125–171.
Miyasaka, Yūsho. (ed.) (1971–72), Pramāṇavārttika-Kārikā (Sanskrit and Tibetan). Acta Indologica, 2, 1–
206.
Oetke, O. (1992), “Doctrine and argument in Vijñānavāda-Buddhism.” Wiener Zeitschrift für die Kunde
Südasiens, 36, 217–225.
Ratié, I. (2014), “On the distinction between epistemic and metaphysical Buddhist idealisms: a Śaiva
perspective.” Journal of Indian Philosophy, 42/2–3, 353–375.
Schmithausen, L. (2001), “Vijñaptimātra(tā),” in Joachim Ritter, Karlfried Gründer, and Gottfried Gabriel
(eds.), Historisches Wörterbuch der Philosophie, Vol. 11. Basel: Schwabe, pp. 1057–1061.
Siderits, M. (forthcoming), “Dignāga’s philosophy of mind.” (To appear in a Korean anthology on philosophy
of mind.)
Siderits, M. (forthcoming), “Dignāga’s reflexivity thesis.” (To appear in Amita Festschrift Kolkata.)
Taber, J. (1992), “What did Kumārilabhaṭṭa mean by ‘SvataḥPrāmāṇya’?,” Journal of the American Oriental
Society, 112, 204–221.
Taber, J. (1994), “Kumārila’s refutation of the dreaming argument: the Nirālambanavāda-Adhikaraṇa,” in R.
C. Dwivedi (ed.), Studies in Mīmāṃsā: Dr. Mandan Mishra Felicitation Volume. Delhi: Motilal
Banarsidass, pp. 27–52.
Taber, J. (2005), A Hindu Critique of Buddhist Epistemology: Kumārila on Perception. The “Determination of
Perception” Chapter of Kumārilabhaṭṭa’s Ślokavārttika, Translation and Commentary. London: Routledge
Curzon.
Taber, J. (2010), “Kumārila’s Buddhist.” Journal of Indian Philosophy, 38/3, 279–296.
Tosaki, H. (1979), Bukkyō ninshikiron no kenkyū. [Studies in Buddhist epistemology]. Tokyo:
Daitōshuppansha. Volume I (jōkan).
Tosaki, H. (1985), Bukkyō ninshikiron no kenkyū. [Studies in Buddhist epistemology]. Tokyo:
Daitōshuppansha. Volume II (gekan).
Watson, A. (2014), “Light as an analogy for cognition in Buddhist idealism (Vijñānavāda).” Journal of Indian
Philosophy, 42/2–3, 401–421.
Zahavi, D. (2005), Subjectivity and Selfhood: Investigating the First-Person Perspective. Cambridge: MIT
Press.
 6
Materialism in Indian Philosophy: The Doctrine and Arguments
Pradeep P. Gokhale

Introduction
In this chapter1 I will discuss the materialist doctrine and the arguments presented for it
in the Lokāyata tradition—both its preclassical and classical phase, with greater emphasis
on the latter. The chapter is divided into four sections:

1. “Indian materialism and Lokāyata”: In this section I will explain what I understand by
these terms and interrelation between the two themes.
2. “Preclassical Lokāyata materialism”: Here I will discuss the materialist doctrines and
arguments advanced by some preclassical “Lokāyata” thinkers.
3. “Classical Lokāyata/Cārvāka materialism”: Here there are two issues: (1) What was the
nature of materialism accepted in the Cārvāka-darśana? And (2) given that there is
diversity in the Cārvāka epistemology, which epistemological model is most suitable for
the Cārvāka materialism?
4. “The defensibility of Cārvāka materialism”: Here I will consider some main objections
raised by opponents against the Lokāyata/Cārvāka materialism and the possible
answers to them.

Indian materialism and Lokāyata

Materialism
Primarily the term “materialism” has two senses: factual and normative. Here I am taking
the term in its former sense, in which materialism is a view about the nature of the world
and the nature of a human being. And though it is not normative in nature, it has some
implications for our perspectives on value. It provides answers to questions like the
following:

A. Ontological (cosmological) questions:

1. Is there consciousness over and above matter in the world?
2. If there is consciousness in the world, what is its status?
3. Are there worlds other than this world?
B. Psychological or anthropological questions:
4. Does consciousness exist apart from body? Are there souls (ātman/jīva/puruṣa)
existing as separable from the body?
5. Does the Self transmigrate? Do actions performed in this life have an impact on the
next life?
6. What is the nature of a human being? Is it just body (or brain or behavior or all
these things together), or is there consciousness in it?
7. What is the self (“I”) of a person? Is it identical with body or consciousness or are
both related with each other? If both are related with each other, then what is their
interrelation?
C. Social and moral questions:
8. Which human values will be permissible within the framework of the materialist
ontology? To use the language of puruṣārthas (goals of life), will dharma and mokṣa
be permissible as the human values? Or will only artha (wealth) and kāma
(pleasure) be permissible?
9. Are social, cultural institutions determined by material conditions or by some divine
or supernatural agencies? Are religions and religious scriptures creations of God or
human creations?

How would a materialist answer these questions? It is not always clear. Of course when
we are concerned with negative answers to some of the questions, we can be more or less
clear. For instance, materialists would agree that consciousness does not exist
independently of matter (Q.1); that there are no other worlds (non-empirical worlds to be
reached by the soul after death) (Q.3); that consciousness does not exist apart from the
body and that there are no souls existing as separable from the body (Q.4); that the so-
called selves do not transmigrate and the question of the actions performed in the present
life having impact on one’s own future lives does not arise (Q.5); if dharma is understood
as a ritualistic action having consequences on other worlds or afterlife, and mokṣa is
understood as the cessation of the cycle of births and deaths, then they are not acceptable
as values (Q.8); social institutions are not creations of divine or supernatural agencies;
Religion is not a divine creation but a human creation (Q.9).
But for some questions materialists may not have unanimous answers. All materialists
may not answer Q.2 in the same way. Some would say that matter is the only reality and
consciousness is reducible to matter; whereas some, namely, non-reductionist materialists
would grant the reality of consciousness but regard it as being dependent on matter.
Hence we can distinguish between reductive and non-reductive (sometimes termed as
eliminative and non-eliminative) forms of materialism.
Similarly with regard to Q.6 and Q.7 some would say that “I” or “self” is represented by
the body alone (or the brain alone or behavior alone) and there is no role of consciousness
in it; whereas some would say that the body (or the brain or behavior or all these things
together) represent “I” or the self only when they are qualified by consciousness. Hence
we can distinguish between reductive and non-reductive forms of
anthropological/psychological materialism. Similarly on Q.8, though the negative answer
is clear, a constructive approach is possible whereby one would reconstruct dharma and
mokṣa as values pertaining to this world and this life without accepting the soul or other
worlds. There is no clear answer to the question as to which value perspective follows
from materialism. The term “materialism” in its normative sense usually refers to
sensuous hedonism. But materialism in this sense does not necessarily follow from
materialism in its factual sense.

Lokāyata/Cārvāka-darśana
There are debates on the terms “Lokāyata” and “Cārvāka.” Here I am not entering into
the debate, but I will make some provisional assertions in order to proceed further.
The term “Lokāyata” has been explained by Chattopadhyaya as proto-materialism
spread among common people, transforming later into the classical materialist school of
Indian philosophy. Ramkrishna Bhattacharya delinks early usage of the term “Lokāyata”
from its later use in the sense of materialist school of Indian philosophy.2 The term in its
early usage according to him meant the science of disputation. Unlike Chattopadhyaya, I
do not derive the term “lokāyata” from lokeṣu āyatam (spread (āyatam) among people
(lokeṣu) or prevalent among people),3 but as lokena āyatam (restrained by the world),
which can be interpreted as “limited by the belief that this is the only world” or “limited
by this-worldly approach” or, to underline the critical aspect of it, “limited by the
approach which disregards other worlds.” And unlike Bhattacharya, I do not interpret the
term “vitaṇḍasatthaṁ” or “vitaṇḍavādasattham” (Sanskrit: vitaṇḍāśāstram or
vitaṇḍāvādaśāstram) associated with Lokāyata by the Buddhists as the science of
disputation or logic as such, but an intellectual activity containing a critical or destructive
form of argumentation used against the dogmas such as God, soul, and other worlds.
Lokāyata in this sense could be understood as a rationalist philosophical movement, which
attempted to solve individual and social issues merely on empirical, rational, and practical
grounds without taking recourse to religion. (Here the word “rationalist” is not to be
contrasted with “empiricist,” but with “religious and dogmatic.”) As a movement rather
than a rigid system, it could accommodate various trends from Skepticism to materialism.
Unlike Bhattacharya, I hold that there is continuity between the early concept of Lokāyata
and the later concept, which was eventually identified with the term “Cārvāka.”
The word “Cārvāka” is taken to mean “one who chews (enjoys)” (carvayati iti) or “one
who has sweet tongue” (cāru-vāk) on the basis of etymology (which is quite artificial) or
the Mahābhārata character called Cārvāka who was Duryodhana’s friend and who, in
disguise of a Brāhmaṇa, opposed the sacrificial rite conducted by Yudhiṣṭhira. None of
these meanings conveys the essence of Cārvāka philosophy. The classical Cārvāka
philosophy has been associated with Bṛhaspati’s aphorisms. But the freedom of thought
and reasoning enjoyed by the Lokāyata thinkers was also enjoyed by the philosophers of
Cārvāka darśana in their interpretation of Bṛhaspati’s aphorisms and we also come across
diversity of trends in the classical Cārvāka darśana. Hence Jayarāśi offers a pro-
Skepticism interpretation of Bṛhaspati’s aphorisms and Jayantabhaṭṭa acknowledges
different schools of Cārvākas.4
The earlier nomenclature Lokāyata and the later nomenclature Cārvāka, both refer to a
diversity of rational, argumentative ways of challenging the traditional dogmatic views.
This puts into question the popular identification of the Cārvāka-darśana with
materialism. In recent scholarship, Debiprasad Chattopadhyaya5 and Ramkrishna
Bhattacharya6 interpret Cārvāka-darśana as essentially a materialist school, whereas
scholars like Eli Franco opine that Cārvāka-darśana may not be one uniform school but an
association of at least two subschools—one materialist and the other skeptic.
Though I agree with Franco in essence, I differ in details. The view that there were only
or mainly two schools of Cārvāka-darśana, namely, materialism and Skepticism, needs to
be qualified further. This is because materialism is primarily an ontological doctrine
whereas Skepticism is primarily an epistemological view. So the diversity in the Cārvāka
view will have to be appreciated at both these levels separately and, maybe, in an
interconnected way. If we just concentrate on the diversity of epistemological views held
by Cārvākas, then we can conceive of at least three schools of Cārvākas: cognitive
Skepticism accepting no pramāṇa (source of knowledge), absolute empiricism accepting
only one pramāṇa, namely, perception, and qualified empiricism accepting perception and
also a certain kind of inference. Hence there was greater diversity in Cārvāka
epistemology as compared to their ontology.
There is also a historical dimension to the relation between Lokāyata and Indian
materialism. Scholars have marked two stages in the development of Indian materialism.
The first stage can be called the early stage or preclassical stage when we come across
materialistic arguments which were called bhūtavāda7 (the doctrine that consciousness
arises from four or five elements) or tajjīvataccharīravāda8 (the doctrine that body and
self are the same). This stage is not necessarily and officially linked with Lokāyata. The
second stage, which starts from sixth century ceonward, can be called the classical stage.
This is the period when the Bṛhaspati-sūtras are formed and commented upon and when
materialism is associated with a pramāṇa theory and hence when the proof of a
materialistic thesis on the basis of a pramāṇa or pramāṇas becomes an important issue. In
what follows we will consider the materialistic arguments in these two stages.

Preclassical Indian materialist thinkers

Prajāpati and Virocana
In Indian philosophical literature we come across personalities which presented
materialist thought and argued for it. Some of them are mythological characters and their
historicity can be doubted. For example, Prajāpati in Chāndogya Upaniṣad (viii: 7–9) and
Bṛhaspati in Padmapurāṇa appear to be mythical figures. Though they present materialist
thought, their identity as materialist thinkers is also dubious. Prajāpati’s story, however,
brings before us an interesting model of anthropological materialism (the doctrine that
body and soul are the same). Prajāpati according to the story is the teacher of both devas
(divine beings) and asuras (demons). Indra and Virocana, their respective leaders,
approach him to know the nature of self. The concept of self, Prajāpati initially introduces
to them is that of the person that is seen in the eye or the self that is reflected in water or
a mirror. Virocana, the king of asuras, is satisfied with this notion of self. He goes back to
his followers and teaches this doctrine to them. The doctrine that body is the self, which is
accepted by asuras, is supposed to be an inferior doctrine. Indra, not satisfied by this
doctrine, returns to Prajāpati for the “true” doctrine—the Upaniṣadic doctrine of the
transcendent self. Hence Prajāpati, who introduces a materialist doctrine, is himself not a
materialist. The same model is carried forward in Padmapurāṇa where Prajāpati is
replaced by Bṛhaspati, who teaches anti-Vedic, anti-sacrificial philosophy to asuras in
order to mislead them.9
Virocana according to the Upaniṣadic story literally follows the idea that body is the self
with its strange implication that body continues to be the self even after death. However,
what is important about this early and crude form of materialism is that it regards
perception as the authority for determining the ultimate nature of the self.
Ajita Keśakambalī
The doctrine that body is the self with its strange implication was not the official doctrine
of the later Lokāyatikas. By dehātmavāda, literally meaning the doctrine that body is the
self (“dehaḥeva ātmā iti vādaḥ”) they implied that body qualified by consciousness is the
self, so that the death of a body is the death of the self, identified with it. The negative
implication of dehātmavāda is that there is no self apart from body. Materialism with
special emphasis on the negative aspect was advocated by the materialist thinkers such as
Ajita Keśakambalī and Paesi (or Pāyāsi) in the preclassical period, references to whom are
found in Jaina and Buddhist canons. Vedic-Upaniṣadic thinkers generally believed in a
transcendent ātman. Jaina thinkers too believed in the transmigrating soul (jīva) and
though the Buddhists believed neither in ātman nor jīva, they believed in life after death,
other worlds, and the beings born in those worlds without parents (Opapāduka/Opapātuka
jīvas). Ajita and Paesi/Pāyāsi, both were opposed to all these beliefs. When Vedic thinkers
believed in them, they tried to justify the institute of Vedic sacrifices on the basis of them.
Though Jainas and Buddhists criticized the institute of sacrifice, they continued to accept
life after death and other worlds for justifying the moral code of conduct. According to
them good conduct was justified because it resulted in happiness in the afterlife and bad
conduct was blameworthy because it resulted in sorrow in the afterlife. Hence the
materialist thinkers, when they rejected life after death, other worlds, and the beings born
in them, also rejected the institute of sacrifices and “morality” which was explained and
justified by referring to life after death. Hence these materialists not only emerged as
antireligious people but also as amoralists. In fact it was conceivable that though the
materialists questioned religious morality with its otherworldly implications, they might
have advocated a secular morality—the analysis of good and bad conduct in terms of their
mundane consequences on the agent and others. References to secular form of good life
advocated by materialists are found in the later literature of Cārvāka/Lokāyata thinkers.10
No such references are attributed to Ajita and Pāyāsi/Paesi; they are depicted as
amoralists and even as immoralists.
The view of Ajita Keśakambalī as described in the Sāmaññaphalasutta of Dīghanikāya
contains both the elements—negative and positive. Negative elements include the denial
of soul, other worlds, life after death, and also otherworldly morality. Positive elements
include the materialist description of birth and death of a person. It says that a human
being is built up of four elements and the four elements which constitute the body go back
to the elements in nature when the body dies. In the instance of Ajita Keśakambalī we find
a statement of a materialist thesis, not an argument in favor of it. In the case of
Pāyāsi/Paesi, we have a series of arguments.

Paesi and Pāyāsi

In Dīghanikya there is a story of a chieftain Pāyāsi who was a nonbeliever and had
confirmed his this-worldly and materialist view through experiments. He had a debate
with the Buddhist monk Kumāra-Kassapa. According to the story, Pāyāsi argues for his
materialist views and Kumāra-Kassapa counters his arguments through analogical and
speculative arguments. Finally Pāyāsi is converted. There is a parallel story in a Jaina
Upāṅga text called Rāyapaseṇiya, where the Prakrit counterpart of the Pali Pāyāsi is
named as Paesi, who has a debate with the Jaina monk Kesī.11
The two dialogues—for brevity let us call them Buddhist dialogue and Jaina dialogue
respectively—contain many arguments in common. It would be an interesting historical
issue as to which story could be original or whether both could have been derived from a
third source. But there are philosophically important differences between the two which
are more interesting. Because the doctrine of jīva (the soul which transmigrates and could
ultimately be liberated) is central to Jainism, the point of debate in the Jaina dialogue is
whether jīva and body are one or different. The whole debate develops around this issue.
And because Buddhism accepts rebirth, other worlds, but no soul, the point of debate in
the Buddhist dialogue is whether there is another world, beings reborn without parents,
and whether there are fruits of good and bad actions. According to Jainism, that all these
things exist is a corollary of the doctrine of soul. According to Buddhism, they are
accepted without accepting the soul. This causes incoherence in the Buddhist dialogue.
The chieftain Pāyāsi provides several instances where a person is killed under
experimental conditions to check whether one can say that a soul-like substance leaves
the body when the body dies, and reports that no such evidence was found. The reader
would expect that the Buddhist monk Kumāra-Kassapa should have clarified his position
on this that the Buddhists do not believe in the existence of soul and that according to
them transmigration and rebirth are possible without there being soul. But Kumāra-
Kassapa does not make any such remark. On the contrary sometimes he uses the
language of a Jaina when he suggests that when a person dreams to be in a garden,
actually his soul goes to the garden (Chattopadhyaya, 1990, p. 19). On another occasion
Kassapa gives an explanation of bodily and mental functions of a person in terms of three
elements—life, heat, and intelligence. A body according to him cannot perform various
functions when it lacks these factors (p. 22). But this can be said even by a materialist. So
sometimes a debate between Pāyāsi and Kassapa turns into an argument between
reductive and non-reductive forms of materialism. The Buddhist dialogue therefore is not
as coherent as the Jaina dialogue and the Buddhist response to the materialist does not
seem to be as consistent as the Jaina response.12
The materialist arguments made by both Pāyāsi and Paesi in the respective dialogues
are similar or of the same type. The arguments are many in number, but they can be
broadly classified into three groups:

1. Arguments against the existence of other worlds.

2. Arguments against the existence of soul, which leaves the body at the time of death.
3. Arguments from reputation and tradition.13

The first group of arguments refers to the expectation that if Pāyāsi’s/Paesi’s close
relative who had performed good or bad deeds died and went to svarga (heaven) or
naraka (hell), respectively, he/she should have come and told him so. Since no such
reports were available to Pāyāsi/Paesi, he had a sufficient ground not to believe in the
existence of svarga or naraka.
This group of arguments anticipated Bṛhaspati’s aphorism paralokino’bhāvāt
paralokābhāvaḥ. In fact it helps us in understanding the meaning of the aphorism. The
literal meaning of the aphorism is “There is no other world because there is no one who
belongs to another world.” Here “one who belongs to another world” really means “one
who comes from another world and reports about it.” The point is that the materialist is
ready to accept the existence of something on the basis of its experience or an authentic
report about it based on the reporter’s experience. Since none of them is available in the
case of the other worlds, the latter’s existence cannot be accepted.
The second group of arguments refers to various experiments made by Pāyāsi/Paesi in
order to find out whether there must be the soul substance which leaves the body at the
time of death. Pāyāsi/Paesi shows on the basis of them that no such soul-substance leaving
the body was seen although all other conditions were kept under control. To cite one such
experiment as Pāyāsi reports it:

Take the case of men, who having taken a felon red-handed, bring him up, saying “This felon, my
lord, was caught in the act. Inflict on him what penalty you wish.” And I should say: “Well then, my
masters, throw this man alive into a jar, close the mouth of it and cover it over with wet leather, put
over that a thick cement of moist clay, put it on to a furnace and kindle a fire.” They saying “Very
Good” would obey me and ... kindle the fire. When we knew that the man was dead, we should take
down the jar, unbind and open the mouth and quickly observe it, with the idea: “Perhaps we may
see the soul of him coming out!” We don’t see the soul of him coming out! This, master Kassapa is
for me evidence that there neither is another world, nor rebirth other than by parentage, nor fruit,
or result of deeds well or ill-done. (Chattopadhyaya, 1990, pp. 18–19)

Through an experiment he also checked whether the changes occurring in body due to
death imply that the soul-substance must have left the body. For example, if a soul-
substance leaves the body at the time of death (presuming that the soul has positive
weight), the dead body should be lighter than the living body. But in fact he found that it
was heavier.
As against Pāyāsi’s/Paesi’s arguments, their Jaina/Buddhist opponents argued on the
following lines:

1. There are some practical and technical difficulties because of which the departed
beings in other worlds do not and cannot come back to see the beings in this world.
2. According to Jainas the soul being too subtle cannot be perceived by us, not because it
does not exist.
3. The phenomena like other worlds, otherworldly beings, and souls (in the case of
 Jainism) cannot be seen by the fleshly eye, but they can be seen by the divine eye
developed through spiritual practice.
4. These transcendent things cannot be tested empirically but can be explained through
similes.

From the above four types of arguments, the arguments 1, 2, and 4 try to give an
explanation of the transcendent phenomena, how they are conceivable, how they can
perhaps be real. Argument 3, the argument from the divine eye, attempts to “prove” the
existence of these things. The pramāṇa by which it is proved is the extraordinary
perception. The authority of such a perception, however, is questionable.
The materialist arguments however express a crude scientific outlook. As
Chattopadhyaya (1964, p. 198) comments on Pāyāsi’s demonstrations of soul-body
identity: “A modern materialist would not of course take resort to such crude
demonstrations in support of his thesis. He has an immeasurably vast stock of scientific
data to substantiate his materialist outlook.”
In spite of inadequacies on both sides, the debate brings the basic methodological
differences between materialism and religious philosophies to the surface. Materialists
are using an empiricist methodology whereas religious philosophies are appealing to
speculation, metaphors, and divine authority.
We find that in the classical Lokāyata/Cārvāka darśana there took place systematization
and development of empiricist epistemology and materialist ontology. Let us consider
them in the next section.

Materialism in the classical Lokāyata/Cārvāka-darśana

When we consider the classical period, two questions about materialism become
important:

1. What is the nature of ontological and psychological materialism accepted in this

period? Are these forms of materialism reductive or non-reductive? By reductive
materialism we mean the doctrine according to which consciousness is reducible to
matter such that in the ultimate analysis only matter exists, consciousness does not.
When applied to the psychological realm it would mean that consciousness is reducible
to body (or to brain) and in the ultimate analysis only bodily processes exist (or brain
processes exist), mental processes do not. By non-reductive materialism we mean the
doctrine that consciousness depends on matter, it cannot exist independently of matter
but it is not reducible to matter. When applied to the psychological realm it would
mean that mental processes depend upon bodily processes or upon brain processes (or
that the former are caused by the latter) but they are not reducible to the latter.
2. What is the most suitable position on pramāṇas on the basis of which the particular
form of materialism can be justified?

The nature of Cārvāka materialism

At the level of both ontological and psychological materialism, we have to distinguish
between its negative and positive aspect. The negative aspect of ontological materialism
simply says that there is no consciousness independently of matter. The positive aspect
says that consciousness originates in matter. Similarly the negative aspect of the
psychological materialism says that there is no disembodied self. The positive aspect
earlier assumed the form that body is the self. Later it was qualified that the body
qualified by consciousness is the self. Hence both the positive aspects of ontological and
psychological materialism (revised) are non-reductive forms of materialism.
Cārvākas were also concerned with the axiological implications of this materialism.
Hence they denied dharma as the human goal understood in ritualistic and otherworldly
sense, and mokṣa understood as a transcendent state beyond the cycle of births and
deaths.

The epistemological ground of Cārvāka materialism

Given that Cārvākas accept a non-reductive form of materialism at two levels—ontological
and psychological—questions like the following become important: which are the
pramāṇas suitable for the knowledge of the materialist thesis of the Cārvākas? Which are
the arguments suitable for its justification?
 I have pointed out earlier that there is greater diversity in Cārvāka epistemology as
compared to their ontology. All Cārvākas, except the skeptics, accept materialism, but try
to support it by different forms of empiricist epistemology. It can be doubted, however,
whether every variety of empiricism attributed to different schools of Cārvākas is equally
capable of supporting their materialism. It is obvious that the popular Cārvāka
epistemology of “perception alone as the means to knowledge” is not suitable for the
Cārvāka style materialism and that a certain type of inference will have to be accepted as
a means to its knowledge and justification. Now the question arises as to what kind of
inference is befitting for the justification of materialism. Here we will refer to the notions
like verifiable inference (utpanna-pratīti anumāna), non-verifiable inference (utpādya-
pratīti anumana), and commonsensically justified inference (lokaprasiddha anumāna) and
draw their implications. I will attempt to show that it is commonsensically justified
inference and the mitigated empiricism based on it, which are most fitting for the
justification of Cārvāka materialism.
Here one has to distinguish between the negative and the positive aspect of
materialism. The negative aspect of materialism simply holds that consciousness cannot
exist outside body/independently of matter. For justifying this, the criterion of
perceptibility may be sufficient. Any epistemological school of Cārvākas can defend it. The
positive aspect, however states that consciousness arises from four elements. The
doctrine in this form asserts a causal relation and hence cannot be defended by naïve
empiricism. This point is relevant mainly to the ontological materialism of the Cārvākas.

Cosmological materialism (Bhūtacaitanyavāda)

The cosmological materialism of Cārvākas assumes the title Bhūtacaitanyavāda, which
means that consciousness arises from the combination of four gross elements. Three
aphorisms of Bṛhaspati seem to be relevant here:

1. Earth, Water, Fire, and Air are the tattvas.14

2. Their combination gets designated as “body,” “sense-organ,” and “object.”15
3. Consciousness arises from them (the elements), which have assumed the form of body,
like the power of intoxication arises16 from molasses and other things (or, like the red
color that arises when betel leaf is chewed with other ingredients of paan17).

The third aphorism mentioned above is the core of the Cārvāka argument for
cosmological materialism. But before considering its relevance for Cārvāka epistemology,
it is necessary to make some points of clarification regarding the first two aphorisms, as
they prepare the background for the third aphorism.

1. The first aphorism refers to the four elements as tattvas. It is interesting to note that it
does not refer to ākāśa (ether/space) as the element. Curiously enough, it is a common
tendency among the heterodox systems of Indian philosophy not to count ākāśa among
physical elements. It is also theoretically more elegant to do so, because accepting
ākāśa as an all-pervasive positive substance is problematic.18
2. In the second aphorism the Cārvāka author is talking about combinations of the four
elements. Cārvākas do not seem to be talking about “how” or “why” such combinations
take place. For other systems this question is very much relevant, because they accept
some kind of divine order or karmic order or a teleological explanation behind
whatever takes place. None of these is acceptable to Cārvākas. They would accept
natural causal factors, but not the alleged factors such as Karma, God’s will, and the
teleology of Puruṣārtha. At that stage they would say that these things are accidental
(yadṛcchā) or rooted in the nature of things (svabhāva).
On this background one can discuss the third aphorism which gives the core of the
materialist doctrine of the Cārvākas. The question is whether it goes well with all the
epistemological schools of Cārvākas. Except the school of the skeptics, that is, Cārvākas
like Jayarāśi, all the epistemological schools of the Cārvākas seem to accept this doctrine
at the ontological level. Even the popular Cārvāka view that anumāna is not pramāṇa, as
found in works like Ṣaḍdarśanasamuccaya and Sarvadarśanasaṅgraha, has been
presented along with cosmological materialism and also psychological materialism. This
appears to be a discrepancy, because though the one who regards only perception as
pramāṇa can claim to know the four mahābhūtas by perception, and also one’s own
consciousness by internal perception, he or she cannot claim to know only on the basis of
perception that consciousness arises from matter. Hence, the extreme empiricism of the
popular Cārvāka school does not support its cosmological materialism. In order to know
and then to justify that consciousness arises from matter, one needs an inference. But
what kind of inference will serve the purpose?
Hence Cārvākas have to distinguish between different types of inferences and then
accept certain types of inferences and reject others. This is the most crucial point in
Cārvāka epistemology. Cārvākas seem to have adopted two different policies about the
acceptability of inferences. One can be called verifiabilism and the other, commonsense
empiricism.19

Verifiabilism (acceptance of utpanna-pratīti anumāna)

This version is given by Jayantabhaṭṭa in his Nyāyamañjari and is attributed by him to the
Cārvākas whom he calls Suśikṣitatara (“more learned”). According to it the inference of
an object which has already been experienced before (utpannapratīti) can be accepted as
pramāṇa; the inference of an object which is yet to be experienced utpādyapratīti cannot.
For instance, God, whose existence the theist wants to infer has not been experienced by
any one. Hence the inference for it should not be accepted as pramāṇa. Acceptance of
utpannapratīti anumana and denial of utpādyapratīti-anumāna helps Cārvākas in
explaining this world and denying otherworldly metaphysics. But now the question is
whether it helps them in establishing their materialist thesis. The materialist thesis of
Cārvākas is brought out clearly by the aphorism which states that consciousness arises
from the four gross elements, like the intoxicating power which arises from molasses.
What the Cārvākas are trying to say here is that although consciousness is heterogeneous
with the four material elements, it could be caused by them. They are saying that
heterogeneous causation is possible, that is, the causal factors combined together can
give rise to a heterogeneous entity.
Here the Cārvākas are making a case for heterogeneous causation by giving the
instance of molasses and intoxicating power. The power of intoxication which was initially
not there in the molasses at all, arises in them through the process of fermentation.
Similarly, the Cārvākas claim, consciousness, which was not there in the four elements at
all, could arise in a particular combination of them.
The above consideration makes us think that the inference for materialism that the
Cārvākas are giving here is utpādyapratīti (non-verifiable) type of inference. The
materialist thesis of the Cārvākas that consciousness arises from matter cannot be
verified either in one’s own case or in the case of others. And the instance of intoxicating
power and molasses is just a metaphor and not an instance proper, because the power of
intoxication is not the same as consciousness. Now the question is: would such non-
verifiable type of inference be acceptable in the Cārvāka epistemological framework, and
if so, under what conditions would it be acceptable?

Commonsense empiricism (acceptance of lokaprasiddha anumāna)

In the Tattvasaṅgrahapañjikā, Kamalaśīla quotes Purandara’s statement that Cārvākas do

admit lokaprasiddha-anumāna, but the kind of anumāna which is made by transcending
the worldly way (laukika mārga) is denied by them.20 Now what does Purandara mean by
the term “lokaprasiddha anumāna”? While asserting lokaprasiddha anumāna Purandara is
also rejecting certain kinds of anumāna in contrast to it. Naturally lokaprasiddha
anumāna is to be contrasted with “inference which transgresses the worldly way.”
Accordingly, I am suggesting, lokaprasiddha can mean “strongly established within the
framework of this-worldly way of life.” Here I am interpreting the term “loka” as related
to this-worldly or secular way of life and the term “prasiddha” as “strongly established”
(pra-siddha). In other words lokaprasiddha anumāna is an inference which proves
something within the framework of this-worldly way of life. I think that this is the most
plausible meaning of the word.
Lokaprasiddha anumāna as interpreted above is wide enough to include verifiable
inferences accepted by the “more learned” Cārvākas. But in addition to verifiable
inferences, it also includes a class of non-verifiable inferences under the condition that
they are lokaprasiddha; that is, that they are strongly supported within the framework of
this-worldly way of life. They are empirically or pragmatically necessary for explaining the
facts of this world. Now let us try to understand the materialist argument of the Cārvākas
in the light of the above notion of lokaprasiddha anumāna.
The third aphorism mentioned above gives one version of the materialist argument of
the Cārvākas: “Consciousness arises from the four gross elements, like the intoxicating
power which arises from molasses.”
Another version can be found in Sarvasiddhāntasaṅgraha: “The consciousness, which is
seen in the modified insentient elements, arises in them, like the red color arises due to
the combination of betel-leaf, betel-nut and (spice-) powder.”21
Both the formulations have the form of an argument, though of an incomplete one. For
example, the first three members of the well-known Nyāya syllogism appear to give the
model of a complete argument:

1. (Thesis:) The mountain has fire

2. (Reason:) Because it has smoke
3. (Instance:) Like the kitchen (or: Whatever has smoke has fire, like the kitchen)

We can compare the Cārvāka arguments with it:

1. The consciousness arises from the four material elements
2. Because . . .
3. Like intoxicating power which arises in molasses
Or: Like red color which arises in the chewed combination of betel leaf etc.

It can be clearly seen that the Cārvāka argument is incomplete because it contains the
statement of the thesis (pratijñā) and instance (dṛṣṭānta) but no reason (hetu). So the
important question persists: what is the reason intended by Cārvākas?
In order to trace the reason, one needs to elaborate the contention of the Cārvākas.
One thing is sure that here the Cārvākas intend to express the idea of inseparability
between body and consciousness. Here one is reminded of two pairs of terms of the so-
called relation of inherence (samavāya) accepted by the Nyāya-Vaiśeṣikas.22 They are
substance-quality (dravya-guṇa) and substance-motion (dravya-kriyā). Whether one
accepts the independent ontological category (padārtha) called samavāya or not, one can
appreciate the point made by them that the quality of a thing or the motion of a thing
cannot exist separately from the thing. Here the Cārvākas are trying to understand
consciousness on the model of a quality (guṇa) of body or function (kriyā) or disposition
(śakti) of a body.23 If a quality or a function or a power arises in a substance, then the
substance is called the material cause (samavāyi-kāraṇa/upādāna-kāraṇa) of that quality
or function or power. In this sense body can be called the material cause of the
consciousness which arises in it. So at this stage, the Cārvākas are simply arguing that
since consciousness is inseparable from corporeal body, it must have the corporeal body
as its material cause. So the reason which could be added to the above incomplete
argument would be that “because consciousness is not found separately from the
corporeal body.” But this may not be sufficient because the opponents of the Cārvākas
have an objection precisely to this. They are thinking that in the case of material cause,
there is homogeneity between cause and effect; the effect cannot be radically different
from its material cause. For example, if the threads are yellow, the cloth will also be
yellow, if the earthen halves are red, the earthen pot will also be red. Here the opponents
are distinguishing between efficient cause (nimitta-kāraṇa) and material cause (upādāna-
kāraṇa). An efficient cause is externally related to its effect and can be heterogeneous to
it. For example, a weaver is an efficient cause of the cloth. He or she is externally related
to it and can be heterogeneous with it. (For example, a weaver is sentient, whereas the
cloth produced is insentient.) But threads are the material cause of the cloth. They are
internally related to it and have to be homogeneous with it. So the opponents think that if
consciousness has a material cause, it must be homogeneous with it. They think that
corporeal body being heterogeneous with consciousness cannot be its material cause
because consciousness is not material, according to them. So they imagine a non-
corporeal substance, which they call ātman, puruṣa, jīva, and so on as the cause or abode
of consciousness. Buddhists do not believe in the substance called ātman, but they accept
a continuum (santāna) of consciousness and regard “immediately preceding
consciousness” as the “immediate cause” (samanantara-pratyaya) of the occurrence of
consciousness at any given moment.24 Hence it becomes the task for Cārvākas to present
their argument in such a strong way that it will answer this objection. Cārvākas are
fulfilling this task by using the above metaphors. The metaphors are modeled to show how
a material cause can produce an effect, which is entirely new, different, and hence
heterogeneous with it. Molasses, that is, the raw material used for producing liquor, does
not have the property of being intoxicating, but this quality arises through the process of
fermentation. Similarly the betel leaf and other ingredients of paan, do not initially have
red color, but the red color arises anew through the process of mixing and chewing them.
Similarly, consciousness, which was not there before in the material elements, can arise in
them if the elements combine in a peculiar way so as to form an organic body. Here the
notions of “processing,” “mixing and chewing,” and “peculiar combination” are
significant, though Cārvākas have only suggested, but not used them explicitly. They may
remind us of the distinction we are taught in school Chemistry between chemical
compounds and mixtures. In a mixture the ingredients which are mixed do not lose their
own properties but only add to each other’s properties. For example, lime juice mixed
with sugar and salt has a taste that is sweet, sour, and salty. It does not have an entirely
different taste, heterogeneous with the original taste, because it is just a mixture and not
a chemical compound. But water (H2O) is a chemical compound of hydrogen and oxygen,
and has an entirely different property (useful for extinguishing fire) than the properties of
its components—hydrogen, which is inflammable, and oxygen, which helps burning.
Similarly, Cārvākas are suggesting, corporeal body is not just four elements put together,
but their biochemical compound which assumes an entirely different property, namely,
sentience which was not there in the material elements.
Hence, Cārvākas through their argument are trying to convey two things:

1. Consciousness must be arising from body and not from a cause external to body,
because it is inseparable from body.
2. Consciousness can arise from body, although it is heterogeneous with the latter
because an entirely new property can arise in a chemical compound like it happens in
the case of intoxicating power of liquor and red color in paan.

Hence if one completes the Cārvāka argument for cosmological materialism, it will look
like the following:

1. Any property found in a substance must have arisen from that substance itself.
2. For example, the power of intoxication (which was not there initially) is found in the
molasses when they are processed, arises from the molasses themselves. Or (2a) For
example, the red color (which was not there initially), found in the betel leaf and other
ingredients of paan, when they are processed (by way of chewing), arises from the
paan-ingredients themselves.
3. Similarly consciousness (which was not there initially) is found in the four elements
when they are configured as a corporeal body.
4. (Conclusion): Therefore consciousness must be arising from the four elements
themselves, configured as a corporeal body.
A question can be raised about the meaning of the word “arises” in view of the two
theories of causation well known in Indian philosophy: satkāryavāda (“The effect inherent
in its cause”) and asatkāryavāda (“The effect non-inherent in its cause”). This gives rise to
two versions of the Cārvāka materialism. According to the satkāryavāda version,
consciousness is already inherent in the four material elements, it only becomes manifest
when the elements are combined in a particular way. According to the asatkāryavāda
version, consciousness is not already present in the matter; it emerges as something new.
In some texts, both the options are kept open.25 However, heterogeneity between
consciousness and matter makes better sense if the former is regarded as non-inherent in
the latter.
The inference for Cārvāka materialism as formulated above falls under the category of
non-verifiable inference (utpādyapratīti anumāna). It is based on heterogeneous material
causation in general, and not about the law of origination of consciousness from material
body which is a special case of heterogeneous material causation. Hence it is a kind of
analogical inference, because intoxicating power, which is generated in molasses, is just a
metaphor or analogy for consciousness and not an instance of it. Now the question is:
would such a non-verifiable, analogical inference be acceptable in the Cārvāka
epistemological framework, and if so, under what conditions would it be acceptable?
Here I want to suggest that although the above argument is a case of non-verifiable
inference, it can still be acceptable to Cārvākas who are mitigated empiricists, on the
ground that it is lokaprasiddha, that is, acceptable within the framework of a worldly way
of life.
The argument for the material origin of consciousness will be a typical case of
lokaprasiddha anumāna. The plausibility of such an argument is not based on “logical
necessity” or “empirical necessity” but “explanatory necessity.” One is claiming here that
just as the power of intoxication in processed molasses cannot be explained without
accepting their origin in the molasses themselves (it cannot be explained satisfactorily as
a transcendent entity having an isolated existence or arising mysteriously as a miracle).
Similarly consciousness, which is found in body (and not found outside it), cannot be
explained satisfactorily without accepting its origin in the material body itself.
Inference for the transmigrating soul as the substratum of consciousness is not
acceptable because it lacks explanatory necessity. But inference for matter as the cause of
consciousness or the living body itself as the locus of consciousness is acceptable because
it has explanatory necessity within the framework of worldly life.

Psychological materialism (Dehātmavāda)

Cosmological materialism is about the relation between consciousness and matter in
general. The issue in dehātmavāda is different. Once consciousness comes into existence
in a combination of matter, a sentient being comes into existence. Now the question can
be asked about the psychical identity of that sentient being. The question about ātman is
the question about self-identification of sentient beings in general and human persons in
particular. Dehātmavāda is a materialist answer to the question of such a psychical self-
identification. In this sense dehātmavāda can be described as psychological materialism.
It should be noted here that the word ātman in the term “dehātmavāda” is used as a
pronoun rather than as a noun. It means “I” or “oneself” and not the metaphysical
substance called ātman.
Two formulations of dehātmavāda are available in ancient discussions on Cārvāka-
darśana: reductive and non-reductive. According to the reductive formulation the body
itself is the self.26 The reason behind this is that only body can be perceived, the Self other
than body cannot.27 Against this the possible objection is that if the body is to be identified
with self, then even a dead body will have to be counted as self.28 But this is impossible
insofar as we mean by ātman the identity of a person-qua-person, or something
representing I-notion of a person. Now the Cārvākas come out with a revised version
which excludes dead bodies from the realm of ātman. According to this formulation the
ātman is body qualified by sentience (caitanyaviśiṣṭa-kāya29).
Dehātmavāda, whether reductive or non-reductive, is criticized and condemned by all
the other systems of Indian philosophy because it goes against their deep-rooted belief in
Karma and rebirth. Personally I feel that the non-reductive version of dehātmavāda is a
strong thesis concerning self-identification, and it can answer the objections of non-
Cārvākas.30 Such a version, however, does not come in the purview of verifiable inference,
but can be supported by commonsense empiricism.
The opponents of the Cārvākas try to make two types of claims against dehātmavāda.
They try to show that the body cannot be qualified by consciousness, but consciousness
must belong to some other locus than body. Second, they try to show that the proper
referent of the term “I” cannot be body but some conscious being other than body. How
Cārvākas can answer these claims of their opponents will be discussed in the next section.

The defensibility of Cārvāka materialism

Let us consider some main objections coming from the non-Cārvāka systems against
materialism and the possible Cārvāka answers to them.

The argument from personal identity

We recognize ourselves in old age as the same (continuous) persons as we were in
childhood. Others also can recognize us in the same way. This personal identity becomes
possible due to the faculties of memory (smṛti) and recognition (pratyabhijñā) which mark
continuity in our life. Now there is nothing common between the body of childhood and
that of old age. If body changes completely but consciousness is continuous, then it must
belong to a locus different from body.31
The Cārvāka response: This argument was difficult to answer earlier because the role
of the brain in memory was not known. Though the Cārvākas would accept that the body
undergoes change, they do not say that everything is momentary. Although all the parts of
the body, which consist of material elements, are subject to change and disintegration,
some parts of the body change more rapidly than others. Memory and recognition are
possible because the relevant parts of the brain which store impressions of past
experiences are more durable than some of the other parts of the body. Hence it is not
necessary to accept the soul in order to explain memory and recognition. Jocularly it may
be said that ancient dehātmavāda could be modified by the modern Cārvāka into brain-
ātmavāda.
The Learned Cārvākas (“suśikṣita-cārvākas”) referred to by Jayanta do not identify
ātman with body but they accept the “knower-principle” (pramātṛ-tattva) which exists
along with the living body so long as the latter exists. (Today we might say that this
knower-principle is a disposition rooted in the brain.) Because of this principle the
memory of the past events in this life becomes possible. But there is no question of
recollecting events of the past life. As the learned Cārvāka of Jayanta says:

The knower-principle does not leave the previous body and enter another body. If it would do so,
then the person would recollect the things experienced in the past life like he recollects the things
experienced in childhood, while being in the present body. If the knower is eternal and distinct
from body, we do not surmise any reason, why he should remember only what he has experienced
in this life and not what he has experienced in another life. Hence the knower does not exist after
the body dies. So leave the repeated story-telling about other worlds based on the doctrine of
eternality of soul and stay happy.32

The argument from karma and rebirth

The main step the opponents take in this regard is to extend the continuity of
consciousness beyond this life. This is done in two ways: (1) according to the doctrine of
karma everyone gets the fruit of each of his action according to the moral/religious worth
of the action. Since the operation of this rule is not fully observed within this life, one has
to accept continuity of consciousness beyond this life; (2) Naiyāyikas believe that a newly
born baby sucks her mother’s nipple because the baby knows that it is a desirable act.
This knowledge must be based on prior experience which the newly born baby must have
had in a previous birth.
The doctrine of karma is also aimed at explanation of the diversity in nature. Different
animals have different physical features and capacities. They are born with those natures
because of their past karma. Some persons are happy; some are unhappy. This too cannot
be explained without accepting past Karma. Opponents generally accept God (īśvara) in
addition to karma. God creates diversity in nature and distributes pleasures and pains to
beings. But he does not do so from his will. In that case God would be biased and partial.
God creates diversity in accordance with the past karma of beings.
The Cārvāka response: (1) The Cārvākas do not accept the doctrine of karma because it
is not supported by experience. They deny the opponents’ claim that the diversity in
nature is due to īśvara or karma. Here Cārvākas adhere to the doctrine of self-nature
(svabhāvavāda) when they say:

As for the pleasure and pain, the opponents should not imagine merit and demerit (as their
causes). People become happy or unhappy by their own nature. There is no other cause.
Who makes peacocks colourful? Who makes cuckoos sing? There is no cause other than self-
nature in this case.33

(2) The Cārvākas would say that the tendency of the child to suck her mother’s nipple is
natural or instinctual. If it is due to the recollection of prior experiences, they argue, why
they do not have explicit speech and recollection revived from previous birth?34

The argument from linguistic usage

Opponents of the Cārvākas draw our attention to certain linguistic usages which indicate
that the first-person singular pronoun “I” refers to ātman and not to the body. They are: “I
perceive the table,” “I am happy,” and so on. Here we don’t say: “This body/visual sense
organ perceives the table” or “This body is happy.” Since the former are genuine and
authentic usages, they clearly show that the one who is happy, the one who perceives, and
so on is not the body but someone other than the body.35
The Cārvāka response: Here the Cārvākas draw our attention to other linguistic usages
which are equally genuine and authentic, such as “I am fat,” “I am young,” “I am old,” and
so on where “I” refers to body and not to ātman. (We don’t say, “My body is fat,” “My body
is young,” etc.) The expression “My body” is meaningful in a secondary sense, like the
expression “The head of Rāhu” (Rāhu being really all head).36
In this controversy both the parties regard certain usages as genuine and primary and
certain others as non-genuine and secondary, when it seems that both are genuine and
primary. Hence both the ways—the reductionist dehātmavāda way of reducing “I” to body
alone and the ātmavāda way of reducing “I” to ātman alone—become problematic. Here
the non-reductive dehātmavāda which regards “I” as “body qualified by consciousness”
fares better because then both physical and mental predicates can be attributed to “I”
without any difficulty. This does not mean, however, that “I” refers sometimes to body
(alone) and sometimes to consciousness (alone). It would be more correct to say that it
refers to the complex of body and its sentience where the two are inseparable. Because
regarding body and mind (consciousness, sentience) as two different substances would
be, to use the terminology of Gilbert Ryle,37 a category mistake.

Concluding remarks
I have tried to show that the Lokāyata/Cārvāka darśana whether in its classical or
preclassical form cannot be identified with materialism, though materialism can be
considered as a dominant trend in the Lokāyata/Cārvāka darśana. Indian materialism has
two aspects: positive and negative. Moreover, it can be understood primarily as having
two forms: cosmological and anthropological (psychological). Cārvāka materialism,
particularly its positive aspect in both these forms, can be better understood as “non-
reductive” rather than “reductive.”
I have argued that neither the popular Cārvāka epistemology of extreme empiricism
nor positivistic empiricism is capable of justifying the non-reductive materialism of the
Cārvākas. For this purpose Cārvākas have to accept a more inclusive epistemology which
Purandara provides by introducing the notion of lokaprasiddha anumāna (which I
understand as commonsensically acceptable inference).
I have suggested that some objections raised against “dehātmavāda” of the Cārvākas
(i.e., its non-reductive version) can be answered by a modern Cārvāka by referring to the
neurological system/brain as sustaining personal identity through memory and
recognition.
Cārvāka-darśana can be looked upon as a living philosophy and not necessarily as a
philosophy of distant past. A question can now be asked: What kind of image of a modern
Cārvāka do we get in the light of the above discussion? A modern Cārvāka may not prefer
to talk in terms of the four gross elements but more and different elements discovered by
physicists. The Cārvāka today would regard consciousness as an outcome of the
biochemical process involving these elements. Similarly instead of causally relating
consciousness with the body the Cārvāka would prefer to relate it with the brain.
But apart from such a modified version of cosmological and psychological materialism,
can we say something definite about the Cārvākas’ position in terms of their value-
perspective? Certainly, the Cārvāka will not buy a religious axiology which attaches value
to God, soul, or other worlds. In his value-perspective there will be no place for dharma
and mokṣa if they are understood in an otherworldly framework. But this does not mean
that he cannot follow these values in any form. For example, Rajendra Prasad, having
explained a reconstructed form of the theory of puruṣārtha in his seminal essay, says in
the concluding section:

As a kāma goal mokṣa can be given a spiritual as well as non-spiritual interpretation, since
theoretically speaking we can find personal peace in various kinds of acquisitions, spiritual and
non-spiritual, religious and non-religious. Therefore even the Cārvāka theory of value can be said
to be an exemplification of the theory proposed here. Dharma will have to be then given a non-
theological, non-otherworldly interpretation and mokṣa will be pleasure, natural pleasure, and not
any kind of heavenly, non-bodily, spiritual bliss.38

I want to suggest that the Cārvāka today could be open to different axiological positions
provided that they are secular in nature. It is intelligible to consider Cārvāka-darśana not
as a complete and closed system, but as an incomplete and open one, provided that the
framework within which it opens itself is secular.
 Notes
1. Some parts of the chapter were given in “Cārvāka Materialism: Some Issues Concerning Its
Nature, Knowledge and Justification,” presented in the National Seminar held at the Asiatic
Society Library, Kolkata, on November 24–25, 2010, and some of them form a part of the author’s
proposed monograph on Lokāyata/Cārvāka philosophy.
2. Of course it should be granted that Bhattacharya (2012, p. 195) also tries to see the link between
the two ideas when he says: “What was common to the older Lokāyatikas and the new Cārvāka
materialists was perhaps disputatiousness: nothing was sacred to them.” Bhattacharya, however,
does not regard this itself—arguing freely and disregarding anything as sacred—as the common
core of the methodological approaches of Lokāyatikas and Cārvākas.
3. The etymology given by Dasgupta following Divyāvadāna, as quoted by Bhattacharya (2012, p.
194).
4. Apart from the popular school of Cārvāka, Jayantabhaṭṭa refers to at least two other schools of
Cārvāka: Cārvāka-dhūrta (cunning Cārvāka) (see NM Part I, pp. 59–600) and Suśikṣitatara (more
learned Cārvāka) (see NM, Part I, p. 113).
5. Through his works such as Chattopadhyaya (1964; 1978).
6. Through Bhattacharya (2012).
7. Bhattacharya (2012, pp. 34–43).
8. Bhattacharya (2012, p. 37).
9. Padmapurāṇa, Sṛṣṭikhaṇḍa, 319–334, 36–38 as included in Pathak (1965, pp. 157–160). In
Viṣṇupurāṇa this role is assigned to Māyāmoha, a deluding character created by the lord Viṣṇu.
See Viṣṇupurāṇa, 3–18, as quoted in Jha (1969, pp. 430–432).
10. For example, kṛṣigorakṣavāṇijyadaṇḍanītyādibhirbudhaḥ/ dṛṣṭaireva sadopāyair-
bhogānanubhaved bhuvi// SSS, 2.9–15 (meaning: A wise person should enjoy in this world only by
following empirical means such as agriculture, cattle-keeping, commerce and governance).
11. In Paesikahāṇayam Kesī is referred to as a young monk (kumāraśramaṇa) belonging to the
Pārśvanātha’s tradition. See Paesikahāṇayam (p. 3). The similarity between the name Kumāra-
Kassapa (Pāyāsisutta) and the description kumāraśramaṇa (Paesikahāṇayam) is striking.
12. This could be stated as an argument in favor of the view that the Jaina dialogue is older and the
Buddhist dialogue is a later, but unscrupulous, adoption of it.
13. The third group can be neglected being an appeal to tradition and reputation.
14. pṛthivyāpastejo vāyuriti tattvāni/ Bhattacharya (2012, p. 78).
15. tatsamudāye śarīrendriyaviṣayasaṁjñāḥ/ Bhattacharya (2012, p. 79).
16. tebhya eva dehākārapariṇatebhyaḥkiṇādibhyo madaśaktivaccaitanyamupajāyate SDS,
Cārvākadarśana.
17. jaḍabhūtavikāreṣu caitanyaṁyattu dṛśyate/ tāmbūlapūgacūrṇānāṁyogādrāga ivotthitam// SSS,
II.7.
18. For example, it is hard to explain that an all-pervasive (and also part-less, indivisible) substance is
a constituent of the mortal physical body, that it constitutes a sense-organ, and so on.
Bhattacharya has pointed out that in Manimekalai there are references to thinkers called
Bhūtavādins who accept five elements including ākāśa and that Guṇaratna refers to some sections
of Cārvākas who consider space as the fifth element. See Bhattacharya (2012, pp. 39–40). It is
possible that some Cārvākas might have done so under the influence of the Nyāya-Vaiśeṣikas or
Sāṅkhyas.
19. For my discussion of the theme, see Gokhale (1993).
20. purandarastu āha, lokapradiddhamanumānaṁcārvākairapi iṣyata eva. Yattu kaiścit
laukikaṁmārgamatikramya anumānamucyate tanniṣidhyate, TSP, on verse no 1481.
21. See footnote 17 above.
22. Nyāya-Vaiśeṣikas acknowledged five such pairs: (1) The product and its parts (avayavin-avayava);
(2) the quality and the qualified (guṇa-guṇin); (3) the motion and the moving (kriyā-kriyāvān); (4)
the common characteristic and the individual (jāti-vyakti); and (5) particularity and the eternal
substance (viśeṣa-nityadravya). See TSa, pp. 96–97.
23. It is not clear whether Cārvākas would treat consciousness as a “quality” of body or
“function/activity” or “disposition” of the corporeal body. The metaphor of the red color arising in
paan suggests that it is like a quality, whereas the metaphor of intoxicating power suggests that it
is a function or a disposition of body.
24. According to the Buddhist theory of causation, samanantara pratyaya, also called
anantarapratyaya, is defined as the immediately preceding cause belonging to the same series as
that of the effect.
25. kecid vṛttikārā vyācakṣate- utpadyate tebhyaścaitanyam, anye abhivyajyata ityāhuḥ, TSP on
verses 1858–59 (meaning: Some, namely vṛttikāras say that consciousness is produced from them;
others say, “it becomes manifest”).
26. sthūlo’haṁtaruṇo vṛddho yuvetyādiviśeṣaṇaiḥ/
viśiṣṭdeha evātmā na tato’nyo vilakṣaṇaḥ// SSS, 2.6
(meaning: Self is nothing but body qualified by the qualifications such as “I am fat,” “I am young,”
“I am old,” “I am a teenager,” etc. It is not something different or excluded from it).
27. dehātirikte ātmani pramāṇābhāvāt. pratyakṣaikapramāṇavāditatā anumānāderanaṅgīkāreṇa
prāmāṇābhāvāt., SDS, (Cārvākadarśana) (meaning: Because there is no authoritative means to the
knowledge of self, additional to the body. Because we accept only perception as the authoritative
means to knowledge, the means such as inference have no authority, as they are not accepted by
us).
28. śarīrasya na caitanyaṁmṛteṣu vyabhicārataḥ, BP (verse 48) (meaning: Consciousness is not the
property of body. Because [in that case it should belong to a dead body also, but] it does not belong
to dead bodies). This argument of Viśvanātha is aimed at showing that consciousness does not
belong to body. But it can be easily applied to the question whether self-identity can be attributed
to the body.
29. caitanyaviśiṣṭaḥkāyaḥpuruṣaḥ, included in many texts. See Bhattacharya (2012, p. 79).
30. The non-reductive version of psychological materialism of Cārvākas that I have taken into account
is the standard version according to which ātman is defined as “body qualified by consciousness.”
But in principle it is possible to formulate it as “consciousness qualified by body (i.e.,
consciousness which is essentially embodied)” or as “the mind-body complex, where both are
inseparable.” Particularly the second version (“essentially embodied consciousness”) seems to be
attributed to “learned Cārvākas” by Jayantabhaṭṭa. See footnote 32 where the essentially
embodied consciousness is termed as pramātā.
31. This is the argument from recognition used for proving the existence of permanent ātman:
ahameva jñātavān ahameva vedmi ityādeḥekakartṛviṣayatayāpratyabhijñānasya bhāvataḥsattvāt
ātmā prasiddhaḥ. TSP on verse no 228 (meaning: Ātman is established from bhāva, that is,
existence, of the recognition such as “I myself have known this (before) and I myself, am knowing
(now)” which has a single subject as its object).
32. na ca pūrvaśarīramapahāya śarīrāntaraṁsaṅkrāmati pramātā/ Yadi hyevaṁbhavet tadiha śarīre
śaiśava-daśānubhūta-padārtha-smaraṇavadatīta-janmānubhūta-padārtha-smaraṇamapi tasya
bhavet/ nahi tasya nityatvāviśeṣe ca śarīrabhedāviśeṣe ca smaraṇa-viśeṣe kāraṇamutpaśyāmaḥ,
yadiha-janmanyevānubhūtaṁsmarati nānya-janmānubhūtam iti/ tasmādūrdhvaṁdehānnāstyeva
pramāteti nityātma-vādamūla-paraloka-kathā-kurukurvīm apāsya yathāsukhamāsyatām/ NM, Part
II, p. 39.
33. na kalpyau sukhaduḥkhābhyāṁdharmādharmau parairiha/
svabhāvena sukhī duḥkhī jano’nyannaiva kāraṇam//4//
śikhinaścitrayetko vā, kokilān kaḥprakūjayet/
svabhāvavyatirekeṇa vidyate nātra kāraṇam//5// SSS 2.4–5
34. nāmābhyāsabalādeva yadi teṣāṁpravartate/
tatkiṁna visphuṭā vācaḥsmṛtirvā vāgmināmiva// TS verse No. 1945
(meaning: If the said conceptual cognition of the new born infants proceeds from the repeated
cognition of names, how is it that they do not have the memory or the clear speech of the eloquent
speakers?), Jha (1986, p. 928).
35. Kumārila, for instance, argues on these lines in SV, “Ātmavāda,” verses 110–138.
36. dehātmavāde ca “sthūlo’ham,” “kṛśo’ham,” “kṛṣṇo’ham” ityādisāmānādhikaraṇyopapattiḥ. mama
śarīramiti vyavahāro rāhoḥśira ityādivadaupacārikaḥ. SDS (Cārvākadarśana).
37. See Ryle (1949).
38. See Prasad (1989, p. 305).

Bibliography
Sanskrit works with abbreviations
BP: Viśvanātha Nyāya Pañcānana, Bhāṣāpariccheda with Siddhāntamuktāvalī. Trans. Swami
Madhavananda (2004). Kolkata: Advaita Ashrama.
CU: Chāndogya Upaniṣad. See Chattopadhyaya (1990).
NM: Jayantabhaṭṭa, Nyāyamañjarī. Varanasi: Chaukhamba Sanskrit Series, 2nd ed., 1971.
SDS: Mādhavācārya, Sarvadarśanasaṅgraha. Ed. V. Abhyankar (1978, rpt). Poona: BORI.
SDSam: Haribhadrasūri, Ṣaḍdarśanasamuccaya. Ed. M. K. Jain (1989). New Delhi: Bharatiya Jnanpith
Publication.
SSS: Śaṅkarācārya, Sarvasiddhāntasaṅgraha. Trans. M. Rangacharya (2006). Delhi: Eastern Book
Linkers.
SV: Kumārila Bhaṭṭa, Ślokavārtikam. Ed. D. Śāstrī (1978). Varanasi: Tara Publications.
TS: Śāntarakṣita, Tattvasaṅgraha. In Tattvasaṅgraha with Commentary of Kamalaśīla. Baroda: Oriental
Institute, 1984.
TSa: Annaṃbhaṭṭa, Tarkasaṃgraha. Ed. and trans. Y. V. Athalye and M. R. Bodas (1963). Pune: BORI.
TSP: Kamalaśīla, Tattva-Saṅgraha-Pañjikā. See TS.
TUS: Shri Jayarāśibhaṭṭa, Tattvopaplavasiṃha. Ed. S. Sukhlalji and P. Rasiklal (1987). Varanasi:
Bauddha Bharati.
VS: Kaṇāda, Vaiśeṣikasūtra. Ed. Muni Jambuvijayaji (1961). Baroda: Gaekwad’s Oriental Series No.
 137.

Secondary and other works

Athavale, S. (1997), Cārvāka, Itihāsa Āṇi Tattvajñāna (Marathi). Wai, Maharashtra: Prājña Pāthaśālā
Maṇḍala.
Bhattacharya, R. (2012), Studies on the Cārvāka/Lokāyata. Delhi: Anthem Press India.
Bollée, W. (2002), The Story of Paesi (Paesi-kahāṇayam), Soul and Body in Ancient India, a Dialogue on
Materialism: Text, Translation, Notes and Glossary. Wiesbaden: Harrassowitz Verlag.
Chattopadhyaya, D. (1964), Indian Philosophy: A Popular Introduction. New Delhi: People’s Publishing
House.
Chattopadhyaya, D. (1978, 4th ed.), Lokāyata: A Study in Ancient Materialism. New Delhi: People’s
Publishing House.
Chattopadhyaya, D., and Gangopadhyay, M. K. (eds.) (1990), Cārvāka/Lokāyata. New Delhi: ICPR
Publications.
Franco, E. (1987), Perception, Knowledge and Disbelief (A Study of Jayarāśi’s Scepticism). Delhi:
Motilal Banarsidass.
Gokhale, P. (1993), “The Cārvāka theory of pramāṇas: a restatement.” Philosophy East and West, 43
(4), 675–682.
Jha, A. (1969), Cārvāka Darśana (Hindi). Lucknow: Hindi Samiti.
Paesikahāṇayam. Trans. J. R. Joshi (1980, Marathi). Pune: Jain Adhyasan Firodiya Publication,
University of Pune.
Pathak, S. (1965), Cārvāka Darśana kī Śāstrīya Samīkṣā (Hindi). Varanasi: Chaukhamba Vidyabhavan.
Prasad, R. (1989), Karma, Causation and Retributive Morality, Conceptual Essays in Ethics and Meta-
ethics. New Delhi: ICPR Publication.
Ryle, G. (1949, rpt 2000), The Concept of Mind. London: Penguin Books.
 7
Sāṁkhya: Dualism without Substances
Ferenc Ruzsa

Dualism may be understood as the view that there are two irreducible substances, soul and matter. They are irreducible
in that neither can produce the other, and as they are substances, they are independent of each other, that is, both could
exist without the other.
Of course the theories of some of the staunchest materialists would meet the above criteria. The ancient Greek
atomists considered the psychê to be made up of atoms, but since these are a distinct class of atoms, and atoms are
uncreated and indestructible, we could say that we have two kinds of substance here: soul atoms and all other atoms.
Presumably we have to add that the spiritual substance must be immaterial in some sense: it may be not in space or at
least not occupying space, not resisting the movement of matter. Unfortunately this restriction would exclude Jainism
(where jīva, soul, is spatially circumscribed and can be stained by karmic matter) and that is an undesirable
consequence. We may try to stipulate instead that a person’s soul must be an absolute unit, thereby excluding Colin
McGinn’s hyperdualism1 as well as those versions of Vedānta where a human soul is just a part of a cosmic entity,
Brahman. Or we may try to emphasize the number two—in atomism there are many kinds of atoms, not only soul and
matter. Unfortunately this would make the atomists pluralists instead of the monists most historians take them to be.
Further, all those theories that postulate, for example, time as a distinct substance would cease to be dualist.
So dualism is far from a clear-cut category and we have to live with it. Old Vaiśeṣika had nine substances (the five
elements, space, time, manas, and soul), and therefore it should best be categorized as pluralism. Later on the
nonmateriality of the soul came to be more important and its liberated state (free from any influence of matter) received
more detailed treatment, so labeling it as dualism should not be objected to.
Classical Sāṁkhya is also a case in point. It is rightly considered the paradigmatic case of ancient Indian dualism yet
all the relevant criteria suggested so far are somewhat problematic. The name of this school is derived from the word
“saṁkhyā,” number, and it is normally supposed that the idea suggested is that here all features of the world are given
in numbered lists. However, in its foundational text, the Sāṁkhya-Kārikā (SK),2 we are never told that there are two
substances. When the fundamental categories of the system are first mentioned, they are “the manifest, the unmanifest
and the knower” (vyaktâvyakta-jña, SK 2). The first two are material principles, the knower is, of course, the soul. In the
next verse a list of the well-known twenty-five entities (tattva) of the system are given in four groups: the unmanifest,
the seven productive and the sixteen unproductive manifest tattvas, and finally the soul. So it appears as the system of
three, or four, or twenty-five principles, not of two!
Still, there can be no doubt that Sāṁkhya is a dualism: puruṣa (the “person,” the standard term here for the soul) and
prakṛti (nature) are regularly contrasted, their relation analyzed, and the aim of the system is the absolute and final
separation of the two. Perhaps it is not a self-conscious dualism: it does not call itself “the system of two,” and it has no
appropriate category for substance (or kind of substance), like res for Descartes.
Sāṁkhya’s conception of the soul (as it is normally interpreted) is again very far from the standard European notion.
It is definitely not the res cogitans, the thinking substance—it does none of the functions of the Cartesian soul. To quote
Larson,3 it is an “eccentric ghost,” a “contentless consciousness.” It does not seem to meet the criteria of substance,
since it has no qualities and cannot act or move—in fact it is absolutely unchangeable. Constantly changing matter and
its more stable formations again cannot be regarded as substances: as I will try to show in this chapter, Sāṁkhya in its
classical version is a strict form of substance reductionism, that is, substances are only constantly changing
combinations of qualities and functions, nothing more.

The Sāṁkhya classic

The Sāṁkhya here presented is Īśvarakṛṣṇa’s version of the system, because it is both the most interesting
philosophically and the most influential formulation. Sāṁkhya itself is immensely old, clearly older than Buddhism, and
had many significantly different forms.4 When, around the beginning of the first millennium CE, many philosophical
schools tried to standardize their teaching, there were several attempts also in Sāṁkhya. Perhaps the extremely short
Tattva-Samāsa (Summary of the principles) is the earliest survivor,5 but it is only a list of key terms without any
recognizable philosophical argument. When the classical Sūtras of the other schools were already written6 and
therefore philosophical style, argumentation, and organization were significantly more developed, there appeared the
SK (c. fourth century CE), a compendium in seventy-two verses written in the āryā meter attributed to Īśvarakṛṣṇa. It
seems to consist of a core text of some fifty verses with several additions and minor reworking.7 I refer to the author of
the core text as “Īśvarakṛṣṇa,” although it is possible that it is in fact the name of the (or a) person who added some
extra verses.
Īśvarakṛṣṇa clearly wanted to write the philosophical classic of the school. Therefore, first and foremost, he construed
a relevant and coherent system where all the tenets are argued for without reference to any kind of traditional
authority. At the same time he continued a long tradition and that is most apparent in his choice of terminology;
however, the at times astonishingly archaic words are given modern definitions and the ancient agricultural/tantric
imagery of the fertile, female nature (Mother Earth) and the passive male spirit ruling over it is consistently understood
as a metaphor only. Third, he tried to produce a text acceptable to all branches of the school and therefore avoided
controversial issues, either not mentioning them at all or at times picking an intentionally ambiguous phrasing.8
Īśvarakṛṣṇa’s success is really amazing. His text became the absolute authority within the school to which additions
were possible (as in the medieval Sāṁkhya-Sūtra) but whose formulations remained standard ever since. As Sāṁkhya
concepts permeate other philosophies (especially Vedānta), Hindu religious thought and literature (notably the Purāṇas)
and also traditional science, his formulations were known and used everywhere.
As a philosopher, he can be compared to the greatest. There are at least three points where his ideas may be
considered fully relevant even in present-day discussions. They are his conception of the guṇas; his version of dualism,
distinguishing the ontological and the epistemic sphere of consciousness; and his substance-reductionism in general and
in the specific cases. His denial of emergence, his causally grounded epistemological theory, and many of his single
arguments are still exciting and may often be felt essentially correct.
Unfortunately he was not well understood in several respects. Already the person(s) extending the SK to its present
form added lots of philosophically uninteresting idiosyncratic material: long lists of possible mental states, or of causes
of non-perception (trivial ones, like “it is too far”), thereby also disturbing the original composition and obscuring the
tight internal logic of the whole. Partly because of this, partly because of his philosophical depth combined with extreme
conciseness, and partly because of the general acceptance of some new ideas (notably the absolute unchangeability of
the soul), even his most faithful commentators often misrepresent his position and fail to grasp some of his arguments.
The voluminous Yukti-Dīpikā, often called “the most important commentary of the SK” is no better; it could be described
much more aptly as “the most important post-SK treatise in the Sāṁkhya tradition (formally a commentary on the SK).”
As has already been noticed by some scholars, there was a definite break in the tradition after the SK, at several places
the commentators plainly just did not know what the author’s intention was.9

The twenty-five entities

Classical Sāṁkhya is widely known as the system of the twenty-five tattvas. Tattva, literally “that-ness” or “being that,”
means reality, principle, element, essence, or substance; here it will be translated as “entity.” No definition of the term
is given, and the concept itself is unimportant in the SK—it occurs but once. Intuitively we could say that a tattva is a
fundamental, fairly stable kind of reality; everything else can be described as a (more or less temporary) combination of
tattvas. It would be very tempting to render tattva with “element”; however, the unanimous practice of the translators
reserve this world for the five bhūtas (earth, water, fire, air, and ether).
Īśvarakṛṣṇa inherited this list and as it was central to the tradition it was out of the question to change it. However,
several signs testify to his less than enthusiastic attitude to it: he never mentions the number twenty-five, his single use
of the term is not technical, and he does not give all the members of the list. Still, it is the fundamental map of the basic
categories of Sāṁkhya, so it is reasonable to start with an overview. (It may be helpful time and again to look at the
summary table at the end of this section.)
Besides the two fundamental, imperceptible entities (soul and unmanifest materiality), there are the twenty-three
empirical entities of the manifest world: thirteen internal (psychic and biological) and ten external (physical) entities.
The latter are the five elements and the five sensibilia; together they make up the inanimate world, but of course they
form also the bodies of living beings. They constitute the “field of activity,” kārya (lit. “what is to be done,” duty, work,
effect).
The elements are called bhūtas or viśeṣas, “beings” and “kinds,” and they are the gross material substances of the
world. It is a question for empirical science what exactly the elements are and how many exist—Īśvarakṛṣṇa seems
unconcerned, he never lists them, although he retains the traditional number five. The category itself is relevant
philosophically, but the members of the group are irrelevant.
The five sensibilia or sense objects are those that the five senses can react to: color, sound, smell, taste, and
tangibility (including hot/cold). They are the tan-mātras or a-viśeṣas, meaning “only that” and “uniform.” Both names
probably express that each of them is an unmixed simple (a color has no sound, etc.), in contrast to the elements that
have more than one sensible property (earth has color, smell, etc.).10 The not exactly natural terminology of tanmātras
(instead of “qualities” or “perceptible qualities”) is justified in Sāṁkhya by its characteristic theory of the three guṇas,
“qualities”—it is so fundamental for the school that it was imperative never to use the word “quality” for anything else
but rajas, tamas, or sattva.
The psychophysiological functions of a living being are together called the instrument, karaṇa (lit. “doing,” action,
means of action). It is conveniently split up into a center and a periphery. The latter, called the external instrument
(bāhya karaṇa), connects the center to the physical world. Inward-moving information comes through the five senses,
outward-moving instructions go through the five active powers.
The senses are not to be confused with the physical organs where they reside: hearing is not the ear—it is the
receptivity to specific entities of the physical world, that is, sound and those things that produce sound. The Sanskrit
term buddhi-indriya, clearly expresses this: it means “power of becoming aware.”
The five active powers (karma-indriya, lit. power of action) represent a surprisingly refined biological theory veiled in
a bafflingly archaic terminology—they are called speech, hand, foot, anus, and lap. No wonder Īśvarakṛṣṇa felt the need
to explain them: “and the function of the five [active powers] is speaking, taking, movement, excretion and joy.”11
Nowadays we would perhaps express the same idea with the concepts of consuming and excretion (including breathing
in and out), movement, reproduction, and communication. A fairly comprehensive list, since perception is dealt with
separately, although growth and healing might perhaps be added.
The innermost core of a living being is the internal instrument, antaḥkaraṇa, consisting of the manas, egoity, and
understanding. Although the concept matches quite well our concept of mind, great care must be exercised in the use of
this term on account of the unfortunate, but consistent practice of Indologists who—mislead by the distant etymological
connection—regularly translate manas with “mind.”12
Manas is, in fact, the lowest function of the mind, directly connected to the senses and the active powers: “Manas is
the coordinator,13 and it is a power reckoned among both kinds [i.e., the senses and the active powers]. It is internal, for
its objects belong to the three times [past, present, and future]; therefore it works in both areas, [internal and
external].”14 As a sense, it takes the isolated pieces of information of each sense and reproduces a coherent internal
picture of the world and passes it on to the conscious awareness; for we normally see an elephant, not a big patch of
grey, and we hear the same elephant, not some unrelated sound. As an active power, it controls the other active powers,
partly automatically, partly in response to a conscious decision “above.” But it is not only the eleventh power, it is also
part of the internal instrument, because the external instrument (i.e., the other powers) can only operate in the present,
while manas can produce images of the past or even imagine things for the future.
Egoity is a very interesting psychological concept15 and it has a central function in the Sāṁkhya meditational practice
leading to the liberation of soul from matter. At first sight it appears superfluous in a system where each individual has
its own soul, but in fact it has a very important role to play. Its name ahaṁ-kāra (“I-maker”) and its synonyms mama-
kāra (“my-maker”) and asmitā (“I am-ness,” i.e., the intuition that “I am [this or such]”),16 together with its succinct
definition abhimāna (“appropriation,” SK 24), are all that we are told, but they are expressive enough and the idea is
widely known in all Hindu schools and in Buddhism as well. Perhaps the best analysis is Śaṅkara’s introduction to his
commentary on the Brahma-Sūtra, where it is called adhyāsa, “superimposition” or rather “projection.” Nature is in
itself single and continuous, but we all split it up into I and not-I (ahaṁ-kāra), mine and not mine (mama-kāra). We feel
certain facts of the world as belonging to our being: I am old, I am fat, I am rich, I am a father, I am a scholar (asmitā).
But in fact the real subject, the soul, is neither fat nor a scholar, it has neither money nor offspring. So egoity is the
“illegal” extension of the ontological subject, the soul, producing thereby our familiar phenomenal selves. Secondarily
we further extend our authority over inanimate objects and even other persons, the similar use of “my” obscuring the
process: my hand, my suggestion, my car, my wife . . .
Understanding, buddhi, comes closest to the Cartesian res cogitans, except for it being a material entity. Its definition
is simply “Understanding is judgment,” but it is added that “virtue, knowledge, dispassion and sovereignty are its good
(sāttvika, sattvic) forms, the dark (tāmasa, tamasic) forms are the opposites of these.”17 So judgment is here not only a
value-neutral cognitive act, but includes also moral judgments and decision-making. This is also the seat of the relevant
motives (passion and power); therefore it appears that buddhi corresponds to our notions of rationality, volition, and
emotions. And it is here, and only here, that soul and matter meet—therefore this is the point where they can part ways
at liberation. “Since understanding prepares all experiences for the soul, again only it can distinguish the subtle
difference between materiality and soul.”18
The last two principles—immaterial consciousness or soul and undetectable unmanifest materiality—are never met
with in experience, only inference can prove their existence. They will be dealt with later in separate sections.
It is worthy of notice that the whole system seems to be basically an anthropology: it represents a person (a human or
any other living being). Although there is room for inanimate objects, provided by the elements and sensibilia, these
physical entities also constitute biological bodies.
Further, this anthropology focuses on information processing, it shows a refined, bidirectional, multilayered system.
To illustrate the point with the inward path: it starts with physical objects (the elements); the sensibilia carry the
information from them, then pass it on to the corresponding senses; manas coordinates the information, arranging it
into a coherent picture; egoity relates it to the subject (“I see,” “he wears my shirt,” “I am afraid of dogs”19);
understanding grasps the situation conceptually, and presents it to the soul, which experiences it.20 It is clear that such
a detailed analysis of cognition is fuller and therefore perhaps superior to and potentially more promising than the more
abstract, widespread two-step model of immediate and conceptualized perception (nirvikalpaka–savikalpaka pratyakṣa).
Here is a table of the twenty-five entities (in the bottom row), normally listed from left to right, together with their
groups and combinations. Mere groups are signified by the plural (like “11 powers”), whereas combinations, that is,
functional higher units are in the singular—like “3 internal instrument,” read “the internal instrument that can further
be analysed into three entities.” In brackets appear those names and numbers that are not actually used in the SK.

[25] entities (tattva)

conscious
[24] nature (prakṛti)
(cetana)
[2]​ inferable only
[23] manifest (vyakta)
(liṅgin)
10 field of activity
13 instrument (karaṇa)
(kārya)
3 internal instrument 10 external instrument (bāhya
(antaḥkaraṇa) [karaṇa])
11 powers (indriya)
5 active 5 5
5 senses
powers sensibilia elements
(buddhîndriya)
(karmêndriya) (tanmātra) (bhūta)
sight (cakṣus)
color
hearing
speech (vāc) (rūpa) [earth]
(śrotra)
unmanifest egoity hand (pāṇi) sound [water]
soul understanding smelling
materiality (ahaṁ- manas foot (pāda) (śabda) [fire]
(puruṣa) (buddhi) (ghrāṇa)
(avyakta) kāra) anus (pāyu) [smell] [air]
tasting
lap (upastha) [taste] [ether]
(rasana)
[touch]
skin (tvac)

Cosmogony or psychology?
A particularly nasty problem in reading Sāṁkhya texts is the unclear distinction between cosmology and cosmogony on
the one hand, and psychology and individual evolution on the other. This is but one appearance of the microcosm–
macrocosm homologization pervading practically the whole of Indian culture, most apparent in the ubiquitous
pantheism, where the cosmos is viewed as the body of God in one sense or another. The classic example is Arjuna’s
vision of Kṛṣṇa in the eleventh canto of the Bhagavad-Gītā (a remarkably Sāṁkhyaistic text). This is a truly archaic
feature of Hinduism and Sāṁkhya, going back to prehistoric ideas and appearing already in the R.gveda,21 although it is
perhaps not of Aryan origin.22 A somewhat similar situation obtains in Vasubandhu’s Abhidharma-Kośa, heavily
influencing Tibetan Buddhism in this respect; here mental (meditational) states and cosmogonical events are
identified.23
The problem, according to Michel Hulin is that “the Sāṁkhya thinkers did not pay much attention to dilemmas that
are crucial to us, like ‘is there only one cosmic buddhi [intellect] or as many buddhis as individual beings?’”24 Eli Franco
points out the absurdity in no unclear terms: “Typical psychological and individual terms like cognition, ego, mind,
sense organs, and even hands, feet, tongue, anus and penis, become trans-individual and obtain cosmological
dimensions.”25
Hulin’s explanation is that since the followers of Sāṁkhya strove to leave this world of individual material existence,
the confusion was not relevant for them:

In the wake of discrimination, there is no ground anymore to contrast the personal with the universal perspective. As for the
“temporary” continuation of individual, psychic experience, the Sāṁkhya thinkers, quite understandably, were prepared toadmit a
certain degree of apparent contradiction within it, as a mark, so to say, of its ultimate lack of authenticity.26

Johannes Bronkhorst found traces in several commentaries indicating that the ambiguity was resolved in some
subschools: the first principles were cosmic, the last ones individual. “If the thinkers of classical Sāṁkhya did indeed
not confuse these two, they must have somewhere drawn a line, in the middle of their evolutionary scheme, to
distinguish between cosmological and psychological (or rather: individual) essences (tattva).”27 His carefully worded
conclusion is,

Does this [ambiguity] still hold true for the main thinkers of classical Sāṁkhya? As we now know, the answer must be a qualified no.
It is true that cognition and ego—i.e. mahat/buddhi and ahaṁkāra—appear to have been shared, and therefore cosmological, entities
for some, though not all Sāṁkhyas. Other elements—in particular mind, sense organs, as well as hands, feet, tongue, anus and penis
—were looked upon as only individual, not trans-individual or cosmological entities. The tanmātras remain enigmatic.28

Bronkhorst did not analyze the SK itself, probably because he thought it self-evident that it is the very source of the
confusion. But that is not true: in the relevant passages Īśvarakṛṣṇa’s meaning is always individual. He just selected a
wording that could be interpreted cosmogonically, thereby accommodating his text to the needs of some traditionally
minded subschools.
If somebody really wanted to combine a cosmological account with the description of individuals, that was easily done
in a coherent way—as, for example, Vijñānabhikṣu (sixteenth century) did in his commentary on Sāṁkhya-Sūtra 3.10:

If there is one subtle body (liṅga, transmigrating entity), how could the experiences of each person be different? On this the Sūtra
says:
“10. Differentiation to individuals is according to the particular karma.”
Even if at the beginning of the creation there is but a single subtle body, connected to the god Hiraṇyagarbha (Golden Womb), still
later it will be differentiated to individuals, its parts will also be separate as individuals. As now the parts of the single subtle body of
the father will be separate as the subtle bodies of his sons and daughters. The Sūtra says the reason: “according to the particular
karma,” and that means that according to the acts (karma) causing experiences etc. of different beings.29

This model was widely used in Vedānta texts, for example, in Sadānanda’s Vedānta-Sāra. But the idea itself is ancient,
already in the Puruṣa-hymn of the Ṛgveda (10.90) everything in the world originates from the different parts of a
primordial cosmic giant.
But clearly Īśvarakṛṣṇa did not want anything like that—there is no sign of a two-step creation, first cosmic, then
individual. We have instead the problematic, ambiguous passage. Is he talking about the origin of the world, or an
individual, or both?

The contact of the soul and the unmanifest: that causes the creation. From nature the Great, from that egoity, from that the group of
sixteen; also from among those sixteen from five, the five elements... . Egoity is appropriation. From it, two kinds of creation proceed:
the eleven powers and the five sensibilia... . The sensibilia are simple; from them, the elements: five from five.30

In Sāṁkhya technical terminology, the Great is just another name of the buddhi, understanding; probably it is just short
for the Great Self (Ātman Mahat). At the beginning of the series we have the clearly single, cosmic unmanifest nature,
at the end the plural, but again cosmic elements (earth, water, etc.) and sensibilia (smell, sound, etc.). In between there
are entities characteristic only of living beings: understanding, egoity, the coordinator manas, the senses, and the
biological faculties.
As an abstract, philosophical creation story it is very attractive. After the first contact of unconscious matter with
soul, there arose sentience (buddhi) in matter; and with it, differentiation started (I and not-I, subject and object—i.e.,
the principle of egoity); with this difference, the subject could react to the object (perception), and then the subject
could influence the object (the karmêndriyas, the powers of action); the object, as perceived, distinguished the sensibilia
sound, color, and so on; and finally, the combinations of the sensibilia resulted in the five basic elements.
If this is the creation of a single Great Person, then it must have started with a single, very special soul—and he would
be God, Īśvara. Then why does everybody contrast godless (nirīśvara) Sāṁkhya with theist (sêśvara) Yoga? Even worse,
did Īśvarakṛṣṇa simply forget about all the simpler individuals like us, since there is not even a mention of the origin of
other beings?
I think that this text is not a creation story at all,31 and its primary reading is just a list of the components of any
living being, indicating the interrelations of the components. Īśvarakṛṣṇa may have belonged to that group in the
Sāṁkhyan fold who taught that the connection of the soul and matter is beginningless; or he may have been agnostic on
the point. But, since he was trying to write the classic for all subschools, he offered this description in a form that can
be read as the beginning of such a connection, either as a new soul entering an eternal world, or even as a creation or
re-creation story. Later tradition preferred the last version, because the idea of the cosmic cycles gained general
acceptance in Hinduism.
The key to the possibility of this multiple interpretation is the term sarga, so far carelessly translated in its usual
sense of “creation.” But it can have other meanings as well: etymologically it is pouring out, emission; and it can mean a
troop and a herd as well. In the SK it is a frequent word; let us consider its occurrences:
In verse 21, sarga is the result of the contact of soul and matter (in the next verses followed by the description of the
twenty-three manifest entities). In verse 24, we have the sarga of the eleven powers and the sarga of the five sensibilia;
in 46, the sarga of mental states (pratyaya); in 52, the sarga of the transmigrating entity (liṅga, consisting of eighteen
tattvas) and the sarga of the dispositions (bhāva). Verse 53 speaks about the divine sarga with eight divisions, the
fivefold animal sarga and one kind of human sarga; taken together, the sarga of living beings. In 54, sarga is clearly the
same, that is, of living beings. Finally, in 66, we find a pun: here sarga is both ejaculation and the continuation of
physical existence.32
It seems that the dominant meaning of sarga in the SK is clearly “group,” especially of some fundamental entities or
categories of the system. And thus the beginning of the apparent “creation-story” will be simply: “The group (of the
twenty-three manifest entities) is the result of the contact of matter with soul.”33 The following series of Ablatives (“from
...”—starting with “From nature the Great,” SK 22 etc. quoted above at note 30) do not necessarily express origination;
rather they suggest direction of dependence.
To sum up: Īśvarakṛṣṇa gave the description of the constituents of a human being in a traditional form, so that those
to whom it was important could read it as an emanation scheme, either of an individual or of the whole cosmos; or of
both. In Īśvarakṛṣṇa’s philosophy such an emanation has no role to play and no attempt is made to substantiate it. Still,
it is a value-ordered presentation starting from the innermost core of a rational human and moving outward, “down” to
the physical constituents of the body and of the external world.

Soul as pure consciousness

In the complex system of the twenty-three entities of manifest nature we have all the functions of the Cartesian res
cogitans fulfilled by understanding, egoity, and manas. It seems to be a closed naturalistic description of the world; but
in Sāṁkhya there is also an immaterial entity, the soul, puruṣa. What could its function be? It is not needed even for
such popular purposes as immortality, since the transmigrating entity is immortal, but it is just a complex of the
intangible material entities.34
Puruṣa is pure consciousness; today it is easier to grasp the concept than half a century ago. If we feel, as most of us
do, that our computers are fully unconscious; and if we think that it would not change even if they were even more
advanced and would be able to do whatever an average human intellect can do; and such computer would control a nice
human-looking artificial body—we got the idea. Consciousness is what a robot does not have, while a dog has it.
This concept is extremely exciting philosophically as Paul Schweizer demonstrated in his classical paper.35 Instead of
the familiar body/mind dualism, we have here a “new” type, mind/consciousness dualism. Schweizer showed that “the
ancient Sānkhya-Yoga version of dualism provides a more felicitous dividing line between substances than does the
Cartesian parsing of mind and matter,” because it “avoids one of the most serious pitfalls of Cartesian dualism, since on
the Indian account, mental causation does not violate physical conservation laws.” This is because the mind (the
internal instrument), in itself unconscious, is a subtle, but material organ, while “consciousness plays no causal role in
the transformation of mental structures, but rather is a passive ‘witness’ to some small portion of these.”36
Another, equally important feature of this analysis is that it neatly separates the two salient problems of the magic of
our internal experience, which are often considered two sides of the same coin: subjectivity and intentionality.
Intentionality, or “intrinsic aboutness,” or the “semantic content” of our minds’ information-processing will largely
appear as an illusion.

It is the conscious, subjective aspect of visual perception which serves to motivate the introduction of a distinct metaphysical
category, not the causally induced representational structure of perception, since it is theoretically feasible that the latter can be
explained in terms of unconscious mechanisms, of generally the same sort that would be applied in the case of robotic “vision.” ... It
is consciousness, rather than content, which provides the most compelling impetus for dualism.37

It is worthy of noticing (as Schweizer does) that this concept of consciousness does not automatically entail a substance
dualism. One might be inclined rather to follow Searle, who in his Chinese room thought-experiment suggested that
consciousness is the result of the brain and so on having special “causal powers.”38
Since consciousness in this interpretation is detached from intentionality, it seems logical to construe it as completely
without content. Indeed, that is what happens in later Sāṁkhya; this tendency is already dominant in the commentaries
of the SK. Now this rather peculiar notion was perhaps best described by Gerald Larson, so I will quote him at length:

[Sāṁkhya is] a dualism between a closed, causal system of reductive materialism (encompassing “awareness” or the “private” life of
the mind), on the one hand, and a non-intentional and contentless consciousness, on the other... . [C]​onsciousness (puruṣa) cannot
think or act and is not ontologically involved or intellectually related in any sense to primordial materiality other than being passively
present. Consciousness, in other words, is sheer contentless presence (sākṣitva).39
It is outside the realm of causality, outside space and time, completely inactive, utterly simple, unrelated apart from its sheer
presence, uninvolved in emergence or transformation, without parts, completely independent ... neither an object nor a subject (in
any conventional sense), verbally uncharacterizable, a pure witness whose only relation to primordial materiality is sheer presence,
utterly isolated, completely indifferent, ... a nonagent.40

Where can this disquieting, anomalous conception of the subject come from? It is difficult not to recognize the Vedāntic
idea of absolute pure consciousness, that is, ātman=Brahman, and I think this influence must have been overwhelming
in the long run. In Advaita Vedānta, however, quite logically there is only one of this pure abstraction of consciousness
shared by all sentient beings, while in Sāṁkhya each person has their own unique soul.
In the Sāṁkhya-Yoga tradition, however, there is a strong introspective evidence for a concept like this, since here
meditation is over-important. The fundamental form of meditation is emptying the mind, which means completely
relinquishing the internal stream of thought, abandoning the normally never-stopping silent debate inside. In this state
the meditator experiences nothing, thinks nothing, feels no change at all—yet they still remain themselves. “Yoga is the
suppression of the functions of the mind. Then the soul remains in its own form.”41 Therefore while for us a nonthinking
self is a very difficult and perhaps unnatural idea, something like this was almost self-evident in the Indian tradition.
The timeless, causally unrelated, contentless consciousness characterized by Larson above, unfortunately, can have
no function in the system (or in a person), and therefore practically destroys the Sāṁkhya position. (Historically it did
disappear.) As Śāntarakṣita, the great Buddhist scholar in the eighth century objected in his Tattva-saṁgraha: “if
consciousness endured always in the same form, how could [it] be the enjoyer of many kinds of objects?”42 And there
can be no doubt that in Sāṁkhya the soul is the “enjoyer,” the experiencer (bhoktṛ), for it is one of the key arguments
used by Īśvarakṛṣṇa to prove the soul’s existence (in SK 17). In fact his concept of the soul was far less abstract than
that of the commentators.43
Īśvarakṛṣṇa clearly maintained that the soul is inactive (a-kartṛ, SK 19–20), but only in the sense that it does not move
in space44 and it cannot move a physical object. It is also “qualityless” (not tri-guṇa, SK 11), but only in the very special
sense that it does not have the three guṇas of nature, which are but the fundamental aspects of materiality. So it means
only that the soul is immaterial; but it can have individual features and changing states. “There are many souls; for
birth, death and the instrument are regulated individually, and we do not act at the same time.”45 This would be clearly
meaningless, if the souls were not responsible somehow for the individual acts.
The soul is the subject—an isolated, neutral spectator or witness.46 This aloofness means only that the soul cannot be
killed or objectively harmed, but it can really feel bad: “In this world the conscious soul suffers from the pain caused by
aging and death.”47 Like someone on learning the loss of their riches from the news can really be pained without any
physical harm done. The soul identifies with the body-mind complex and therefore feels the processes of life as its own.
It is clearly stated to be the knower and experiencer.48
Although it is repeatedly said that the puruṣa is not the agent, this can mean only that it lacks the ability to physically
move an object. The terminology of the two great systems of the material world—the “instrument” (karaṇa), the thirteen
psychophysiological entities; and the “field of activity” (kārya), the ten physical entities—seems to require an agent
(kartṛ), but we are told that “no-one operates the instrument”49; only “the passive one seems like the agent, although
the guṇas are the agent.”50 At another place, however, the soul seems to govern or superintend (adhi-ṣṭhā, SK 17) the
person. How is this possible, since it cannot move or change anything that is material? The only answer that we get
again and again is that nature follows the soul’s aim or purpose (SK 13, 17, 31, 42, 56, 58, 60, 63, 65, 68, 69).
To sum up: Īśvarakṛṣṇa’s soul or consciousness, the source of subjectivity and inherently private experiences is an
immaterial entity not in space, but in time. It has temporal states, it is conscious of our experiences: it reacts to
materiality (more exactly, to understanding or perhaps the whole of the mind). It “sees” and it suffers. It cannot move
physical objects, but nature (more exactly, understanding) reacts to its states, “follows its aim.”51

Materiality, the three qualities, and the unmanifest

The Sāṁkhya analysis of matter is another highlight of the system. Prakṛti, “nature” is not like the everyday European
mechanical concept of dead matter, modeled more or less on billiard balls. It is not dead stuff, rather an alive creative
process: it is essentially productive. It is not a loose heap of individual substances like atoms but continuous or
inseparable. Most importantly it is made up of the three qualities.52
The theory of the three guṇas or qualities is the most important contribution of Sāṁkhya to Indian culture. In
Īśvarakṛṣṇa’s presentation it is a refined, grand vision of everything there is, except for the soul. It is a relevant
description of everyday physical processes as well as mental phenomena. Here again we meet with the problem: are the
guṇas psychological or cosmological? In this case, however, the answer is clearly “both.” And it is not due to
Īśvarakṛṣṇa’s failure to grasp the difference, rather to a deep insight into the fundamental, common structure.
The three qualities, tamas, rajas, and sattva53 represent three aspects of any material phenomenon; as some
reflection shows, it is impossible to conceive of any material object where any of the three is missing. Tamas is
inertness, the tendency to remain the same; therefore mass, conservation and stability, rigidness. Rajas is energy,
mobility, the impulse to move and change, repulsion, and separation. Sattva is information, ordering, structure,
coherence, cohesion, and attraction. These are the three threads (another meaning of the word guṇa) of which reality is
spun. They are omnipresent, but not distributed equally—lightning is more rajasic than a tree, a book is more sattvic
than a stone. Their inherent dynamism, cooperation, and rivalry drive all processes from evolution to decay.
There is a facile, ubiquitous tendency to identify the guṇas (at least as psychic factors) with the triad happiness, pain,
and bewilderment (sukha, duḥkha, and moha). Although the commentaries frequently do this, thereby divesting the
theory from much of its explanatory power (but making it easier to grasp), there is nothing in the SK that could justify
or even allow this simplification.
There is another material entity beyond the twenty-three already discussed, namely, the unmanifest (avyakta; or
principal, pradhāna). This is a theoretical construct, it can never be experienced. Its supposition is necessitated by the
unity of the world: there must be something invisible that keeps the whole thing together, that guarantees that
everywhere the same things or events are possible, that provides for the universal validity of natural laws. The idea is
somewhat similar to empty space, but more alive, perhaps it could be compared to a physical force field, especially a
gravitational field in the general theory of relativity.
Since the unmanifest is the underlying cause of all material entities, its properties can be deduced from its effects.
The Sāṁkhya theory of causation is about the relation of two things (not events). It is a mildly deterministic theory
stating that nothing “unexpected” appears in the process, the features of the effect are grounded in the features of the
cause(s). Basically this is a rejection of the possibility of emergence (in the strong sense).54 This theory necessitated the
acceptance of the nonmaterial soul: since there is no consciousness in the guṇas, but we are conscious, without the soul
it would be a case of emergence.55
As all material entities are characterized by the three qualities, the same must be true of the unmanifest: “the
unmanifest is proven to be such [as described in SK 11, see note 52], since the effect consists of the qualities of the
cause.”56 Perhaps the difference is that since the unmanifest is unmanifest (i.e., imperceptible), here the guṇas must be
present in their potential form.

Substance reductionism
The three qualities are not properties or attributes of something else (a substance or substrate): forgetting about the
soul for the moment, there is nothing beyond them. They do not characterize an object different from them—they
themselves make up the whole word and all the things within it. This well-known position of Sāṁkhya is, in fact, not
distinct from the basic concept of the three qualities. It logically follows from the statement that everything is
characterized by the three guṇas. For if they were properties of a substance different from them, that substance in itself
should be without the three qualities, and that is ex hypothesi an impossibility.
So the three qualities do not qualify substances, they constitute them. There is nothing in the substances but the
three guṇas. This seems to be a particularly clear case of a strong substance reductionist theory, yet Sāṁkhya is not
normally understood in this way in modern research. The reason is not very clear for me; I suspect that perhaps the
very idea is so alien to our regular European ways of thinking that most students could not take this statement at face
value. Since the position of Sāṁkhya (stating that everything is the three guṇas and nothing else) could not be called
into question, the only way open was to reinterpret the guṇas as not qualities at all. Perhaps few scholars would now go
so far as directly stating that they are substances (as Dasgupta or Hiriyanna did57); rather they opt for an innovative
terminology, like Eliade’s “modes of being of prakṛti” or Larson’s “constituent process.” The most widespread solution is
to use the word “constituent.”58
One difficulty with this approach is that constituents can occur singly, and you can mix them, but in standard
Sāṁkhya there is no such thing as “pure sattva” to which some “unmixed rajas” could be added. The guṇas always
appear together; they could be called “constituent qualities” (for qualities cannot appear alone), but perhaps “aspects of
materiality” is more directly illuminating.
In any case, if the guṇas were something else and not qualities (in some sense), why would they be called that? The
usual answer is similar to Wezler’s: “The conception of the three guṇas ... was developed at a time when Indian thinkers
‘had not yet learned’ to distinguish between substance as such and its qualities or properties.”59 I find this suggestion
implausible, considering Īśvarakṛṣṇa’s extremely refined conception of the three guṇas and at the same time the plain
impossibility of the supposition that he was not perfectly aware of the Vaiśeṣika analysis of substance and quality
(dravya and guṇa).
This consensus of modern scholarship started to change with Wezler’s fundamental paper from 1985 quoted above,
where he identified some Sāṁkhya tradition where substance is considered nothing but a combination of qualities—this
time, however, of undisputable qualities, that is, the sensibilia color, sound, and so on. He found this doctrine mentioned
both after Īśvarakṛṣṇa (in the Jaina Mallavādin’s Dvādaśāra-Naya-Cakra and its commentary by Siṁhasūri) and also
much earlier (second century BCE) in Patañjali’s Mahā-Bhāṣya. Patañjali’s commentator Kaiyaṭa even connects this
theory with the three standard qualities of Sāṁkhya: “Sattva, rajas and tamas are the qualities. The five qualities sound
etc. are developed from them and so consist of them. Pots and the like are compounds made of the [five], and there is no
substance as a whole beyond them.”60
Bronkhorst in two papers61 added further witnesses of the doctrine (Bhartṛhari, his commentator Puṇyarāja,
Vasubandhu, and Dharmapāla), and connected it to an earlier phase of Sāṁkhya where the sensibilia were among the
twenty-five entities. He suggested that possibly even Īśvarakṛṣṇa belonged to this phase, provided that he understood
the tanmātras as the sensibilia.62 Alex Watson in his 2006 book aptly summarized previous research, adding
Rāmakaṇṭha (tenth century) to the known witnesses.

Despite all these references in the texts of other schools, neither Wezler nor Bronkhorst were able to find a statement of this view in
a surviving Sāṅkhya source. It is a measure of how much that was central to this tradition is lost ... It is striking that an author as late
as Rāmakaṇṭha refers to this as Sāṅkhya view centuries after it had been abandoned by Sāṅkhyas.63

Now it seems that this idea was not at all given up by Sāṁkhya, at least not before Vijñānabhikṣu in the sixteenth
century. In the Yukti-Dīpikā itself there are at least four references to precisely this form of the doctrine, that is, where
it is stated that substances are but combinations of the qualities and these are not the usual Sāṁkhya guṇas (tamas,
rajas, sattva).64 So now we see that this sort of substance reductionism is proper Sāṁkhya, well attested in the classical
period—still, it is atypical. It is like saying that a house is made of clay. A house is built of bricks, and bricks consist of
clay; and clay is a kind of matter. So it is true, a house is made of clay, but Sāṁkhya normally would say that it is made
of bricks. Or, of matter.
Sāṁkhya, like most Indian traditions, thinks that everyday objects are made up of the five elements (earth, water,
fire, air, ether).65 What is peculiar to Sāṁkhya is the position that the elements consist of the five sensibilia.66 This is a
surprising theory; however, it is perfectly convincing, as can be seen through the following argument: All that I know of
the world I learn through the senses. So I have absolutely no ground to suppose that there is anything in a pot beyond
its sensible qualities, color, shape, sound, and so on.
This theory is unambiguously present in the SK “emanation theory” where it states that the five elements arise out of
the five tanmātras, which clearly stand for the sensibilia.67 Therefore it is not particularly important to state separately
that everyday objects (as consisting of the elements) consist of the sensibilia.
Once we have convinced ourselves that substance reductionism is an important element in Sāṁkhya, we are prepared
to realize that the guṇas are indeed what their name says—they are qualities. Sāṁkhya is substance reductionist at the
deepest level: matter itself, prakṛti, is nothing but its constituent qualities, the three fundamental aspects of any
material reality, inertia, energy, and information.
Not only nature itself, but all its manifestations, all material objects and phenomena are reducible to the three
qualities. “Therefore, recognising that bodies, even up to Brahmā, cannot abide, for they are but compounds of sattva
etc., like pots.”68 Now this is the fundamental insight of Sāṁkhya, compared to which the particular cases (like the
reduction of the objects to the elements, or directly to the sensibilia; or the reduction of the sensibilia to the three
guṇas69) are of little importance and are therefore seldom mentioned in Sāṁkhya texts proper. Perhaps some of these
seemed more unusual to other traditions and therefore they remarked on them pointedly, making these particular forms
of Sāṁkhya reductionism more conspicuous in non-Sāṁkhyan texts.
Īśvarakṛṣṇa himself was a conscious and consistent substance reductionist: he never uses a word denoting substance,
like dravya, vastu, dharmin, or svabhāva.70 The tattvas, “entities,” are again either not substances or clearly reducible;
even their names are often action nouns like buddhi, “understanding,” or adjectives like avyakta, “unmanifest.” We have
seen that he derives the elements from the sensibilia, and he defines all the psychophysiological factors as functions,
not things. Understanding is judgment, egoity is appropriation, and so on. He even says that the three factors of mind
can be viewed either separately or as a unit: “The particularity of the three is their function; this is not common. Their
common function as an instrument is the five winds (breath etc.) [i.e., the life functions].”71 And this is not an exception
—in the table of the entities above it is clear that there are several functional groups among them and most of these can
be seen as a unit (while distinctness is supposed to be a characteristic feature of substances). Only nature as a whole,
the Universe, or its abstraction, the unmanifest could be seen as proper substances: but they, too, are nothing but the
combination of the three qualities. Even agency, which is normally thought to be a privilege of substances, is attributed
to the qualities (SK 20, guṇa-kartṛtve).
Substance reductionism is an idea that is difficult to grasp for most people in the European tradition while it is a
fundamental and stable feature of Sāṁkhya, Śaiva Siddhānta, and Buddhism. Why could these people accept this
concept so easily? Since these philosophies seem to be independent of the Vedic tradition, it is tempting to speculate
that a different linguistic background may have been at work here. In most European languages and also in Sanskrit the
opposition noun–adjective is particularly strong. We can only speak in substance–quality language, and it makes it next
to impossible to think in another way. In the Dravidian languages (as in Tibetan and Chinese: that may be relevant for
Buddhism) there is no separate grammatical category for adjectives, so for people who can think in these languages
substance reductionism (to qualities) is no more alien than analyzing a substance in terms of its components, for
example, elements.72

The Buddha’s master

Insubstantialism is a very characteristic philosophical doctrine; and we see now that it is shared by Sāṁkhya and
Buddhism.73 Is this a mere coincidence? Clearly not, for there are a lot more identical or similar teachings in the two
systems. The innumerable common points between early Buddhism and classical Yoga—which is, after all, a version of
Sāṁkhya—have been repeatedly pointed out.74 Let us mention a few central features shared by the SK and the Buddha.
The starting point of both philosophies is the universality and unavoidability of suffering, duḥkha. This is the first
noble truth (ārya-satya) of the Buddha, and the very first word of the SK; its ground is the transience of all existence:
aging and death.75 And, of course, both traditions promise to show the way out.
They give an analysis of the world only inasmuch as it is essential for liberation. They do not give a description of the
cosmos, the Earth, the Sun, or the seas; they do not talk about the gods, although their existence seems to be naturally
accepted.76 In fact their description of reality is basically an anthropology only. Clearly the picture is almost identical
(with the obvious difference of the puruṣa, the nonmaterial soul in Sāṁkhya), although the terminology differs. Both
analyze the human being as a compound of five entities; listing these, I first give the Sāṁkhya concept, then the
Buddha’s term taken from the frequently mentioned group of the “five skandhas.” (1) Body (and all external physical
realities): kārya–rūpa; (2) the connection between external and internal: bāhyakaraṇa–vedanā77; then the mind analyzed
into three factors, that is, (3) coordinator: manas–saṁjñā; (4) personal coloring: ahaṁkāra–saṁskārāḥ; (5) conceptual
thinking: buddhi–vijñāna.
In the liberating praxis of both, a key feature is the disidentification of the subject with exactly these factors.78 Even
the wording is practically identical: “Practicing so with the entities: ‘I am not [this], [this] is not mine, [this] is not the I’
...” in the SK; “He considers [the five skandhas] so: ‘this is not mine, I am not this, this is not my self’” in the extremely
important sermon of the Buddha, The simile of the cobra.79
Last, probably as a logical consequence of their rejecting static substances, both traditions emphasize causally
determined processes of change, in Sāṁkhya called pariṇāma, in Buddhism pratītya samutpāda (typically translated
“evolution” or “change,” and “dependent origination” or “conditioned arising,” respectively).
Now with so many points in common, it is natural to ask: who learned from whom? Although the two schools must
have been mutually influencing each other for centuries, still I think that the basic direction can be identified. The
Buddha was not shy to point out if something was his own discovery or innovation, like the Middle Way (rejecting
ascetic practices) or his gentle meditation technique (“letting go” instead of Yoga’s “oppressing”). In case of his
anthropology, the five skandhas, he clearly uses a system he inherited; for in his very first sermon he merely mentions it
without feeling the need to explain the idea.80
After Gautama had left his home seeking the way to liberation from suffering, he had two teachers. Arāḍa Kālāma, his
first teacher, was—according to his earliest detailed biography81—a Sāṁkhya philosopher. Of course, this late tradition
could be mistaken; the older sources are less explicit. From the Buddha’s account of his own education, we learn only
that Kālāma had a school (gaṇa) with two texts (ñāṇavāda and theravāda, “knowledge-speech” and “speech of the
elders,” probably a sūtra and its commentary), where they practised awareness and meditation (sati and samādhi). The
deepest, last teaching, that Gautama himself directly realized (sacchi-kata) was ākiñcañña-āyatana, that seems to mean
“the state of not having anything.”82 And this sounds suspiciously similar to the disidentification-meditation mentioned
above.
 According to unanimous tradition, Gautama came from Kapila-vastu, “the ground of Kapila”—and Kapila is, of course,
the legendary founder of the Sāṁkhya school. So it seems perfectly consistent with all that we know of him that the
inherited part of the Buddha’s doctrine comes from an early form of Sāṁkhya.
What remains to be proven is that substance-reductionism is an old enough feature of Sāṁkhya to make it plausible
that this is actually the source of the Buddha’s (and Buddhism’s) insubstantionalist doctrine. Part of it is already clear,
since the five-skandhaanthropology (and especially the term rūpa, “form” for the body) is in itself a reductionist
approach, dissolving the unified concept of person as one substance—and we saw that the Buddha inherited this
concept.
Pañcaśikha is one of the most famous ancient Sāṁkhya masters. In the Mahā-Bhārata (12.211–212) he leads king
Janaka to liberation; this text is one of our most important sources for early Sāṁkhya.83 The text is somewhat unclear
(as many philosophical texts in the Mahā-Bhārata), but the key motifs of substance-reductionism and of disidentification
are unmistakable:

Endless suffering will not cease for him who, through wrong views, sees this collection of qualities as being himself. But where could
the continuous flow of suffering, clinging to the word, find a place, when the view is “not-self,” and also “I am not” and “not mine”?84

Janaka, the king of Videha (where the Buddha also spent some time) is more famous as the patron of Yājñavalkya, the
sage of the Bṛhad-Āraṇyaka Upaniṣad. Many of his questions and even words are the same as those of Pañcaśikha here.
Yājñavalkya’s master was Uddālaka Āruṇi,85 the greatest teacher of the Chāndogya Upaniṣad. Now Āruṇi’s teaching has
many significant Sāṁkhya features,86 most notably a substance-reductionism to three qualities! He calls them “forms”
or “colors” (rūpa), not guṇas, but the idea is exactly the same. They are present in everything, and there is nothing else
in the objects.87
And, as luck would have it, it can be proven that the Buddha knew this teaching. For he refers to a very unusual view,
shared only by Āruṇi and Yājñavalkya,88 according to which a person at death simply dissolves into the Great Being, also
called the Self, ātman. The Buddha, of course, thinks that this is a harmful theory:

There is this wrong view: “The Self and the world are the same. After death I will be that: I will stay identical for eternal times,
constant, stable, eternal, essentially unchanging.” He considers it so: “this is mine, I am this, this is my self.”89

So the Buddha inherited his substance-reductionist ideas from the proto-Sāṁkhya circles of Āruṇi and Yājñavalkya.
What his own innovation was, he made perfectly clear: rejecting their ātman-substance, he created a system entirely
without substances, and so—to the dismay of many—a religion with no soul.
Notes

1. McGinn (1993).
2. The SK will be quoted based on my edition, where some text-critical information is also given. When a variant reading specified in
the edition seems relevant for the purposes of this chapter, it will be marked here with “v.l.” All Sanskrit texts quoted will be
standardized, with hyphens, sandhi-markers, and punctuation added. All translations are mine except when noted explicitly.
3. Larson (1987, p. 77).
4. “Though the Mokṣadharma as we have it offers already a bewildering diversity of often contradicting views, the historical reality at
the time of its first composition was still more complex—each ashram, so to speak, having its own competing version of proto-
Sāṁkhya philosophy and being keen on having it canonized.” Bakker and Bisschop (1999, p. 468).
5. Most scholars consider this text very late (c. fourteenth century). I have tried to show elsewhere that this opinion is not well founded
and that in all probability it is earlier than the second century ce. Ruzsa (2014, pp. 101–107).
6. At least the Nyāya-Sūtra is clearly referred to.
SK 5: trividham anumānam ākhyātam, “it has been explained that inference has three kinds.”
SK 6: sāmānyatas tu dṛṣṭād atîndriyāṇām prasiddhir anumānāt, “Things beyond the senses are proved through inference by
analogy.”
Nyāya-Sūtra 1.1.5: atha tat-pūrvakaṁtrividham anumānam: pūrvavat, śeṣavat, sāmānyato dṛṣṭaṁca, “Then inference is based on that
[i.e., perception]; it has three kinds: like before, like the rest, and by analogy [lit., seen by genus].”
A quasi-quotation from Nāgārjuna’s Mūla-Madhyamaka-Kārikā (16.1&5) again seems unmistakable:
1 saṁskārāḥsaṁsaranti cen, na nityāḥsaṁsaranti te; saṁsaranti ca nânityāḥ. sattve ‘py eṣa samaḥkramaḥ.
5 na badhyante, na mucyanta udaya-vyaya-dharmiṇaḥsaṁskārāḥ; pūrvavat sattvo badhyate na, na mucyate.
“1. If composites transmigrate, they do not transmigrate as eternal; but temporary things do not transmigrate. The logic is the same
with a living being. 5. Composites, whose nature is to rise and pass away, cannot be bound or freed. Similarly a living being is not
bound and not freed.”
Compare SK 62: tasmān na badhyate ‘ddhā na mucyate nâpi saṁsarati kaś-cit. “Therefore no-one is bound, nor freed, and also does
not transmigrate.”
Also the fact that the plurality of souls needed proof (SK 18) indicates that a system of Vedānta was known to Īśvarakṛṣṇa.
7. I have argued for this in Ruzsa (2010, pp. 428–430), and in more detail in Ruzsa (1997a). A possible reconstruction of the core text
will be found in my edition of the SK.
8. A particularly nasty example from SK 41: vinā viśeṣair na tiṣṭhati nir-āśrayaṁliṅgam. “The transmigrating entity cannot stay with no
substrate, without the elements,” that is, without a gross material body. The message is perhaps that even gods must have a gross
body and there is no Bardo, antara-bhāva, intermediate existence between dying and rebirth. But if we read it as vinâviśeṣair (i.e.,
sandhi of vinā + aviśeṣair), the meaning will be “without the subtle elements,” which is but a scholastic insistence on the precise
build-up of the liṅga, the transmigrating entity.
9. Larson (1987, pp. 18–19); Bronkhorst (1994, p. 315); Ruzsa (1997b), in more detail in Ruzsa (2014, pp. 97–114).
10. In many traditions, including many commentaries of the SK, the elements are called mahā- or sthūla-bhūtas (great or gross
elements), while the tanmātras are referred to as sūkṣma-bhūtas, “subtle elements.” These subtle elements are theoretical entities
(although advanced yogis supposedly can see them) with no clear function or justification; they are subtle earth, subtle water and so
on. They make up the subtle body, which we carry on while transmigrating, while the gross elements (the gross body) is left behind.
This interpretation is alien to the SK and therefore it is seriously misleading to translate—in this text—these entities as “gross and
subtle elements.” Although no list of the tanmātras is given, it is perfectly clear that they are the sensibilia:
28 rūpâdiṣu pañcānām ālocana-mātram iṣyate vṛttiḥ
34 buddhîndriyāṇi teṣām pañca, viśeṣâviśeṣa-viṣayāṇi
“The function of the five [senses] is merely observing colour etc... . Among them [i.e., among the ‘powers’] there are the five senses;
their objects are the elements and the aviśeṣas [= tanmātras].” This makes sense only if the end of the list of the twenty-five entities,
the last ten, that is, the tanmātras and the elements, starts with “color,” and further, the list should contain the objects for the other
senses as well. It follows that the tanmātras refer to the sensibilia.
But Īśvarakṛṣṇa, as discussed above in note 8, with one of his characteristic ambiguous phrases makes room for the approach
where the tanmātras are constituents of the transmigrating entity (liṅga), and are therefore best understood as subtle elements. In
SK 40 the ambiguity remains: mahad-ādi sūkṣma-paryantam saṁsarati ... liṅgam, “the liṅga [consisting of the entities] from the
Great (i.e. buddhi, understanding) down to the subtle [elements] transmigrates.” However, the sentence could also mean only that
“the liṅga transmigrates from the great down to the tiny.”
11. SK 28: vṛttiḥ/ vacanâdāna-viharaṇôtsargânandāś ca pañcānām.
12. I am not aware of any philosophical tradition where manas would be a regular term for “mind.” In Yoga citta is used, in Nyāya and
Vaiśeṣika ātman (working together with manas) is the mind. Vedānta seems mostly to follow Sāṁkhya usage.
13. Saṁkalpaka, “who fits together,” “who joins,” “connector” is often misunderstood (already by the Indian tradition), based on the
everyday meaning of saṁkalpa, “intention,” leading to an interpretation of manas as volition.
14. This is the less frequent, but probably more original version of SK 27:
saṁkalpakam atra manaḥ, tac cêndriyam ubhayathā samākhyātam. antas tri-kāla-viṣayaṁ; tasmād ubhaya-pracāraṁtat.
15. Quite independent of the Indian tradition, Wilhelm Dilthey’s (1890) unique analysis seems to grasp the problem in some of its depth.
He substitutes for the usual model (the I, the subject, as already given cognizes the external world) a dynamic approach where the I
arises in the process of frustrating interactions with external reality; the frustration defines it as alien, as not-I, as object.
16. In the SK we have only ahaṁkāra, but the expression in SK 64 clearly shows that at least these three aspects, but most probably
also the three terms were known: evaṁtattvâbhyāsān: ‘nâsmi, na me, nâham’ ity, “Practicing so with the entities: ‘I am not [this],
[this] is not mine, [this] is not the I.’”
17. SK 23: adhyavasāyo buddhir. dharmo, jñānaṁ, virāga, aiśvaryam
sāttvikam etad-rūpaṁ. tāmasam asmād viparyastam.
It is interesting to note that the four sattvic forms of buddhi are the four basic religious aims (as clearly spelled out in SK 44 and 45):
dharma for Vedism, knowledge for Sāṁkhya, dispassion for Buddhism and the other śramaṇa religions, sovereignty (or magic
powers) of Yoga.
18. SK 37: sarvam pratyupabhogaṁyasmāt puruṣasya sādhayati buddhiḥ
sâiva ca viśinaṣṭi punaḥpradhāna-puruṣântaraṁsūkṣmam.
19. The last example is somewhat hypothetical, suggesting that egoity adds the coloring of our personalities (saṁskāras: past
impressions, memories, inclinations, unconscious traumas, instincts) to the information processed. This idea was inspired by Tibor
Körtvélyesi’s (2011, pp. 49–50) analysis of the fourth Buddhist skandha, that is, saṁskāras.
20. This information-processing series is not my idea, it is Īśvarakṛṣṇa’s. He uses the verb “to see” (dṛś) in connection with both the
senses (SK 4, 5, 30) and the soul (19, 21, 65, 66, 59, 61). He calls the external instrument (which includes the senses) the sense
object (viṣaya) of the mind (the internal instrument; SK 33). The whole idea is nicely expressed in SK 35–36. See also the quotation in
note 43.
21. Ruzsa (forthcoming).
22. Ruzsa (2007).
23. Szegedi (2009, pp. 42–52).
24. Hulin (2007, p. 59).
25. Franco (1991, p. 123).
26. Hulin (2007, p. 60).
27. Bronkhorst (1999, p. 683).
28. Bronkhorst (1999, p. 688).
29. nanu liṅgaṁced ekaṁ, tarhi kathaṁpuruṣa-bhedena vilakṣaṇā bhogāḥsyuḥ? tatrâha: “vyakti-bhedaḥkarma-viśeṣāt. 10.” yady api
sargâdau Hiraṇyagarbhôpādhi-rūpam ekam eva liṅgam, tathâpi tasya paścād vyakti-bhedo, vyakti-rūpeṇâṁśato nānātvam api
bhavati; yathêdānīm ekasya pitṛ-liṅgadehasya nānātvam aṁśato bhavati putra-kanyâdi-liṅgadeha-rūpeṇa. tatra kāraṇam āha: karma-
viśeṣād iti; jīvântarāṇāṁbhoga-hetu-karmâder ity arthaḥ. (p. 90)
Bronkhorst (1999, p. 688) also refers to this example.
30. SK 21, 22, 24, and 38.—The highly suspect verse 25 would even more increase the cosmogonic feel of the text: “The sattvic eleven
proceeds from the Vaikrita (‘Modified’) egoity; the group of the sensibilia from the Bhūtādi (‘Beginning of the Elements’); both from
the Taijasa (‘Brilliant’)”—The Sanskrit original:
21 puruṣasya ... pradhānasya ... saṁyogas: tat-kṛtaḥsargaḥ.
22 prakṛter mahāṁs, tato ’haṁkāras, tasmād gaṇaś ca ṣoḍaśakaḥ; tasmād api ṣoḍaśakāt pañcabhyaḥpañca bhūtāni.
24 abhimāno ’haṁkāras. tasmād dvi-vidhaḥpravartate sargaḥ: aindriya ekādaśakas, tānmātraḥ (vv.ll. ekādaśakaś ca gaṇas,
tanmātraḥ) pañcakaś câiva.
25 sāttvika ekādaśakaḥpravartate Vaikṛtād ahaṁ-kārāt; Bhūtâdes tānmātraḥ(v.l. tanmātraḥ), sa tāmasas; Taijasād ubhayaṁ.
38 tan-mātrāṇy a-viśeṣās; tebhyo bhūtāni, pañca pañcabhyaḥ.
31. Not surprisingly, the regular counterpart of creation, that is, an apocalypse or cosmic dissolution, pralaya is not even mentioned. SK
69 is clearly a later addition; in any case, it does not speak about the dissolution of the world, but of the beings or elements (bhūta),
and it says only that it occurred in the system of the highest ṛṣi (Kapila): sthity-utpatti-pralayāś cintyante yatra bhūtānām, “where
the abiding, arising and dissolution of the beings are thought of.”
32. Verses 21 and 24 have been quoted above, note 30.
46 eṣa pratyaya-sargo: viparyayâśakti-tuṣṭi-siddhy-ākhyaḥ. guṇa-vaiṣamya-vimardāt tasya ca bhedās tu pañcāśat.
52 na vinā bhāvair liṅgaṁ; na vinā liṅgena bhāva-nirvṛttiḥ. liṅgâkhyo bhāvâkhyas tasmād bhavati dvidhā (v.l. dvividhaḥpravartate)
sargaḥ.
53 aṣṭa-vikalpo daivas, tairyagyonaś ca pañcadhā bhavati, mānuṣyaś câikavidhaḥ—samāsato bhautikaḥsargaḥ.
54 ūrdhvaṁsattva-viśālas, tamo-viśālaś ca mūlataḥsargaḥ, madhye rajo-viśālo: Brahmâdi-stamba-paryantaḥ.
66 dṛṣṭā mayêty upekṣaka eko. dṛṣṭâham ity uparatânyā. sati saṁyoge ’pi tayoḥprayojanaṁṇâsti sargasya.
33. This shows, incidentally, that it is somewhat superficial to call understanding or mind (buddhi and antaḥkaraṇa) material entities.
Without the soul there is no mind: “buddhi and ahaṁkāra ... are mixtures of matter and consciousness.” “The buddhi, the first
material product, has two causes, the prakṛti and the puruṣa principle, and is like a knot made of two ropes.” Jacobsen (1999, pp.
225–226).
34. At death only the gross body made up of the five elements dies. The subtler, invisible parts of a living being (the subtle body,
sūkṣma-śarīra or liṅga) move on to build a new gross body. So this transmigrating entity is immortal; and it consists of material
entities (tattvas of prakṛti)—the thirteen-fold instrument (i.e., understanding, egoity, manas, the five senses and the five active
powers) and perhaps the five tanmātras.
35. Schweizer (1993).
36. Schweizer (1993, pp. 858, 849, 850).
37. Schweizer (1993, p. 852).
38. Searle (1980, 422). Searle, however, focuses on intentionality, not on consciousness.
39. Larson (1987, p. 77).
40. Larson (1987, p. 81).
41. Yoga-Sūtra 1.2–3: yogaś citta-vṛtti-nirodhaḥ. tadā draṣṭuḥsvarūpe ’vasthānam.
42. Eka-rūpe ca caitanye sarva-kālam avasthite / nānā-vidhârtha-bhoktṛtvaṁkathaṁnāmôpapadyate? (Tattva-saṁgraha 288). The
translation and the Sanskrit text is taken from Watson (2010, p. 92). This paper is an excellent comparison of four positions about the
eternal, immaterial self: Nyāya (self without essential consciousness), Sāṁkhya (self as pure consciousness), Vedānta (self as
unchanging knowledge), and Śaiva Siddhānta (active self).
43. Candrakīrti, another great Buddhist scholar in the seventh century, still knows this more natural idea, for he writes about the
Sāṁkhya theory of cognition: “When the sound etc. has been grasped by the hearing etc. superintended by the manas,
understanding makes a judgment. Then the soul becomes conscious of the object as judged by the understanding” (śabdâdiṣu
śrotrâdi-vṛttibhir manasâdhiṣṭhitābhiḥparigṛhīteṣu buddhir adhyavasāyaṁkaroti. tato buddhy-avasitam arthaṁpuruṣaś cetayate).
Lang (2010, p. 56).
44. So it does not transmigrate (na saṁsarati, SK 62).
45. SK 18: janma-maraṇa-karaṇānāṁprati-niyamād, a-yugapat pravṛtteś ca
puruṣa-bahutvaṁsiddhaṁ; tri-guṇâdi (v.l. trai-guṇya)-viparyayāc câiva.
46. Viṣayin and vivekin, SK 11; kevalin, madhya-stha, sâkṣi, and draṣṭṛ, SK 19.
47. SK 55: tatra jarā-maraṇa-kṛtaṁduḥkham prâpnoti cetanaḥpuruṣaḥ.
48. jña, SK 2, and bhoktṛ, SK 17, 37, 40.
49. SK 31: na kena cit kāryate karaṇam.
50. SK 20: guṇa-kartṛtve ’pi tathā kartêva bhavaty udāsīnaḥ.
51. The historical process is more complex than an unidirectional movement toward a more abstract conception of the soul, for
Kundakunda, an important Jain author before Īśvarakṛṣṇa, had a conception much like him, while already attributing to the
Sāṁkhyas the idea of absolutely unchanging soul: “If the soul does not change by states like anger, then transmigration will be
impossible—or the Sāṁkhya position will be true.”
apariṇamaṁte hi sayaṁjīve kohâdiehi bhāvehiṁsaṁsārassa abhāvo pasajjade, saṁkha-samao vā.
Bronkhorst (2010, p. 219).
52. SK 11: “The manifest is made up of the three qualities, continuous, object, common, unconscious, essentially productive. The
unmanifest also. Soul is its opposite and also similar [to some other properties of the unmanifest].”
tri-guṇam, aviveki, viṣayaḥ, sāmānyam, acetanaṁ, prasava-dharmi vyaktaṁ; tathā pradhānaṁ. tad-viparītas, tathā ca pumān.
53. These traditional names are better left untranslated. Tamas means darkness, rajas the atmosphere, sattva existence, essence, or
living being.
54. The name of the theory, sat-kārya (“existent-effect”), is wrongly construed as “the preexistence of the effect in the cause”; it could
rather be understood as “effect of existent(s),” meaning that there must be a feature in the cause(s) explaining a feature of the
effect. Ruzsa (2003a, pp. 286–287).
55. SK 17: tri-guṇâdi-viparyayād ... puruṣo ’sti, “There is the soul, for there is the opposite of ‘made up of the three qualities’ [SK 11]
etc.,” that is, for we find in the world features opposite to those characterizing matter (like “unconscious”).—The argument is not
begging the question, for it is not an arbitrary postulate that the guṇas lack consciousness: the guṇas are (abstractions and
generalizations) based on experience, and in non-sentient objects we do not find the slightest trace of consciousness.
56. SK 14: kāraṇa-guṇâtmakatvāt kāryasyâvyaktam api siddham.
The way this inference works has been more fully treated in Ruzsa (2003a, pp. 293–295).
57. Dasgupta (1922–55, pp. 241–245); Hiriyanna (1932, pp. 271–272).
58. Hacker’s (1985, p. 112) position may appear representative: “Diese Lehre hat europäischen Forschern zunächst einmal einige
Schwierigkeiten bereitet, weil wir sonst das Wort guṇa überall mit ‘Eigenschaft’ übersetzen können, und hier scheint nach unseren
Denkmöglichkeiten nur ein Wort wie etwa ‘Konstituente’ zu passen ... Frauwallner aber, der hier tiefer sah ..., hat die Bedeutung
‘Eigenschaften’ beibehalten und hat ganz deutlich ausgesprochen, daß eben die Eigenschaften hier als Substanzen aufgefaßt
werden. Das ist uns Fremd, aber wir müssen uns an diese Fremdheit gewöhnen, sonst können wir eben indische Philosophie nicht
studieren.”
59. Wezler (1985, p. 26).
60. sattva-rajas-tamāṁsi guṇāḥ, tat-pariṇāma-rūpāś ca tad-ātmakā eva śabdâdayaḥpañca guṇāḥ. tat-saṅghāta-rūpaṁca ghaṭâdi, na tu
tad-vyatiriktam avayavi-dravyam astîti sāṁkhyānāṁsiddhântaḥ. Wezler (1985, p. 10).
61. Bronkhorst (1994; 1997). In the second paper he brings to our notice Vasubandhu’s testimony in his Abhidharma-Kośa-Bhāṣya
(III.50a). Although Bronkhorst himself believes that the passage presents “Sāṁkhya in its classical form, and not in its pre-classical
shape, in which no unchanging substrate of properties had yet been introduced,” he clearly overlooked the unequivocal statement of
the Sāṁkhya position denying a substrate beyond the qualities: kaś câivam āha, dharmebhyo ’nyo dharmîti? “Who said that the
substrate is different from the properties”? Bronkhorst (1997, p. 394).
62. Bronkhorst (1994, pp. 311–312). I think that it is the only possible reading of the SK (see note 10). This understanding of the
tanmātras as the sensibilia lived on in the commentaries at the side of the new, “subtle elements” interpretation. See, for example,
Bronkhorst’s main source for “classical” Sāṁkhya, the Yukti-Dīpikā on SK 24cd: tanmātrāṇāṁśabda-sparśâdīnām
ayaṁtānmātraḥsargaḥ, “This tānmātra creation is of the tanmātras, sound, touch etc.” Wezler and Motegi (1998, p. 195).
Bronkhorst’s key argument for the substantiality of the tanmātras is that they have more than one quality. This is, however, not
standard “classical” Sāṁkhya; it appears only in the Yukti-Dīpikā. Besides the passage that Bronkhorst and Wezler analyzed, it is
mentioned cursorily in the introduction to SK 22 (p. 187): eka-rūpāṇi tanmātrāṇîty anye; ekôttarāṇîti vārṣagaṇyaḥ. “Others say that
the tanmātras have one form each; Vārṣagaṇya says that each has one more [than the previous tanmātra].” So this is clearly the
unique position of a single master, not accepted by others.
63. Watson (2006, pp. 187–188). In footnote 196 he also mentions (but does not comment on) the references to the substance
reductionist position in later Sāṁkhya that I sent him.
64. Yukti-Dīpikā (YD) on SK 16c (pp. 163–164): āha: ... na hi vo dharmebhyo ’nyo dharmī! ... ucyate: ... yathā senâṅgebhyo ’nanyatvam
senāyāḥ. “Opponent: For you, there is no substrate other than the properties! ... Answer: [True,] as an army is not different from its
parts.”
YD on SK 23a (p. 189), a Buddhist objects: dharma-dharmiṇor ananyatvād ... vânyatvam iti dosaḥ. “... the properties and the
substrate are not different. Or ... they would be different, and that is unacceptable [for you Sāṁkhyas].”
YD on SK 15b (p. 144): na câikaiko rūpâdīnāṁdravyâkāraḥ, samudāya-dharmatvāt. “Colour etc. singly do not have the form of a
substance, for that is a property of their combination.” A note in ms. D adds: rūpâdi-pañcaka-vyatiriktaṁtāvad dravyaṁnâsti, “there
is no substance beyond the five: colour etc.”
YD on SK 34d (p. 218): sabda-sparśa-rasa-rūpa-gandha-samudāya-rūpā mūrtīr, “physical bodies are combinations of sound, touch,
taste, colour and smell.”
65. It is so obvious that it is seldom mentioned; one example is Gauḍapāda on SK 17 (p. 91):
idaṁśarīraṁpañcānāṁmahābhūtānāṁsaṁghāto vartate, “this human body is a compound of the five great elements.”
66. How exactly they are related is not specified by Īśvarakṛṣṇa. The commentators’ preferred version is that from sound arises ether, if
we add touch, we have air; adding color, fire; adding taste, water; and earth has all the five sensibilia.
67. In most commentaries the “subtle element” interpretation of the tanmātras appears only in the discussion of their role in the liṅga,
the transmigrating entity.
68. tasmāt saṅghāta-mātratvāt sattvâdīnāṁghaṭâdivat / ā brahmaṇaḥparijñāya dehānām anavasthitim. YD 50cd (p. 248).
69. śabda-sparśa-rūpa-rasa-gandhāḥpañca trayāṇāṁsukha-duḥkha-mohānāṁsaṁniveśa-viśeṣāḥ, “The five (sound, touch, colour, taste
and smell) are particular arrangements of the three, happiness, pain and bewilderment.” Steinkellner (1999, p. 670, fr. 4).
70. Actually dharmin and svabhāva do occur in SK 11 and 55, but only in the sense of “essentially.” Āśraya, “substrate” or “support,” is
freely used, but it is not necessarily a substance: in SK 12 we are told that the qualities are the substrates for each other
(anyo’nyâbhibhavâśraya-janana-mithuna-vṛttayaś ca guṇāḥ).
71. SK 29: svālakṣaṇyaṁvṛttis trayasya; sâiṣā bhavaty asāmānyā.
sāmānya-karaṇa-vṛttiḥprāṇâdyā vāyavaḥpañca.
72. It is not suggested that Sāṁkhya has any particular connection with the Dravidian speaking South (although many important
philosophers writing in Sanskrit, e.g., Śaṅkara and Rāmānuja, belong there). It seems that Dravidian languages were spoken in
Northern India before the arrival of the Indo-Aryans, and the continuous substrate influence of Dravidian is apparent in the
development of all Indo-Aryan languages (including Sanskrit) for much more than a millennium. So there is nothing inherently
improbable in the supposition that some thinkers, known to us only through Sanskrit texts, could speak and think in a Dravidian
language—I have argued for the influence of Dravidian on Aryan languages starting already at the time of the composition of the
R.gvedic hymns and continuing at least to the period of late Prakrits (c. 500 CE) in Ruzsa (2013).
73. That the Buddha himself already had a strong anti-substantialist attitude is clearly shown beyond the well-known an-ātman, “no-
self,” doctrine by his philosophical term for the human body: rūpa, “form,” that is, a quality, not a substance.
74. Tandon (1995).
75. Ruzsa (2003b; 2014, pp. 97–114), that is, chapter IX: “Pain and its cure. The aim of philosophy in Sāṁkhya.”
76. SK 54 mentions Brahmā as the highest being within creation. The Yoga-Sūtra and the commentators of the SK speak about Īśvara,
the Lord, that is, God, although he is not the creator. See Bronkhorst (1983). Brahmā Sahaṁ-pati, Sakka the king of the gods, and
Māra the evil appear frequently in stories about the Buddha.
77. Here, however, the Buddhist approach is more passive. The Sāṁkhya concept includes the powers of action as well, while vedanā is
only sensation.
78. As Marzenna Jakubczak (2012, p. 42) so clearly emphasized in her nicely perceptive paper, in Sāṁkhya “the aim is not to identify
directly with puruṣa, but rather to keep disidentifying with the present phenomenal self by means of constant realisation: ‘I am not,
not mine, not I’ (nāsmi na me nāham; cf. SK 64; M[ajjhima]N[ikāya] 109.15–16).”
79. evaṁtattvâbhyāsān: ’nâsmi, na me, nâham’ ity (SK 64)—rūpaṁ (etc.) ’n’ etaṁmama, n’ eso ’ham asmi, na m’ eso attā’ ti
samanupassati (Alagaddûpama-Sutta, Majjhima-Nikāya 22).
80. Saṁkhittena, pañc’ upādāna-kkhandhā dukkhā. “Summarily, the five skandhas of appropriation are painful.” Dhamma-Cakka-
Ppavattana-Sutta, Saṁyutta-Nikāya 5.12.2.1 (= 56.11) = Vinaya-Piṭaka, Mahā-Vagga 1.6.
81. This is Aśvaghoṣa’s Buddha-carita, written half a millennium after the Buddha’s time.
82. Mahā-Saccaka-Sutta, Majjhima-Nikāya 36.
83. Motegi (1999).
84. Mahā-Bhārata 12.212.14–15:
14 imaṁguṇa-samāhāram ātma-bhāvena paśyataḥasamyag-darśanair duḥkham anantaṁnôpaśāmyati.
15 anātmêti ca yad dṛṣṭaṁ, te nâhaṁna mamêty api: vartate kim-adhiṣṭhānā prasaktā duḥkha-saṁtatiḥ?
85. Bronkhorst (2007, pp. 226–231)—On the curious inversion of motifs between the two great sages I have written in Ruzsa (2009).
86. As it was perfectly clear in antiquity: the first important discussion in the Brahma-Sūtra (1.1.5–11) tries to prove that this
interpretation is false—thereby showing that it was widespread to quote the Sad-Vidyā, that is, Āruṇi’s teaching to his son Śvetaketu
in the sixth chapter of the Chāndogya Upaniṣad, as a scriptural authority for Sāṁkhya.
 87. See Ruzsa (2004). It is interesting to note that Āruṇi teaches sat-kārya (“the effect of existent”) in the most literal sense, since his
fundamental entity is not called prakṛti but sat, “existent.”
88. Chāndogya Upaniṣad 6.8–10; Bṛhad-Āraṇyaka Upaniṣad 2.4 and 4.5.
89. Yam p’ idaṁdiṭṭhi-ṭṭhānaṁ’so loko so attā, so pecca bhavissāmi: nicco dhuvo sassato a-vipariṇāma-dhammo, sassati-samaṁtath’eva
ṭhassāmî’ ti tam pi ’etaṁmama, eso ’ham asmi, eso me attā’ ti samanupassati. From The simile of the cobra (Alagaddûpama-Sutta,
Majjhima-Nikāya 22).

Bibliography
Indic texts quoted
Gauḍapāda’s commentary on the SK:
Mainkar, T. G. (2004), Sāṁkhyakārikā of Īśvarakṛṣṇa: With the Commentary of Gauḍapāda. (The Vrajajivan Indological Series 36.) Delhi:
Chaukhamba Sanskrit Pratishthan. (Includes text and translation.)
Mahā-Bhārata:
Dandekar, R. N. (ed.) (1971–76), The Mahābhārata: Text as Constituted in the Critical Edition. Poona: Bhandarkar Oriental Research
Institute.
Mūla-Madhyamaka-Kārikā:
Kalupahana, D. J. (1986), Mūlamadhyamakakārikā of Nāgārjuna: The Philosophy of the Middle Way. New York: State University of New
York. (Includes text and translation.)
Nyāya-Sūtra:
Nyaya-Tarkatirtha, Taranatha and Tarkatirtha, Amarendramohan (eds.) (1936–44), Nyāyadarśanam with Vātsyāyana’s Bhāṣya,
Uddyotakara’s Vārttika, Vācaspati Miśra’s Tātparyaṭīkā & Viśvanātha’s Vṛtti. Calcutta: Metropolitan Printing & Publishing House.
(Repr. Delhi: Munshiram Manoharlal, 1985.)
Translation:
Vidyābhūṣana, Satisa Chandra (tr.) and Sinha, Nandalal (rev. ed.) (1930): The Nyāya Sūtras of Gotama. Allahabad: Pâṇini Office. (Repr.
Delhi: Motilal Banarsidass, 1990.)
Vijñānabhikṣu’s commentary on the Sāṁkhya-Sūtra:
Garbe, Richard (ed.) (1895), The Sāṁkhya-Pravacana-Bhāṣya or Commentary on the Exposition of the Sānkhya Philosophy by
Vijñānabhikṣu (Harvard Oriental Series 2). Cambridge, MA: Harvard University.

SK, Sāṁkhya-Kārikā:
Ruzsa, Ferenc (ed.) (2014b), Īśvarakṛṣṇa’s Sāṁkhya-Kārikā. https://www.academia.edu/4757287/Sa_khya-Karika_-
_reordered_text_and_words.
Translation:
Larson, G. J. (1979), Classical Sāṁkhya. An Interpretation of Its History and Meaning (2nd ed.). Santa Barbara: Ross/Erikson, pp. 255–
277.
YD, Yukti-Dīpikā:
Wezler, A., and Motegi, S. (eds.) (1998), Yuktidīpikā: The Most Significant Commentary of the Sāṁkhyakārikā. Stuttgart: Franz Steiner
Verlag.
Yoga-Sūtra:
Maas, Philipp André (2006), Samādhipāda: Das erste Kapitel des Pātañjalayogaśāstra zum ersten Mal kritisch ediert. Indologica Halensis
—Geisteskultur Indiens. Texte und Studien, 9. Shaker Verlag, Aachen.
Pali texts are quoted from the Vipassana Research Institute’s online edition, http://www.tipitaka.org.

Secondary sources
Bakker, H., and Bisschop, P. (1999), ‘Mokṣadharma 187 and 239–241 reconsidered,” Asiatische Studien/Études Asiatiques, 53/3, 459–472.
Bronkhorst, J. (1983), “God in Sāṁkhya.” Wiener Zeitschrift für die Kunde Südasiens, 27, 149–164.
Bronkhorst, J. (1994), “The qualities of Sāṁkhya,” Wiener Zeitschrift für die Kunde Südasiens, 38 (Orbis Indicus, Festschrift G.
Oberhammer), 309–322.
Bronkhorst, J. (1997), “Sāṁkhya in the Abhidharmakośa Bhāṣya,” Journal of Indian Philosophy, 25, 393–400.
Bronkhorst, J. (1999), “The contradiction of Sāṁkhya. On the number and the size of the different tattvas,” Asiatische Studien/Études
Asiatiques, 53/3, 679–691.
Bronkhorst, J. (2007), Greater Magadha. Studies in the Culture of Early India. (Handbook of Oriental Studies, Section 2: South Asia, 19.)
Leiden: Brill.
Bronkhorst, J. (2010), “Kundakunda versus Sāṁkhya on the soul” in Nalini Balbir (ed.), Svasti: Essays in Honour of Prof. Hampa
Nāgarājaiah for His 75th Birthday. Krishnapuradoddi: K. S. Muddapa Smaraka Trust, pp. 215–226.
Dasgupta, S. (1922–55), A History of Indian Philosophy. Cambridge: Cambridge University Press. (Repr. Delhi: Motilal Banarsidass,
1988.)
Dilthey, W. (1890), “Beiträge zur Lösung der Frage vom Ursprung unseres Glaubens an die Realität der Außenwelt und seinem Recht”
(The origin of our belief in the reality of the external world and its justification). Sitzungsberichte der königlich preußischen Akademie
der Wissenschaften zu Berlin, 1890/2, 977–1024.
Franco, E. (1991), “Whatever happened to the Yuktidīpikā?,” Wiener Zeitschrift für die Kunde Südasiens und Archiv für Indische
Philosophie, 35, 123–138.
Hacker, P. (1985), Grundlagen indischer Dichtung und indischen Denkens. Ed. Klaus Rüping. Wien: Publications of the De Nobili
Research Library 12.
Hiriyanna, M. (1932), Outlines of Indian Philosophy. London: G. Allen & Unwin Ltd. (Repr. Delhi: Motilal Banarsidass, 1994.)
Hulin, M. (2007), “The ego-principle (Ahaṁkāra) as a key concept in the Sāṁkhyakārikā,” in Victoria Lysenko and Michel Hulin, Classical
Indian Philosophy Reinterpreted. Delhi: Decent Books (Centre for Sri Aurobindo Studies), pp. 47–60.
Jacobsen, K. A. (1999), Prakṛti in Sāṁkhya-Yoga (Asian Thought and Culture 30). New York etc.: Peter Lang.
Jakubczak, M. (2012), “Why didn’t Siddhārtha Gautama become a Sāṁkhya philosopher, after all?” in Irina Kuznetsova, Jonardon Ganeri,
and Chakravarthi Ram-Prasad (eds.), Hindu and Buddhist Ideas in Dialogue: Self and No-Self. Farnham: Ashgate, pp. 29–45.
Körtvélyesi, T. (2011), “Ötböl áll az ember. Listákba szedett fogalmak a páli kánonban” (Man is made up of five. Concepts arranged in
lists in the Pali Canon). Vallástudományi Szemle, 7/1, 42–66.
Lang, K. C. (2010), “Candrakīrti’s critique of the Sāṁkhya concepts of puruṣa and prakṛti,” in Johannes Bronkhorst and Karin
Preisendanz (eds.), From Vasubandhu to Caitanya. Studies in Indian Philosophy and Its Textual History (Papers of the 12th World
Sanskrit Conference. Volume 10.1). Delhi: Motilal Banarsidass, pp. 53–59.
Larson, G. J. (1987), “Introduction to the philosophy of Sāṁkhya,” in Gerald James Larson and Ram Shankar Bhattacharya (eds.),
Sāṁkhya: A Dualist Tradition in Indian Philosophy (Encyclopedia of Indian Philosophies 4). Princeton: Princeton University Press, pp.
1–103.
McGinn, C. (1993), “Consciousness and cosmology: hyperdualism ventilated,” in Martin Davies and Glyn W. Humphreys (eds.),
Consciousness: Psychological and Philosophical Essays. Oxford: Blackwell, pp. 155–177.
Motegi, S. (1999), “The teachings of Pañcaśikha in Mokṣadharma.” Asiatische Studien/Études Asiatiques, 53/3, 513–535.
Ruzsa, F. (1997a), A klasszikus szánkhja filozófiája (The philosophy of classical Sāṁkhya). Budapest: Farkas Lőrinc Imre Könyvkiadó.
Ruzsa, F. (1997b), “The triple suffering. A note on the Sāṁkhya-kārikā,” paper read at the Xth World Sanskrit Conference, Bangalore.
https://www.academia.edu/472497/The_triple_suffering._A_note_on_the_Sa_khya-karika.
Ruzsa, F. (2003a), “Inference, reasoning and causality in the Sāṁkhya-kārikā.” Journal of Indian Philosophy, 31/1–3, 285–301.
Ruzsa, F. (2003b), “The types of suffering in the Mahāvyutpatti and the Pāli Canon.” Acta Orientalia Academiae Scientiarum Hungaricae,
56/1, 49–56.
Ruzsa, F. (2004), “The meaning of Āruṇi’s promise.” Indologica Taurinensia, 30, 229–235.
Ruzsa, F. (2007), “The fertile clash. The rise of philosophy in India,” in Csaba Dezső (ed.), Indian Languages and Texts through the Ages.
Essays of Hungarian Indologists in Honour of Prof. Csaba Töttössy. Delhi: Manohar, pp. 63–85.
Ruzsa, F. (2009), “A só-hasonlat az upaniṣadokban” (The simile of salt in the Upaniṣads), in Bangha Imre and Mészáros Tamás (eds.), A
Danubio ad usque Gangen. Klasszika-filológiai, indológiai és magyar kultúrtörténeti tanulmányok a 80 éves Vekerdi József
köszöntésére. Budapest: Typotex, pp. 201–211.
Ruzsa, F. (2010), “The authorlessness of the philosophical sūtras.” Acta Orientalia Academiae Scientiarum Hungaricae, 63/4, 427–442.
Ruzsa, F. (2013), “The influence of Dravidian on Indo-Aryan phonetics,” in Jared S. Klein and Kazuhiko Yoshida (eds.), Indic across the
Millennia: From the Rigveda to Modern Indo-Aryan. Bremen: Hempen Verlag, pp. 145–152.
Ruzsa, F. (2014), “Key Issues in Indian Philosophy” (unpublished DSc thesis). Budapest: Hungarian Academy of Sciences.
https://www.academia.edu/4531513/Key_issues_in_Indian_philosophy.
Ruzsa, F. (forthcoming), “Is the cosmic giant an Indo-European myth?” in Joel Brereton and Theodore Proferes (eds.), Papers of the Veda
section of the 13th World Sanskrit Conference. http://www.academia.edu/472488/Is_the_Cosmic_Giant_an_Indo-European_myth.
Schweizer, P. (1993), “Mind/consciousness dualism in Sāṅkhya-Yoga philosophy.” Philosophy and Phenomenological Research, 53/4, 845–
859.
Searle, J. R. (1980), “Minds, brains, and programs.” Behavioral and Brain Sciences, 3/3, 417–424.
Steinkellner, E. (1999), “The Ṣaṣṭitantra on perception. A collection of fragments,” Asiatische Studien /Études Asiatiques, 53/3, 667–677.
Szegedi, M. (2009), “Szándéketika és pszicho-kozmikus összeomlás: apokaliptika a buddhizmusban” (Intention based ethics and psycho-
cosmic collapse: apocalypse in Buddhism). Vallástudományi Szemle, 5/3, 32–53.
Tandon, S. N. (1995), A Re-appraisal of Patañjali’s Yoga-Sutras in the Light of the Buddha’s Teaching. Dhammagiri: Vipassana Research
Institute.
Watson, A. (2006), The Self’s Awareness of Itself. Bhaṭṭa Rāmakaṇṭhas’s Arguments against the Buddhist Doctrine of No-Self. Wien:
Publications of the De Nobili Research Library 32.
Watson, A. (2010), “Rāmakaṇṭha’s concept of unchanging cognition (nityajñāna). Influence from Buddhism, Sāṁkhya and Vedānta,” in
Johannes Bronkhorst and Karin Preisendanz (eds.), From Vasubandhu to Caitanya: Studies in Indian Philosophy and Its Textual History
(Papers of the 12th World Sanskrit Conference. Volume 10.1). Delhi: Motilal Banarsidass, pp. 79–120.
Wezler, A. (1985), “A note on Mahābhāṣya II 366.26: guṇasaṁdrāvo dravyam. (Studies on Mallavādin’s Dvādaśāranayacakra II),” in
Buddhism and Its Relation to Other Religions: Essays in Honour of Dr. Shozen Kumoi on His Seventieth Birthday. Kyoto: Heirakuji
Shoten, pp. 1–33.
Wezler, A., and Motegi, S. (eds.) (1998), Yuktidīpikā. The Most Significant Commentary of the Sāṁkhyakārikā. Stuttgart: Franz Steiner
Verlag.
 8
Śaiva Nondualism
Raffaele Torella

“Nondualism” being a correlative term, we should start by defining “dualism.” Not

dualism in general, however, but the particular dualism of the Tantric Śaiva tradition of
medieval Kashmir, starting probably around VII c., with which nondual Śaivism competes
and from which it ultimately derives. This is evident from the anti-dualist tension we can
find in the (comparatively) late Kaula scriptures, while no anti-nondualist tension can be
detected in the early Śaiva scriptures (Sanderson, 1992, pp. 307–308). Cf., among others,
statements like the one Kṣemarāja puts in the mouth of the supreme Śiva himself at the
outset of the Śivasūtra-vimarśinī: “Now that mankind is mostly adhering to dualist views
the secret tradition should not break.”1 Or in the maṅgalastotra of Kṣemarāja’s
Netratantra-uddyota: “In order to do service to the world he illustrates this Netra, a Śaiva
(tantra) in unity with the three states, burning the fuel of dualism, overflowing with the
nectar of supreme nondualism, eliminating the blindness of dualism-cum-nondualism.”2
As far as ontology is concerned, dualism is characterized by a definite position
regarding a few crucial points: in primis, relationship between the supreme God and
individual subjects, and between the supreme God and the manifested world (cf.
Sanderson, 1992). The individual souls (aṇu) are potentially endowed with the same
powers as the Lord (apart from the cosmogonic one), but in the saṃsāra condition these
powers are partly or totally inhibited by three kinds of impurity (mala), conceived of as
“physically” covering the individual soul. An analogously physical action, that of ritual, is
required for the soul to get rid of such impurities. Once the process of liberation is
achieved, an eternal distinction remains between the liberated souls (muktaśivas) and
Śiva. Śiva and the universe belong to two different orders. Even though the Lord pervades
the universe, the universe comes from a different source, usually called māyā; from it the
universe arises and finally dissolves. Their basic otherness is also demonstrated by the
fact that Śiva is often shown to entrust a minor god, Ananta, with the task of creating and
governing the material world. It is to be added that ontological dualism is significantly
flanked by epistemological dualism, less openly declared but clearly presupposed, which
hinges on the unbridgeable gap between knowing subject and knowable object. If we look
at full-fledged nondualism as expressed in the works of the four great masters
Somānanda, Utpaladeva, Abhinavagupta, and Kṣemarāja, we find complete coincidence
between Śiva and the individual soul; the material world is the free manifestation of Śiva
himself; māyā is not an autonomous reality, but a power of Śiva; the impurities (mala) are
by no means substantial realities, but erroneous attitudes of the subject based on his lack
of knowledge; the opposition knower-knowable is only provisional, and finally everything
shines as absolute I-ness.
As a worldview based on reason and revelation (yukti and āgama), Śaiva nondualism is
expected to have its roots in Śaiva scriptures. Indeed, one of the most popular division of
the Śaiva scriptures (cf. Tantrāloka [TĀ] I.18) presents three sets of texts, characterized
by dualism, dualism-cum-nondualism and nondualism, promulgated by Śiva, Rudra, and
Bhairava, respectively, in the number of ten, eighteen, and sixty-four. The scrutiny of what
is extant from the sixty-four Bhairavatantras risks disappointing the seeker for
unequivocal nondualist lines. As Sanderson (1992) and Törzsök (2014) have shown, even
the very occurrence of terms like advaita, advaya, and so on hardly refers to
straightforward affirmation of ontological and epistemological nondualism, but rather
concerns ritual practice. A pratice is termed “advaita” when it programmatically rejects
the mainstream Brahmanical opposition between what is in itself pure (normal, socially
acceptable, clear, etc.) and what is impure (abnormal, despised by the generality of Hindu
society, obscure, etc.). The impure is also often felt as more effective, quickly
transformative, although directly or potentially dangerous. We can envisage several pairs
of opposite terms, which can roughly be reduced to two: ritual in a pure environment
(pure ingredients, pure behaviors, defense of self-identity) versus ritual in an impure
environment (impure ingredients, illicit practices, possession); and ritual versus
knowledge. This is not to be taken in a too schematic way: within the same school,
different kind of attitudes can be detected (for one, see the knowledge temptation found
in a Saiddhāntic text such as the Mataṅgapārameśvara) or the unwillingness to get rid of
rituals in a champion of the primacy of knowledge over ritual, like Abhinavagupta.
The main question might also be formulated in the following terms: does advaitācāra
necessarily presuppose a noetic nondualist framework? This seems particularly true for
the Kaula tradition within nondualist tantras. Then, why are statements expressing
ontological and epistemological nondualism so rare even in the Kaula texts, or at least in
the early ones?3
It is an undoubted fact that in the economy of a Tantric scripture ritual holds by far the
most prominent position. The (mostly theoretical) division into four pādas (kriyā, caryā,
jñāna or vidyā, and yoga) has given the commentators the occasion to delve into the
relationship among these four domains. According to Nārāyaṇakaṇṭha’s commentary on
the Mṛgendra-tantra (yogapāda, pp. 1–2), the ideal order is vidyāpāda, kriyā-, yoga-, and
caryā-. Without vidyā, which is the means to describe the own nature of the individual
soul, the bonds and the Lord, which shows the extent of all paths, defines the rank of
Mantras, Lord of Mantras, and so on, it is impossible to bring about initiation successfully.
Therefore, the vidyāpāda, which provides knowledge of all this, should be expounded first
in a thoroughly detailed way in order to describe the own nature of the individual soul, the
bonds, and the Lord. Then, also the kriyāpāda has to be taught in order to present the
procedure of the different kinds of initiation along with their various limbs. Now, however,
even if they know both knowledge and ritual action, the various kinds of initiands
(sādhaka, putra, etc.) cannot bring about the performing of their respective activities as
they have been taught to them unless they know their own inner self by virtue of repeated
practice of perfect absorption, and so on. That is why the yogapāda follows along with the
description of the various limbs, such as prāṇāyāma, in order to impart knowledge about
yoga. And, Nārāyaṇakaṇṭha adds (caryāpāda, p. 207), without caryā, consisting of the
practice of what is prescribed and the avoidance of what is forbidden, the adept cannot
obtain the desired results without obstacles.
Rāmakaṇṭha in his commentary on Mataṅgapārameśvara (vidyāpāda, p. 567) gives a
partly different picture.

Caryā and yoga are not possible without knowledge, for neither yoga (in the sense of samādhi) nor
caryā, which has previously described, are possible if their respective aims are unknown.
Knowledge, however, is not sufficient to liberate the soul from the bonds, basic impurity, etc.—
which are substances—without the factual action performed by initiation... . Therefore, only
initiation is the (direct) cause of liberation, because knowledge and the other factors are seen to be
applied only within the context of initiation.4

The subordinate position of knowledge with respect to ritual may explain the absence of
explicit nondualistic statements in nondual scriptures. On the other hand, we can admit
that an anti-dualist attack, seen as an attack on the Brahmanical or semi-Brahmanical
establishment, is much more effective if it concerns praxis rather than mere knowledge.
The need for philosophical awareness comes only later, that is, when nondual Śaivism
engages in a śāstric discourse with adversaries (within and without the Tantric context).
This happens in concomitance with the emergence from the dimension of restricted
circles, and the attempt at establishing itself in the stratum of social normality, by
internalizing, or in any case circumscribing the specific differences. We can hypothesize
that the first step after negating the pure-impure opposition in ritual and behavioral
contexts is to elaborate on the basic equality of all by stressing the presence of Śiva in the
universe. For this purpose, stating that the all is pervaded by Śiva may prove not to be
sufficient as this is also upheld in Saiddhāntic circles. There is no impurity since Śiva “is”
the universe. This is precisely what will constitute the core of Somānanda’s teaching.
In explaining the sameness of all with the universal Śiva-nature, Abhinavagupta (TĀ
IV.274) refers to a passage on samatā from a comparatively early Trika scripture, the
Trikaśāsana, quoted in full by Jayaratha in the Viveka thereon: “There is sameness of all
beings and, by all means, of all conditions. There is sameness of all philosophies and, by
all means, of all substances. All the stages are the same, and also all the lineages, all the
goddesses, and by all means, all the classes.”5 The locus classicus for Somānanda’s
concept of universal samatā because of everything having the same Śiva-nature is
Śivadṛṣṭi [ŚD] I.48, to be read in the light of Utpaladeva’s comments:

Thus, it is firmly established that the Śiva-nature is the same for all entities. A differentiation in
them in terms of higher, lower, etc. may be maintained only by those who are ready to think
anything true.
Vṛtti: This is the meaning: Starting from Paramaśiva down to all objects, such as a jar, etc., the
Śiva-nature is the same, in the sense that it is neither more nor less, and it is definitely present in
everything with no exception (niyatā), since the nature of full consciousness is never exceeded.
Due to such experience of unity with the Śiva-nature, everything possesses a marvellous and
indefinable (kāpi) state. Thus, since everything has intimate unity with the Śiva-nature, we can
speak of things as differentiated into higher, lower, etc., of their having a pure or impure nature,
etc., only on account of our non-awareness of such intimate unity. This may take place in people
just owing to mere belief, that is, without sound reasons. In things there is no purity or impurity
whatsoever.6

The purity-impurity issue often emerges in Abhinavagupta’s works. In TĀ XV.163cd-164ab

he is confronted with a dubious statement found in the very core text of the TĀ, the
Mālinīvijayottara-tantra [MVU]: “One should not think that there may be anything that is
not purified by it [the arghapātra]. By it, everything is purified and what is impure
becomes pure.”7 After quoting this passage, Abhinavagupta points out that “impurity is to
be considered as such only from the point of view of limited souls and their teachings, for
everything resides in its own state either after a previous (impure) state or after a pure
state.”8 Later on in this chapter (417cd-418ab), Abhinavagupta takes up the issue again:
“The Lord in the MVU has not prescribed the purification of the two sacrificial ladles in
order to suggest that essence of purification is none other than full perception of the true
reality.”9
But the MVU also has passages in which the opposition purity/impurity is negated, like
the well-known XVIII.74: “Here there is neither purity nor impurity, nor deliberation about
what may be eaten or not, neither duality nor nonduality, and not even adoration of liṅga,
and so on.”10
The passage is commented on at length by Abhinavagupta in TĀ IV.212ff. “Even if we
consider things as existing externally,” Abhinavagupta continues, “purity and impurity are
not comparable to [existing objects, like] the blue colour. Purity and impurity are
qualifications pertaining to the knower depending on whether he perceives the object as
united with consciousness or not.”11
This point had been anticipated by Jayaratha in his Viveka on TĀ IV.221cd-222ab:

In fact, we do not deny the practical use in everyday life of the notion of purity and impurity, but
we point out that purity and impurity are not properties of the object, for it is the knower who
ascertains “this is pure, this is impure.” Had purity and impurity been properties of the object,
something impure could never become pure, and vice versa, because what is blue could not ever
become non-blue.12

This echoes Utpaladeva’s ŚDVṛtti, quoted above (na tu vastūnāṃśuddhir aśuddhir vā

kācit). Abhinavagupta’s position can be summarized by his own words: “Impure is what
has fallen away from consciousness: therefore everything stands pure if it has achieved
identity with consciousness.”13
Among the most quoted statements on the matter found in Trika scriptures stands
Vijñānabhairava 121 kiṃcijjñair yā smṛtāśuddhiḥsā śuddhiḥśaṃbhudarśane | na śucir hy
aśucis. The meaning of this verse, whose readings are much fluctuating both in
manuscripts and in quotations, I find most likely is: “What men of limited knowledge
traditionally14 consider as impurity, this in the Śaiva worldview is taken as purity. For
purity cannot become impurity [or vice versa].”15
We may close this topic by referring to the position of another important Trika Kaula
tantra, the Vīrāvalī, now lost. The passage is quoted in TĀ IV.242: “The life principle is
what sets in motion all entities; nothing exists that is destitute of such life principle.
Whatever is destitute of such life principle you should consider as ‘impure.’”16 Needless to
say, for Abhinavagupta the life principle (jīva) is the supreme light of consciousness. He
concludes: “Therefore what is not exceedingly distant from consciousness brings about
purity.”17
This may be compared to an analogous statement (in fact, many more could be quoted)
already found in an early nondualistic text, the Spanda-kārikā by Bhaṭṭa Kallaṭa (or, for
some, by Vasugupta), particularly referring to the domain of language: “There is no state
in words, meanings and mental elaborations that is not Śiva.”18 “No state,” says
Kṣemarāja in Spandanirṇaya p. 48, is meant to include the initial, medial, and final parts
of all these realities.
The coincidence of Śiva and the world might be taken in an “illusionistic” sense, as in
some Advaita Vedānta or Vijñānavada approaches. If having the nature of Śiva in a sense
“enhances” the reality of the world, in another it risks derealizing it, that is, flattening its
multifarious aspects and finally making it fade altogether. Realizing the ultimate Śiva
nature might lead to the very disappearance of the universe as such.
Since its very beginnings nondual Śaivism tackles this crucial issue. According to
Somānanda, it cannot be said that the universe is “imagined” as Śiva, or vice versa,
because the one is directly the other. Just as gold is not “imagined” as such neither in the
simple jewel of solid gold nor in the earring in which the craftsmanship is so refined as to
set aside, as it were, its nature of pure gold, so Śiva is “formed, arranged” as universe—in
the sense that he has become such and such, that is, freely presents himself in this form.
An original and very subtle treatment of the issue can be found in the ŚD, which is worth
quoting in full. The passage hinges on the distinction between kḷpti and kalpanā: Śiva is
not just conceived (kalpita) as having the form of the world, and vice versa, but he is
indeed (auto-)formed (kḷpta) as having the form of the world.

It is not mental construction that operates as regards this [universe] consisting of the Earth
principle, etc., since the [Śiva principle] is formed precisely in this manner. If something is
conceived of as different from what it is, then we can speak of mental construction. But is that
[Earth, etc.] conceived of differently from what it is? If [mental construction] concerns something
real,19 then mental construction is just a word [without content].
Vṛtti—We cannot say that, with respect to the multitude of entities that are perceived as having the
form of earth, etc., the fact of having Śiva as their own form is a mental construction. For in actual
fact it is the very Śiva principle that is formed having earth, etc., as its own form. If something that
does not possess a certain form is ascertained as possessing that form, as in the case of fantasy,
then this would be a [case of] mental construction, i.e. this would be something constructed
mentally. But, since everything is directly Śiva himself, the earth, etc., are in actual fact nothing
but Śiva, [so] will the earth, etc., be “mentally constructed” as being Śiva? The meaning [of the
rhetorical question] is: they are not! If we speak of mental construct with regard to [the cognition
of] the earth, etc. which are really so [i.e. of the nature of Śiva], then “mental construct” in the
present case is just a verbal expression.20

Utpaladeva’s Vṛtti adds some welcome clarifications:

On certain occasions, one can mentally construct the presence of real fire with respect to a fire
represented in a picture or to the red colour found in the Aśoka flower, etc. Śiva is formed having
the earth, etc. as his form; he has really assumed such a form, namely, he spontaneously presents
himself as such: this is the meaning. Mentally constructed would be said of somebody /something
that is only assumed arbitrarily to be such. There is a difference between the meaning of these two
words—namely, a) autonomously being by oneself formed in a certain way (kḷpti), i.e. one’s
becoming such and such, referring to somebody /something that is autonomously formed in that
form, and b) being mentally constructed (kalpanā) as being such and such—in the sense of merely
imagining as being such and such somebody/something that instead by himself/itself is not so.
Those who are not acquainted with such a difference are indeed to be worshipped: this is to be
taken in a derisory sense.21

The Śiva-nature embraces everything (ŚD VI.127 ab tato jñeyā śivatā sarvagocarā).
Somānanda further specifies that the world’s having the nature of Śiva involves that all
objects, like the jar, have the same powers as Śiva (icchā, jñāna, kriyā) and possess
sentiency (sacittvam). Somānanda denies that Śiva may be destituted of his powers even
when he is not active, that is, in his transcendent and quiescent state (śānta). Incidentally,
this proves that even insentient objects may not be without the powers of Śiva.

If you admit that the nature of Śiva exists in the quiescent state and in the gross state as well,
then, tell me, what is the state of quiescence? Does it not possess any reality for you? If it is real,
then there indeed is within it the meeting of such three powers.
Vṛtti—[Obj.:] If we admit—yatra is to be taken in the sense abhyupagama “admitting”—that the
Śiva-nature which is present in the quiescent state is to be considered the same as the Śiva-nature
in the gross state, i.e. when Śiva takes the form of the universe, then please tell me what is the
state of quiescence, i.e. the state in which Śiva is there [alone]? According to you, quiescence too is
a real entity, and a real entity is made of the fact of existing (sattāmayam), and the fact of existing,
 of being, is the fact of being the agent of the action “to be,” and agency, which has freedom as its
essence, belongs to consciousness, possessing the powers of will, etc. Therefore, every real entity
is Śiva.22

Being is, actually, being united with the manifestation of consciousness (cf. ŚD IV.29b
cidvyaktiyogitā; IV.7ab sarvabhāveṣu cidvyakteḥsthitaiva paramārthatā). Thus everything
is pervasive, incorporeal, and endowed with will, like consciousness (V.1). If things can be
efficient, it is because they “want” one particular action that is peculiar to them (V.16, 37).
And if they want it, they must also know it, in other words be conscious—first and
foremost, of themselves. All things are in all conditions knowing their own self (V.105ab
sarve bhāvāḥsvam ātmānaṃjānantaḥsarvataḥsthitāḥ). This dignifies all levels of reality,
including the surface level, made of human transactions and related verbal behavior, in a
word “vyavahāra.”

If the opponent asks why there is acceptation and rejection [of the doctrine maintaining the Śiva
nature of everything], [we reply] because he manifests himself in this way. Or, [because] everything
serves the course of ordinary life. [Obj.:] But the ordinary course of life has nothing to do with
actual reality. [Reply:] You should know that the own form [of anything] is related to actual reality.
[Obj.:] The ordinary course of life is never [related to actual reality]. [Reply:] [It is], since the Lord
manifests himself in this way inasmuch as he takes the form of the appearing of the unreal [, too].
Vṛtti—[Obj.:] How is it possible that there are upholders and opponents [of a certain doctrine]?
For there can be no different worldviews, given that there is only one Śiva. [Reply:] What has been
said earlier, namely that this is just meant to defeat those who maintain a different view [with
respect to the view that all things have the same Śiva nature], this too holds true because it is Śiva
himself who manifests himself in this way. Alternatively, we may say that everything, i.e. speaking
pro or contra the authoritative doctrine [i.e. maintaining the Śiva nature of everything], serves the
ordinary course of life. [Obj.:] But the ordinary course of life, being informed by non-perception of
identity with Śiva, has nothing to do with the real. [Reply:] On the contrary, what shines according
to its own form is indeed real, a reality which derives from its having [ultimately] Śiva as its own
form. [Obj.:] But the ordinary course of life is never real; on the contrary, it is just illusion. [Reply:]
Even the unreal has the Lord as its own form, since it is the Lord that manifests himself in the form
of the appearing of the unreal. Precisely owing to this, according to the principle expressed in the
Īśvarapratyabhijñā, even the unreal, inasmuch as it is being manifested, has consciousness as its
own form. With this specification only: something is called “unreal” when it is not manifested
externally.23

The absolute identity of Śiva and the universe, being the outcome of his free self-
expression, involves a reinterpretation of the role and status of māyā. The latter is seen by
the Saiddhāntic scriptures as an irreducible counterpart of Śiva, from which the universe
emanates and finally dissolves. This status is still conspicuously present in the MVU,
where māyā is defined in the following terms: “Māyā is one, pervasive, subtle, without
parts, receptacle of the universe, with neither beginning nor end, non-Śiva, sovereign,
imperishable” (I.26 sā caikā vyāpinī sūkṣmā niṣkalā jagato nidhiḥ| anādyantāśiveśānī
vyayahīnā ca kathyate ||). The reading aśivā (anādyantāśivā-ī°) is apparently the original,
also shared by other Śaiva scriptures and Saiddhāntic treatises; it is moreover
presupposed by the term with which Abhinavagupta glosses it in the TĀ, that is, jaḍā (cf. a
thorough analysis of this passage in Sanderson [1992, pp. 300–306; see also Vasudeva
[2004, pp. 182–184]). This does not prevent Abhinavagupta from choosing instead the
reading śivā while discussing the same passage in MVUVārttika, this being just one of the
many occasions in which he tries to reinterpret nondualistically the not few dualistic
statements found in the MVU. The otherness of māyā, according to Abhinavagupta and all
nondual Śaivism, is only apparent: “Māyā is the power of the god, inseparable from him,
the freedom to make differentiation appear. In fact, this [appearing of differentiation] is
caused by it” (TĀ IX.149cd-150ab māyā ca nāma devasya śaktir avyatirekiṇī ||
bhedāvabhāsasvātantryaṃtathāhi sa tayā kṛtaḥ|). This echoes Utpaladeva’s statement:
“By the power of māyā of Parameśvara, whose essence is light, the world—which consists
of his own self—is manifested as differentiated.”24
The role of māyā as material cause of the universe, stated by Saiddhāntic scriptures, is
rejected. As the central text of nondual Śaivism, Utpaladeva’s ĪPK says: “Indeed, the
Conscious Being, God, like the yogin, independently of material causes, in virtue of his
volition alone, renders externally manifest the multitude of objects that reside within it”
(Torella, 2002, p. 116).25 Not only does Śiva not need a material cause, but according to
ŚD III.80–82ab, he is the material cause himself as well as the efficient cause and the
noninherent cause.
Despite the strong emphasis on the reality of the world as self-expression of the Lord,
we may come across opposing statements about the status of the manifested world in the
scriptures and in the works of exegetes, but these need not concern us too much. As
Somānanda says:

There is no fault as regards the intention of the [scriptural] sentences expressing the
worthlessness of the universe, the aim of whose statements being that of arousing dispassion, etc.
A multiplicity of the consciousness principle cannot be admitted.
Vṛtti—There is no contradiction [in our view] with sentences—also uttered by Śiva—like “the
universe is unreal, similar to a bubble,” because there is no fault in them, expressing as they do the
goal of dispassion, of non-egoity, etc., inasmuch as their final aim is to prove that everything
possesses the nature of the one Śiva. Nor, on the other hand, does the fact of maintaining that
everything is Śiva involve a plurality of the consciousness principle.26

Just as the universe ultimately coincides with Śiva, so there is no real difference between
Śiva and the individual subjects (Spandanirṇaya p. 48 jīvaśivayor vāstavo na ko ’pi
bhedah). The individual I is no other than the I of Śiva. Conceiving of the supreme reality
as an absolute I, which is destined to become one of the central and most peculiar tenets
of nondual Śaivism, is due to Utpaladeva, whose ĪPK and commentaries constitute the
basis of the Pratyabhijñā philosophy, the very core of the philosophical dimension of
nondual Śaivism. However, there is no substantial difference between Utpaladeva’s I and
Somānanda’s dynamic Self-Śiva which underlies the whole universe and express himself
in it. Utpaladeva is the one who chose to use this word and concept regardless of the
negative associations generally attached to it in Indian thought, being aware of the fact
that the risk of a reification that has always weighed heavily on the word ātman was even
more concerning, and that this makes it less suitable for expressing the unpredictable
overflowing of the divine personality. Somānanda remarks: “Once an action has been
accomplished along with its result, immediately after, the will for another action arises,
the infiniteness of the powers of Śiva being the cause for this. These powers, which are
perennially present, flow according to their own being. Therefore, Śiva is one whose
nature is ‘flowing.’”27 The term “I” is implicitly aimed against the two conceptions that are
after all closest to the Pratyabhijñā and which it most aspires to differentiate itself from:
the consciousness devoid of a subject of the Vijñānavāda and the static ātman of the
Nyāya-Vaiśeṣika (or the ātman-brahman of the Vedānta). As Utpaladeva puts it in the
Ajaḍapramātṛsiddhi: “The I-ness is the resting of the light in itself; this resting in itself is
freedom in that dependence from any other reality is excluded, it is agency in the primary
sense, Lordship.”28
Utpaladeva’s ĪPK is very clear about the basic identity of the individual subject and
Śiva, but at the same time accounts for the difference that can be seen in ordinary reality,
and the way to become fully aware of such identity. The divine subject (pati)—whose body
is constituted, as it were, by the universe—is counterposed with the “beast” (paśu), the
fettered soul, in its various forms, depending on the impurities that characterize it. In the
conception outlined by Utpaladeva there are two components coming from different
sources: the hierarchy of subjects—which is a peculiar element of the Śaivasiddhānta, but
also included in Trika texts, like the MVU—and a version of the three impurities which,
though deriving from the analoguous doctrine of the Śaivasiddhānta, has an utterly
nondualistic qualification. The āṇava, māyīya, and kārma impurities completely lose their
original nature of “substances” that physically obstruct the self of the paśu from without,
and turn out to be erroneous attitudes of the individual consciousness.29 The āṇava
impurity, with its obliterating the one or the other of the components of subjectivity
(consciousness and freedom, bodha and svātantrya) determines that identity crisis onto
which the other two are grafted: the māyic one—which causes the I to see the world of
objects as separate from himself—and the kārmic one—which makes him consider his own
actions as the causes of the series of rebirths, miring him in the saṃsāra. As Somanānda
says, it is the very belief in the actual existence of bond and liberation that constitutes the
impurity (VII.cd na me bandho na me mokṣas tau malatvena saṃsthitau). The cause of all
three is the power of māyā, which has its roots in the very will of the Lord (ĪPV III.2.5 and
Vṛtti). The individual knower, variously contaminated by these three impurities, is then
distinguished according to four levels which he tends to identify with (void, vital breath,
mind, and body) and the conditions in which his experience of reality may take place
(waking, sleep and deep sleep, corresponding to direct perception, mental construct, and
partial or total suspension of all activity and cognition). The fourth state, in which duality
is overcome, corresponds gradually to the conditions of Vijñānākala, Mantra,
Mantreśvara, and Īśa: the state beyond the fourth is the one in which every trace of the
knowable is dissolved in the absolute I of Śiva in which the individual I finally merges.30
The aim of the “new and easy” way expounded by the Pratyabhijñā school is merely to
trigger an act of identification in the devout, which does not reveal anything new but only
rends the veils that hid the I from himself: a cognition is not created, but only the blur
that prevented its use, its entering into life, is instantly removed. In Abhinavagupta’s
words, commenting on the penultimate stanza of ĪPK:

There is a fascinating young girl, increasingly enamoured of a heroic character merely on hearing
his many qualities praised; beside herself with passion and unable to bear the pain of separation
from her beloved, she informs him of her state, sending him messages of love. But once she
encounters him [without knowing who he is] as he passes fleetingly by, the sight of him,
notwithstanding everything, does not penetrate her heart, because now that his qualities are not
manifest he seems to be a man like any other. But when, thanks to the words of a go-between, she
becomes aware of that man’s celebrated qualities, she instantly enters a state of fullness. In the
same way, the manifestation of the self, even though it shines constantly as Lord of all, does not
determine any state of fullness until one is aware of its qualities. When, however, through the
words of a master or by other means, awareness of the greatness of Maheśvara arises,
characterized by omnipotence, omniscience, etc., the state liberation-while-living made of absolute
fullness is immediately achieved.31

The centrality of knowledge in the path to liberation had already been emphasized by
Somānanda in the ŚD. It is true that the Śiva principle, whether it is known or not, does
not suffer any real obstruction (VII.1ff.). Fire, whether externally perceived or not, still
generates light, and gold, if it is not known, does not therefore become a stone. All this is
doubtless true; it is true that the gem of desires remains such even if it is not known, but
only if man knows it as such can he enjoy its effects (VII.4). The same may be affirmed
regarding individuals’ identification with Śiva. It has been said that even a fire that is not
known generates light, but so much more will be generated by a fire aptly arranged for
this purpose, such as a lamp in a house (VII.10cd ff.). Thus, a means must be taught
whereby the attained awareness of the nature of Śiva produces visible effects in the souls
(VII.12ab). This means is represented by logical argumentation, the scriptures, and the
teacher (VII.5).
All in all, Somānanda does not go beyond a powerful affirmation of the identity of Śiva
and the universe, the task of defining their relationship being assumed only later by his
disciple Utpaladeva, who starts precisely from where Somānanda had stopped. The latter
had not developed his own ideas about the ontological status of the manifested world, but
had only asserted, strongly and repeatedly, its reality (satyatā) and its having the nature
of Śiva (śivarūpatā). Things are “states” (avasthā, bhāva) of Śiva, and their emergence is
due solely to his will, brought about by nothing other than a natural overflowing of his
energies, whose characteristic feature is “joy” (āmoda) as well as “play” (krīḍā). In the
sixth chapter of the ŚD, Somānanda attacked those that in various ways claim that the
external world is unreal, especially various types of Vedāntins (VI.3ff.), who consider it as
an illusory manifestation (vivarta) of Brahman, caused by nescience, or the Vijñānavādins
(VI.33–34), who affirm the reality of consciousness but make unreal objects arise from it,
and, moreover, do not admit an agent subject of this consciousness, whereas for
Somānanda every action, and therefore also the action of knowing, is necessarily
dependent on an agent subject. Even those who claim the reality of the external object,
like the Buddhist bāhyavādins, are wrong because they do not admit a unifying principle
of reality, and thus make the passage from the moment of sensation to the moment of
mental elaboration impossible (IV.80cd ff.), and the same holds for the operation of apoha
through which they seek to elude the universal.32 The later construction of the ontological-
epistemological edifice of Pratyabhijñā by Utpaladeva has to be seen within the context of
his appointing precisely the Buddhists as the main adversaries.33 While for Somānanda
the Buddhists are opponents just like many others, they are given a special status in the
work of Utpaladeva, for whom they, admired and attacked in an equally strong way, are so
to speak the most intimate enemies. The criticism of their positions is to Utpaladeva of a
substantial help in building and refining the Pratyabhijñā philosophy. This also holds for
the model chosen for defining the relationship between Śiva and the world. Instead of
resorting to one or another Brahmanical model, Utpaladeva basically refers to the
Vijñānavāda doctrine explaining the emergence of the external world by the multiform
awakening of latent impressions within consciousness. Just as the ākāras of the
Vijñānavāda34 do not have any separate existence from the consciousness in/from which
they emerge, likewise for Utpaladeva the objects are nothing but “reflections,” or
“manifestations” (ābhāsas) in the mirror of supreme consciousness.35

The objects that are manifested in the present can be manifested as external only if they reside
within.
Vṛtti: Even in direct perception, however, the manifestation of objects as separate is admissible
only if they are absorbed in the cognizer.
... Seeing that ordinary worldy activity can be accomplished on the basis of such
“manifestations” alone, what sense is there in wanting to resort an external reality other [than
consciousness], which is not supported by reason?36

While the Buddhist model is clearly visible in Utpaladeva’s conception, the former, as
usually occurs in his philosophical strategy, acts only as a raw material to be aptly
modified and adapted to an utterly different worldview: thus, the impersonal
consciousness of the Vijñānavāda is substituted by the dynamic I-ness of Śiva, and,
consequently, the divine will takes the place of the mechanical emergence of the
vāsanās.37 This is shown in all evidence in the verse already quoted: “Indeed, the
Conscious Being, God, like the yogin, independently of material causes, in virtue of his
volition alone, renders externally manifest the multitude of objects that reside within it.”38
The multitude of objects (arthajāta) is described by Abhinavagupta (ĪPV I p. 184) as
ābhāsavaicitryarūpa “consisting of the multiformity of reflections.”
Utpaladeva presents these “reflections-manifestations” as having the nature of
universal (in fact, each one of them is connected with a word). They can appear in
isolation or aggregated around a dominant manifestation, and are provided with a single
or complex causal efficiency,39 on the basis of a compatibility that has its ultimate
foundation in the law of necessity established by the Lord. United among themselves, thus
becoming more and more particularized and finally combined with three manifestations
endowed with a special individualizing force—space, time, and form (these too derived
from Buddhism)—they constitute the world of everyday experience. It is easy to glimpse in
this conception elements drawn from the Vijñānavādins (consciousness as a receptacle of
everything that is gradually manifested), from the Vaiyākaraṇas and Vaiśeṣikas (the idea
of a hierarchy of universals) and more generally from the Buddhist pramāṇa tradition. In
this case, too, the word ābhāsa as a philosophical term—from which Utpaladeva’s doctrine
draws its most widespread denomination (ābhāsavāda)—was of course not invented by
him, but was commonly used in the Vedāntic and Buddhist schools. It also occurs in the
ŚD, but, even when it does not appear in a context where explicit reference is made to
opposing doctrines (various kinds of vivarta), its use seems to be merely sporadic and
casual, and in any case devoid of a precise technical connotation.
In one respect, ābhāsa is not distinguished from prakāśa (and related terms), and
sometimes the two terms seem interchangeable (cf. svābhāsa-svaprakāśa). In another
(namely, in its more technical use, as in the so-called ābhāsa theory), though their
essential unity of nature remains, ābhāsa is seen as a particle, an individualized and
extroverted form of the “great light,” “cut out” in it. This fragmentation of the light is
accompanied by, and also presupposes, an analogous descent of consciousness to the
state of fragmented subjectivity of the manifold individual subjects (the true subject is
avicchinnābhāsa; cf. ĪPV II p. 125); see in particular Abhinavagupta’s commentaries on
ĪPK II.3.1–2. The relationship between consciousness and ābhāsas is that between the
mirror and the reflected image, subtly analyzed in chapter III of the TĀ.40 The ontological
status of the ābhāsa is therefore a mixture of autonomy and heteronomy, without its basic
reality ever being called into question41 (ĪPK I.5.4 Vṛtti). Though Utpaladeva does not
explicitly repeat Somānanda’s extreme formulation (“the jar exists, knowing itself”), by
using a typical Vijñānavāda argument he underlies the necessarily common nature of
consciousness and its object: only that which is itself light, that is, sentience, can shine in
knowledge. “If it were not essentially light, the object would remain non-light as before
[the cognitive act]; and light is not differentiated [from the object]: being light constitutes
the very essence of the object.”42 The object cannot receive such “light” from outside:
light must already be the very self (ātman) of the object, its own form (svarūpabhūta; cf.
Vṛtti thereon). The being of the object consists in its becoming manifest:
prakāśamānatātmikā sattā (Vivṛti on ĪPK I.5.1; Torella, 2007c, p. 934). To say that
something is insentient is like saying it is inexistent (Ajaḍapramātṛsiddhi 13ac). Even
when an ābhāsa is viewed as external, as “this”—the “this” continues to have its
foundation in the I; it may also be said that reflective awareness of something in terms of
separation, of “this,” has become fully achieved only when it rests in its innermost being,
thus becoming the reflective awareness “I” (Ajaḍapramātṛsiddhi 15).43 It is the same light
of the self that is manifested as self and as other (Ajaḍapramātṛsiddhi 13cd). So light, in
its essence, is the knower itself: it is the contact with the knower’s light that, so to speak,
kindles the latent, inner luminous nature of the object. Thus, if it is true that both subject
and object are essentially light, we are not allowed to say that the light-knower is the
light-object, but only the other way round. To explicate this concept, Utpaladeva in the
Vivṛti (Torella, 2007c, p. 936) makes a rare exception to his usual dislike for quotations:
for the second time, he cites a passage from the Bhagavadgītā (now, VII.12d na tv
ahaṃteṣu te mayi “But I am not in them, [whereas] they are in me”).
From what has been said so far, we see that Utpaladeva envisages in consciousness/
Śiva a dual pole, prakāśa-vimarśa—the first understood as the motionless cognitive light
that constitutes the basic fabric, the founding structure of reality, of the “given”; the
second as the spark that causes this luminous structure to pulsate by introducing self-
awareness, dynamism, freedom of intervention, of self-assertion, thus expressing in
theoretical terms what is the nature of an unpredictable divine personality, like that of the
violent and loving Śiva handed down in the scriptures and with whom Utpaladeva
dialogues in his mystical hymns. The two polarities are not to be seen as separate
realities, but merely as two sides of a coin, like Śiva and Śakti: reflective awareness is the
very own nature (svabhāva) of light (ĪPK I.5.11; see above). Prakāśa forms, together with
a large group of synonyms or quasi-synonyms, a close-knit constellation of “luminous”
terms indicating the notions of being manifested, emerging from the dark, coming to
consciousness or, more in general, of being the object of knowledge and finally simply
“being,” whose use was already firmly established especially in Vedāntic and Buddhist
contexts (on the metaphor light/knowledge, see Watson [2014]); prakāśa and synonyms
frequently occur in the Vākyapadīya. Apart from isolated and uncertain cases in the Śaiva
scriptural tradition,44 vimarśa,45 in the pregnant sense Utpaladeva attributes to it, cannot
but derive from Bhartṛhari’s teaching, especially if we consider its link with light, on the
one hand, and the word, on the other. I am referring to the two very famous and most
quoted stanzas Vākyapadīya I.131–32, whose influence, though extending over the whole
structure of the Pratyabhijñā (and nondual Śaiva philosophy as a whole), we find
concentrated particularly on two closely connected aspects. One (ĪPK I.5.19) concerns the
only way deemed possible to account for a common fact in everyday experience, such as
the immediate and seemingly thoughtless action that still achieves its purpose—namely,
that of affirming the presence of a subtle reflective awareness even within the sensation
or movement captured at its most direct and undifferentiated moment. “Even at the
moment of direct perception there is a reflective awareness. How otherwise could one
account for such actions as running and so on, if they were thought of as being devoid of
determinate awareness?” (cf. Torella, 2002, p. 125).46 The other regards the two solemn
general formulations that define “vimarśa” as the essential nature of light and
indissolubly link consciousness, reflective awareness, freedom, and the supreme word:

The essential nature of light is reflective awareness; otherwise light, though “coloured” by objects,
would be similar to an insentient reality, such as the crystal and so on. (Cf. Torella, 2002, p. 118)
Consciousness has as its essential nature reflective awareness; it is the supreme Word that arises
freely. It is freedom in the absolute sense, the sovereignty of the supreme Self. (Cf. Torella, 2002, p.
120).47

The importance of Bhartṛhari in the structure of Śaiva nondualism may not be

undervalued. This may be surprising if we think how he had been heavily attacked by the
very father of the Pratyabhijñā, Utpaladeva’s guru Somānanda (cf. Torella, 2009). In order
to undermine the discontinuous universe of the Buddhists, Utpaladeva decides to avail
himself precisely of Bhartṛhari’s main doctrine, the language-imbued nature of
knowledge, which is meant to demolish one of the main foundation stones of the Buddhist
edifice, the unsurpassable gulf between the moment of sensation and that of conceptual
elaboration, representing, as it were, the very archetype of the Buddhist segmented
reality. The omnipervadence of language is the epistemological version of the
omnipervadence of Śiva, and at the same time calls for integration into the spiritually
dynamic Śaiva universe.
Utpaladeva, referring to an enigmatic statement in the Bhagavadgītā (XV.15b)̄, had
identified three powers (śakti) in the Lord: Cognition, Memory, and Exclusion. His aim is
to show that cognition, memory, and exclusion, which constitute the very basis of the
knowledge process in the human mind, are indirectly also a proof of the coinciding of the
individual subject with universal consciousness. None of these phenomena can be really
explained and their complex functioning accounted for satisfactorily in merely “mechanic”
terms, as first of all the Buddhists do.

On the contrary, such a world of practical experience will become possible if all the various
cognitions—direct perception, memory etc.—have only one intrinsic nature, being in essence
consciousness only. Of this [ātman], namely the Supreme Lord, whose essence is all, these
cognitions represent the various and multiform powers of knowledge, etc. The world of practical
experience would otherwise be impossible.48
... if there were no Maheśvara who contains within himself all the infinite forms, who is one,
whose essence is consciousness, possessing the powers of knowledge, memory and exclusion.49

As the starting point for a broad presentation of the position of Śaiva nondualism
regarding epistemology we may take Utpaladeva’s affirmation of the basic identity of the
I, consciousness and any cognitive activity.

Therefore the real nature of cognition would not be respected if it were presented in terms of
objectification “here is this cognition”; consciousness (bodha) in fact is illuminated only by itself
and is able to shine autonomously merely as I ... . The same consciousness is called cognition,
when it is turned outwards, towards objects, and is [apparently] differentiated because of them.
When instead it is turned inwards, then it is called the knower.50
For the knower consists only of the light of I, while cognition is nothing but the light of the I
turned towards objects, without any additional form of its own other than this.51

A valid cognitive process is based on the attainment of “conformity” between the

“apprehended object” (grāhya) part and the “apprehending subject (or cognition)”
(grāhaka) part. Utpaladeva’s discourse is based on full acceptation of the epistemological
scheme provided by Dignāga: the twofold aspect of cognition.52 The “apprehending
cognition” part assumes the form of the “apprehended object” part; the cognitive process
consists precisely in the conformity or likeness (sārūpya) between the two.53 In the
classical definition by Mookerjee (1935, p. 77), this doctrine, known as sākāravāda, holds
that “knowledge of external reality is made possible by virtue of the objective reality
leaving an impression of its likeness on the mirror of consciousness.” On the contrary, the
competing doctrine known as nirākāravāda maintains that “our consciousness is clear like
a clean slate and does not depart an inch from its intrinsic purity even when it apprehends
the external reality. Consciousness is an amorphous substance and remains so in all its
activities. It is like light and reveals the object with its form and qualities without
undergoing any morphological articulation in its constitution.” (p. 77)
The essential of Utpaladeva’s position is expressed in two terse stanzas of the ĪPK,
defining pramāṇa:

The means of knowledge is that thanks to which the object is situated within its own confines “this
thing, with these characteristics.” This means of knowledge is an ever freshly arising light related
to a subject. This light, whose essence is the inner reflective awareness of that which is thus
manifested, becomes—as regards the object without spatio-temporal differentiations, etc., and
expressed by a single name—knowledge (miti), [provided it is not] invalidated.54

Among the many aspects of this complex definition (for which I refer the reader to Torella
[2002, pp. 161–163]) one may be considered particularly significant: the non-
differentiation between the means of knowledge and its result (pramāṇa and pramā (miti
in the kārikā). Utpaladeva’s starting point is once again a Buddhist doctrine. The
distinction between pramāṇa and pramā—the Buddhist epistemologist says—is only the
outcome of the analytic consideration of a reality, cognition, which is in itself one. The two
terms which are thus foregrounded cannot in any case represent a relationship of cause
and effect—because this would require the actual otherness of the two terms—but at most
a relation of establisher-established (vyavasthāpya-vyavasthāpaka), with a division of roles
within the same reality. While up to this point Utpaladeva shares the Buddhist view, he
strongly departs from it regarding the concept of “function, activity” (vyāpāra) performed
by the elements occurring in cognition. Vyāpāra is denied by the Buddhists, who consider
every distinction on this basis completely imaginary (in the act of piercing with an arrow
the arrow may be attributed at will a variety of vyāparas, like that of kartṛ or karaṇa or
apadāna); furthermore, vyāpara conflicts with their doctrine of momentariness. Cognition,
Dharmakīrti concludes (Pramāṇavārttika III.308 ab), only “appears” to be endowed with a
function. Utpaladeva and Abhinavagupta’s reply that vyāpāra not only exists but
constitutes the very essence of pramā, and the non-differentiation of pramāṇa and pramā
hinges on this: vyāpāra is not a different reality from the subject that acts and from the
instrument that is set in action. All this has already been essentially contained in the
laconic pramātṛvyāpāraḥ that follows pramitiḥ in the Vṛtti. For his part, Abhinavagupta
goes on to say (ĪPV II pp. 69ff.) that this does not mean that pramāṇa and pramā are
simply two ways of saying the same thing; the cognitive light that is the essential nature
of both is turned toward the external object in the pramāṇa, whereas in the pramā it is
turned inward as pure determinate awareness, contracted due to the influence of the
object assumed in it, having as its essence the word.
But our practical reality is apparently made of concrete objects, belonging to a specific
time and space, having a particular form, not of single or multiple manifestations or
ensembles of manifestations. Utpaladeva is well aware that the primary aim of
philosophical reflection is to account for ordinary experience, and in ordinary experience
we are indeed confronted with what the Buddhists call a “unique particular” (svalakṣaṇa).
Thus, also in this case, Utpaladeva starts from a well-known Buddhist doctrine, but his
own elaboration of the concept of svalakṣaṇa takes a fairly different direction.55 For the
Buddhist epistemologist the starting point is the particular; perception grasps it in its
entirety but is also inexpressible and uncommunicable; many different ascertainments
(niścaya) may stem from this single perceptual content, each of them capturing a part of it
and connecting it with a word, which therefore denotes a certain sāmānya (universal) (or
rather the negation of what is other than that feature). For Utpaladeva, each pramāṇa
grasps an individual ābhāsa (which is a sāmānya), expressed by a determinate word,
depending on a determinate reflective awareness, or grasps (in the perception itself and
not in a later cognitive act) a group of ābhāsas coordinated by the Lord’s power of
necessity around a dominant ābhāsa, which allows the perception to remain unitary. The
group of ābhāsas taken in its totality ultimately corresponds to the Buddhist svalakṣaṇa.
The two conceptions are after all not so much opposed to each other: Utpaladeva’s
svalakṣaṇa clearly derives from the svalakṣaṇa of the Buddhists, but with a significantly
inverted perspective. The difference in the treatment of the important theme of
“vividness” or “dimness” of a cognitive act by the Buddhists and the Śaiva philosophers
depends on the different assumptions above (Torella, 2002, p. 147, n. 2; 2007b, pp. 546–
548, 556–561).
The individual subject can cognize, remember, and exclude only if it is conceived of as
inscribed within an eternal and, at the same time, dynamic universal I-ness, that is, Śiva.

Notes
1. p. 1 dvaitadarśanādhivāsitaprāye jīvaloke rahasyasampradāyo mā vicchedīti.
2. p. 1 yad dvaitendhanadāhi yac ca paramādvaitāmṛtenocchalat |
dvaitādvaitadṛgandhakāraharaṇaṃdhāmatrayaikātmakaṃśaivaṃnetram anugrahāya jagato
’mutraitad uddyotate || 4 ||
3. I might quote not more than two texts showing straightforward nondualistic statements, the
Ucchuṣmatantra and the Kālikākrama. “How is it possible, o Dear one,” says a stanza of the former
(quoted by Kṣemarāja in the Śivasūtravimarśinī p. 8; and the Svacchandoddyota, vol. II p. 55, ad
VII.249) “that these can be object of knowledge without being [at the same time] subject of
knowledge? Object of knowledge and subject of knowledge constitute a single reality. That is why
there is no impurity.” (My understanding of this verse considerably differs from Bisschop and
Griffith [2007, p. 4]) (yāvan na vedakā ete tāvan vedyāḥkathaṃpriye | vedakaṃvedyam ekaṃtu
tattvaṃnāsty aśucis tataḥ ||). Several verses from the lost Kālikākrama are quoted by Kṣemarāja
again in the Śivasūtravimarśinī. For example: “Knowledge shines in various forms, externally and
internally. Without knowledge there is no existence of object, therefore the world is made of
knowledge. Without knowledge things cannot become object of cognition. From this it ensues that
knowledge constitutes the essential nature of the object” (p. 118 tattadrūpatayā jñānaṃbahir
antaḥprakāṣate | jñānād ṛte nārthasattā jñānarūpaṃtato jagat || na hi jñānād ṛte bhāvāḥkenacid
viṣayīkṛtāḥ| jñānaṃtadātmatāṃprāptam etasmād avasīyate ||). The Kālikākrama, being quoted for
the first time by Kṣemarāja, is likely to be a post-Abhinava scripture. The date of the
Ucchuṣmatantra (on this problematic title, see Sanderson [2009, p. 194; Hatley, 2007, pp. 275–
 281) is unsettled.
4. This is meant to nuance the statement found in XXVI.53cd: mokṣo vātha catuṣṭayāt “Liberation
comes from the tetrad” (Vṛtti: “from all the four pādas taken together, not just by one of them”).
However, the Mataṅgapārameśvara (kriyāpāda, V.3cd-4ab) clearly says that a guru is expected to
be expert in all the four domains.
5. samatā sarvabhāvānāṃvṛttīnāṃcaiva sarvaśaḥ| samatā sarvadṛṣṭīnāṃdravyāṇāṃcaiva sarvaśaḥ||
bhūmikānāṃca sarvāsām ovallīnāṃtathaiva ca | samatā sarvadevīnāṃvarṇānāṃcaiva sarvaśaḥ||.
6. evaṃsarvapadārthānāṃsamaiva śivatā sthitā | parāparādibhedo ’tra śraddadhānair udāhṛtaḥ||
Utpaladeva’s Vṛtti: paramaśivāt prabhṛti ghaṭādyantānām api padārthānāṃsamaivānyūnānatiriktā
ca śivatā pūrṇacidrūpānatirekān niyatā sarveṣāṃtathā sāmarasyāsvādanāt kāpi sthitiḥsyād ity
arthaḥ| evaṃca sarvasya śivarūpasāmarasyāt tadakhyātimayaśuddhyaśuddhirūpaparāparādibhedo
bhāvānām uktaḥ| tatra śraddhāmātreṇopapattirahitena jantūnāṃ, na tu vastūnāṃśuddhir aśuddhir
vā kācit.
7. XVI.52 na cāsaṃśodhitaṃvastu kiñcid apy upakalpayet | tena śuddhaṃtu sarvaṃyad aśuddham
api tac chuci ||
8. aśuddhatā ca vijñeyā paśutacchāsanāśayāt || svatādavasthyāt pūrvasmād athavāpy upakalpitāt |
9. 417cd-418ab tattvasaṃdarśanān nānyat saṃskārasyāsti jīvitam || iti vaktuṃsruvādīśaḥśrīpūrve na
samaskarot.
10. nātra śuddhir na cāśuddhir na bhakṣyādivicāraṇam | na dvaitaṃnāpi cādvaitaṃliṅgapūjādikaṃna
ca ||
11. TĀ IV.244cd-245ab bahiḥsatsv api bhāveṣu śuddhyaśuddhī na nīlavat || pramātṛdharma
evāyaṃcidaikyānaikyavedanāt |
12. na hi vayaṃśuddhyaśuddhivyavahāram apahnumahe, kiṃtu “te vastudharmatayā na bhavataḥ” ity
ucyate, pramātā hi vyavasyati—idaṃśuddham idam aśuddham iti, vastudharmatve hi anayor
aśuddhaṃna kadācid api śuddhyet śuddham api vā nāśuddhaṃsyāt, na hi nīlam anīlam api kadācid
bhavet.
13. TĀ IV.240cd-241ab [. . .] aśuddhaṃsaṃvidaś cyutam || saṃvittādātmyam āpannaṃsarvaṃśuddham
ataḥsthitam |
14. Śivopādhyāya glosses: dharmaśāstrajñaiḥ.
15. See above n.12.
16. sarveṣāṃvāhako jīvo nāsti kiñcid ajīvakam | yat kiṃcij jīvarahitam aśuddhaṃtad vijānata ||
17. 243ab tasmād yat saṃvido nātidūre tac chuddhim āvahet. For a wider treatment of the
purity/impurity issue in Śaiva Tantrism, see Torella (2015).
18. II.4ab tasmāc chabdārthacintāsu na sāvasthā na yā śivaḥ |.
19. That is, if it presents something just as it really is.
20. III.82cd-83 na pṛthivyādike tasmin kalpanā saṃpravartate || tathātvenaiva kḷptatvāt tad atat [tad
atat em.; cf. Torella, 2014, p. 593] kalpitaṃkiṃ, satye nāmāstu kalpanā. Vṛtti—na ca
pṛthivyādirūpatayā pratīyamāne bhāvajāte śivarūpitā kalpitā bhavati, vastusthityaiva
pṛthivyādirūpeṇa (I delete vā; cf. Torella, 2014, p. 593) śivatattvasyaiva kḷptatvāt.
atadrūpaṃtadrūpaṃyadāvasīyate yathā manorājyādi, tadā tat kalpanā kalpyamānaṃbhavet. yāvatā
sākṣāc chiva eva sarvam iti vastusthityā pṛthivyādi śiva eva, śivatvena kiṃkalpitaṃbhavati na
bhavaty evety arthaḥ. atha satyataḥsati (satyataḥsati em.; cf. Torella, 2014, p. 593) tasminn eva
kalpaneti kathyate tadā vastutas tatra kalpaneti nāmakṛtiḥ.
21. p. 133 citragate vahnau, aśokapuṣpādau rakte vahnikalpanā śobhate kvacid avasare
pṛthivyādirūpeṇa śivaḥkḷptaḥsaṃpannaḥsvayaṃsthita ity arthaḥ. kalpitaḥ (kalpitaḥ em.; cf. Torella,
2014, p. 594) punaḥsvayaṃsa tathā kevalam iṣyate tena rūpeṇeti kḷptasya svayaṃkḷptes
tathāsaṃpatteḥ, kalpanāyāś ca svayam atathābhūtasya tathā saṃbhāvanāmātrarūpāyā ye
śabdārthatve na bhedarūḍhās te vandyā ity upahāsaḥ.
22. III.55cd-56 śānte śivatvaṃsthūle ’pi śivatvaṃyatra varṇitam || tatra kā śāntatā brūhi śānte
kiṃvastutā na te | vastutā cet tathābhūtaśaktitritayasaṃgamaḥ. Vṛtti p. 119 yatrābhyupagame
śānte ’pi śivatvaṃsthūle ’pi jagadrūpatve śivatvam eva varṇitaṃtatra śāntatā śivarūpatā kā syāt,
brūhy etat. śāntam api hi tava vastu, vastu ca sattāmayaṃsattā bhavattā bhavanakartṛtā,
svātantryātmakaṃca kartṛtvaṃcidrūpasyecchādiśaktimata iti sarvaṃvastu śivaḥ.
23. III.76cd-78ab vāditvaprativāditve kasmāc cet tasya tatsthiteḥ|| vyavahārāya vā sarvaṃvyavahāro
na vastugaḥ| svarūpaṃvastugaṃviddhi vyavahāro na jātucit || tatheśvaravyavasthānād
avastvābhāsarūpataḥ. Vṛtti—tadaikye darśanabhedābhāvād vāditvaṃprativāditvaṃca kutaḥ, yad
uktaṃvivādihananāya iti, tad etad api tasyaiva tathā sthiteḥ. lokavyavahārāya vā
sarvaṃśāstravādiprativādyādi. lokavyavahāraś ca tadabhedākhyātimayo na vastugataḥ. yat
punaḥsvarūpeṇa prakāśate tac chivarūpavastvātmakam eva, vyavahāras tu na kadācit vāstavo ’pi
tu bhrama eva, avastv api īśvarātmaivāvastvābhāsarūpeṇa tasyāvasthānāt. ata
eveśvarapratyabhijñoktanītyā tad avastv api prakāśamānaṃcidrūpam eva, kevalaṃbāhye
prakāśanābhāvād avastūcyate.
24. ĪPK I.5.17 Vṛtti prakāśātmanaḥparameśvarasya māyāśaktyā
svātmarūpaṃviśvaṃbhedenābhāsyate.
25. I.5.7 cidātmaiva hi devo ’ntaḥsthitam icchāvaśād bahiḥ| yogīva nirupādānam
arthajātaṃprakāśayet ||
26. III.95cd-96ab viśvatucchatvavākyānāṃvairāgyādyarthavādinām || tātparyeṇa na doṣo ’sti
 nānācittvaṃna kalpate. Vṛtti—viśvam asatyaṃbudbudopamam iti śivoktair api vākyair na virodhaḥ,
yata eṣāṃvairāgyanirātmatādiprayojanābhidhānavatām ekaśivātmatāsamarthana- paratvena na
doṣo ’sti, na ca sarvaśivatve nānācittvaprasaṅgaḥ.
27. III.92cd-94ab saphalāyāṃsamāptāyāṃkriyāyāṃsamanantaram || kriyāntarecchāsaṃbhūtis
tannimittam anantatā | yato ’sti śivaśaktīnāṃtāś ca nityam avasthitāḥ|| saranty eva svabhāvena tat
saratprakṛtiḥśivaḥ.
28. 22cd-23 prakāśasyātmaviśrāntir ahaṃbhāvo hi kīrtitaḥ|| uktā saiva ca
viśrāntiḥsarvāpekṣānirodhataḥ | svātantryam atha kartṛtvaṃmukhyam īśvaratāpi ca || See also
Utpaladeva’s criticism of the inert self (of the Nyāya, etc.) in an extant passage of his IPVivṛti:
“However, mere acknowledgement that there is a single self (with respect to these two cognitions)
does not amount to also affirming its (necessary) sovereignty (aiśvaryam), for such a self, if
conceived of in terms of detached idleness (audāsīnyāt), cannot act as the possessor of the powers
of perception, memory, etc. (ananubhavasmaraṇādiśaktimattvāt)” (Torella, 2012, p. 296).
29. On āṇava as nescience, see MVU 1.23cd malam ajñānam icchanti; Śivasūtra I.2
(a)jñānaṃbandhaḥ.
30. On the alchemic metaphor used by Abhinavagupta (and Utpaladeva), see Torella [2002, pp. XXXII–
XXXIV). In the relevant passages, both the edited texts of ĪPV and ĪPVV need significant
emendations.
31. ĪPV II pp. 274–275 yadā nāyakaguṇasaṃśravaṇapravṛddhānurāgā kāminī taddarśanam eva
paramupādeyam ākāṅkṣantī divāniśam avaśahṛdayā devatopayācitāni dūtīsaṃpreṣaṇāni
madalekhadvārakātmāvasthānivedanāni kurvāṇā virahakṣāmībhavadgātralatikā tiṣṭhati, tadā
tadupayācitavaśāt aśaṅkitam eva savidhavartini priyatame ’valokite tais tair
utkarṣaviśeṣaiḥparāmarśapadavīm agacchadbhir janasādhāraṇatām āpādite saṃpannam api
priyatamāvalokanaṃna hṛdayaṃpūrṇīkaroti; tathā svātmani viśveśvare satataṃnirbhāsamāne ’pi
tannirbhāsanaṃna hṛḍayasya pūrṇatām ādhatte; yataḥso ’py ātmā
viśvajñatvakartṛtvādyapratihatasvaśaktilakṣaṇapāramaiśvaryotkarṣayogena na parāmṛṣṭaḥ—iti
bhāsamānaghaṭāditulyavṛttānto jātaḥ. yadā tu dūtīvacanād vā tallakṣaṇābhijñānād vopāyāntarād
vā tānutkarṣān hṛdayaṅgamīkaraṇenāmṛśati, tadā tat kṣaṇam adbhutaphullanyāyenaiva tāvat kām
api pūrṇatām abhyeti [...]; tadvad ātmani guruvacanāj jñānakriyālakṣaṇaśaktyabhijñānāder vā yadā
pāramaiśvaryotkarṣahṛdayaṅgamībhāvo jāyate, tadā tatkṣaṇam eva pūrṇatātmikā jīvanmuktiḥ.
32. These themes will be masterfully developed by Utpaladeva, especially in his ĪPVivṛti (cf. Torella,
2007a).
33. “Through this subtle play of a declared basic disagreement with the doctrines of Buddhist
philosophers, a limited acceptance and purely instrumental (or thought to be such) use of them,
the masters of the Pratyabhijñā end up being somehow drawn into their orbit. The architecture of
the Pratyabhijñā feels the effect of this. The very fact that many problems are posed, more or less
unwittingly, in Buddhist terms to a certain extent prefigures their development and reduces
possible alternatives as regards solutions.” Torella (1992, p. 329).
34. On ākāra, see recently Kellner (2014) and McClintock (2014).
35. The term ābhāsa can be variously translated as “(luminous) manifestation, reflection, appearance,
image.”
36. ĪPK I.5.1 vartamānāvabhāsānāṃbhāvānām avabhāsanam | antaḥsthitavatām eva ghaṭate bahir
ātmanā—Vṛtti pratyakṣe ’pi yāvad arthānāṃbhedenāvabhāsaḥpramātrantarlīnānām eva
satāṃyuktaḥ (cf. Torella, 2002, p. 111). I.5.6 syād etad avabhāseṣu teṣv evāvasite sati | vyavahāre
kim anyena bāhyenānupapattinā || (cf. Torella, 2002, p. 114).
37. See above Somānanda’s analogous statement. On this point, see Ratié (2010); cf. also Ratié
(2014).
38. See also I.6.7 “Thus also in the course of ordinary reality the Lord, entering the body, etc.,
renders externally manifest by his volition the multitude of objects that shine within him” (tad
evaṃvyavahāre ’pi prabhur dehādim āviśan | bhāntam evāntar arthaugham icchayā bhāsayed
bahiḥ ||). Cf. Torella (2002, p. 133).
39. This causal efficiency is not intrinsic to the object, and consequently may not act as the proof of
its truth-reality, as in the Buddhist doctrine, but it is just a particular ābhāsa which may or may not
be associated with the other ābhāsas constituting the object, without this affecting their already
established intrinsic reality. The same holds for externality (see below).
40. See also Mahārthamañjarī-parimala, p. 147.
41. Nondual Śaivism, and above all Pratyabhijñā philosophy, keeps the concepts of reality (sattā) and
externality (bāhyatā) rigourously distinct. As ĪPK I.8.5 says, “Existing externally is to be considered
an accessory condition (upādhi) and not the very essence of the manifestations of being and non-
being. These, therefore, insofar as they are inner manifestations, always exist”
(bhāvābhāvāvabhāsānāṃbāhyatopādhir iṣyate | nātmā sattā tatas teṣām āntarāṇāṃsatāṃsadā ||)
(cf. Torella, 2002, p. 148). And again (ĪPK I.8.7): “Insofar as they are essentially constituted by
consciousness the manifestations permanently reside internally; insofar as they are manifested as
external owing to the power of māyā, they also exist externally” (cinmayatve ’vabhāsānām
antareva antar eva sthitiḥsadā | māyayā bhāsamānānāṃbāhyatvād bahir apy asau ||) (cf. Torella,
2002, p. 149).
42. ĪPK I.5.2 prāg ivārtho ’prakāśaḥsyāt prakāśātmatayā vinā | na ca prakāśo bhinnaḥsyād
 ātmārthasya prakāśatā || (cf. Torella, 2002, p. 111). The Vṛtti thereon glosses “light” (prakāśa) with
“cognizer” (pramātṛ). See also passages from Vāmanadatta’s Saṃvitprakāśa, on which
Utpaladeva’s influence is evident (e.g., I.10–12 yathāgninā samāviṣṭaṃsarvaṃtadrūpam īkṣyate |
tathā jñānasamāviṣṭaṃsarvaṃtadrūpam īkṣyatām || pramātrapekṣabhāveṣu na hy
avasthāvakalpate | yatas tataḥprakāśantāṃsvayam eva tadātmanā ||
tvadātmakatvaṃbhāvānāṃvivadanti na kecana | yat prakāśyadaśāṃyātā nāprakāśaḥprakāśyate ||).
43. According to Vāmanadatta’s Ātmasaptati, the very concept of viṣaya is groundless unless we
admit the ultimate identification of the object with consciousness: II.8–9 svīkāro viṣayīkāraḥsa
tatrodghoṣyate budhaiḥ| yad anyatra prasiddhaṃtat saṃvidaḥkim apohyate || saṃvidā svīkṛtaṃyac
ca na tad viṣayasaṃjñitam | yatsvīkṛtaṃtadātmaiva viṣayoktiḥkva tiṣṭhatām ||.
44. See also ŚD II.83d, 84c.
45. On this term and its synonyms, cf. Torella (2002, p. XXIV, n. 32).
46. I.V.19 sākṣātkārakṣaṇe ’py asti vimarśaḥkatham anyathā | dhāvanādy upapadyeta
pratisaṃdhānavarjitam ||.
47. I.V.11 svabhāvam avabhāsasya vimarśaṃvidur anyathā | prakāśo ’rthoparakto ’pi
sphaṭikādijaḍopamaḥ ||. I.V.13 citiḥpratyavamarśātmā parā vāk svarasoditā | svātantryam etan
mukhyaṃtad aiśvaryaṃparamātmanaḥ
48. Vivṛti introducing ĪPK I.3.7 (cf. Torella, 2007a, p. 477) yujyeta punar etādṛśo vyavahāro yady eṣām
anubhavasmṛtyādīnāṃcinmātrasāratvenaika evātmā syād yasyaiva viśvātmanaḥparameśvarasyaitā
jñānādikā vicitrāḥśaktayo ’nyathā tu na syād iti.
49. ĪPK I.3.7 na ced antaḥkṛtānantaviśvarūpo maheśvaraḥ| syād ekaś cidvapur
jñānasmṛtyapohanaśaktimān ||.
50. Vivṛti on I.3.6 (cf. Torella, 1988, pp. 146–147) tad idaṃjñānam iti pratītau
svaprakāśaikarūpaṃjñānaṃna prakāśitaṃsyāt, bodho hi svaprakāśaikarūpo ’ham ity eva
prakāśārha aham ity asyaiva prakāśavāditvāt [?]​, sa eva hi yadā tv arthonmukhas tadārthabhedād
bhidyamāna iva jñānavyapadeśya | yadā tv antarmukhatayā vyavasthitas tadā pramātā kathyate.
Then Utpaladeva remarks that the Buddhists say the same thing except that they consider the
permanence of the subject as being uniquely the product of a wrong superimposition brought
about by discursive thought.
51. Vivṛti on I.3.6 ahaṃprakāśamātrarūpo hi pramātā, viṣayonmukhāhaṃprakāśamātrarūpaṃtu
jñānam nādhikyarūpam (cf. Torella, 1988, p. 148).
52. Pramāṇasamuccaya-vṛtti p. 4 (ad I.9a) “Cognition arises having two manifestations: it contains the
manifestation of itself and that of the object. The self-awareness of both manifestations constitutes
the result [of cognition].” (dvyābhāsaṃhi jñānam utpadyate svābhāsaṃviṣayābhāsaṃca.
tasyobhayābhāsasya yat svasamvedanaṃtat phalam). On this crucial passage, see recently Kellner
(2011).
53. Utpaladeva adopts the conception, also upheld by Dignāga and Dharmakīrti—and, more generally,
by Sautrāntikas and Vijñānavādins (along with Sāṃkhya and Vedānta)—known as sākāravāda. The
nirākāravāda is followed by Vaibhāṣikas (and Nyāya-Vaiśeṣikas, Mīmāṃsakas and Jaina); cf
Kajiyama (1989). Expectedly, the dualist Śaivasiddhānta is also nirākāravādin.
54. II.2.1–2 (cf. Torella, 2002, p. 161) idam etādṛg ity evaṃyadvaśād vyavatiṣṭhate | vastu
pramāṇaṃtat so ’pi svābhāso ’bhinavodayaḥ|| so ’ntastathāvimarśātmā deśakālādyabhedini |
ekābhidhānaviṣaye mitir vastuny abādhitā ||.
55. For a detailed discussion, see Torella (1992, pp. 332–336).

Bibliography
Texts
Abhinavagupta, Īśvarapratyabhijñāvimarśinī, edited by Mukund Rām Shastri, vols. I–II, KSTS XXII
XXXIII, Bombay, 1918–21.
Abhinavagupta, Īśvarapratyabhijñāvivṛtivimarśinī, edited by Madhusudan Kaul Shastri, vols. I–III,
KSTS LX LXII LXV, Bombay, 1938–43.
Abhinavagupta, Tantrāloka with Commentary by Rājānaka Jayaratha, edited with notes by Madhusudan
Kaul Shastri, vols. I–XII, KSTS XXIII, XXVIII, XXX, XXXVI, XXXV, XXIX, XLI, XLVII,, LIX,, LII, LVII,
LVIII, Allahabad-Srinagar-Bombay, 1918–38.
Bhartṛhari, Vākyapadīya (mūlakārikās), Bhartṛharis Vākyapadīya, edited by W. Rau, Abhandlungen für
die Kunde des Morgenlandes, 42, Wiesbaden, 1977.
Dharmakīrti, Pramāṇavārttika with the Commentary “Vṛtti” of Acharya Manorathanandin, critically
edited by Swami Dwarikadas Shastri, Varanasi, 1968.
Dharmakīrti, Pramāṇavārttika I: Pramāṇavārttikam. The First Chapter with the Autocommentary, text
and critical notes by R. Gnoli, Serie Orientale Roma XXIII, IsMEO, Roma, 1960.
Dignāga, Pramāṇasamuccaya, see Steinkellner (2005) and Jinendrabuddhi, Pramāṇasamuccayaṭīkā.
Jinendrabuddhi, Viśālāmalavatī Pramāṇasamuccayaṭīkā, chapter I. Part I: Critical edition, by E.
Steinkellner, H. Krasser, H. Lasic, China Tibetology Research Centre/Austrian Academy of Sciences,
Beijing/Vienna, 2005.
Kṣemarāja, Śivasūtravimarśinī, edited by J. C. Chatterji, KSTS I, 1911.
Kṣemarāja, Spandanirṇaya, edited by Madhusudan Kaul Shastri, KSTS XLII, Srinagar, 1925.
Maheśvarānanda, Mahārthamañjarī with the Autocommentary Parimala, edited by Pt. V. V. Dvivedi,
Yogatantra-Ratnamālā 5, Varanasi, 1972.
Mālinīvijayottaratantram, edited by Pt. Madhusūdan Kaul Shastri, KSTS XXXVII, Bombay, 1922. See
also Vasudeva (2005).
Mataṅgapārameśvarāgama (Kriyāpāda, Yogapāda et Caryāpāda), avec le commentaire de Bhaṭṭa
Rāmakaṇṭha, édition critique par N. R. Bhatt, Publications de l’Institut Français d’Indologie No. 65,
Pondichéry, 1982.
Mataṅgapārameśvarāgama (Vidyāpāda), avec le commentaire de Bhaṭṭa Rāmakaṇṭha, édition critique
par N. R. Bhatt, Publications de l’Institut Français d’Indologie No. 56, Pondichéry, 1977.
Mṛgendratantra (vidyāpāda and yogapāda) with commentary of Nārāyaṇakaṇṭha, edited by
Madhusudan Kaul Shastri, KSTS L, Bombay, 1930.
Mṛgendra-āgama (kriyāpāda et caryāpāda), avec le commentaire de Bhaṭṭa Rāmakaṇṭha, édition
critique par N. R. Bhatt, Publications de l’Institut Français d’Indologie No. 23, Pondichéry, 1962.
Somānanda, Śivadṛṣṭi with the vṛtti by Utpaladeva, edited by Pandit Madhusudan Kaul Shastri, KSTS
LIV, Srinagar, 1934.
Utpaladeva, Īśvarapratyabhijñākārikā with svavṛtti (see Torella [2002]).
Utpaladeva, Ajaḍapramātṛsiddhi. In: Siddhitrayī, edited by Pandit Madhusudan Kaul Shastri, KSTS LIV,
Srinagar, 1921.
Vāmanadatta, Saṃvitprakāśa and Ātmasaptati, critical edition and annotated translation by Raffaele
Torella (forthcoming).
Vijñānabhairava with the Commentary Partly by Kṣemarāja and Partly by Śivopādhyāya, edited by
Mukunda Ram Shastri, KSTS VIII; with Commentary Kaumudī by Ānanda Bhaṭṭa, KSTS IX, Bombay,
1918.

Translations and studies

Bisschop P., and Griffith, A. (2007), “The practice involving the Ucchuṣmas (Atharvavedapariśiṣṭa 36).”
Studien zur Indologie und Iranistik, 24, 1–46.
Hatley, Sh. (2007), The Brahmayāmalatantra and Early Śaiva Cult of the Yoginīs. A dissertation in
religious studies presented to the Faculties of the University of Pennsylvania in partial fulfilment of
the requirements for the degree of doctor of philosophy (unpublished).
Kajiyama, Y. (1989), “Controversy between the sākāra- and nirākāravādins of the yogācāra school:
some materials,” in Id., Studies in Buddhist Philosophy (selected papers). Kyoto, pp. 429–418.
Kellner, B. (2011), “Self-awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and vṛtti. A close
reading.” Journal of Indian Philosophy, 38, 203–231.
Kellner, B. (2014), “Changing frames in Buddhist thought: the concept of ākāra in Abhidharma and in
Buddhist epistemological analysis.” Journal of Indian Philosophy, 42, 275–295.
McClintock, S. (2014), “Kamalaśīla on the nature of phenomenal content (ākāra) in cognition: a close
reading of TSP ad TS 3626 and related passages.” Journal of Indian Philosophy, 42, 327–337.
Mookerjee, S. (1935), The Buddhist Philosophy of Universal Flux: An Exposition of the Philosophy of
Critical Realism as Expounded by the School of Dignāga, University of Calcutta, Calcutta (repr.
Delhi: Motilal Banarsidass,. 1980).
Ratié, I. (2010), “The dreamer and the yogin: on the relationship between Buddhist and Śaiva
idealisms.” Bulletin of the School of Oriental and African Studies, 73 (3), 437–478.
Ratié, I. (2014), “On the distinction between epistemic and metaphysical Buddhist idealisms: a Śaiva
perspective.” Journal of Indian Philosophy, 42, 353–375.
Sanderson, A. (1992), “The doctrine of the Mālinīvijayottaratantra,” In T. Goudriaan (ed.), Ritual and
Speculation in Early Tantrism: Studies in Honor of André Padoux. Albany: SUNY Press, pp. 281–312.
Steinkellner, E. (2005), “Dignāga’s Pramāṇasamuccaya, Chapter I. A hypothetical reconstruction ...”
www.oeaw.ac.at/ias/Mat/dignaga_PS_1.pdf.
Torella, R. (1988), “A fragment of Utpaladeva’s Īśvarapratyabhijñā-vivṛti.” Philosophy East and West,
38, 137–174.
Torella, R. (1992), “The Pratyabhijñā and the logical-epistemological school of Buddhism,” in T.
Goudriaan (ed.), Ritual and Speculation in Early Tantrism: Studies in Honor of André Padoux. Albany:
SUNY Press, pp. 327–345.
Torella, R. (2002), The Īśvarapratyabhijñākārikā of Utpaladeva with the Author’s Vṛtti, critical edition
and annotated translation, Motilal Banarsidass, Delhi (I ed. Serie Orientale Roma LXXI, IsMEO,
Roma, 1994).
Torella, R. (2007a), “Studies in Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part I. Apoha and anupalabdhi
in a Śaiva garb,” in K. Preisendanz (ed.), Expanding and Merging Horizons: Contributions to South
Asian and Cross-Cultural Studies in Commemoration of Wilhelm Halbfass. Vienna, pp. 473–490.
Torella, R. (2007b), “Studies in Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part II. What is memory?,” in
Indica et Tibetica. Festschrift für Michael Hahn zum 65. Geburtstag von Freunden und Schülern
überreicht, herausgegeben von Konrad Klaus und Jens-Uwe Hartmann, Wiener Studien zur
Tibetologie und Buddhismuskunde, Heft 76, Wien, pp. 539–563.
Torella, R. (2007c), “Studies in Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part III. Can a cognition become
the object of another cognition?,” in D. Goodall and A. Padoux (eds.), Mélanges tantriques à la
mémoire d’Hélène Brunner. Pondicherry, pp. 475–484.
Torella, R. (2009), “From an adversary to the main ally: the place of Bhartṛhari in the Kashmirian
Śaivādvaita,” in M. Chaturvedi (ed.), Bhartṛhari: Language, Thought and Reality. Delhi, pp. 343–354.
Torella, R. (2012), “Studies in Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part V. Self-awareness and yogic
perception.” in F. Voegeli et al. (eds.), Devadattīyam. Johannes Bronkhorst Felicitation Volume.
Worlds of South and Inner Asia 5. Berne: Peter Lang, pp. 275–300.
Torella, R. (2014), “Notes on the Śivadṛṣṭi by Somānanda and its commentary.” Journal of Indian
Philosophy, 42, 551–601.
Torella, R. (2015), “Purity and impurity in non-dualistic Śaiva Tantrism”, in Proceedings of the
International Conference “Religions: Fields of Research, Method and Perspectives,” Studia
Religiologica, 48 (2), 101–115.
Törzsök, J. (2014), “Nondualism in early Śākta Tantras: transgressive rites and their ontological
justification in a historical perspective.” Journal of Indian Philosophy, 42, 195–223.
Vasudeva, S. (2004), The Yoga of the Mālinīvijayottaratantra, Chapters 1–4, 7, 11–17. Critical Edition,
Translation and Notes. Pondicherry: Institut Français de Pondichéry-Ecole Française d’Extrême-
Orient.
Watson, A. (2014). “Light as an analogy for cognition in Buddhist idealism (Vijñānavāda).” Journal of
Indian Philosophy, 42, 401–421.
 9
An Indian Debate on Optical Reflections and Its Metaphysical
Implications
Śaiva Nondualism and the Mirror of Consciousness*
Isabelle Ratié

Mirrors and other reflecting entities are banal yet puzzling objects: endowed with the
extraordinary ability to make us perceive what we would otherwise remain unable to see
(such as our own face), they also offer a somewhat disturbing vision of the world, showing
things where they are not (our face in the mirror) or providing them with properties that
they cannot possess—by virtue of their strange power, tiny elephants appear in glittering
jewels and multiple suns shine on a rough water. As Karin Preisendanz once noted, if
Indian optics “did not emerge as a distinct science or pre-science in India ..., optical
theories were developed, problems realized, and solutions attempted,” but always within
the larger frame of philosophical speculations,1 and optical reflections (pratibimba in
Sanskrit) are no exception: in classical India, the question of how they occur was
primarily discussed in connection with that of their problematic ontological status. Are
reflections as unsubstantial as mirages, or do they hold some reality, and if so, in what
way do they exist? Moreover, this ontological discussion itself was often embedded in a
more encompassing metaphysical debate, one that had as its stake the very reality of the
phenomenal world. For if consciousness can be compared (as many Indian thinkers
believed) to a mirror reflecting the objects of which it is aware, then we are also entitled
to ask in what way the reflections within that mirror can be said to exist.
The following pages are an attempt to understand how real the perceived world is
according to the Śaiva nondualist philosophers Utpaladeva (c. 925–975 CE) and
Abhinavagupta (c. 975–1025 CE) on the basis of their position in a debate on the nature of
optical reflections. This debate might at first sight appear as a minor digression in
Utpaladeva’s magnum opus—the Pratyabhijñā (Recognition) treatise2—and
3
Abhinavagupta’s two commentaries thereon. It is, however, crucial for the understanding
of the Śaiva nondualists’ ontology, and later Indian authors often present it as a defining
feature of Utpaladeva’s system.4 According to the Śaiva nondualistic scriptures, the
universe only exists within Śiva understood as a single, all-encompassing, and all-powerful
consciousness manifesting itself in an infinite variety of forms; and the Śaiva philosophers
express this absolute idealism by saying that the perceived universe is nothing but
reflections on the mirror of consciousness.5
Modern scholarship has already pointed out this recurring use of the mirror analogy in
Śaiva nondualistic literature and offered interpretations of it.6 The goal of this chapter is
to highlight its philosophical import while emphasizing two often overlooked points: first,
contrary to what is usually assumed in secondary literature, this analogy is probably no
novelty on Abhinavagupta’s part7 and was in all likelihood already fully present in
Utpaladeva’s works.8 Second, in order to understand the meaning of the Śaiva contention
that phenomena are like reflections, we must strive to understand what reflections are not
according to the Śaivas. For the Pratyabhijñā treatise and its commentaries endeavor to
provide the dogmas contained in the Śaiva nondualistic scriptures with a rational
justification, and they do so by systematically engaging in a philosophical dialogue with
rival schools of thought. Utpaladeva’s work is profoundly polemical9 and we can hope to
gain a fuller understanding of the Śaiva comparison of phenomena with reflections if we
are able to determine in what way this comparison is a response to other, non-Śaiva
theories regarding the nature of reflections and the reason(s) why phenomena might be
compared to reflections.

Reflections are nothing but the reflecting entity: Vasubandhu’s thesis

Giving an overview of all the positions held in the Indian debate over the nature of
reflections is far beyond the scope of this chapter; but we can start our enquiry by
outlining a few theses that were in all probability known of the Śaivas. Among Buddhist
philosophers,10 Vasubandhu (fourth to fifth centuries CE?), the author of the AKBh (a work
that Abhinavagupta undoubtedly knew),11 played a particularly important role in this
debate.
Determining whether reflections have some kind of reality was already an important
issue in Vasubandhu’s time, particularly in intra-Buddhistic debates involving
Sarvāstivādins (who contended that all things exist in some form at all times and
therefore defended the reality of reflections) and Dārṣṭāntikas (who claimed that reflected
images do not actually exist because the reflected object cannot enter the reflecting
surface although it seems to do so).12 Thus in the AKBh,13 Vasubandhu sets out to show
that reflections have no reality whatsoever and gives several reasons why they must be
mere illusions: two distinct entities (the mirror and its reflection) cannot coexist in the
same place14; two observers watching the same mirror from different points of view do not
see the same reflections in it15; whereas light and darkness are incompatible, the sun can
be reflected in a mirror placed in the shade16; and although the mirror has no depth, we
do not see the reflected moon on the surface of the mirror, but rather, somewhere beyond
it.17 Vasubandhu concludes that reflections are “nothing at all that would exist,” that is,
nothing that would be distinct from the mirror18: when we see reflections in fact we see
nothing but the reflecting entity, but we are deluded into seeing something besides the
mirror because of a particular causal complex involving the presence of a mirror, light, a
specific angle, and so on.19

Reflections are nothing but the reflected object: the position held by
Naiyāyikas and Mīmāṃsakas
On the Brahmanical side, the reality of reflections is also denied. Thus the Naiyāyikas
agree with Vasubandhu that we are mistaken in thinking that we actually perceive such
things as reflections.20 But according to Brahmanical authors, reflections are illusions not
because they would be nothing apart from the reflecting entity itself, but rather, because
upon seeing a reflection in a mirror, we see nothing but the reflected object (bimba) itself.
In order to understand this thesis we should keep in mind that according to the
Naiyāyikas vision occurs thanks to some invisible rays emanating from the eyes and
coming into direct contact with the object to be perceived.21 Pakṣilasvāmin (fifth century
CE?) therefore explains that when we see our face in a mirror, the visual rays bounce back
upon reaching the surface of the mirror and end up directly touching the face.22 As
Uddyotakara (sixth century CE?) specifies, in such a case we mistakenly believe that we
perceive an object consisting in an image of our face and residing in the mirror because
our perception of the mirror and the immediately posterior perception of our face occur
too quickly for us to realize that these are two different objects in two different places.23
The Mīmāṃsakas seem to have adopted the Naiyāyikas’ general theory regarding
optics24 and their understanding of the phenomenon of optical reflection is quite similar.25
Thus the Mīmāṃsaka Kumārila (sixth century CE?) also claims that when perceiving what
we believe to be a reflection of the sun, in fact we grasp nothing but the sun itself, but we
are deluded into thinking that we see a reflected image that has an existence of its own
because the visual rays are diverted upon reaching the surface of the reflecting entity:
when we contemplate a reflection of the sun in water, in fact our visual rays are in contact
with the sun itself.26 Similarly, Pārthasārathimiśra (who lived at some point between the
tenth and thirteenth centuries CE) insists that reflections do not exist as such. Quite
amusingly, in order to demonstrate this he invokes arguments very close to those used by
the Buddhist Vasubandhu27 before offering a conclusion that has a much more orthodox
ring from a Brahmanical point of view: reflections are nothing because when we perceive
them we only grasp the reflected object.28

A Buddhist critique of the Brahmanical position: Śāntarakṣita’s claim that

reflections have no objective support
Vasubandhu’s position was defended and refined in the eighth century by Śāntarakṣita
and his disciple Kamalaśīla, who both emphasize their debt toward the author of the AKBh
by ostentatiously paraphrasing his arguments against the reality of reflections.29
Śāntarakṣita (whose TS might well have influenced Utpaladeva)30 thus explains that
reflections are a mere illusion (bhrānti)—that is, in fact they are nothing at all31—on the
grounds that (1) two distinct material objects cannot exist in the same place32; (2) two
observers differently located do not see the same reflection in the same place33; and (3)
although the mirror has no depth, we see reflected objects as if they were below the
mirror’s surface.34 Śāntarakṣita and Kamalaśīla also criticize the view that the image seen
in the mirror has a reality of its own by showing that whether we consider the mirror as a
lasting entity or whether we accept the Buddhist view of universal momentariness, the
mirror can never start bearing a reflection. For if the mirror has a lasting existence it
cannot be endowed with changing reflections (which would alter its nature and make it
impermanent); it must therefore reflect the same image (or no image at all) forever.35 But
if we concede its momentariness, we must also admit that at every moment there arises a
different mirror bearing a different reflection, so that the mirror, lasting or not, can never
acquire a reflection—an absurdity which vanishes if we accept that the reflection is a
mere illusion.36 Finally, contrary to what the Brahmanical authors (i.e., here, mainly
Kumārila’s followers)37 claim, the perception of the reflection cannot have an external,
objective support (ālambana) in the form of the reflected object itself, because the
reflection and the reflected object have strikingly different appearances38: the reflection
of my face in a mirror is located in front of my own body instead of being part of it, it
faces the opposite direction and has a different size, color, and so on according to the
particular properties of the object that reflects it.39 And Kumārila cannot rightfully claim
that these two very different appearances share a single objective support, because this
contradicts another thesis of his according to which the aspect or appearance (ākāra) of
objects does not belong to consciousness. For the Mīmāṃsaka denies that a cognition
grasps the object that it perceives by taking on its form: according to him the appearance
of the perceived object is, rather, a property inherent in the object itself.40 But by
defending such a thesis he loses the ability to explain the occurrence of erroneous
perceptions in which we mistakenly apprehend an object as located in a place where in
fact it is not found, whereas he himself describes the perception of reflections as a
mistaken apprehension of this type.41

The metaphysical stake in this optical debate: phenomena as reflections

within consciousness
As noted above, in classical India optical theories were formulated in connection with
philosophical debates, and the discussion on the nature of reflections is no exception. One
of its main philosophical contexts42 is the issue of the ontological status of the phenomenal
world. Thus the contention that the reality of the perceived universe is comparable to that
of a reflection is a topos of Buddhist texts43 both in Madhyamaka literature (where it
usually conveys the idea that all things are “devoid of intrinsic nature,” niḥsvabhāva)44
and in the Vijñānavāda. In this latter tradition, which propounds the thesis that everything
is consciousness,45 it is often found along with the more famous dream analogy (according
to which perceived objects have no more reality than dreamt objects)46 and it is used to
show that all ordinary perceptions are devoid of any external support.47 Vasubandhu and
his Buddhist contemporaries were certainly already aware of the implications of the
discussion over the reality of reflected images for the ontological status of the perceived
universe48; and in later Indian philosophical texts these implications become more and
more transparent as the Vijñānavādins elaborate their critique of a rival Buddhist thesis
(usually ascribed to the Sautrāntikas) according to which the reason why we perceive
various objects is the existence of an external world casting a reflection of itself onto
consciousness.49 By way of contrast, according to the Vijñānavādins reflections are mere
illusions (when perceiving them we wrongly believe that we are grasping an object called
“reflection” independent of our perception of the mirror), just as, when perceiving any
object of the world, we mistakenly assume that this object has an existence of its own,
independently of the cognition manifesting it. While defending the Vijñānavāda, Buddhist
authors such as Śāntarakṣita and Kamalaśīla50 therefore explain that precisely because
cognitions of reflections appear to present an external object while reflections are in fact
nonexistent, we must consider that similarly, all perceived objects have no reality outside
of consciousness.51
 As for the theory of reflections held by the Naiyāyikas and Mīmāṃsakas, whether or not
it was primarily designed as an answer to the Buddhist idealists, it is from this
perspective that Śāntarakṣita and Kamalaśīla understand it. Thus according to them
Kumārila is mainly concerned with showing that all cognitions must have a substrate
external to consciousness: our dreams only appear not to correspond to any reality, since
dream objects really are combinations of previously perceived entities; and similarly,
reflections only seem to be unsubstantial, but in fact the perception of the sun’s reflection
in the water, far from apprehending a nonexistent entity, grasps a real, existing sun in the
sky. And as a consequence the Vijñānavādin cannot rely on such cognitions to argue that
all perceived objects are devoid of any external basis.52
Finally, it should be noted that Kumārila’s explanation of optical reflections was
adopted by some of the proponents of Advaita Vedānta.53 However, while in the Mīmāṃsā
the theory of optical reflection served to emphasize that even illusions such as reflections
have an external support (so that one cannot deny the reality of the external world on the
grounds that such illusions occur), the Vedāntins rather exploited the analogy of
phenomena with reflections to establish their nondualism54: they used it to show that the
variety of the phenomenal world is an illusion55 since it is only the result of our distorted
apprehension of reality,56 and since in fact this unreal appearance of diversity occurs on
the basis of the absolute unity of the Brahman.57

What reflections (and phenomena) are not according to the Śaivas: neither
illusions, nor the reflected object itself
As for the Śaiva nondualists, paradoxically for authors who came to be known as the
proponents of a “doctrine [that phenomena are similar to] reflections” (pratibimbavāda),
they often seem to mistrust this analogy of reflections. The reason for this is that they
defend an absolute idealism according to which reality is a unitary, omnipotent, and
infinite consciousness playfully manifesting itself in the form of a diverse and external
universe (whereas in fact nothing exists outside of this consciousness). This means that
they cannot accept any representation of consciousness as a mirror faithfully reproducing
images of the external world: consciousness is not like a mirror because a mirror
passively reflects a given, while according to the Śaivas, the essence of consciousness is
its dynamism. When imagining an object, for instance, we become aware of, for example,
a pot because our consciousness pictures up the pot by taking its form at will. And not
only is consciousness capable of thus taking the form of infinitely varied objects:
according to whether it focuses on the imagined object or on its own creative activity it
can grasp itself in an objective form, as being “the pot,” or in a subjective form, as being
the consciousness taking the form of the pot (“I am imagining this pot”). The Śaiva
nondualists call “realization” (vimarśa or parāmarśa)58 this extraordinary ability of
consciousness to freely apprehend itself as being this or that; and they consider that Śiva
manifests the universe by virtue of this very power, that is, by merely imagining himself in
the form of an external world: the innumerable objects and even the various individuals
living and dying in that world are nothing but Śiva imagining himself in the form of such
limited entities.59
The Śaivas therefore emphasize that consciousness is not like a mirror in that it does
not passively reflect a preexisting and independent universe.60 Yet they do have recourse
to the analogy of phenomena as reflections. This analogy appears in several of
Abhinavagupta’s works,61 but Abhinavagupta himself specifies that its source is
Utpaladeva’s Pratyabhijñā treatise62; and indeed, it occurs in this work as Utpaladeva
emphasizes that consciousness is capable of manifesting diversity while remaining
unitary,63 just as a mirror can show multiple reflections without being shattered.64 While
commenting on this passage, Abhinavagupta explains:

And [the reflected objects] such as the mountain do not occur in the mirror in the way [illusory
appearances] such as silver or a double moon [occur, i.e.] by concealing the nature of the [real]
mother-of-pearl or single moon; for when such a manifestation of the mirror [bearing reflections
occurs], the very fact that [the object in front of us] is a mirror is all the more obvious, since [we
then] think: “this is a stainless, an excellent, a pure mirror!” For the mountain that is external [to
the mirror] does not enter the mirror, since [if it were the case, then] as a consequence this
[external mountain] would leave its own place[, which is absurd]; nor is the [mountain] manifest on
the surface of the [mirror], since [if it were the case, then] as a consequence the mirror would not
 be manifest [as it would be hidden by the mountain]; nor is [the mountain manifest] inside [this
mirror, since] the nature [of the mirror] is dense, solid and resistant, so that there is no possibility
of penetrating it; nor [is it manifested] behind [the mirror,] since [we] do not see [it] there[, behind
the mirror], and since it is manifested only as being distant [from the mirror and not right behind
it]. Nor do the visual rays, being diverted [from the mirror] because upon reaching it they have
bounced back, grasp the mountain itself, since when a mirror located on the side of a mountain
manifests [a reflection of the mountain, we can] see both the reflected [object] and its reflection.65

The passage is important because it shows what reflections (and therefore phenomena)
are not according to the Śaivas.
First of all, we cannot hold that they are pure and simple illusions because they are not
false appearances concealing reality. Error is defined by the Śaivas as an incomplete
manifestation (apūrṇakhyāti), and indeed, when we mistake mother-of-pearl for silver due
to their common property of glittering, we fail to grasp the existence of mother-of-pearl as
long as we are convinced that what we see is silver; similarly, when we see two moons
instead of one due to some malfunction of our visual organ, we remain unaware of the
single moon as long as we believe that there are two moons in the sky. But in the case at
hand, nothing remains veiled or unmanifest. Reflections do not prevent us from being
aware of the mirror’s presence; in fact they even lead us to realize that the surface on
which they appear is indeed a mirror, because when seeing a reflection we understand
that it is only a reflection, and we do not surmise that the reflected object itself has
somehow entered the mirror: we know that we are only contemplating an image of this
object, and while apprehending this image we remain conscious of the mirror’s presence.
The argument already appeared in Utpaladeva’s lost commentary on his own verses66
(along with a criticism of the objection that reflections cannot be real because of the
contradiction between their place and size and that of the corresponding reflected
object)67; it is also found in later Śaiva literature.68
But Abhinavagupta also argues that reflections cannot be reduced to the reflected
objects themselves: the thesis that visual rays are diverted upon reaching the surface of
mirrors and thus come into direct contact with the object that they reflect is not sound
because in some cases we can grasp simultaneously the reflected object and its reflection
in the mirror. In the TĀ, Abhinavagupta criticizes the same position by saying that if we
really perceive thanks to a visual organ capable of leaving our body to touch distant
objects, then mirrors should be useless,69 and we should be aware of perceiving our face
itself rather than an image of it in a mirror.70
Who is thus targeted by Abhinavagupta? In the ĪPVV, when mentioning the argument
that reflections must be illusions because of the contradiction affecting their place and
size, Abhinavagupta’s opponent invokes Bhartṛhari’s authority71; but Bhartṛhari does not
seem to have concerned himself with the ontological status of reflections,72 and the rest of
the discussion in the ĪPVV shows that the thesis that reflections are mere illusions is
rather defended by a Buddhist.73 Abhinavagupta obviously has in mind arguments of the
type formulated by Vasubandhu and Śāntarakṣita, and the passage could even be read as
a pastiche of these authors when they try to show that reflections are illusions because
reflected objects are seen as possessed of properties that cannot belong to them:
according to Abhinavagupta, reflections are not illusions precisely because we are
immediately aware, upon seeing a mirror, that what we see is not the object itself having
penetrated the mirror’s surface, but something else that has an existence of its own.
As for the thesis that visual rays bounce back, while commenting on Abhinavagupta’s
TĀ, Jayaratha (thirteenth century CE) ascribes it to some Naiyāyikas74 whom he deems
“whimsical” since even in the Nyāya their position was far from being unanimously
shared, and seems to have been ignored in particular by his Kashmiri fellow
countrymen.75 There are indeed good reasons to think that in these passages
Abhinavagupta targets the Naiyāyikas Pakṣilasvāmin and Uddyotakara: as seen above,
their works (which Abhinavagupta knew)76 defend an explanation of optical reflection very
similar to the thesis criticized here. Besides, the Śaiva nondualists argue at length that
reflecting objects are characterized by a property called limpidity (nairmalya or
svacchatā) which consists in a power of manifesting other things while manifesting
itself,77 and which, in its purest form, only belongs to consciousness78; and this theory is
certainly a response to the Naiyāyikas’ definition of limpidity as a purely material
property,79 which, according to the Śaivas, results from a confusion between limpidity and
resistance (the latter belonging to all material bodies and having nothing to do with the
ability to reflect).80
It is nonetheless possible and even probable that Abhinavagupta was not exclusively
targeting the Naiyāyikas. Thus Utpaladeva and Abhinavagupta very likely knew
Śāntarakṣita’s TS (or at least the Buddhist arguments formulated in this work against the
Mīmāṃsā’s main theses)81 so that they were certainly aware of the debate on reflections
between Vijñānavādins and Mīmāṃsakas and of its implications regarding the status of
phenomena.
Finally, Abhinavagupta could be alluding here to the Advaita Vedānta’s position
regarding reflections. Determining whether this is likely is an arduous task though. Many
a fanciful story links Abhinavagupta with Śaṅkara; thus the Vedāntin is said to have died
of an ulcerous disease as a result of a curse by Abhinavagupta, who in turn purportedly
died of the same disease transferred back to him by Śaṅkara’s pupil Padmapāda.82 But
finding scientific evidence of a relationship between Śaiva nondualism and Vedāntic
literature proves particularly difficult. Since Śaiva dualists as well as nondualist
predecessors of Utpaladeva were undoubtedly acquainted with Vedāntic doctrines,83 there
is no reason to think that Utpaladeva and his commentator were ignorant of them. But
most of the time Vedāntic theories are conspicuously absent in the Pratyabhijñā treatise:84
it is as though Utpaladeva and Abhinavagupta had felt that strategically speaking,
targeting such prestigious and well-established rivals as the Buddhist epistemologists or
the Naiyāyikas and Mīmāṃsakas should be much more rewarding than discussing with
proponents of Advaita Vedānta. One important clue with respect to this debate on
reflections is the fact that at least one later Śaiva author, namely, Maheśvarānanda
(fourteenth century CE?), sees it first and foremost as a discussion with Vedāntins.85
However, Śaiva nondualism seems to have been progressively affected by what we might
call a “vedānticization” and it has been noted that the originality of the Pratyabhijñā texts
was sometimes lost by later commentators projecting onto them the monistic ontology of
Advaita Vedānta86: it is quite possible that the same tendency affects Maheśvarānanda’s
understanding of this debate. Nonetheless, Utpaladeva’s master, Somānanda (900–950
CE?), already mentions a Vedāntic theory according to which the phenomenal world with
all its differences is a manifestation “as a reflection” of the unitary Brahman.87 It seems
very unlikely that Utpaladeva, who has authored a commentary on Somānanda’s ŚD,88 was
unaware of the Vedāntins’ position in this debate, all the more since the only important
debate with Vedāntins in the whole Pratyabhijñā treatise occurs precisely in the verse
following that devoted to the discussion on reflections, and revolves around the
ontological status of the differences pervading the phenomenal world (which, Utpaladeva
argues, cannot be considered illusory or contradictory with the unitary nature of the
absolute consciousness).89 So although it is unclear which Vedāntic work Utpaladeva and
Abhinavagupta might have in mind,90 we can safely assume that in this debate they are
taking into account not only the thesis held by the Naiyāyikas, Mīmāṃsakas, and
Vijñānavādins, but also that of the Vedāntins.91

The Śaivas’ “teaching of the mirror”: phenomena as reflections without a

reflected universe
To sum up, when discussing the nature of reflections or comparing phenomena with
reflections, Utpaladeva and Abhinavagupta have in mind several rival theories that we
must take into account if we wish to understand the Śaivas’ position. They reject the
Naiyāyikas’ and Mīmāṃsakas’ view that just as reflections (which are supposedly nothing
but objects external to the reflecting entity), phenomena are aspects belonging to a reality
external to, and independent of, consciousness. However, while they share with the
Vijñānavādins the idea that nothing exists apart from consciousness, and with the
Vedāntins the belief that the ultimate reality is unitary, they refuse to admit with both
interlocutors that reflections and phenomena are illusory appearances. According to the
Śaivas, reflections, just as the differentiated universe of our perceptions, exist in some
way that cannot be denied on the mere grounds that the ultimate reality is consciousness
(as the Vijñānavādins contend) or that it must be purely unitary (as the Vedāntins argue).
But then what are reflections according to the Śaivas, and in what way can the perceived
world be compared to them?
The answer to that question is that in the Pratyabhijñā system, reflections cannot be
reduced to illusory appearances because they are various ways for the mirror—which is
real, and the essence of which is precisely to manifest itself in a variety of ways—of
manifesting itself; but for this very reason, they have no autonomous existence.
Reflections only exist insofar as mirrors manifest them, and this absolute ontological
dependence constitutes their specificity.92 Abhinavagupta explains (here again, while
relying on Utpaladeva’s lost Vivṛti)93 that this is precisely why there is a point in
comparing phenomena with reflections:

Just as reflections are manifest [only] while adhering to the [surface of the] mirror, but do not exist
by themselves, regardless of their identity with the mirror, in the same way, it is while adhering to
the manifesting [consciousness, i.e.] while being entirely dependent on it—that the [various
perceived objects] such as the pot are manifest; [but] they do not exist by themselves.94

Reflections are real, and yet they are only real inasmuch as they belong to the mirror from
which they seem to be distinct95; they appear to be distinct from the mirror, and yet they
can be manifest only because they are fundamentally one with the mirror.96 Similarly,
perceived objects only exist insofar as consciousness takes on their form, and they too
seem distinct from consciousness (to be aware of an object in front of us is to apprehend
the presence of something external to us), but they can only seem so because our
consciousness manifests them by manifesting itself as if it were distinct from itself (our
consciousness of the external pot is our consciousness taking on the form of a pot external
to consciousness).97
Admittedly, there is something strikingly odd about this comparison of consciousness
with a mirror in an idealistic system, since reflections do not only depend on the existence
of a reflecting surface but also on the presence of an object external to that reflecting
surface. The Śaivas systematically downplay this aspect of the problem.98 They are able to
do so because they have shown that consciousness cannot be compared to a mirror
reflecting external entities (for the simple reason that the distinction between “inside”
and “outside” becomes meaningless when it comes to consciousness),99 but also because
as shown above, they subject the Naiyāyika notion of limpidity to a complete conceptual
metamorphosis. In the Pratyabhijñā system, the characteristic of reflecting entities
becomes the capacity to manifest oneself as something else while remaining oneself
(which is the very essence of consciousness according to the Śaivas). This profoundly
transforms the meaning of the mirror analogy, since it is no longer consciousness that is
seen as functioning like a mirror (while the latter requires an external entity of which it
produces an image): rather, mirrors are to some extent comparable to consciousness, and
the fact that mirrors need an external reality to produce reflections is no longer a defining
feature of the act of reflecting, but only a sign that mirrors do not possess in full the
ability to reflect—mirrors are not mirrors enough when compared to consciousness.
The Śaivas’ “teaching of the mirror”100 is thus meant as a way to point out the infinite
plasticity of consciousness: it is first and foremost an invitation to recover one’s identity
with Śiva by systematically recognizing phenomena as limited aspects that the absolute
consciousness freely takes on. For although Utpaladeva and Abhinavagupta insist that
reflections are not illusions (because we are usually aware that they are only reflections),
they also compare the mistake of ordinary individuals—who have lost the awareness of
their identity with the unitary and all-encompassing consciousness—with that of
somebody who, in some particular circumstances, would cease to be aware that what
(s)he sees is a mere reflection in a mirror: the ordinary subjects are metaphysically
deluded precisely insofar as they do not realize that the various perceived objects
(including those that they wrongly consider as themselves, such as, e.g., their bodies) are
in fact nothing but reflections in consciousness, that is, ways for consciousness of
manifesting itself as what it is not.101 According to the Śaivas, however, such an illusion is
possible only because the absolute consciousness never really loses the awareness that
phenomena are nothing but reflections, just as, when we are daydreaming, for instance,
we can get momentarily engrossed in our fantasies, but only because somehow we always
remain aware that it is just a fantasy.102
In the Śaiva nondualistic perspective the goal of the mirror analogy is therefore to
enable us to realize our identity with the absolute consciousness by making us pay
attention to the freedom of consciousness involved in any ordinary perception:103 against
the Vedāntins, the Śaiva nondualists refuse to discard the phenomenal world as an illusory
appearance, explaining that phenomena can reveal the power of consciousness, just as,
upon realizing that we are seeing a reflection, we become aware of the power of the
surface bearing them. Abhinavagupta even playfully reverts the Vedāntins’ use of the
mirror analogy (i.e., the idea that phenomenal differences are only adventitious properties
that do not belong to reality, just as a face reflected on various surfaces can show multiple
distortions that do not really affect it)104 to point out that phenomena, insofar as they are
manifestations of consciousness, must be pervaded with the power of freely manifesting
oneself that characterizes consciousness:

And just as a smell, a visual form, tactile sensation, taste and so on,105 when reflected, are manifest
while having their appearance affected by [the properties] of their [reflecting] support, as [in the
case of] e.g. a face [reflected] in a sword, in the same way, this [entire] universe, which is reflected
in consciousness, must receive all [its] properties: being a manifesting entity, being autonomous,
etc.106

So this mirror analogy is primarily used by the Śaivas as a soteriological device. Yet it is
pervaded with polemical concerns. The previous pages have certainly given the reader a
glimpse of the many difficulties and uncertainties involved in any attempt to outline the
various positions against which this analogy was designed. Hopefully, however, they have
also contributed to show that, because Indian concepts were elaborated in a constant
dialogue between rival currents, our best chance to understand them is to try and replace
them in their proper philosophical context. This is a difficult task, but also an exciting one,
and much remains to be done in this respect—particularly as regards the Śaiva
philosophical corpus.

Notes
* Heartfelt thanks are due to Vincent Eltschinger, who read an earlier version of this chapter and
provided insightful remarks.
1. Preisendanz (1989, pp. 143–146).
2. That is, the ĪPK, a series of verses on which Utpaladeva himself has written two commentaries: a
short Vṛtti (on which, see Torella [2002]) and a more detailed Vivṛti of which only fragments are
available to date (see Torella, 1988, 2007a, b, c, and d; Kawajiri, 2016; and Ratié 2016a, b and c).
3. The discussion on the nature of reflections appears in both of Abhinavagupta’s commentaries (the
ĪPV and ĪPVV) on ĪPK 2.4.19 and was certainly the main topic of Utpaladeva’s lost Vivṛti ad loc.
4. See, for example, Maheśvarānanda’s MMP (written around 1300 ce?), pp. 153–160; cf. the SDS, a
fourteenth-century doxographical work according to which in the Pratyabhijñā system, Śiva
“manifests the [objective] entities in the mirror of his Self, like reflections” (svātmadarpaṇe bhāvān
pratibimbavad avabhāsayati) (p. 190).
5. Although in TĀV, vol. II, pp. 26–27, Jayaratha mentions a few scriptural passages conveying (or
seeming to convey) the idea that phenomena are similar to reflections, this mirror analogy appears
to be a largely post-scriptural phenomenon within the Śaiva corpus.
6. See, for example, Rastogi (1984, pp. 28–31), Lawrence (2005), Ratié (2011a, pp. 273–289), Padoux
and Ratié (2013), and Kaul (forthcoming).
7. Thus Rastogi (1984, p. 28), states that Abhinavagupta “follows ... Utpala in using this analogy but
with added dimensions” (without specifying the nature of these added dimensions). Similarly, while
conceding that Utpaladeva makes use of this analogy, Lawrence (2005, p. 586) adds that
“nevertheless, Utpaladeva does not thematize reflection as a basic theoretical and practical
approach to the Ultimate Reality” while “it is Abhinavagupta who may be credited with making the
metaphor of reflection into a favored trope of monistic Śaiva theological discourse.”
8. See in particular nn. 62, 64, 66, 67, and 93 below.
9. See, for example, Frauwallner (1962a, p. 22), Sanderson (1985, p. 203; 1988, p. 694), Torella
(2002, p. xiii), and Ratié (2011a, pp. 6–14).
10. On the profound influence of the Buddhist philosophical traditions on Utpaladeva and
Abhinavagupta, see, for example, Torella (1992), Ratié (2010a, 2011b, 2014a and 2016a and b),
and McCrea (2016).
11. It is quoted, for example, in ĪPVV, vol. II, p. 225 (= AKBh on AK 2.24, p. 54); vol. II, p. 227 (= AK
2.50a); vol. III, p. 191 (= AK 5.48b); see also the allusion in ĪPVV, vol. I, p. 175 (abhidharmādau . .
.).
12. On this debate, see, for example, Cox (1988, pp. 53–55).
13. On the context of this passage (a discussion of the Buddhist notion of intermediate existence,
antarābhava, between the moment of death and that of rebirth), see Kritzer (2000, particularly pp.
243–247) and Fukuda (2003, p. 261).
14. See AKBh, p. 120: kathaṃtāvad asiddham—sahaikatra dvayābhāvāt. tatraiva hi deśa
ādarśarūpaṃdṛśyate pratibimbaṃca, na caikatra deśe rūpadvayasyāsti sahabhāva
āśrayabhūtabhedāt. “To begin with, how [can we claim that the existence of the reflection] is not
established? [It is the case] because two [things] cannot exist together in the same [place]. For
[we] see the form of the mirror and the reflection in the same place; but two forms cannot exist
 together in the same place, because the elements that [constitute their] substrates must be
different[, otherwise they would be the same thing].”
15. AKBh, p. 120: tathā digbhedavyavasthitair* ekasmin vāpyambudeśe
svābhimukhadeśasthānāṃrūpāṇām anyonyaṃpratibimbakam upalabhyate na tv ekatra rūpe
dvayoḥpaśyatoḥsahadarśanaṃna bhavatīti na tatra rūpāntaropapattir yuktā.
[*digbhedavyavasthitair AKVy: digbhedavyavasthiter AKBh.] “In the same way, several [observers]
located in different places [around a pond] perceive in the same water of the pond a reflection of
the other [observers’] forms that are found in the places faced by each [observer]; yet two
observers must simultaneously see [the same thing] in the same form[, which is not the case here];
so the existence of a form distinct [from the water, i.e. the reflection], cannot be legitimate[ly said
to exist] in the [water].”
16. AKBh, pp. 120–121: chāyātapayoś ca dvayoḥsahaikatra bhāvo na dṛṣṭaḥ. upalabhyate ca
chāyāstha ādarśe sūryasya pratibimbakam iti na yukto’sya tatra prādurbhāvaḥ. “Moreover, [we] do
not see that these two [things,] shade and sunlight, [can] exist in the same [place]; and [yet we]
perceive the reflection of the sun in a mirror located in the shade—so the arising of the [reflection
of the sun] in the [mirror] cannot be rationally justified.”
17. AKBh, p. 121: anyatraiva hi deśa ādarśatalaṃbhavaty
anyatraivāntaragataṃcandrapratibimbakaṃdṛśyate kūpa ivodakam. tac ca tatrotpadyamānaṃ*
nānyatropalabhyeta**. [*tatrotpadyamānaṃ AKVy: tatropapadyamānaṃ AKBh.
**nānyatropalabhyeta AKVy: nānyatropalabhyate AKBh.] “[Or again, the reflection does not exist]
because the surface of the mirror exists in one particular place, [whereas we] perceive the
reflection of the moon as being located in another place, [i.e. deep] inside [the mirror], just as
water in a well [that we perceive as being at the bottom of the well and not on its surface]. And
this [reflection, if it really] arose there[, on the surface of the mirror,] would not be perceived
elsewhere[, beyond this surface.]”
18. Cf. AKBh, p. 120: pratibimbaṃnāmānyad evotpadyate dharmāntaram ity asiddham etat. “[The
existence of the reflection is not established, i.e.] it is not established that what [we] call a
‘reflection’ arises while being distinct, [i.e., as] a distinct entity.”
19. See AKBh, p. 121: ato nāsty eva tat kiṃcit. sāmagryās tu sa tasyās tādṛśaḥprabhāvo yat tathā
darśanaṃbhavati, acintyo hi dharmāṇāṃśaktibhedaḥ. “As a consequence this [reflection] is nothing
at all that would exist; but the power of the causal complex (sāmagrī) is such that it produces an
appearance of this sort, for entities have a diversity of powers that cannot be fathomed (acintya).”
Kritzer (2002) shows that here as elsewhere in the AKBh, Vasubandhu uses the term “acintya” as a
hint that this cannot be explained by the orthodox Sarvāstivādin doctrine (according to which
reflections are real).
20. See, for example, NVTṬ, p. 485: ...prasaṅgāt pratibimbabhramotpādakramam āha ...
“[Uddyotakara] explains as a consequence [of the point just discussed] the process through which
the illusions that are reflections arise.”
21. On this theory, see, for example, Frauwallner ([1956] 1973, pp. 29–33) and Preisendanz (1989); on
its criticism by Buddhist authors, see, for example, Hattori (1968, pp. 36–37 and 124–126).
22. See NBh, p. 161: yathādarśe pratihatasya parāvṛttasya nayanaraśmeḥsvena mukhena sannikarṣe
sati svamukhopalambhanaṃpratibimbagrahaṇākhyam... bhavati. “For example, in a mirror, when
the visual rays, being diverted [and] bouncing back, come into contact with [our] own face, there
occurs a perception that [we usually] call the ‘apprehension of a reflection’ [although in fact it is a
perception of our] own face...” (Following Preisendanz (1989, fn. 31, p. 148), I understand raśmi in
the singular as referring to a “bundle of single rays.”)
23. See NV, pp. 362–363 : ... prasādasvabhāvatvād ādarśodakādiṣu nayanaraśmiḥpratihanyate. sa ca
pratihataḥpratinivṛtya svamukhādinā sambadhyate. tasya cāgrasambandhād yad abhimukham
agraṃtad abhimukhaṃmukhādi paśyatīti yathāgrato vyavasthitasya puruṣasyeti.
ādarśamukhagrahaṇaṃtu krameṇa bhavad apy āśubhāvān na vibhāvyate. “In such [things] as a
mirror or water, because [their] nature is limpidity, the visual rays are diverted. And when the
[visual rays] that are diverted bounce back, they come into contact with e.g. one’s own face. And
because [in this process they] come into contact with what[ever is] in front [of the mirror], they
see what[ever] is in front, such as a face; for instance [they see the face] of the man standing in
front [of the mirror]. But although the apprehension of the face and [that of] the mirror occur one
after the other, due to the speed [of this process] it does not appear [to be so].”
24. On the Mīmāṃsakas’ defense of the Nyāya thesis that visual rays emanate from the eyes and
come into contact with the object, see, for example, Bhatt (1962, pp. 174–177) and Taber (2005,
pp. 58–64).
25. It is not identical however (as far as I understand it): whereas according to the Naiyāyikas, the
visual rays bounce back because the mirror’s surface is endowed with a particular property called
prasāda (“limpidity,” on which see n. 79 below), according to the Mīmāṃsakas, it is a light ray on
the mirror’s surface which is responsible for the visual ray bouncing back (see below, nn. 26, 28,
and 53).
26. See, for example, ŚV, Śabdanityatādhikaraṇa 180cd–181ac: atra brūmo yadā tāvaj jale saureṇa
tejasā // sphuratā cākṣuṣaṃtejaḥpratisrotaḥpravartitam / svadeśa eva gṛhṇāti savitāram ... “To this
[objection] we reply [the following]: to begin with, when a sunray shining on water makes the
 visual ray bounce back, [what this visual ray] grasps is the sun [existing] in its own place [i.e., the
sky].”
27. See Śāstradīpikā, p. 399: nanu pratibimbaṃnāmārthāntaraṃtatra dṛśyate. na, tasyābhāvād
anupalambhāc ca, na hi mūrtamadhye mūrtāntaraṃsambhavati. kiṃca śarāvastham
udakaṃbhūmer upari nābhidaghne dhārayitvā tasyopariṣṭād aratnidaghne svamukhaṃkurvann
udakasyādhastād aratnimātre mukhapratibimbaṃpaśyati, tasmiṃś ca deśe pārśvasthāḥpuruṣā na
kiṃcid api paśyanti tena dṛśyādarśananirastaḥpratibimbākhyo’rtho na śakyo’bhyupagantum. “[–
Objection :] But in the [case when something is reflected in a mirror or water, we] perceive
[something] called ‘reflection’ which is an object distinct [from the reflected entity. To this we
answer]: no, because the [reflection] does not exist and because [we] do not [really] perceive it.
For a material [thing] cannot exist within another material [thing]. Besides, [someone] who carries
water above the ground in a plate, at the level of his navel, placing his face above the [water] at a
maximum distance of one cubit, sees a reflection of his face only within that [distance of one] cubit
[and sees it as if it were] under the [surface of the] water, whereas the persons standing on the
side do not see anything at that very place [where he sees the reflection]. Therefore [we] cannot
admit the [existence of the] thing called ‘reflection’: [it is] refuted by the fact that [we] do not
perceive [it whereas] it should be perceived [if it existed].” Cf. AKBh, p. 120, quoted above, nn. 14,
15, and 17.
28. See Śāstradīpikā, pp. 399–400: tasmād ādarśatejasā jalena ca pratihataṃnāyaṇaṃtejaḥparāvṛttya
gṛhṇātīti yuktam. “Therefore it is right to hold [rather] that the visual rays, which have been
diverted by the [light] ray from the mirror and by water and have bounced back, grasp [our very
face and not an image of it somewhere else].”
29. Compare, for example, AKBh, p. 121 (quoted above, n. 19) with TS 262ab (quoted below, n. 36)
stating that things have diverse “powers” (śakti) that “cannot be fathomed” (acintya). See also
Kamalaśīla’s allusion to Vasubandhu’s analogy with the water in a well: compare AKBh, p. 121
(quoted above, n. 17) with TSP, vol. I, p. 133 (quoted below, n. 34).
30. See Ratié (2014c, pp. 210–218).
31. See below, nn. 32, 34, and 36. Note that the terms bhrānti/bhrama can denote both the illusion
understood as the appearance having no correspondence with reality and the cognition
apprehending this appearance.
32. See TS 259b–d: ... mūrtānām asahasthiteḥ/ bibharti darpaṇatalaṃnaiva chāyāṃkadācana //
“Because material [objects] cannot exist together [in the same place], the surface of the mirror
never bears any reflection at all in any circumstance.” Kamalaśīla comments (TSP, vol. I, p. 133):
tathā hi darpaṇatale taddeśāny eva parvatādipratibimbāny upalabhyante, na ca
mūrtāḥpadārthāḥkadācid ekadeśatām āpadyante, aikātmyaprasaṅgāt. “To explain: on the surface
of the mirror, the reflections of e.g. a mountain are perceived as being located precisely in the
[place occupied by the mirror]; but material objects cannot share the same place in any
[circumstance], because [if they did] as a consequence they would be one and the same [thing].”
See also TS 2592: pratibimbodayas tv atra prāg eva vinivāritaḥ/ sahaikatra dvayāyogān
mūrtānāṃpratighātataḥ// But [we] have already dismissed in this [treatise the possibility] that a
reflection might [really] arise, on the grounds that two [things] cannot exist together in the same
[place], because material [objects] offer resistance [to each other].” Cf. Vasubandhu’s argument
quoted above, n. 14.
33. See TS 260: pārśvadvitayasaṃsthāś ca suśuklaṃsphaṭikopalam / samīkṣante* tad eṣo’pi na
chāyāṃpratipannavān // [*samīkṣante correction: samīkṣyante TS.] “And [those] who stand on both
sides [of a piece of crystal that seems red due to the red flower behind it] see that piece of crystal
as perfectly colorless; so this [piece of crystal] does not bear any reflection either.” Cf.
Vasubandhu’s argument quoted above, n. 15.
34. See TSP, vol. I, p. 133: tathā hi kūpāntargatodakavad darpaṇatale pratibimbakam antargatam
upalabhyate, na ca darpaṇatalasya vibhāgaḥ—randhram asti nibiḍatarāvayavasanniveśāt, tasmād
bhrāntir iyam. “To explain: on the surface of the mirror, [we] perceive the reflection as being inside
[the mirror], just as water [that we perceive as being deep] inside the well [and not on its surface];
but the surface of the mirror is undivided, [says Śāntarakṣita, i.e.] it has no empty space since its
parts hold together very tight—so this [arising of the reflection] is an illusion.” See also TSP, vol. II,
pp. 709–710 (quoted below, n. 39). Cf. Vasubandhu’s argument quoted above, n. 17.
35. See TS 258: sopadhānetarāvastha eka eveti sarvadā / tacchāyas tadviyukto vā sa dṛśyetānyathā
punaḥ// “... Whereas if it were not the case [that a reflecting entity such as a mirror or a crystal is
momentary], given that [this entity] would remain one [and the same,] whether in the state where
[a blue lotus] is placed [next to it] or in the state where [this blue lotus] is not [placed next to it, as
a consequence we] would always see it possessed of this reflection, or [we would always see it]
devoid of this reflection!”
36. See TS 261: bhedaḥpratyupadhānaṃca sphaṭikādeḥprasajyate / tacchāyāpratipattiś cet tasya
vidyeta tāttvikī // “And there [would] follow that [a reflecting entity] such as a crystal would be
different according to each [object] placed [next to it] if this [crystal] really acquired a reflection of
the [object placed next to it].” Kamalaśīla explains (TSP, vol. I, p. 133): yadi hi
paramārthataḥsphaṭikāder upadhānoparāgapratipattir bhavet tadā yathā kramabhāvinīnām
upadhānachāyānāṃsvabhāvabhedān naikātmyam, tadvat tadātmanaḥsphaṭikāder apy upadhānam
 upadhānaṃprati pratyupadhānaṃbhedaḥprasajyeta. yadi punar bhrāntir iyam ity aṅgīkriyate
tadāyam adoṣa iti jñāpanārthaṃtāttvikīty āha. “To explain: if the [reflecting entity] such as a
crystal really acquired a coloration from the [object] placed [next to it], then just as the [various
momentary] reflections arising in succession from the [object] placed [next to the crystal] would
not all be the same [reflection] because of the difference between their [respective] natures
[existing at different moments], in the same way, there would also follow that [a reflecting entity]
such as crystal would be different according to each [object] placed [next to it, since the crystal]
would have as its nature these [momentary reflections]. But if [we] admit that this [arising of a
reflection in the reflecting entity] is an illusion, then there is no [more] fault—it is in order to
convey this [idea] that [Śāntarakṣita] says ‘[if this crystal] really [acquired a reflection]...’” In TS
262ab Śāntarakṣita concludes: tasmād bhrāntir iyaṃteṣu vicitrācintyaśaktiṣu / “Therefore this
[arising of a reflection] is an illusion [that occurs] in [the presence of entities] that have diverse
powers which cannot be fathomed (acintya).” Kamalaśīla explains (TSP, vol. I, p. 133): yataś
caivaṃpakṣadvaye’pi chāyāpratipattir na yujyate tasmād bhrāntir iyam iti sthitam. “And since thus,
whichever thesis [we choose] among the two [just mentioned concerning the capacity of things to
last], acquiring a reflection is impossible, it is [now] established that the [arising of reflections in
mirrors, etc.] is an illusion.”
37. Kumārila is explicitly presented as Śāntarakṣita’s target in several passages of the TS and TSP
criticizing the Brahmanical theory of reflections. TS 258–262 (quoted above, nn. 35 and 36) is part
of a critical examination (TS 222–284) of Kumārila’s thesis regarding the nature of the Self; and TS
2586–2589 (quoted below, n. 38) is presented by Kamalaśīla as an answer to TS 2223, which is a
quotation of ŚV, Śabdanityatādhikaraṇa 189cd–190ab (it might as well be a fragment from
Kumārila’s lost Bṛhaṭṭīkā, which, as shown in Frauwallner (1962b), contained a number of verses
identical to those found in the ŚV). As for TS 2080, it is part of a discussion with Kumārila on the
existence of external objects: it seems to me that Jha (1937, vol. II, p. 986) is wrong in supposing
that it answers an objection formulated in the previous verse by “another party” (than the
Mīmāṃsaka). In doing so he follows the assumption of the TSP editor, who surmises (TSP, vol. II, p.
709) a euphonic combination (sandhi) of the word aparaḥ (“another [opponent]”) with the
preceding compound, but given the context it seems likely that one should rather read paraḥ, that
is, “the opponent [dealt with so far]” (namely, Kumārila).
38. See TS 2586–2589ac: pratibimbakavijñānaṃsvāsyādyālambanaṃna tat / tadvilakṣaṇanirbhāsād
rasaśabdādivittivat // alpīyasy āsyam alpīyo darpaṇe pratibhāti hi / viparyastaś ca vṛkṣādir
jalamagnaḥpratīyate // darpaṇābhimukhaṃbimbaṃnaivaṃtu pratibimbakam /
jalādyantargataṃcedaṃbimbaṃtv ārād avasthitam // āśrayānuvidhānena sthūlasūkṣmādibhedi ca /
pratibimbaṃna bimbaṃtu ... “The cognition of a reflection does not have as its [objective] support
(ālambana) e.g. the own face [of the person observing a mirror], because the [objective]
appearance [of this cognition] differs from the [face], just as the cognition of e.g. a taste or a sound
[differs from the cognition of a visual form]. For the face appears smaller in a smaller mirror; and a
tree [on a bank] for instance is apprehended as being upside down [and] immersed in the water
[that reflects it]. The reflected object is turned towards the mirror, but the reflection is not; and
this [reflection] is e.g. in the water whereas the reflected object stands in the distance; and the
reflection is different as regards its being large or minute according to its substrate, whereas this
is not the case of the reflected object.”
39. See TS 2080: nābhimukhyena taddṛṣṭeḥsvamukhādes tathekṣaṇam / pramāṇadeśabhedādidṛṣṭeś
cānyapadārthavat // “It is not e.g. [our] own face that [we] see thus [when observing a mirror],
because [we] see it in front [of us], and because [we] see it as different [from our face] as regards
the size, place and so on, just as when [we see] a [completely] different object.” Cf. TSP, vol. II, pp.
709–710: etad uktaṃbhavati. yadi mukhādigrāhakaṃtajjñānaṃsyāt tadā yathaiva tanmukhādi
vyavasthitaṃtathaiva gṛhṇīyāt, na hy anyākārasya jñānasyānyad grāhyaṃyuktam atiprasaṅgāt,
yāvatā dakṣiṇābhimukhasthito darpaṇatalaṃnibhālayann uttarābhimukhaṃsvamukhaṃpaśyati,
tathālpīyasi* darpaṇatale mahato’pi svamukhasyālpapratibimbakam upalabhyate tathā
darpaṇatalasambaddhaṃdūrādhaḥpraviṣṭam ivekṣyate. na ca tāvad bahalaṃtathādarśatalaṃnāpi
mukhādi tatsambaddham. tathā vimalasalile sarasi taṭāntasthitaśākhiśikhariṇāṃpratibimbāny
adhogataśākhādiśikharaśekharāṇy upalabhyante, na ca te tathā sthitāḥ. tasmāt
pratibimbajñānaṃna svamukhādigrāhakaṃtadvilakṣaṇapratibhāsitvāc chabdajñānavat.
[*tathālpīyasi TSPK: yathālpīyasi TSP.] “This is what [Śāntarakṣita] means: if the cognition of the
[reflection] apprehended e.g. [our own] face, then [we] would apprehend it exactly as our face is;
for it is not right [to assume that] a cognition endowed with a particular appearance [of a given
object] might have an object different [from that objective appearance], because this would lead to
absurd consequences. [But] someone who is facing southwards while observing the surface of the
mirror sees his face turned northwards; in the same way, on the surface of a small mirror, the
reflection of his face is apprehended [as being] small although [his face is] big; and again, [the
reflected face] is perceived as being in contact with the surface of the mirror [and yet] as if it had
penetrated deep underneath [this surface]. Now, to begin with, the surface of the mirror has no
such depth; nor is the face or [any other reflected object really] in contact with the [surface].
Similarly, in the clear waters of a lake, [we] perceive the reflections of the tops of the trees
standing on the bank of the river[, and in these reflections we see] the [tree] tops as well as the
 highest parts of the branches bending downwards, whereas they do not exist in this way. For all
these reasons, the cognition of the reflection does not apprehend e.g. the own face [of the observer
standing in front of the mirror,] because it has an appearance [of object] that differs from the [face
itself], just as the cognition of a sound [differs from that of a visual form].”
40. See, for example, Bhatt (1962, pp. 23–25 and 45) and Taber (2010, p. 283).
41. See TS 252: bhavanmate hi nākāro buddher bāhyas tu varṇyate / na vivakṣitadeśe ca
gajavāsyādayaḥsthitāḥ// “For in your doctrine, the appearance [of object] does not belong to the
cognition: rather, [you] describe it as external [to consciousness]. And [in the case of erroneous
cognitions, the objects that we wrongly believe to exist in one place], such as e.g. the elephant
[perceived as located] in a blade [reflecting it], do not exist in the place about which [the
cognition] is.” (On my interpretation of the compound gajavāsyādayaḥ, see Ratié [2014c, p. 293,
fn. 1053].) Cf. TSP, vol. I, p. 130: kiṃca bhavato mīmāṃsakasya mate yo bhāsamānaḥsa ākāro na
buddheḥ, kintv asau bāhyārthasvabhāvo varṇyate, ākāravān bāhyo’rtho nirākārā buddhir iti
vacanāt. “Moreover, in your doctrine—you Mīmāṃsaka—the [objective] appearance that is
manifest [when we perceive] does not belong to the cognition: rather, [you] describe it as having as
its nature the external object, since [you] state that the external object possesses an appearance,
whereas the cognition is devoid of appearance (nirākāra).”
42. Other such contexts include the role of the sense organs in perception (see, e.g., the Nyāya texts
quoted above, nn. 20, 22, and 23), the existence of an intermediate state between death and
rebirth (as in the case of the AKBh passage quoted here: see above, n. 13) or the nature of speech
(see, e.g., Kumārila’s thesis quoted above, n. 26).
43. See, for example, Lamotte (1949, vol. I, pp. 357–360) and May (1959, p. 75, fn. 110). On the
various uses of the mirror simile in Buddhist literature, see also Wayman (1971 and 1974).
44. See, for example, MMK 23.9cd, according to which all perceptible forms (visual, auditory etc.) are
“like an [illusory] person [appearing thanks to some] magic trick, and similar to reflections”
(māyāpuruṣakalpeṣu pratibimbasameṣu ca). Cf., for example, Prasannapadā, p. 295: niḥsvabhāveṣu
sarvabhāveṣu pratibimbamarīcikājalālātacakrasvapnamāyendrajālasadṛśeṣu ... “With respect to all
things, which are devoid of intrinsic nature, [i.e.] similar to reflections, the water of a mirage, a
fire-brand [appearing as] a circle [when moved very fast], a dream, an illusion, a magic trick . . .”
45. As to whether this Buddhist idealism should be understood as simply stating that we cannot have
access to any external reality, or rather, as denying the existence of anything outside of
consciousness, see, for example, Kellner and Taber (2014); on the Śaivas’ view regarding this
issue, see Ratié (2014b).
46. On this analogy, see, for example, Taber (1994 and 2010, p. 281); see also Ratié (2010a, pp. 453–
460).
47. See, for example, TVBh, p. 126: tatpṛṣṭhalabdhena jñānena māyāmarīcisvapnapratiśrut-
kodakacandranirmitasamān sarvadharmān pratyeti. “Through the knowledge acquired
subsequently one knows all things to be similar to illusions, mirages, dreams, echos, the moon
[reflected] in the water or [fictional] creations.” Cf. the remark found in the Mahāyānasaṅgraha
that the image is meant as a way to explain how karmic law or phenomenal diversity can occur if
there is no external object (see Lamotte, 1938, p. 123), or MVBh, p. 66 and MVṬ, p. 220, which
specify that the reflection of the moon in the water “exists only [inasmuch as it is] an illusion”
(bhrāntimātrāstitva). It should be noted that some of these Yogācāra texts invoke the argument
(later developed by Naiyāyikas and Mīmāṃsakas!) that reflections do not exist as such because
when perceiving a reflection, in fact all we see is the reflected object (see Fukuda [2003, p. 264],
on this argument in the Sandhinirmocanasūtra and Asaṅga’s interpretation of it in the
Mahāyānasaṅgraha).
48. See Kritzer (2002, pp. 72–73).
49. On this thesis and its presentation by Abhinavagupta, see Ratié (2011b) and McCrea (2016, pp.
266–272).
50. Although these authors belong to the Madhyamaka (see, e.g., Seyfort Ruegg, 1981, pp. 87–93),
this affiliation remains mostly implicit in the TS and TSP, where the Vijñānavāda is presented as
the highest perspective (see, e.g., McClintock, 2010, pp. 85–91).
51. See the Vijñānavādin’s thesis formulated in TS 2078: vivādāspadam ārūḍhaṃvijñānatvād ato
matam* / advayaṃvedyakartṛtvaviyogāt pratibimbavat // [*matam conjecture: mataḥ TS.] “So [we]
consider that [the cognition] which constitutes the subject of this debate [i.e., ordinary
perception], involves no duality [between consciousness and an external object], since it is devoid
of [the duality between] being an object of knowledge and being a [knowing] agent; and this is the
case] because [this perception] is a cognition, just as [the cognition] of a reflection.”
52. See, for example, the objection (which might come from a Naiyāyika but is more probably to be
ascribed to Kumārila: see above, n. 37) put forward in TS 2079 in answer to the Vijñānavādin’s
thesis: nanu ca pratibimbe’pi jñānaṃsālambanaṃmatam / cakṣūraśminivṛttau hi svamukhādes
tathekṣaṇāt // “But even in [the case of] a reflection, [we] consider that the cognition has an
[external objective] support; for [it is] because the visual rays bounce back [that we] see e.g. [our]
own face thus [i.e., as being inside the mirror].” Cf. TSP, vol. II, p. 709: yasmān nāyanā raśmayo
darpaṇāditalapratihatā nivartamānāḥsvamukhādinā sambadhyante tatas te tathā
mukhādipratītihetavo bhavanti. ataḥsvamukhāder eva tathā
 darpaṇādyantargatādirūpeṇekṣaṇaṃbhavati. tataś ca na
pratibimbajñānaṃgrāhyagrāhakadvayarahitaṃsiddham. “Because the visual rays, when diverted
by the surface of [the reflecting entity] such as a mirror, bounce back [and] come into contact with
e.g. [our] own face, they are the causes [making us] apprehend e.g. [our own] face thus. Therefore
the perception [of the reflection] is [in fact the perception] of e.g. [our] own face itself thus, [i.e.] in
a form that is e.g. inside [the reflecting entity] such as the mirror. And therefore [contrary to what
you claim, you] have not established that the cognition of a reflection is devoid of the duality of the
apprehended object and apprehending subject [since the object of this cognition exists
independently of the cognition itself.]”
53. See, for example, Bhāmatī, vol. I, p. 21: evaṃvijñātṛpuruṣābhimukheṣv ādarśodakādiṣu svaccheṣu
cākṣuṣaṃtejo lagnam api balīyasā sauryeṇa tejasā
pratisrotaḥpravartitaṃmukhasaṃyuktaṃmukhaṃgrāhayad, doṣavaśāt taddeśatām
anabhimukhatāṃca mukhasyāgrāhayat, pūrvadṛṣṭābhimukhādarśodakadeśatām ābhimukhyaṃca
mukhasyāropayatīti pratibimbavibhramo’pi lakṣito bhavati. “In the same way, in limpid [reflecting
entities] such as a mirror or water [when they] face an observer, the visual rays, although in
contact [with the surface of the mirror or water], are diverted by a stronger sunray, [and] being in
contact with the face [itself], they make [us] grasp the face [itself; but] because of a defect [in this
process], they do not make [us] grasp where the face is [really] located and the fact that it does not
[really] face [us. In other words, in this case they make us] superimpose on the face [the property
of] being located in the mirror or water, which face [us] and which [we] have previously seen, and
[the property of] facing [us]—therefore this illusion of reflection too complies with the definition
[of superimposition just given, according to which one superimposes something previously seen on
some other locus].”
54. The Vedāntins seem to have been aware of this paradox. See, for example, BSBh, vol. II, p. 710,
highlighting the limits of the analogy in a nondualistic context: sūryādibhyo hi
mūrtebhyaḥpṛthagbhūtaṃviprakṛṣṭadeśaṃmūrtaṃjalaṃgṛhyate, tatra
yuktaḥsūryādipratibimbodayaḥ, na tv ātmā mūrto na cāsmāt pṛthagbhūtā viprakṛṣṭadeśāś
copādhayaḥ, sarvagatatvāt sarvānanyatvāc ca. “For water is grasped [as being] a material [thing]
distinct from [other] material [things] such as the sun, [and] located in a place distant [from them];
[and] a reflection of e.g. the sun can arise on [the surface of] that [water]. But the Self is not
material and the adventitious properties are not distinct from it and located in a place distant
[from it], because [the Self] is omnipresent and not different from anything.” Śaṅkara points out
that the analogy is nonetheless valid because analogies do not require a complete similarity
(sarvasārūpya) of the two relata involved (ibid.: yukta eva tv ayaṃdṛṣṭānto vivakṣitāṃśasambhavāt
... “However, this analogy is valid, because [an analogy] can be [only] partly relevant . . .”).
55. They also had frequent recourse to the analogy of reflections in a mirror to explain, for example,
the problematic relationship between the Brahman and the individual souls (jīva).
56. See, for example, BSBh, vol. II, p. 710: jalagataṃhi sūryapratibimbaṃjalavṛddhau vardhate
jalahrāse hrasati jalacalane calati jalabhede bhidyata ity evaṃjaladharmānuyāyi bhavati, na tu
paramārthataḥsūryasya tathātvam asti. evaṃparamārthato’vikṛtam ekarūpam api sad brahma
dehādyupādhyantarbhāvād bhajata ivopādhidharmān vṛddhihrāsādīn. “For the reflection of the sun
which is in the water gets larger when the [surface of the] water is larger, gets smaller when the
[surface of the] water gets smaller, trembles when the [surface of the] water trembles, is scattered
[into different parts] when [the surface of the] water is scattered—thus [this reflection] conforms
to the properties of the water, but in reality the sun does not exist in such a way. Similarly, the
Brahman, which in reality remains unchanged, seems to possess adventitious properties such as
getting larger and smaller, etc., although in fact it is unitary, because of its being concealed by
such adventitious properties as the body, etc.”
57. See, for example, BSi, p. 72: api ca bhedatvād eva viśvasya bhedo’bhedopādāna iti
śakyate’numātum. dṛṣṭo hi maṇikṛpāṇadarpaṇādiṣv abhinnamukhopādānas tadbhedaḥ. “Morover,
[we] can [legitimately] infer that the universe’s difference, precisely because it is a difference, has
as its basis nondifference. For the difference [affecting] a [face reflected] in jewels, swords,
mirrors and so on has as its basis the unitary face [thus reflected].” Cf. Tattvasamīkṣā ad loc., p.
133: yo yo’bhedaḥsa sarva īdṛśaḥ, yathā bimbābhedopādānaḥpratibimbānāṃbheda iti. “All
nondifference is such [that it is the basis from which we conceptualize difference,] just as the
difference of [various] reflections [of a single object] has as its basis the unity of the reflected
object.”
58. On this important Śaiva notion (interpreted and translated in sometimes strikingly different
ways), see, for example, Alper (1987, pp. 188 ff.), Torella (2002, pp. xxiii–xxv), and Ratié (2011a,
pp. 158–162).
59. On Utpaladeva’s thesis that imagination is a free creativity and that as such it constitutes the
essence of consciousness, see Ratié (2010b).
60. See in particular ĪPK 1.5.11: svabhāvam avabhāsasya vimarśaṃvidur anyathā /
prakāśo’rthoparakto’pi sphaṭikādijaḍopamaḥ// “[The wise] know that the nature of [conscious]
manifestation is realization (vimarśa): if this were not the case, although the manifesting
[consciousness] would have its appearance affected (uparakta) by the object, it would be similar to
an insentient [reflecting entity] such as a crystal.” See Torella (2002, p. 118). In his commentaries,
 Abhinavagupta insists that while such objects as a crystal can manifest other objects by bearing
their reflection, they remain insentient, contrary to consciousness, precisely because they are not
capable of realization. See ĪPV, vol. I, pp. 197–198 (as edited in Ratié [2011a, pp. 501–502]):
athānyenāpi satā ghaṭena yato’vabhāsasya pratibimbarūpā cchāyā dattā, tām asāv avabhāso
bibhrad ghaṭasyety ucyate, tataś cājaḍaḥ, tarhi sphaṭikasalilamakurādir apy evambhūta evety
ajaḍa eva syāt. atha tathābhūtam apy ātmānaṃtaṃca ghaṭādikaṃsphaṭikādir na
parāmraṣṭuṃsamartha iti jaḍas tathāparāmarśanam eva tarhy ajāḍyajīvitam
antarbahiṣkaraṇasvātantryarūpaṃsvābhāvikam avabhāsasya ... “If [you say that] the pot, while
being distinct [from its manifestation within consciousness], provides this manifestation with an
appearance consisting in a reflection (pratibimba), [and if you add] that the manifestation bearing
this [reflection] is what [we] call [the manifestation] of the pot, and that it is for this reason that
[this manifestation] is sentient—then [all reflecting objects] as well, such as a crystal, water or a
mirror, must necessarily be sentient since they are perfectly similar [to consciousness as you have
just described it]! If [you reply] that [reflecting objects] such as a crystal are insentient because
although they thus [bear a reflection of the object], they are unable to produce a realization
(parāmṛś-) either of themselves or of the [reflected objects] such as a pot, then [we must conclude
that] it is such a realization, which is the [very] life of sentiency, [and] which consists in freedom
with respect to the internal and external sense organs, that is essential to [conscious]
manifestation ...” On this famous passage see, for example, Alper (1987) and Ratié (2011a, pp.
495–507).
61. See the long passage on reflections in TĀ 3, or PS 12–13: darpaṇabimbe yadvan nagaragrāmādi
citram avibhāgi / bhāti vibhāgenaiva ca parasparaṃdarpaṇād api ca //
vimalatamaparamabhairavabodhāt tadvad vibhāgaśūnyam api / anyonyaṃca tato’pi ca vibhaktam
ābhāti jagad etat // “Just as, in the orb of a mirror, the variety of e.g. a city or a village is manifest
as if it were distinct [both] with respect to its own [components] and with respect to the mirror,
[although in fact] it involves no [such] distinction; in the same way, this universe is manifest [as
being] distinct [both] with respect to its own [parts] and with respect to the absolutely limpid
consciousness that is the Supreme Bhairava, although [in fact] it involves no [such] distinction.”
See Bansat-Boudon and Tripathi (2011, pp. 112–117).
62. See TĀ 3.8, where Abhinavagupta defines the property called limpidity that makes the
manifestation of reflections possible: svasminn abhedād bhinnasya darśanakṣamataiva yā /
atyaktasvaprakāśasya nairmalyaṃtad gurūditam // “Limpidity is nothing but the capacity of [an
entity such as a mirror] to manifest in oneself [a reflection that is apparently] distinct while not
ceasing to manifest oneself[; and is possible] because [the reflecting entity actually remains] one
[with the reflection]: this has been said by the Master.” Cf. TĀV, vol. II, p. 10: gurūditam iti guruṇā
paramaguruṇā śrīmadutpaladevenāthārthasya yathārūpam ity ādi, tathā na ca yuktaṃjaḍasyaivam
ity ādi śrīpratyabhijñākārikādvayaṭīkāyām etan nikhilam eva pratibimbasatattvam uditam uktam
ity arthaḥ. “‘[This has been] said by the Master’, [i.e.] by the master of [Abhinavagupta’s] master
[namely] the venerable Utpaladeva [in ĪPK 1.2.8—a verse] which begins with [the words]
athārthasya yathārūpam ...—and [in ĪPK 2.4.19—a verse] which begins with [the words] na ca
yuktaṃjaḍasyaivam ... This nature of reflections has been exhaustively (nikhilam eva) explained in
[Utpaladeva’s] detailed commentary (ṭīkā) on these two verses from the Pratyabhijñā [treatise]—
this is [what Abhinavagupta] means [here].” The commentary alluded to by Jayaratha is
Utpaladeva’s lost Vivṛti; as for ĪPK 1.2.8, it has to do with a rather different debate on the
relationship between the intellect and consciousness in Sāṃkhya (on this debate, see Ratié [2011a,
pp. 94–106 and 276–280]).
63. See ĪPK 2.4.19: na ca yuktaṃjaḍasyaivaṃbhedābhedavirodhataḥ/ ābhāsabhedād ekatra cidātmani
tu yujyate // “And [agency defined] in this way [as a capacity to transform oneself] cannot belong to
an insentient [entity], because [if it were the case] there would be a contradiction between
difference and identity, since [transformation involves] a difference between [various]
manifestations; whereas [such an agency] is possible in the unitary [entity] that has consciousness
as its essence.” See Torella (2002, p. 186).
64. The analogy, which was certainly stated at length in Utpaladeva’s lost Vivṛti (see above, n. 62 and
below, n. 93), also appears in his Vṛtti, p. 60: jaḍasyābhinnātmano bhedenāvasthiter virodhād
ayuktam, svacche cidātmany ekasminn evam anekapratibimbadhāraṇenāvirodhād yujyate.
“[Agency] cannot [belong to an insentient entity] because [if it were the case] there would be a
contradiction between the unitary essence of the insentient [entity] and the differentiated
existence [that transformation involves. However,] it is possible in a limpid (svaccha) unitary
[entity] that has consciousness as its essence, because in this case there is no contradiction
[between its unity and] its bearing various reflections.” See Torella (2002, p. 186). Abhinavagupta
explains that while an ordinary object cannot change without ceasing to be itself (a triangle that
would cease to have three angles would cease to exist), consciousness, like a mirror, is capable of
manifesting itself in different ways while remaining itself. See, for example, ĪPV, vol. II, p. 177 (as
edited in Ratié [forthcoming]): yat tu prameyadaśāpatitaṃna bhavati kiṃtu cidrūpatayā
prakāśaparamārtharūpaṃcidekasvabhāvaṃsvacchaṃtatra bhedābhedarūpatopapadyate.
anubhavād eva hi svacchasyādarśāder akhaṇḍitasvasvabhāvasyaiva
parvatamataṅgajādirūpasahasrasambhinnaṃvapur upalabhyate. “However, having [both of these]
 forms[, i.e.,] differentiation and undifferentiation, is possible in that which has not fallen into the
state of an object of knowledge, but rather, is limpid (svaccha) [because] its nature is one with
consciousness—[i.e., in that which] consists in the ultimate reality that is the manifesting entity
(prakāśa) because its nature is consciousness (cit). For it is through [immediate] experience that
[we] perceive the form of a limpid [reflecting entity] such as a mirror as divided into countless
[various] forms—a mountain, an elephant and so on—whereas its own nature remains intact.”
65. ĪPV, vol. II, pp. 177–178 (as edited in Ratié [forthcoming]): na ca rajatadvicandrādi yathā
śuktikaikacandrasvarūpatirodhānena vartate tathā darpaṇe parvatādi, darpaṇasya hi tathāvabhāse
darpaṇataiva sutarām unmīlati nirmalo’yam utkṛṣṭo’yaṃśuddho’yaṃdarpaṇa ity abhimānāt. na hi
parvato bāhyas tatra saṅkrāmati svadeśatyāgaprasaṅgād asya, na cāsya pṛṣṭhe’sau bhāti
darpaṇānavabhāsaprasaṅgāt, na ca madhye nibiḍakaṭhinasapratighasvabhāvasya
tatrānupraveśasambhāvanābhāvāt, na paścāt tatrādarśanād dūratayaiva ca bhāsanāt, na ca
tannipatanotphalitapratyā-vṛttāś cākṣuṣā mayūkhāḥparvatam eva gṛhṇanti, bimbapratibimbayor
ubhayor api parvatapārśvagatadarpaṇāvabhāse’valokanāt.
66. See ĪPVV, vol. III, p. 242: akhyātir eva cāpūrṇakhyātimayī bhrāntes tattvam ity upapāditam eva,
prakṛte tu nākhyātiḥkācid iti darśayaty ubhayor apīti darpaṇaparvatayoḥ. “And [we] have already
demonstrated that the essence of illusion is non-manifestation (akhyāti), that is to say, an
incomplete manifestation (apūrṇakhyāti); but as regards the subject at hand, there is no non-
manifestation whatsoever. This is what [Utpaladeva] shows [in his Vivṛti] with the words ‘both of
them [are manifest],’ [which mean] ‘both the mirror and the mountain [reflected in it are
manifest].’”
67. See ĪPVV, vol. III, p. 242: nanu darpaṇo’lpadeśo’lpaparimāṇaś ca, tadviparītas tu parvataḥ. “But
the mirror has a place and size that are small, whereas it is not the case of the mountain [reflected
in the mirror]!” As for Utpaladeva’s refutation of the objection, see ibid.: etan nirākaroti
deśaparimāṇavirodha eva nāstīti vadan vitatam apīty ādinā. “With [the passage of the Vivṛti
beginning with the words] ‘even though it is large,’ [Utpaladeva] refutes this [thesis] by explaining
that [in fact] there is no contradiction whatsoever between the places and sizes [of the reflected
object and its reflection].” This refutation was based on the fact that the place and size of the
reflected object remain the same (the mountain does not become small simply because it is
manifested in a mirror, nor does it recover its size when it is no longer reflected in the mirror). See
ibid.: tasya deśāder iti parvatadeśasya parvataparimāṇasya ca darpaṇadeśaparimāṇollaṅghanena
prasaraṇaṃyato nāsti. “With [the passage beginning with the words] ‘the place and so on,’
[Utpaladeva explains that there is no such contradiction] because the place and size of the
mountain [outside of the mirror] do not result from the fact that [the mountain] would abandon the
place and size of the mirror [once it is no longer seen in the mirror].” And this shows that the
reflection cannot be a mere illusion (ibid.): nanv aprasaraṇaṃcet, bhrāntatvam astu pītasyeva
śuklollaṅghanenāprasarataḥ. atrocyate’tirodhanāc ceti pītena tu śuklaṃtirodhīyata eva yato na
khyātiḥ. “[– Objection:] But if [the place and size of the mountain] do not result [from that, then]
let [us] admit that [the reflection] is illusory, just as the yellow [seen on a conch by someone who
has jaundice is illusory and] does not result from the fact that [the conch would really] abandon
[its being] white! To this [objection Utpaladeva] replies with [the passage beginning with] ‘and
because there is no concealment’; [that is to say: there is no concealment of the mountain by its
reflection,] whereas [in the case of the white mistaken for yellow,] because the yellow does conceal
the white, there is no manifestation [of the white, and this non-manifestation is the characteristic
of an erroneous cognition.]”
68. See, for example, PSV, p. 37: ... darpaṇas tattatpratibimbamayo’pi
tebhyaḥpratibimbebhyaḥsamuttīrṇasvarūpatayā cakāsti, na punas tanmayaḥsampadyate yena ca
na darpaṇa iti pratītiḥsyāt. sarvasya punas tattatpratibimbagrahaṇe’pi darpaṇo’yam ity abādhitā
pratipattiḥ... “... Although the mirror is made of various reflections, it is manifest as having a form
that transcends these reflections, and it does not amount to nothing but these [reflections, in]
which [case] as a consequence, [upon seeing reflections in a mirror we] would have the cognition
‘[this is] not a mirror’: on the contrary, even when [we] grasp various reflections, [we] all have the
cognition, uncontradicted [by some later cognition that would invalidate it], ‘this is a mirror.’”See
Bansat-Boudon and Tripathi (2011, p. 114). Cf. MMP, p. 157: nanūktam eva darpaṇe puruṣo nāstīti
bādho’nubhūyata iti cet, na; darpaṇe yadi puruṣa āropyate, tadā sa tatra nāstīti bādho’pi syāt,
naivam anubhavaḥ, tatra tacchāyāmātrasyaivopalambhaḥ, na tu tadvataḥpuṃsa ity
anubhūyamānatvāt. ataś ca darpaṇe puruṣo’stīti yuktimatāṃpratītyabhāvād etadbādho’pi nāsti.
“[The opponent] might [object the following:] But [we] have already stated that [the perception
apprehending the reflection of e.g. a man in a mirror is an illusion because we] experience that
[this cognition is] contradicted [by a later invalidating cognition taking the form] ‘there is no man
in the mirror’. [To this we reply]: no. If a man were [indeed wrongly] superimposed onto the
mirror, then there could also be [later] an [invalidating] contradiction [in the form] ‘this [man] is
not in the mirror’. But experience is not so, because [we rather] experience that what [we]
perceive is a mere reflection of the [man in the mirror,] but [in that mirror we do] not [perceive]
the man [himself] to whom this [reflection] belongs. And so, since sensible [people] do not have a
cognition [in the form] ‘there is a man in the mirror’ [when seeing the reflection of a man, they] do
not [experience] either any [later] contradiction of this [thought]!”
69. See TĀ 3.12–13: yas tv āha netratejāṃsi svacchāt pratiphalanty alam / viparyasya
svakaṃvaktraṃgṛhṇantīti sa pṛcchyate // dehād anyatra yat tejas tadadhiṣṭhātur ātmanaḥ/ tenaiva
tejasā jñatve ko’rthaḥsyād darpaṇena tu // “As for the proponent [of the thesis] that [when we
stand in front of a mirror,] the visual rays are diverted by the limpid [surface of the mirror] enough
to bounce back [and] grasp our own face, [we] ask him [the following question:] if one knows
thanks to this very visual ray [which is capable of leaving] the body of the Self that governs it [so
as to reside] in some other place, then what could be the use of the mirror?” Abhinavagupta does
not specify why mirrors become useless if the theory of visual rays is adopted. One might assume
that according to him, if we admit that visual organs can freely roam about outside of our body,
there is no particular reason why they should not go backwards by themselves, so that the mirror
becomes superfluous. According to Jayaratha, however, Abhinavagupta rather means that if the
theory of visual rays is adopted, then any object capable of resistance should make the visual rays
bounce back, and as a result we should see our faces reflected in walls as well as in mirrors. See
TĀV, vol. II, pp. 14–15: ... puraḥpratiphalanahetūnām anyeṣām api kuḍyādīnāṃtatra saṃbhavāt.
atha darpaṇādaya eva pratiphalanahetavo na kuḍyādaya iti cet, svacchandābhidhānam etat,
yataḥsamāne’pi pratighātahetutve darpaṇādaya eva tathā na kuḍyādaya ity atra na kiṃcin
nimittam utpaśyāmaḥ. “[Abhinavagupta says so] because in that case, other [things] too, such as a
wall, could be the causes of [the visual rays’] bouncing back towards [our face]. If [our opponent
replies]: ‘But only such [things] as a mirror can be the causes of the [visual rays’] bouncing back[,
and] not such [things as] a wall,’ [we answer that] this is idle talk, since, given that [both the
mirror and the wall] have in common the [property of] causing a resistance [which makes things
bounce back], we do not see any reason why such [things] as a mirror would be thus and not such
[things] as a wall!”
70. TĀ 3.14: viparyastais tu tejobhir grāhakātmatvam āgataiḥ/ rūpaṃdṛśyeta vadane nije na
makurāntare // “Let [us admit] that [when we look at our own face in a mirror,] the [entity]
grasping [the face] is the [visual] rays once they have bounced back; [but if this were the case, we]
should see the visual form in our own face [and] not in something else[, i.e.] the mirror.” This
criticism seems to echo Śāntarakṣita’s (cf. above, n. 39).
71. See ĪPVV, vol. III, p. 242 (quoting VP 1.100): yad āha: viruddhaparimāṇeṣu vajrādarśatalādiṣu /
parvatādisarūpāṇāṃbhāvānāṃnāsti sambhavaḥ// “As [Bhartṛhari] has stated: ‘[Objective] entities
having the form of e.g. a mountain cannot exist on the surface of such [reflecting entities as] a
diamond or a mirror, [since this surface] has a size that is contradictory [with objective entities
such as the mountain].’”
72. In the passage mentioned here Bhartṛhari merely shows that just as reflected entities seem to
take some of the properties of the reflecting entity but do not really possess them (the mountain
does not really become as small as the diamond reflecting it), speech seems to be temporally
differentiated whereas in fact it is not. See, for example, VPV, p. 101: upalabdhiviṣayatāpattau tu
teṣām abhinnakālānām upalabdhisthityabhimānaḥ. “But when these [linguistic realities] that do
not have a differentiated temporality become the objects of a perception, the belief [arises in
ordinary people] that these [syllables, words, sentences, etc.] have the [temporally differentiated]
existence of the perception [that grasps them].”
73. See, for example, ĪPVV, vol. III, pp. 243–244, where the interlocutor is explicitly designated as a
Vijñānavādin, and where Utpaladeva’s goal is still to show “the difference [between the
manifestation of reflections] and an illusion” (bhrānter vailakṣaṇyam, ibid., p. 244).
74. See his introduction to TĀ 3.12 in TĀV, vol. II, p. 13: ... keṣāṃcana naiyāyikānāṃpratyāvṛttair
nayanaraśmibhiḥsvasyaiva mukhasya grahaṇe’pi darpaṇe mukham* iti bhrāntir iyam na
punaḥsatyatvabhrāntatvavyatirekeṇa tṛtīyasya rāśyantarasyābhāvāt pratibimbaṃnāma kiṃcid
astīti mataṃnirākartum āha ... [*darpaṇe mukham conjecture: darpaṇamukham TĀV.]
“[Abhinavagupta now] states [the following verses] in order to refute the opinion of some
Naiyāyikas according to whom even though [when seeing one’s face in a mirror,] one grasps one’s
own face thanks to the visual rays that have bounced back, the [cognition] of the face [as being] in
the mirror is an illusion, and since there is no third option besides reality and illusion, what [we]
call a reflection is nothing at all.”
75. See TĀV, vol. II, pp. 13–14 on TĀ 3.12: ya ity ekavacanena sūtrakārāsūtritatvāt
sarveṣāṃnaiyāyikānāṃnaitan matam iti sūcitam, kaiścid eva hy āgrahapravṛttair etad uktam iti
bhāvaḥ. ata eva vṛttikārabhūṣaṇakārādibhir etan nāmāpi na spṛṣṭam. “With the [use of the]
singular in ‘the one who says ...’, [Abhinavagupta] hints at the fact that this is not the opinion of all
Naiyāyikas, since it has not been declared in any aphorism by the author of the [Nyāya]sūtras; for
this has only been stated by some whimsical ones—this is the implicit idea. For this very reason,
this [thesis] has not been mentioned at all by the author of the Vṛtti, [i.e., Jayanta Bhaṭṭa], or by
the author of the [Nyāya]bhūṣaṇa, [i.e., Bhāsarvajña,] etc.” On Jayanta Bhaṭṭa’s identity with the
“author of the Vṛtti,” see, for example, Steinkellner (1961, p. 159) and Gupta (1963, p. 15).
76. Pakṣilasvāmin is mentioned as “the author of the Nyāyabhāṣya” (nyāyabhāṣyakṛt) in ĪPV, vol. II, p.
84, and Uddyotakara’s name appears in ĪPVV, vol. I, p. 40.
77. On this thesis (certainly already formulated in Utpaladeva’s Vivṛti on ĪPK 1.2.8), see Ratié (2011a,
pp. 276–289).
78. See, for example, TĀ 3.9: nairmalyaṃmukhyam ekasya saṃvinnāthasya sarvataḥ/
 aṃśāṃśikātaḥkvāpy anyad vimalaṃtat tadicchayā // “Limpidity belongs fundamentally [and]
entirely to the unitary Lord that is consciousness; any other limpid [entity] is so only in some
respects, in part, and by virtue of this [Lord’s] will.” The reason for this is that limpidity is defined
in the Śaiva nondualistic system as a capacity to be manifest without being limited to any specific
form, and only consciousness is endowed with a capacity to manifest itself in countless forms (as
we all experience, for example, in imagination: see Ratié [2010b]). See in particular TĀ 3.48:
atyantasvacchatā sā yat svākṛtyanavabhāsanam / ataḥsvacchatamo bodho na ratnaṃtv ākṛtigrahāt
// “Absolute limpidity is [the ability] not to be manifest [only] in one’s own individual form; as a
consequence, only consciousness is absolutely limpid, whereas a crystal is not, because it is
possessed of an individual form.”
79. On the Naiyāyikas’ limpidity (prasāda) defined as the material property (i.e. according to most
interpretations, a particular color) distinguishing reflecting entities from other objects, see, for
example, Preisendanz (1989, pp. 193–199; 1994, pp. 587–589).
80. See, for example, TĀV, vol. II, p. 15 (immediately after arguing that the theory of diverted visual
rays does not explain the difference between mirrors and, for example, walls: see above, n. 69):
athātrādhikaḥsvacchatvākhyo dharmo’sti nimittam iti cet, naitat, svacchatvaṃhi na pratighāte
nimittam, evaṃhy ālokasya svacchatvāt tasmin sati nabhasi na kasyāpy avakāśaḥsyāt pratyuta tat
pratibimbagrahaṇe nimittam. “If [the opponent replies] that there is a cause for the [fact that only
things such as mirrors can reflect something else, i.e.] an additional property [that only belongs to
these things and is] called limpidity, this is not [right]; for limpidity is not the cause of resistance[,
contrary to what your theory of diverted rays presupposes]. For if it were the case, [then] since
there is [light] in the sky, given that light is limpid there should be no room for anything [but light
in the sky]! Rather, this [limpidity must be defined as] the cause of [our] apprehending a
reflection.”
81. See Ratié (2014c, pp. 210–218).
82. See, for example, Potter (1981, pp. 16 and 118); Granoff (1985, p. 465) mentions a similar story
(although in the latter Śaṅkara triumphs on his own and survives).
83. As regards the dualistic Śaivasiddhānta, a telling example is found in Sanderson (2006, pp. 68ff.)
which establishes a probable date for Sadyojyotis notably from the fact that he was not aware of
the doctrine of “illusionism” (vivartavāda) propounded by Śaṅkara and Maṇḍanamiśra, contrary to
later Saiddhāntikas who criticize both the latter and the earlier Vedāntic “transformationism”
(pariṇāmavāda). As for Śaiva nondualism, see ŚD 6.3 ff. where Somānanda attacks several types of
Vedānta.
84. Only one verse in Utpaladeva’s work explicitly criticizes the Vedānta’s monism, namely, ĪPK
2.4.20.
85. While presenting the Śaiva pratibimbavāda, Maheśvarānanda engages in a lengthy criticism of
the main Vedāntin theories regarding the ontological status of the phenomenal world in a passage
(MMP, p. 155) beginning thus: nanv asti vivartaḥpariṇāmādir veti cet, na ... “If [an opponent
objects the following to this doctrine that phenomena are similar to reflections:] ‘But there is an
illusory transformation (vivarta) [of the Brahman], or a real transformation (pariṇāma), etc.’, [we
reply:] No ...” See also the conclusion of this discussion in MMP, p. 159: iti vivartādivyati-
rekeṇāsmadupakṣiptaḥ-pratibimbapakṣa eva prauḍhim āḍhaukate. “Thus this thesis propounded
by us that [phenomena are similar to] reflections attains perfection when distinguished from [the
Vedāntins’ theory of] illusory transformation, etc.”
86. On the vedānticization of Abhinavagupta’s ontology by his commentator Bhāskarakaṇṭha
(seventeenth century?), see Ratié (2013).
87. See ŚD 6.11: pratibimbatayā cānye ye vā sargamukhe svayam / brahmaiva gṛhṇāty ātmānaṃato
bhedopapādanam // “And other [Vedāntins say that] it is the Brahman itself that grasps itself by
itself in the form of a reflection at the beginning of the [cosmic] emission; [and] this is how
difference is possible [despite the fact that only the unitary Brahman exists].”
88. Unfortunately this part of the ŚDV has not come down to us.
89. On the critique of Vedānta monism in ĪPK 2.4.20 and its commentaries, see Ratié (2011a, pp. 668–
713).
90. Modern scholarship has often assumed that Abhinavagupta knew Śaṅkara’s works (or even had
the “same” conception of ultimate reality: see Pandey [1936, p. 151]). According to Sanderson
(1985, p. 210, en. 41), however, in the Kashmiri Śaiva works of the tenth and eleventh centuries “it
is the doctrine of Maṇḍanamiśra which is generally in mind,” while “no source betrays familiarity
with the doctrines of Śaṅkara.” As noted in Bansat-Boudon and Tripathi (2011, p. 8, fn. 41),
determining whether they were acquainted with Maṇḍana’s tradition rather than Śaṅkara’s is no
easy task though. What is certain is that we should be wary of assuming that Śaṅkara’s works had
to be known by any posterior author somehow acquainted with Advaita Vedānta, because
Śaṅkara’s fame might well be a rather late phenomenon. Besides, as shown in Ratié (2011a, pp.
669–680; 2010b, pp. 364–369), Abhinavagupta’s commentaries sometimes seem to paraphrase
Maṇḍanamiśra’s BSi; and since Abhinavagupta at least (and maybe Utpaladeva as well) probably
knew some of Vācaspatimiśra’s works (see Ratié 2014a, p. 134, fn. 30), it is quite possible that one
of their main sources regarding Vedāntic ontology was the beginning (alas lacking to date) of
Vācaspatimiśra’s Tattvasamīkṣā.
91. See also nn. 104 and 106 below.
92. See, for example, ĪPVV, vol. III, p. 238: svacchatvasyāyaṃprabhāva ukto bhavati, na tu
pratibimbatāyāḥpratibimbādirūpasya svayam asāratvāt. “[Utpaladeva] explains that this is the
power of limpidity, as opposed to [the property of] being a reflection, because what consists in
such a [thing] as a reflection has no essence by itself.” Cf., for example, Tantrasāra, p. 10: ...
sarvam idaṃbhāvajātaṃbodhagagane pratibimbamātraṃpratibimbalakṣaṇopetatvād idaṃhi
pratibimbasya lakṣaṇaṃyad bhedena bhāsitum* aśaktam anyavyāmiśratvenaiva bhāti tat
pratibimbam ... [*bhāsitum conjecture: bhāsitam Tantrasāra.] “All these [objective] entities
[constituting the perceived universe] are nothing but reflections on the firmament of
consciousness, because they have the status that characterizes reflections. For this is the
characteristic of any reflection: that which, while it cannot be manifested independently, is only
manifest as being mixed with [a reflecting entity that appears to be] distinct [from it], is a
reflection.” Cf., for example, TĀV, vol. II, p. 65 (explaining TĀ 3.56 quoted below, n. 97):
tato’nyasmāt tadākāragrahaṇasahiṣṇor darpaṇāder bhedena pṛthak
svātantryeṇāśakyaṃbhāsanaṃyasya tat, tat paratantram ity arthaḥ. “Therefore [a reflection is]
that whose manifestation cannot occur ‘independently’, [i.e.] separately [or] autonomously with
respect to another [entity] such as a mirror, which is capable of taking on its aspect; i.e. this
[reflection] is heteronomous (paratantra).”
93. See the testimony of Utpaladeva himself in ŚDV, p. 14: viśvātmatvaṃca cinmayasya
pratibimbānām iva darpaṇaparamārthatvena bhāvānāṃsvacchacinmātrasatattvatayāvasthānāt.
etac ca sarvam īśvarapratyabhijñāṭīkāyāṃnipuṇam ālocitam. “And that which consists in
consciousness takes the form of the universe, because entities exist while having as their reality
nothing but a limpid consciousness, just as reflections [only exist] while having the mirror as their
ultimate reality; and [I]​ have exhaustively shown all this in [my] detailed commentary on [my]
Īśvarapratyabhijñā [treatise].” For somewhat different interpretations of this passage, see Nemec
(2011, pp. 117–118) and Ratié (2014a, p. 163, fn. 116).
94. ĪPVV, vol. II, p. 81 : darpaṇasya lagnatvena yathā pratibimbāni bhānti, na tu darpaṇābhedam
anādṛtya svātmapariniṣṭhitāni, tathāmī ghaṭādayas tasyaiva prakāśasya lagnās tadanujīvitā
nirbhāsante; na svaniṣṭhitā iti.
95. See, for example, TĀ 3.23c: na cāvastutvaṃsyān na ca kim api sāraṃnijam iti. “And [this
reflection] cannot be unreal; yet neither can [it] have any essence of its own.”
96. Abhinavagupta often explains that phenomena seem to be “over and above” consciousness or
“distinct from” it (adhika, lit. “[something] more”) although in fact they are not so, just as
reflections seem to be distinct from the mirror while their manifestation depends on their unity
with the mirror. See, for example, ĪPVV, vol. III, p. 343: etad eva hi tan nirmalatvaṃyat
svātmānatiriktasyāpy atiriktasyevāvabhāsanam. “For this is precisely what limpidity is: the
manifestation of [the reflecting entity] as if it were distinct, although [in fact] it is not distinct from
itself.” Cf., for example, ĪPVV, vol. II, p. 404: ... pratibimbakalpatayānatiriktair apy atiriktābhāsair
arthair nīlasukhādibhiḥ... “... Objects, such as blue or pleasure, which, in the manner of reflections,
appear as distinct [from consciousness] although [in fact] they are not ...” Abhinavagupta’s
disciple, Kṣemarāja, is particularly fond of this idea (see, e.g., Ratié [2011a, fn. 54, p. 668 and p.
693]).
97. See, for example, ĪPVV, vol. III, p. 239: yad vinā yan nīrūpam apṛthagbhūtam api
pṛthagbhūtatayeva nirbhāsate, tat tatra pratibimbam iti
pratibimbasāmānyalakṣaṇaṃcidapekṣayāsti bhāvānām. “An [entity A] which is unsubstantial
(nīrūpa) without an [entity B, and] is manifest as if it were distinct [from the entity B] whereas [in
fact] it is not distinct [from it], is a reflection inside that [entity B]; therefore [the analogy is
legitimate because] with respect to consciousness, [objective] entities have this characteristic in
common with reflections.” See also TĀ 3.56–57: anyavyāmiśraṇāyogāt tadbhedāśakyabhāsanam /
pratibimbam iti prāhur darpaṇe vadanaṃyathā // bodhamiśram idaṃbodhād
bhedenāśakyabhāsanam / paratattvādi bodhe kiṃpratibimbaṃna bhaṇyate // “What [we] call a
‘reflection’ is a manifestation that, because it is mixed with [an entity that appears to be] distinct
[from it], cannot exist independently of this [other entity]—just as a face in a mirror. The
manifestation which, from the supreme ontological principle (tattva) [from which cosmic
emanation proceeds to the last of them] is within consciousness, mixed with consciousness, [and]
cannot exist independently [of consciousness], why shouldn’t [we] call it a reflection?”
98. See, for example, Abhinavagupta’s treatment of this objection in TĀ 3.52–53: nanu bimbasya
virahe pratibimbaṃkathaṃbhavet / kiṃkurmo dṛśyate tad dhi nanu tad bimbam ucyatām //
naivaṃtallakṣaṇābhāvād bimbaṃkila kim ucyate / anyāmiśraṃsvatantraṃsad
bhāsamānaṃmukhaṃyathā // “ [– An objector:] But how could there be a reflection if there is no
reflected object? [– Abhinavagupta:] What shall we do [then]? For we do perceive this [reflection
that is the phenomenal world]!—But [then] let [us] call this [phenomenal world] a reflected object
—It [can]not [be] so; for how can it be said that [the phenomenal world] is a reflected object? For it
does not have the characteristics of this [reflected object, namely the properties of] existing [and]
being manifest while not being mixed with anything else [and] while being autonomous, contrary
to e.g. a face [that is manifest even in the absence of a mirror].”
99. On this demonstration, see Ratié (2011b).
100. darpaṇavidhiḥ (TĀ 3.23d).
101. This point was explained by Utpaladeva in his lost Vivṛti on ĪPK 2.4.19 while he expounded the
mirror analogy. See ĪPVV, vol. III, p. 244: nanu cidrūpābhedena prakāśanam asiddham
ahambhāvabahiṣkāreṇa nīlādīnāṃpratyayād ity āśaṅkyāha māyāpramātur iti ... “Having
anticipated the [following objection, Utpaladeva] states [the passage beginning with] ‘for the
subject [deluded] by māyā ...’ [in his Vivṛti]: ‘But [although reflections are seen not to be distinct
from the mirror manifesting them,] it is not established that [the phenomenal world] is manifest
while not being distinct from what consists in consciousness; for [objects] such as blue are
perceived as being external to [the perceiver’s] subjectivity [expressed as ‘I’, while objects are
apprehended as ‘this’].”
102. On this point made by Utpaladeva in a somewhat different context, see Ratié (2014a, pp. 161–
163). Cf., for example, PSV, p. 40: itthaṃparameśvarāpekṣayā svāṅganirmite bhāvarāśau na kācid
bhedabhrāntir māyāpramātrapekṣayā tu yo’yaṃbhedāvabhāsaḥ. eṣāpūrṇatvākhyātirūpā bhrāntiḥ,
pūrṇasyādvayātmano rūpasyākhyānam aprathā, pūrṇaṃna bhāsate, kiṃtv
apūrṇaṃdvayarūpaṃbhāsate; bheda eva pratīyata iti yāvat. tasmān
niravadyo’yaṃpratibimbavādaḥ. “Thus from the Highest Lord’s point of view there is no illusion
whatsoever of a difference [between consciousness and] all the [objective] entities that are created
out of parts of [consciousness] itself; but from the point of view of the subject deluded by māyā,
there is such a manifestation [of the universe] as different [from the consciousness reflecting it].
This [differentiated manifestation] is an illusion consisting in an incomplete manifestation. [That is,
this illusion is only] the non-manifestation of the full, [i.e.] nondual nature [of consciousness
manifesting the universe. This nature] is not fully manifest, rather it is manifest as having an
incomplete, [i.e.] a dual form; that is to say, only difference is apprehended [and not the unity on
which it rests]. Therefore this thesis that [phenomena are similar to] reflections (pratibimbavāda)
is blameless.” See Bansat-Boudon and Tripathi (2011, pp. 116–117).
103. See, for example, TĀ 3.44, which points out that the freedom of consciousness is manifest in the
very midst of the phenomenal universe reflected in it: tena saṃvittimakure viśvam ātmānam
arpayat / nāthasya vadate’muṣya vimalāṃviśvarūpatām // “Therefore the universe, while projecting
itself [as a reflection] on the mirror of consciousness, expresses this Lord’s limpid existence in the
form of the universe.”
104. On this idea in Maṇḍanamiśra’s BSi, see above, n. 57. Cf. TĀV, vol. II, p. 55, which also points to a
Vedāntin context (compare with, e.g., the BSBh passage quoted above, n. 56): iha khalu
rūpādayaḥpratibimbitāḥsantaḥsvādhāropādhivaiśiṣṭyenai-vāvabhāsante, yathā khaḍge
taddharmordhvatādyuparaktatayā mukham ... “In this [world], for sure, a visual form and so on,
while being reflected, are manifest only as being particularized by the adventitious properties of
their [reflecting] support, just as a face [reflected] in a sword [that is manifest] while having its
appearance affected by properties of this [sword] such as its [particular] length . . .”
105. According to TĀ 3.24–43, the objects of the five senses can be reflected (thus echoes, for
instance, are presented as reflections of sounds).
106. TĀ 3.45–46: yathā ca gandharūpaspṛgrasādyāḥpratibimbitāḥ/ tadādhāroparāgeṇa bhānti khaḍge
mukhādivat // tathā viśvam idaṃbodhe pratibimbitam āśrayet /
prakāśatvasvatantratvaprabhṛtiṃdharmavistaram //

Bibliographical references
Editions of Sanskrit texts
[AK = Abhidharmakośa] See AKBh.
[AKBh = Abhidharmakośabhāṣya] Abhidharṃ-koshabhāṣya of Vasubandhu, edited by P. Pradhan, K. P.
Jayaswal Research Institute, Patna, 1967.
[AKVy = Abhidharmakośavyākhyā] Sphuṭārthā Abhidharmakośavyākhyā, the Work of Yaśomitra [1936],
edited by U. Wogihara, Sankibo Buddhist Book Store, Tokyo, 1989.
[Bhāmatī] See BSBh.
[BSBh = Brahmasūtrabhāṣya] The Brahmasūtra Śaṅkara Bhāṣya, with the Commentaries Bhāmatī,
Kalpataru and Parimala, edited by A. Śāstrī, 2 vols [2nd edn], Nirnaya Sagar Press, Bombay, 1938.
[BSi] Brahmasiddhi by ācārya Maṇḍanamiśra with commentary by Śaṅkhapāṇi, edited by S. K. Sastri,
Madras Government Oriental Manuscripts Series 4, Madras, 1937.
[ĪPK = Īśvarapratyabhijñākārikā] See Torella (2002).
[ĪPV] Īśvarapratyabhijñāvimarśinī, edited with notes by M. R. Shāstrī/M. K. Shāstrī, Kashmir Series of
Texts and Studies 22 and 33, Nirnaya Sagar Press, 2 vols, Srinagar, 1918–21.
[ĪPVV] Īśvarapratyabhijñāvivṛtivimarśinī by Abhinavagupta, edited by M. K. Shāstrī, Kashmir Series of
Texts and Studies 60, 62, and 65, Nirnaya Sagar Press, 3 vols, Bombay, 1938–43.
[MMK] Nāgārjuna. Mūlamadhyamakakārikāḥ, edited by J. W. de Jong, The Adyar Library and Research
Centre, Madras, 1977.
[MMP = Mahārthamañjarīparimala] The Mahārthamañjarī with the Commentary Parimala of
Maheśvarānanda, edited by T. G. Śāstrī, Trivandrum Sanskrit Series 66, Trivandrum, 1919.
[MVBh] Madhyāntavibhāgabhāṣya. A Buddhist Philosophical Treatise edited for the first time from a
 Sanskrit manuscript by G. M. Nagao, Suzuki Research Foundation, Tokyo, 1964.
[MVṬ] Sthiramati—Madhyāntavibhāgaṭīkā. Exposition systématique du Yogācāravijñaptivāda, édition
par S. Yagaguchi, publiée sous les auspices du Keimeikwai, tome I, Nagoya, 1934.
[NBh = Nyāyabhāṣya] Gautamīyanyāyadarśana with Bhāṣya of Vātsyāyana, edited by A. Thakur,
Nyāyacaturgranthikā 1, Indian Council of Philosophical Research, New Delhi, 1997.
[NV] Nyāyabhāṣyavārttika of Bhāradvāja Uddyotakara, edited by A. Thakur, Nyāyacaturgranthikā 2,
Indian Council of Philosophical Research, New Delhi, 1997.
[NVTṬ] Nyāyavārttikatātparyaṭīkā of Vācaspatimiśra, edited by A. Thakur, Nyāyacaturgranthikā 3,
Indian Council of Philosophical Research, New Delhi, 1996.
[Prasannapadā] Madhyamakavṛttiḥ. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna, avec
la Prasannapadā, Commentaire de Candrakīrti, publiée par L. de la Vallée Poussin (Saint Petersburg,
1903–13), Biblio Verlag, Osnabrück, 1970.
[PS] The Paramārthasāra by Abhinavagupta with the Commentary of Yogarāja, edited by J. C. Chatterji,
Kashmir Series of Texts and Studies 7, Srinagar, 1916.
[PSV = Paramārthasāravṛtti] See PS.
Śāstradīpikā of Pārtha Sarathi Miśra, with the Commentary called Yuktisneha Pra-pūrani, by Pandit
Rama Krishna Miśra, edited by L. S. Dravid, Chowkhamba Sanskrit Series 188, 189, 190, 225, and
226, Benares, 1916.
[ŚD] Śivadṛṣṭi of Śrīsomānandanātha with the Vṛtti by Utpaladeva, edited by M. K. Shastri, Kashmir
Series of Texts and Studies 54, Srinagar, 1934.
[SDS] Sarva-darśana-saṃgraha of Sāyaṇa-Mādhava, edited with an original commentary in Sanskrit by
V. S. Abhyankar, Bhandarkar Oriental Research Institute, Poona, [1924], 1978.
[ŚDV = Śivadṛṣṭivṛtti] See ŚD.
[ŚV] Ślokavārttika of Śrī Kumārila Bhaṭṭa, with the Commentary Nyāyaratnākara of Śrī
Pārthasārathimiśra, edited and revised by G. S. Rai, Ratna Publications, Varanasi, 1993.
[TĀ] Tantrāloka of Abhinavagupta with commentary by Rājānaka Jayaratha, edited by M. K. Shāstrī,
Kashmir Series of Texts and Studies 23, 28, 29, 30, 35, 36, 41, 47, 52, 57, 58, and 59, 12 vols,
Allahabad/Srinagar/Bombay, 1918–38.
[Tantrasāra] The Tantrasāra of Abhinava Gupta, edited by M. R. Shāstrī, Kashmir Series of Texts and
Studies 17, Nirnaya-Sagar Press, Bombay, 1918.
[Tattvasamīkṣā] Vācaspatimiśra’s Tattvasamīkṣā. The Earliest Commentary on Maṇḍanamiśra’s
Brahmasiddhi, critically edited by D. Acharya, Franz Steiner Verlag, Stuttgart, 2006.
[TĀV = Tantrālokaviveka] See TĀ.
[TS] Tattvasaṅgraha of Ācārya Shāntarakṣita with the Commentary “Pañjikā” of Shri Kamalshīla, edited
by D. Śastri, 2 vols, Bauddha Bharati Series 1, Varanasi, 1968.
[TSP = Tattvasaṅgrahapañjikā] See TS.
[TVBh] Sthiramati’s Triṃśikāvijñaptibhāṣya. Critical Editions of the Sanskrit Text and Its Tibetan
Translation, by H. Büscher, Verlag der Österreichischen Akademie der Wissenschaften, Wien, 2007.
[VP 1 = Vākyapadīya, Chapter 1] Vākyapadīya of Bhartṛhari with the Vṛtti and the Paddhati of
Vṛṣabhadeva, Kāṇḍa I, edited by K. A. S. Iyer, Deccan College Monograph Series 32, Poona, 1966.
[VPV = Vākyapadīyavṛtti] See VP 1.
[Vṛtti = Īśvarapratyabhijñākārikāvṛtti] See Torella (2002).

Secondary literature
Alper, H. P. (1987), “svabhāvam avabhāsasya vimarśam: judgment as a transcendental category in
Utpaladeva’s Śaiva theology.” The Adyar Library Bulletin, 51, 176–241.
Bansat-Boudon, L., and Tripathi, K. D. (2011), An Introduction to Tantric Philosophy: The
Paramārthasāra of Abhinavagupta with the Commentary of Yogarāja. London: Routledge.
Bhatt, G. P. (1962), The Basic Ways of Knowing: An In-depth Study of Kumārila’s Contribution to Indian
Epistemology [Varanasi].Delhi: Motilal Banarsidass, 1989, revised ed.
Cox, C. (1988), “On the possibility of a nonexistent object of consciousness: Sarvāstivādin and
Dārṣṭāntika theories.” Journal of the International Association of Buddhist Studies, 11 (1), 31–87.
Frauwallner, E. ([1956] 1973), History of Indian Philosophy, vol. II: The Nature-Philosophical Schools
and the Vaiśeṣika System, the System of the Jaina, the Materialism, translated from the original
German (Geschichte der indischen Philosophie, II. Band: Die naturphilosophischen Schulen und das
Vaiśeṣika-System. Das System der Jaina. Der Materialismus, Salzburg, 1953) by V. M. Bedekar. Delhi:
Motilal Banarsidass.
Frauwallner, E. (1962a), Aus der Philosophie der śivaitischen Systeme, Deutsche Akademie der
Wissenschaften zu Berlin, Vorträge und Schriften 78, Akademie-Verlag, Berlin.
Frauwallner, E. (1962b), “Kumārila’s Bṛhaṭṭīkā.” Wiener Zeitschrift für die Kunde Süd- und Ostasiens
6, 78–90.
Fukuda, T. (2003), “Bhadanta Rāma: a Sautrāntika before Vasubandhu.” Journal of the International
Association of Buddhist Studies, 26 (2), 255–287.
Granoff, P. (1985), “Scholars and wonder-workers: some remarks on the role of the supernatural in
philosophical contests in Vedānta hagiographies.” Journal of the American Oriental Society, 105 (3),
 459–467.
Gupta, B. (1963), Die Wahrnehmungslehre in der Nyāyamañjarī, Verlag für Orientkunde Dr. H.
Vorndran, Walldorf-Hessen.
Hattori, M. (1968), Dignāga, On Perception, being the Pratyakṣapariccheda of Dignāga’s
Pramāṇasamuccaya. Cambridge, MA: Harvard University Press.
Jha, G. (1937), The Tattvasaṅgraha of Śāntarakṣita with the Commentary of Kamalaśīla. Translated into
English, 2 vols. Baroda: Oriental Institute.
Kaul, M. (forthcoming), “Can a reflected image exist separately outside the mirror? An exploration into
Abhinavagupta’s theory of reflection,” paper read at the conference Around Abhinavagupta: Aspects
of the Intellectual History of Kashmir from the 9th to the 11th Century, University of Leipzig, June
2013.
Kawajiri, Y. (2016), “New fragments of the Īśvarapratyabhijñā-Ṭīkā,” in B. Bäumer and R. Torella (eds.),
Utpaladeva, Philosopher of Recognition, pp. 77–101.
Kellner, B., and Taber, J. (2014), “Studies in Yogācāra-Vijñānavāda idealism I: the interpretation of
Vasubandhu’s Viṃśikā.” Asiatische Studien/Études Asiatiques 68 (3), 709–756.
Kritzer, R. (2000), “rūpa and the antarābhava.” Journal of Indian Philosophy, 28, 235–272.
Kritzer, R. (2002), “Unthinkable matters: the term acintya in the Abhidharmakośabhāṣya,” in Sakurabe
Ronshu Committee (ed.), Early Buddhism and Abhidharma Thought—in Honor of Doctor Hajime
Sakurabe on His Seventy-Seventh Birthday. Kyoto: Heirakuji Shoten, pp. 65–86.
Lamotte É. (1938), La Somme du Grand Véhicule d’Asaṅga (Mahāyānasaṃgraha), tome II: Traduction
et commentaire (reedition). Louvain: Institut Orientaliste, 1973.
Lamotte, É. (1949), Le Traité de la grande vertu de sagesse de Nāgārjuna
(Mahāprajñāpāramitāśāstra), 3 vols. Louvain: Institut Orientaliste.
Lawrence, D. P. (2005), “Remarks on Abhinavagupta’s use of the analogy of reflection.” Journal of
Indian Philosophy, 33, 583–599.
May, J. (1959), Candrakīrti—Prasannapadā madhyamakavṛtti. Douze chapitres traduits du sanscrit et
du tibétain. Paris: Adrien Maisonneuve.
McClintock, S. (2010), Omniscience and the Rhetoric of Reason. Śāntarakṣita and Kamalaśīla on
Rationality, Argumentation, and Religious Authority. Boston: Wisdom Publications.
McCrea, L. (2016), “Abhinavagupta as intellectual historian of Buddhism,” in E. Franco and I. Ratié
(eds.), Around Abhinavagupta: Aspects of the Intellectual History of Kashmir from the Ninth to the
Eleventh Century. Berlin: Lit Verlag, pp. 263–286.
Nemec, J. (2011), The Ubiquitous Śiva. Somānanda’s Śivadṛṣṭi and His Tantric Interlocutors. Oxford
and New York: Oxford University Press.
Padoux, A., and Ratié, I. (2013), “pratibimba,” in D. Goodall and M. Rastelli (eds.),
Tāntrikābhidhānakośa. A Dictionary of Technical Terms from Hindu Tantric Literature, vol. III. Wien:
Verlag der Österreichischen Akademie der Wissenschaften, pp. 513–514.
Pandey, K. C. (1936), Abhinavagupta: An Historical and Philosophical Study, second edn. Varanasi:
Chowkhamba Sanskrit Series, 1963.
Potter, K. (ed.) (1981), Encyclopedia of Indian Philosophies, vol. III: Advaita Vedānta up to Śaṃkara and
his Pupils. Delhi: Motilal Banarsidass.
Preisendanz, K. (1989), “On ātmendriyamanorthasannikarṣa and the Nyāya-Vaiśeṣika Theory of
Vision.” Berliner Indologische Studien, 4–5, 141–213.
Preisendanz, K. (1994), Studien zu Nyāyasūtra III.1 mit dem Nyāyatattvāloka Vācaspati Miśras II, 2
vols. Stuttgart: Franz Steiner Verlag.
Rastogi, N. (1984), “Some more Nyāyas as employed by Abhinavagupta.” Annals of the Bhandarkar
Oriental Research Institute, 65 (1), 27–42.
Ratié, I. (2010a), “The dreamer and the yogin—on the relationship between Buddhist and Śaiva
idealisms.” Bulletin of the School of Oriental and African Studies, 73 (3), 437–478.
Ratié, I. (2010b), “‘A five-trunked, four-tusked elephant is running in the sky’: how free is imagination
according to Utpaladeva and Abhinavagupta?” Asiatische Studien/Études Asiatiques, 64 (2), 341–
385.
Ratié, I. (2011a), Le Soi et l’Autre. Identité, différence et altérité dans la philosophie de la
Pratyabhijñā. Leiden-Boston: Brill.
Ratié, I. (2011b), “Can one prove that something exists beyond consciousness? A Śaiva criticism of the
Sautrāntika inference of external objects.” Journal of Indian Philosophy, 39 (4–5), 479–501.
Ratié, I. (2013), “Pāramārthika or apāramārthika? On the ontological status of separation according to
Abhinavagupta,” in N. Mirnig, P. Szántó, and M. Williams, Puṣpikā. Tracing Ancient India through
Texts and Traditions: Contributions to Current Research in Indology, vol. I. Oxford: Oxbow Books, pp.
381–405.
Ratié, I. (2014a), “A Śaiva interpretation of the satkāryavāda: the Sāṃkhya notion of abhivyakti and its
transformation in the Pratyabhijñā treatise.” Journal of Indian Philosophy, 42, 127–172.
Ratié, I. (2014b), “On the distinction between epistemic and metaphysical Buddhist idealisms: a Śaiva
perspective.” Journal of Indian Philosophy 42, pp. 353–375.
Ratié, I. (2014c), Une critique bouddhique du Soi selon la Mīmāṃsā. Présentation, édition critique et
traduction de la Mīmāṃsakaparikalpitātmaparīkṣā de Śāntarakṣita (Tattvasaṅgraha 222–284 et
Pañjikā). Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Ratié, I. (2016a), “Some hitherto unknown fragments of Utpaladeva’s Vivṛti (I): on the Buddhist
controversy over the existence of other conscious streams,” in R. Torella and B. Bäumer (eds.),
Utpaladeva, Philosopher of Recognition. Delhi: DK Printworld, pp. 224–256.
Ratié, I. (2016b), “Some hitherto unknown fragments of Utpaladeva’s Vivṛti (III): on memory and
error,” in E. Franco and I. Ratié (eds.), Around Abhinavagupta. Aspects of the Intellectual History of
Kashmir from the Ninth to the Eleventh Century. Berlin: Lit Verlag, pp. 375-400.
Ratié, I. (2016c), “In Search of Utpaladeva’s Lost Vivṛti on the Pratyabhijñā Treatise: A Report on the
Latest Discoveries (with the Vivṛti on the End of Chapter 1.8).” Journal of Indian Philosophy,
published online at http://link.springer.com/article/10.1007/s10781-016-9302-2
Ratié, I. (forthcoming), Abhinavagupta on causality: A critical edition and annotated translation of
Īśvarapratyabhijñāvimarśinī 2.4.
Sanderson, A. (1985), “Purity and power among the Brahmans of Kashmir,” in M. Carithers, S. Collins,
and S. Lukes (eds.), The Category of the Person: Anthropology, Philosophy, History. Cambridge, MA:
Cambridge University Press, pp. 190–216.
Sanderson, A. (1988), “Śaivism and the Tantric tradition,” in S. Sutherland (ed.), The World’s Religions.
London: Routledge and Kegan Paul, pp. 660–704.
Sanderson, A. (2006), “The date of Sadyojyotis and Bṛhaspati.” Cracow Indological Studies, 8, 39–91.
Seyfort Ruegg, D. (1981), The Literature of the Madhyamaka School of Philosophy in India. Wiesbaden:
Otto Harrassowitz.
Steinkellner, E. (1961), “Die Literatur des älteren Nyāya.” Wiener Zeitschrift für die Kunde Süd- und
Ostasiens, 5, 149–162.
Taber, J. (1994), “Kumārila’s refutation of the dreaming argument: the Nirālambanavāda-Adhikaraṇa,”
in R. C. Dwivedi (ed.), Studies in Mīmāṃsā. Dr. Mandan Mishra Felicitation Volume. Delhi: Motilal
Banarsidass, pp. 27–52.
Taber, J. (2005), A Hindu Critique of Buddhist Epistemology. Kumārila on perception. The
“Determination of Perception” Chapter of Kumārila Bhaṭṭa’s Ślokavārttika, Translation and
Commentary. London and New York: Routledge Curzon.
Taber, J. (2010), “Kumārila’s Buddhist.” Journal of Indian Philosophy, 38 (3), 279–296.
Torella, R. (1988), “A fragment of Utpaladeva’s Īśvarapratyabhijñā-vivṛti.” East and West, 38, 137–174.
Torella, R. (1992), “The Pratyabhijñā and the logical-epistemological school of Buddhism,” in T.
Goudriaan (ed.), Ritual and Speculation in Early Tantrism, Studies in Honor of André Padoux. Albany:
SUNY Series in Tantric Studies, State University of New York Press, pp. 327–345.
Torella, R. (2002), Īśvarapratyabhijñākārikā of Utpaladeva with the Author’s Vṛtti, critical edition and
annotated translation [Roma, 1994]. Delhi: Motilal Banarsidass, corrected edition.
Torella, R. (2007a), “Studies on Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part I: anupalabdhi and apoha
in a Śaiva Garb,” in K. Preisendanz (ed.), Expanding and Merging Horizons. Contributions to South
Asian and Cross-Cultural Studies in Commemoration of Wilhelm Halbfass. Wien: Verlag der
Österreichischen Akademie der Wissenschaften, pp. 473–490.
Torella, R. (2007b), “Studies on Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part II: What is memory?,” in K.
Klaus and J. U. Hartmann (eds.), Indica et Tibetica. Festschrift für Michael Hahn zum 65. Geburtstag
von Freunden und Schülern überreicht. Wien: Wiener Studien zur Tibetologie und Buddhismuskunde
66, Arbeitskreis für tibetische und buddhistische Studien Universität Wien, pp. 539–563.
Torella, R. (2007c), “Studies on Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part III. Can a cognition
become the object of another cognition?,” in D. Goodall and A. Padoux (eds.), Mélanges tantriques à
la mémoire d’Hélène Brunner. Pondichéry: Institut Français d’Indologie/École Française d’Extrême-
Orient, Collection Indologie 106, pp. 475–484.
Torella, R. (2007d), “Studies on Utpaladeva’s Īśvarapratyabhijñā-vivṛti. Part IV. Light of the subject,
light of the object,” in B. Kellner, H. Krasser et al. (eds.), Pramāṇakīrtiḥ. Papers Dedicated to Ernst
Steinkellner on the Occasion of His 70th Birthday. Wien: Wiener Studien zur Tibetologie und
Buddhismuskunde 70.2, Arbeitskreis für tibetische und buddhistische Studien Universität Wien, pp.
925–940.
Torella, R. (2013), “Inherited cognitions: prasiddhi, āgama, pratibhā, śabdana. Bhartṛhari, Utpaladeva,
Abhinavagupta, Kumārila and Dharmakīrti in dialogue,” in V. Elstschinger and H. Krasser (eds.),
Scriptural Authority, Reason and Action. Proceedings of a Panel at the 14th World Sanskrit
Conference, Kyoto, September 1st–5th 2009. Wien: Verlag der Österreichischen Akademie der
Wissenschaften, pp. 455–480.
Wayman, A. (1971), “The mirror-like knowledge in Mahāyāna Buddhist literature.” Asiatische Studien/
Études Asiatiques, 25, 353–363.
Wayman, A. (1974), “The mirror as a pan-Buddhist metaphor simile.” History of Religions, 13 (4), 251–
269.
 Part Three

Universals and Momentary Existence

 10
A Road Not Taken in Indian Epistemology: Kumārila’s Defense of
the Perceptibility of Universals
John Taber

One of the most striking aspects of the debate about universals in Indian philosophy, for
someone who is familiar with the treatment of universals in Western philosophy, is that
those who defended the reality of universals—I shall call them “realists”—insisted that
they were perceptible. Why, one might ask, would a philosopher who wants to defend the
reality of universals adopt this position as part of his/her strategy? For it seems prima
facie that universals are not perceptible. This was quickly pointed out by the Buddhists:
one does not distinctly apprehend the qualifier one attributes to something, be it a
universal, a quality, or a motion, in the same way one apprehends the stick one attributes
to “a man carrying a stick.”1 Indeed, it would seem that if universals were perceptible,
one could settle the dispute about the existence of real, as opposed to imaginary or
mentally constructed, universals simply by drawing attention to that fact. The question, in
other words, for the scholar with a background in Western philosophy approaching the
Indian debate about universals is this: Why did the realist take the bait offered by the
Buddhist? The criterion of reality for the Buddhist was the capacity to produce an effect
(arthakriyāsāmarthya) or causal efficacy, and the least causal efficacy a thing can have is
the capacity to produce a cognition.2 Why did the realist epistemologist go along with this
criterion, as if to say: We’ll show you that universals can cause cognitions of themselves!
Why, they can even produce perceptual cognitions!
This question occurs to the scholar who comes to Indian philosophy from Western
philosophy, because Western philosophers have typically taken a different view of
universals. They are not objects of perception but objects of the intellect, understanding,
Nous—whatever one wishes to call the non-sensory, “rational” faculty that apprehends
them; and, not being sensible things, they are not subject to change. As Plato put it, they
do not occupy a realm between being and nonbeing like changing sensible things, but
they are. They remain always the same, eternally themselves, and that is what it is to be
fully real. In other words, again, why did Indian realist philosophers accept as terms of
debate assumptions that put them at a distinct disadvantage in their endeavor to defend
the reality of universals, in particular, the assumption that for something to be real it must
be, let us say, concrete: located in time and space and able to affect other things, above all
to produce cognitions of themselves, ideally perceptual cognitions, through some natural
causal mechanism?
Is it possible that Indian philosophers simply were unable to conceive of abstract
entities, that they were, so to speak, stuck on the lower rungs of Plato’s epistemic ladder,
the Divided Line? This seems highly unlikely given that Indian linguistics and
mathematics are obviously concerned with abstract entities. Pāṇinian grammar describes
the behavior of linguistic units, their combination, deletion, and replacement, by means of
an algebraic calculus. Yet the recognition of abstract entities in these sciences did not
carry over into metaphysics and epistemology. Indian epistemology seems to be
relentlessly empiricist in orientation. To be sure, scripture (śabda, āgama) is widely
accepted as a means of knowledge of transcendent things, such as Dharma or Brahman,
but it is the senses that hold sway over the natural world. Although inference is almost
universally recognized as a means of knowledge as well, in Indian philosophy it seems to
work almost exclusively with empirical generalizations. It seems, upon an initial
examination—I put emphasis on the word “seems”—that no Indian system points clearly to
a higher faculty of a priori knowledge. This would determine the kinds of entities that are
accepted as real in Indian metaphysics. Although, to be sure, some Indian theorists
postulated supersensible entities—pradhāna, Īśvara (god), atoms, and so forth—they are
not by nature abstract or intelligible, like Plato’s forms. Nor is the natural world
populated by formal essences. Although things are said to have “natures,” a nature is not
usually depicted as the sort of thing that is comprehended when properly formulated in a
definition,3 as it has been in Western philosophy since Aristotle.4
Such tendencies of Indian philosophy, whether real or imagined, have been commented
on by others. Thus, as J. N. Mohanty (1992, p. 269) says:

There is a strong empiricist strand in Indian thought. This is testified by the primacy of perception,
by the importance of an “exemplifying instance” (dṛṣṭānta) in the syllogistic theory, and by a
conspicuous lack of modal thinking (“possible worlds,” “necessity,” etc.). However, some of the
ruinous consequences of empiricism are avoided by extending the scope of “perception” to include
intuitive grasp of universals and relations (and in some cases to extraordinary (alaukika)
perception of the yogis of all time, past, present, and future). As a matter of fact, even if the
philosophical positions were never classified as empiricism and rationalism (or their kin),
empiricism has a stronger claim in the Indian tradition than rationalism has. While the word
“experience” goes over, with some loss of meaning, into pratyakṣa (perception), the word “reason”
has no Sanskrit synonym. Buddhi may translate into “intellect,” but the principal epistemological
and metaphysical associations of “reason” are missing.5

In this chapter I would like to go deeper into the question why realist Indian
epistemologists chose to defend the perceptibility of universals. In my view, they were
compelled to do so by some very strong philosophical considerations that emerged from
the debate they were carrying on with the Buddhists about meaning and the nature of
inference. The particular realist philosopher I shall focus on is the seventh-century
Mīmāṃsā philosopher Kumārila, who I believe to be one of the giants of world philosophy;
in any case, I believe he was more than a match for his Buddhist opponents.
First, however, I would like to take a short detour into Western philosophy to remind us
of the emphasis placed on the imperceptibility, or conversely, the intelligibility of
universals. I cite below a couple of passages from Plato’s Phaedo. Plato, of course, is
talking in these passages about the forms, and there is some question about whether
Plato’s forms should indeed be considered universals. (Aristotle, for instance, in one place
refers to Plato’s forms as “particulars.” He also distinguishes his formal causes, which are
obviously derived from Plato’s forms, from universals.6) The most widespread view among
modern interpreters of Plato, however, is that his forms are universals. Indeed, the term
“Platonism” is virtually synonymous with the belief in universals.
The Phaedo concerns the immortality of the soul. In the first passage we shall consider,
Socrates is arguing against the view that the psyche (soul) after death could just be
dispersed upon separating from the body, like a puff of smoke, as his interlocutors
Simmias and Cebes fear. Is it indeed the kind of thing that is subject to disintegration?

(Socrates:) We must then ask ourselves something like this: what kind of thing is likely to be
scattered? On behalf of what kind of thing should one fear this, and for what kind of thing should
one not fear it? We should then examine to which class the soul belongs, and as a result either
fear for the soul or be of good cheer.
(Cebes:) What you say is true.
Is not anything that is composite and a compound by nature liable to be split up into its component
parts, and only that which is noncomposite, if anything, is not likely to be split up?
I think that is the case, said Cebes.
Are not the things that always remain the same and in the same state most likely not to be
composite, whereas those that vary from one time to another and are never the same are
composite?
I think that is so.
Let us then return to those same things with which we were dealing earlier, to that reality of whose
existence we are giving an account in our questions and answers; are they ever the same and in
the same state, or do they vary from one time to another; can the Equal itself, the Beautiful itself,
each thing in itself, the real, ever be effected by any change whatever? Or does each of them that
really is, being uniform by itself, remain the same and never in any way tolerate any change
whatever?
It must remain the same, said Cebes, and in the same state, Socrates.
What of the many beautiful particulars, be they men, horses, clothes, or other such things, or the
many equal particulars, and all those which bear the same name as those others? Do they remain
the same or, in total contrast to those other realities, one might say, never in any way remain the
same as themselves or in relation to each other?
The latter is the case, they are never in the same state.
These latter you could touch and see and perceive with the other senses, but those that always
remain the same can be grasped only by the reasoning power of the mind? They are not seen but
are invisible?
 That is altogether true, he said.
Do you want us to assume two kinds of existences, the visible and the invisible?
Let us assume this. (Phaedo 78b–79a; Cohen et al., 2011, pp. 286–287)

Here, Socrates attributes the invisibility of the forms to the fact that they are
apprehended by the intellect, “the reasoning power of the mind,” and not the senses.
Earlier in the dialogue this point came up with the first reference to the forms as what
philosophers above all seek to know:

(Socrates:) What about the following, Simmias? Do we say that there is such a thing as the Just
itself, or not?
(Simmias:) We do say so, by Zeus.
And the Beautiful, and the Good?
Of course.
And have you ever seen any of these things with your eyes?
In no way, he said.
Or have you ever grasped them with any of your bodily senses? I am speaking of all things such as
Bigness, Health, Strength and, in a word, the reality of all other things, that which each of them
essentially is. Is what is most true in them contemplated through the body, or is this the position:
whoever of us prepares himself best and most accurately to grasp that thing itself which he is
investigating will come closest to the knowledge of it?
Obviously.
Then he will do this most perfectly who approaches the object with thought alone, without
associating any sight with his thought, or dragging in any sense perception with his reasoning,
but who, using pure thought alone, tries to track down each reality pure and by itself, freeing
himself as far as possible from eyes and ears and, in a word, from the whole body, because the
body confuses the soul and does not allow it to acquire truth and wisdom whenever it is
associated with it. Will not that man reach reality, Simmias, if anyone does?
What you say, said Simmias, is indeed true. (Phaedo 65d–66a; Cohen et al., 2011, p. 274)

Thus, Plato declared the imperceptibility of universals in no uncertain terms at the

beginning of the Western philosophy, and this declaration has been immensely influential.
That universals are imperceptible has been the dominant view until modern times.
Universals, that is to say, are abstract and intelligible entities. By abstract is usually
meant non-spatiotemporal and causally inert. Now, obviously, Plato thought that forms
were causally efficacious in a specific sense. The locus classicus for this view is a later
passage in the Phaedo (99d–102a) where Socrates argues that the “safest” explanation for
why anything is what it is is that it “shares in” its form. Aristotle inherited this view from
Plato but, as is well known, modified it by bringing the forms down from a separate realm
and placing them in things, as their essences. Since modern times, however, when
philosophers realized they could do away with formal causes in science, abstract objects,
for example, mathematical objects and universals, have typically been considered to be
causally inert; and their being causally inert is connected with their being considered by
most philosophers not to be perceptible: they cannot produce cognitions in us by any
known perceptual mechanism. Therefore, if we know them, we must know them in some
other way. For Plato and the seventeenth-century rationalists it was evident that we arrive
at a comprehension of the essences of things through investigation, perhaps culminating
in a kind of rational intuition.
In more recent times universals have become problematic precisely because of their
non-spatiotemporality and causal inertness. With the ascendency of naturalism in
contemporary philosophy, that is, the acceptance of the view that all phenomena should
be analyzable by the methods of the natural sciences, universals, considered as abstracta,
have become questionable on the grounds that it seems impossible to give a naturalistic
account of how they are known.7 In such an environment, if one wished to defend the
reality of universals, it would not be unreasonable to attempt to defend the position that,
contrary to the traditional view, they are causally efficacious and perceptible. (Thus, they
are neither abstract objects nor exclusively intelligible.) Such a position in fact was
adopted by one of the leading modern defenders of universals, D. M. Armstrong.8
In what follows, then, I shall be sketching Kumārila’s views about the way in which
universals are cognized. The relevant passages are scattered across several chapters of
his Ślokavārttika (ŚV). I shall be drawing mainly from the Ākṛtivāda chapter, but I shall
also be referring to the Pratyakṣasūtra chapter, the Anumānapariccheda, the Vanavāda,
and the Apohavāda. None of these chapters has been critically edited, so what I offer here
is, for me,9 a first pass at understanding Kumārila’s arguments, with the assistance
however of the commentaries of Sucaritamiśra (where available),10 Jayamiśra, and
Pārthasārathimiśra. I have collected preliminary translations of a few of the verses I will
be discussing in an appendix. I shall also, in the course of my discussion, have occasion
here and there to draw attention to some of the ideas and arguments of the Buddhist
philosopher Dharmakīrti that seem particularly to clash with the sort of position Kumārila
is developing. I shall not go into the question of the chronological order of Kumārila and
Dharmakīrti here. My overall impression, however, is that Kumārila in his Ślokavārttika,
where he is addressing Buddhist views that seem to go beyond those of Dignāga,11 is
aware of developments that are already pointing in the direction of Dharmakīrti’s
system.12
The first verse of the Ākṛtivāda announces what is at stake for the Mīmāṃsaka in the
debate about universals.13 For the connection of word and meaning to be eternal—hence,
for language to be eternal, which is required for the Veda to be eternal—not only must
words be eternal but also their meanings. If the meaning of a word were something other
than the universal—if it were, say, all the particulars of a certain type (for instance, in the
case of the word “cow,” all particular cows)—then there could neither be a connection of
the word with them, since there are infinitely many particulars (they are “endless”); nor, if
such a connection were possible, could it be eternal, since particulars are perishable
(Ākṛti 1).14 The task of the Ākṛtivāda is to prove that such things as universals exist.
(Kumārila explains, Ākṛti 3, that the expression ākṛti, “shape” or “configuration,” used by
his predecessor Śabara, on whose text he is commenting, simply means genus [jāti] or
universal [sāmānya].) Later, in the discussion of Mīmāṃsāsūtra 1.3.33, it will be
established that universals are indeed what words mean (Ākṛti 2). Objections to the
existence of universals and their perceptibility will be dealt with in the Vanavāda chapter.
The refutation of the Buddhist view that the meanings of words are “exclusions” (apoha)
will be carried out in the Apohavāda, which immediately follows the Ākṛtivāda. Thus, with
the Vanavāda following the Apohavāda, we have a sequence of three chapters in the
Ślokavārttika devoted to the discussion of universals, Ākṛti-, Apoha-, and Vanavāda, and
then in Tantravārttika 1.3.33 a very extensive concluding defense of the idea that
universals, as opposed to particulars, are the meanings of words even though the Veda
enjoins actions involving individual objects. This is a considerable amount of material one
must master in order to present a complete picture of Kumārila’s understanding of
universals and how they are known. I will not attempt to do any such thing here.
We should note at the outset that the framework for the introduction of universals in
Indian philosophy—for this is true not only of Kumārila—was reflection on the question of
meaning. In Greek philosophy, however, the framework for the introduction of forms
(Plato) and formal causes (Aristotle), which came to be viewed as universals, was
reflection on first principles, the ultimate causes of things being what they are.15 This
difference in approach resulted, I believe, in universals being less evidently essences in
Indian philosophy. It could be argued that that is what allowed not just the Buddhists but
also Jaina philosophers to raise problems about how they could relate to the particulars
that have them.16 Nevertheless, I believe that Kumārila comes close to conceiving
universals as essences by arguing, apparently in part to counter the objection that
universals could not relate to particulars if they were different from them, that they in fact
are not different from them: universal and particular are united together in an entity
which is both universal and particular in nature. Unfortunately, it would take me too far
afield in this chapter to bring out all the aspects of Kumārila’s theory that suggest that
universals have for him some of the features associated with Aristotelian essences, but I
do indeed believe that to be the case—they have some of the features.
On one level, Kumārila suggests, universals are not controversial. Everyone agrees that
the referent of a word is “some universal” that is the basis of a cognition of many
particulars as a single thing—that is, of all of them as “cow” or “horse,” and so on (Ākṛti
4).17 The question is whether this “some universal” or other is a real universal as opposed
to, say, a causal property that all the otherwise completely distinct particulars happen to
share. For the Mīmāṃsaka, however, cognition itself is always the criterion. We should
take our experience at face value—for what else do we have to go by?18 And our
cognitions of objects, Kumārila maintains—we shall see that he attempts to introduce
specific evidence for this claim—present them as both distinct from other things and
common with other things: I cognize the cow as distinct from all other cows, yet also as
the same as other cows. “And that would not occur without dualness of nature,” that is,
without the things we cognize in that way really being both distinct and common (Ākṛti 5).
If a thing were just a particular, the cognition of it as a universal would not occur; and
vice versa, if it were just universal in nature,19 the cognition of it as a particular would be
without any basis (Ākṛti 6).20
Moreover, there is no reason to think that one or the other of the aspects of this
cognition, the particular aspect or the universal aspect, is false or “figurative,” that is,
that it does not literally belong to the object. The cognition of each is firm and constant,
never corrected by a subsequent cognition; so it cannot be an error (Ākṛti 7).21 However,
the particular could not be seen only figuratively as the universal or vice versa, because to
apply the concept of one thing figuratively to another one must have been directly
acquainted, on some previous occasion, with both; but that could not be the case if only
one or the other (i.e., particular or universal but not both) were real (Ākṛti 8ab).22 And in
general, Kumārila reminds us, he has shown in his refutation of idealism in the Śūnyavāda
chapter of the ŚV that our cognitions do correspond to real objects outside us; they are
not simply apprehensions of forms residing within our consciousness (Ākṛti 8cd). Contrary
to what the Buddhists are inclined to think, our cognitions accurately represent the world
around us and are valid unless and until they are overturned by other cognitions.
The veracity of our cognitions of objects as both particular and universal in nature,
however, can be supported by the basic metaphysical consideration that universals and
particulars/differences mutually depend on one another (Ākṛti 9).23 The universal is the
commonness of many particulars/differences, whereas particulars/differences are the
differentiae of a universal. Both, therefore, must always be found together. Kumārila even
formulates this as a syllogism: there cannot be a universal without a difference, just as
there cannot be a rabbit’s horn, because being without a difference, it would not be a
universal. Conversely, a difference without a universal would not exist, precisely because
it would be without a universal (Ākṛti 10).24 In other words, neither could occur without
the other because then it would cease to be what it is. Therefore, there cannot be a
complete difference of universal and difference (Ākṛti 11). (Obviously, they cannot be
completely identical either, as we shall see below.) Thus, our cognitions of objects as both
are confirmed.
Kumārila then proceeds to refute the view that the cognition of a single thing in regard
to many particulars that is in need of a cause—this single cognition we need to explain—is
caused by a particular capacity (śakti) of distinct individuals and not by a real universal.25
This is one of the passages in Kumārila’s discussion of universals where he appears to be
criticizing a theory that at least approaches one expounded by Dharmakīrti.26 In PV 1
Dharmakīrti famously develops the idea that the basis for applying the same word or
concept to different particulars could be because they differ from other particulars in
producing a certain effect. The type of effect he highlights as serving this function is “a
single judgment” (ekapratyavamarśa); that is, we would associate different particulars
with each other on the basis of their all producing a judgment having the same content,
such as “cow,” “horse,” and so forth.27 A universal could not produce such an effect, he
says, because a universal is eternal. In general, Dharmakīrti insists on the principle that
only momentary things can have causal efficacy.28 By virtue of producing this same effect,
multiple particulars that are really inherently different from each other can fall under the
same word or concept insofar as they have the same difference from particulars that do
not have that effect. Sameness is not a property of the things themselves. This is, roughly,
Dharmakīrti’s interpretation of Dignāga’s idea of apoha, exclusion, as the referent of
words and concepts. In fact, it is a bit more complicated than this, as we shall soon see.
With Ākṛti 12 ff. Kumārila poses problems for this type of theory (though not
necessarily Dharmakīrti’s theory). He who would maintain, he says, that particulars have
the capacity to produce a cognition of what is common (sāmānyabuddhi)29 must say
whether that capacity itself is grasped or not, and whether it is one or many. If it is
grasped and it is one, then it is just the universal by another name. If is not grasped but
somehow gives rise to a cognition of one thing nevertheless, then the latter cognition
would not have an object (nirviṣaya). The śakti could not be its object, because it does not
appear in the cognition. Nor, obviously, could the particulars be the objects of cognitions
of one thing, since they are diverse. Moreover, if the śakti differed for each particular,
then the apprehension of it would be tantamount to the apprehension of the particulars
(Ākṛti 12–16). “Therefore,” Kumārila concludes, “a universal, too [along with the
particular] is to be accepted, which is distinct from the particulars and their śaktis, which
is also common to all of them, and which inheres in each” (Ākṛti 17).
Thus, Kumārila has no patience for a theory that proposes that a cognition of one thing
can be generated by many things. At PV(SV) 1.107–109 Dharmakīrti suggests, I think, a
clever, if rather elaborate and possibly ultimately incoherent,30 way to get around this
objection: namely, particulars of the same type do not produce a cognition that has a
single appearance,31 but rather distinct cognitions that are apprehended as similar.32
Nevertheless, on the basis of producing indirectly, via the cognitions they cause, the same
(erroneous) judgment (ekapratyavamarśa), the cognitions they (those particulars) cause
are viewed as the same and those particulars in turn, insofar as they are judged as
different from other things that do not have that effect, are viewed as the same.33 In other
words, Dharmakīrti sees the cognition of the sameness of particulars as occurring on a
higher level than the cognitions immediately produced by them. It is the ascertainment
that they produce cognitions that in turn produce the same judgment that grounds the
belief that they are the same—if I understand him correctly. The term ekapratyavamarśa,
which Dharmakīrti uses but Kumārila does not, suggests an awareness that is conceptual
or proto-conceptual in nature, which does not mirror something that is in the
particulars.34
In the passage from Dharmakīrti that I have been discussing, he really seems to be
directly answering Kumārila.35 Kumārila returns, however, to the problem of ekabuddhi,
the cognition of one thing produced by many particulars, at Ākṛti 31cd–38. Here, the
question is why universals are manifested only by certain particulars, not all. Why do we
see cowness just in cows and not also, say, in chairs? There must be something
responsible for causing that manifestation only in those things; what is it? Presumably
such a cause would have to be universal in nature itself; for it is unlikely that some
particular could restrict a universal to certain particulars. But if another universal has to
be postulated to account for a universal’s connection with only certain particulars, hence
its manifestation only by them, then one will have to postulate yet another universal to
account for the connection of the second universal with the particulars, and we are off on
a regress.36 Kumārila, however, suggests that the connection of the particular with the
universal is “natural” (svābhāvika) and not externally caused (Ākṛti31cd–32ab). That is, as
Jayamiśra puts it, it is due to its own causes that a certain sort of particular arises such
that it is connected with a certain universal.37 (Here, by the way, is one place in
Kumārila’s discussion where the universal starts to seem like an essence.)38 Kumārila goes
on to say that just as some think a thing’s nature can account for its capacity (śakti) to
produce a certain idea, so could its nature account for the connection of the particular
with a certain universal (Ākṛti 32cd).39 Conversely, a thing’s nature can explain why it
does not manifest a certain universal; no other cause has to be sought (Ākṛti 34).40 The
(Buddhist) opponent, however, at this point is allowed to come back with the following
objection:

If it is not held that another universal is the cause of restricting its occurrence [in certain
particulars], why is the cognition “cow” not restricted [to certain particulars] without cowness?
Just as there is the occurrence or non-occurrence of cowness,41 etc., only in certain [particulars],
even though they are equally different, so the cognition [“cow”, etc.] will arise [in regard to certain
particulars] even without [another] cause [i.e., just due to the nature of the particulars]. (Ākṛti 35–
36)42

To this, Kumārila responds: Although the manifestation of a universal only in certain

particulars is due just to the nature of those particulars, the cognition that has as its
content “cowness” as something distinct from the particular cannot occur without an
object, namely, cowness that is distinct from the particular. Indeed, we believe that there
is a general property distinct from the particular; so surely there is such a thing!
Otherwise, such a cognition would just be false; but there is no evidence, in the form of a
sublating cognition, that it is (Ākṛti 38ab).43

Indeed, the arising of cognitions without an object is not accepted. They44 hold [rather] there to be
a universal distinct from the particular. Therefore, certainly, there is that. (Ākṛti 37)

This seems to be Kumārila’s main argument, already announced at Ākṛti 5–6. We have a
clear cognition of an object not only as a particular but also as something extending over
many particulars that is not a particular. This cognition, if it is true, requires a proper
object, precisely the sort of thing it represents. If it were not true, it would somehow be
overturned, cancelled by another, more accurate cognition. Since that does not occur,
however, the thing it represents, a universal distinct from particulars, must exist.
Clearly, this argument is an application of Kumārila’s theory of intrinsic validity,
svataḥprāmāṇyam: a cognition presents itself as true unless and until it is proven invalid
by another cognition or cognitions that initially present themselves as true.45 Although
central to Kumārila’s philosophy—and already evident in the thought of Kumārila’s
predecessors, so he must have been aware of it—this is a theory that Dharmakīrti for
some reason chose not to extensively attack (he may well have been sympathetic to it). As
for the cognition of an object as the same as others, that is, as having an aspect that is
common, he may well have thought that it is invalidated by the consideration that a
universal, being eternal and not subject to change, could not possibly produce a cognition
of itself. In any case, it is this sort of consideration that is brought up in the modern
debate about abstract objects, as already mentioned: not being located in time or space,
they would be causally inert. I do not at this time know of any passage where Kumārila
addresses specifically this objection, but we can see that his theory potentially provides
an answer to it. If the universal really is not different from the particular, and the
particular is located in time and space and so can causally influence other things, then so
can the universal.46
It is, rather, a different sort of defeating consideration that Kumārila takes up toward
the end of his Ākṛtivāda. Something cannot be both universal and particular in nature,
because those two kinds of things conflict: a universal extends over many things while a
particular is located in only one place (Ākṛti 51ac).47 If they were both the nature of the
same thing, the universal would become a distinct thing different from other things, like a
particular, and the particular would become just one thing (extending over many) (Ākṛti
52).48 How is it possible that a thing is both one and many, both other and the same (Ākṛti
53ab)?49 In short, universal and particular are contradictory in nature; so a cognition that
apprehends the same thing as both must, once again, be either false or “figurative,” that
is, not literally true (Ākṛti 51d).50
Kumārila’s response to this is, essentially, to say that our experience should trump our
analysis. One cannot say that there is a contradiction between universal and particular
absolutely. When the particular is known as not different from the universal, then indeed
it does not occur in just one place; and, insofar as it is identical with the particular, the
universal does not occur in other places (Ākṛti 54–55ab). The problem of being both same
and different can be explained in the same way: something can be one by virtue of one of
its natures, many by virtue of the other (Ākṛti 55cd–56ab).51 Just as one can choose to
separate out this or that color of a variegated object, one can focus on the universal or
particular aspect of an object (Ākṛti 57cd–58ab).52 When the thing is apprehended just as
a particular, the universal continues to exist “in conformity with that cognition”; that is to
say, it does not cease to exist but, overcome by the particular aspect, it is simply not
evident (Ākṛti 60cd–61ab).53 Likewise, when one knows the particular as identical with the
universal, then the fact of being just the universal is cognized (Ākṛti 61cd–62ab).54 “When,
however, the diverse/variegated thing”—that is, the thing possessing both universal and
particular aspects—“is simultaneously cognized”—in an initial non-conceptual perception
that precedes the conceptual one—“then indeed everything [namely, that it is] other, not
other, particular, etc., disappears” (Ākṛti 62cd–63ab).55
I do not know if Dharmakīrti anywhere critiques precisely this theory, which appears to
be a variety of what is usually called “perspectivism” or “perspectivalism.”56 He is,
however, occupied with a related problem at the beginning of the apoha section of the
first chapter of his Pramāṇavārttika. There, he asks, how do we know that an exclusion is
what is made known by a word or an inferential mark? Because, he answers, a subject of
inference (dharmin) or the referent of a word, which is already known in its entirety in a
positive way by perception—for the particular is one (eka), without parts (niraṃśa)—
remains to be known only in the ways in which it differs from other things. That is to say,
it is in order to exclude properties that could be erroneously attributed to or
“superimposed” upon the object that one endeavors to comprehend it, by means of
inference or language, as different from other things in certain respects.57 As he says, “As
many extraneous natures as this [object] has [in contrast to other things], there are that
many [possible] superimpositions” due to a deficiency in the conditions of
ascertainment.58 Thus, the conceptual awareness of the object as having a certain
property actually pertains to the exclusion of the object from others.59 For Dharmakīrti
and Kumārila both, a non-conceptual, perceptual awareness grasps the object in its
totality, but not explicitly. For Dharmakīrti, properties surface in our awareness by a
mental process of differentiating the object from others with which it could be confused,
whereas for Kumārila different aspects of the object, both particular and universal,
emerge when one focuses on one or the other. For Dharmakīrti, the properties are
dependent on the activity of the cognizer60; for Kumārila, they are in the object, waiting
for us to direct our attention to them. The contrast between realism and anti-realism
could not be starker.
Obviously, I cannot present all of the arguments for the existence of universals
Kumārila develops in his Ākṛtivāda. It is far too rich a text. It merits an entire monograph,
which should include not just a new edition and translation but also a philosophical study,
much deeper and more thorough than what I have attempted here. I have barely
scratched the surface. I shall move on to consider a couple of passages from the Vanavāda
that support the doctrine of the perceptibility of universals that I have been highlighting.
In the passage of Śabara’s commentary on Mīmāṃsāsūtra 1.1.5 that Kumārila is
discussing in his Vanavāda, Śabara61 makes the assertion that a universal62 is not
something to be proved because it is perceptible.63 Kumārila explains that the purpose of
this statement is to reject any inference others might introduce to refute the existence of
universals (Vana 10cd).64 For example, some have argued that there can be no universal
separate from individuals, because when the latter are not cognized a universal is not
cognized, just like a herd, an army, or a forest is not cognized when there is no cognition
of the individual things—cows, soldiers, and trees—that make it up (Vana 11). Since
universals are known to everyone there is no point in demonstrating their existence to
those who accept them. Nevertheless, “a contradiction according to common
usage/practice,” pointing out that we perceive them, should be stated—as Śabara in fact
does—for the benefit of those who perversely attempt to refute them (Vana 12).65 Having
affirmed that we perceive universals, any attempt to prove their nonexistence will not
even be able to get off the ground, the very thesis of any such proof being already
contradicted by what is accepted by ordinary people at the outset.
An opponent, however, is allowed to object that there would be no dispute about
universals if they were perceptible (Vana 13ab)! To which Kumārila responds: Those
logicians (tārkika) who are inclined to dispute universals also dispute such things as the
pramāṇas: their definitions, how many there are, and so on. If the mere fact that there is a
dispute about something meant that it does not exist, there would not even be pramāṇas
(Vana 13cd)!66 Such people—Buddhists, of course—even dispute the existence of material
form, and so forth, that is, physical objects, which are considered by everyone else to be
perceptible (Vana 14ac’). As Sucarita puts it, the thoughts of stupid people are
uncontrolled.67 Thus, the fact that there is a dispute about universals has nothing to do
with whether or not they are perceived. Indeed, ordinary people do not at all dispute
universals (Vana 14c’d). They are always taking action based on the ascertainment of
things of certain types that fulfil certain prescriptions, for instance, giving sour milk,
buttermilk, and so on to Kauṇḍinya Brahmins (Vana 15). This is one of the ways in which
Kumārila implies that universals have specific causal properties: it makes a difference
what kind of gift one gives to a priest!
A bit further on in the Vanavāda Kumārila actually does offer supporting evidence, not
for the existence of universals, but for their perceptibility. I am sure the irony will not be
lost on the reader that, all the while insisting that universals are perceptible, Kumārila in
the end has to concede that their perceptibility is not as perceptible as he would like it to
be and resorts to an argument to bring that fact to our attention. Namely, he points out
that when we see mung beans, māṣa beans, or sesame seeds—apparently he is thinking of
a pile of beans one might see in the market—we do not perceive the individual beans, but
rather “a universal which is an object of the senses is apprehended by means of a
cognition of one thing (ekabuddhinirgrāhyā jātir indriyagocarā),” that is, as Sucarita
explains,68 by an awareness that presents a single form (ekākārasaṃvit) (Vana 24).69
Similarly, one sees a man in the distance and wonders whether he is a Brahmin or some
other kind of man. That would not be possible if one did not apprehend “with the sense of
vision, etc.” the universal humanity without apprehending the difference (Vana 25). So
much for the stock Buddhist argument against universals, that there are no real
universals distinct from particulars because one does not cognize them when one does not
cognize the particulars!
I am not sure these are very good examples, however. If I were at the market and saw a
pile of beans, I would not be able to identify them as any type at all, mung, lentil, or
whatever. To me, they would just be “beans.” But I think the general point Kumārila is
getting at is valid: we can indeed see that something has a certain property before we see
what particular thing it is. That is a church I see across the Neckar, which I recognize by
its spire. I do not know if it is Catholic, Lutheran, or Baptist. I see a type of object first—
even if I lack the word for it—without apprehending the particular.
One of the somewhat surprising aspects of Kumārila’s theory of universals revealed by
these two examples, which was not fully evident until now, is that he believes universals
are perceived by the external senses, typically the sense of vision. Thus, we cannot
ascribe to him belief in some kind of mental perception (mānasapratyakṣa), which could
possibly be intellectual in nature, perhaps a kind of intellectual intuition, thereby
assuaging our concern (if one had such a concern) that he was insufficiently Platonist.
Kumārila does not take the route of Jayanta Bhaṭṭa, who did hold, at least some of the
time, that universals are perceivable by a “mental perception.”70 He is, rather, closer to
Praśastapāda in thinking that universals are made accessible immediately (or at least in
the second moment) in the act of perceiving objects with the external senses.71
Having satisfied ourselves that Kumārila does believe that universals are perceptible,
we must finally return to the question posed at the beginning: Why would Kumārila take
this path? Why did he emphasize the perceptibility of universals and not their
intelligibility? Why did he seem to go along with the Buddhist assumption that only that
which is causally efficacious is real; hence, to establish the existence of universals, one
must show that they are at least capable of producing cognitions of themselves, even
perceptual cognitions that arise upon the initial encounter of one’s faculties with the
object, like particulars? Why did he think that our conceptually cognizing universals had
to be preceded by a perceiving of them, and that it is the perceiving of them that assures
us that they really belong to the things we attribute them to? The perceptibility of
universals is, as I said before, a prima facie implausible position, despite Kumārila’s
protestations to the contrary. I think we have seen that he has to use considerable
ingenuity to make it seem at all attractive. He had to have had a deeper motive for
adopting this view. What was it?
I think the first thing to be said in answer to these questions is that, above all else, as I
have tried to argue, Kumārila was a realist, and the criterion of causal efficacy is
essentially a realist criterion. One of the foremost modern defenders of universals, D. M.
Armstrong, appeals to this same criterion in distinguishing real properties from putative
ones. He refers to it as “the mark of being” and finds it, not in any Buddhist text of course,
but in a statement made by the Eleatic Stranger in Plato’s Sophist:

I suggest that anything has real being that is so constituted as to possess any sort of power either
to effect anything else or to be effected, in however small a degree, by the most insignificant
agent, though it be only once. I am proposing as a mark to distinguish real things, that they are
nothing but power. (247d–e; quoted in Armstrong, 1978, p. 46)

This is an altogether plausible criterion of being, though of course not the only option,
that Kumārila could very well have accepted without giving voice to it, and—since it really
is a criterion of common sense72—without having known Dharmakīrti’s formulation of it.73
A more specifically epistemological reason for insisting on the perceptibility of
universals that Kumārila is quite explicit about, however, is that universals must not only
exist, but they must be perceivable in order for language and inference to function.
Kumārila asserts the necessity of universals for language and inference at Ākṛti 39: if
there were no universals language and inference could not function, “for there could not
be a connection with particulars, because they are infinite.”74 Language operates on the
basis of a known connection of a word with its meaning and inference on the basis of a
known connection of the inferential mark with what possesses it. If a word or an
inferential mark referred to or indicated particulars, its relation with what it means or
what it indicates could never be learned, since the particulars that fall under what a word
means or what an inferential mark indicates are unlimited. Of course, the Buddhists
thought they could solve this problem without positing universals: words and inferential
marks are related, not to particulars, but to apohas, exclusions. Apohas can play the role
of universals insofar as they extend over many particulars, but they are not “real things”
(vastu); they are absences or non-entities (abhāva), that is, negations. Kumārila, however,
in his Apohavāda poses the same problem for apohas: if they were the meanings of words
and what inferential marks indicate, how could one learn the connection between words
and what they mean and inferential marks and what possesses them? For it seems that we
cannot cognize apohas independently of words and inferential marks.
Kumārila develops this argument at Apoha 71–82, which I shall summarize here rather
freely, deviating somewhat from the sequence of the verses. Kumārila asks, how is an
apoha known? He takes the Buddhist position to be that an apoha is known only by means
of language or inference, since it is not a real thing; certainly it could not be known by
perception. A word or an inferential mark can only make something known, however, if
there is an established relation between it and the thing it indicates. If an apoha is not
somehow evident independently of language and inference, then the relation between a
word and its meaning—the apoha—or an inferential mark and what it indicates—again, an
apoha—could never be established in the first place so that a word or inferential mark
could make it known (Apoha 73–74).75 The prospects for a direct apprehension of an
apoha, meanwhile, are not promising, either. Knowing a specific apoha distinct from other
apohas, Kumārila suggests, would entail either knowing all the things it is based on, that
is, all the things that fall under the exclusion (i.e., in the case of the word or inferential
mark “cow,” all cows), as non-different (aviśeṣataḥ), or all the things that are excluded
(i.e., all non-cows, i.e., horses, lions, camels, etc.) as non-different. One could only know
that all the things it is based on or all the things that are excluded are non-different
insofar as they share the same property, and that of course would amount to knowing a
universal (Apoha 71–72).76
Now Dignāga’s view was actually that one learned the meaning of a word, that is, the
apoha, by observing what the word is not used for. Simply by observing that the word
“cow” is not used for horse, lion, camel, and so on—that is, non-cows—one establishes
that it refers to what is not a non-cow, namely, cows.77 But, how is this actually supposed
to work in practice? Kumārila points out, if there is no positive concomitance established
for a word and what it does mean—if, that is to say, there is no example at all of what the
word “cow” refers to—and one only noticed what it is not applied to, then in fact
everything would be excluded by the apoha. A word would mean nothing (Apoha 75). If
one did, on the other hand, take a particular type of cow, say a spotted cow, as the
example of what is not a non-cow, then other types of cow—black cows, white cows, and
so on—along with horses, lions, and camels would be different from that and not be
included in the meaning of the word “cow” (Apoha 77). If one learned that “cow”
expresses the exclusion of non-cow on the basis of not observing that it is employed for
anything other than one specific, individual cow—perhaps the cow that the teacher is
pointing to—then everything other than that particular would be excluded and the word
“cow” would refer only to that specific, individual cow (Apoha 82)! Moreover, not only
would it be impossible to learn the relation of a word with its meaning, but there would be
no occasion for employing words; since an exclusion is not ascertained by the senses, “on
the condition of perceiving what, is it to be used?” (Apoha 78).
Thus, Kumārila argues that if an apoha cannot be cognized independently of language
and inference, it would not be suitable as the meaning of a word. By implication, then, the
meaning of a word must be something that is accessible to the language learner
independently of language and inference; that is to say, it must be perceptible. Since it
must also be a universal—for, once again, a relation cannot be established between a word
and infinitely many particulars—there must, therefore, be perceptible universals. Being
perceptible, universals will be real things (vastu), not mere absences or negations.
This passage in the Apohavāda is to be placed alongside a discussion in the
Pratyakṣasūtra chapter of the Ślokavārttika, which comes before the Ākṛti-, Apoha-, and
Vanavāda chapters. At Pratyakṣa 200–206, arguing against the view that all conceptual
awareness consists just in the superimposition of words onto what one is experiencing
and that one never experiences anything directly, unfiltered by language—this of course
was Bhartṛhari’s position—Kumārila maintains that one must be distinctly aware of word-
meanings, which must be universals, independently of an awareness of words in order to
be able to learn the relation of word and meaning.78 In fact, however, the entire second
half of the Pratyakṣasūtra, from v. 111 on, which is dedicated to proving that there is such
a thing as conceptualized (savikalpaka) perception, can be seen as establishing the
perceptibility of universals. The awareness of an object that is in contact with the senses
as a cow, brown, walking, and so forth, is a perception precisely because the words or
concepts being applied depict real features—that is, not only the genus but also
properties, which are a kind of universal—of the object that are perceived.
At Anumānapariccheda 149–153, finally, Kumārila argues in anticipation of his
statement at Ākṛti 39 that if universals were not perceptible inference would not be
possible, at least not for the Buddhist. The Buddhist accepts only two pramāṇas,
perception and inference. If the inferential mark of an inference, the liṅga—which
Kumārila says everyone takes to be a universal (e.g., the property of possessing smoke)79
—were not known by perception, then another inference would be required in order to
know it. The liṅga of that inference, however, could be known only by a further inference.
Thus, because infinitely many inferences would have to be carried out in order to know
the liṅga, and because a liṅga cannot indicate anything if it is not known, one could never
infer anything. For inference to be possible the liṅga—which, again, is a universal—must
be perceptible. Of course, the perceptibility of universals will also make the establishing
of a pervasion (vyāpti), the relation of invariable concomitance between the inferential
mark and the property to be proved that constitutes the major premise of an inference,
from repeated observation of their co-occurrence—this was Kumārila’s theory of the
ascertainment of the pervasion—much less problematic than otherwise.
Thus, for Kumārila it is not enough for universals simply to exist in order for language
and inference to function. They must also be perceived. This, surely, is his deeper motive
for adopting what, for someone trained in Western philosophy, is an unusual position. We
see, however, that it lies at the very heart of Kumārila’s system; for not only is it required
for his understanding of language and inference but, together with the fundamental
epistemological principle of intrinsic validity, it also anchors his realism. For Kumārila, we
can be confident that the properties and types we predicate of objects really belong to
them because we perceive them, that is, they cause perceptual cognitions in us. Thus, we
live in a real, mind-independent world populated by physical objects with real properties,
and we can assume that our experience represents them accurately.
Sometimes in philosophy one has to adopt a position and try it out, so to speak—
develop it completely and attempt to defend it—before one can realize that it is untenable.
Although I have hardly been able to do that here, that is, give the strongest possible
statement of Kumārila’s position and defend it against all objections, I feel that I have
explored it enough to know that I would be inclined to move in a different direction in
developing an epistemological argument for the existence of universals, though an
argument that is indebted to Kumārila in one key respect.
It remains very implausible that you can see, say, cowness. What makes something a
cow is a certain set of facts: it gives milk, it can moo, it chews the cud, it can give birth
only to other cows, and so on. On the ultimate biological level, what makes something a
cow is that its DNA contains a certain sequence of genes. A collection of facts is not
something one can see. Similarly, when I think I recognize a church off in the distance, I
see a distinctive visible feature most churches have—a steeple. But what is essential to a
church is that it is a place of Christian worship, and that is not something you can see.
Thus, when Kumārila is talking about seeing that something is a man, I believe he is
confusing the universal with the shape or configuration of an object, which is one of the
three things Nyāya philosophers held to be the meaning of a word—the ākṛti in the usual
sense. According to the Naiyāyikas, a word can refer either to the individual (vyakti), the
shape/configuration (ākṛti), or the genus/universal (jāti), depending on the
circumstances.80 Now, Kumārila insists that, while the configuration indicates (is an
upalakṣaṇa of) a universal because both universal and configuration inhere in the same
individual, they are not identical.81 This is another aspect of his theory of universals that
needs to be investigated thoroughly. Regardless of that, however, it seems more
reasonable that if there are universals, they would not be things you could see. They
would, most plausibly, be certain sets of facts about things that we comprehend
intellectually and formulate in a definition or a theory. In that sense, being a cow would be
something abstract.82
Nevertheless, these abstract entities may still produce cognitions by some mechanism
we do not understand. Plato famously thought that one could go through a dialectical
process that culminates in a rational intuition of universals; an alternative view would be
that they are discoverable through science. If it is indeed the case that we can achieve a
clear and distinct cognition of universals, then Kumārila may be right that we should base
our beliefs about what does or does not exist on our experience. We should not let our
judgments about what is possible or impossible dictate whether our cognitions are valid
or invalid; rather, our cognitions, valid and invalid, should determine what is possible and
impossible. If we somehow are able to have valid cognitions of universals, whether
perceptual or not, then that would by itself establish their existence, even if in the end we
must hold them to be abstract, or at least non-spatiotemporal.
I believe that essentially this idea is expressed in the following quotation from the
mathematician Kurt Gödel, though he is of course talking about a different type of
abstract entity, namely, mathematical objects. Without realizing it, he is appealing to
Kumārila’s principle of intrinsic validity.

The objects of transfinite set theory ... clearly do not belong to the physical world and even their
indirect connection with physical experience is very loose... But, despite their remoteness from
sense experience, we do have a perception also of the objects of set theory, as is seen from the fact
that the axioms force themselves upon us as being true. I don’t see why we should have less
confidence in this kind of perception, i.e., in mathematical intuition, than in sense perception,
which induces us to build up physical theories and to expect that future sense perceptions will
agree with them and, moreover, to believe that a question not decidable now has meaning and may
be decided in the future.83

I conclude with Kumārila’s own statement of his principle from the Codanāsūtra chapter
of his Ślokavārttika:

tasmād dṛdhaṃyad utpannaṃna visaṃvādam ṛcchati /

jñānāntareṇa vijñānaṃtat pramāṇaṃpratīyatām // ŚV, Codanā 80//
Therefore, that cognition which is firm, which has [actually] arisen, which is not contradicted by
another cognition—that is to be regarded as a means of knowledge.

Appendix: Selected Verses from ŚV Ākṛti-, Vana-, and Apohavāda

Ākṛtivāda
ākṛtivyatirikte ’rthe sambandho nityatāsya ca /
na siddhyetām iti jñātvā tadvācyatvam ihocyate //Ākṛti 1//
If the meaning were distinct from the ākṛti, then a connection and its eternality could not
be established. Realizing this, it is stated here [in this treatise, or more generally, in
Mīmāṃsā] that that is what is meant.
jātim evākṛtiṃprāha84 vyaktir ākriyate ’nayā /
sāmānyaṃtac ca piṇḍānām ekabuddhinibandhanam // Ākṛti 3//
He [Śabara] calls the genus ākṛti [with the following derivation in mind:] the particular is
formed (ākriyate) by this. And that is the universal of particulars which is the cause of a
cognition of one thing.
tannimittaṃca yat kiṃcit sāmānyaṃśabdagocaram /
sarva evecchatīty evam avirodho ’tra vādinām // Ākṛti 4//
And everyone holds that the cause of that [single cognition] is some universal which is the
referent of words. Thus, there is no conflict of the experts in regard to this.
sarvavastuṣu buddhiś ca vyāvṛttyanugamātimikā /
jāyate dyātamakatvena vināsau85 ca na sidhyati // Ākṛti 5//
Moreover, in regard to all things a cognition arises which has the nature of exclusion and
commonness; and without [those things] having a dual nature that is not admissible.
nirviśeṣaṃna sāmānyaṃbhavec chaśaviṣāṇavat /
sāmānyarahitatvāc ca viśeṣās tadvad eva hi // Ākṛti 10//
There could not be a universal without a particular/difference, like a hare’s horn; and
particulars/differences [without a universal] are indeed just like that [i.e., like a hare’s
horn], because they are devoid of a universal.
tadanātmakarūpeṇa hetū vācyāv imau punaḥ /
tena nātyantabhedo ’pi syāt sāmānyaviśeṣayoḥ // Ākṛti 11//
On the other hand, these two reasons are to be explained as “not having that nature.”
[That is, because in the one case it would no longer have the nature of a universal, in the
other case, it would no longer have the nature of a difference.] Therefore, there could be
no absolute difference, either, of universal and difference.
bhinnā viśeṣaśaktibhyaḥsarvatrānugatāpi ca /
pratyekaṃsamavetā ca tasmāj jātir apīṣyatām // Ākṛti 17//
Therefore, a universal, too [along with the particular] is to be accepted, which is distinct
from the particulars and their śaktis, which is also common to all of them, and which
inheres in each.
sāmānyaṃnānyad iṣṭaṃcet tasya vṛtter niyāmakam /
gotvenāpi vinā kasmād gobuddhir na niyamyate // Ākṛti 35//
yathā tulye ’pi bhinnatve keṣucit vṛttyavṛttitā86 /
gotvāder animittāpi87 tathā buddhir bhaviṣyati // Ākṛti 36//
[Objection:] If it is not held that another universal is the cause of restricting its
occurrence [in certain particulars], why is the cognition “cow” not restricted [to certain
particulars] without cowness? Just as there is the occurrence or non-occurrence of
cowness, etc., in certain [particulars], even though they are equally different, so the
cognition [“cow”] will arise [in regard to certain particulars] even without [another]
cause.
viṣayeṇa hi buddhīnāṃvinā notpattir iṣyate /
viśeṣād anyad icchanti sāmānyaṃtena tad dhruvam// Ākṛti 37//
Indeed, the arising of cognitions without an object is not accepted. They hold [rather]
there to be a universal distinct from the particular. Therefore, certainly, there is that.
yadā tu śabalaṃvastu yugapat pratipadyate /
tadānyānanyabhedādi sarvam eva pralīyate // Ākṛti 62cd-63ab//
When, however, the diverse thing is simultaneously cognized, then indeed everything
[namely, that it is] other, not other, particular, and so on, disappears.

Vanavāda
mudgamāṣatilādau ca yatra bhedo na lakṣyate /
tatraikabuddhinirgrāhyā jātir indriyagocarā88 //Vana 24//
When the particular is not perceived, when there are mung beans, māṣa beans, sesame
seeds, and so forth, then the universal which is an object of the senses is apprehended by
means of a cognition of one thing.
ārād dṛṣṭe ca puruṣe sandeho brāhmaṇādiṣu /
na syād yadi na gṛhyate sāmānyaṃcakṣurādinā // Vana 25//
When a man is seen from a distance, there would not be doubt about whether he is a
Brahmin, etc., if the universal [humanity] were not apprehended with the sense of vision,
etc.

Apohavāda
na cānvayavinirmuktā pravṛttir liṅgaśabdayoḥ /
tābhyāṃca na vināpoho na cāsādharaṇe ’nvayaḥ //Apoha 73//
apohaś cāpy aniṣpannaḥsāhacaryaṃkva dṛśyatām /
tasminn adṛśyamāne ca na tayoḥsyāt pramāṇatā //Apoha 74//
And there is no functioning of an inferential mark or a word without a concomitance
[between the word or inferential mark and what it indicates]. And without those two an
apoha is not [cognized]. But there is no positive concomitance with something unique.
Moreover, an apoha is not evident [before knowing it through a word or an inferential
mark, so] with regard to what will one observe the coexistence of [a word or inferential
mark]? And if that [coexistence] is not observed those two [namely, word and inferential
mark] would not be means of knowing [it].89

Notes
1. PV 3.145–146. Cf. PV 1.71ab; PV 3.48–49.
2. Cf. PV 3.50.
3. Definitions, however, specifically definitions of the pramāṇas, the means of knowledge, became
the focus of epistemological debate.
 4. By “Western philosophy” I mean the various traditions of thought that evolved from the
investigations of the ancient Greeks, in particular the Presocratics, Plato, and Aristotle, in
continental Europe and Britain over the past twenty-five hundred years, and which since the
nineteenth century have spread to other Commonwealth countries and the Americas and much of
the rest of the world.
5. My italics. We also find in Indian ethics a reluctance to derive morality from a single a priori
principle. See Hacker (2006, p. 486): the definition of dharma in Hinduism is “radically empirical.”
6. See, however, Fine (1999, p. 12).
7. This view is often attributed to Quine. See, for instance, Sen (2006, p. 39). For Quine, however,
the main question about universals is whether it is necessary to postulate them in order to explain
the reference of predicates. The debate about whether a naturalistic account can be given of
knowledge of abstract entities in contemporary philosophy has mainly concerned mathematical
objects. For a helpful overview of the various positions taken, see Balaguer (2009), especially
section 5, “The Epistemological Argument against Platonism”. See Benacerraf (1973) for a classic
statement of the view that our knowledge of mathematics is problematic on a Platonist
interpretation. The epistemological argument against the existence of mathematical objects can be
applied mutatis mutandis to universals.
8. See Armstrong (1978). See especially pp. 54–55. As Armstrong prefers to put it, particulars act on
our senses in virtue of universals: “Consider a typical case of the application of the predicate
‘crimson.’ Objects having the property, crimson (or, perhaps better, having one of the disjunctive
range of properties covered by that predicate) act upon our sense-organs. They act in virtue of the
crimsonness of the object” (p. 65; my italics).
9. Scharf (1996) is valuable but does not bring out what I consider the most interesting and
important aspects of Kumārila’s discussion. Harikai (1997) is very helpful in elucidating Kumārila’s
understanding of the non-difference of universal and particular. It is not as concerned with the
epistemological aspects of Kumārila’s position as I am, nor with his struggles with the Buddhists.
Bhatt (1962, pp. 405–421), as usual, presents Kumārila’s main ideas accurately, and in a
philosophically astute manner, but does not follow the text very closely.
10. Kei Kataoka has kindly provided me with scans of Adyar Library manuscripts of Sucaritamiśra’s
Kāśikā on Ākṛtivāda and Vanavāda. These chapters of the Kāśikā have not yet been edited and
published.
11. Which, however, is impossible to determine with certainty, since not all of Dignāga’s works have
been preserved. Dignāga’s lost Sāmānyaparīkṣāvyāsa (Pind, 2009, p. 9) almost certainly would
have been relevant to understanding Kumārila’s discussion of universals.
12. See Taber (2010).
13. My readings of the verses are based on a comparison of two printed editions, ŚVŚ and ŚVK, and
the available commentaries. Where the editions do not agree, or a reading supported by a
commentary is preferred, the variants are noted.
14. The point that a word could not refer to the particulars of a certain type because they are
“endless” and a connection with infinitely many things could not be made is expressed by Dignāga
at PS(V) 5.2ab. He also states there that the particulars could not be the meaning of a word
“because of deviation” (vyabhicārataḥ). That is to say, a word in that case would potentially apply
to many different particulars, so that one would not know which particular is being referred to
when the word is used. See Pind (2009, pp. 76–77).
15. The orientation toward first philosophy, however, did not carry into the modern period. In Locke,
for instance, after belief in the existence of real universals, that is formal causes, had been widely
abandoned, the question of universals becomes the question of the origin of “general concepts”
which serve as the meaning of general terms. Armstrong (1978, p. 11) laments the inclination to
see universals as meanings: “Why is it that philosophers have thought, or have been tempted to
think, that to each distinct predicate-type there corresponds its own peculiar universal? I think
that the answer is clear. It is the influence of the Argument from Meaning which has so often, and
so fatally, distorted the Problem of Universals. If universals are conceived of as predicates, then it
follows at once that each predicate-type is associated with its own universal. Realists have then
put an inflationary, Nominalists a deflationary, interpretation on this situation.”
16. See, for example, PV 3.25cd–27ab; cf. Franco and Notake (2014, pp. 80–85). The problem arises
for the Jainas if universals are conceived as “absolutely different” from particulars. See Trikha
(2012).
17. I take the expression ekabuddhi in Ākṛti 3d and elsewhere, with some uncertainty, to mean “a
cognition of one/a single thing” in the sense of one thing extending over all the particulars, as
opposed to a cognition of something that is distinct (bhinna) from other things. At Ākṛti 12a, for
instance, the expression sāmānyabuddhi seems to be standing in for ekabuddhi. Kumārila’s
ensuing discussion, vv. 13–16, of whether a “single” (ekā) or a “diverse” (bhinnā) śaktiḥ is
apprehended also supports this interpretation. (See, e.g., 15ac’: parasparabhinnatvād viśeṣā
naikabuddhibhiḥ / gṛhyante ...) I also see confirmation of this reading in Kumārila’s discussion of
the problem how a thing could be “both one and many,” that is, at once universal and particular in
nature, in Ākṛti 51–63ab. Thus, insofar as one cognizes a universal, one cognizes one thing as
opposed to many things. Sucarita and Jayamiśra gloss ekabuddhi as ekākārā buddhiḥ, “a cognition
 having a single form,” that is, a cognition that presents one thing. It is also possible to translate
ekabuddhi as “a single/the same cognition,” provided one understands it equivalently: a single type
of cognition defined by a certain content, say “cow,” that extends over many things, every time it
occurs. (Obviously, the ekabuddhi cannot be a numerically identical cognition produced by all
particulars of a certain type.)
18. This principle is articulated by the Vṛttikāra, an earlier Mīmāṃsaka who is quoted extensively by
Śabara in connection with his explanation of MS 1.1.5: “For us [Mīmāṃsakas] in all matters
[direct] experience is the authority”; sarvatra no darśanaṃpramāṇam. Frauwallner (1968, pp. 36,
15).
19. The commentators Sucaritamiśra and Pārthasārathimiśra take this to mean that perception
initially apprehends “mere being,” as certain Advaitins believed.
20. Dharmakīrti attacks the idea of a perception of two forms, particular and universal or common, at
PVSV 26, 16–20: na hi śuktau dve rūpe samānaṃviśiṣṭaṃca tathāpratipattiprasaṅgāt. apratipattau
vā vivekena dvitvavikalpāyogāt. atiprasaṅgāc ca. tasmāt paśyan śuktirūpaṃviśiṣṭam eva paśyati.
niścayapratyayavaikalyāt tv aniścinvan tatsāmānyaṃpaśyāmīti manyate. “For there are not two
forms in mother-of-pearl, a common one and a distinct one; for then it would follow that it is
[always] perceived in that way, or because if they were not perceived distinctly, the notion [that
mother-of-pearl] has two forms would be incorrect and there would be unwanted consequences
[namely, one could suppose that anything, whether one perceives two distinct forms of it or not,
has distinct forms!]. Therefore, seeing the form of mother-of-pearl, one sees only that which is
distinct. But not ascertaining [the particular thing] due to a weakness/defectiveness in the
conditions for ascertainment, one thinks, ‘I see what it has in common.’”
21. Reading with ŚVK veṣyate in 7b; ŚVŚ ceṣyate.
22. A stock example of a figurative expression is, “The boy is a fire,” where “fire” applied figuratively
connotes brilliance. One must be previously directly acquainted with both of the terms in order to
compare them, or more correctly, to superimpose one upon the other.
23. Although it seems clear in most instances that Kumārila employs the term viśeṣa to mean
particular, that is, interchangeably with vyakti, sometimes he employs it to mean difference or
even a subtype. In Ākṛti 9–11 it is especially ambiguous. Kumārila could be using it in the broader
sense, meaning difference, with the particular intended as a special case.
24. The commentators supply the reasons (hetu) for these two syllogisms. Thus, Pārthasārathi on
Ākṛti 10: nirviśeṣaṃparikalpitaṃsāmānyaṃdharmī nāstīti sādhyam, tathā niḥsāmānyā viśeṣā na
santīti viśeṣarahitatvāt sāmānyarahitatvād iti hetudvayam. Is this a stock Mīmāṃsā argument
Kumārila inherited from his predecessors? His attempt to clarify the hetu in the following verse,
Ākṛti 11, suggests that it may not have been his own.
25. Cf. Kataoka (2010, pp. 182, 5–181, 3 [Nyāyamañjarī text], 217–216 [Kataoka’s explanation]).
26. See, for example, PV 1.68–77.
27. The same ambiguity that affects Kumārila’s expression ekabuddhi attends Dharmakīrti’s term
ekapratyavamarśa. See note 17 above. However, since particulars are judged the same on the
basis of having the same effect, which is the ekapratyavamarśa, it seems that the emphasis for
Dharmakīrti is on its being the same judgment, even though its being the same is due to its
presenting the same thing, for instance, “cow.” Thus, it seems most appropriate to me to translate
this term as “a single/the same judgment.”
28. In his proof of momentariness Dharmakīrti sets up the dilemma: what is not momentary cannot
have causal efficacy either gradually, because that would involve some change in its nature over
time, or at once, because then, insofar as it is by nature a causal agent in a single moment, it
would be constantly producing that effect. See Steinkellner (1968).
29. See PV 1.106, the first half of which Jayamiśra cites on Ākṛti 12:

tad ekam upakuryus tāḥkatham ekāṃdhiyaṃca na /

kāryaṃca tāsāṃprāpto ’sau jananaṃyad upakriyā //
They [the particulars] would influence that one [thing, the universal, on the view that the universal does not
bring about the cognition of one thing by itself]. Why [could they] not also [influence] the single cognition
[directly]? And that [universal, etc.] would become their effect, since influencing is producing.
(Note: Gnoli reads kāryaś ca in 106c.) See Frauwallner (1936, pp. 279 and 281).

30. The main problem for Dharmakīrti is that the account he wants to give of our cognition of general
entities is, at bottom, an error theory, and it is notoriously difficult to give a coherent account of
error.
31. abhinnapratibhāsinī dhīḥ, PV 107ab’. Frauwallner translates, “eine Erkenntnis, welche das
gleiche Bild zeigt,” p. 281. See note 17 above.
32. PV 1.107cd: pratibhāso dhiyāṃbhinnaḥsamānā iti tadgrahāt //
33. PV(SV) 1.107–109, especially PVSV 56, 18–57, 6. Cf. Frauwallner (1936, pp. 281–284). According
to Frauwallner’s enumeration these are stanzas 109–111.
34. See Kataoka (2010, pp. 181–176) (sec. 3.4.5) for Jayanta’s critique of the ekapratyavamarśa
proposal and 216–214 for Kataoka’s explanation of the passage.
35. See, for example, PVSV 55, 6–13.
36. This characterization of the objection is taken from Sucarita’s commentary on Ākṛti 31cd, ŚVKāś,
pp. 2574–75. See PV 3.47cd and PVV ad loc. The objection is reminiscent of Plato’s Third Man
Argument, since it involves a regress, but is not the same. It concerns whether another universal
must be postulated to explain the connection of universals with particulars and not whether
another universal has to be postulated to explain why the same predicate (“man”) would apply to
both universal and particular.
37. svakāraṇata evedṛśī vyaktir utpadyate yena sā jātyā sambadhyate, ŚVŚ 10, 8.
38. Commenting on 31cd Sucaritamiśra says: nātra rajjughaṭayor iva jātivyaktyor
atyantabhinnayoḥsambandhaḥapi tarhi tādātmyam eva. vyaktir hi svakāraṇato jāyamānaiva
tattadākārā niṣpadyate. ataḥsvābhāvike tādātmye kasya kena sambandho yo ’nyanibandhano
bhavet; “Here, there is not a connection of a universal and a particular that are completely distinct
from each other, like a rope and a pot. Rather, there is identity. For the particular arising from its
own causes comes forth with this or that form (ākāra). Therefore, [their] identity being natural, of
what is there a connection with what that could have something else as its cause?” (ŚVKāś, p.
2575).
Cf. Ākṛti 46–47, discussed by Harikai (1997, pp. 401–402). Kumārila explains that because the
features of a cow, such as the dewlap, are not distinct from the individual cow and the universal
“cow” also is not distinct from it, the answer to the question, “How can cowness exist only in
things possessed of dewlap, etc.?,” so that the latter serve as indicators (upalakṣaṇa) of cowness,
is that it has the nature of the individual endowed with dewlap (tadātmaka). If one were to ask,
“Why is there that identity?,” the answer would be: Due to the nature of the individual (svabhāvāt,
47d).
39. See ŚVŚ ad loc.
40. Reading with ŚVŚ kasmāt in 34a, which seems to be supported by ŚVKāś; ŚVK tasmāt.
41. Reading with ŚVK vṛttyavṛttitā in 36b; ŚVŚ vṛttyapekṣitā.
42. Cf. Dharmakīrti’s discussion PV 3.45–47 and PVV ad loc. See Franco and Notake (2014, pp. 123–
126).
43. Cf. ŚVŚ ad loc.: yadi paraṃbhavatā punaḥpunar etāvad eva vaktavyam avyatirikta evāyam,
vyatiriktāvabhāsapratyayo mithyābhūtaḥiti na ca tatra kiṃcana pramāṇam ity āha.
44. Sucarita interprets icchanti, “they hold,” as “for us” (naḥ) (ŚVKāś, p. 2579), which could mean
ordinary people, but usually means Mīmāṃsakas. Jayamiśra’s discussion seems to imply that it is
the view of the Mīmāṃsakas that Kumārila is referring to here. But surely the point would not
carry much weight if this were just a belief shared by a relatively small circle of philosophers!
45. See Kataoka (2011, pp. 60–98) for a fresh account of Kumārila’s theory of intrinsic validity.
46. Chakrabarti (2006) offers an analogous proposal from the Nyāya perspective.
47. Cf. PV 3.25ac’: na jātir jātimad vyaktirūpaṃyenāparāśrayam / siddham ...; “The universal is not
the thing possessing the universal, since that which has the form of a particular is established as
not having another substratum.” Cf. Franco and Notake (2014, pp. 80–81). That is to say, it does
not extend over other things.
48. Reading with ŚVK prasajyate in 52d; ŚVŚ pratīyate. Cf. PVSV 24,25–25,2.
49. Cf. PV 3.41: parasparaviśiṣṭānām aviśiṣṭaṃkathaṃbhavet / rūpaṃdvirūpatāyāṃvā tad vastv
ekaṃkatham bhavet // “How could things that are different from each other have a nature that is
not different? Or, if it has two natures, how could that thing be one?” Cf. Franco and Notake (2014,
p. 41).
50. Reading with ŚVŚ varam in 51d, which seems to be supported by ŚVKāś, p. 2591, 5–6; ŚVK kiṃcit.
51. Reading with ŚVŚ hi in 56a; ŚVK tu.
52. Reading with ŚVK varṇavigrahaḥ in 57d; ŚVŚ varṇanigrahaḥ. Cf. TS 1745, which however also
attests varṇanigrahaḥ.
53. Reading with ŚVK tadbodhānuguṇam in 60d.; ŚVŚ tadbhāvānuguṇam.
54. Reading yadābhedena in 61d; ŚVK and ŚVŚ yadā bhedena. Reading with ŚVŚ sāmānyamātratvam in
62a; ŚVK sāmānyatantratvam.
55. Cf. TS 1746.
56. The philosophical and historical relation of the theory of Kumārila sketched here and Jaina
perspectivalism is beyond the scope of this chapter.
57. See PV(SV) 25,26–27,6. See Kellner (2004, esp. pp. 7–9, 11–15).
58. yāvanto ’sya parabhāvās tāvanta eva ... samāropā iti, PVSV 26, 22–23.
59. PV 1.40–41.
60. Although, in the passage of PV(SV) discussed here, he seems to say that the object has properties
(guṇa, dharma) or aspects (ākāra) (see, e.g., PV 1.45 and 46), he also says that the object is “one”
and “without parts,” and that “different types” (jātibhedāḥ) are “conceived” or “imagined” (PV
1.41).
61. Or perhaps the Vṛttikāra. Śabara earlier announced that in this section of his commentary on MS
1.1.3–5 he is presenting the views of a predecessor, whom he calls simply “the commentator,”
Vṛttikāra. Whether he is directly quoting him or merely paraphrasing or summarizing his ideas is
uncertain.
62. Which, again, he refers to by the word ākṛti. See p. 12 above.
63. Frauwallner (1968, pp. 40, 14–15): nanv ākṭiḥsādhyāsti na veti? na pratyakṣā satī sādhyā
 bhavitum arhati.
64. Reading with ŚVK cocyate in 10d; ŚVŚ bodhyate.
65. Reading with ŚVK sarvalokaprasiddhatvāt in 12a; ŚVŚ sarvatra lokasiddhatvāt.
66. Reading with ŚVK idam in 13d; ŚVŚ iyam. Following Sucarita’s interpretation, ŚVKāś, p. 2815.
67. ŚVKāś, p. 2816, 9, on Vana 14: niraṅkuśā hi jaḍadhiyāṃvṛttiḥ.
68. Ms. p. 2823.
69. Reading with ŚVK indriyagocarā in 24d; ŚVŚ indriyagocaraḥ. Jayanta uses this same example; see
Kataoka (2010, pp. 197, 5–196, 3).
70. NM I, 319; 322–323. Jayanta, however, also argues that they are objects of sense perception,
given in “non-conceptual” perception; see, cf. Kataoka (2010, pp. 201–194).
71. Bronkhorst and Ramseier (1994, pp. 45, 3–5).
72. And one should keep in mind that Dharmakīrti says that it applies only to the realm of vyavahāra,
everyday practice.
73. It may have already been established as a principle in Sarvāstivāda thought. See Dhammajoti
(2007, pp. 76–86, esp. p. 86). It seems implied by Vasubandhu’s mentioning of “the causing of an
effect” (kṛtyakriyā) as a criterion of objective reality in Viṃśikā 2.
74. Reading with ŚVK na hi syān in 39c; ŚVŚ naiva syān.
75. From this point on I shall refrain from noting variants, in anticipation of the publication of a new
edition and translation of ŚV Apoha by Kataoka and Taber.
76. These verses (Apoha 71–72) conclude a long discussion, beginning with Apoha 42, of the problem.
How can one tell a difference between apohas? If one cannot do that, Kumārila charges, then on
the Buddhist apoha theory all words would be synonymous. Specifically, Kumārila argues, one
could not cognize different apohas in terms either of the things they are based on or the things
they exclude.
77. PS 5.34. See Pind (2009, pp. 103–104).
78. See Taber (2005, pp. 131–133).
79. Anumāna 150ab.
80. Nyāyasūtra 2.2.66.
81. Vana 16–19. Pārthasārathimiśra, interestingly, sees the discussion of the example of gold and the
different objects made from it, vv. 20–23, as further supporting the point that the universal is
different from the shape. Jayamiśra, meanwhile, takes it to be showing that the universal is
perceptible, like the gold itself in various ornaments.
82. One can also, of course, adopt the reasonable position of Armstrong, that while there are property
universals that are perceptible, there are no “substantival universals”; that is, being gold or being
an electron are not properties. See Armstrong (1978, pp. 61–67).
83. Gödel (1964, p. 271).
84. ŚVŚ prāhur.
85. ŚVŚ vinā sā.
86. ŚVŚ vṛttyapekṣitā.
87. ŚVŚ animitte ’pi.
88. ŚVŚ indriyagocaraḥ.
89. Text and translation in preparation by Kei Kataoka and John Taber.

Bibliography
Armstrong, D. M. (1978), Universals and Scientific Realism, Vol. 2: A Theory of Universals. Cambridge:
Cambridge University Press.
Balaguer, M. (2009), “Platonism in metaphysics.” Stanford Encyclopedia of Philosophy.
plato.stanford.edu/entries/platonism/#5.
Benacerraf, P. (1973), “Mathematical truth.” Journal of Philosophy, 70, 661–679.
Bhatt, G. P. (1962), Epistemology of the Bhāṭṭa School of Pūrva Mīmāṃsā. Varanasi: Chowkhamba
Sanskrit Series Office.
Bronkhorst, J., and Ramseier, Y. (1994), Word Index to the Praśastapādabhāṣya. Delhi: Motilal
Banarsidass.
Chakrabarti, A. (2006), “On perceiving properties,” in Strawson and Chakrabarti, pp. 309–318.
Cohen, S. M., Curd, P., and Reeve, C. D. C. (eds.) (2011), Readings in Ancient Greek Philosophy. 4th ed.
Indianapolis: Hackett Publishing.
Dhammajoti, K. L. (2007), Sarvāstivāda Abhidharma. Hong Kong: Centre of Buddhist Studies.
Fine, G. (1999), “Introduction,” in Gail Fine (ed.), Plato 1: Metaphysics and Epistemology. Oxford:
Oxford University Press.
Franco, E., and Notake, M. (2014), Dharmakīrti on the Duality of the Object: Pramāṇavārttika III 1–63.
Vienna: LIT Verlag.
Frauwallner, E. (1932), “Beiträge zur Apohalehre. I. Dharmakīrti. Übersetzung.” Wiener Zeitschrift für
die Kunde des Morgenlandes, 39, 247–285.
Frauwallner, E. (1968), Materialien zur ältesten Erkenntnislehre der Karmamīmāṃsā. Vienna:
Österreichische Akademie der Wissenschaften.
Gödel, K. (1964), “What is Cantor’s continuum problem?” in Paul Benacerraf and Hilary Putnam (eds.),
Philosophy of Mathematics. Englewood Cliffs, NJ: Prentice Hall, pp. 258–273.
Hacker, P. (2006): “Dharma in Hinduism.” Journal of Indian Philosophy, 34, 479–496.
Harikai, K. (1997), “Kumārila’s acceptance and modification of categories of the Vaiśeṣika school,” in
Eli Franco and Karin Preisendanz (eds.), Beyond Orientalism: The Work of Wilhelm Halbfass and Its
Impact on Indian and Cross-Cultural Studies. Amsterdam: Rodopi, pp. 395–415.
Kataoka, K. (2010), “A critical edition of Bhaṭṭa Jayanta’s Nyāmañjarī: Jayanta’s view on jāti and
apoha.” The Memoirs of Institute for Advanced Studies on Asia, 158, 220(61)–168(113).
Kataoka, K. (2011), Kumārila on Truth, Omniscience, and Killing, Part 2: An Annotated Translation of
Mīmāṃsā-Ślokavārttika ad 1.1.2 (Codanāsūtra). Vienna: Verlag der österreichischen Akademie der
Wissenschaften.
Kellner, B. (2004), “Why infer and not just look? Dharmakīrti on the psychology of the inferential
process,” in Shoryu Katsura and Ernst Steinkellner (eds.), The Role of the Example (Dṛṣṭānta) in
Classical Indian Logic. Wiener Studien zur Tibetologie und Buddhismuskunde 58. Vienna: Universität
Wien, pp. 1–51.
Mohanty, J. N. (1992), Reason and Tradition in Indian Thought. Oxford: Clarendon Press.
NM: Nyāyamañjarī of Jayantabhaṭṭa. 2 vols. Edited by K. S. Varadacharya. Mysore: Oriental Research
Institute, 1969.
Pind, O. H. (2009), “Dignāga’s philosophy of language: Dignāga on anyāpoha.” Dissertation, Universität
Wien.
PS 5: Pramāṇasamuccaya of Dignāga, Chapter 5. In Pind (2009).
PV 1: In PVSV.
PV 3: Pramāṇavārttika of Acharya Dharmakirti, with the Commentary ‘Vritti’ of Acharya
Manorathanandin. Bauddha Bharati Series 3. Edited by Swami Dwarikadas Shastri (Svāmī
Dvārikadāsa Śāstrī). Varanasi: Bauddha Bharati, 1968.
PVSV: The Pramāṇavārttikam of Dharmakīrti: The First Chapter with the Autocommentary. Serie
Orientale Roma 23. Edited by Raniero Gnoli. Rome: Istituto Italiano per il Medio ed Estremo Oriente,
1960.
PVV: Pramāṇavārttikavṛtti of Manorathanandin. In: PV 3.
Scharf, P. (1996), The Denotation of Generic Terms in Ancient Indian Philosophy: Grammar, Nyāya, and
Mīmāṃsā. Philadelphia: American Philosophical Society.
Sen, P. K. (2006), “Strawson on universals,” in Strawson and Chakrabarti, 17–47.
Steinkellner, E. (1968), “Die Entwicklung des Kṣaṇikatvānumānam bei Dharmakīrti.” Wiener Zeitschrift
für die Kunde Süd- und Ostasiens, 12/13, 261–377.
Strawson, P. F., and Chakrabarti, A. (eds.) (2006), Universals, Concepts and Qualities: New Essays on
the Meaning of Predicates. Aldershot, UK: Ashgate.
ŚVK: Ślokavārtikaṭīkā (Śarkarikā) of Bhaṭṭa Jayamiśra. Edited by C. Kunhan Raja. Madras: University of
Madras, 1946.
ŚVŚ: Ślokavārttika of Śrī Kumārila Bhaṭṭa, with the Commentary Nyāyaratnākara of Śrī Pārthasārathi
Miśra. Edited by Svāmī Dvārikadāsa Śāstrī. Varanasi: Tara Publications, 1978.
ŚVKāś: Ślokavārttikakāśikā of Sucaritamiśra.
For Ākṛtivāda: A manuscript preserved in the Adyar Library, Chennai, No. 63359 (TR 66–5). Paper.
Devanāgarī. pp. 2543–2612.
For Vanavāda: A manuscript preserved in the Adyar Library, Chennai, No. 63360 (TR 66–6). Paper.
Devanāgarī. pp. 2802–2908.
ŚVŚ: Ślokavārttikaśarkarikā of Jayamiśra. In: ŚVK.
Taber, J. (2005), A Hindu Critique of Buddhist Epistemology: Kumārila on Perception. London:
RoutledgeCurzon.
Taber, J. (2010), “Kumārila’s Buddhist.” Journal of Indian Philosophy, 38, 279–296.
Trikha, H. (2012),: Perspektivismus und Kritik. Das pluralistische Erkenntnismodell der Jainas
angesichts der Polemik gegen das Vaiśeṣika in Vidyānandins Satyaśāsanaparīkṣā. Publications of the
De Nobili Research Library 36. Vienna: Institut für Südasien-, Tibet- und Buddhismuskunde.
TS: Tattvasaṅgraha of Śāntarakṣita. With the Commentary of Kamalaśīla. Vol. 1. Gaekwad’s Oriental
Series 30. Edited by Embar Krishnamacharya. Baroda: Central Library, 1926.
 11
The Role of Causality in Ratnakīrti’s Argument for
Momentariness
Joel Feldman

Whether there are things that endure over time is an issue that has provoked at least as
much debate in the history of Indian philosophy as it has in the history of Western
philosophy. The issue has been a perennial subject of dispute in the European tradition
going back to Heraclitus and Parmenides, and the controversy over endurance persists to
this day in the debate over the metaphysics of temporal parts. In the classical Indian
tradition, the view that things do not endure is defended by a long line of Buddhist
philosophers who argue for the doctrine of momentariness (kṣaṇikatva), a view that is
vigorously challenged by a long line of philosophers of the Nyāya school (among others),
who maintain that things endure. An intricate debate over the issue unfolds over the first
millennium of the Common Era as the Buddhists refine their arguments to meet the
objections of their Nyāya critics. Ratnakīrti, an eleventh-century Buddhist philosopher and
one of the last great Buddhist philosophers in the history of Indian philosophy, takes up
the issue in his Kṣaṇabhaṅgasiddhi which, being one of the last Buddhist texts on the
subject, stands as the Buddhists final word in this debate with the Naiyāyikas. By
examining the argument in this text, we can discern the key issues that motivate the
debate over endurance in the Indian tradition.
In his many and wide-ranging philosophical works, Ratnakīrti defends an ontology of
self-characterized particulars (svalakṣaṇa), each of which is momentary (kṣaṇika) and has
its own self-nature (svabhāva) by means of which it produces an effect and then is
destroyed in a moment (kṣaṇabhaṅga) by that very nature. Our ordinary experience of the
world as consisting of enduring objects is on this view a mistaken cognition resulting from
the imposition of concepts upon these particulars, imagining them to have universal
characteristics (sāmānyalakṣaṇa), when they are actually merely grouped together
according to our desires by means of their exclusion (apoha) from those things that do not
fulfil our purposes. Ratnakīrti employs rigorous logical methods to defend these views,
demonstrating his mastery of the logical techniques of the Nyāya school by framing his
arguments in such a way as to conform to Nyāya standards.1
The view that things exist for only a single moment gains currency among some
Buddhists in the first centuries of the Common Era.2 The earliest arguments derive the
conclusion that things are momentary from the mere fact that they change, but this
argument depends upon the controversial assumption that things are merely collections of
properties (dharma).3 Vātsyāyana criticizes this argument, insisting that a substance
(dravya), being a property-bearer (dharmin) distinct from its properties, can endure even
when it undergoes a change of properties.4 Buddhists later begin to shift to the argument
that things are momentary because they are destructible. If being destructible is inherent
in the self-nature of any thing, they argue, then it must be destroyed immediately just by
its own nature as soon as it comes to exist.5 This argument depends on the controversial
premise that things cannot be destroyed by an external cause. Uddyotakara, who
criticized this argument extensively, rejects this premise and furthermore insists that a
thing can have the capacity to be destroyed and yet fail to exercise that capacity in the
absence of auxiliary causes.6 In this way, the nature of causality becomes relevant to the
dispute over momentariness.
The issue of causality becomes even more central to the debate when the Buddhists
offer their most sophisticated argument for momentariness: the argument from causal
efficiency. This argument was first proposed by Dharmakīrti in the Hetubindu, where he
defends a version of the argument from spontaneous destruction against the objection
that not everything is destructible. To this Dharmakīrti responds that such a thing could
not exist because a non-momentary entity “lacks the capacity to produce an effect either
successively or simultaneously.”7 After Dharmakīrti, the debate focuses exclusively on this
argument and the nature of causality becomes the main issue of contention, leading both
sides to develop complex and detailed accounts of causality. The issue of momentariness
was taken up by Ratnakīrti’s teacher Jñānaśrīmitra, in his Kṣaṇabhaṅgādhyāya, where he
defends the doctrine against such Naiyāyikas as Vācaspati Miśra and Bhāsarvañja.8
Ratnakīrti formulates Jñānaśrīmitra’s argument in a concise and rigorous form, which
represents the most sophisticated statement ever offered by the Buddhists on the issue of
momentariness.9
In this chapter, I will discuss Ratnakīrti’s formulation of the argument, focusing on the
role of causality in his reasoning. I will first present Ratnakīrti’s main argument, an
inference from existence, showing how the issue of whether there are unrealized
capacities becomes the central point of contention. I will then discuss Ratnakīrti’s defense
of the denial of unrealized capacity and of his claim that capacity and incapacity are
incompatible properties. Finally, I will turn to Ratnakīrti’s own account of causality,
showing how he attempts to address the objections of his Nyāya critics.

Inference from existence

Both the advocates and opponents of momentariness adhere to a common inferential
pattern, whereby an inferential subject (pakṣa) is known to be qualified by the probandum
(sādhya) because it is known to be qualified by a prover (hetu), which is able to serve as a
reliable sign of the probandum because the two properties stand in a relationship of
universal pervasion (vyāpti). This universal pervasion can be supported by evidence in two
ways: by positive correlation (anvaya) and negative correlation (vyatireka). Positive
correlations are cases which are “known to be qualified by the probandum” (sapakṣa), and
which are also qualified by the prover. Negative correlations are “cases known not to be
qualified by the probandum” (vipakṣa), and which are also not qualified by the prover. For
instance, we may infer that the mountain is fiery from the observation that it is smoky.
The universal relation between smoke and fire is supported both by positive correlations
(smoky things that are also fiery, such as the kitchen hearth), and negative correlations
(non-fiery things that are not smoky, like the lake).10 A single counterexample—a case
where the prover is found, but the probandum is absent—can serve to refute such an
inference, because an inference is only valid if all cases qualified by the prover are also
qualified by the probandum. The inferential subject itself obviously cannot be cited as a
counterexample, but in fact any example may be disputed by an opponent and can thus
come to be included in the inferential subject.11 This principle, which Kisor Chakrabarti
has called the “general acceptability of inductive examples,” is important because
Ratnakīrti deploys it masterfully at key points in his argument to deprive his Nyāya
opponent of all possible counterexamples.12
Ratnakīrti frames his argument as an inference from existence: “What exists is
momentary, like a pot.”13 The prover here is existence (sattva) and the probandum is
momentariness (kṣaṇikatva), with a clay pot serving as the example supporting the
universal pervasion between existence and momentariness. The choice of the pot as the
supporting example, which is obviously question-begging from the point of view of the
opponent (given that the opponent will insist that the pot endures), reveals a few key
characteristics of Ratnakīrti’s argument. First, it is not an empirical argument following
from the observation of particular cases, despite its adherence to the formal requirement
of presenting a supporting example. The example of the pot is a mere arbitrary case of
something that exists and the proof of its momentariness will follow purely from the
definition of existence by means of the supporting inferences to come.14 The example also
brings up the problem of the scope of the inferential subject. If the inferential subject is
“everything that exists,” then the pot would have to come from within the inferential
subject (antarvyāpti), a procedure not accepted by Naiyāyikas, but advocated by some
Buddhists, most notably Ratnākaraśānti.15 Ratnakīrti avoids this dubious procedure by
designating the inferential subject as “these things which have become the subject of
dispute,” leaving open for the moment whether the pot will be something that will come to
be disputed, but aiming the argument at whatever entities the opponent might claim to
endure. In this way he is able to technically avoid relying on antarvyāpti, while still
fulfilling the requirements of a formal inference. He has his opponent immediately dispute
the example, and subsequent inferences are intended to infer the momentariness of the
pot, not the momentariness of everything. In this way, Ratnakīrti is able to prove the
momentariness of everything without running afoul of the problem that an argument
concerning everything is deprived of all possible confirming instances (asādhāraṇa),
which will by definition fall within the inferential subject. The focus of the argument thus
shifts to the pot, which now becomes the inferential subject.
Ratnakīrti proposes to demonstrate the momentariness of the pot by pointing out the
problematic consequences (prasaṅga) following from the assumption that the pot endures.
Ratnakīrti defines existence as “capacity to produce an effect”(arthakriyākāritva). It
follows that the pot as an existing thing must be capable of producing an effect at any
given moment. If the pot exists in the next moment, he argues, there are three
possibilities: (1) it produces the same effect, (2) it produces a different effect, or (3) it
produces no effect at all. It cannot produce the same effect because that effect has
already been produced, but it must produce some effect or else it will not exist at all. That
leaves only option (2), but Ratnakīrti argues that this option implies momentariness: in
the next moment the pot is capable of producing an effect that it was not capable of
producing in the earlier moment. To regard the same object as both capable and
incapable of producing a certain effect would require us to superimpose incompatible
properties (virūddhadharma) upon a single thing. It follows that the pot at any given
moment must be distinct from the pot at future moments.16 The unfortunate consequence
(prasaṅga) that follows from the endurance of the pot is that it will both be capable and
incapable of producing its future effects right now. Through a transformation of this
unfortunate consequence (prasaṅgaviparyaya), Ratnakīrti constructs an inference with
“non-production of future effects” as the prover and “being distinct from the pot in the
future” as the probandum. In this way, he demonstrates the momentariness of the pot and
secures the example for his inference from existence.
This argument for the momentariness of the pot depends on the crucial and
controversial assumption that capacity (śaktatva) amounts to production (karaṇatva).
Ratnakīrti declares that there can be no delay between the capacity and the production,
because it would lead to the unfortunate consequence that the effect would not be
produced in the present moment. He rejects the theory that an auxiliary cause is required
for the capable thing to exercise its capacity, stating that it would be “improper that there
need be an auxiliary cause for something that is capable,”17 adding that “something not
producing something else is not entitled to be spoken of as producing that other thing.”18
Ratnakīrti reconstructs this as a formal inference: “That x at time y appropriately spoken
of as productive of z, that x at that time y produces z, like the final complete collection of
causal factors (sāmagrī) produces its effect.”19 Here the pot is again the inferential
subject, the prover is “being appropriately spoken of as productive of something,” and the
probandum is “producing that thing.” The universal pervasion is supported by the
example of the complete collection of causal factors, which on the Nyāya view is indeed
immediately productive given that it is by definition a collection of causal factors
sufficient for production of the effect. This second inference shores up Ratnakīrti’s main
argument by offering justification for the assumption that capacity must be exercised
immediately.
Although the Naiyāyika cannot deny that the sāmagrī provides positive evidence in
favor of immediate production, they need only identify a single mutually acceptable
counterexample to show that the inference is inconclusive (anaikāntika). Anticipating this
objection, Ratnakīrti has his opponent offer the example of a seed in a granary as
something capable of producing a sprout and yet not producing it until the arrival of
auxiliary causes such as soil, water, and so on.20 Ratnakīrti, however, immediately
disputes the example, insisting that the seed in the granary is only called capable
metaphorically (aupacārika), because it is one cause in a series that eventually produces
the sprout. In the literal (pāramārthika) sense, he maintains, it is not appropriate to call
the seed capable.21 Relying on the principle of the general acceptability of inductive
examples, he is thus able to help himself to the positive evidence supporting his universal
pervasion (given that such examples are accepted by both sides) while nonetheless
excluding all possible counterexamples, which remain under dispute and thus part of the
inferential subject. In the absence of any undisputed counterexamples, the opponent is
unable to refute the inference.
Is the distinction between metaphorical and literal capacity question-begging at this
stage of the argument? Ratnakīrti may seem to be assuming the momentariness of the
seed in order to justify his rejection of the counterexample. Yet he does not assume that
the seed is momentary: he merely insists that it too be included in the inferential subject,
which cannot be used as evidence by either party to the discussion. Our metaphorical way
of speaking about something as capable when it is not literally capable is merely an
explanation of why we call the seed in the granary capable despite its lack of capacity in
the literal sense. Such an explanation does not require the assumption that the seed is
momentary, it only needs to show that if the seed were momentary, then our conventional
discourse regarding its capacity could be explained in this way.
Ratnakīrti also provides further support for the distinction by arguing that if we do not
restrict the literal meaning of capacity to immediate production, then “there would be no
restriction with respect to proper talk of productiveness for anything anywhere.”22 If we
say that a thing is capable despite the fact that it cannot produce until the arrival of an
auxiliary cause, then anything could be said to be productive of anything simply by
stipulating that some auxiliary condition would be needed for it to actually produce that
thing. For instance, I cannot now lift a weight of two hundred pounds, and I am generally
not regarded as capable of doing so. Nonetheless with some training and effort I might
someday successfully lift a weight of two hundred pounds. Perhaps, then, I can be called
capable now after all, and regard my lack of training as an auxiliary cause that has not yet
arrived. On the other hand, I am probably not capable of lifting one thousand pounds no
matter how diligently I train. Yet if I were supplied with sufficiently powerful bionic arms,
and I then lifted one thousand pounds, who could deny that I was capable at that
moment? Perhaps we should say that I am right now capable of lifting one thousand
pounds, but unable to exercise that capacity due to my lack of bionic arms. If the proper
use of the word “capacity” is not restricted to its most literal sense of immediate
production, then it seems to follow that I can properly be called capable of lifting any
amount of weight.
We use a variety of metaphorical senses of the word “capacity” in various contexts and
this is unproblematic as long as our meaning is clear. If we are asking whether I could
right now pick up that two hundred pound weight, then I am incapable. If we are asking
whether I have the long-term aptitude to lift two hundred pounds, then perhaps I am
capable. If we are asking what I am capable of given technology that could increase my
strength beyond normal human limits, then I am capable of almost anything. None of
these is the one and only proper application of the concept of capacity: each one in its
context employs a different metaphorical concept of capacity depending on which missing
factors are relevant in that context. All of these metaphorical senses of capacity, however,
rely on literal capacity—the immediate productiveness generated when the complete
collection of casual factors is assembled.
This problem of the unrestricted application of the concept of capacity can be
reconstructed as a formal inference as follows: “That x which at time y does not produce
z, that x at that time y is not suitable to be properly spoken of as capable of producing
with respect to z. Just as barley, which does not produce a rice-sprout, is not spoken of as
the cause of a rice-sprout.”23 This version of the argument relies upon the negative
correlations (vyatireka) supporting the inference of immediate production from being
appropriately spoken of as capable. The pot does not produce its future effects right now.
Anything that does not actually produce a certain thing is not appropriately spoken of as
capable of producing that thing. Therefore the pot is not appropriately spoken of as
capable of producing those effects. Here the evidence is a barley seed, which everyone
agrees is not appropriately spoken of as capable of producing a rice-sprout and which also
does not produce a rice-sprout. The rice seed would not be an effective counterexample
because it is under dispute as well, with Ratnakīrti insisting that it too can only be spoken
of as capable in a metaphorical sense. Once again, under the principle of the general
acceptability of inductive examples, the opponent is deprived of all possible
counterexamples, but must accept the evidence of the undisputed cases as support for the
pervasion between capacity and immediate production.
Despite the complexity of its structure, Ratnakīrti’s argument rests on a small number
of assumptions: (1) Existence is capacity to produce an effect. (2) Anything appropriately
regarded as capable of producing a certain effect must produce that effect immediately.
(3) Capacity and incapacity are incompatible properties which cannot qualify the same
object. The pot can be shown to be momentary because it exists, and so by (1) it must
produce some effect now, but it does not produce its future effects now. By (2) then it is
not now capable of producing those effects and by (3) it is distinct from the future pot
which is capable. Since by (1) anything that exists will have some capacity, it follows that
all things that exist are momentary. Of these three controversial assumptions, (1) does not
generate much debate, and it is generally accepted that existence is pervaded by capacity
to produce an effect. The other two assumptions remain the subject of intense dispute
between Ratnakīrti and his Nyāya opponents, and Ratnakīrti defends both against various
objections. I will next examine Ratnakīrti’s defense of the assumption that causality is
immediate productivity, then I will turn to his defense of the claim that capacity and
incapacity are incompatible properties. In the final section I will discuss Ratnakīrti’s
positive account of causality and the role of auxiliary causality in that account.

The denial of unrealized capacity

At the heart of this dispute is a disagreement over whether causality should be
understood primarily in terms of necessary or sufficient conditions. Ratnakīrti has insisted
all along that the sense of “capacity” in his definition of existence as “capacity to produce
an effect” must be understood as requiring that each thing that exists be a sufficient
condition for the production of something else. The Nyāya school on the other hand
maintains that a thing can be considered capable of producing an effect if it is merely a
necessary condition for the production of that effect. On this view, the arrival of auxiliary
causes (sahakārin) trigger the exercising of a previously unrealized capacity. Ratnakīrti
defends his account of causality in terms of sufficient conditions by examining the Nyāya
account of auxiliary causality more carefully, arguing that it cannot undermine the force
of his argument from causal efficiency.
Ratnakīrti allows his opponent to elaborate the theory of auxiliary causality, arguing
that “even though the pot is capable, it will produce an effect in a sequence through its
dependence on a sequence of auxiliary causes.”24 The opponent defines capacity
(samartha) as the entire collection of auxiliary causes (sahakāri-sākalyam): although there
is no active production in the absence of some auxiliary condition, he argues we should
still regard the rest of the collection as capable, because it would produce if those missing
factors were to arrive.25 Ratnakīrti replies that even if we accept the definition of capacity
as the total collection of causal factors, such a collection produces if and only if it has the
nature of producing at that moment. Consequently, if it lacks some causal factors, then it
is not yet capable, and if it were capable without those causal factors, it would produce
without delay.26 This is a direct consequence of locating the causal capacity in the self-
nature (svabhāva) of a thing: whenever anything has such a self-nature it must produce
immediately. If the earlier collections lacking some auxiliary factor fail to produce, then
they cannot have such a self-nature.
The opponent nonetheless continues to insist that an entity can produce by its own
nature (svarūpa) and yet not produce “by that nature alone,” because it can have the
nature of producing only in conjunction with certain other things. To defend this view, the
opponent draws a distinction between two kinds of capacity: innate (nija) and acquired
(āgantuka). The innate capacities are those possessed by a thing without the assistance of
auxiliary causes and the acquired are those brought about by the auxiliary factors.
Consequently, a thing with a certain innate capacity could continue to endure with that
capacity while acquiring other capacities from the auxiliary factors.27 In this way an
enduring thing could produce a series of different effects all while retaining its basic
innate capacities, acquiring the further power to discharge those capacities as the
auxiliary causes arrive in sequence. For instance, a piano may produce a series of notes as
the finger strikes each key in sequence: it retains the innate capacity to play each note,
but acquires the immediate power to produce each note at the moment of the arrival of
the finger on each key.
Ratnakīrti does not deny that we could draw this kind of distinction between innate and
acquired capacity, but he insists that it cannot apply to capacity in the literal sense.
Ratnakīrti holds that the producer is a self-characterized particular (svalakṣaṇa) with its
own unique self-nature, which includes its capacity to produce various effects. We group
together such individuals in various ways according to our purposes and desires, but only
the individuals exist, and the capacity of the groups derives ultimately from the capacity
of these individuals. We can only make a distinction between innate and acquired capacity
in a case where a group of such individuals is lacking some other individual whose arrival
would trigger the production of the effect. Depending on our desires and purposes we can
then adopt various metaphorical ways of speaking about whether a given group is capable
or not. When we are speaking literally about what happens at the moment of production,
however, we are talking about the self-nature of the producer. When causal power is
actually exercised, it is always some individual with the inherent power to produce that
actually produces. This individual, whatever it is, either has the power to produce or it
does not: it cannot both have and not have that power.28 If it has an innate capacity which
is not exercised until another capacity is acquired, then it does not yet have the capacity
to produce the effect; it merely has the capacity to produce the cause of the effect. For
instance, we may speak of the piano metaphorically as having the innate capacity to
produce a series of different effects depending on the arrival various auxiliary causes.
Nonetheless, speaking literally, we must admit that in order to trigger the production of
each note, the auxiliary causes must generate a new capacity in the piano making it
capable of producing that note at that moment. It still cannot produce the effect unless it
actually has that new capacity.
The opponent attempts to escape the force of this reasoning by maintaining that
although the effect depends on the whole collection of causal factors, the cause does not.
The cause can be a cause independently of the auxiliary conditions even if the effect is not
produced without those conditions. The problem here is that we have to imagine that the
cause can acquire the capacity to produce the effect from the auxiliary factors and yet
remain independent of those very auxiliary factors which give it the power to produce that
effect. It then no longer seems to play any role in the causal process, and it would thus
become possible that an effect could come to be without a cause.29 The opponent tries to
account for this anomaly (aparādha) by suggesting that the capable cause can exist and
yet the effect still not come to be. Ratnakīrti argues that this would undermine the
fundamental principle of causality: whenever the cause exists, the effect comes to be.30
Without this principle, there would be no causal regularity, because causes would not
always produce their effects. This fundamental principle of causality depends upon the
immediate productiveness of the fully capable cause, because only the fully capable cause
invariably produces its effect. Metaphorical senses of capacity do not conform to this
principle of causal regularity.
The opponent attempts to escape this line of reasoning by shifting from a definition of
causality in terms of positive correlation (anvaya) to a definition by negative correlation
(vyatireka). Ratnakīrti has defined causality in positive terms: when the cause occurs, the
effect occurs. This definition stresses that causes are sufficient conditions. The opponent
suggests defining causality in negative terms: when the cause does not occur, the effect
does not occur. This stresses that causality can be understood in terms of necessary
conditions.31 This maneuver, however, explains the anomaly of something being a cause
even though it does not produce the effect only at the cost of completely disconnecting
cause and effect from each other: we would have no reason to expect the effect given the
cause. Moreover, Ratnakīrti argues, there is no guarantee that absence of the effect is due
to the absence of the cause, since under this definition the absence of the effect can occur
when the cause is present.32 The strategy of shifting to a negative definition of causality
cannot evade Ratnakīrti’s argument that causality must be understood primarily in terms
of sufficient conditions for production of an effect.
Perhaps the most interesting move considered by Ratnakīrti is the possibility of
indexing the capacity to time, so that even at the current moment a thing can have the
self-nature of producing the later effect at that later time. If something has the “nature of
producing something later on,” asks the opponent “why should it produce that effect in
the beginning?”33 Here the opponent, who stands in for Bhāsarvañja,34 a Nyāya
philosopher who advanced this line of reasoning, suggests that each capacity of a thing is
a capacity to produce at a certain time. Consequently, there is no contradiction between
having the capacity to produce later and not producing at the present moment.
Ratnakīrti’s usual rejoinder that something must exhibit its self-nature as long as it exists
does not so easily deflect this move. The opponent stands firm, insisting that “given its
endurance, its self-nature is just that it produces at a later time.”35 There is no problem
with the object enduring and continuing to possess a capacity to produce the later effect
later. It does not have to repeatedly produce a current effect because at each moment it
has only the capacity to produce the future effect in the future.
Ratnakīrti identifies several problems with the strategy of the time-indexing of capacity.
First, he argues, it would undermine our confidence in inference, because we would be
unable to rule out that current smoke might not have been caused by past fire: we would
thus be unable to infer fire from smoke.36 If smoke observed today could have been
caused by a fire that occurred years ago, then I would have no reason to expect to find
fire on the mountain when I observe smoke there. He also presents a dilemma: is the
effect produced in the earlier moment or the later moment? If it is produced in the earlier
moment, then it will exist in that moment, contradicting the opponent’s claim that it does
not exist until later. If the effect is produced later, however, then the cause, which exists
at the earlier time, cannot be producing it. Ratnakīrti maintains that actual production of
an effect must take place at the moment the effect comes to be and that there could be no
causal connection between the putative earlier cause and the later thing which is alleged
to be its effect.37
As we will see, Ratnakīrti’s rejection of the Nyāya account of auxiliary causality does
not preclude a role for auxiliary causes in his final account of causality as sufficient
capability. Nonetheless, he has argued that the role of the auxiliary causes as necessary
conditions that arrive sequentially is still derivative upon the sufficiency of the producer
of the effect at the moment of production. The auxiliary causes may be necessary in the
sense that they are needed to generate a producer completely capable of producing the
effect, but at the moment of production, the job of those auxiliary causes is already
finished and the producer acts alone in producing the effect. The primacy of sufficient
conditions over necessary conditions is a consequence of understanding causality in terms
of causal powers which are inherent to the self-nature of the cause. This inherence
excludes the possibility of outside assistance at the moment of production, although it still
allows the auxiliary causes to play a role in bringing about something with a productive
self-nature. However, even the role of the auxiliary causes must be understood in terms of
their inherent capacity to bring about the capable cause. Thus we cannot have an account
of auxiliary causality purely in terms of necessary conditions: the necessity of those
conditions must be understood in terms of their collective sufficiency.

The incompatibility of capacity and incapacity

Before turning to Ratnakīrti’s positive account of causality, it is worth pausing to examine
his defence of his second controversial premise: capacity and incapacity are incompatible
properties, which cannot qualify one and the same object. It is on the basis of this premise
that Ratnakīrti is able to draw the conclusion that the seed in the granary is distinct from
the seed that produces the sprout: the seed in the granary must be distinct from the seed
that gives rise to the sprout, because the latter is capable and the former is not. These
qualities, Ratnakīrti argues, are not just different from each other: they are truly
incompatible. This is not a case where a blue object is mistakenly perceived as yellow. In
that case we superimpose the property of being yellow on something that is actually blue,
but producing blue and producing yellow are not contrary properties. The same object can
produce multiple different effects; what it cannot do is both produce and not produce the
same effect.38 Ratnakīrti considers two interesting objections to this claim: (1) by the
same reasoning things at a time would be split into multiple entities and (2) just as
capacity and non-capacity are not contrary in different places, they are not contrary at
different times. Both of these objections are based upon an analogy between space and
time which Ratnakīrti rejects, stressing that the problem in each case is that the
imposition of such properties would lead to a contradiction.
The opponent first argues that if contrary properties show that we must distinguish
between the thing at different times, then the presence of multiple capacities in a single
thing at a given time would require us to also distinguish a distinct thing for each
capacity. The opponent gives the example of a lamp, which is consuming both the wick
and the oil. If earlier and later stages must be distinguished because they have different
capacities, then even at a single time things will have to be regarded as multiple. The
example of the lamp again stresses the role of the auxiliary causes, which in the example
become “implicated in the nature” of the main cause, resulting in its not producing as a
single thing.39 The charge of the opponent is that Ratnakīrti will not be able to make sense
out of production by a self-nature if he is forced to admit that the producer has multiple
self-natures.
In his reply to this objection, Ratnakīrti stresses that his argument only depends on the
assumption that contrary properties may not qualify one and the same thing. It does not
depend on the assumption that different qualities and capacities cannot qualify one and
the same thing: “Production of the sprout is contrary to non-production of it, but it is not
contrary to the production of another.”40 Ratnakīrti does not deny that things like the lamp
and the seed have multiple self-natures insofar as they are indeed collections of many
different properties (dharmas), each of which is itself a self-characterized particular
(svalakṣaṇa) which produces by means of its own self nature.41 However, consuming the
wick and consuming the oil are not incompatible properties: they can qualify one and the
same object. It does not matter whether we adopt the Nyāya view that substances exist as
distinct from their various properties or the Buddhist view that things are merely
collections of properties. Objects may have multiple properties, but they cannot have
contrary properties. The multiple properties that things exhibit in exercising their various
capacities may be due to multiple natures or may be aspects of a single nature, but in
neither case can they have contrary properties or natures.
The second objection deploys the analogy between space and time in another way.
Capacity and incapacity are not contrary properties when they occur at different times,
the opponent argues, just as they are not contrary when they occur in different places.
This objection can again be traced to Bhāsarvañja, who earlier proposed indexing the
capacity to time.42 Here he suggests something similar: the capacity to produce a certain
effect at a certain time and the non-capacity to produce that effect at a different time are
not really contrary. This is supported by the analogy to place: we do not say that capacity
to produce at a certain place is contrary to non-capacity in a different place. A seed
planted in my front yard does not give rise to a sprout in the backyard. Ratnakīrti rejects
this analogy between space and time, arguing that the two cases are different. Whether
the properties are contrary is based upon whether there is or is not opposition. There is
no opposition between capacity to produce in one place and non-capacity in another place
and so the properties are not contrary.43
The opponent, however, presses the point, insisting that there is also no opposition
between capacity to produce an effect at a certain time and non-capacity to produce that
effect at another time.44 The analogy does indeed have some strength if we accept the
indexing of capacity to time. Ratnakīrti has effectively admitted that capacity is indexed to
place, because the effect is produced at a certain location and not another. This is why
there is no contradiction between being capable of producing the effect in one location
and not capable of producing that effect in another place. If capacity is also indexed to
time, so that every capacity is a capacity to produce at a certain time, then a similar result
would follow with respect to time. Capacity to produce at one time and non-capacity to
produce at another would not stand in opposition.
Ratnakīrti begins his reply by defining “opposition” as “a state of mutual avoidance,
whose rule is the non-presence of two contrary properties in a single property-bearer,”
offering existence and nonexistence as an example.45 There are positive and undisputed
examples of things that produce a certain effect in one place but not at others, such as the
seed producing its sprout in one particular location and not others: we all agree that it is
the same seed that produces here and not there. There are not, however, similar
undisputed examples of a single thing that produces a certain effect at one time, but not
at another. All such examples that would tend to support the view of endurance are at
issue in this debate and cannot be appealed to as evidence. Ratnakīrti argues that there is
opposition in the case of being capable at one time, but not at another, because it has
been established by inference. If a cause endured it would have to produce the same
effect again in the next moment, but even if it produced a similar effect, it could not be
that same effect, because that effect has already been produced. Ratnakīrti has already
rejected the idea that capacity can be indexed to time on the grounds that it would render
inference impossible. Indeed, Ratnakīrti seems to suggest that it is the effect that should
be indexed to time, arguing that “the opposition between capacity and non-capacity
cannot be thrown out on the presumption that they pertain to different times, because
there is only the thing’s own occasion (sva-samaya)”46 To be a certain effect is to come to
be at a certain time. A similar effect at a different time would not be the same effect. To
be incapable of producing that thing, fixed as it is at its time, is contrary to that capacity,
whether at the time of production or at any other time. There is thus an opposition
between capacity to produce an effect and non-capacity to produce that effect even at
different times.
 Ratnakīrti’s account of causality
Ratnakīrti begins to develop his own account of causality by presenting himself with a
pair of dilemmas: does the fully capable cause produce the effect dependently or
independently and does it produce the effect as a unity or a multiplicity?47 The opponent
will charge that a momentary thing can produce neither dependently nor independently
and neither as a unity nor as a multiplicity. Ratnakīrti will argue that the momentary self-
characterized particular produces independently and as a single thing. The question will
then be what is the role of the auxiliary conditions in an account where the cause
produces independently and by itself? Ratnakīrti will locate the role of the auxiliary
factors in their contribution to the generation of the sufficient cause, which then acts
alone at the moment of production. Ratnakīrti’s account will thus distinguish between two
levels of discourse regarding capacity: an ultimate level where capacity is taken literally
as sufficiency for production of the effect and a conventional level at which we speak
metaphorically about something as a cause because through a series of encounters with
auxiliary conditions it produces a successor which produces yet another in a series
(santāna) which eventually yields the fully capable cause and then the effect.
The opponent first asks whether the fully capable cause produces the effect
independently or dependently. If it relies on the assistance of the auxiliary causes, he
argues, then it cannot be momentary because there is no time in a single moment for the
auxiliary cause to act upon the producer of the effect.48 However, a momentary thing
cannot produce independently because it needs the auxiliary conditions to produce the
effect. If the final seed-moment acts independently, then why should it produce a sprout,
rather than another seed-moment like the others earlier in the series? This would deprive
the auxiliary conditions of any meaningful role in the process.49 The opponent then raises
a second dilemma: does the fully capable cause act as a unity or as a multiplicity? If the
auxiliary conditions act together with the seed at the moment of production to produce
the sprout, then the cause is deprived of its unity.50 Consequently, if we regard the
auxiliary conditions as distinct from the main cause, we should see multiple effects, but if
we do not regard them as distinct, then the auxiliary causes are implicated in the nature
of the producer: the presence of soil, water, etc. would be part of the nature of the seed.51
A momentary thing thus could not produce dependently nor independently nor as a unity
nor a multiplicity.
Ratnakīrti first rejects the thesis that the fully capable cause produces dependently. At
the ultimate level of discourse, where we are talking about the capacity in the literal
sense, there can be no assistance provided to the fully capable cause by anything else at
the moment of production, whether by means of some kind of conjunctive causal power
(sambhūyakaraṇa) or even just by virtue of being situated together with the auxiliary
factors. In either case, Ratnakīrti acknowledges there is the problem that there is no time
for a momentary thing to interact with anything else.52 At the ultimate level of discourse,
Ratnakīrti thus comes down firmly on the independence horn of the dilemma: at the
moment of production, the fully capable cause produces independently.53
The problem for an account of causality which maintains that the fully competent cause
acts independently is finding a meaningful role for the auxiliary conditions in spite of their
lack of assistance at the moment of production. The opponent immediately raises the
issue, asking, “If it is independent why does it not produce even in the absence of the
gathering conditions?”54 Just because the capable cause produces independently, however,
does not mean that it is not itself dependent upon previous causes for its capable nature.
At the conventional level of discourse, where we are talking metaphorically, Ratnakīrti
maintains, we will be able to say that the auxiliary causes do assist in the production of
the effect, because they contribute to the production of the fully competent cause.
Ratnakīrti argues that the auxiliary conditions such as soil and water play a meaningful
role in the production of the sprout because it is a sequence of encounters with such
conditions that makes possible the production of the capable seed that actually produces
the sprout.55 The earlier seed-moments in the series are not capable of producing a
sprout, but in conjunction with the auxiliary conditions their successors can obtain new
capacities. The seed-moment in conjunction with soil and water will generate a successor
and another in a chain that eventually generates the fully competent cause which in turn
produces the sprout. Without the previous encounters with these auxiliary factors, the
capable seed-moment could not arise. The auxiliary factors and the earlier seed-moments
are thus all necessary conditions for the eventual production of the sprout and can thus be
called causes of the sprout metaphorically. When speaking literally about what is true in
the final analysis, however, capacity can only be causal sufficiency, and the final seed-
moment produces by itself, having been generated as fully competent by the previous
influence of the auxiliary conditions. This account provides a meaningful role for the
auxiliary conditions and one which explains our conventional designation of the earlier
non-productive seed-moment as having capacity to produce the sprout.
Ratnakīrti considers several problems with his account of causality. First, if the earlier
seed-moment is to be metaphorically designated a cause because it is a member of a
series which actually produces a sprout, then it follows that not all of the seeds in the
granary can be considered causes of sprouts even at the conventional level, because not
all series will actually encounter all of the necessary conditions. We designate things as
seeds because we expect them to be useful in fulfilling our desire for sprouts, and this is
why we store them in the granary, but if some are not capable even in the conventional
sense, then why do we treat them all similarly? The opponent points out that all the seeds
in the granary are seen to be similar and are prepared and stored in same way and asks
why there is a difference between seeds that produce and those that do not.56 Here the
issue is whether this account of causality can find a meaningful role for the seedhood of
the seed. Since Ratnakīrti rejects the existence of universals such as seedhood, he needs
to explain why we store seeds in the granary rather than other things, such as stones.57
Ratnakīrti responds to this objection by admitting that we don’t as a matter of fact
usually know which seed-series will and which will not lead to the production of a
sprout.58 We group seeds together and store them in the granary because they are similar
to things that have in the past actually given rise to sprouts through a series. There is no
guarantee that every seed in the granary would produce a sprout even if they were all
planted in the ground, nor can we be sure anyone will ever plant any given seed. Although
there is no universal seedhood that dwells in all seeds giving them a latent capacity for
sprout-production, things are conveniently designated as seeds because of our
expectation that they will fulfil our desire for sprouts, unlike stones which will not fulfil
that desire even if they encounter the auxiliary conditions.
Another objection to this account is that there is no justification for positing the fully
competent cause in the first place. Given that we cannot know the full set of needed
conditions for generation of the fully competent cause, the opponent asks, how can we
know that it has actually been generated?59 We cannot know by perception that the seed
is fully capable: it looks more or less like the earlier seeds in the series, and we can never
be sure if more auxiliary conditions are still needed. How then can we know that there is a
fully competent cause at the moment of production at all? Ratnakīrti replies that we can
infer that there must have been a cause fully capable of producing the effect from the
effect itself.60 Although it is true that we can never be sure when the fully competent
cause will arise, once we have the effect then we know something must have existed in
the previous moment which had the full capacity to produce the effect. This fully
competent cause produces independently at the moment of production without assistance
from the auxiliary conditions, but it only arises in the seed series that encounters these
auxiliary conditions, securing their role in the process.
Crucial to Ratnakīrti’s account is the distinction between a conventional and an
ultimate level of discourse with respect to causality. At the ultimate level capacity is being
a sufficient condition for production of the effect, and it follows that the fully competent
cause can only produce independently. At the conventional level, capacity is the ability to
produce with the addition of various missing necessary conditions, and whether
something is to be designated capable depends on the context. At this level of discourse,
we can talk about the producer depending on the auxiliary conditions because only by
encountering them did it itself come to be produced. Ratnakīrti can thus say that in one
sense the fully competent cause produces independently and in another sense
dependently. If dependence means “producing conjunctively,” he says, then the final seed-
moment just produces independently. If dependence means being “properly situated with
respect to the auxiliaries having been encountered,” then we should say that the final
seed-moment produces dependently, given that its own sufficient causal power is in turn
dependent on the previous encounters between earlier seed-moments and the auxiliary
conditions.61
Ratnakīrti deploys a similar response to the question of whether the fully competent
cause produces as a single thing or a multiplicity. Although we speak metaphorically
about the sprout having many different causes, at the ultimate level of discourse the fully
capable cause produces as a single thing with a single nature.62 There is consequently no
question of the cause being disunified and requiring the production of multiple effects.
Since the auxiliary factors do not assist at the moment of production they are not obliged
to produce additional sprouts.63 These factors nevertheless remain necessary conditions
for the production of the sprout and can be included among its causes at the conventional
level of discourse. The soil and water do not thereby become “implicated in the nature” of
the seed, making it a collection at the moment of production; rather the action of the soil
and water become “integrated into the effect” through a series of encounters with
previous seed-moments.64 Although its causal power does come indirectly from these
previous contacts, it has a single nature, which is defined by its particular set of causal
powers. Various causal powers may be the result of the contribution of various auxiliary
conditions, but they are all powers of the same single producer, which exercises them
independently as a single thing at the moment of production.
Ratnakīrti’s account of causality addresses the central objections of his Nyāya critics,
who complain that it cannot give a meaningful role to both the auxiliary conditions and to
seedhood. The auxiliary conditions, however, are given a prominent role in the causal
process, although not at the moment of production. As they arrive, they condition the
production of the next seed-moment in the series in such a way that it has new causal
powers the previous seed-moment did not have. This way of explaining the role of the
auxiliary factors also makes sense of why causality can be understood in terms of
necessary conditions: we metaphorically say that earlier stages are capable because if
certain auxiliary factors were encountered in sequence a later stage would produce some
effect. Although Buddhists do not posit universals like seedhood that all seeds might
share, the seeds in the granary can nonetheless be distinguished from things that will
definitely not produce sprouts whether the auxiliary factors are encountered or not.
Stones, for instance, are not stored in the granary, because of their exclusion (apoha) from
sprout production. It is because we regard seeds as capable in the derivative sense that
we store them in the granary, anticipating that we might bring about the encounter with
the auxiliary factors that would produce a sprout. As a matter of fact, however, seedhood
is pervaded by sprout production at neither the ultimate nor at the conventional level of
discourse. At the ultimate level, the seeds in the granary are simply not capable of
producing sprouts. At the conventional level, we can never be sure that a given seed will
actually encounter the proper sequence of conditions to generate a series that gives rise
to the immediate producer of the sprout. However, once the sprout is produced we can
always infer that a sufficient cause existed, and so the completely competent cause is not
posited without justification.

Conclusion
Ratnakīrti brilliantly takes advantage of the inferential pattern accepted across the
classical Indian tradition to construct an argument that meets the rigorous standards of
that tradition. By introducing the example of the pot in support of his main inference from
existence, and then offering another set of inferences in favor of the momentariness of the
pot, he is able to avoid the problem that an argument concerning everything that exists
would be deprived of all positive examples (asādhāraṇa). His argument does depend upon
the controversial claim that there can be no unrealized capacity to produce an effect, but
he offers both positive and negative instances to support the pervasion between capacity
and production, and he successfully deploys the principle of the general acceptability of
inductive examples to show that all possible counterexamples are unavailable as evidence
to the contrary. He furthermore offers a coherent account of causality as sufficiency that
accounts for our ordinary conventional discourse regarding capacity, while arguing that
such conventional discourse is only possible on the basis of causal sufficiency.
Ratnakīrti’s argument, despite its technical cleverness, did not persuade his Nyāya
opponents, and the debate over momentariness ends in a kind of stalemate, with each side
adopting an internally coherent but mutually contradictory set of views. Udayana, a
Naiyāyika who was a contemporary of Ratnakīrti, repackaged and sharpened many of the
Nyāya objections to the argument from causal efficiency in his Ātmatattvaviveka.65
Unfortunately, because of the decline of Buddhism and its eventually extinction on the
Indian subcontinent in the century following Ratnakīrti, no great Buddhist author ever
offered a reply to Udayana. Although the results are hardly decisive, the exchange
between the Naiyāyikas and the Buddhists over the doctrine of momentariness does
reveal a close connection between the issue of identity over time and the nature of
causality. Whether one is willing to accept the Ratnakīrti’s argument for momentariness
depends in the end on whether one is willing to accept his account of causality.

Notes
1. Many of Ratnakīrti’s philosophical works in Sanskrit can be found in Thakur (1975). Several of
these works are also included in Śāstrī (1910). English translations of some of these works are
available in McDermott (1969), Patil (2009; 2011), Feldman and Phillips (2011), and Ganeri (2012).
2. Rospatt (1995, pp. 15–28). Although the doctrine of momentariness becomes extremely influential
among many schools of Buddhism, it is not accepted by all Buddhists. Nāgārjuna, for instance, the
founder of the Mādhyamika school, rejects the doctrine and argues against it extensively in his
masterwork, the Mūlamadhyamakakārikā. For a translation and discussion of that argument, see
Garfield (1995, pp. 267–274).
3. The argument from change is found in many sources including Asaṅga’s Śrāvakabhūmi and
Mahāyānasūtrālaṅkārabhāṣya, and the Nyāyānusāra of Saṃghabadra. See Rospatt (1995, pp. 154–
155).
4. Vātsyāyana, Nyāyasūtrabhāṣya 4.1.34–36; 3.2.10–14. See Gangopadhyaya (1982).
5. The argument from destruction is also found in many sources, including the
Abhidharmasamuccayabhāṣya and Viniścayasaṃgrahaṇī of Asaṅga, and Vasubandhu’s
Abhidharmakośabhāṣya. See Rospatt (1995, pp. 182–238). For a full translation of the
Abhidharmakośabhāṣya, where the argument is elaborated in detail, see Pruden (1988).
6. Uddyotakara, Nyāyavārṭika 3.2.14. See Jha (1999, pp. 1303–1315).
7. Dharmakīrti, Hetubindu 2.28. See Gokhale (1997, p. 50). See Gupta (1990), for a discussion of the
role of causality in Dharmakīrti’s argument.
8. For a discussion of Jñānaśrīmitra’s works, see Thakur (1987), which also contains the Sanskrit text
of the Kṣaṇabhaṅgādhyāya.
9. Thakur (1975, pp. 1–25).
10. The Nyāya theory of inference is discussed extensively in the Nyāyasūtra and its commentaries.
See especially 1.1.33–39, in Gangopadhyaya (1982) and Jha (1999). See also Chatterjee (1939, pp.
299–316). The Buddhist theory of inference was first developed by Dignāga in his
Pramāṇasamuccaya. For a translation, see Hattori (1968). For a discussion of Dignāga’s theory of
inference, see Hayes (1986). For further discussion of Indian theories of inference, see Matilal
(1998).
11. Nyāyasūtra 1.2.5. See also Chatterjee (1939, p. 309).
12. Chakrabarti (1999, pp. xii–xvi). Chakrabarti considers the plausibility of the principle, noting
possible objections to it, but he also makes clear that Nyàya authors rely on it frequently in
defense of their own realist positions.
13. Kṣaṇabhaṅgasiddhi (KBS) 67.6. All references to KBS give the page and line numbers from
Thakur (1975). The translation is from Feldman and Phillips (2011, p. 42). Ratnakīrti divides his
text into two parts. In the first part, he presents an inference based on positive correlation
(anvayātmikā): everything that exists is momentary. In the second part, he presents an inference
based on negative correlation (vyatirekātmikā): everything that is non-momentary is nonexistent.
For an English translation of the Vyatirekātmikā and a discussion of the peculiar epistemological
issues it raises, see McDermott (1969). See also Gupta (1990).
14. In the Kṣaṇabhaṅgādhyāya, Jñānaśrīmitra uses the example of a cloud, which is far more unstable
than a pot, possibly giving the impression that the argument is an empirical generalization.
Jñānaśrīmitra’s argument is also purely conceptual, but by using the example of the pot, Ratnakīrti
makes this much clearer.
15. For a discussion of Ratnakīrti’s position on the issue of antarvyāpti and his relationship to
Ratnākaraśānti, see Ruegg (1970) and McDermott (1972). See also Tani (2004) in Hino and Wada
(2004, pp. 375–383).
16. KBS 67.23–68.16. See Feldman and Phillips (2011, pp. 48–54).
17. KBS 68.4–10. See Feldman and Phillips (2011, pp. 50–52).
18. KBS 68.14–16. See Feldman and Phillips (2011, pp. 52–53).
19. KBS 68.17–18. See Feldman and Phillips (2011, pp. 53–54).
20. KBS 68.25–26. See Feldman and Phillips (2011, pp. 55–56).
21. KBS 68.26–30. See Feldman and Phillips (2011, pp. 56–57).
22. KBS 69.5–7. Translation by Feldman and Phillips (2011, p. 59).
23. KBS 69.11–13. Translation by Feldman and Phillips (2011, p. 60).
24. KBS 74.19–20. See Feldman and Phillips (2011, p. 97).
25. KBS 74.22–26. See Feldman and Phillips (2011, pp. 98–99).
26. KBS 74.27–75.4. See Feldman and Phillips (2011, pp. 99–100).
27. KBS 76.6–9. See Feldman and Phillips (2011, pp. 100–101).
28. KBS 76.9–13. See Feldman and Phillips (2011, pp. 108–109).
29. KBS 76.13–16. See Feldman and Phillips (2011, pp. 108–109).
30. KBS 76.17–20. See Feldman and Phillips (2011, pp. 109–110).
31. KBS 76.21–22. See Feldman and Phillips (2011, pp. 111–112).
32. KBS 76.22–29. See Feldman and Phillips (2011, pp. 112–113).
33. KBS 77.1–5. See Feldman and Phillips (2011, pp. 113–114).
34. Woo (1999, p. 211).
35. KBS 77.11. Translation by Feldman and Phillips (2011, p. 115).
36. KBS 77.11–12. See Feldman and Phillips (2011, pp. 114–115).
37. KBS 77.12–20. See Feldman and Phillips (2011, pp. 115–117).
38. KBS 77.20–22. See Feldman and Phillips (2011, pp. 117–118).
39. KBS 77.23–78.4. See Feldman and Phillips (2011, pp. 118–120).
40. KBS 78.10. Translation by Feldman and Phillips (2011, p. 121).
41. KBS 78.5–15. See Feldman and Phillips (2011, pp. 120–121).
42. Woo (1999, p. 223).
43. KBS 78.19–20. See Feldman and Phillips (2011, p. 123).
44. KBS 78.23–25. See Feldman and Phillips (2011, pp. 123–124).
45. KBS 78.25. Translation by Feldman and Phillips (2011, p. 124).
46. KBS 79.7–8. Translation by Feldman and Phillips (2011, p. 127).
47. This objection can be traced to Vācaspati Miśra; see Woo (1999, p. 229).
48. KBS 79.24–26. See Feldman and Phillips (2011, pp. 132–133).
49. KBS 79.13–24. See Feldman and Phillips (2011, pp. 129–132).
50. KBS 79.26–28. See Feldman and Phillips (2011, p. 133).
51. KBS 80.1–3. See Feldman and Phillips (2011, pp. 134–135).
52. KBS 80.15–19. See Feldman and Phillips (2011, pp. 139–140).
53. By taking this position, Ratnakīrti makes a sharp break with Dharmakīrti, who argued in the
Hetubindu that at the moment of production the seed acts together with its auxiliary causes to
produce the sprout. Dharmakīrti held that the seed and the auxiliary conditions are all individually
fully capable of producing the sprout at the moment of production. Given their individual
sufficiency, he argued that their cooperative causal power did not require the production of any
new feature in either the main cause or the auxiliary cause. They could thus produce together
without requiring any additional time for interaction (Hetubindu 2.21–2.25). See Gokhale (1997,
pp. 40–48).
54. KBS 80.19–23. See Feldman and Phillips (2011, pp. 140–141). The gathering conditions
(upasarpaṇapratyaya) are the various conditions which arrive sequentially over the course of time,
eventually bringing about the generation of the fully competent producer (kurvadrupa).
55. KBS 80.23–80.25. See Feldman and Phillips (2011, pp. 141–142).
56. KBS 80.26–28. See Feldman and Phillips (2011, pp. 142–143).
57. Ratnakīrti defends his rejection of universals and develops an account of generality in terms of
exclusion (apoha) in his Apohasiddhi. An English translation of that text can be found in Patil
(2011).
58. KBS 80.28–81.4. See Feldman and Phillips (2011, pp. 143–144).
59. KBS 81.5–6. See Feldman and Phillips (2011, p. 144).
60. KBS 80.6–9. See Feldman and Phillips (2011, p. 145).
61. KBS 81.15–18. See Feldman and Phillips (2011, pp. 148–150).
62. KBS 81.19–21. See Feldman and Phillips (2011, pp. 150–151).
63. Dharmakīrti argued that despite simultaneous production by many fully competent causes,
multiple effects would not need to be produced, because each cause would have only the power to
produce that same sprout (Hetubindu 2.11–2.14). See Gokhale (1997, pp. 26–31). By rejecting the
doctrine of conjunctive causal power, Ratnakīrti avoids the need for such an argument.
64. KBS 81.10–14. See Feldman and Phillips (2011, pp. 145–148).
65. See Dravid (1995) and Laine (1998).

Bibliography
Bagchi, S. (ed.) (1970), Mahāyānasūtrālaṅkāra of Asaṅga. Darbhaṅga: Mithila Institute.
Chakrabarti, K. (1999), Classical Indian Philosophy of Mind. Albany: SUNY Press.
Chatterjee, S. C. (1939), The Nyāya Theory of Knowledge: A Critical Study of Some Problems in Logic
and Metaphysics. Calcutta: Calcutta University Press.
Dravid, N. S. (trans.) (1995), Ātmatattvaviveka by Udayanācarya. Shimla: Indian Institute of Advanced
Study.
Dvivedin, M. V., and Dravida, L. S. (eds.) (1986), Ātmatattvaviveka. Calcutta: The Asiatic Society.
Feldman, J., and Phillips, S. (2011), Ratnakīrti’s Proof of Momentariness by Positive Correlation. New
York: American Institute of Buddhist Studies.
Ganeri, J. (2012), The Self: Naturalism, Consciousness, and the First Person Stance. Oxford: Oxford
 University Press.
Gangopadhyaya, M. (trans.) (1982), Gautama’s Nyāyasūtra with Vātsyāyana’s Commentary. Calcutta:
Indian Studies.
Garfield, J. (1995), The Fundamental Wisdom of the Middle Way. Oxford: Oxford University Press.
Gokhale, P. (trans.) (1997), Hetubindu of Dharmakīrti (a Point on Probans). Delhi: Sri Satguru
Publications.
Gupta, R. (1990), Essays on Dependent Origination and Momentariness. Calcutta: Sanskrit Pustak
Bhandar.
Hattori, M. (trans.) (1968), Dignāga on Perception. Cambridge: Harvard University Press.
Hayes, R. (1988), Dignāga on the Interpretation of Signs. Dordrecht: Kluwer Academic Publishers.
Hino, S. and Toshihiro Wada (2004), Three Mountains and Seven Rivers: Prof. Musashi Tachikawa’s
Felicitation Volume. Delhi: Motilal Banarsidass.
Jha, G. (trans.) (1999), The Nyāyasūtras of Gautama with the Bhāṣya of Vātsyāyana and the Vārṭika of
Uddyotakara. Vols. 1–4. Delhi: Motilal Banarsidass.
Laine, J. (1998), “Udayana’s refutation of the Buddhist thesis of momentariness in the
Ātmatattvaviveka.” Journal of Indian Philosophy, 26, 51–97.
Limaye, S. V. (trans.) (1992), Mahāyānasūtrālaṅkāra of Asaṅga. Delhi: Sri Satguru Publications.
Matilal, B. (1998), The Character of Logic in India. Ed. Jonardon Ganeri and Heeraman Tiwari. Albany:
SUNY Press.
McDermott, A. C. S. (trans.) (1969), Ratnakīrti’s Kṣaṇabhaṅgasiddhi Vyatirekātmikā: An Eleventh
Century Logic of “Exists.” Dordrecht: D. Reidel.
McDermott, A. C. S. (1972), “Mr. Ruegg on Ratnakīrti.” Journal of Indian Philosophy, 2, 16–20.
Patil, P. (2009), Against a Hindu God: Buddhist Philosophy of Religion in India. New York: Columbia
University Press.
Patil, P. (2011), “Without brackets: a minimally annotated translation of Ratnakīrti’s Demonstration of
Exclusion.” Available on the website for Siderits (2011) at
https://cup.columbia.edu/media/7285/side15360_patil.pdf.
Pradhan, P. (ed.) (1975), Abhidharmakośabhāṣya of Vasubandhu. Revised 2nd ed. Patna: K. P. Jayaswal
Research Institute.
Pruden, L. (trans.) (1988), Abhidharmakośabhāṣya, from the French Translation by Poussin. Berkeley:
Asian Humanities Press.
Rospatt, A. (1995), The Buddhist Doctrine of Momentariness. Stuttgart: Franz Steiner Verlag.
Ruegg, D. S. (1970), “On Ratnakīrti.” Journal of Indian Philosophy, 1, 300–309.
Sanghavi, S., and Jinavijaya, M. S. (eds.) (1949), Hetubinduṭīkā. Baroda: Oriental Institute.
Śàstrī, H. (ed.) (1910), Six Buddhist Nyāya Tracts. Calcutta: The Asiatic Society of Bengal.
Shastri, Swami D. (ed.) (1968), Pramāṇavārttika of Ācārya Dharmakīrti. Varanasi: Bauddha Bharati.
Shukla, K. (ed.) (1973), Śrāvakabhūmi of Ācārya Asaṅga. Patna: K. P. Jayaswal Research Institute.
Siderits, M., Tillemans, T., and Chakrabarti, A. (eds.) (2011), Apoha: Buddhist Nominalism and Human
Cognition. New York: Columbia University Press.
Tani, T. (2004), “Jñānaśrīmitra’s proof of momentary destruction of existence: a philosophical
investigation from the standpoint of intuitionistic logic,” in Hino and Wada, pp. 375–392.
Tarkatirtha, A. (1985), Nyāyadarśanam with Vātsyāyana’s Bhāṣya, Uddyotakara’s Vārttika, Vācaspati
Miśra’s Tātpāryaṭīkā, and Viśvanātha’s Vṛtti. 2nd ed. Edited by Taranatha Nyayatarkatirtha and H. K.
Tarkatirtha. Calcutta Sanskrit Series 18. 1936–44. Rpt, New Delhi: Munshiram Manoharlal.
Tatia, N. (ed.) (1976), Abhidharmasamuccayabhāṣyam. Patna: K. P. Jayaswal Research Institute.
Thakur, A. (1975), Ratnakīrtinibandhāvali. Patna: Kashi Prasad Jayaswal Research Institute.
Thakur, A. (1987), Jñānaśrīmitrabandhāvali. Patna: Kashi Prasad Jayaswal Research Institute.
Woo, J. (1999), The “Kṣaṇabhaṅgasiddhi-Anvayātmikā”: An Eleventh Century Buddhist Work on
Existence and Casual Theory. University of Pennsylvania Dissertation. UMI microform #9926216.
 Part Four

Self, No-Self, and Self-Knowledge

 12
Self or No-Self? The Ātman Debate in Classical Indian
Philosophy*
Alex Watson

Introduction
What is the (essential/ultimate) nature of sentient beings such as people? The main
traditions of classical Indian philosophy could be divided into four groups according to the
answer they give to this metaphysical question. The first group, containing just one
member, the Cārvākas, held that a person is just a body and the powers or properties of
that body. They thus denied the possibility of the continuation of life after death. All other
traditions claimed that people include a nonphysical constituent, which is their core
identity and which survives the death of the body. Do these immaterial entities remain
permanently separate from each other or do they—at the time of liberation—lose their
separate identities and merge into a greater whole? The latter answer was given by those
in the second group: Advaita Vedāntins, Nondualistic Śaivas, and certain Pāñcarātrika
Vaiṣṇavas.1 For them, individual souls/selves are identical with, or parts, emanations,
evolutes, effects, or contractions of, an Oversoul or Absolute Self, named by the respective
traditions as Brahman, Śiva, and Nārāyaṇa. The two remaining groups agree that the
nonphysical parts of people remain forever distinct from each other; they disagree over
whether they should be characterized as souls/selves or not. For the Buddhists they
should not; for those in the final fourth group—for example, Nyāya, Vaiśeṣika, Mīmāṃsā,
Sāṅkhya, Śaiva Siddhānta, Jainism—they should.
This chapter does not concern itself at all with the first two groups. It looks at some
debates between the last two—between, on the one hand, the Buddhists, and, on the
other, those traditions that posited individual selves that remain permanently numerically
distinct, there being no sense in which these selves are ultimately one. What precisely
was the issue here? What was at stake in the question of whether that part of us that
survives the death of the body should be termed a “self” or not? Section 1 provides an
answer to that question by identifying key points of dispute in the debate between Nyāya
and Buddhism. Section 2 introduces the Śaiva Siddhānta view, honoring its self-
representation as falling in the middle ground between Nyāya and Buddhism.2 Section 3
first argues that this self-representation is misleading, that Śaiva Siddhānta’s position is
just as extreme as Nyāya’s, and then diagnoses this polarization of the debate as resulting
from a shared presupposition. Section 4 identifies some better candidates for the middle
ground—Bhāṭṭa Mīmāṃsā, Jainism, Personalist Buddhism (pudgalavāda), and “Buddhism
without momentariness”—and explicates their views by placing them on a spectrum.

1 Nyāya versus Buddhism

In this first section, then, we will observe three ways in which the self was conceived of by
the Naiyāyikas, and in each case we will discern how the Buddhists denied such a
conception.

1.1 Self as unitary essence

The Naiyāyikas conceived of the self as the unitary essence of a person. It is unitary in the
sense that it is one thing over time: it endures without ceasing to exist and without its
nature changing in any way.
For the Buddhists we have no unchanging essence: we are something different in every
moment. In this moment I am an association of particular mental and bodily states, and in
the next moment I am a different association of mental and bodily states. By the time of
the second moment, the first mental and bodily states have ceased to exist; and there is
nothing that continues to exist from the first moment to the second.
 This means that as a person moves across a room, it is inaccurate to speak of
movement; rather than there being one thing that moves, there is a plurality of things
arising in very quick succession, in neighboring locations.3 It is like a film of a person
projected on a screen, which actually consists of a plurality of separate frames, but each
one follows the previous so rapidly, and resembles it so closely, that it produces the
illusion of one continuous person.
This is the Buddhist doctrine of momentariness; it encourages us to see ourselves not
as unitary and permanent, but plural and momentary.4 Figure 12.1 depicts the contrast
between the Brahmanical notion of an enduring, unchanging self (represented by a line)
and the Buddhist idea that what we are in one moment is not what we are in next
(illustrated by circles that are distinct but touch each other, as Buddhist moments are
distinct but temporally contiguous).

12.1

Furthermore, even at one point of time, for Buddhism, we are not one thing but an
association of five: a bodily state and four mental states.5 See Figure 12.2.

12.2

Thus the Brahmanical self, with its permanently unchanging essence, dissolves in
Buddhism into a diachronic and synchronic plurality. The self, since it endures
permanently, beyond death, is what explains reincarnation for the Brahmanical schools. In
other words, it is that which explains how we continue to be the same thing when we have
a different body, in a different incarnation, or no body, between incarnations. How then
can the Buddhists, in whose teachings reincarnation occupies an important place, explain
the process?
During life, each moment of consciousness (which is one of the four kinds of mental
constituents of a person) is linked to the next moment of consciousness in that it causes it
to arise. The same goes for the other three kinds of mental constituent, and the physical
constituent. The way it works at death is similar to the way it works during life: the last
moment of consciousness before death gives rise to a new consciousness in the first
moment after death. The same goes for the other three kinds of mental constituent. But
whereas during life these four mental constituents were always associated with a
momentary configuration of the body, at death the four can separate from the bodily
constituent and can reproduce themselves sequentially until such a time as they become
associated with a new body, a new embryo. See Figure 12.3, in which the vertical lines
represent the point of death and the point of the beginning of a new life. To the left of the
first vertical line, at the bottom, is the body of the present life; to the right of the second
vertical line, at the bottom, is the body of the next life.

12.3

So both sides in this debate are dualists, in that for both there is a nonphysical part of
us that exists beyond the body and senses and is not brought to an end by death. Only the
Cārvākas denied that. But the nonphysical part was conceived of very differently: by one
side as eternally unchanging and by the other as momentary (and as fourfold even in one
moment).

1.2 Self as substance
The Naiyāyikas and Vaiśeṣikas distinguished substances (dravyas) from qualities (guṇas),
the former being property-possessors (dharmins) and the latter properties (dharmas). A
thing, such as a pot or a mango, is a property-possessor, and it has five qualities—taste,
smell, color, and so on—corresponding to our five senses.
The thing was regarded as a separate ontological entity from its qualities, as indicated
by our use of language when we say, “the smell of the mango,” implying that the mango is
something that exists over and above its smell. Nevertheless a quality is inextricably
linked to a substance. It cannot exist without one. We do not find a color existing alone in
midair. There must be some substantial object to which it belongs, some substrate
(āśraya) that locates it.
The Naiyāyikas and Vaiśeṣikas use this principle to argue for the existence of the self.
Just as colors or smells presuppose substances to which they belong, so consciousness
presupposes a substance to which it belongs, that substance being the self.6
The Buddhists denied the existence of a self conceived of as the substance to which
consciousness belongs. This was part of a more general denial of the existence of
substances over and above qualities.7 Whereas to a Naiyāyika a mango is one thing with
five qualities, to a Buddhist it is five things occurring together, that is, at the same time
and in the same place.8 This is illustrated in Figure 12.4, taking the large circle to refer to
a mango, and the small circles to refer to its smell, taste, color, and so on. Or the large
circle can equally well represent a self, in which case the diagram illustrates that for
Nyāya consciousness and so on belong to a self, whereas for Buddhism consciousness and
the other four constituents (skandha) of a person exist together, as part of a
conglomeration, without belonging to anything else.9

12.4

By disputing that colors, smells, and so on belong to a substance, Buddhism calls into
question the very concept of a quality (guṇatva). Inasmuch as the concept itself implies
the concept of a substance, being one incomplete half of a substance-quality distinction,
Buddhism does away with talk of qualities.10 It refers to the things that are termed
“qualities” by Nyāya as simply parts (deśa) of a conglomeration (samudāya, samūha,
saṅghāta).11 Furthermore, whenever we use expressions that might seem to describe
parts as belonging to a whole, such as “the trees of the forest” or “the color of the
mango,” the term for the whole should not be taken to imply the existence of anything
other than the sum of the parts. It refers not to a unity, but to a conglomeration of
elements: a particular group of trees in the first case, a particular group of five sensible
properties in the second.

1.3 Self as agent
The Naiyāyikas and Vaiśeṣikas also conceived of the self as the agent of physical actions
(kartṛ), and the agent/subject of cognitions (jñātṛ). (In Figure 12.5 the continuous line on
the left to which all of the circles are attached represents the agent; the circles represent
either physical actions or cognitions.) On the one hand it is that which, through the
impulse of its will/effort (prayatna), initiates all of our physical actions. On the other it is
the perceiver of our perceptions, the thinker of our thoughts, and so on. The perception of
a pot, say, lasts just for an instant but its perceiver outlives that perception and is the
perceiver of the next and subsequent perceptions.
For Buddhists that which brings about a physical action is just that which causes it,
which for them is the intention that occurred in the stream of consciousness in the
moment preceding the action. The Vaiśeṣikas had compared the self as instigator of bodily
movements to a puppeteer instigating the bodily movements of a puppet below.12 Such a
notion of an agent standing above the sequence of mental and physical actions is precisely
what is denied by the Buddhists. The intention that brings about my present action of
touching the computer keyboard was itself caused by the previous moment of
consciousness, and so on. There is no part of a person standing outside this chain of
mental and physical events; each event is conditioned by the previous ones and brings the
next one into existence, and there is nothing over and above this causal chain that is
unconditioned. So Buddhism, by bringing the agent down from its lofty position, dividing
it up into discrete moments of intention, and dispersing them into the psychophysical
stream, replaces a two-tier model with a one-tier one.13

12.5

How did the Buddhists dispute the Naiyāyika and Vaiśeṣika notion of the self as the
agent/subject of cognition? For Buddhism the agent of a cognition (jñātṛ/grāhaka) is
simply the cognition itself (jñāna/grahaṇa). That which is conscious of a pot is
consciousness at that particular moment. So if two consecutive cognitions occur to me,
verbalizable as “I see a pot” and “I see a cloth,” the two occurrences of “I” have two
different referents: two different instances of consciousness.
No two physical actions share a common agent, because each has its own separate
prior intention; no two cognitions, or mental actions, share a common agent, because
each is its own agent. In both cases the agent of the first action exhausted itself with that
action and then ceased to exist, so it is not available to be the agent of the second action.
In the case of cognitions, just as in the case of physical actions, we have a two-tier
model replaced by a one-tier one. The subject of consciousness is dissolved into
consciousness itself. The existence of a thinker separate from thoughts, or a perceiver
separate from perceptions, is denied. Neither cognitions, nor physical movements, are
seen as actions that require an ontologically distinct actor, but rather simply as events
that occur in a particular psychophysical stream.
The absence of a continuous agent was unacceptable to the Naiyāyikas and the other
Brahmanical schools, because of its corollary that the thing that performs an action is not
the same as the thing that experiences the fruit of that action subsequently. This seemed
unjust: why should one thing experience the positive or negative consequences of an
action performed by something else?
For Buddhism, that is just the way it is. A planted seed turns into a shoot, a stalk,
leaves, a flower, and then a fruit. No one would say that it is unjust for the fruit to accrue
to the flower and not the seed. It is in the nature of things that the seed has turned into
something different by the time the fruit comes along. Similarly an action is performed
and by the time the fruit of that action occurs, the stream that performed it has become
something different.14 The result does not occur in a different stream however. That would
be unjust.

* * *

Each of the three Buddhist positions that we have just observed results from applying
more general Buddhist principles to the specific case of the self. The denial of a
permanent, unchanging self is a special case of the conception of the momentariness of
everything. The denial of the self as a substance possessing qualities is a special case of
the denial of substances over and above qualities. The denial of the self as autonomous
agent is a special case of the general position that nothing stands outside the chains of
causes and effects that make up the world.

1.4 Buddhist argument against self as substance

Before moving to Section 2, we will look in a little more detail at two of the principal
arguments in the Buddhist-Brahmanical debate over the existence of a self. How precisely
does the Buddhist reject the Naiyāyika argument, given in Section 1.2 and endnote 6, for
the self as substance? As already mentioned, the Buddhist denies in general the existence
of substances over and above qualities. His argument focuses on the evidence of our
experience. Addressing the mango example, he points out that all we can experience
there are five separate qualities.15 Through our eyes we can see a visible form (rūpa),
consisting of a shape and color, through our faculty of taste we can experience a taste,
and so on. But we do not experience some further possessor of those qualities, lurking
behind them. None of our faculties apprehends such a thing. Neither could they in
principle, given that our eyes can only sense form, our noses smells, and so on.
The case of the self is analogous. We experience states of consciousness (perceptions,
desires, thoughts, etc.), so we can grant reality to them; but we do not perceive some
substrate of those states of consciousness, underpinning them, or some substance uniting
them, in which they inhere. Such a thing is an ontological extravagance that results from
going beyond the evidence of our experience and multiplying entities.
Could we not infer the existence of a self as the thing to which consciousness belongs?
No, for the inference would require for its validity an example illustrating the existence of
substances over and above qualities, of property-possessors over and above properties;
but the fact that we do not need to assume the existence of a unitary mango to which its
qualities belong indicates that such examples are not forthcoming.
Could we not infer the existence of a unitary mango to which the smell, taste, color, and
so on belong? This could then serve as the example in the inference of a substance to
which consciousness belongs. A unitary mango could be inferred as, for example, the only
plausible explanation of the fact that the mango’s smell, taste, color, and so on occur
together. They never split off from each other. Does this not indicate that they all belong
to the same thing? No, argues the Buddhist, drawing on the principle of parsimony of
postulation (kalpanālāghava), the Indian version of “Ockham’s razor.” Rather than
postulating an imperceptible substance to which the five properties “stick,” it is more
parsimonious, argues the Buddhist, to assume that they stick to each other. For how this
“sticking to each other” was elaborated in terms of their forming a causal complex in
which they function as co-operating causes (sahakāripratyayas) for each other, see Watson
(2006, pp. 57–58).
To conclude: if it cannot even be proved that a mango-substance exists as the substrate
of taste and so on, obviously it cannot be proved that a self-substance exists as the
substrate of consciousness.16

1.5 Buddhist argument against self as agent of cognition

The principal argument for the existence of the self given by such Naiyāyika authors as
Vātsyayana and Bhaṭṭa Jayanta is as follows. We are asked to envisage a situation where
someone experiences pleasure from a particular kind of object, and on seeing the same
kind of object later and remembering the earlier pleasure, feels desire for the object (see
Figure 12.6, where P = pleasure, S = seeing, and D = desire).

12.6

Unless the earlier pleasure and the later seeing of the object had the same subject, the
desire would not arise. After all, points out the Naiyāyika, we do not find desire arising in
one person (Y) as a result of pleasure in another (X) (Figure 12.7).

12.7

So if one person were not one subject, as assumed in Figure 12.6, but a plurality of
subjects (as depicted in Figure 12.8), surely desire would not arise. Why would a
subsequent subject of experience desire something that caused pleasure not to it, but to
some totally different subject of experience?

12.8

The fact that people do desire things that have previously given them pleasure
indicates that it is the same thing that is the agent of both the desire and the earlier
pleasure.
That is the Naiyāyika argument. It gains its plausibility from likening the situations
depicted in Figures 12.7 and 12.8. It rests on the claim that since desire does not arise in
the first of these situations, it would not arise in the second. But the Buddhist has
perfectly adequate means at his disposal for distinguishing the two. Two people are not
analogous to two moments within the same stream: the latter are joined by a causal chain;
the former are not. Thus in the situation represented in Figure 12.8, the final subject is
linked by a chain of cause and effect back to the earlier pleasure, such that it has access
to memory traces (saṃskāras) of the pleasure. Person Y, by contrast, does not have access
to memory traces of person X’s pleasure, and that is why—according to the Buddhist—
desire does not arise in person Y.
The validity of the argument requires that the reason desire does not arise when there
are two people (Figure 12.7) is because of a lack of sameness of subject. But the Buddhist
has a plausible alternative: that it is due to a lack of a chain of causation along which
traces can be transmitted. So long as this alternative remains unrefuted, difference of
subject will not be sufficient to logically preclude the rise of desire. Hence the occurrence
of desire will not entail sameness of subject.
In order for the argument to work, the Naiyāyika has to prove that desire can only arise
in the same subject that experienced the earlier pleasure. He tries to do that by pointing
out that when desire does occur, it is in the same subject as that of the pleasure (Figure
12.6), not a different one (Figure 12.7). But the Buddhist just replies that a single person
is not a single subject, but a plurality of different ones (as in Figure 12.8).
Thus this is not an argument that forces any shift in the Buddhist position; it requires
for its validity that a single person is a single subject, but that is exactly what is in
question. Neither the argument from consciousness as a quality requiring a support nor
this one from desire as requiring the same subject as the pleasure that gave rise to it
oblige the Buddhist to rethink.17

2 Śaiva Siddhānta
Having observed the Naiyāyika and the Buddhist positions, we will now introduce Śaiva
Siddhānta. First we will see how that tradition differentiates itself from Nyāya. As
representative of Śaiva Siddhānta we will take Bhaṭṭa Rāmakaṇṭha (950–1000), who was
the most influential and prolific of the early Saiddhāntika exegetes, that is to say those
writers belonging to the phase of this tradition that came to an end in the twelfth century,
after which it survived only in the Tamil-speaking south, where it was transformed under
the influence of Vedānta and devotionalism (bhakti). Rāmakaṇṭha was Kashmirian, as
were most of the early exegetes of this tradition.

2.1 Śaiva Siddhānta against Nyāya

Rāmakaṇṭha does not think that Naiyāyika arguments are capable of establishing a self.
He counters them by agreeing with Buddhist arguments against them. His responses to
the Naiyāyika argument for the self as substance to which qualities belong and the
Naiyāyika argument for the self as subject of cognition are exactly the Buddhist responses
outlined in the last two sections.18
He agrees with Buddhism that there is no self as substance over and above
consciousness, and no self as agent over and above consciousness. For him, as for
Buddhism, consciousness does not require something other than itself in which to inhere.
He concurs with Buddhism that the perceiver of our perceptions, the thinker of our
thoughts, is just consciousness (grāhaka/jñātṛ = jñāna).
2.2 Śaiva Siddhānta between Nyāya and Buddhism
How then does he preserve the self? For him consciousness is the self. He equates the self
and consciousness, or to put it another way, he characterizes consciousness as the nature
(svabhāva) of the self.19 This means that he holds consciousness to be permanent, not
momentary, as it is for both Buddhism and Nyāya.20 Although consciousness for Nyāya
belongs to a permanent self, it itself consists of discrete, momentary instances, as in
Buddhism. The difference between the three views is represented in Figure 12.9.

12.9

For Nyāya, there is a self that is separate from consciousness. For Buddhism there is no
self. For Rāmakaṇṭha, there is a self but it is just consciousness. Rāmakaṇṭha crosses out
the line but joins up the dots into a line. He travels down the path of Buddhist
argumentation quite a long way: he reduces the self to the stream of consciousness. But
he then argues that the stream is unchanging.
So between Buddhism and Nyāya it was a debate about the existence or nonexistence
of an entity. Between Buddhism and Rāmakaṇṭha there is agreement about what exists; it
is just a question of how to classify that: whether as something plural or unitary, changing
or unchanging.
For Rāmakaṇṭha it is unitary and unchanging, but it is not a static entity like the self of
the Naiyāyikas. It is dynamic, yet constant. Dynamic in that it is a process, the process of
the shining forth of consciousness. Constant in that (1) the light of consciousness pours
out always in the same form, and (2) there are no breaks in the process. Consciousness as
envisaged by Rāmakaṇṭha, then, differs in two ways from consciousness as envisaged by
Buddhism: it is differentiated neither qualitatively nor temporally. Consciousness for
Buddhism, divided up as it is into dissimilar discrete entities, each one ceasing to exist
before the next one comes into existence, resembles a light forever going on and off, and
each time producing a different colored light; consciousness for Rāmakaṇṭha resembles a
light that is permanently on, forever sending out light of the same color. This constant
pouring forth of the illuminating light of consciousness is precisely what the self is, just as
the sun is nothing more than a constant pouring forth of light.
The difference of Rāmakaṇṭha’s position from that of Buddhism will now be further
elaborated.

2.3 Śaiva Siddhānta against Buddhism

Rāmakaṇṭha argues that as we proceed through life, experiencing various objects of
perception, we never lose a sense that it is me who is the experiencer of those objects.21
The Buddhist accepts that we have this sense of unbroken personal identity, but he claims
that it is mistaken. To be precise, he maintains that our direct, nonconceptual
(nirvikalpaka) experience of ourselves is of a sequence of distinct momentary perceivers,
but that we superimpose the concept of oneness on to that plurality. It is like the example
mentioned earlier of a plurality of distinct film images appearing as one continuous
image. Or, to use the Buddhists’ own example, like a very calm river that looks like one
stable, unchanging piece of water, despite being lots of different bits of water rushing
by.22
I give here two of Rāmakaṇṭha’s arguments against the coherence of this Buddhist
position.

1. It is possible that with regard to a sequence of entities external to us, we might be

fooled by the rapidity with which they succeed each other and the similarity of each
one to the previous, into mistaking the plurality for a unity. But the Buddhist is asking
us to believe that a sequence of momentary perceivers could fool themselves, as
though the distinct film images, or the bits of water in the river, could (if they were
conscious) deceive themselves into thinking that they are one unbroken thing, or could
 experience themselves as one unbroken thing. This seems much less plausible.23
2. In the standard example of superimposition, namely, that of silver on to mother-of-
pearl, it is clear that we would not mistake the mother-of-pearl for silver unless we had
experienced silver previously. But in a Dharmakīrtian universe, in which everything—
both perceivers and perceived objects—is momentary, how could a perceiver ever have
experienced something enduring, in order to acquire the concept of duration, in order
to superimpose it on to what is momentary?24
So Rāmakaṇṭha’s view is that although the objects of our experience change,
consciousness itself, the perceiver, is constant. This asymmetry was problematic for
Dharmakīrtian Buddhism, committed as it was to the non-difference of perceiver and
perceived objects (grāhyagrāhakābheda). This non-difference entails that any change in
what is perceived necessitates a change in the perceiver. If the object of our experience
changed from a pot to a cloth, and yet there was no change in consciousness, how would
consciousness have registered the change in the object? The very way in which
consciousness perceives an object is by being marked by that object.
Rāmakaṇṭha does not see why that should be so. When light—an analogy that the
Buddhists also use to elucidate the nature of consciousness—illuminates an object, we do
not regard that object as coloring the light, marking it or modifying it in any way.25 So
similarly for Rāmakaṇṭha consciousness is the illuminator of objects but is itself
unaffected by them. Experience changes because different objects come within the range
of the light of consciousness, not because consciousness itself changes.26
Rāmakaṇṭha points out that the Buddhists themselves do not hold consistently to the
position that consciousness is differentiated by its objects. For they accept that at one
moment a single, undivided perceiver can perceive a multicolored object. If in one
moment consciousness can be single and yet perceive a plurality of different objects, why
cannot one temporally extended consciousness perceive a plurality of objects? If a
relationship of one perceiver to many objects is possible at one point of time, what reason
is there to deny the possibility of such a relationship over time?27 See Figure 12.10, where
consciousness is represented above, and its perceived object(s) below.

12.10

3.1 Questioning Śaiva Siddhānta’s location in the middle ground

We have seen, then, that there are ways in which Śaiva Siddhānta falls closer to Buddhism
than Nyāya does. That is how Rāmakaṇṭha himself presents it. He sides with Buddhism
against Nyāya, presenting the Naiyāyika position as one of ontological extravagance. In
effect he says: We Śaivas, just like you Buddhists, recognize that a Naiyāyika self beyond
consciousness is a fiction. We don’t postulate such an unperceived entity. For us, as for
you, consciousness does not require something other than itself in which to inhere. We
accept no agent of consciousness separate from consciousness: for us as for you, the
perceiver of our perceptions, the thinker of our thoughts, is just consciousness.
But insofar as Śaiva Siddhānta too postulates a self that is completely unchanging and
unmodifiable (avikārya), its view still occupies an extreme position. There is a sense, in
fact, in which Śaiva Siddhānta’s view is even more extreme than Nyāya’s. Śaiva
Siddhānta’s self is arguably even more removed from change than Nyāya’s, since
individual cognitions inhere in the latter, whereas for Śaiva Siddhānta (as for Sāṅkhya)
individual cognitions belong not to the self but to the buddhi, the faculty of intellect
(responsible for conceptualization [vikalpa] and determining [adhyavasāya]).
I would now like to point to some problems with Śaiva Siddhānta’s idea of an
unchanging, unmodifiable self—an idea which I see as shared also by Nyāya and Sāṅkhya
(and Advaita Vedānta, although its distinction between the individual self, jīva, and the
absolute Self places it in a slightly different position). Below I will respond to possible
objections against it being attributable to Nyāya.

3.2 Critique of an unchanging self

As we saw above, Buddhists argue against the idea of an unchanging perceiver on the
grounds that if it is unaffected by changes in its objects, it would not be able to register
those changes, that is, would not be able to perceive them. Śaiva Siddhānta (like Sāṅkhya)
replies that the buddhi is modified. Since the buddhi, a faculty internal to the subject,
registers object-changes, the subject can perceive them. But that just relocates the
problem from the boundary between self and objects to the boundary between self and
buddhi. How can the self, if it is unmodifiable, detect changes in the buddhi?28 It is not
sufficient for Śaiva Siddhānta to adduce the example of light, an illuminator which is
unaffected by the objects it illuminates. For light does not perceive the objects it
illuminates, it enables them to be perceived by someone’s consciousness. If that
consciousness is in no way affected by the objects and the light, it remains mysterious
how it could perceive them.
Śaiva Siddhānta claims, furthermore, that the self is itself perceived (at all times). But
can any other example be given of something that is perceived and permanently
unchanging? Do not all objects of perception change over time? Are not the only things
that don’t change concepts, universals? Yet none of the defenders of the self want it to be
a mere concept. To claim that the self is available to introspection is hard to reconcile
with the claim that it exists beyond all change.
Those two objections are applicable also to Sāṅkhya and Advaita Vedānta. But Śaiva
Siddhānta, unlike those two, maintains that the self is an agent, and that lays it open to a
third objection. How can the agent, that which brings about action, exist always in the
same form? If it is unmodifiable, how could it produce a certain response at one time and
a different response at a subsequent time?
Rāmakaṇṭha can respond that though the self does not change, the conditioning
(bhāvanā) that operates on the mind does.29 This change in mental conditioning means
that the self’s habits can change. At one time it can be pulled in one direction, at another
time in another. But this response is not compelling. If the self’s choices are made not by
it, but by forces external to it, then it is just a passive plaything that is manipulated and
that offers no input into decision-making. If it is the self that chooses what to do, it is very
difficult to see how it can be unmodified. The fact that it chooses one thing at one time
and one thing at another implies a difference in it.
An analogy that Rāmakaṇṭha is fond of using to explain how something unchanging can
be an agent is that of a magnet.30 The magnet, despite not moving, can cause movement
in iron filings; similarly a self, despite never being modified, can cause the body with
which it is currently associated to move in various ways. But the analogy would only
render plausible that an unmodifiable self could act in different ways at different times if
the magnet were capable of making (identical) iron filings move in one way in one
situation and in another way in another (identical) situation.
Those three objections are all of the kind that we meet in the classical Indian
discussions themselves. I now mention some from a more contemporary perspective. One
of the problems with the concept of a subject of experience that is completely
unmodifiable is that it is of use in so few discourses. Psychology, history, and (auto-
)biographical literature all require a concept of the individual as something that is capable
of changing, growing, developing, regressing. And an unchanging self will be of little use
to those who believe that the postulates of philosophy should be naturalizable, that is,
should be capable of taking their place in the empirical sciences. What use would they
have for something that undergoes no change itself and hence can contribute to no
change in anything else? Such an entity looks a priori undiscoverable.
These considerations carry only so much weight. Why, after all, should the postulates of
philosophy need to be sanctioned by disciplines outside philosophy? But an unmodifiable
self, rather like a person-like God, is susceptible to a two-pronged attack: not only can its
plausibility be challenged, but also deep-rooted psychological motives for belief in it can
be identified, the combination of these two making it appear as a piece of wishful
thinking. What are these psychological motives? It answers a need for some post to cling
to in a world full of unwanted change, a reliable counterpoint to the unpredictable flow of
life and its unwelcome twists. It offers hope to that part of us that would rather not be
muddied by life’s torments, that prefers to be uninvolved and clean rather than enmeshed
and bruised. By affirming an eternally calm, still part of us, it proclaims victory over the
painful and turbulent experiences we dislike. As such, it can be seen as a reaction to and
compensation for feeling insecure and insignificant.
3.3 Diagnosis
Inasmuch as Śaiva Siddhānta postulates a self that is completely unmodifiable, it too can
be placed with Nyāya. At one end of the spectrum we have the Kṣaṇikavādins, the
Buddhist proponents of momentariness, according to whom we are different in every
single moment—not only qualitatively but also numerically. At the other end of the
spectrum we can place Nyāya, Śaiva Siddhānta, and Sāṅkhya for whom what we are,
essentially, is a self that is not only unitary over time, but also eternally unchanging. We
thus have a polarized debate with each of two extremes attacking the other extreme, and
no one adopting or attacking the middle ground. The view that we could be a self that
changes—that has numerical identity but qualitative change—is mostly ignored.
Some will here object that that is precisely the Nyāya position. They will claim that a
Naiyāyika self changes over time in the sense that it has changing cognitions and so on
inhering in it. But Naiyāyikas assert that the self and its cognitions are quite separate
from each other. Any substance is a separate entity from its qualities, for Nyāya, in
accordance with its doctrine of guṇaguṇibheda, the difference of qualities from that which
possesses the qualities. Moreover, the ontological distance between a self and its qualities
—cognitions, pleasure, pain, desire, and so on—is even greater than that between a
physical substance and its qualities. For the self is eternal, its particular qualities are
momentary; it is omnipresent, and they are restricted to a particular place. This firm
separation between the self and its qualities means that Nyāya ends up with the view that
despite cognitions and so on inhering in the self, changes in the former do not affect the
nature of the latter.31 That it is correct to attribute to Nyāya the view that changes in the
self’s qualities do not affect the self’s nature is confirmed by passages dealing with
liberation (apavarga, mokṣa). In such passages Naiyāyika authors assert that the self’s
nature is, and always has been, free of all of its particular qualities (sakalaguṇāpoḍha).32
These qualities are thus irrelevant to its nature.33 They are part of the “not-self” with
which it confuses itself while in saṃsāra,34 and they are “to be abandoned” (hātavya,
heya).35 At liberation it becomes free of them and in so doing rests in its true nature alone.
I suggest that the explanation for this polarization of the debate—for the reluctance of
these traditions to occupy the middle ground—lies in a shared assumption: the
assumption that if an entity changes, it can no longer be the same thing, that is, that at
the level of the fundamental constituents of the world there can be no qualitative change
without numerical change.
This is an explicitly stated presupposition of the Kṣaṇikavādins. It is what carries much
of the weight in Dharmakīrti’s inference of momentariness—leading to the postulation of
consciousnesses that are numerically differentiated down to the level of every single
moment. It can also be detected in Śaiva Siddhānta36 and Nyāya and is what explains their
refusal to allow any change on the part of the self. For if it is accepted, then any change in
the nature of a self will entail that that self ceases to exist and is succeeded by another.
The concept “self” would then no longer be applicable; the Buddhist position would have
been lapsed into.

4 Between an unchanging self and momentariness

4.1 Bhāṭṭa Mīmāṃsā and Jainism

Does no one assert the existence of a self that is changing? Is the genuinely middle
ground totally unoccupied? No, two traditions can be placed there: Bhāṭṭa Mīmāṃsā and
Jainism.
The originator of the former, Kumārila, explicitly rejects the prevalent presupposition
that qualitative change entails numerical change. He puts the following objection into the
mouth of an opponent: surely if the self is transformed, it cannot be eternal.37 He then
responds that if non-eternal (anitya) means just being liable to transformation (vikriyā), he
has no problem calling the self non-eternal. But the self is certainly not subject to
destruction (uccheda, nāśa); for to be modified (and to be non-eternal in that sense) is not
to cease to exist but rather just to “assume another state” (avasthāntaraprāpti).38 Some
aspects of the self are permanent and some (its states or qualities) are impermanent.
Examples of the former are its consciousness (caitanya), its existence (sattā), and the
particular substance (dravya) that comprises it; examples of the latter are its pleasures
and pains. It is compared to a snake coiling into different positions, or a piece of gold that
is remolded from a dish to a necklace to an earring.39 The snake itself and the gold atoms
stand for its unchanging aspects; the different positions of the snake and the different
shapes of the gold stand for its changing aspects.
Kumārila and those in his tradition refer to the self’s pleasure, pain, and so on not just
as its “states” (avasthā), but also its “qualities” (guṇa) or “properties” (dharma). In that
case how is this view different from that of the Naiyāyikas? In both cases we have a
permanent substance (dravya) with changing qualities/properties. It is different because
Kumārila has a different take on the relation between substances and their qualities; he
specifies the relation not as difference/separateness (bheda), but rather as both difference
and non-difference/separateness and non-separateness (bhedābheda). This closer
connection, or blurred boundary, between a substance and its qualities means that—
unlike for the Naiyāyikas—modification of the latter does entail modification of the former.
Kumārila has no problem accepting that the self is modified.
The same goes for Jainism. It distinguishes between the essence (bhāva, jāti) of the self
and its modes (paryāya).40 But the two sides of this distinction are (unlike for the
Naiyāyikas and as for Kumārila) not completely different/separate from each other; they
are rather different aspects of the same thing. So one and the same self-substance is
permanent and unchanging when viewed from one point of view, and impermanent and
changing when viewed from another. Its permanence must be indexed to one aspect of it,
namely, its essence; if it were completely permanent (sarvathā nityatve),41 it could not be
transformed, so the good conduct which causes someone to cease transmigrating would
not be able to have any effect on it.42
There is much similarity between the Bhāṭṭa and the Jaina views, as brought out by Uno
(1999). But one dissimilarity is that for the Jainas the self, though immaterial, changes its
size; it occupies the same dimensions as the body with which it is currently associated
(svadehaparimāṇa). It is thus subject to contraction (saṃharaṇa) and expansion
(visarpaṇa).43
For both the Bhāṭṭas and the Jainas the self is one numerically identical thing that
changes. Although these two traditions allow more change in the self than any of the
other self-theorists, they only allow so much. In order to see what I mean by this, consider
the example of the boat that over time has had all of its parts replaced. This would not
serve as a valid analogy for a Jaina or Bhāṭṭa self, because the boat’s numerical identity—
if it is even considered numerically identical—consists not in it being the very same
substance, composed of the same stuff, but in other factors such as continuity of structure
and an uninterrupted spatiotemporal path. For Kumārila the stuff out of which the self is
composed is eternal (that is the point of the gold analogy with its eternal gold atoms),
whereas in the case of the boat there is nothing that continues to exist throughout the
entire span of its life. Both the boat and a Bhāṭṭa/Jaina self are “one thing that changes,”
but such a definition is not sufficient to capture the Bhāṭṭa or Jaina conception of self, for
selfhood was taken by both to require a strong sense of numerical identity. Yes it can
change, but to count as a self it must also be numerically identical in the strong sense of
being the very same substance, composed of the same stuff, with no change whatsoever in
its essence. With that we reach the limit of the self-theorists; any attempt to preserve
numerical identity but in a weak sense—analogous to that of the boat—will count as a
Buddhist view rather than a self-view.44 This worry that too much change cannot be
admitted of the self can be observed in the Jaina view that changes in the self’s size—in
the amount of space it takes up to fit its current body—do not involve any change in its
“weight” (it has agurulaghutva = never gets heavier or lighter) or “innate extent”:
“whether a given body is as large as the entire loka-ākāśa or as small as the tiniest object
imaginable, the number of the soul’s space-points remains the same.”45 This is explained
by the analogy of a cloth, whose mass never alters however many shapes it is folded
into.46 Here again, then, we see that the self’s changes must coexist with no change in its
essence, compositional stuff, or nature as substance.
We have reached the location on the spectrum we are delineating at which we pass
from the self-views to the Buddhist views.

4.2 Two more Buddhist views

We have so far been using the expression “the Buddhist view” to refer to that of
momentariness-theorists (Kṣaṇikavādins) such as Vasubandhu, Dharmakīrti, and their
followers. That is because non-Buddhists, when they set about proving a self, took the
momentariness-theorists to be representative of Buddhism. But two other Buddhist views
will be mentioned here. The first view we meet after crossing the boundary into the
Buddhist side of the spectrum is that of the Personalists (Pudgalavādins). They felt that
the unqualified denial of a self on the part of their fellow Buddhists was not true to the
Buddha’s teaching, especially to those passages in which he is depicted as rejecting both
the view that there is a self and the view that there is not. They thus postulated a
“person” (pudgala) that cannot be said to be either the same as or different from the
psychophysical constituents (skandha).
If it were the same as the constituents, they reasoned, then it would be as momentary
as them, and memory, rebirth, and moral responsibility would be difficult to account for. If
it were independent from them, then it would be as eternal and unconditioned as a
Sāṅkhya or Naiyāyika self (atman), and hence all the problems that Buddhists see with
such an entity ensue: it cannot enter into a mutual relationship with psychophysical
reality, it would seem to be already liberated and so makes the religious life redundant,
and so on. The Personalists regarded their view as taking the proper middle way between
the two extremes of eternalism and annihilationism.47
They compared the relationship between the constituents and the person to that
between a tree and its shadow, or fuel and the fire rising from that. As Eltschinger and
Ratié (2013, pp. 73–75) perceptively note, four aspects of the analogies seem to have been
intended. (1) The shadow is neither the same as nor different from the tree, and the fire is
neither the same as nor different from the fuel. (2) The shadow and the fire exist, but in a
less substantial and determinate way than the tree and the fuel. (3) There is no shadow
without the tree and no fire without the fuel. (4) The shadow and the fire are caused by,
respectively, the tree and the fuel. The person, then, is a kind of epiphenomenon thrown
up by the constituents; it cannot exist without them, but it is not reducible to them.
It is not difficult to distinguish this “person” from the self of the Naiyāyikas, Śaiva
Siddhāntins, and so on; but there is some overlap between it and the self of the Bhāṭṭas
and Jainas. As the former is neither the same as nor different from the constituents, so the
latter is neither the same as nor different from (or rather both the same as and different
from) the self’s qualities such as its pleasures, pains, and cognitions. In what way, then,
are the two concepts distinct? (1) A Bhāṭṭa or a Jaina self is not caused by its qualities. (2)
It can exist without them—in the state of liberation and between incarnations. (3) It is not
less substantial or determinate than its qualities. (4) It is a substance; the person is not.
(5) It is eternal; the person is neither eternal nor momentary.
The “person” of the Pudgalavādins thus falls between an enduring self-substance (as
upheld by the Bhāṭṭas and Jainas) and the transient constituents (skandha). The next view
on the spectrum—let us call it “Buddhism without momentariness”—is one that was not,
to my knowledge, actually put forward by any Buddhists. It is a product of my reflection
and a sense that there is conceptual space here for a “Buddhistic” view that fits between
the two properly Buddhist views dealt with in this chapter. It and the momentariness view
of the Kṣaṇikavādins fall together, against the Personalists, in asserting that an individual
consists in nothing more than the constituents. What separates the two of them is that for
the former the constituents are temporally extended, for the latter they are momentary.
The “Buddhist-without-momentariness” view is represented in Figure 12.11.

12.11

The five rows represent the five kinds of constituent. If we take the top row as
consciousness (vijñāna-skandha), each rectangle in that row denotes an instance of
consciousness. All of these instances will be transient; some may last only for a moment,
most for longer, but none forever. The same goes for the other four constituents. One kind
of constituent, a feeling say, may or may not begin and end at the same time as another,
an impulse or an instance of consciousness say. It is very unlikely that instances of all five
constituents will stop and start at the same time. Thus this view avoids what some held to
be a problematic feature of momentariness—that there are breaks in the process, that an
individual is completely destroyed (in every moment) before arising again in the next, and
that this seems equivalent to annihilationism. On this view there are no breaks in the
process, no destruction of an individual, because at the point where one kind of
constituent ceases, others will be existent. The overlapping of the constituents avoids
annihilationism.

Conclusion
The chapter began by distinguishing four groups of classical Indian philosophical
traditions. Having looked at several of the traditions that belong in the third and fourth
groups—at the positions they take on the issue of selfhood and personal identity—we see
that they can be arranged along the following spectrum.

Notes
* I am extremely grateful to John Taber, Isabelle Ratié, Chakravarthi Ram-Prasad, Luke O’Sullivan
and Jim Duerlinger for comments on an earlier draft of this article. I have also benefitted greatly
from audience responses to versions presented in 2013 at St. Stephens College (Delhi), Harvard,
Berkeley, Columbia and the University of Chicago; I would like to mention in particular Mark
Siderits and Michael Allen.
1. See Watson, Goodall, and Sarma (2013, pp. 27–35).
2. I thank the Journal of Indian Philosophy for allowing me to reproduce material in sections 1 and 2
that I have already published there in Watson (2014b).
3. See Abhidharmakośabhāṣya pp. 473,20–23; Duerlinger (2003, p. 99).
4. As to why Buddhists asserted the momentariness of both mental and physical entities—why they
explained change not as one thing becoming modified, but rather as a succession of distinct
momentary things—see Dharmakīrti’s Pramāṇaviniścaya 2:53–55, Vādanyāya p. 2,1–3,13 and
Hetubindu p. 4*,6–7, p. 19,10–13; and Dharmottara’s Pramāṇaviniścayaṭīkā ad 2:53–55 and
Kṣaṇabhaṅgasiddhi. See also Steinkellner (1963; 1968/69), Mimaki (1976), von Rospatt (1995),
Yoshimizu (1999), and Sakai (2010a, b; 2011).
5. These four kinds of mental state are: feelings (vedanā), ideation (sañjñā), impulses (saṃskāra),
and consciousness (vijñāna); see Vetter (2000).
6. The argument involves three contentions, each of which had their own supporting arguments: (1)
Qualities cannot exist without substances to which they belong; (2) consciousness, desire,
aversion, pleasure, pain, volition are qualities; (3) the self is the only possible substance to which
these qualities could belong. See Nyāyavārttika ad 1.1.10, p. 62,12–18, and Praśastapādabhāṣya p.
16,3–7. For the second stage of the argument in particular, see Nyāyavārttika ad 3.2.18,
Nyāyamañjarī vol. 2, p. 278,14–15, and Candrānanda ad Vaiśeṣikasūtra 2.2.28. For the third stage
of the argument, see, for example, Nyāyamañjarī vol. 2, pp. 284,6–293,2 and Nyāyasūtra 3.2.47
with the commentaries ad loc. See also Chakrabarti (1982), Oetke (1988, pp. 255–256, 258–260,
280, 286–300, 359–360, 464), Matilal (1989, pp. 74, 77; 1994, p. 286), Preisendanz (1994, pp. 187,
209, 278–281), Kano (2001), and Watson (2006, pp. 174–184).
7. See Abhidharmakośabhāṣya pp. 475,14–16 and 475,22–476,3; Duerlinger (2003, pp. 103 and 104).
8. The concept “mango,” for Buddhism, corresponds to no reality, but is a false unity that we
superimpose on to something plural, like the concept “forest.” To use Vasubandhu’s terminology in
the Abhidharmakośabhāṣya, forests and mangoes are prajñaptisat, not dravyasat (see, e.g., p.
461,14ff.): they have merely conceptual, not substantial, existence.
9. The four skandhas that Buddhism groups with vijñāna (consciousness)—rūpa, vedanā, sañjñā,
saṃskāra—are of course not the same as the qualities of the self that Nyāya groups with jñāna
(consciousness): icchā, dveṣa, prayatna, sukha, duḥkha (Nyāyasūtra 1.1.10).
10. See Abhidharmakośabhāṣya p. 476,1–2, Mataṅgavṛtti, vidyāpāda p. 153,8–11 and
Nareśvaraparīkṣāprakāśa introducing 1:5, p. 11,1–4.
11. See Nareśvaraparīkṣāprakāśa introducing 1:5, p. 11,4–6.
12. See Praśastapādabhāṣya p. 15,12 and Candrānanda ad Vaiśeṣikasūtra 3.2.4, p. 28,18–19.
13. See Abhidharmakośabhāṣya pp. 476,19–477,3; Duerlinger (2003, p. 107).
14. See Abhidharmakośabhāṣya p. 477,11–17; Duerlinger (2003, p. 108).
15. See Nareśvaraparīkṣāprakāśa, avatārikā to 1.5 (p. 11,1–6), Kiraṇavṛtti ad 2:25ab (p. 53,4–8),
Mataṅgavṛtti, vidyāpāda p. 153,8–11, and Watson (2006, pp. 184–192; 2010a, pp. 87–89).
16. This strategy of Buddhist argument goes back to the Abhidharmakośabhāṣya. The opponent there
states that the self is required as the support (āśraya) of consciousness (citta) and traces
(saṃskāra), in the way that earth supports its qualities such as smell. Vasubandhu replies that the
example of earth is exactly what convinces him that there is no self. The fact that we perceive only
a certain combination of smell and other properties, not some extra entity “earth” supporting
them, indicates that there is no such entity, and this indicates that, analogously, there is no self
 supporting consciousness and traces (Abhidharmakośabhāṣya pp. 475,14–16; Duerlinger [2003, p.
103]). On Śāntarakṣita and Kamalaśīla’s refutation of the view that desire and other states of
consciousness require a support, see Hulin (1978, pp. 100–101).
17. Versions of this argument are found at Nyāyabhāṣya ad 1.1.10, p. 16,5–20, Nyāyavārttika ad
1.1.10, pp. 60,12–63,2, Nyāyamañjarī pp. 278,4–284,5; see also Oetke (1988, pp. 345–352 and 256–
258), Matilal (1989, pp. 74–77; 1994, pp. 286 and 289–291), Taber (1990, pp. 36–37), Preisendanz
(1994, pp. 202, 306), and Kapstein (2001, pp. 146–151 and 375–383). Apart from the earliest, that
of the Vṛttikāra in the Śābarabhāṣya, they all appeal to the concept of synthesis (pratisandhāna,
anusandhāna). Desire would not arise, it is argued, were it not for the subject’s ability to
synthesize the earlier pleasure with the present seeing of the object. One can only synthesize
cognitions of which one is the subject. Therefore the earlier pleasure and the present seeing must
have the same subject. But, replies the Buddhist, why is it the case that only cognitions having the
same subject can be synthesized? What is required for synthesis, as the Naiyāyika also recognizes,
is the activation of a memory trace of the earlier cognition. Why is this not enough? Why do the
Naiyāyikas also insist on a further requirement, namely, sameness of subject?
At this point certain Naiyāyikas give a verbalization of the synthesis, such as “Earlier I derived
pleasure from this object, and now this same I am experiencing it again”; and argue that such a
cognition would not arise unless I were indeed the subject of both the earlier pleasure and the
present seeing. It is true, replies the Buddhist, that such a cognition implies that I sense myself as
the subject of both the present and the past experiences of the object. But this sense of sameness
is easily explainable as resulting from the rapidity with which consecutive momentary subjects
succeed each other, and the similarity of each one to the previous; this fools us into superimposing
oneness on to what is actually multiple. Some Naiyāyikas put forward synthesis as verbalized
above not as a means of inferring the self, but as including a direct perception (pratyakṣa) of the
self. But this is countered on the grounds that it assumes what the argument sets out to prove: the
validity of such seeming experiences of an enduring subject.
Even some Naiyāyikas regard the argument as a failure (those that Jayanta refers to as
svayūthya, “those of our own fold”), pointing out that it can only work if it asserts that synthesis
includes direct perception of an enduring subject. But if such a subject is available to pratyakṣa,
then this whole inference from desire becomes pointless (Nyāyamañjarī p. 277,14–17).
Furthermore even proponents of the argument such as Uddyotakara and Jayanta allow their
Buddhist interlocutors (pūrvapakṣins) to overcome the various Naiyāyika strategies they put
forward. They see its success as dependent on an independent refutation of the coherence of
momentariness. Thus Uddyotakara allows his opponent to answer each of his points until at the
end he argues that a momentary entity would not be able to leave a trace on another momentary
entity, whether the latter existed contemporaneously with it or immediately after it (Nyāyavārttika
p. 62,19–63,2; see also Taber [2012]). Similarly, throughout Jayanta’s long discussion the opponent
is able to answer all of Jayanta’s assertions, and the debate is only closed when Jayanta asserts
that he will explain later in the chapter that there can be no relation of cause and effect between
momentary cognitions (Nyāyamañjarī vol. 2, p. 284,3–4).
For a more detailed analysis of the argument and its history, see Watson (2006, pp. 138–157 and
159–165). Note that it is different from, though sometimes mistakenly conflated with, the
arguments we find in the commentaries to Nyāyasūtra 3.1.1, which are also often put in terms of
synthesis (pratisandhāna). The argument that we have been examining is based on cognitions at
different points of time being synthesized by a single entity, who must therefore exist continually
over that time span. The arguments in the commentaries to 3.1.1 are based on perceptions from
different sense-faculties being synthesized by a single entity, who must therefore exist over and
above the individual sense-faculties. What is aimed to be proved is not necessarily an entity that
endures over time, but one that exists above and beyond the plurality of sense-faculties. On these
arguments based on 3.1.1, see Halbfass (1976, p 163), Matilal (1986, pp. 252–254, 372), Oetke
(1988, pp. 260–269), Taber (1990, pp. 39–42), Laine (1993), Preisendanz (1994, pp. 183–187),
Chakrabarti (1992) and Ganeri (2000; 2007, pp. 180–181).
18. For his response to the first Naiyāyika argument, see Nareśvaraparīkṣāprakāśa, avatārikā to 1.5,
p. 11,1–6, Kiraṇavṛtti ad 2:25ab (p. 53,4–8), Mataṅgavṛtti, vidyāpāda pp. 153,8–11, and Watson
(2006, pp. 184–192; 2010a, pp. 87–89). For his response to the second, see
Nareśvaraparīkṣāprakāśa introducing 1:5, pp. 9,10–10,8, and Watson (2006, pp. 138–159 and 240,
n. 99).
19. The view that consciousness is the nature of the self may appear to some as not so different from
the view that it is a quality/property of the self; talk of a thing’s “properties” in English can seem
more or less synonymous with talk of its “nature.” But in Indian philosophical discourse whereas
the relation between a thing and its nature was held to be identity, sameness (tādātmya), the
relation between a thing and its qualities (guṇas) or properties (dharmas) was held to be inherence
(samavāya). A thing and its nature are the same thing; a thing and its qualities/properties are not—
the latter belong to the thing, but are of a different nature.
20. His view may remind some readers of either Sāṅkhya or Advaita Vedānta; for an analysis of the
differences of his view from both of these, see Watson (2010a).
21. Nareśvaraparīkṣāprakāśa ad 1:5, pp. 13,20–14,18; Watson (2006, pp. 220–230).
22. Nareśvaraparīkṣāprakāśa ad 1:5, pp. 14,18–15,5; Watson (2006, pp. 230–236).
23. That a sequence of momentary perceivers could deceive themselves is also contradicted—
according to Rāmakaṇṭha—by the Buddhist assertion that all cognition is nonconceptual with
regard to itself: see Watson (2010b, pp. 302–303; 2006, pp. 237–238, 245–251). For a discussion of
what precisely Dharmakīrti means by cognition being nonconceptual with regard to itself, see
Watson (2010b, pp. 317–319).
24. This point is an extrapolation of Rāmakaṇṭha’s thinking, rather than a close report of what he has
written. His account of why superimposition would be impossible if everything were momentary
(see sarveṣāṃkṣaṇikatvena yojanānupapatteḥ at Nareśvaraparīkṣāprakāśa ad 1:5, p. 15,17–18, in a
passage translated and analyzed at Watson [2006, pp. 238–245]) focuses more on superimposition
requiring an enduring perceiver than on it requiring experience of duration.
25. For discussion of the light analogy, see Watson (2010b, p. 305; 2014a) and Watson and Kataoka
(2010, pp. 304–306).
26. See, for example, Nareśvaraparīkṣāprakāśa ad 1.6ab, p. 26,4–13, Watson (2006, pp. 333–382;
2010a, esp. pp. 111–112).
27. For the full argument, see Nareśvaraparīkṣāprakāśa ad 1.6ab, pp. 26,19–28,11, and Watson
(2006, pp. 335–348).
28. We find Śāntarakṣita (c. 725–788) arguing in exactly this way (against a Sāṅkhya opponent) in the
Tattvasaṅgraha (294ff.); see Watson (2010a, pp. 90–95). See also Siderits (2011, p. 421).
29. Rāmakaṇṭha appeals to these changes in mental predisposition (bhāvanā) in the Mataṅgavṛtti (ad
6:34c–35a, pp. 173,11–174,1) during a defense of the unchanging nature of consciousness.
30. See index entry for “magnets” in Watson, Goodall, and Sarma (2013).
31. For an account of the evolution of the increasing distance that developed between the self and its
qualities in Nyāya and Vaiśeṣika, see Frauwallner (1956, pp. 91–104; 1984, pp. 61–71).
32. Nyāyamañjarī vol. 2, pp. 359,6: sakalaguṇāpoḍham evāsya rūpam.
33. They are described as extrinsic to it, not innate (na naisargika): Nyāyamañjarī vol. 2, p. 359,5.
34. Nyāyabhāṣya p. 6,9–10.
35. Nyāyabhāṣya p. 6,11; Nyāyamañjarī vol. 2, pp. 265,10–12 and 430,3–4.
36. Śaiva Saiddhāntika authors reveal their acceptance of this presupposition in the way that they
respond to the Dharmakīrtian inference of momentariness. For the Dharmakīrtian Buddhists the
seed that produces the sprout cannot be the same thing as the seed when it was in the granary and
not producing a sprout, because the former has as its nature the ability to produce a sprout and
the latter does not. Saiddhāntika authors agree that these two seed phases would be numerically
distinct entities if they had different natures (svabhāva), and so they are forced into making the
counterintuitive claim that there is no difference at all in the nature of the two-seed phases, the
only difference between the two situations being the presence or absence of auxiliary causes such
as earth and moisture that allow the sprout to be produced. See Watson, Goodall, and Sarma
(2013, pp. 378–390) and Paramokṣanirāsakārikāvṛtti pp. 177–181.
37. Ślokavārttika, ātmavāda 21.
38. Ślokavārttika, ātmavāda 22–23.
39. See Ślokavārttika, ātmavāda 26–28, Pārthasārathimiśra and Jayamiśra (Ślokavārttikaṭīkā—
Śarkarikā) ad loc., and Umbeka (Ślokavārttikavyākhyā—Tātparyaṭīkā) ad pratyakṣasūtra 53.
40. Sarvārthasiddhi ad 5.29, Uno (1999, p. 424).
41. Sarvārthasiddhi ad 5.31.
42. Uno (1999, p. 425). See also Jaini (1979, p. 103): “the Jaina suggestion—indeed requirement—of
some form of change in the soul-substance constitutes a unique and significant departure from the
mainstream of Indian thought.” It does indeed constitute a departure from such mainstream
traditions as Nyāya, Vaiśeṣika, and Sāṅkhya, but not from Bhāṭṭa Mīmāṃsā.
43. Jaini (1979, p. 58), citing the Rājapraśnīyasūtra as the earliest source for this idea.
44. My thoughts here were partly derived from and partly stimulated by John Taber’s comments on an
earlier version of this chapter.
45. Jaini (1979, p. 103), Sarvārthasiddhi § 557.
46. Jaini (1979, pp. 58–59), Rājapraśnīyasūtra § 67.
47. Eltschinger and Ratie (2013, p. 84).

References
Primary sources
Abhidharmakośabhāṣya: Abhidharmakośabhāṣya of Vasubandhu, edited by P. Pradhan, Tibetan Sanskrit
Works Series 8. K. P. Jayaswal Research Institute, Patna, 1975.
Hetubindu: Dharmakīrti’s Hetubinduḥ, Teil 1, Tibetischer Text und rekonstruierter Sanskrit Text,
edited by Ernst Steinkellner, Wien, 1967.
Kiraṇavṛtti: Kiraṇavṛttiḥ. Bhaṭṭa Rāmakaṇṭha’s Commentary on the Kiraṇatantra. Volume 1: Chapters
1–6, edited by Dominic Goodall, Publications du département d’indologie, Institut français de
Pondichéry/École française d’Extrême-Orient 86 (1). Pondichéry, 1998. (In references to this edition,
line numbers are not counted from the top of the page; they rather follow the line numbers printed in
 the edition, which start counting from the previous verse-segment.)
Kṣaṇabhaṅgasiddhi: see Frauwallner (1935).
Mataṅgapārameśvara: Mataṅgapārameśvarāgama, Vidyāpāda, avec le commentaire de Bhaṭṭa
Rāmakaṇṭha, edited by N. R. Bhatt, Publications de l’Institut français d’Indologie 56. Pondichéry,
1977.
Mataṅgavṛtti: in Mataṅgapārameśvara.
Nareśvaraparīkṣāprakāśa: Nareśvaraparīkṣā of Sadyojyotis with the Prakāśa Commentary of
Rāmakaṇṭha, edited by Madhusūdan Kaul Śāstrī.,KSTS 45, Delhi, 1989. (Original edition Śrīnagar,
1926).
Nyāyabhāṣya: Gautamīyanyāyadarśana with Bhāṣya of Vātsyāyana, edited by Anantalal Thakur,
Nyāyacaturgranthikā 1, New Delhi, 1997.
Nyāyamañjarī: Nyāyamañjarī of Jayantabhaṭṭa, with Ṭippaṇi—Nyāyasaurabha by the Editor, edited by
K. S. Varadacharya, 2 vols, Oriental Research Institute Series 116, 139. Mysore, 1969 and 1983.
Nyāyasūtra: in Nyāyabhāṣya.
Nyāyavārttika: Nyāyabhāṣyavārttika of Bhāradvāja Uddyotakara, edited by Anantalal Thakur,
Nyāyacaturgranthikā 2, New Delhi, 1997.
Paramokṣanirāsakārikāvṛtti: in Watson, Goodall, and Sarma (2013).
Pramāṇaviniścaya: Pramāṇaviniścaya, Chapters 1 and 2, edited by Ernst Steinkellner. China Tibetology
Publishing House/Austrian Academy of Sciences Press, Beijing/Vienna, 2007.
Pramāṇaviniścayaṭīkā: see Sakai (2010a).
Praśastapādabhāṣya: Word Index to the Praśastapādabhāṣya: A Complete Word Index to the Printed
Editions of the Praśastapādabhāṣya, edited by Johannes Bronkhorst and Yves Ramseier, Delhi, 1994.
Rājapraśnīyasūtra: in Suttāgame (Vol. II, 41–103).
Sarvārthasiddhi: Sarvārthasiddhi of Pūjyapāda (with Tattvārtha-sūtra of Umāsvāti), Sanskrit text with
Hindi translation by Phoolchandra Siddhāntaśāstrī, Bhāratīya Jñānapīṭha Publications, Varanasi,
1971.
Ślokavārttika: Ślokavārttika of Śrī Kumārila Bhaṭṭa with the Commentary Nyāyaratnākara of Śrī
Pārthasārathi Miśra, edited by Dvārikādāsa Śāstrī, Prāchyabhārati Series 10, Varanasi, 1978.
Ślokavārttikaṭīkā (Śarkarikā): Ślokavārttikaṭīkā (Śarkarikā) of Bhaṭṭaputra Jayamiśra, edited by
Kunhan Raja, Madras University Sanskrit Series, 17, University of Madras, 1946.
Ślokavārttikavyākhyā (Tātparyaṭīkā): Ślokavārttikavyākhyā Tātparyaṭīkā of Uṃveka Bhaṭṭa, edited by
S. K. Ramanatha Shastri, Rev. K. Kunjunni Raja, and R. Thangaswamy, Madras University Sanskrit
Series, 13, University of Madras 1971. (1st ed. 1940.)
Suttāgame: Prakrit texts, edited by Puppha Bhikkhu, 2 vols, Sūtrāgama-prakāśaka-samiti, Gurgaon-
Delhi, 1953.
Tattvasaṃgraha: Tattvasangraha of Ācārya Shāntarakṣita with the Commentary “Pañjikā” of Shri
Kamalashīla, edited by Dwarikadas Shastri, 2 vols, Bauddha Bharati series 1, 2, Vārāṇasī, 1981–82.
Vādanyāya: Dharmakīrtis VādanyāyaḥTeil I, edited by M. T. Much, Philosophisch-Historische Klasse,
Sitzungsberichte 581. Österreichische Akademie der Wissenschaften, Wien, 1991.
Vaiśeṣikasūtra: Vaiśeṣikasūtra of Kaṇāda with the Commentary of Candrānanda, edited by Muni Sri
Jambuvijayaji, Gaekwad’s Oriental Series 136, Baroda, 1961.

Secondary sources
Chakrabarti, A. (1982), “The Nyāya proofs for the existence of the soul.” Journal of Indian Philosophy,
10, 211–238.
Chakrabarti, A. (1992), “I touch what I saw.” Philosophy and Phenomenological Research, 52(1), 103–
116.
Duerlinger, J. (2003), Indian Buddhist Theories of Persons: Vasubandhu’s “Refutation of the Theory of a
Self.” London and New York: RoutledgeCurzon.
Eltschinger, V., and Ratie, I. (2013), Self, No-Self, and Salvation: Dharmakīrti’s Critique of the Notions
of Self and Person. Beiträge zur Kultur- und Geistesgeschichte Asiens 75. Sitzungsberichte
(Österreichische Akademie der Wissenschaften, Philosophisch-Historische Klasse) 837. Verlag der
Österreichischen Akademie der Wissenschaften, Wien.
Frauwallner, E. (1935), “Dharmottaras Kṣaṇabhaṅgasiddhiḥ: Text und Übersetzung.” Wiener
Zeitschrift für die Kunde des Morgenlandes 42: 217–258 [=Kleine Schriften 530–571].
Frauwallner, E. (1956), Geschichte der indischen Philosophie. Volume 2. Salzburg.
Frauwallner, E. (1984), History of Indian Philosophy. Volume 2. Translated from original German into
English by V. M. Bedekar. Delhi.
Ganeri, J. (2000), “Cross-modality and the self.” Philosophy and Phenomenological Research, 61(3),
639–657.
Ganeri, J. (2007), The Concealed Art of the Soul: Theories of Self and Practices of Truth in Indian
Ethics and Epistemology. Oxford: Oxford University Press.
Halbfass, W. (1976), “Zum Begriff der Substanz (dravya) in Vaiśeṣika.” Wiener Zeitschrift für die Kunde
Südasiens, 20, 141–166.
Hulin, M. (1978), Le Principe de l’Ego dans la Pensée Indienne Classique. La Notion d’Ahaṃkāra.
Publications de l’Institut de Civilisation Indienne 8 (44). Paris.
Jaini, P. S. (1979), The Jaina Path of Purification. Delhi, Varanasi, Patna: Motilal Banarsidass.
Kano, K. (2001), “Pariśeṣa, Prasaṅga, and Kevalavyatirekin—the logical structure of the proof of
Ātman.” Journal of Indian Philosophy, 29, 405–422.
Kapstein, M. (2001), Reason’s Traces; Identity and Interpretation in Indian and Tibetan Buddhist
Thought. Studies in Indian and Tibetan Buddhism. Boston: Wisdom.
Laine, J. (1993), “Some remarks on the Guṇaguṇibhedabhaṅga chapter in Udayana’s
Ātmatattvaviveka.” Journal of Indian Philosophy, 21, 261–294.
Matilal, B. K. (1986), Perception: An Essay on Classical Indian Theories of Knowledge. Oxford:
Clarendon Press.
Matilal, B. K. (1989), “Nyāya critique of the Buddhist doctrine of non-soul.” Journal of Indian
Philosophy, 17(1), 61–79.
Matilal, B. K. (1994), “The perception of the self in Indian tradition”, in R. Ames et al. (eds.), Self as
Person in Asian Theory and Practice. New York: SUNY Press, pp. 279–295.
Mimaki, K. (1976), La Réfutation bouddhique de la permanence des choses (sthirasiddhidūṣaṇa) et la
preuve de la momentanéité des choses (kṣaṇabhaṅgasiddhi). Publications de l’Institut de Civilisation
Indienne 41. Paris: de Boccard.
Oetke, C. (1988), “Ich” und das Ich. Alt- und Neu-Indische Studien 33. Stuttgart.
Preisendanz, K. (1994), Studien zu Nyāyasūtra III.1 mit dem Nyāyatattvāloka Vācaspati Miśras II. Alt-
und Neu-Indische Studien 46. Stuttgart.
von Rospatt, A. (1995), The Buddhist Doctrine of Momentariness: A Survey of the Origins and Early
Phase of This Doctrine up to Vasubandhu. Alt- und neu-indische Studien 47. Stuttgart: Franz Steiner
Verlag.
Sakai, M. (2010a), Dharmottaras Erklärung von Dharmakīrtis kṣaṇikatvānumāna
(Pramāṇaviniścayaṭīkā zu Pramāṇaviniścaya 2 vv. 53–55 mit Prosa). Doctoral thesis, University of
Vienna.
Sakai, M. (2010b), “Arcaṭa ni yoru Dharmakīrti no Setsunametsu-ronshō Kaishaku: kṛtakatvānumāṇa to
sattvānumāna” (“Arcaṭa’s Interpretation of Dharmakīrti’s kṣaṇikatvānumāna: kṛtakatvānumāna and
sattvānumāna”). Nanto Bukkyo, 93, 38–62.
Sakai, M. (2011), “Śākyabuddhi and Dharmottara on the inference of momentariness based on the
absence of external causes of destruction,” in H. Krasser, H. Lasic, E. Franco, and B. Kellner (eds.),
Religion and Logic in Buddhist Philosophical Analysis. Proceedings of the Fourth International
Dharmakīrti Conference, Vienna, August 23–27, 2005. Vienna, pp. 407–421.
Siderits, M. (2011), “Review of The Self’s Awareness of Itself, by Alex Watson.” Orientalistische
Literaturzeitung, 106(6), 419–423.
Steinkellner, E. (1963), Augenblicklichkeitsbeweise und Gottesbeweis bei Śaṅkarasvāmin. Doctoral
thesis, University of Vienna.
Steinkellner, E. (1968/69), “Die Entwicklung des kṣaṇikatvānumānam bei Dharmakīrti.” Wiener
Zeitschrift für die Kunde Südasiens, 12–13, 361–377.
Taber, J. (1990), “The Mīmāṃsā theory of self-recognition.” Philosophy East and West, 40(1), 36–57.
Taber, J. (2012), “Uddyotakara’s defense of a self,” in I. Kuznetsova, J. Ganeri, and C. Ram-Prasad
(eds.), Hindu and Buddhist Ideas in Dialogue; Self and No-Self. Farnham, Surrey: Ashgate, pp. 97–
114.
Uno, T. (1999), “Ontological affinity between the Jainas and the Mīmāṃsakas,” in S. Katsura (ed.),
Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the Third
International Dharmakīrti Conference, Hiroshima, November 4–6, 1997. Österreichische Akademie
der Wissenschaften, Philosophisch-Historische Klasse, Denkschriften 281. Wien: pp. 419–31.
Vetter, T. (2000), The “Khandha Passages” in the Vinayapiṭaka and the Four Main Nikāyas.
Österreichische Akademie der Wissenschaften, Philosophisch-Historische Klasse, Sitzungsberichte
682. Wien.
Watson, A. (2006), The Self’s Awareness of Itself: Bhaṭṭa Rāmakaṇṭha’s Arguments Against the
Buddhist Doctrine of No-Self. Publications of the De Nobili Research Library 32. Wien.
Watson, A. (2010a), “Rāmakaṇṭha’s concept of unchanging cognition (nityajñāna): influence from
Buddhism, Sāṃkhya, Vedānta,” in J. Bronkhorst and K. Preisendanz (eds.), From Vasubandhu to
Caitanya. Studies in Indian Philosophy and Its Textual History. Papers of the 12th World Sanskrit
Conference. Volume 10.1. Delhi: Motilal Banarsidass, pp. 79–120.
Watson, A. (2010b), “Bhaṭṭa Rāmakaṇṭha’s elaboration of self-awareness (svasaṃvedana), and how it
differs from Dharmakīrti’s exposition of the concept.” Journal of Indian Philosophy, 38, 297–321.
DOI: 10.1007/s10781-010-9094-8.
Watson, A. (2014a), “Light as an analogy for cognition in Buddhist idealism (vijñānavāda).” Journal of
Indian Philosophy, 42(2), 401–421. DOI: 10.1007/s10781-013-9192-5.
Watson, A. (2014b), “The self as a dynamic constant: Rāmakaṇṭha’s middle ground between a
Naiyāyika eternal self-substance and a Buddhist stream of consciousness-moments.” Journal of
Indian Philosophy, 42(1), 173–193. DOI: 10.1007/s10781-013-9215-2.
Watson, A., Goodall, D., and Sarma, S. L. P. A. (2013). An Enquiry into the Nature of Liberation. Critical
Edition and First Translation of the Paramokṣanirāsakārikā and -vṛtti.
Watson, A., and Kataoka, K. (2010), “Bhaṭṭa Jayanta’s refutation of the Yogācāra Buddhist doctrine of
Vijñānavāda: annotated translation and analysis.” South Asian Classical Studies, 5, 285–352.
Yoshimizu, C. (1999), “The development of sattvānumāna from the refutation of a permanent existent
in the Sautrāntika tradition.” Wiener Zeitschrift für die Kunde Südasiens, 43, 231–254.
 13
Where the Self and Other Meet: Early Indian Yogācāra Buddhist
Approaches to Intersubjectivity*
Roy Tzohar

How does our understanding of others and of the external world rely upon and in turn
affect our understanding of ourselves? This question underlies the long-standing
philosophical engagement with the issue of intersubjectivity—that is, the shared nature of
our experiences, in particular of the external world. Intersubjective experiences are
important not just for their role in bridging the self and others, but also because they
present an emergent notion of objectivity: because the world is available to all of us, it
seems we can assume that it exists independently of any of us. For Buddhist non-realist
schools such as the early Indian Yogācāra, who have argued that the phenomenal world is
mind-dependent, this assumption posed a particularly tenacious philosophical challenge.
In this chapter, I wish to explore the ways in which these Buddhist Yogācāra thinkers, and
especially one of them, Vasubandhu, who lived around the fourth century CE, attempted
to explain the possibility of intersubjective agreement under a “mind only” view. In what
follows, I present Vasubandhu’s arguments on this issue and those of some of his
commentators as a way of drawing out their broader understanding of intersubjectivity, as
well as its attendant conceptions of personal identity, of the life world, and of otherness.

* * *

Vasubandhu (typically dated to around the late fourth to early fifth century CE) is one of
the most influential philosophical Buddhist thinkers. His work left an enduring mark both
on subsequent generations of Indian Buddhist philosophers and on the traditions of the
Mahāyāna outside India. He is ascribed the authorship of seminal textual works in a
variety of genres,1 and traditionally considered to be one of the founders, along with his
half-brother Asaṅga, of the Mahāyāna Buddhist Yogācāra school. The name Yogācāra
literally means the path or the application of yoga, marking the school’s interest in
practical and meditational aspects, and other, later epithets of this school are
“Vijñānavāda” (a doctrine of consciousness), and “Cittamātra” (“mind only” or
“consciousness only”), both of which reflect its emphasis on the analysis of cognitive
processes and the activity of consciousness.
The question of the proper interpretation of the early Indian Yogācāra philosophical
worldview, and especially of its claim that all phenomena are merely mental
representations (vijñapti-mātra), has been the subject of an ongoing scholarly
controversy.2 Roughly speaking, on the one side of this debate are scholars who maintain
that the Yogācāra arguments do indeed aim to establish a kind of subjective or
metaphysical Idealism, according to which nothing exists outside of the perceiving
consciousness,3 and on the other are scholars who question the adequacy of attributing
any “idealistic” ontological claims to the Yogācāra and instead interpret the school as
offering a variant of epistemic idealism,4 a stance by which direct knowledge of
externality as such is impossible, and hence all knowledge is given within a mental realm.
A detailed discussion of this controversy is beyond the scope of this chapter, however, as
for the proper interpretation of Vasubandhu’s Yogācāra views in relation to the terms of
this broader debate, there seems to be a scholarly consensus regarding the following
premises: (1) that he holds that external objects do not exist as such (i.e., as they appear
to us), and (2) that he holds that all phenomena can be known or discussed solely in terms
of mental representations or impressions (vijñapti). These mental representations, in turn,
are explained as the outcome of the ongoing activity of consciousness, understood as a
constant flux of causal mental events (I will return to this below).
Bearing this in mind, let us now turn to examine Vasubandhu’s explanation of
intersubjective agreement in one of his shortest but highly influential philosophical works,
the “Treatise in twenty verses” (Viṃśika). Like many other works of Indian philosophical
literature, this short piece was written in a polemical style, its first verse stating
Vasubandhu’s thesis as follows: “All this [the three realms of existence] are merely mental
representations, because of the appearance of non-existent objects, just like in the seeing
of non-existent hair-nets etc. by one suffering from an eye disease (timira).”5 Here,
Vasubandhu uses the paradigmatic example of a person whose eyesight affliction causes
him to see objects that he takes to be real and external while in fact they are neither.
Similarly, Vasubandhu asserts, while we conceive of the world and of ourselves as
objectively existent phenomena, in fact these are merely mental impressions or
appearances—projected, so to speak, by our mind—which erroneously appear to us as if
they exist externally and independently of the mind that conceives them.
Such a forthright ontological denial of the existence of external reality, however, is of
course very hard to defend philosophically. The realist can rightly point out that in order
to refute the objectivity of external reality, Vasubandhu must first accept it, in practice, as
a necessary epistemic point of reference both for his experience and for his arguments,
and in this sense, his stance involves a practical contradiction.6
Avoiding such a critique, therefore, calls for a different philosophical strategy. And
indeed, Vasubandhu’s argument, as we will shortly see, shifts ground from an ontological
to an epistemic context: instead of providing a proof (arguably impossible) for the
nonexistence of external objects, he aims to show that our view of external objective
reality as self-evident is unfounded and does not withstand philosophical scrutiny. His
argument can be reconstructed as follows: since all our knowledge of the world is by
definition mediated by our perceptions, we can never know an external object as such,
that is, independently of our perceptions of it; under these conditions, we have no definite
way of knowing that what we see is indeed external, and no definite criteria for
differentiating between the internal and the external, between physical objects and
mental constructions. As we will see below, to support this line of argument Vasubandhu
offers the state of dreaming as a paradigmatic example of circumstances in which
“internal” objects are falsely imagined to be external. In the same way, he suggests, it may
well be that we are fundamentally mistaken in our current waking judgment of the
external world, without being aware of it. Although this example cannot establish the
nonexistence of external reality, it can shake our self-assured dogmatism about the
existence of such a reality, and this is Vasubandhu’s goal at this stage.
His opponent,7 however, does not easily succumb to doubt. In a series of objections, he
tries to show that even though our knowledge of external objects is necessarily mediated
by our perceptions, there nonetheless exists a set of definite criteria for distinguishing
between external objects and internal cognitions. He then goes on to point out three
features of our experience that, according to him, cannot be explained satisfactorily
unless we presuppose the existence of external physical objects as the source of our
mental images.
The first of these is spatial and temporal determinacy, that is, the fact that our
experiences of objects are all subject to certain spatial and temporal limitations. This
feature can be explained, contends the realist, if we accept the physical existence of
objects, that is, their existence in a particular time and space. To this Vasubandhu replies
that, in dreams, objects similarly appear to be limited to particular locations and times,
yet they are nonetheless unreal. Spatial-temporal determinacy, therefore, does not
indicate the existence of an external reality.8
Another objection9 stated by the opponent is that real objects, unlike objects in a
dream, have real efficacy, that is, they perform actual work, whereas actions in a dream
do not carry real consequences. In response, Vasubandhu points out that it is not true that
dream actions are never efficacious, offering as a counterexample the phenomenon of
nocturnal emission. If the example appears unconvincing insofar as it marks an exception
rather than the norm of dream states, we should recall that the opponent seeks a definite,
unerring criterion for distinguishing between reality and mental creation, so that even a
single counterexample suffices to undermine his efforts.
The most challenging objection raised by the opponent, however, concerns the
possibility of intersubjective agreement, that is, of different perceivers having
simultaneous similar experiences. This kind of intersubjective agreement is absent, the
opponent argues, when the relevant experiences are known to be mere mental
impressions, as in Vasubandhu’s own examples of the eyesight affliction, or in a dream
state, both of which are private, not shared. Intersubjective agreement, therefore, can
serve us to differentiate between dreams and the waking state, between reality and
mental creation.10 Furthermore, the opponent argues, the possibility of intersubjective
agreement seems to support the existence of external physical objects—otherwise, how
can we explain the fact that we all see the same things simultaneously? We therefore must
conclude that the objects of our perception exist independently of the minds that perceive
them.
This last objection poses a serious difficulty for Vasubandhu insofar as he cannot appeal
again to the dream analogy. But he is not at a loss for arguments. He responds by using
two examples drawn from Buddhist cosmology, the first concerning the realm of the
“ghosts” (pretas) and the second involving the beings that populate the Buddhist hells
(naraka).11
The pretas12 are beings who in previous lives had been overcome by lust and greed, and
whose current state is marked by unquenched hunger and thirst.13 They are said by
Vasubandhu to have shared visions—induced by their common previous karma—of rivers
of pus and filth guarded by violent guards, images that are in fact only imagined by them
rather than actually existent, thereby demonstrating that intersubjective agreement is
possible even with respect to nonexistent objects.
But why are the pretas’ visions said to be merely a figment of their imagination?
Vasubandhu does not elaborate, perhaps assuming that his audience would already be
familiar with this example. Other Yogācāra sources (see n. 12), however, clarify that the
pretas hover on the peripheries of the human world, and where humans see rivers of
crystal clear water they see only rivers of pus and so on. It is therefore this discrepancy
that attest to the unreality of the content of the pretas’ shared experience. It should be
noted, however, that this argument alone cannot satisfactorily undermine intersubjective
agreement as evidence for the existence of external physical objects, since doubt in
respect to the validity of the pretas’ shared experience is dependent upon an implicit
reaffirmation of the validity of our own shared experience (in this case, of the clean rivers)
as a criterion for truth, which is precisely the premise that Vasubandhu seeks to undercut.
So Vasubandhu needs a better argument regarding intersubjective agreement, one that
seeks not just to falsify an experience by another one14 but rather to undermine the entire
experiential framework as such. To this end, Vasubandhu presents the example of the
hells (naraka), in which a group of hell beings, that is, beings whose actions in past lives
caused them to be reborn into a hell state, are subjected to various torments by ferocious
hell guardians. According to Vasubandhu, however, these hell guardians, vivid as they may
seem, are unreal and merely collectively and simultaneously imagined by all the hell
beings.
Why does he assume this to be so? The answer has to do with the particular nature of
the Buddhist hell state. Unlike the hells of the Abrahamic traditions, where one arrives
following a divine judgment, arriving at a Buddhist hell is seen as the inevitable outcome
of a certain way of life, an eventuality governed by karma—the same general causal
principle that regulates the relations between causes and their effects among all living
creatures and which entails that certain causes are bound to lead to certain effects, if not
in this life then in the next. In this sense, every being will experience the fruition of his or
her volitions and actions—not through the meditation of a punishing judge, but in a
causal, almost mechanical manner.
According to this dogma, the hell state by definition contains only beings who are there
by right, that is, as the outcome of their personal karmic retribution, which they
experience as various forms of torture. But how, then, to account for the presence there of
the hell guardians, who are not tormented? Vasubandhu answers quite simply: such
guards cannot really exist and therefore are only imagined to be there. Note that in this
case, in contrast to the example of the pretas, the unreality of the shared experience is
indicated not by its discrepancy with some other contradictory experience but, rather, by
its status as internally logically incoherent (this is an important point to which I will
return below). If the hell guardians are unreal, how then can they be simultaneously
imagined by all the hell beings? Vasubandhu answers that, as in the case with the pretas,
this is simply the outcome of the hell beings’ shared karma—their similar past
circumstances bring about similar present experiences, in which they imagine the hell
guardians as the tangible cause of their otherwise inexplicable suffering. Intersubjective
agreement, he concludes, is therefore possible even with respect to imagined objects, and
hence cannot serve as a criterion for establishing the existence of a physical external
reality.15
What are we to make of the seemingly dogmatic way in which this conclusion has been
reached? Even if Vasubandhu’s opponent shares his premises about the hell states, still,
this justification for the radical epistemic claim that Vasubandhu wants to establish can
seem somewhat narrow. As we will shortly see, however, Vasubandhu’s argument is more
broadly and subtly designed than it first appears. His choice of the hell example, apart
from conveying in condensed form an account of how we commonly construct our reality,
also allows him to reflect on the inevitable suffering and ignorance that are linked to this
process, and on the role of philosophical discourse in easing this suffering. Yet in order to
appreciate these broader intentions, we must first make sense of his statement that the
hell beings all see the same visions because of their “shared past karma.” To this end, we
need to deepen our understanding of the workings of karma and of Buddhist hells under
the Yogācāra causal mentalistic framework. Specifically, we need to consider how the
notion of karma shapes the Buddhist understanding of what it means to be a sentient
being and of the various worlds that these beings inhabit.

* * *

As mentioned earlier, according to the Yogācāra, the enlightened mind conceives

phenomena in terms of a reduction to mental representations or appearances, the
outcome of a causal flow of mental events, which cannot be traced to any particular
chronological or teleological point of departure. This is not, of course, the way we
ordinarily view reality. According to the Yogācāra, this is because our primordial
ignorance leads us to apply to this causal mental flow a dualist distinction between a
grasper and a thing that is grasped (roughly, the subjective and objective aspects of our
experience respectively), and as a consequence, to apply to it many further
conceptualizations and reifications, culminating in a view of the self and the world of
objects as distinct entities existing objectively and independently of the mind that
perceives them. This process is systematically described in great detail by the Yogācāra
account of the activity of consciousness in terms of its “transformation” (pariṇāma) from
the most fundamental and subliminal level up to our everyday awareness, and which,
insofar as it is casual and interdependent, is understood by Yogācāra thinkers as the
Buddhist fundamental notion of dependent origination (pratityasamutpanna).16 I explicate
this scheme in greater detail below, but for now, let us turn to consider the way in which,
within this framework, karma is involved in the construction of our lifeworld. In his
commentary to the treasury of the Abhidharma (Abhidharmakośabhāṣya), Vasubandhu
famously asserts that the cause of the arising of the “world of sentient beings and the
‘receptacle’ inanimate world (sattva-bhājana-loka) is none other than the accumulated
karma of its beings.”17 To unpack this claim, we will proceed gradually, considering first
the category of a sentient being, then of worlds, and finally, the possibility of
intersubjective agreement—all under the mentalist framework presented above.
In Buddhist cosmology, to be a sentient being means it is necessary to belong to one of
the five18 realms or forms of existence—that is, the various classes of gods, humans,
animals, ghosts, and hell beings. Under the Yogācāra’s mentalist causal framework, each
life-form—for instance, the life-form of humans—is explained as constituting a particular
mental causal continuum, a certain mind-stream (saṃtāna)19 within the general causal
mental flow, which is how the Yogācāra understands karma. Thus, humans differ from the
hell beings insofar as these two life-forms constitute different patterns of mental-
continuums, they are the outcome of different karmic causal strands.20
The same principle is at work with respect to the various realms or worlds that these
life-forms are said to occupy. This is because under a mentalist causal framework, being a
certain life-form, insofar as it is delimited by a particular pattern of causal mental
continuum, implies a certain limited range of possible experiences. Thus, being human,
for example, means having at one’s disposal a certain range of possible experiences—
certain mental appearances, but not others. This range of possible experiences is what we
call the human realm, and it is shared, by definition, by all human life-forms. Being a hell
being implies a whole different range of possible experiences, as prescribed by that
particular mental continuum, and this range is what we call “the hells.”
The world of animals provides another example, perhaps more easily understandable
insofar as animals are more familiar to us than the hells but nonetheless representing a
separate realm of existence according to Buddhist cosmology. Being a dog, for instance,
means experiencing reality in a radically different way than we humans experience it (in
terms of sensory data, emotional reactions, conceptual categories, etc.), as prescribed by
the causal continuum that we call “a dog,”21 and in this sense alone the animal is said to
occupy a different realm, or world.
It is now evident, then, that the notion of a “world” designates more than a mere
physical locus. Different beings can have shared experiences and even occupy the same
physical environment and yet occupy entirely different realms of existence and worlds.
Take the pretas, for instance: as described above, they share some experiential content
with humans and appear to occupy the same physical space as humans—both see rivers at
the same time in the same place—and in this respect their causal continuum and our own
(i.e., our respective karmic strands) have certain shared features; but they also differ in
many respects, and these differences explain the deep generic discrepancies between
their experiences and ours (pus versus clear water). These differences justify regarding us
as occupying different worlds.
Instances such as these seem to indicate that for Buddhist cosmology the various
realms of existence are a matter not merely of loci but also of a commonly experienced
state of mind. Rupert Gethin has described this feature in theoretical terms as amounting
to a fundamental equivalence in Buddhist thought between cosmology and psychology,
and a blurring of the distinction between physical loci and mental states.22 This
equivalence may explain, for instance, why we find that while the Buddhist realms of
existence mark a concrete place—the hells are said to exist in the spaces between the
world spheres (cakra-vāḍa), humans occupy the southern continent, Jambudvīpa, the gods
dwell in heavenly realms that surround the slope and peak of mount Meru, and so on—
these worlds are seen to be entirely dependent on their beings insofar as a world comes
into existence only when its beings are born into it and disappears if they all cease to
exist.23
To recap our track thus far: We have seen that under the framework described above,
our subjective experiences, the kind of life-forms that we are and the worlds we inhabit
are all nothing but mental appearances, the end products of a causal mental flow. A self is
just a certain mental causal continuum; a life-form stands for certain similar patterns
among these continuums; and the totality of shared experiences available to these life-
forms, either as a possibility or in actuality, is what we call a world. Thus, insofar as these
categories are all reduced to causal and mental factors, the differences between persons,
life-forms, and worlds are not essential but merely causally circumstantial (which is,
ultimately, what it means to say that everything is constituted and governed by karma).
While this framework explains how we, as humans, share a world, and why this world
can only be as it is and not otherwise—why in the human realm of existence there can be
people and chairs and rivers of clear water, and not rivers of pus—it still needs to explain
how, within this shared world, there can be intersubjective agreement about the content
of certain particular experiences while other experiences are accessible only to the
subject. In other words, given that both we and the human world around us are merely
mental appearances, what is it that allows all of us simultaneously to have the same, or
almost the same, appearances?
The answer to this question is found again in the notion of shared karmic strands,
explained however by the Yogācāra in greater resolution in terms of the activity of
consciousness and its role in the construction of the lifeworld. The aforementioned notion
that the world arises from the actions of its beings was further developed and explicated
in the various Yogācāra treatises, where it is traced to the function of the most
fundamental and unconscious level of awareness, the storehouse consciousness (ālaya-
vijñāna). This level of consciousness has a pivotal karmic role in the sense that it is the
locus of what the Yogācāra refers to, metaphorically, as karmic causal “seeds” (bīja) and
their “impressions” (vāsanā), that is, their karmic consequences. The general contours of
this account are that any past experience leaves a certain “impression” in the storehouse
consciousness, which then serves as a karmic causal “seed” that will ripen and give rise to
a certain experience, which will in turn leave its own impression on the storehouse
consciousness, and so on, recursively.24
One of the various explanatory roles performed by the notion of the storehouse
consciousness is that of explaining how our lifeworld can be causally accounted for by
karma. An account of this process appears, for instance, in the first chapter of Asaṅga’s
Mahāyānasaṃgraha (MS), where our common surrounding “receptacle world” (bhājana-
loka) and personal sense sphere (prātyatmikāyatana)—respectively, our shared
intersubjective experiences and what will be described for now, for lack of a more
accurate translation, provisionally as “private” experiential content—are traced,
respectively, to the maturation (vipāka) of similar and dissimilar karmic seeds (bīja) and
impressions (vāsanā) in the storehouse consciousness.25 So, simply put, whatever causal
mental activity is shared at any given moment by our respective mind-streams will appear
as intersubjective, and whatever causal mental activity is not shared will be experienced
privately. We can all have a simultaneous perception of the same object because of our
shared karmic seeds and impressions, but we do not perceive it in exactly the same way
(in terms of visual perspective, for instance) and we do not know one another’s
accompanying thoughts because that portion or activity of our mind-streams is not
shared.
It should be noted, however, that the categories of “private” and “shared” experience
require further qualification in this context, since according to the Yogācāra, all of our
experiences, even the most private, are pervaded by conceptual categories, which are by
definition shared and constitutive of our general life-form. The Yogācāra assigns a pivotal
role in the creation of intersubjective content, within the storehouse consciousness, to the
mediating activity of conceptual distinctions manifesting as particular “impressions of
speech” (abhilāpavāsanā), which are understood to be causally efficacious. In other
words, according to the Yogācāra, conceptual linguistic activity is itself a causal
phenomenon: it is causally induced and causally effective. These impressions of speech
are seen both to bring about and in turn to be recursively informed by linguistic
conventions, categories, and conceptual proliferations (prapañca), which are responsible
for the shared elements in our cognitive experiences.26
According to this view, our experience, which is initially an undifferentiated causal
mental flow, necessarily passes through certain conceptual filters. At the most
fundamental level, this manifests in a basic conceptual distinction—vikalpa—that
separates all experience into the categories of a subjective aspect (a grasper) and an
objective aspect (what is grasped). This first and most basic distinction is the original sin,
so to speak, following which many other conceptual categories are imposed on our
otherwise undifferentiated experience so as to organize it into meaningful units. It is
therefore the inescapably intersubjective nature of our conceptual activity and language
use, seen as casually efficacious, that accounts for the common content of our experiences
and of the world that we inhabit.
We are now in a position to revisit our initial question of how the workings of karma
induce intersubjective agreement among the hell beings. We saw that for the Yogācāra,
intersubjectivity is constituted by the shared portions of individual karmic streams. This
accounts for a wide array of phenomena, ranging from the shared aspects of our forms of
life and our shared worlds, to the more narrow content of particular intersubjective
experiences. With respect to the latter, especially, our conceptual activity and language
play a constitutive role. Now, applying this picture to the case of the hells and hell beings,
we can see that it is the shared karmic seeds and impressions mediated by specific
conceptual and linguistic cultural baggage that account for these life-forms’ particular
mode of being, for the physical-experiential world that they share, and for the specific
content of their experiences—including, in particular, their shared vision of the hell
guardians and more generally of a Buddhist hell, and not, say, of Dante’s inferno.
Furthermore, and here we come full circle to the original purpose of the pretas’
example, this framework also provides the Yogācāra with a more phenomenologically
sophisticated description of intersubjective experiences, one that accounts not merely for
experiential agreement (for that which we and the pretas all see alike, i.e., the rivers) but
also for the discrepancies in different perceivers’ experiences of the same object (clear
water versus pus). Thus, explaining both the shared and private aspects of experience
through the same function of causal patterns (their similarity or dissimilarity
respectively), the Yogācāra account treats intersubjective agreement as indicating nothing
more than the fact that some experiences were commonly and casually constructed.
Accordingly, the discrepancies between the experiences of different perceivers with
respect to a single object are seen simply as the flip side of the same inherent aspect of
this process of construction, that is, they are seen as phenomenologically “built into” our
shared experiences.
 Having presented Vasubandhu’s response to the objection regarding intersubjective
agreement, let us finally summarize his argument: Ordinarily, we presuppose the
existence of an external reality as the basis of all our knowledge of the world. However,
since our knowledge of existence is necessarily given within a mental realm (to exist
means to be perceived), and since Vasubandhu has demonstrated that no criterion (not
determination in space and time, not efficacy, and not intersubjective agreement) can
differentiate definitively between the internal and the external, we have no way of
knowing externality as such and no way of vouching for its objective existence. (In order
to know with certainty that external objects exist we would have to step outside of our
own epistemic framework, but this we cannot do, by definition.) This form of skepticism
about the external world might not suffice to bring the realist opponent to admit a
universal error and embrace the Yogācāra account of experience in causal mental terms,
but it suggests the uncritical nature of his position, which is founded on common sense
and habit more than on reasoning, and by extension establishes the possibility that we are
mistaken in all of our ordinary perceptions.
How does this argument and the doubt it fosters fit into the larger Yogācāra and
general Buddhist worldview? What does it have to do with suffering and its cessation, and
with the Buddhist soteriological horizon? A possible answer is implied by Vasubandhu’s
choice and careful rendering of the hell example, and by the kind of reflection it
encourages about the role of philosophy and the limits of our understanding. The hell
example shows us not just that different perceivers can share the same illusory vision
(seeing the hell guards) but also that this vision is shared by them despite being logically
impossible. This highlights an interesting feature of the hell example, namely, that the hell
guardians’ presence is deemed impossible by the same basic logic that renders it essential
(their presence, among other factors, is what makes hell “hellish”). A condition that is
necessary but at the same time also impossible is therefore the mark of the hell state.27
But recalling the initial premises of this polemic, we can see that this paradox is also the
mark of the realist’s position: on the one hand, he takes external objective reality as the
necessary and self-evident basis for our experience, while on the other he is forced to
admit that insofar as direct knowledge of this objective ground is impossible, its existence
remains a matter of conjecture.
The analogy is rather straightforward: like the hell beings, we too are trapped in a
world of delusion, unaware of the ways in which our mind is involved in the creation of
both our samsaric prison and our torments, which we feel powerless to resist. This
samsaric epistemic “hell” is the only reality we know and feel we can know, but then
comes along a philosopher, and from within the confines of this epistemic trap, points out
using reasoning that what we take to be self-evident and unquestionable is in fact
unfounded. While this reasoning does not cause our hell and its various guardians to
disappear—and to this extent, philosophy alone cannot liberate us—it has the power to
shake our assurance of their status as given and creates an opening through which the
Yogācāra doctrine may enter our awareness, challenging and realigning the boundaries
between ourselves, the world, and others.

Abbreviations
AKBh Abhidharmakośabhāṣya
MAVṬ Madhyāntavibhāgaṭīkā
MS Mahāyānasaṃgraha
SNS Saṃdhinirmocanasūtra
TD Derge Edition of the Tibetan Buddhist Canon (Bstan ‘Gyur Sde Dge)
Viṃś Viṃśika

Notes
* This chapter is part of a broader research project on the concept of “world” (Sattva-bhājana-loka)
in Early Indian Yogācāra Buddhism, which I have been pursuing with the generous support of the
Marie Curie IRG fellowship of the EU (CORDIS). An early version of the chapter was presented as
part of the Aspects of No-Self lecture series at the Numata Center for Buddhist Studies, University
of Hamburg, Germany, and published in Tzohar (2016. DOI 10.1007/s11841-016-0544-y). My
discussion here closely follows and summarizes this source. I am grateful to the center’s director,
Michael Zimmerman, for his valuable feedback, as well as to my colleague Sonam Kachru of the
Max Planck Institute for the History of Science, Berlin, with whom I had many stimulating
 conversations about Vasubandhu and whose comments on an early version of this chapter made it
better.
1. For a general bibliographical survey of Vasubandhu’s work’s and the scholarly controversies
regarding his dating, authorship, and doctrinal affiliation, see Tzohar (2013).
2. Apart from its debate about the proper interpretation of the Yogācāra doctrines, much of this
controversy turns on the precise semantic and philosophical meaning attributed to the term
“idealism.” For opposing interpretations of Vasubandhu and a thorough picture of the current state
of this debate, see Garfield and Gold’s (2011) relatively recent public polemic: “On the Reality of
the Mind in Yogācāra: A Constructive Debate on Vasubandhu’s Trisvabhāvanirdeśa”; Lusthaus
(2003); and Schmithausen (2005).
3. This sort of interpretation of the Yogācāra ideas can be found, for instance, in the early
translations and interpretations by La Vallée Poussin (1928) and Suzuki (1930). Later, it is found
in, among others, Matilal (1974), Griffiths (1986), Wood (1991), Hopkins (1999; 2002), Siderits
(2007), and Schmithausen (2005).
4. Works that feature such an interpretation—while varying in their ontological commitment—
include Wayman (1965), Ueda (1967), Willis (1979), Kochumuttom (1982), Kalupahana (1987),
Lusthaus (2003), and Gold (2006).
5. vijñaptimātram evaitad asadarthāvabhāsanāt |yathā taimirikasyāsat
keśacandrādidarśanaṃ||Viṃśika 1. Lévi (1925).
6. This is a variant of a standard realist response to the skeptic: showing that his skepticism is
necessarily parasitic upon knowledge that appears as self-evident, and as such, renders his doubt
incoherent.
7. Vasubandhu does not state the opponent’s identity anywhere in the work, but Vinītadeva’s
subcommentary (rab tu byed pa nyi shu pa’i 'grel bshad T.4065) takes him in this case to be a
“proponent of [the existence of] external objects” (phyi rol gyi don du smra ba), that is, a direct
realist. In the ensuing arguments, which deal with intersubjective agreement, Vinītadeva
attributes some objections to a Vaibhāṣika Buddhist realist (bye brag tu smra ba), or a Sautrāntika
(mdo sde pa) whom he takes to be promoting a form of representationalism. See Derge (199–?), TD
175a1, 178b3, 180a7, respectively.
8. For an assessment of this objection, see Siderits (2007, pp. 150–151).
9. For the sake of clarity, I present the objections and their responses in an order other than the one
in which they appear in the Viṃśikā. For a presentation that preserves their original order, see
Lévi (1925, pp. 3–5).
10. | taddeśakālapratiṣṭhitānāṃsarveṣāṃsaṃtāna utpadyate na kevalam ekasya | yathā
taimirikāṇāṃsaṃtāne keśādyābhāso nānyeṣāṃ | Lévi (1925, p. 3). In discussing intersubjective
agreement, both the objection and the ensuing response use the term “streams” (saṃtāna) rather
than “persons” (whose existence both Vasubandhu and the opponent apparently deny). These
“streams” are said to stand for the mental continuums that underlie and come to constitute such
imagined entities as persons. This already represents a highly phenomenal perspective on
experience, according to which the opponent questions the ability of a “mere representations”
view to account for the epistemic boundaries between one such mental stream and another—
boundaries manifested, for instance, in the sharp distinction we draw between shared and strictly
private experiential content. As we will see below, the Yogācāra response is to account for this
distinction—and indeed for any experiential particularity—with causal explanations. This
understanding of saṃtāna is clearly reflected in a definition provided by Vasubandhu elsewhere—
in the tenth chapter of his seminal “Treasury of Abhidharma and Commentary”
(Abhidharmakośabhāṣya), which is dedicated to a refutation of the existence of persons. As
translated by Kapstein (2001, p. 374):
That which, preceded by deeds, is an on-going coming-to-be of mental events, is a continuum
[saṃtāna R.T]. Its arising otherwise is transformation. And moreover, that potency, which
immediately produces the fruit, being distinct from [that involved in] other transformations, is the
distinctive feature of the transformation.
This definition, which equates saṃtāna with a causal mental process, indicates that the
distinctiveness of each such stream is accounted for by the particularity of the specific causal
chain that constitutes it. Similarly, the particularity and distinctiveness of any experience is
accounted for by a certain “restriction” or “delimitation” (niyamaḥ) imposed upon it, so to speak,
by the particularity of the causal chain of mental events that constitutes it.
11. Some contemporary studies conflate these two examples (often by stripping them off their
Buddhist cosmological particularities) and view them both as making a general case, applicable
even to our times, for the possibility of collective hallucination or psychosis. See Matilal (1974, pp.
142–143), Wood (1991, pp. 165–167), and Siderits (2007, pp. 152–153, 155–156). While
Vinītadeva’s subcommentary on the Viṃśika also appears to consider both examples as making
more or less the same point (TD 4065 D. 177a4–6), a number of other, closely related Yogācāra
sources all appear to treat the preta and the naraka examples as separate (if complementary)
arguments. See MS II 14, 14b in Lamotte (1973, pp. 30–31, 104–106 n.14; and the
Madhyāntavibhāgaṭīkā (MAVṬ) on chapter I verse 3, in Yamaguchi and Lévi (1934, 18 [line 25]–19
 [line 14]).
12. Pra-ita, literally “gone forth,” “departed.”
13. See Jones (1949, pp. 22–24).
14. In the sense of treating the pretas’ shared vision as a perceptual illusion of sorts—a collective
hallucination or a mirage shared by several perceivers undergoing similar circumstances, whose
fictional nature is revealed once its efficacy is put to the test. As will be demonstrated below,
Vasubandhu did not make do with this sort of argument from illusion, and it appears that other
Yogācāra sources also affirm that this was not the aim of this example. For instance, Sthiramati’s
MAVṬ (a subcommentary on a work and commentary ascribed to Maitreya/Asaṅga and to
Vasubandhu respectively) uses the preta example to point out a fundamental oversight in the
realist’s account of intersubjective experiences. Briefly, Sthiramati argues that the realist
framework—which traces intersubjective agreement to mind-independent objects—cannot account
for the inevitable perceptual discrepancies between the experiences of various perceivers with
respect to the “same” object (as in the case of the pretas’ perception as compared to that of
humans). The Yogācāra, however, he argues, can provide an alternative explanation for both
intersubjective agreement and experiential discrepancies in strictly causal mental terms, that is,
without recourse to externally existing physical objects. Sthiramati’s argument thus capitalizes on
the realist’s failure to identify the full phenomenological complexity of intersubjectivity—which is
characterized not merely by agreement but also by inherent discrepancies—and in turn presents
what he takes to be a more complete account of such experiences. Rather than using the pretas
example to establish the possibility of a collective hallucination, his argument presents an ironic
inversion of the realist premise, showing that intersubjective agreement not only does not require
the existence of mind-independent objects but is in fact incompatible with their existence. For
Sthiramati’s comments, see Yamaguchi and Lévi (1934, 18 [line 25]–19 [line 14]).
15. yathā hi narakeṣu nārakāṇāṃnarakapālādidarśanaṃdeśakālaniyamena
siddhaṃśvavāyasāyasaparvatādyāgamanagamanadarśanaṃcety ādigrahaṇena sarveṣāṃca
naikasyaiva taiś ca tadbādhanaṃsiddham asatsv api narakapālādiṣu
samānasvakarmavipākādhipatyāt| ... Lévi (1925, p. 4). The passage indicates that insofar as the
hell guardians are perceived in a certain time and a place, and effect tortures, this example also
serves as a response to the objections regarding spatial and temporal determinacy and efficacy.
16. This framework is outlined in Vasubandhu’s Triṃśikā and further developed in Sthiramati’s
commentary, the Triṃśikā-bhāṣya. See Tzohar (2017).
17. “By whom was the manifold variety of the sentient and in-sentient [receptacle] world created? ...
The variety of the world arises from action of beings...” sattva-bhājana-lokasya bahudhā vaicitryam
uktaṃtat kena kṛtaṃ [?] ... sattvānāṃkarmajaṃloka-vaicitryaṃAKBh IV.1a, Pradhan (1975, pp. 192
[3–5]).
18. According to Vasubandhu’s AKbh III.4ab. Pradhan (1975, pp. 114 [4–7])). Other sources
enumerate six realms of rebirth, adding also the asuras. See La Vallée Poussin (1988), 500 n.26.
19. For a definition of saṃtāna, see n. 11.
20. In this respect, every form of life, insofar as it constitutes a certain pattern of causal mental
events, is seen as the outcome of a respective chain of past causal mental events, that is, as the
outcome of a particular past karma. For an account of the Buddhist conception of the category of a
species and of the world in terms of karma, see Waldron (2003, p. 165), and Kachru (2015). Kachru
provides a detailed account of the way in which these concepts map unto a notion of volitional
sentience in Vasubandhu’s Yogācāra and Abhidharma thought.
21. See Siderits (2007, p. 155).
22. Gethin (1997, p. 211).
23. Kajiyama (2000, p. 188); and La Vallée Poussin (1988, p. 476).
24. Of course, the notion of the storehouse consciousness as the storage place of seeds and
impressions is itself a metaphor rather than an actual locus or essence. This type of consciousness
was an innovation of the Yogācāra and played an important explanatory role in various other
contexts pertaining to meditational practices, personal continuity in death and rebirth, and the
shared construction of the lifeworld. For a summary of these various roles, see Yamabe (2004); and
for a detailed discussion, see Waldron (2003, especially chapter 1 and appendix II).
25. In Asaṅga’s Abhidharmasamuccaya, the distinction between shared and unshared karma appears
to explain in the first place the distinction between sentience and insentience, in this case between
what is understood to pertain to living beings (i.e., karma producing) and to the surrounding world
of objects. However, the ensuing use of shared karma to explain also the reciprocal relations
between living beings (their seeing of one another, for example) seems to interpret this distinction
along more phenomenal lines, aligning it with the distinction between shared and unshared
elements in our experience. In the MS, this tendency is even more pronounced as the distinction
between shared and unshared karma is further qualified by the commentators so as to
differentiate between the world and what pertains to the individual and between the common and
private aspects of experience, respectively. See Walpola (2000, p. 118); MS. I.59–60 ibid., 81–84;
Lamotte (1973).
26. See Waldron’s (2003) analysis of the MS 1.58–61 discussion of shared and unshared karma, and
the Saṃdhinirmocanasūtra (SNS), chapter 5 section 2, understanding of the storehouse
 consciousness with respect to the “the appropriation which consists of the predispositions toward
profuse imaginings in terms of conventional usage of images, names, and concepts” (nimitta-nāma-
vikalpa-vyavahāra-prapañca-vāsanā-upādāna). Waldron (2003, pp. 158–161, 164–169). For a
discussion of the possible interpretation of vāsanā-upādāna in the context of the SNS, see
Schmithausen (1987, pp. 71–74 [4.4.2–4.5.2]).
27. Biderman (2008, pp. 273–274). Biderman reads the logical untenability of the hell guardian’s
argument as serving Vasubandhu to establish a coherentist notion of truth, with which he counters
the realist opponent.

Bibliography
Biderman, S. (2008), Crossing Horizons: World, Self, and Language in Indian and Western Thought.
New York: Columbia University Press.
Derge, T. (199–?). Bstan ‘Gyur Sde Dge’i Par Ma: Commentaries on the Buddha’s Word by Indian
Masters (Electronic ed., 11 CDs). (Reproduced from Editions of the Individual Sections Published in
Delhi at the Delhi Karmapae Chodhey, Gyalwae Sungrab Partun Khang, 1986, Itself a Reproduction
from Prints from the Eighteenth-Century Sde-Dge Blocks; 213 vols.) New York: Tibetan Buddhist
Resource Center.
Garfield, J., and Gold, J. (2011), “On the reality of the mind in Yogācāra: a constructive debate on
Vasubandhu’s Trisvabhāvanirdeśa,”
http://religion.princeton.edu/events/2011/aarpanel/AAR_Conference_Panel_2011.html.
Gethin, R. (1997), “Cosmology and meditation: from the Aggañña-Sutta to the Mahāyāna.” History of
Religions, 36 (3), 183–217.
Gold, J. C. (2006). “No outside, no inside: duality, reality and Vasubandhu’s illusory elephant.” Asian
Philosophy, 16 (1), 1–38.
Griffiths, P. J. (1986), On Being Mindless: Buddhist Meditation and the Mind-Body Problem. La Salle,
IL: Open Court.
Hopkins, J. (1999), Emptiness in the Mind-Only School of Buddhism. Berkeley: University of California
Press.
Hopkins, J. (2002), Reflections on Reality: The Three Natures and Non-natures in the Mind-Only
School: Dynamic Responses to Dzong-Ka-Ba’s the Essence of Eloquence. Berkeley: University of
California Press.
Jones, J. J. (1949), The Mahavastu Vol.1. London: Luzac.
Kachru, S. (2015), “Minds, bodies and worlds: philosophy beyond empiricism in Vasubandhu’s twenty
verses.” PhD dissertation, University of Chicago.
Kajiyama, Y. (2000), “Buddhist cosmology as presented in the Yogācārabhūmi,” in Jonathan A. Silk
(ed.), Wisdom, Compassion, and the Search for Understanding: The Buddhist Studies Legacy of
Gadjin M. Nagao. Honolulu: University of Hawaii Press, pp. lx, 420.
Kalupahana, D. J. (1987), The Principles of Buddhist Psychology. Albany: State University of New York
Press.
Kapstein, M. (2001), Reason’s Traces: Identity and Interpretation in Indian & Tibetan Buddhist
Thought, Studies in Indian and Tibetan Buddhism. Boston: Wisdom Publications.
Kochumuttom, T. A. (1982), A Buddhist Doctrine of Experience: A New Translation and Interpretation
of the Works of Vasubandhu, the Yogacarin. Delhi: Motilal Banarsidass.
La Vallée Poussin, L. de (1928), Vijnaptimatratasiddhi. Paris: P. Geuthner.
La Vallée Poussin, L. de (1988), Abhidharmakośabhāṣyam. Translated by Leo Pruden. Berkeley, CA:
Asian Humanities Press.
Lamotte, É. (1973), La Somme Du Grand Véhicule d’Asaṅga (Mahāyānasaṁgraha). Louvain-La-Neuve,
Belgium: Université de Louvain, Institut Orientaliste.
Lévi, S. (1925), Vijñāptimātratāsiddhi. Deux Traités De Vasubandhu: Vimśatika Et Trimśika. 1 v. Paris:
H. Champion.
Lusthaus, D. (2002), Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Buddhism
and the Ch’eng Wei-Shih Lun. New York: RoutledgeCurzon.
Matilal, B. K. (1974), “A critique of Buddhist idealism,” in A. Kunst, L. Cousins, and K. R. Norman
(eds.), Buddhist Studies in Honour of I. B. Horner. Dordrecht and Boston: D. Reidel.
Pradhan, P. (ed.) (1975), Abhidharmakośabhāṣyam of Vasubandhu. rev. 2nd ed. Patna: K. P. Jayaswal
Research Center.
Schmithausen, L. (1987), Ālayavijñāna: On the Origin and the Early Development of a Central Concept
of Yogācāra Philosophy. Tokyo: International Institute for Buddhist Studies.
Schmithausen, L. (2005), On the Problem of the External World in the Ch’eng Wei Shih Lun, Studia
Philologica Buddhica. Occasional Paper Series, 13. Tokyo: International Institute for Buddhist
Studies of the International College for Advanced Buddhist Studies.
Siderits, M. (2007), Buddhism as Philosophy: An Introduction. Indianapolis: Hackett.
Suzuki, D. T. (1930), Studies in the Lankavatara Sutra, One of the Most Important Texts of Mahayana
Buddhism, in Which Almost All Its Principal Tenets Are Presented, Including the Teaching of Zen.
London: G. Routledge.
Tzohar, R. (2013), “Vasubandhu,” in Richard Payne (ed.), “Buddhism,” Oxford Bibliographies. New
York: Oxford University Press, April, http://www.oxfordbibliographies.com/obo/page/buddhism.
Tzohar, R. Published online: September 15, 2016. DOI 10.1007/s11841-016-0544-y. “Imagine being a
Preta: Early Indian Yogācāra approaches to intersubjectivity.” Sophia International Journal of
Philosophy and Traditions.
Tzohar, R. (2017), “Does early Yogācāra have a theory of meaning? Sthiramati’s arguments on
metaphor in the Triṃśikā-Bhāṣya.” Journal of Indian Philosophy 45, no. 1, 99–120.
Ueda, Y. (1967), “Two main streams of thought in Yogācāra idealism.” Philosophy East and West, 17,
155–165.
Waldron, W. S. (2003), The Buddhist Unconscious: The Ālaya-Vijñāna in the Context of Indian Buddhist
Thought. London and New York: RoutledgeCurzon.
Walpola, R. (2000), Abhidharmasamuccaya: The Compendium of the Higher Teaching. Trans. Sara
Boin-Webb. Fremont, CA: Asian Humanities Press.
Wayman, A. (1965), “The Yogācāra idealism.” Philosophy East and West, 15, no. 1, 65–73.
Willis, J. D. (1979), On Knowing Reality: The Tattvārtha Chapter of Asaṅga’s Bodhisattvabhūmi. New
York: Columbia University Press.
Wood, T. E. (1991), Mind Only: A Philosophical and Doctrinal Analysis of the Vijñānavāda. Honolulu:
University of Hawaii Press.
Yamabe, N. (2004), “Theories of consciousness,” in Robert E. Buswell Jr. (ed.), Encyclopedia of
Buddhism. New York: Macmillan, pp. 174–178.
Yamaguchi, S., and Lévi, S. (1934), Madhyantavibhagatika: Exposition Systématique Du
Yogacaravijñaptivada. Nagoya: Librairie Hajinkaku.
 14
Luminous Mind: Self-Luminosity versus Other-Luminosity in
Indian Philosophy of Mind
Matthew MacKenzie

Luminosity
The metaphor of consciousness as light (prakāśa) or luminosity (prakāśatā) is at the heart
of Indian thinking about the nature of the mind going back at least as far as the early
prose Upaniṣads. In a well-known dialogue in the Bṛhadāraṇyaka Upaniṣad between King
Janaka and the sage Yājñavalkya, the King asks, “Yājñavalkya, what is the source of light
for a person here?” In response, Yājñavalkya mentions the external sources of
illumination, such as the sun, moon, and fire, as well as the illumination provided by a
voice in the darkness. The King then asks,

“But when both the sun and the moon have set, the fire has died out, and the voice is stilled,
Yājñavalkya, what then is the source of light for a person here?”
“The self (ātman) is then his source of light [svayaṃjyoti]. It is by the light of the self that a
person sits down, goes about, does his work, and returns.”
“Which self is that?”
“It is this person—the one that consists of perception among the vital functions (prāṇa), the one
that is the inner light within the heart.” (BAU 4.3.1–8; Olivelle, 1998, p. 111)

Like a light, consciousness has (or is) the capacity to shine forth (prakāśate) and
illuminate (prakāśayati) its object. Indeed, just as, without illumination, no objects could
be visible, without the light of consciousness, no object could be experienced. Thus
luminosity comes to denote the capacity to disclose, present, or make manifest. Physical
light, of course, can reveal or make visible objects, but, as later Indian philosophers
pointed out, it can only do so to a perceiver. The luminosity of consciousness, however, is
that original capacity to make experientially present some object to some subject. Yet this
inner light that makes possible all experience and knowledge is itself quite elusive. Earlier
in the text, Yājñavalkya remarks, “You can’t see the seer who does the seeing; you can’t
hear the hearer who does the hearing; you can’t think of the thinker who does the
thinking; and you can’t perceive the perceiver who does the perceiving. The self within all
is this self of yours. All else besides this is grief!” (BAU 3.5.2; Olivelle, 1998, p. 83). On
this view—widely, but by no means universally held in Indian thought—the conscious
subject (or consciousness itself) is not knowable in the same way as its objects. Despite its
association with the elusiveness of the subject, however, luminosity also comes to be
associated with the distinctive flavors (rasa) or qualitative features of conscious
experience—that is, with the phenomenality of consciousness (Ram-Prasad, 2007, p. 54).
By the classical period, luminosity comes to denote the distinctive mark or feature of
consciousness as that which reveals or discloses (to a subject), particularly in the context
of distinct episodes of conscious cognition. As Chakravarthi Ram-Prasad (Ram-Prasad,
2007, p. 54) characterizes it:

Luminosity is the rendering of an event as subjective. It is that by which there is an occurrence,

which it is like something to undergo. The subjective is the having of the experience (anubhava).
Luminosity is the Indian metaphor for phenomenality, the undergoing by the subject of something
else (its object). The philosophers are agreed on all sides that consciousness is phenomenological;
it is luminous. The debate is over the constitution of the phenomenality of consciousness. The
debate is about what it is for there to be subjectivity.

While we should not simply identify phenomenality and subjectivity, it is important to note
here that the question of luminosity constitutively involves questions of intentionality,
phenomenality, and subjectivity, as well as their interrelations.
The basic divide in Indian accounts of the luminosity of consciousness is between other-
illumination (paraprakāśa) and self-illumination (svaprakāśa) theories. For advocates of
other-illumination, the luminosity of consciousness consists in its capacity to present a
distinct object. Thus, transitive, object-directed intentionality is the mark of
consciousness. Conscious states, in order to be states the subject is conscious of, must be
presented by a distinct, higher-order cognition. Hence, consciousness illuminates that
which is other than itself, and conscious states themselves are apperceived by another
state. In contrast, for advocates of self-illumination theories, the luminosity of
consciousness consists in its being reflexive or self-presenting. Consciousness presents
itself in the process of presenting its object. Moreover, just as light does not need a
second light in order to be revealed, so consciousness does not need a distinct state to
present itself—it is self-intimating.
In what follows I will discuss some of these different accounts of luminosity developed
and defended in the Indian context, as well as some relevant connections to contemporary
issues in the philosophy of consciousness.

Other-luminosity
As mentioned above, the basic divide with which we are concerned here is between other-
luminosity (paraprakāśa) and self-luminosity (svaprakāśa) accounts of consciousness.
However, these are broad classifications that cover a variety of distinct and often
incompatible views, each of which is developed within the context of a particular,
systematic philosophical viewpoint (darśana). So, for instance, while the Nyāya, Bhaṭṭa
Mīmāṃsā, and Madhyamaka schools have defended other-luminosity, the specific views
defended are often quite different and form only a part of dramatically different
integrated philosophies of consciousness. For my purposes in this chapter, I will
concentrate on the Nyāya tradition’s account of other-luminosity as a paradigm case,
though I will also discuss some relevant features of the Bhaṭṭa Mīmāṃsā view.
The Nyāya school defends a resolutely realist, intentionalist, and first-order view of
consciousness. It is realist in three relevant respects. First, Nyāya is fully ontologically
realist about consciousness. Consciousness is a real, unique, and irreducible capacity of
the self.1 Second, Nyāya is epistemologically realist in that veridical episodes of conscious
cognition disclose real, mind-independent objects and properties in the word. Third, the
school defends a direct realist account of perception in that veridical episodes of
perception reveal mind-independent physical objects, without the intermediary of
conscious mental images or representations (ākāra). Nyāya realism is, in turn, tightly
bound up with their intentionalist analysis of consciousness. It is the very nature of
consciousness to be of or about an object (svābhāvika viṣayapravaṇatvam). Episodes of
conscious cognition are individuated by their object, and the intentional object of
cognition is (in the standard case) an object in the world, not any kind of mental
intermediary. Thus, intentionality is the mark of the mental and the intrinsic intentionality
of consciousness is understood in externalist terms. Moreover, since consciousness is
transparent (nirākāra, “without image/aspect”), the phenomenal features of a conscious
state are grounded in (and perhaps reducible to) the way the object is presented through
the state. Finally, Nyāya defended a first-order view of consciousness. The luminosity of a
cognition consists in its revealing its object (arthaprakāśobuddhi), not itself. That is, it
illuminates that which is other than itself (its object) and requires a distinct, second-order
cognition for a first-order cognition to be revealed.
On the Nyāya view, then, when one has a perception of a tree, one is in a conscious
state with the content “there is a tree.” The first-order cognition derives its content from
the worldly object it discloses and it makes no reference to either the cognition itself or its
subject. The term for a first-order cognition in this context is vyavasāya. It presents the
tree in a particular way (in terms of color, shape, and so on), but this is ultimately a
function of the properties of the tree. But in what sense is this cognition conscious? It is
conscious in that it makes its subject aware of the object. When one has a perception of a
tree, that perception makes one conscious of the tree. On this view, one need not be
aware of the cognition in any way in order for it to make one conscious of its object. More
formally, we can call this the independence condition: the presence of a conscious mental
state does not depend on the subject’s being aware of that state (or aware of being in it).
As Fred Dretske (1995, p. 100) remarks,
 There are, to be sure, states in (or of) us without which we would not be conscious of trees and
pianos. We call these states experiences. Since these experiences make us conscious of things ...
the states themselves can be described as conscious. But we must be careful not to conclude from
this that because the states are conscious, we must, perforce, be conscious of them.

The Nyāya school is in complete agreement with this point.

How, then, do we come to be aware of our conscious states, when we are? According to
Nyāya, our first-order cognitions (vyavasāya) are cognized by a subsequent, higher-order
cognition (anuvyavasāya) or apperception. In the typical case of perception, a first-order
cognition triggers a second-order cognition that takes the first-order cognition as its
intentional object. There are several features of this view worth noting. First, this is
thought to be an automatic causal process. It is causal in that, in minds like ours, the
occurrence of a first-order conscious perception triggers (causes) an apperceptive meta-
cognition. This implies, on the Nyāya view of causality, that first- and second-order
cognitions (qua vehicles of content) are distinct, because causation occurs between
distinct existences. Second, anuvyavasāya as automatic meta-cognition should be
distinguished from voluntary reflective introspection. While we do typically apperceive
our basic perceptions, we do not usually go around introspecting our cognitions. Yet,
insofar as the relation between vyavasāya and anuvyavasāya is causal, the Nyāya school
holds that this process can be blocked or simply fail to occur. For instance, when one is
fully absorbed in an object, one might have no meta-cognition of the perceptual cognition,
or when one hears a loud noise, the sudden shift in attention might disrupt the formation
of a meta-cognition directed at a prior visual perception.2 Third, while the vyavasāya and
the anuvyavasāya are distinct cognitions, their content is intimately related. When a
perception with the content “there is a tree” is apperceived, the content of the
apperception will be “I perceive a tree.” The direct object of the meta-cognition is the
first-order cognition, while the indirect objects of the meta-cognition are the first-order
cognition’s object (the tree) and the self. Hence, while the meta-cognition is about another
cognition, it retains reference to the first-order cognition’s object of perception. In
addition, the meta-cognition involves self-reference in that it indirectly refers to the self as
the subject of the state. It does not, however, refer to itself, as all cognitions reveal only
that which is other than the cognition itself.
As I pointed out above, Nyāya is centrally concerned with articulating and defending a
robustly realist account of consciousness and cognition. With this in mind, we can see why
the various elements of their other-luminosity theory are tightly connected. As
intrinsically intentional, consciousness is essentially world-directed. Indeed, on their view,
a conscious cognition is, in a sense, exhausted by its revealing its object. It exists only as
this directedness at the world. However, it is also no less than the unique capacity to
make the self aware of the world. So while Nyāya (like most Indian schools) gives a causal
account of episodes of cognition, it is an account of the causes and conditions of the
occurrence of an intrinsically intentional (luminous) state of the self, irreducible to any
physical or otherwise non-intentional quality. Further, if the nature of intentionality is to
reveal transcendent objects—that is, objects that are independent of the cognitive state
itself—then it would seem to follow that consciousness is transparent or formless.
Anything revealed by a conscious state, on this account of intentionality, is ipso facto
other than that state. As simply the illuminating of an object, consciousness itself has no
form of its own.3 For Nyāya, then, because consciousness is essentially saviṣayatā (with-
an-object, i.e., intentional), it must also be nirākāra (formless or without aspect).
This line of thinking is strikingly similar to one developed by Jean-Paul Sartre. In the
introduction to Being and Nothingness, he writes, “All consciousness, Husserl has shown,
is consciousness of something. This means that there is no consciousness that is not a
positing of a transcendent object, or if you prefer, that consciousness has no ‘content’”
(Sartre, 1956, p. 11). For Sartre here the paradigm case of “content” would be something
like an immanent mental image that serves as an intermediary between mind and world,
such as the “ideas” of the British empiricists. For both Nyāya and Sartre, the theory of
intentionality as pure world-directedness leads to a rejection of these intermediaries (the
ākāra in the Indian context) and therefore to the view that consciousness is radically
transparent or formless. That a cognition is formless in this sense does not preclude its
having a contentful intentional structure (viṣayatā). Rather, the point is that this form of
intentionality is direct and presentational as opposed to the indirect and re-presentational
account given by proponents of sākāravada (with-aspect-ism).4 For Nyāya, then, a proper
understanding of the luminosity of consciousness entails the rejection of its supposed self-
luminosity, because, if a cognition only reveals a transcendent object, it cannot reveal
itself and it must be revealed by another cognition.5 We may call this the intentionality
argument against self-luminosity.
In summary, we can discern three interwoven strands in the Nyāya case for other-
luminosity.6 There are causal considerations in support of other-luminosity. The
Naiyāyikas are asatkāryavādins concerning causality, meaning that they take cause and
effect to be ontologically distinct. A sensory-perceptual cognition arises from an
appropriate causal link between an object in the cognizer’s vicinity, the sense organs, and
the self. A cognition, being caused by its object, must be distinct from its object. In the
case of apperception, a meta-cognition arises when there is an appropriate causal link
between the prior cognition, the inner sense (manas), and the self. So again, there must
be two distinct cognitions. Call this the causal argument for other-luminosity. There are
also broadly phenomenological considerations in support of the view. That is, Nyāya
points out ways in which perception and apperception can diverge in our own experience,
such as sudden shifts of attention. Call this the phenomenological argument. Finally, as
we have seen in the intentionality argument, the Nyāya analysis of the nature of cognition
takes it to be pure directedness at a distinct object. These three strands make for a
powerful and integrated approach to the nature of consciousness.
The Bhaṭṭa Mīmāṃsā school also defended an other-luminosity account of
consciousness, but one that was in some ways quite different from Nyāya. Bhaṭṭa
Mīmāṃsā, like Nyāya, holds that cognitions are other-luminous in that their function is to
reveal a distinct object. However, Nyāya holds a first-order view of consciousness in that
an object-directed first-order cognition is sufficient for the subject to be in a fully
conscious cognitive state. That is, a first-order cognition yields a personal-level conscious
state,7 such as seeing a tree. In contrast, for Bhaṭṭa Mīmāṃsā there is a two-stage process
leading to full-blown person-level conscious states. In the first stage, one has a perceptual
state that presents an independent object. In the second stage, one has a meta-cognition
that makes available on the person-level the object of the first cognition. Hence, we might
think about this meta-cognition as yielding a form of access-consciousness of the contents
of the first-order state. In this way the Bhaṭṭa Mīmāṃsāview is similar to higher-order
representation (HOR) views of consciousness. However, it is important to note that, unlike
standard HOR views, Bhaṭṭa Mīmāṃsā holds that the first-order cognition is luminous and
therefore both intentional and phenomenal. Hence, the luminosity of the first-order state
is not explained by the occurrence of the higher-order state. Rather, the higher-order
state explains the subject’s conscious access to the contents of the first-order state—the
incorporation of that content into the person-level stream of consciousness. Moreover, as
distinct states, one could have a first-order perception of a tree in the absence of a higher-
order cognition, yielding a case of phenomenal consciousness without access-
consciousness. Here again we see a difference between Nyāya and Bhaṭṭa Mīmāṃsā views
of other-luminosity. For Nyāya, a first-order perceptual state is a full-blown conscious
state in the subject’s mental life and is available for belief-formation, attention, and
action-guidance. For Bhaṭṭa Mīmāṃsā, however, the first-order state in the absence of
meta-cognition will not be fully integrated with other functions and would therefore be
more like a subliminal or very marginal perception. This puts their view somewhere
between pure first-order and higher-order views of consciousness.
Perhaps the most unique feature of the Bhaṭṭa Mīmāṃsā view is its account of the
nature of the meta-cognition. On their account, the meta-cognition does not take the first-
order cognition as its object. In fact, first-order cognitions never become direct objects of
cognition in the causal-perceptual sense. Rather, the intentional object of the meta-
cognition is the object of the first-order cognition. When one has a first-order perceptual
cognition of a tree, the meta-cognition’s object is also the tree, not the perceptual
cognition. But if the meta-cognition also takes the tree as its object, what is its function
over and above mere perception of the tree? For the Bhaṭṭas, the meta-cognition serves to
ascribe the awareness of the perceptual object to the subject, thereby incorporating it
fully into the subject’s explicit mental life. Note, then, that we have here a distinction
between the conscious intentionality (luminosity) of a state and its subjectivity, its for-me-
ness. It seems that the role of the meta-cognition here is to bring the intentional state into
the domain of full-blown subjective experience. Again, this is similar to higher-order views
of consciousness. But standard higher-order views still take the meta-cognition to be
about another cognition, whereas the Bhaṭṭas hold that it too is about a worldly object.
The basic idea here is that the meta-cognition grasps its object as cognized (jñātatā)
and having the quality of being cognized presupposes a cognition of the object. Thus the
subject becomes aware of the object directly (non-inferentially) and the first-order
cognition indirectly by way of analytic presupposition (arthāpatti) or by inference
(anumiti) (Ram-Prasad, 2007, pp. 62–63). Thus Bhaṭṭa rejects the inner perception model
of self-awareness in favor of an indirect model. The meta-cognition reveals that one has
cognized the object, but is not directly of the first-order cognition. Now, in the standard
model of inference in Indian philosophy, inferential warrant is based on universal
concomitance (vyāpti), for example, “where there is smoke, there is fire.” In this case,
however, a problem arises because the cognition known on the basis of the cognizedness
of the object is never itself perceived. What then could establish the existence of the
cognition? According to Kumārila, the move from the cognized object to the cognition is
based on analytic presupposition. The property of “cognizedness”is definitionally or
conceptually related to the occurrence of cognition. Thus, the cognizedness of the object
already presupposes the cognition. Other Bhaṭṭa thinkers (and their critics) take the move
to be a form of abductive inference, wherein one infers from the observable to the
unobservable. In either case, the upshot is that, when one sees a tree, one does not see
one’s seeing, but one comes to know that she is seeing by way of the “seenness” of the
tree itself. Therefore, self-knowledge of this kind is parasitic on knowledge of the world.
Yet what is this property of “cognizedness”? It has been compared to the “cookedness”
of rice, though I am not sure how helpful is that analogy. More importantly, it is explained
in terms of the object’s availability for speech and action. When I have a perceptual
cognition of a pot, the pot is experientially available for reference (“there’s a pot”) and
actions such as grabbing the pot. Recently, in his contemporary appropriation of the
Bhaṭṭa model, Mark Siderits (2011) has suggested that cognizedness be understood in
terms of global availability. That is, when an object is available for sensation, speech,
action, memory, belief-formation, and so on, it has the higher-order property of
cognizedness. On this view, a first-order cognition has both phenomenal character and
intentionality, but it is not yet fully subjective or fully conscious. It does, in some minimal
sense, make its subject aware of its object. It also triggers a meta-cognition that directly
takes the object as qualified by cognizedness (global availability). This meta-cognition also
indirectly takes it that the subject is cognizing the object. The result is a cognition of the
form “I am cognizing the object.” In contemporary terms the Bhaṭṭa Mīmāṃsā view
appears to be a unique form of a non-reductive higher-order thought account of
consciousness.

Self-luminosity
In the Pramāṇavārttika, the Buddhist philosopher Dharmakīrti remarks that, “the mind is
by nature luminous cognizance (prabhāsvara)” (PV 2.210cd-211ab; Dunne, 2004, p. 372).
Just as light reveals those objects upon which it falls, consciousness has the unique
capacity to make experientially present those objects to which it is directed. It is that by
which there can be any phenomenal appearance. As the capacity for experiential
manifestation, luminosity, as we have seen, is linked to the phenomenality of
consciousness. Consciousness has no color, shape, size, weight, and so on, yet it is the
condition of the possibility of experiencing these qualities. Indeed, when one searches for
consciousness itself as an object or content, one is unable to find it.8 The cognizance of
consciousness is its capacity to grasp or apprehend its object. This aspect, then, is linked
to the intentionality of consciousness. Conscious states not only present phenomena, but
also are intentionally directed to objects (viṣaya). Further, consciousness involves not just
the capacity to reveal or phenomenally manifest, but also to identify, reidentify, and
understand its objects.
According to the Buddhist proponents of self-luminosity, the luminosity of
consciousness consists in its being reflexive or self-presenting. Consciousness presents
itself in the process of presenting its object. In some Buddhist schools, the term
“svasaṃvedana” (self-awareness) denotes this self-luminosity or pre-reflective self-
awareness that is an invariant aspect of conscious experience. On this view, individual
conscious states simultaneously disclose both the object of consciousness and (aspects of)
the conscious state itself. Thus, when a subject is aware of an object, she is also (pre-
reflectively) aware of her own experiencing. Buddhist philosophers such as Dignāga,
Dharmakīrti, and Śāntarakṣita defended the idea that consciousness is reflexive or self-
presenting in this way.
According to Dignāga, “Every cognition is produced with a twofold appearance, namely
that of itself (svābhāsa) and that of the object (viṣayābhāsa)” (PS(V) 1.9a; Hattori, 1968, p.
28). The object-appearance or object-aspect is the presentation of the intentional object in
cognition. It is what the experience is as of. Whatever the further status of the intentional
object, insofar as it is given in experience, there is an object-appearance. There are then
further questions such as the relation between the object-aspect and the object that is the
cause of the cognition. Dignāga argues for a causal theory of experience whereby the
intentional object of an experience (its object-aspect) is logically independent of its cause
and wherein the experience can be as of an object that does not in fact exist, as in, for
instance, a vivid hallucination of a pink elephant in the room when there is no pink
elephant in the room. Further, the identity of the cognition is partly constituted by its
object-aspect and so it can be individuated in terms of its intrinsic intentionality. That is, a
cognitive episode’s being as of a tree is intrinsic to it, whether or not there is a cognition-
independent tree. Yet a cognition is not exhausted by its presentation of an intentional
object. It also presents a subject-aspect (svābhāsa), which for Dignāga means the way the
cognition presents itself. When I have an experience as of a tree, on this view, the
experience presents both the tree (the object-aspect) and the experiencing of the tree (the
subject-aspect). And since I grasp both the object-appearance and the self-appearance of
the cognition in which the object is presented, the dual-aspect structure of cognition
implies pre-reflective self-awareness (svasaṃvedana). Importantly, the viṣayābhāsa,
svābhāsa, and svasaṃvedana are features of a single episode.9 Hence, the pre-reflective
self-awareness here is not a distinct higher-order cognition, but rather an intrinsic feature
of the first-order cognition itself. This means that Dignāga’s view can be classified as a
same-order theory of self-awareness.
Dignāga’s dual-aspect view of cognition, then, is similar to that of Colin McGinn. On
McGinn’s view, experiences are “Janus-faced.” He writes:

Subjective aspects of experience involve reference to the subject undergoing the experience—this
is what their subjectivity consists in. But we can also say that ... experiences have a world-directed
aspect: they present the world in a certain way, say as containing a scarlet sphere against a blue
background. This is their representational content, what states of affairs they are as of. Thus ...
experiences are Janus-faced: they point outward to the external world but they also present a
subjective face to their subject; they are of something other than the subject and they are like
something for the subject. (McGinn, 1991, p. 29, emphasis in the original)

The world-directed aspect or objective face here corresponds closely to Dignāga’s

viṣayābhāsa, while the subjective face corresponds to the svābhāsa. Note further that this
involves a rejection of the transparency of experience. For Dignāga, as for McGinn, there
is more to experience than how the object is presented. The phenomenal character of
experience also involves how the experience itself is presented. Thus we should make a
phenomenological distinction between what the object is like and what it is like to cognize
the object. For instance, what it is like to see a bright yellow lemon may be different from
what it is like to imagine the lemon or to remember it, even if the lemon itself is presented
in the same way (as bright yellow, etc.) in each cognition. Jonardon Ganeri (2012, p. 170)
characterizes the subjective-aspect simply as, “whatever it is in virtue of which attending
to one’s experience does not collapse into attending to the world as presented in
experience.”
Indeed, Dignāga’s main argument for the dual-aspect view (dvyābhāsatā) is that if
cognitions are transparent, there would be no distinction between a cognition and the
cognition of that cognition (PS 1.11; Kellner, 2010, p. 210). That is, if a cognition of a tree
(C1) has no form other than its object and a meta-cognition of C1 (C2) is also transparent,
then, while C2 has as its object C1, C1 has no form other than the form of the tree.
Therefore, C2 will collapse into just another cognition of the tree. However, if C1 has two
faces, then we can make sense of what C2 grasps when it cognizes C1, namely, the
subject-aspect of C1. C2 will also grasp C1’s object in grasping C1, but it will not collapse
into it. Call this the meta-cognition argument.
Dignāga does not say very much about the nature of the subject-aspect, but he does say
that we are aware of the various “mental factors” (caitta) that are built into our cognitive
episodes (PS(V) 1.6ab; Kellner, 2010, p. 207). In much Buddhist philosophy of mind,
cognitive episodes are analyzed according to a citta-caitta model. On this model, the
primary cognition is the citta (consciousness or conscious event) and this basic event can
be characterized in terms of a variety of mental factors (the caitta). One common
taxonomy lists fifty-one distinct mental factors, divided into six groups. Within this
taxonomy, however, there are five omnipresent factors: feeling (vedanā), recognition
(saṃjñā), intention (cetanā), attention (manasikāra), and contact (sparśa). Each episode of
conscious cognition involves some permutation of these five factors. In addition to being
luminous and cognizant, an episode of cognition may be pleasant or unpleasant, focused
or unfocused, calm or agitated, and so on (Dreyfus, 2011, p. 119). These mental factors
contribute to the global phenomenal character of the episode and are available for
reflection by the subject—that is, they are pre-reflective aspects of the experience. These
features of the experiencing, rather than the object experienced, may be part of what
Dignāga terms the svābhāsa, the “self-appearance” or subject-aspect of the cognition.
On this interpretation of Dignāga’s reflexivist theory of consciousness, the object- and
subject-aspects of experiences are phenomenal modes of presentation. The object-aspect
presents the object (e.g., as being red and spherical), while the subjective-aspect presents
the experience itself (e.g., as being a pleasant, focused, visual experience, as well as being
a cognition of a red sphere). They are phenomenal modes of presentation in that the
object and the experience are presented qualitatively. In modern parlance, we can say
that there is something it is like to be aware of a red sphere and there is also something it
is like to live through an involuntary, attentive, pleasant, visual experience of a red
sphere.10
Note that, in the above McGinn passage he remarks that the subjective face is
presented to the subject. Likewise, in Dignāga, the dual aspects of experience are
appearances (ābhāsa) and, arguably, it is a feature of the conceptual grammar of
“appearance” that it is always “appearance-to.” Indeed, insofar as our discussion of
luminosity has taken it to be a form of phenomenal presentation, the very idea of
consciousness as presentation seems to presuppose presentation-to. This fits well with
views that accept the self (ātman) as the enduring subject of experience, but seems to
present a problem for Buddhist philosophers for whom anātmavāda (the doctrine of no-
self) is foundational. In short, if both the objective and subjective aspects are
presentations, to whom or to what are they presenting?
For Buddhist reflexivists like Dignāga, the answer is svasaṃvedana (self-awareness,
svasaṃvitti). Here “self-awareness” or “reflexive awareness” denotes the primitive, direct
acquaintance one has with one’s own experience. To be presented an object in experience
is to be aware of the object as it is given in and through that experience, whether or not
one thematizes the object as experienced. To experience an object is also to live through
the experiencing directly. In both cases, we may say, there is something it is like to have
the experience and that is a function of how the experience presents the object and how it
presents itself. Reflexive awareness, then, is the direct awareness one has of that which is
presented (objectively or subjectively) in experience. It is, in other words, the basic
awareness of the objective and subjective faces of each cognitive episode.
The Buddhist reflexivists, then, are committed to something like what David Rosenthal
(1997) has called the Transitivity Principle, according to which a subject is in a conscious
state M only if the subject is, in some suitable way, aware of M (or of being in M). The
subject need not be reflectively or attentively aware of M, but must only be aware of it “in
some suitable way.” The Transitivity Principle does not entail reflexivity, because one
might, like Rosenthal, hold the higher-order representationalist (HOR) view that M is
represented by a distinct, second-order state. It does, however, entail a rejection of the
independence condition maintained by Nyāya. What is distinctive about the reflexivist
view is that self-awareness is not a distinct higher-order state, but an intrinsic feature of
consciousness. It is a same-order view of pre-reflective self-awareness.
Dignāga’s central argument for self-luminosity has come to be known as the memory
argument. Dignāga states that one cannot remember what one has not experienced
before. But we can remember our own previous experiences—that is, we can remember
not just the object of a previous experience, but also the previous experience itself. Thus
the experience must itself have been experienced. If the prior cognition is cognized by a
distinct cognition, as in the higher-order view, then there would occur an infinite regress.
This is because, according to Dignāga, the higher-order experience too can be
remembered and so must itself have been experienced, and so on. To avoid the regress, he
argues, we must hold that experiences are reflexive—that is, that the awareness of the
experience is not separate from the experience itself.
For example, in order for me to genuinely remember a particular sunset, I must have
actually seen it—“memory” here is a success term. On Dignāga’s view, I can also
remember my experiencing the sunset. I can remember the vividness of the perceiving,
the way my attention was grabbed by the explosion of color and shape, the mix of joy and
melancholy that accompanied seeing the sunset, and so on.11 But if I can only remember
that of which I have previously been aware, then I must have been (however pre-
reflectively or inattentively) aware of my perceiving the sunset. I have access through
memory to both the object-appearance (the beautiful red sunset) and the self-appearance
(what it was like for me to be perceiving the sunset on that occasion). That access
presupposes awareness of both aspects of the experience. Note, then, that Dignāga is
arguing for something similar to the transitivity principle of consciousness, but with
regard to memory. A subject can only remember a state M if the subject was in some
suitable way aware of M. But if my awareness of seeing the sunset is a separate state M*
and I can also remember my awareness of my seeing of the sunset then M* would need a
further state M**, and so on. Dignāga concludes that reflexivism wins by eliminating the
alternative. However, it is open to the anti-reflexivist to respond that the regress is
blocked by making M* unavailable to memory. Perhaps, following HOR theories of
consciousness, M* is an unconscious cognition of M. Dignāga that does not address this
response.
The upshot of the meta-cognition argument and the memory argument, for Dignāga, is
that experiential episodes are complex events12 involving an object-appearance, a subject-
appearance, and the basic awareness of these appearances, where these features are
inseparable aspects of experience. Now, in one respect, the role of svasaṃvedana is
epistemic. Indeed it comes to be treated by Buddhist epistemologists as the most basic
and secure means of knowledge (pramāṇa). The idea here is that we have a direct (i.e.,
immediate, non-inferential) and perhaps even infallible acquaintance with the phenomenal
contents of our experience. To have a conscious pain is to be aware of the qualitative pain
directly, just by having it. Moreover, even if there is no scarlet sphere in my immediate
environment, I am still directly aware that I am having an experience as of a scarlet
sphere. On this view, then, there is no phenomenal presentation—no presentation of
either the subjective or objective face of an experience—without the basic awareness of
those faces. Thus Dharmakīrti argues, “The seeing of objects is not established for one
whose apprehension thereof is itself imperceptible” (PV 1.54). That is, if one is not at all
aware of the experience in and through which the object is presented, then the object is
not phenomenally present at all. Furthermore, note that the svābhāsa and viṣayābhāsa are
given to or given within a conscious, first-person point of view. For the Buddhist
reflexivists, svasaṃvedana constitutes the conscious point of view within which the two
faces of cognition are given. Reflexivity, therefore, constitutes a minimal form of
subjectivity in the phenomenological sense of a dative of manifestation that to which the
phenomenally present is presented. Yet, crucially, this point of view is not a distinct or
enduring subject of experience existing over and above the interconnected episodes of
experience constituting individual streams of consciousness (cittasantāna). Rather, as we
have seen, reflexive awareness is a basic feature of each individual experiential episode.
At bottom, each episode of experience is its own subject.
In addition to the epistemic role of svasaṃvedana, we also find it playing a
transcendental role—that is, self-luminosity comes to be seen as the distinguishing mark
(svalakṣaṇa) or very nature of consciousness. In Madhyamakālaṃkāra 16, Śāntarakṣita
(2005, p. 53) argues:
Consciousness rises as the contrary
Of matter, gross, inanimate.
By nature, mind is immaterial
And it is self-aware.
On this view, matter is inherently inanimate and insentient (jaḍa), while consciousness is
inherently luminous and cognizant—that is, reflexive and intentional. There is nothing it is
like to be a stone, and it has no states that are intentionally directed toward an object. In
contrast, dynamic sentience is the very mode of being of consciousness, and for
Śāntarakṣita, the sentience or phenomenality of consciousness is understood in terms of
its reflexivity. As his Tibetan commentator, Jamgon Mipham, remarks in this context:

Objects like pots, being material, are devoid of clarity [luminosity] and awareness [cognizance].
For them to be cognized, it is necessary to rely on something that is quite different from them,
namely, the luminous and knowing mind. The nature of consciousness, on the other hand, is unlike
matter. For it to be known, it depends on no condition other than itself... . In the very instant that
consciousness arises, the factors of clarity and knowing are present to it. Although other things are
known by it, it is not itself known by something else and is never without self-awareness (it is never
“self-unaware”). (Śāntarakṣita, 2005, p. 202)

The distinction drawn here is similar to Searle’s distinction between those things with a
first-person ontology and those with a third-person ontology. On Searle’s (2002, p. 60)
view, “consciousness has a first-person ontology; that is, it only exists as experienced by
some human or animal, and therefore, it cannot be reduced to something that has a third-
person ontology, something that exists independently of experiences.” Objects like pots do
not have experiences and apparently exist independently of their being experienced.
Conscious states, however, do not exist independently of being experienced—their very
mode of being is to be experienced
Mipham goes on to argue, following the Indian Buddhist reflexivists, that:

It is thanks to reflexive awareness that, conventionally, phenomenal appearances are established

as the mind, and the mind [i.e., a cognitive episode] is in turn undeniably established as the object-
experiencer. If reflexive awareness is not accepted, the mind would be disconnected from its own
experience of phenomena and the experience of “outer objects” would be impossible.
(Śāntarakṣita. 2005, p. 123)

That is, for these thinkers, the experiential object is recognized to be a phenomenal
appearance (ābhāsa) or representation (ākāra) that is not distinct from the cognition
within which it is presented. In other words, one sees that the supposed external object is
in fact merely the objective face of an experience and thus an aspect of the experience
itself. Further, on this reflexivist view, absence of pre-reflective self-awareness would yield
a kind of mind-blindness (“the mind disconnected from its own experiences”) wherein at
any given time one might be having any number of phenomenal experiences without any
awareness that one was having them, in absence of which their intentional objects would
not be phenomenally present. In such cases, one’s cognition would be more like blindsight
than phenomenal consciousness. That is, if conscious states have a first-person ontology—
they only exist insofar as they are experienced or undergone—then they presuppose a
subjective or first-person point of view within which they appear.13 This basic conscious
point of view is reflexive awareness, in the absence of which one would have no access to
one’s own states and their contents. Hence, on this view, reflexivity or luminosity
constitutes the necessary condition of any phenomenal appearance, subjective or
objective. In this sense, reflexivity or self-luminosity comes to be seen as transcendental.
Whereas the Buddhist epistemologists recognize both reflexivity and intentionality as
fundamental aspects of consciousness, the Advaita Vedānta school takes only reflexivity to
be the essence of consciousness.14 Consciousness (cit) in its fundamental nature is pure
reflexive subjectivity. What we normally think of as the intentionality of consciousness
itself actually arises from the association of pure non-intentional consciousness with
certain non-cognitive mental states (vṛtti). As Ram-Prasad (2007, p. 80) characterizes it:

This Advaitic conception of consciousness as essentially reflexive in fact is tantamount to saying

that it is purely reflexive. Indeed, this is the idea behind the conception of consciousness as
“witness” (sākṣin ...). Just as onlookers do not engage in the events they are witnessing, so
witnessing-consciousness does not engage with objects. It is present, but it is transparent to
content, not itself intentionally directed towards (i.e. “engaged with”) objects.

Moreover, it is important to note that, for the Advaitin, the ātman is not a substantial self
such as was defended by the Naiyāyikas, but rather, the self is pure reflexive
transcendental subjectivity. That is, the ātman is not an individuated, enduring entity. It is
the witnessing subjectivity that can never be objectified. Thus, in contrast to the Buddhist
theory of reflexive awareness, for the Advaitin it makes no sense to say that individual
mental states or events are self-luminous. Conscious mental states are immediately
present, not needing an additional second-order mental state to reveal them. They are,
therefore, self-presenting in a derivative sense in that they need no further cognitions to
reveal them. They are not, however, present to themselves, as in the Buddhist view.
Rather, conscious states are immediately present to the self as pure witnessing
subjectivity. The ātman, as pure consciousness, is the self-luminous source of illumination
for any phenomenon whatsoever, “internal” or “external,” and cannot itself become an
object of cognition. As the condition of the possibility of any presentation of an object,
consciousness is not one object among others, yet it is indubitably present. So, while the
Buddhist view of luminosity focuses on the internal structure of individual, empirical
cognitive events, the Advaita account of luminosity focuses on a transcendental notion of
subjectivity in which the subject is distinct from any empirical entity.15
The Advaita philosopher Citsukha, for instance, carefully defines self-luminous as that
which is immediately evident (aparokṣa), but not an object of knowledge (Ram-Prasad,
2007, p. 75). It is immediately evident in the sense that we cannot be mistaken about
whether we are conscious. The earlier Advaitin, Śrī Harṣa, unpacks this familiar Cartesian
point in terms of three absences (p. 75). First, there is the absence of doubt over whether
one is having an occurrent conscious cognition. Second is the absence of a meta-cognition
that one has no first-order cognition, when one does have a first-order cognition. Third is
the absence of a meta-cognition that one does not have a first-order cognition when one
has no first-order cognition. Whatever we make of Śrī Harṣa’s view, the basic point is
straightforward: we have an immediate and indubitable acquaintance with at least the
existence of our occurrent consciousness. For the Advaitin, we can be mistaken about the
contents of consciousness in a variety of ways, but the fact that we cannot be mistaken
about whether we are conscious at all just supports the distinction between consciousness
itself (cit) and its various contents (viṣaya) and modifications (vṛtti). Citsukha is keen to
maintain that consciousness is never its own object, because it is never an object at all.
On this view, an object is that which is revealed by consciousness, and that which is
revealed distinct from that which reveals. Indeed, on this account, what it is to be an
object is to be presented as distinct from the experience of it. In this way it is similar to
the phenomenological notion of an object as a Gegenstand—that which stands against
one’s awareness of it. If consciousness were to know itself as an object, it would thereby
falsify itself by occluding the revealing of that object in the knowing itself. Hence, cit is
self-conscious without positing itself as an object. As the source of all revealing, then,
consciousness is pure unobjectifiable subjectivity.
On the one hand, consciousness cannot be known as an object, while, on the other
hand, its occurrence cannot coherently be doubted. If consciousness is known as an
object, they argue, there would be an infinite regress, owing to the distinction between
awareness and objects of awareness. If it is denied that consciousness is self-evident, then
one would, it seems, be forced into the performative self-contradiction of denying that one
has experience. In fact, Śaṅkara argues that because consciousness is the condition of
experiencing any object of knowledge, its existence cannot be overturned—that is, it
cannot be denied on the basis of any future experience, since any future experience will
itself presuppose consciousness. Any object of consciousness can be doubted, but
consciousness itself cannot.
So, whereas the Yogācārin view of self-luminosity plays a central role in an account of
experience that unifies reflexive awareness and the subjective and objective aspects of
experience, the Advaitins sharply distinguish the self as pure reflexive consciousness from
all objects and modifications of experience. Moreover, while the Buddhists hold that
episodes of experience are impermanent occurrences within an ever-changing stream of
consciousness (cittasantāna), the Advaitins maintain that self-luminous consciousness is
changeless and permanent. As the changeless background of all changing experience, this
witness-consciousness is taken to account for the diachronic unity of experience.
According to Advaita, the Buddhist view of experience as made up of causally connected
moments of experience—self-luminous or not—cannot account for either the diachronic
cognition of objects or of the stream of consciousness itself.16 Śaṅkara argues:

The perception of similarity takes the form of “This is like that.” “That” refers to the remembrance
of something seen: “this” to the perception of something present. If after remembering the past
experience denoted by “that,” consciousness should linger till the present moment referred to by
“this,” then the doctrine of momentariness would be gone. If, however, the remembrance
terminates with the notion of “that,” and a different perception relating to the present (arises and)
dies with the notion of “this,” then no perception expressed by, “This is like that,” will result, as
there will be no single consciousness perceiving more than one thing. Moreover, it will be
impossible to describe our experiences. Since consciousness ceases to be just after seeing what
 was to be seen, we cannot use such expressions as, “I see this,” or “I saw that,” for the person who
has seen them will not exist till the moment of making these utterances. (BAU-B 4.3.7; Deutsch and
Dalvi, 2004, p. 138)

The Buddhist is faced with a dilemma. Because perception of similarity (or identity)
requires comparison between an earlier and a later perception, either there is a single
enduring consciousness that has both perceptions and the doctrine of momentariness
(kṣaṇikavāda) is false, or there is no enduring consciousness and each perceptual event is
locked in the solipsism of the present moment. In the latter case, no perception of
similarity is possible. Third, Śaṅkara argues that if consciousness is momentary, then
there can be no diachronic continuity of the first-person perspective. Moreover, note that
Śaṅkara sees very clearly the deep phenomenological connection between the experience
of persisting objects and the experience of oneself as a persisting subject. Thus, on
Śaṅkara’s view, one must either accept an enduring self or consciousness, or be faced
with an experientially disconnected series of mental events.17
The standard Buddhist response to this objection is to appeal to the causal and
functional connections between mental events within a single stream of consciousness. Of
course, the mere fact that one event causes another (even when those events are mental)
does not entail that the two events are experientially continuous. In order for the Buddhist
response to be plausible, the causal-functional connections must ground phenomenal
continuity. So phenomenally continuous mental events must not only be causally
connected, they must both be part of the same phenomenal point of view. In the Advaita
account this point of view is provided by witness-consciousness as an enduring subject of
experience over and above the various mental events (vṛttis) that constitute the stream of
experience. However, Buddhist philosophers will appeal to a continuous, but not
enduringly self-identical, point of view as a feature of the connection mental events. For
Buddhist reflexivists, svasaṃvedana can be pressed into service here. First, reflexive
awareness constitutes the synchronic phenomenal point of view in that it is that feature of
awareness to which the subject-aspect and object-aspect are present. Second, as Dignāga
argues, reflexive awareness plays a central role in the diachronic relations of access-
consciousness in memory. That is, svasaṃvedana allows for the apprehension of both
aspects of experience by a later experience and this later apprehension is from the inside
(Ganeri, 2012). Hence, on this view, it is the self-luminosity of consciousness that
constitutes the synchronic and diachronic phenomenal point of view, not an enduring
subject.18 The Advaitin’s mistake, it might be said, is to slide from a continuous point of
view within a stream to an enduring subject above or behind the stream (MacKenzie,
2012).

Conclusion
By way of conclusion, it will be useful to draw some further connections between the
Indian debates around the nature of luminosity and some contemporary issues in
philosophy of mind. The first and most obvious connection involves the long-running
Western debates over the connection between consciousness and self-consciousness. This
debate goes at least as far back as Aristotle and continues up through the present
(Caston, 2002). In the classical phenomenological tradition (including Brentano, 1995) the
orthodox view is that consciousness constitutively involves pre-reflective self-awareness
(Zahavi, 2006). Further, the connections between phenomenological and Indian accounts
are the subject of a thriving current literature (MacKenzie, 2007; Thompson et al., 2011).
In particular, the connection between reflexive awareness, subjectivity, and the self are
ripe for cross-traditional exploration. Similarly, in analytic philosophy debates between
first-order representationalist, higher-order, and self-representationalist views of
consciousness present a fruitful area of connection (Kriegel and Williford, 2006). Indeed,
the self-representationalism of Kriegel (2009) bares remarkable resemblance to that of
Dignāga, while the first-order representationalism of Dretske is in some important
respects similar to that of Nyāya. Also worth exploration would be the connections
between Bhaṭṭa views and the dispositional higher-order views of Carruthers (2000) and
others. Further areas of connection include debates over the nature of phenomenal
consciousness, the transparency of experience, the relation between self-awareness and
intentionality, the temporality of experience, the senses of ownership and agency, personal
identity, externalism and self-knowledge, and bodily self-awareness.
 Despite the antiquity of some of the Indian views and thinkers discussed above, I
believe they offer important insights and resources for contemporary thinking. Not only
do we find interesting—and in some cases, remarkably contemporary—views and
arguments among classical Indian accounts of the luminosity of consciousness, we find
well-integrated models, covering a range of important aspects of the mind and its
operations. These models are worthy of considerations as part of the global history of
philosophy, but also, in some cases, as live options for contemporary philosophy of mind.
Moreover, cross-traditional philosophy of mind, it seems to me, is especially valuable
insofar as it allows us to engage with models and traditions that may have quite different
background assumptions and commitments, and deals with a different or perhaps wider
range of experiences, than our current discourse.

Abbreviations
PS(V) Pramāṇasamuccayavṛtti of Dignāga. See Hattori (1968) and Kellner (2010).
PV Pramāṇavārttikam of Dharmakīrti. See Dharmakīrti (1960).
BAU Bṛhadāraṇyaka Upaniṣad. See Olivelle (1998).
BAU-B Bṛhadāraṇyakopaniṣadbhāṣya of Śaṅkara. See Deutsch and Dalvi (2004).

Notes
1. This capacity is in the Nyāya ontology technically a quality (guṇa), as distinct from a substance or
action.
2. It is important to note that while anuvyavasāya is a function of the inner sense (manas), which is
also responsible for voluntary selective attention, apperception here is not a case of voluntary
attention.
3. The formlessness of consciousness here refers to its lack of an immanent object or intermediating
image (ākāra), not its general lack of properties or nature.
4. The sākāravadins held the view that we are aware of external objects, not directly, but by way of a
mental image or representation of that object in consciousness.
5. Sartre himself does not draw this conclusion and endorses the idea that consciousness is self-
luminous.
6. There are, of course, other strands in the Nyāya view, particularly having to do with epistemology.
7. By “person-level” here I mean, roughly, a fully, attentively conscious mental state, as opposed to
either a subliminal or merely background state, such as inattentively hearing the hum of a
refrigerator.
8. This is one of the phenomenological uses of śūnyatā (emptiness).
9. Note that the relation between svābhāsa and svasaṃvedana is controversial. On some
interpretations of Dignāga, self-awareness just is the subject-aspect’s apprehension of the object-
aspect. On my view, self-awareness is the apprehension of both faces. Yet, insofar as the subject-
aspect is the presentation of the cognition as that very cognition itself, there will be an intimate
connection between the svābhāsa and svasaṃvedana.
10. For an interpretation of the subject-aspect that does not appeal to phenomenal character, see
Ganeri (2012). Though, I must admit that, on my notion of phenomenal character, anything that
presents a “subjective face” to a first-person point of view is, by definition, an aspect of
phenomenal character.
11. These are, of course, my examples, not Dignāga’s.
12. The events are not mereologically complex, but rather aspectually complex.
13. This point is linked to the transitivity principle and is therefore consistent with either a reflexivist
or a higher-order view.
14. Actually, things are a bit more complicated here. Because the Yogācārins hold that, ultimately,
there are no objects apart from consciousness, it can be said that reflexivity is more fundamental
than intentionality. The point here, though, is that for the Yogācārin, both grasper and grasped are
part of consciousness, whereas for the Advaitin, consciousness only appears to be intentional.
15. The ātman or witness here is the dative, the “for whom,” of any phenomenal manifestation.
16. The following draws from MacKenzie (2012).
17. As Śaṅkara argues later in the same passage, mere causal connection between mental events is
not sufficient to give experiential continuity.
18. Note here that for the Buddhist reflexivist, the phenomenal point of view is parasitic on the
moments of consciousness that form the causal continuum, rather than belonging to an enduring
subject—like a series of beads arranged so that the hole in each bead is aligned with others to
form an opening through the whole series.

References
Brentano, F. (1995), Psychology from an Empirical Standpoint. Translated by A. Rancurello, D. Terrell,
and L. McAlister. London: Routledge.
Carruthers, P. (2000), Phenomenal Consciousness: A Naturalistic Theory. Cambridge: Cambridge
University Press.
Caston, V. (2002), “Aristotle on Consciousness.” Mind, 111 (444), 751–815.
Deutsch, E., and Dalvi, R. (2004), The Essential Vedānta: A New Source Book of Advaita Vedānta.
Bloomington: World Wisdom.
Dharmakīrti (1960), The Pramāṇavārttikam of Dharmakīrti: the first chapter with the Autocommentary,
trans. R Gnoli, Serie Orientale Roma 23, Istituto Italiano per Il Medio ed Estremo Oriente, Rome.
Dretske, F. (1995), Naturalizing the Mind. Cambridge: MIT Press/Bradford Books.
Dreyfus, G. (2011), “Self and subjectivity: a middle way approach,” in M. Siderits, E. Thompson, and D.
Zahavi (eds.), Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions.
Oxford: Oxford University Press, pp. 114–156.
Dunne, J. (2004), Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom Publications.
Ganeri, J. (2012), The Self: Naturalism, Consciousness, and the First-Person Stance. Oxford: Oxford
University Press.
Hattori, M. (1968), Dignaga, on Perception. Cambridge: Harvard University Press.
Kellner, B. (2010), “Self-awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti: A close
reading.” Journal of Indian Philosophy, 38 (3), 203–231.
Kriegel, U. (2009), Subjective Consciousness: A Self-Representational Theory. Oxford: Oxford
University Press.
Kriegel, U., and Williford, K. (eds.) (2006), Self-Representational Approaches to Consciousness.
Cambridge: MIT Press/Bradford Books.
MacKenzie, M. (2007), “The illumination of consciousness: approaches to self-awareness in the Indian
and Western traditions.” Philosophy East and West, 57 (1), 40–62.
MacKenzie, M. (2012), “Luminosity, subjectivity, and temporality: an examination of Buddhist and
Advaita views of consciousness,” in I. Kuznetsova, J. Ganeri, and C. Ram-Prasad (eds.), Hindu and
Buddhist Ideas in Dialogue: Self and No Self. Farnham: Ashgate, pp. 181–198.
McGinn, C. (1991), The Problem of Consciousness. Oxford: Blackwell.
Olivelle, P. (1998), The Early Upaniṣads: Annotated Text and Translation. Oxford: Oxford University
Press.
Ram-Prasad, C. (2007), Indian Philosophy and the Consequences of Knowledge. London: Ashgate.
Rosenthal, D. (1997), “A theory of consciousness,” in N. Block, O. Flanagan, and G. Güzeldere (eds.),
The Nature of Consciousness: Philosophical Debates. Cambridge: MIT Press.
Śāntarakṣita (2005), The Adornment of the Middle Way: Śāntarakṣita’s Madhyamakālaṃkāra with
Commentary by Jamgon Mipham, trans. Padmakara Translation Group, Shambhala Publications,
Boston.
Sartre, J.-P. (1956), Being and Nothingness. Trans. Hazel Barnes. New York: Philosophical Library.
Searle, J. (2002), “Why I am not a property dualist: Journal of Consciousness Studies, 9 (12), 57–64.
Siderits, M. (2011), “Buddhas as zombies: a Buddhist reduction of subjectivity,” in M. Siderits, E.
Thompson, and D. Zahavi, Self, No Self? Perspectives from Analytical, Phenomenological, and Indian
Traditions. Oxford: Oxford University Press, pp. 308–331.
Thompson, E., Siderits, M., and Zahavi, D (eds.) (2011), Self, No Self? Perspectives from Analytical,
Phenomenological, and Indian Traditions. Oxford: Oxford University Press.
Zahavi, D. (2006), Subjectivity and Selfhood: Investigating the First-Person Perspective. Cambridge:
MIT Press/Bradford Books.
 15
Solving Kant’s Problem: K. C. Bhattacharyya on Self-Knowledge*
Jay L. Garfield

Kant’s problem as K. C. Bhattacharyya sees it

Krishnachandra Bhattacharyya’s (1875–1949) most important systematic work, The
Subject as Freedom (1923), is first and foremost a sustained engagement, from the
standpoint of Vedānta, with Kant’s discussion of self-knowledge in the Critique of Pure
Reason. In the Critique, Kant argues that while we can think the transcendental subject—
and indeed necessarily must think it as a condition of the possibility of subjectivity itself—
we cannot know the subject, or self. Because knowledge requires intuition, and the forms
of intuition are spatiotemporal, and because the self lies outside of space and time as their
transcendental condition, Kant argues, the self lies outside of the domain of knowledge. It
cannot fall under any category; it cannot be schematized; it cannot be the object of any
judgment. Nonetheless, he argues, it must be possible for the “I think” to accompany any
representation, and so we must think ourselves as unitary subjects in order for any
experience to count as the experience of a subject.
While Kant is one of the most important influences on Bhattacharyya’s thought, this
central doctrine of the Kantian critical philosophy is anathema to him. From the
standpoint of any of the major Indian traditions, including prominently the Vedānta and
Vaishnava tantric traditions that form the backdrop of Bhattacharyya’s thought, Kant gets
things completely backwards. From the perspective of Vedānta, knowledge of the self is
the very goal of philosophical and spiritual practice, and the self, being that with which
we are most intimately involved, must be knowable, if indeed anything is truly knowable—
since anything that is known as object must be known in relation to the self. On the other
hand, given that the self is never object, but only subject, and given that thought is always
objective—that is, directed upon an object—the self, from the standpoint of this tradition,
cannot be thought.
So, there is broad agreement between the Kantian and the Vedānta perspectives that
the self is a kind of epistemic singularity: it is the transcendental condition of discursive
thought yet cannot be the object of discursive thought. This is the deep affinity that leads
Bhattacharyya to explore the points of contact between the Kantian and the Vedānta
frameworks. Nonetheless, there is a sharp disagreement about the nature of this
singularity: while Kant sees the self as in the domain of thought, but not in the domain of
knowledge, Vedānta sees it as falling within the domain of knowledge but not within the
domain of thought.1
So much for a tension between two traditions. But why does Bhattacharyya defend the
Vedānta side of this dispute? I believe that this is primarily because he sees a deep
tension in Kantian philosophy that can only, on his view, be resolved from the perspective
of Vedānta2: Bhattacharyya sees the Kantian view as committed to a series of claims about
the self that undermine its own commitment to the self’s unknowability.
The first of these is the obvious claim that it is unknowable. To assert this is to assert
something about it, and to know that it is unknowable is to know something about it.3 But
more importantly, Bhattacharyya takes seriously Kant’s own association of transcendental
subjectivity and freedom, especially as that doctrine is developed in the second and third
Critiques and in the Foundations of the Metaphysics of Morals, but also as it emerges in
the doctrine of the spontaneity of thought in the first Critique. Indeed, this connection is
the focus of The Subject as Freedom. The awareness of our acts—including our act of
thought—as our own, is at the same time the awareness of our freedom as thinkers, as
subjects and as actors. And it is a condition of our subjectivity that we know that these
acts are ours; hence that we know that we are free; hence that we know the self.
This knowledge of the self is not knowledge by acquaintance, but rather a direct
(though as we will see, in an important sense nondiscursive and intuitive) awareness of
the fact that we are selves, a knowledge of who we are, and of our freedom. For these
reasons, Bhattacharyya takes it that on Kant’s own terms, self-knowledge must be
possible. Vedānta, because of the affinities we have just noted to the broader Kantian
perspective, provides the entrée for the explanation of how this is possible. Here is how
Bhattacharyya himself puts the predicament:

11. The metaphysical controversy about the reality of the subject is only about the subject viewed
in some sense as object. The thinnest sense in which it is objectified is “being taken as meant.”
Ordinarily the validity of this degree of objectification of the subject is not questioned, nor
therefore the possibility of a dispute about its reality. If, however, the subject is taken, as
explained, to be what is expressed by the word I as expressing itself, it is not meant or at best
meant as unmeant and is accordingly above metaphysical dispute. There is properly no metaphysic
of the subject, if by metaphysic is understood an enquiry into the reality conceived as meanable.
Even the unknowable thing-in-itself of Spencer and Kant is not taken to be unmeanable. It is at
worst taken to be a problem in meaning. The knowable is meant and the negation of the knowable
is, if not meant, tried to be meant, being not a gratuitous combination of words but a believed
content that is problematically formulated. The subject which is also believed is formulated as I
which is, however, understood as unmeanable though not as a mere word like abracadabra. The
understanding here is not a mystical intuition though it may point to its possibility, nor an intuition
of a meaning that can be a term of a judgment, nor yet the thought of a meaning that is not known
because not intuited or that is known without being intuited. It is somewhere midway between a
mystic intuition and the consciousness of a meaning, being the believing awareness of a speakable
content, the negation of which is unmeaning and which, therefore, is not a meaning. What is
claimed to mystically intuited is speakable only in metaphor which represents a contradiction in
meaning and what is affirmed or denied in metaphysic is a meanable. The subject as I is neither
contradictory nor meanable and the exposition of it accordingly is intermediate between mysticism
and metaphysic. As, however, the subject is communicable by speech without metaphor, it cannot
be taken as falling outside philosophical inquiry. (93)4

Let us pause to unpack a few important ideas that run through this discussion. First, all of
this trades on Bhattacharyya’s distinction between the speakable and the meanable. The
meanable roughly coincides with Kant’s knowable. Whatever can be designated
intersubjectively as an object falls, for Bhattacharyya, under the head of the “meanable.”5
In fact in ¶¶2–3 (87–88), Bhattacharyya explicitly ties meaning to intersubjective
agreement and availability of referents for terms.6 This anticipation of Wittgenstein and
Sellars takes him a bit beyond Kant, of course, but the ideas are nonetheless congruent.
The speakable, on the other hand, is whatever can be spoken of or communicated through
language. It is a broader category than the meanable, since there may be some things we
can communicate—that are not nonsense—even though we cannot assign them
meanings.7 So, we can talk about ourselves, even though there is no term that can mean
the self.
With this distinction in mind, we can return to the dilemma Bhattacharyya poses for the
Kantian view: The subject cannot be taken to be meant, for it is not intersubjectively
available as the referent for I. Nobody but me is aware of my own subjectivity, and so
there is no way to establish a convention of reference or meaning.8 And the first-person
pronoun has a unique role in designating the self. Were I to refer to myself using a name
or a description, in the third person, the possibility of error through misidentification
intrudes.9 But the first-person indexical gets immediately, directly, at the speaking subject,
and is so understood by addressees as well as by the speaker.
So, although the word “I” has no meaning in this strict sense, it is not meaningless. It
conveys something, and is understood; indeed, it is indispensable. It is therefore
speakable, but not meanable. But it is therefore not nonsense, and hence denotes a
possible object of knowledge. But knowledge of what kind? Not discursive, or
“metaphysical” knowledge, for that would suggest that the self is an entity among
entities, an object, and not the subject we wish to know. Nonetheless, it is communicable,
but communicable as a kind of “intuition,” not entirely mystical, but not entirely empirical
either. To answer these questions and to explain the manner in which the self is known is
the goal of Bhattacharyya’s inquiry.
Reading The Subject as Freedom is challenging in part because of the forbidding
density and terseness of the text itself and because of Bhattacharyya’s idiosyncratic and
often opaque prose style. This opacity in part arises from Bhattacharyya’s peculiar
philosophical neologisms. It also emerges from the fact that he is always thinking, even
while writing in English, with Sanskrit senses and contrasts in the background, but never
making these Sanskrit references explicit. But reading this text is also challenging
because Bhattacharyya does not signal the objects of his frequent anaphoric discussions.
It is left to the reader to figure out whether he has Husserl in mind, a particular Indian
school, or whether he is working out his own ideas. Interpretation of this book is hence
always fraught.
My aim here is not to provide a reading of The Subject as Freedom as a whole. That
would require a book-length study. Instead, I simply intend to focus on the structure of
Bhattacharyya’s account of self-knowledge. I will begin with a brief discussion of his
account of the relationship between subjectivity and freedom. I then turn to his hierarchy
of grades of subjectivity, developing the relation between the various levels of bodily
subjectivity, psychic subjectivity, and finally spiritual subjectivity, showing how each
implicates a greater degree of freedom. I will then turn to the account of self-knowledge
scaffolded by that hierarchy, an account according to which self-knowledge is complex
and multileveled. We will then consider how that account of self-knowledge squares with
Bhattacharyya’s view that the subject cannot be thought, before concluding with some
thoughts about the view of freedom that emerges from this discussion and the respect in
which Bhattacharyya takes himself to have solved Kant’s problem.10 My aim is neither to
defend nor to criticize Bhattacharyya’s framework, but rather to articulate it as clearly
and as sympathetically as possible so as to make it available for critical reflection and
consideration by contemporary philosophers.

Subjectivity and freedom

At the end of the first chapter of The Subject as Freedom, Bhattacharyya returns to the
Kantian problem. Here he develops the direct connection between subjectivity and
freedom.

21. The persisting objective attitude of Kant in his first Critique explains not only his admission of
the thing-in-itself and his denial of self-knowledge, but also his disbelief in the possibility of a
spiritual discipline of the theoretic reason through which self-knowledge may be attainable. From
the subjective standpoint, object beyond knownness, this beyond this-ness is, as explained,
meaningless. It may be that, wedded as we are to our body, we cannot get rid of the objective
attitude and the tendency to look beyond the constructed object to the purely given. But not to be
able to deny need not imply admission and though the Kantian disclaimer of idealism as
accomplished knowledge is intelligible, his admission of the unknowable reality appears to be an
unwarrantable surrender to realism... . (100; emphasis in the original)
22. Self-knowledge is denied by Kant: the self cannot be known but can only be thought through
the objective categories ... there being no intuition of it. (101)

This is the summation of Bhattacharyya’s diagnosis of the Kantian predicament. Kant

allows the reality of the self, and indeed its necessity, but denies us any knowledge of it,
including, presumably, the knowledge that it lies beyond knowledge. The “surrender to
realism” is the commitment—incoherent on Kant’s own grounds—to something that is
real, yet in its nature independent of our mode of intuition and knowledge. We will see
that when Bhattacharyya examines the self as an object of knowledge, it will importantly
not be real in this sense, but will turn out to be transcendentally ideal, not given
independent of our modes of subjectivity, but determined by those very modes. In this
sense, as we will see, Bhattacharyya takes himself to be even more of a transcendental
idealist—more relentlessly consistent in this commitment—than Kant himself.
Bhattacharyya continues later in this paragraph:

The subject is thus known by itself, as not meant but speakable and not as either related or
relating to the object. It is, however, believed as relating to object and symbolized as such by the
objective relations. The modes of relating are at the same time the modes of freeing from
objectivity, the forms of the spiritual discipline by which, it may be conceived, the outgoing
reference to the object is turned backwards and the immediate knowledge of the I as content is
realized in an ecstatic intuition. (101)

Self-knowledge, that is, is knowledge of the self as it exists independent of its objects,
even though that must be knowledge of a self that is essentially capable of objective
relations. And this is the first link of subjectivity to freedom. The self must be capable of
being understood simply as a self, free of any relation to a particular object. That
knowledge must be immediate, on pain of turning the self into an object, but can only be
realized through an act of ecstatic transcendence in which subjectivity stands outside of
itself. Bhattacharyya emphasizes this in the next paragraph:

23. Spiritual progress means the realization of the subject as free... . One demand among others—
all being absolute demands—is that the subjective function being essentially the knowing of the
object as distinct from it, this knowing which is only believed and not known as fact has to be
known as fact, as the self-evidencing reality of the subject itself. (101)

The plan of The Subject of Freedom is to develop this self-knowledge gradually, moving
through progressively more abstract and complete levels of freedom, each corresponding
to a more adequate form of self-knowledge. As we will see, complete self-knowledge, while
achieved at the final stage of this hierarchy, comprises all of the stages, and depends on
each sense of freedom to be adumbrated. Here is Bhattacharyya’s outline of the plan:

24... . The steps ... correspond to a gradation of subjective functions, of modes of freedom from the
object. Identified as we are with our body, our freedom from the perceived object is actually
realized only in our bodily consciousness, though even this, as well appear later, is only imperfectly
realized ... The next stage of freedom is suggested by the distinction of the perceived object
including the body from the ghostly object in the form of the image, idea, and meaning, which may
be all designated “presentation.” Consciousness as undissociated from such presentation, but
dissociated from the perceived and felt body, may be called presentational or psychic subjectivity.
The dissociation of the subject of consciousness from this presentation conceived as a kind of
object would be the next stage of freedom, which may be called non-presentational or spiritual
subjectivity. The three broad stages of subjectivity would then be the bodily, the psychical and the
spiritual ... Wedded as we are to our body, actual freedom is felt only in bodily subjectivity and
freedom in the higher stages as suggested by psychology is believed not as what is actual but as
what has to be achieved or realized... . The elaboration of these stages of freedom in spiritual
psychology would suggest the possibility of a consecutive method of realizing the subject as
absolute freedom, of retracting the felt positive freedom towards the object into pure intuition of
the self. (102)

Let us be clear about this plan, as it structures the remainder of the account.
Bhattacharyya identifies three broad stages of subjectivity, each consisting in a distinctive
level of freedom. The first is bodily subjectivity. In being aware of ourselves as bodies in
space, we are aware of our determinate location in relation to other objects, and so our
freedom to consider or to disengage with other objects in space and time. In psychic
subjectivity we are aware of ourselves as mental subjects, whose direct intentional objects
are representations. In this awareness, we recognize our freedom from our bodies and
from our location in space and time, and the fact that we can entertain representations in
the absence of any external object to which they correspond.
In the final level of subjectivity, spiritual subjectivity, we recognize our freedom from
those representations. We come to realize that our existence is not dependent upon our
objects, but they depend upon us. At this point we intuit ourselves as spiritual subjects per
se. We complete this process of self-knowledge, Bhattacharyya intimates, when we adopt
the same cognitive attitude of freedom toward ourselves that we are able to develop in
relation to our objects, an unmeanable sense of ourselves as pure subjects. We will turn
shortly to the account of the successive grades of subjectivity and freedom, but first we
must turn to Bhattacharyya’s general account of introspection.

Interlude: the structure of introspection

Bhattacharyya’s characterization of introspection and its objects is fundamental to his
understanding of self-knowledge, and it is articulated through a rather unusual
vocabulary. He refers to psychological phenomena as psychic facts, and he takes psychic
fact to consist in relations of the subject to its objects. Introspection, then, is a form of
abstraction in which I first become aware of an object of consciousness and then abstract
from the object the way I, as subject, am related to it, distilling the psychological state
that mediates my awareness. Bhattacharyya writes, “What is called psychological
introspection is apparently a process of abstraction from the object of its modes of
relatedness to the subject” (103).
The first thing to note about this account is that for Bhattacharyya introspection is not
a direct sensation of my inner episodes, but rather a theoretical exercise. He hence
rejects the direct givenness of the inner from the outset. He emphasizes (¶26, p. 104) that
this applies to feelings just as much as it does to abstract thoughts. In any case, while
psychic facts or cognitive episodes are in one sense subjective, as modes of relatedness to
objects, they are, in introspective awareness, also objective, and hence are presented
through higher-order psychic facts that take them as objects.
Just as in introspection we achieve a certain degree of subjective freedom from
involvement with the object of the psychological state we consider, in reflection on
introspection we can achieve a similar detachment from that state itself. Freedom hence
emerges at each successive stage of reflective introspection, but is never complete so long
as an objective attitude is maintained toward the state introspected. Bhattacharyya refers
to the pure cognitive state abstracted from its object—the attitude itself—as the fringe of
the psychic fact. This, he emphasizes (¶32, p. 107), never occurs alone, but always in the
context of the relevant psychic fact. So, while I may become aware of believing because
belief is the fringe of my belief that Bengal is verdant, I am never aware of belief, per se,
with no content of belief. Reflection on these fringes, or contentful state-types that are
constitutive of our psychology, is hence always theoretical, never observational, simply
because we never observe the fringes themselves, only the states in which they figure. We
only know the fringes through subsequent reflection on the complex state of which we are
directly aware.
Bhattacharyya draws an interesting corollary from his account of introspection:
introspective awareness, or self-knowledge, is essential to knowledge itself (¶¶35–37, ff.).
This is because knowledge requires the distinction between perception and illusion, which
in turn requires the distinction between believing in the content of a perceptual state and
not believing in it.11 For me to be aware of something as an illusion is for me to be aware
that I have a certain presentation and that I do not believe in the existence of that which
is presented; and knowledge, for Bhattacharyya as well as for Kant, requires the
awareness that we know; and to take myself to know something is to be aware of my
reflective belief in what is presented. This is not a trivial matter: Bhattacharyya is
pointing out that the subject and its relation to its objects cannot be excluded from the
domain of knowledge, as that would be to eviscerate the entire structure of knowledge
itself.
Bhattacharyya takes this to be a serious critique of the Kantian conception of
knowledge. He anticipates the Kantian objection:

43. To such a view the Kantian may be supposed to object that the metaphysical reality thus
adumbrated is only subjective though it appears real in the object by illusion, by a permanent
illusion which we can critically correct without being able to remove. The critical correction may
only be sought to be strengthened in a non-cognitive way—the moral or aesthetic way—with the
entertainment of the metaphysical reality in faith. (113)

That is, the Kantian is taken to reply, the apparent knowledge of the subject that emerges
from the kind of reflection that Bhattacharyya characterizes as introspection is not
genuine knowledge, but transcendental illusion. In knowing an object, we necessarily
have faith in the existence of the subject that considers it, as well as in its modes of
subjectivity, but no knowledge of it, as that knowledge would have to be unmediated
knowledge of the thing as it is in itself (or, we might add, mere psychological knowledge
of the empirical ego).
Bhattacharyya responds as follows:

44. That metaphysical reality is subjective is admitted in the sense that it is not an object behind
the perceivable object... . But this unknown something that is known as such and formulated is not
merely subjective in the sense of being illusory or in the sense of being the content of a faith that
can never be turned into knowledge. To take it as merely subjective is to assume the object to be
alone knowable and to be incapable of being known as put forth subjectively or freely. It is to deny
the facthood of the constructive function by which the perceivable object comes to be for the
subject. The epistemological functions are indeed believed and not known but they are not believed
as merely subjective. (113–114; emphases in the original)

This is the crux of the matter. When Kant excludes the subjective side from the domain of
knowledge, Bhattacharyya argues, he excludes what must be presupposed even as a
ground of the argument for its exclusion, sawing off the metaphysical branch on which the
transcendental philosophy rests. The very fact that we can only know what is subject to
the constructive activity of the mind entails that if knowledge of that is possible,
knowledge of the subject that conditions it must be possible as well, and this in two
respects: first, to know that our knowledge is always conditioned by the subject is to know
something about the subject; and second, to really understand the objects of our
knowledge, to understand their limitations to the conditions of our subjectivity, and to
understand them as our objects is to be aware of ourselves as subjects. Bhattacharyya
sums this up as follows:

51. Thus we meet the Kantian difficulty. Psychic fact ... is object and more than object. It is more in
the sense of being a metaphysical reality constitutive of the object which is its phenomenon, a
reality that is known as unknown and as knowable ... [it] is at once real and realizing, realizing as
being already real, this being the objective counterpart of knowing the object as unknown. To Kant,
metaphysical reality ... is only thought and believed ... We agree that the introspective awareness
of the presentation ... is not knowledge of knowing but only imagination of knowing the
metaphysical. The imagination, however, is not an illusion, but only incomplete or unrealized
knowledge... . Cognitive realization of the metaphysical reality as subjective has to be admitted, at
least, as an alternative spiritual possibility. (115–116; emphasis in the original)

This quest for the cognitive realization of the nature of subjectivity and hence the
understanding of the conditions of subjective freedom is the project of The Subject as
Freedom. Achievement of this knowledge is, for Bhattacharyya, the achievement of
genuine self-knowledge. Having argued for the transcendental necessity of this kind of
self-knowledge for any knowledge whatsoever, and having challenged Kant’s claim that
this kind of self-knowledge must be impossible, Bhattacharyya begins the project of
constructing this knowledge of the subject.

The grades of bodily subjectivity

The first grade of subjectivity is the physical: awareness of and knowledge of oneself as a
body. Bhattacharyya distinguishes three successive moments of this subjectivity, each
involving a distinct aspect of self-knowledge, and each implicating a distinct mode of
freedom. The first of these is the awareness of the body as an external object; the second
is the awareness of the body as a felt immediate object; the third and most abstract, the
awareness of absence. Let us consider each of these in turn. In each case we will consider
each of these three aspects: subjectivity, self-knowledge, and freedom. We will then turn
to psychical and spiritual subjectivity and the modes of self-knowledge and freedom they
each enable.
It is important to note when we consider physical subjectivity that we always identify
ourselves with our bodies, and that part of self-knowledge is knowledge of our own
bodies. We recognize ourselves in the mirror; we recognize and ostend others as bodies.
But more than this, our bodies constitute the perspective from which we are perceptually
engaged with others, the mode under which we act, and the loci of our sensations. They
also provide the spatial reference point from which we experience the world—the here
that makes it the case that I am always here. All of this is involved in Bhattacharyya’s
account of physical self-knowledge, an account that recalls some of Schopenhauer’s
reflections on the body as immediate object in The Fourfold Root of the Principle of
Sufficient Reason and which anticipates important insights of Merleau-Ponty in The
Phenomenology of Perception.
Bhattacharyya begins by emphasizing this centrality of the body to self-experience:

58. The materialistic view that the subject is but the body is true insofar as the body represents a
stage of being of the subject. But it ignores the unique singularity of one’s own body even as a
perceived object. No merely objectivist account can do justice to this singularity. The objectivity of
other perceived objects is constituted by their position relative to the percipient’s body, which
itself, therefore, cannot be taken to be so constituted. To the percipient, the body is an object
situated relatively to some other percipient’s body as imagined, being not perceived by himself in a
space-position though not known, therefore, as non-spatial. The percipient as in his body or as his
body is in this sense, dissociated from the external world, being what his perceived world is
distinct from. At the same time he cannot help imagining himself as included in the world though it
may be as a privileged object. (122–123)

There is a lot going on in this rich paragraph, and we have the resources here to unpack
Bhattacharyya’s account of this first grade of subjectivity and of self-knowledge. First, at a
basic but nonetheless essential level, the subject is the body. When I use the first-person
singular pronoun to refer to my physical incarnation, I am correct. Nonetheless, one way
in which I know my body is to perceive it as an object using external senses including
sight, touch, and even smell and taste. I am hence perceivable, and am hence, as body, a
kind of fusion of subject and object. My senses give me knowledge of my body as object,
but although the mechanism of their doing so is the same as that by means of which they
deliver other objects, they also do so in a way importantly different from that in which
they give me those others: for I lie at the origin of the spatial coordinate system that
structures my knowledge of the external world, and all other bodies are spatially located
relative to my body. The only way that I locate my body in subjective space is by reference
to the imagined gaze of another, as to assign a determinate location (as opposed to a
subjective origin) presupposes another origin for the coordinate system (as Merleau-Ponty
was to argue later).
Self-knowledge at this level of subjectivity is hence in part perceptual, and in part an
immediate knowledge of myself as spatial origin. Without the former, I cannot represent
myself as a physically instantiated subject in a physical world, and so cannot even
represent my own sensory knowledge as mine; without the latter I cannot distinguish
myself as a subject from all else in the world that is object. And at this level of subjectivity
I already distinguish myself as subject precisely by a kind of freedom—in this instance,
the freedom from being simply another object located in the external world, and hence
the freedom to posit the loci of the objects of my Lebenswelt in relation to me, to my body.
The second moment of bodily subjectivity concerns the body not as perceived in
external sense, but as known immediately. This immediate knowledge might at first be
thought to be merely proprioception, but it is more than that. For, as we will see, it is not
merely the immediate apperceptive awareness of the position or sensations of the body,
but the awareness of the body from the inside, as subjective. The account of this
subjectivity, which is the first level at which, Bhattacharyya argues, a genuine sense of
freedom emerges, and at which subjectivity is first experienced as subjectivity, is
complex. Let us work through it with care. Bhattacharyya draws the distinction between
the perceived and the felt body as follows:

60. One’s own body is not only perceived from the outside; one is immediately or sensuously aware
of it also from within in what is called “feeling of the body.” This feeling is not, like the feeling of an
object, a psychic fact from which the object known is distinguished. The bodily feeling is but the
felt body, which is not known to be other than the perceived body. Yet the perceived body is distinct
from it so far as it is an “interior” that is never perceived and cannot be imagined to be perceived
from the outside... . The interior cannot be understood here as the interior that one may imagine
oneself seeing. (123–124)

The first distinction here is the distinction between an awareness in which the object is
distinct from the psychic fact of which it is an object, on the one hand, and feeling, in
which there is no such distinction, on the other. When I perceive any object—say, when I
see my hand—we can distinguish between the act of perception, in this case, perhaps, a
visual perception, and the object, my hand. The former is psychic fact; the latter object.
But when I feel my body as a physical interiority there is no such distinction. There is not
an act of feeling distinct from my being my body. Second, Bhattacharyya emphasizes, this
interiority is not simply a distinct perspective on the same object. The interiority of my felt
body is not an imagined spatial interior that I might see, for instance, in a laparoscope,
but rather a position that can never be imagined to be perceived. It is in this sense, while
physical, purely subjective. (Once again, the anticipation of Merleau-Ponty in this entire
account is striking.)
Bhattacharyya draws this distinction in yet another way, pointing out that the kind of
space represented in the interior of the felt body is different from the kind of space the
perceived body occupies. He puts this in terms of a kind of indefiniteness. The guiding
idea here is that while the interior space of the felt body is not experienced as having
definite dimensions or spatial location, that indefiniteness is not the same kind of
indefiniteness that we might find in an indefinite awareness of the location of a sensation,
such as an itch in our back, or in a hazy awareness of our posture at the end of a long day:

61. Objective space that is indefinitely perceived is the same as the objective space that is
definitely perceived ... But felt space is indefinite in the sense that it is more than the objective
space it is defined into, ... (124)

When my awareness of the space of the perceived body is indefinite, that space will be the
same as that into which it might be resolved in a more definite awareness. So, for
instance, if I can’t say where exactly that itch on my back is, when I locate it with my
finger, I locate it not in another space, but more precisely in the same space in which I
originally only located it with some approximation. But the indefiniteness of spatial
representation in the felt body is not an absence of precision; it cannot be precisified at
all, in fact. Instead, even when we limn perfectly the volume of the interior of the body, we
leave out the interiority of the body, which, while spatially oriented, outruns any attempt
at location.
Bhattacharyya now turns to the implications of these differences for the nature of
subjectivity itself and the freedom it implicates:

64. We may consider body-feeling in relation to psychic fact and introspection into psychic fact on
the one hand and to the perceived body and perceived object on the other. The perceived body is
only potentially dissociated from the perceived object inasmuch as it is not merely like
presentation not denied to be object but is positively known as object... . The object, however, is
fully distinguished from the felt body: the perceived object presents exterior surface only... .
Corresponding to this full distinction from the felt interior, there is the actual but imperfect
dissociation of freedom of the felt body from the perceived environment, The felt body, however,
does not appear even imperfectly dissociated form the perceived body. (125)

The perceived body is, he points out, not all that different from other perceived objects.
While to be sure, it has, as he argued earlier, a subjective dimension, it is also represented
as an object from which, like all other objects, the subjective awareness of it is
dissociated. The felt body is entirely different in this respect. Even though, as
Bhattacharyya notes at the end of this passage, the felt body is in one sense the same
thing as the perceived body, in its mode of presentation as felt, it is entirely distinct from
the object. Perceived objects are only surfaces—they are essentially exterior; the felt body,
as opposed to the perceive body, has no surface—it is essentially interior. Bhattacharyya
now brings this point to bear in order to draw another important distinction, in terms of
psychic fact and identification:

65. Again, the perceived body is fully distinguished from psychic fact ... There may be
consciousness of the body as mine and at the same time as not other than myself, unlike the
consciousness of the object which if felt as mine is felt as not me. The felt body, however, is only
half distinguished from psychic fact, since it is the feeling of the body on the one hand is not
actually dissociated from the perceived body on the other. (126; emphases in the original)

When we perceive objects, including our own bodies, there is, as noted above, a
distinction between object and cognitive act. The object is hence alien to the self, and this
is true even of the perceived body, as perceived. But the felt body is not mine, but me; not
alien, but intimate. For that reason, the felt body is more like a psychic fact than the
object of one; it is hence, unlike the perceived body, on the subject side of the subject-
object duality, not on the objective side. This has important consequences for subjectivity
and freedom:

66. The facthood of the subjective is constituted by the feeling of detachment or freedom. The first
hint of this freedom is reached in the feeling of the body... . When the perceived body is
distinguished from the felt body, the exterior from the interior, we have an explicit feeling of
distinction, detachment or freedom from the perceived object. (127)

While there is indeed, as we saw above, a simple level of freedom in the perceptual
awareness of the body, there can be, Bhattacharyya, argues, no awareness of that freedom
in that perceptual consciousness of body, simply because without the awareness of
interiority, there is no awareness of the distinction between psychic fact and object, and
hence of subjectivity itself. One cannot look down and develop awareness of subjectivity,
and hence of freedom until one reaches the second rung of the ladder; while the first rung
might in part constitute subjectivity, it cannot constitute awareness of that subjectivity.
For that reason, while perceptual awareness is a mode of self-knowledge, it is not a mode
of knowledge of subjectivity or of freedom. Only when we have this feeling of body do we
rise to the level of true self-consciousness, and at that, only at the most basic level. We
climb one step further when we enter the third and final moment of bodily awareness: the
awareness of absence.
Just as it was a few decades later for Sartre, the awareness of absence constitutes an
essential mode of subjectivity for Bhattacharyya. Unlike Sartre, however, he argues that
this mode of subjectivity is an aspect of bodily self-consciousness, and indeed is the most
abstract and profound mode of that consciousness. Let us see how that goes.
Bhattacharyya asks us to consider the awareness of absence. The examples he gives us
are the awareness of the absence of a tree in a field in which the tree once stood, and the
absence of a book we seek in a room where we expected to find it. In each case, a specific
absence becomes the object of our awareness. Now, Bhattacharyya concedes (¶74) that
our awareness of the absence is not entirely perceptual. After all, we see an empty field,
not an absent tree, and we see a space on the table where we expected the book, not an
absent book. There is hence an essentially inferential aspect to this awareness. So, unlike
the perception of the body, or even the feeling of the body in perception of an external
object, here the object of our awareness is not a perceived particular, but rather an
inferred abstraction.
Bhattacharyya characterizes the mode of our awareness of the absent object as
imaginative. Using an example closer to that Sartre was to mobilize in his discussion of
the absence of Pierre in the café, Bhattacharyya writes:

77. Consider the absence of a beloved person ... When such a person is missed or imaginatively
perceived as now absent, there may not be any relevant reference to the locus, namely the room.
But one may come to imagine the room as with the person and then realize his absence in
reference to this imagined content. To imagine an object in a perceived locus is a special form of
imagination in which the present locus I viewed as characterizing and not as characterized by the
imagined content. The belief in the absence of the object as thus characterized by the locus, the
absence here of the imagined room as sentimentally associated with the beloved person, is
immediate knowledge but not perception. The absence is not taken to be fact in the present locus;
and as the presentness of the absence is not the presentness of any concrete thing, it cannot be
said to be perceived. The secondary cognition is conscious non-perception, the room that is
perceived by sense being turned into the imagined character of the location of the imagined
person. (133)

What is going on here, and why is this so important to bodily self-consciousness? First,
note that while Bhattacharyya regards the awareness of absence as in a certain sense
immediate—that is, we are not first aware of seeing something, and then aware of
inferring an absence from it—that is not the immediacy of perception, but rather of an
automatic act of imagination. Sartre sees the empty café, but he is instantly aware of the
absence of Pierre. And he is not thereby perceptually aware of Pierre, but rather
imaginatively aware of the café avec Pierre, while perceptually aware of it sans Pierre,
and at the same time aware that that is mere imagination, or, as Bhattacharyya puts it,
conscious non-perception.
But this conscious non-perception requires more of us than would the actual perception
of Pierre. The latter requires awareness of the object, and so immediately of its relation to
our body in space. To become aware of that awareness, in turn requires attention to our
own bodily interiority—to the fact that our subjectivity is in our body, even though it is not
perceivable, as is our body as it is presented in the most basic mode of perceptual
consciousness. But to become aware of the absence requires us to be immediately aware
of the fact that we are perceiving one thing and imagining another, and hence of the
position of the body with respect not only to that which impinges upon it and to which it is
perceptually related, but also with respect to what we merely imagine. We imagine the
absent object—even though it actually bears no determinate relation to our body—in
relation to our body. The awareness is hence bound up with the body, but free of the
actual nexus of our body with its surrounds. Bhattacharyya puts it this way:

78. In the imaginative perception of absence and the absent, there is no explicitly felt dissociation
from the position of the perceived body, which however is imperfectly distinguished from the
imagined position of absence or of the absent. In conscious non-perception, there is the explicitly
felt dissociation from the perceived body but not from the felt body, though the felt body has begun
to be distinguished from the absence of the absent. The relation of the perceived body in the
former case and the felt body in the latter to the known absence is like the relation of the
perceived body to the felt body. The perceived body is half distinguished from the felt body which,
however, is not felt to be dissociated from the perceived body. Absence imaginatively perceived is
thus on a level with the felt body, both being felt undissociated from the perceived body which
however is half distinguished from them. Absence known by conscious non-perception is on a
higher level. (134)

This paragraph is far from transparent. But once we see what is going on here, we will
see why this form of consciousness is, according to Bhattacharyya, physical, and why it is
so important in the hierarchy of modes of self-consciousness and self-knowledge. First,
Bhattacharyya notes, when we are aware of an absence, we are aware of that as an
absence in a particular locus, and that locus is identified in relation to our body. It is an
absence here or there. But second, we do not relate the absent thing to our perceived
body. After all, the perceived body occupies a particular space, and the absent thing does
not. Instead, we locate it with respect to the felt body; it is not here, in our subjective
space (despite the fact that we do not literally locate the absence spatially, as we do, at
least indefinitely, locate our felt body).
Now, Bhattacharyya calls attention to a strange asymmetry in the relation between the
perceived and the felt body: When we are aware of the perceived body, it is “half
distinguished” from the felt body; that is, it is present as mine, not as me in perception,
even though I identify myself with it in other respects, taking it to be the same as the felt
body. Nonetheless, the same is not the case for my experience of the felt body. I do not
represent it also as mine, and so do not consciously associate it with the perceived body.12
Now, he points out, in the same way, the absence is represented as distinct from the
perceived body—it is represented as an absence in a space outside of the perceived body
—but it is not dissociated from the felt body, for it is not represented as a real concrete
thing, but rather as a cognitive act of imagination carried out by the embodied subject.
For this reason, the awareness of absence is a higher level of consciousness, and
implicates a higher level of self-consciousness, despite remaining tied to an embodied
perspective. Bhattacharyya concludes this discussion with the following observation:

79. Conscious non-perception then is a transitional stage between body-feeling and imagination
with which psychic fact begins. It is the consciousness of presentness without space-position ... It
is free from space but not from the present and accordingly does not imply a presentation of the
object as dissociated from the object. Psychic fact begins with the distinguishing of what the
present is not ... Were one to start with object-perception, ... the first clear hint of the subjective
fact would be realized in the knowledge of absence through conscious non-perception. (135)

The awareness of absence is hence the pivot point in self-knowledge. It allows us to be

aware of an object, but not at a particular place. Nonetheless, by virtue of the temporality
of that awareness, it is an awareness of that object in relation to the physical self. And it is
a direct awareness of the object, albeit as absent, not an awareness of a representation of
that object. So, once again, it is tied to the physical, to embodied reality, the world of
objects that exist in relation to the physical self. Nonetheless, because of the awareness of
the distinction between what is perceived and what is imagined, reflection on this mode of
awareness takes us for the first time beyond the physical into the realm of psychic fact.
For the distinction between perceiving Pierre and imagining Pierre is a psychic, not a
physical distinction. Self-knowledge here then rises to the apperceptive awareness of
myself as a being who perceives in distinct modes. It is on this basis that I can come to be
aware of myself as a mind, and of the distinction between my representations and reality.
It is to that mode of awareness and that level of self-knowledge that we now turn.

Psychic and spiritual subjectivity

Bhattacharyya begins the transition to the discussion of psychic subjectivity—the
subjectivity that takes the self to be a mind, and hence that which makes introspective
self-knowledge possible—with this observation:

80. Psychology does not begin till the perceived object is distinguished from the half-perceived
body... . To those who would not go further in psychology, introspection is only observation of the
indefinite body-interior and psychic fact is only a bodily attitude, the beginning of the behaviour of
an organism to the environment. Some, however, would go one step further and admit the image as
a unique fact, appearing as a quasi-object from which object including the body is distinguished... .
The image may be functional in character as a reference to the object, ... but that it appears
presented as a substantive something from which the object is distinct and exists in a sense in
which the object does not exist cannot be denied. (136)

Here we see the transition from the final mode of bodily consciousness to psychic
consciousness. Bhattacharyya uses the term “image” as Kant does “Vorstellung” or we do
“representation.”13 He is noting that while we can make sense of the activity of
introspection into somatic self-consciousness, we also, upon reflection, recognize the
presence in our psychological life of thought mediated by representations, and hence of
those representations themselves. While representations share with felt somatic states (as
opposed to perceived somatic states) the absence of any determinate spatial location, and
like them are on the subjective, as opposed to the objective side of experience, unlike
somatic states they lack both spatial temporal determinateness. Our beliefs or imaginings
need not be occurrent; they need not have fixed temporal boundaries. And unlike felt
somatic states, he urges (¶¶86 ff), these are not experienced as internal to the body, but
rather to the mind. Introspection into our cognitive activity finds not felt states but rather
thoughts, and these thoughts are all intentional in structure. Throughout this discussion,
Bhattacharyya’s debt to Husserl, and in particular the first volume of Ideas, which
Bhattacharyya read closely, is evident.
Moreover, Bhattacharyya argues (¶¶90 ff), unlike the conscious non-perception of the
absent (the mere awareness of absence) our awareness of our representations, even when
the objects of those representations are absent, is not merely the non-perception of
objects, but an awareness of the actual facticity of the representations themselves. The
ability to dissociate the representation from the object relies upon the final stage of body-
consciousness, but, he argues, the positive awareness of the representation (image) as a
psychic fact represents a new stage in consciousness and in self-knowledge.
A second moment of psychic subjectivity, Bhattacharyya argues, emerges when we
move from the awareness of images, or representations of objects, to ideas. Ideas are non-
imagistic, discursive symbols that do not represent concrete objects. Bhattacharyya’s
principal examples of ideational thought are logical thoughts, and thoughts expressed in
words. Bhattacharyya draws the distinction between the representational and the
ideational in two ways: in terms of their respective vehicles of thought and in terms of
their respective objects of thought.14 The vehicle of imagistic thought is the
representation of an object, and its object is a particular; the vehicle of ideational thought
is the word, and its object is a universal.
Corresponding to each of these moments of subjectivity is a new degree of freedom. In
imagistic thought the subject is conscious of its freedom from the object. Unlike
perception—even “perception” of absence—there is no requirement in representational
thought of the representation of the body, or of the presence in thought of any external
object or space whatsoever. And when we move to ideational thought there is a further
freedom—a freedom from the particular as an object of thought, together with a freedom
from any sensory component of thought whatsoever. The purely symbolic frees thought
from any reference to the concrete at all, even in intentional content. To be conscious of
oneself as a thinking subject is hence to be conscious of oneself as free in a sense far
greater than that involved in thinking of oneself as an embodied subject—it is to represent
one’s cognitive subjectivity as absolutely independent not only of the external world, but
also of the modes of appearance of that world to physical senses.
Reflection on this mode of subjectivity yields yet another level of self-knowledge. Even
at the level of imagistic representational thought, Bhattacharyya claims, introspection
finds not somatic states or feelings, but intentionality. And once the climb has been made
to ideational subjectivity, introspection finds intentionality directed to the abstract and not
the concrete. We come to know ourselves at this level of subjectivity not as conscious
bodies, but as intentionally directed, concept-and-language-wielding thinking things.
Again, this self-knowledge does not replace, but supplements that developed earlier,
layering our self-understanding as we layer our subjectivity.
The final moment of subjectivity for Bhattacharyya is the spiritual. In developing his
account of this kind of subjectivity, the level at which complete freedom emerges, as well
as the most complete self-knowledge, he begins with the concept of feeling. Importantly,
this term must be understood not in the sense of somatic feeling that is in play in the
discussion of the second level of bodily subjectivity, but rather in the sense of aesthetic, as
well as ethical feeling. In approaching spiritual subjectivity in this way, Bhattacharyya is
following not Kant’s path to the third Critique but the Vedānta emphasis on aesthetic
sensibility as the path to the understanding of Brahman, a track he also treads in his
important essay, “The Concept of Rasa.” Bhattacharyya argues in that essay that it is
essential to aesthetic experience not only that we are affected by the aesthetic object, but
that we free ourselves from that affection by contemplating that affection, and so
achieving reflective awareness of ourselves as subjects.
In ethical experience we address one another as subjects in dialogue. In this discussion
at the close of The Subject as Freedom, Bhattacharyya recurs to an important insight he
defends near the beginning of the book: to take oneself as the referent of I is to take
addressees as you, others as he or she. In short, he argues in the first chapter of the book,
the possibility of speech—and hence subjectivity—is conditional upon intersubjectivity,
simply because speech presupposes both addresses and conventions that constitute
meaning. He deploys that insight at the denouement of the discussion to argue that to
understand oneself as a subject is to understand oneself as a member of a class of those
capable of introspective self-awareness:

120. The realization of what a speaker means by the word I is the hearer’s awareness of a possible
introspection. Such awareness is as much knowledge as actual introspection. The speaker calls
himself I and may be understood by the hearer as you. As thus understood, the introspective self is
individual, not an individual being—for introspection is not a subjective being like feeling—but the
function of addressing another self. The speaker does not understand himself through the meaning
of the word I: his introspection is through the word and not through its meaning and is less a self-
knowing than a self-revealing, revealing to a possible understander of the word I. Yet as the
addressing attitude is only implicit, it is to him accidental and posterior to his self-knowing. To the
understanding self, however, although he understands the speaker’s self-knowing because he is
himself self-knowing, his understanding of the other I is primary while his own self-knowing is
accidental and secondary. The speaker knows himself in implicitly revealing to the hearer and the
hearer knows the speaker in implicitly knowing himself... . There are thus two cases—self-intuition
with other-intuition implicit in it and other-intuition with self-intuition implicit in it. Both are actual
knowledge... Because the word I is at once the symbol and the symbolized, it cannot be said to
have simply the symbolizing function... . (161–162; emphases in the original)
121. Actual introspection is implicitly social, being a speaking or addressing or self-evidencing to
another possible introspection or self. (162)

This is dramatic stuff, and it is hard to miss the anticipations of Heidegger, Wittgenstein,
and Sellars, whose respective emphases on the necessarily social nature of self-
consciousness, language, and thought were to transform twentieth-century philosophy of
mind. Let us work through these ideas to come to an understanding of Bhattacharyya’s
view of the highest level of subjectivity and of the kind of self-knowledge and freedom it
suggests.
First, Bhattacharyya notes, the term “I” does not denote an object. It is, in the language
of the first chapter of the text, a term expressing a speakable, but not a meanable. When
we use the first-person pronoun, we signal that we are introspectors—that we are capable
of self-consciousness—but we do not denote that which is the ultimate content of
introspection, for that is subjectivity itself, which, if denoted, becomes object, and not
subject.
Second, in virtue of the role of I as a vocable, but non-denoting term (here note as well
the anticipations of Anscombe on the first person),15 this speaking of the self, and hence
self-consciousness itself, is parasitic on the very possibility of language, and so the
existence of addressees who are also capable of using the first- and second-person
pronouns. So, self-knowledge and therefore also subjectivity are essentially
intersubjective phenomena, not private. There is no knowledge of subjectivity whatsoever
outside of the context of social interaction and discourse.
Spiritual subjectivity, the awareness of oneself as pure subject, capable of action,
reflection, and judgment is then not the awareness of an isolated ego, but the awareness
of a self among selves, and for this reason can rise from the level of mere awareness to
that of knowledge. Bhattacharyya concludes his investigation with this reflection on the
nature of freedom as it emerges from this collective notion of subjectivity:

135. I am never positively conscious of my present individuality, being conscious of it only as that
which is or can be outgrown, only as I feel freeing myself from it and am free to the extent implied
by such as feeling. I do not know myself as free but I conceive that I can be free successively as
body from the perceived object, as presentation from the body, as feeling from presentation and as
introspective function from feeling... . [I]​ may be free even from this distinctness, may be freedom
itself that is de-individualized but not therefore indefinite—absolute freedom that is to be evident.
(171)

Absolute freedom, like absolute subjectivity, Bhattacharyya concludes, is not an object of

immediate awareness, not something of which I am positively conscious as an entity.
Instead, it is something that I know as a potential; the potential to ascend in reflection at
any time through reflection on my identity as a body to reflection on my identity as a
thinker, and finally to reflection on my self as that which can be aware of itself either as
body or as cognitive subject. The cognitive subject is transcendental and, like Kant’s
transcendental subject, is absolutely free in aesthetic or ethical experience. On the other
hand, contra Kant, I can speak intelligibly about it, even if that self about which I speak
remains beyond denotation.
That self is not pure individual, but a social subjective position of which I have
knowledge whenever I speak with others as a person among persons. While the absolute
subjectivity, with its special mode of transcendental access to the self is inspired by that
articulated in Vedānta, the insight that the social turn is necessary for its intelligibility is
Bhattacharyya’s.

The ladder of self-knowledge

Bhattacharyya, in The Subject as Freedom, presents a model of self-knowledge that is far
from simple or immediate. The self he presents is multilayered, and the modes of self-
knowledge corresponding to each layer are importantly different from one another. The
first layer of the self is the physical body, an embodied self that we know in three distinct
ways: at the most basic level, through external sense-perception; at a more abstract level
through our first-person embodiment and the experience of the body from the inside, and
finally through reflection on the relation of the body to that which is not even perceived,
but only imagined. In each case we know as aspect of our embodiment, but also in each
case, an aspect of the self. The second principal layer of the self is the cognitive, and this
we know in two ways. First, we know ourselves as the subjects of representations of the
individual objects we perceive or imagine, in thought that, if not imagistic, is at least
singular. But we also know ourselves as the subjects of abstract thought, like that thought
undertaken as we do philosophy. In this aspect, we know ourselves as users of arbitrary
symbols that can relate us to universals as well as to particulars, of ideas as well as
images.
It is important to note that these are not distinct selves. Bhattacharyya is not arguing
that we are bodies plus minds, but rather that the self is both embodied and cognitive.
The self we encounter at each of these levels of subjectivity and in each of these modes of
self-knowledge is the same self, but manifest in a different way, known through a different
modality.
In each case, the self is the subject, that which can be spoken, but never ostended, and
there is only one such entity (or nonentity). Finally, there is the spiritual layer of the self,
the self as that which can reflect both on its cognitive and physical functions, and so is
free of both, and of their objects, what would be in Vedānta ātman in its manifestation as
sākṣin, the pure witnessing consciousness. This self is known in our use of the first-person
pronoun in its purest sense, and is therefore known only in discourse, and hence only in
interaction with other selves. Dasein, we might say, reflecting on Bhattacharyya’s
anticipation of Heidegger, demands Mitsein. Or, perhaps less grandly, meaning requires a
linguistic community, and thought demands meaning. We can only know ourselves as
things of a kind, not as ontological surds. And self-knowledge is complete only when it
integrates each of these levels of self-understanding as modes of knowledge of the same
self, the referent of I.
It is also instructive to note that at each of these levels of self-knowledge, even though
the self and its subjectivity are in one sense present immediately in introspection, they are
nonetheless always known through some epistemic mediation. Another way to put this is
that Bhattacharyya does not succumb in his account of introspection or self-knowledge to
any crude version of the Myth of the Given according to which we simply appear to
ourselves just as we are. In fact one of the more remarkable aspects of Bhattacharyya’s
entire account of self-knowledge is his avoidance of this epistemological framework,
despite its dominance of the philosophical milieu in which he worked—both Indian and
Western.
In bodily subjectivity, we know ourselves first through the mediation of perception, and
hence subject to the conditions imposed by our sense-faculty, then through an inner self-
awareness, which, though it might appear to be more immediate, is nonetheless mediated
through being non-dissociated with the object of perception as well as the object of the
third level of bodily awareness. That is, even though we may be immediately aware of
ourselves as bodily interiors, to know that body is to know it as the perceived body, and as
the body that stands in determinate relations to the perceived and to the imagined.
Cognitive subjectivity is also conceptually mediated, as it requires not only the
awareness of representations and ideas, but also the awareness of ourselves as the
subjects of, but as distinct from, those representations and ideas, and of that which they
intend. This is obviously a highly conceptually and linguistically enriched and mediated
subjectivity. And most dramatically, even the most rarified form of self-knowledge of all,
spiritual self-knowledge, turns out to be entirely linguistically and socially mediated. This
relentless rejection of self-knowledge as simply the taking of what is given in subjectivity
is what sets Bhattacharyya apart from his contemporaries and antecedents, and may be
the most original aspect of his philosophical program in The Subject as Freedom. Indeed,
it is not until Wittgenstein writes The Philosophical Investigations and Sellars “Empiricism
and the Philosophy of Mind” that Western philosophers really catch up.

Kicking away the ladder: freedom from what? For what?

When we assemble this complex and sophisticated form of self-knowledge, we can see the
shape of Bhattacharyya’s response to Kant. While Kant insisted that we could think, but
could never know the subject, Bhattacharyya shows that we know the subject in a variety
of modalities: perceptual, cognitive-introspective, and reflective; cognitive, ethical, and
aesthetic. Nonetheless, we do not know the subject as an object among objects; that
would be to deny its subjectivity and its transcendental status. Instead, while we know
and can even speak of the subject, we can never directly refer to, or mean it. Rather, we
engage with it as a mode of freedom, and as a mode of our engagement with other
subjects.
Kant argues that the subject is entirely non-spatiotemporal. Bhattacharyya responds
that in virtue of its embodiment it must have spatiotemporal locations and figure in
spatiotemporal relations; nonetheless, as we have seen, that location is not one among
many—not the location of an object (except at the lowest grade of self-knowledge, which
is indeed, important), but location at the origin of the spatiotemporal coordinate system.
Without that location, agency and perception, he argues, would be impossible. While Kant
argues that we cannot know ourselves as thinkers, Bhattacharyya shows that in psychic
subjectivity we know ourselves as minds, and necessarily as minds free to take alternative
attitudes towards our cognitive contents. Without such self-knowledge, we could not think
our own thoughts as ours. And even at the level of spiritual subjectivity, Bhattacharyya
argues contra Kant that although we cannot articulate this as knowledge as knowledge of
a thing, we must be aware of ourselves and be able to speak of ourselves as subjects of
our psychophysical lives.
So, Bhattacharyya goes halfway with Kant. Like Kant, he argues that the subject can
never be object. Like Kant, he argues that it has a special ontological and epistemological
status, and like Kant he believes that we cannot know it in the way that we know objects.
Unlike Kant, however, Bhattacharyya argues that not all knowledge is objective, and he
argues that non-objective self-knowledge and the ability to speak one’s position as subject
are necessary conditions of the possibility of any objective knowledge. Hence, even
though he only goes halfway down one road with Kant, after parting company, he takes a
road to a much more ambitious and complex account of subjectivity and self-knowledge
still well within the transcendental idealist tradition, although clearly inflected with
Husserlian phenomenology, and well on the way to the phenomenology of Heidegger and
Merleau-Ponty.
Bhattacharyya hence makes good on the promise to vindicate a central insight of the
Vedānta tradition—the insight that the self as subject is knowable, and that knowledge of
it is a necessary context for all other knowledge. And he does so both through the
surprising route of a detailed examination of bodily consciousness, undoubtedly inspired
by Vaishnava tantric ideas. But as we have seen, it is not a mere appropriation. There is a
dramatic linguistic and communitarian twist. This transcendent self-knowledge is not
immediate, but is mediated through our linguistic interactions with others.
Knowledge of the self is also knowledge of its freedom, again, a knowledge
transcending the Kantian epistemic bounds. This freedom is a freedom from immediate
involvement with our objects of experience and of thought; our freedom as transcendental
entities of a specific kind who can engage or disengage as we choose from particular
kinds of contemplation. But it is also a freedom to engage; a freedom for aesthetic and
ethical engagement, and the freedom to engage with others in the collective epistemic
activity that makes subjectivity possible in the first place.
Notes

* My reading of K. C. Bhattacharyya’s thought, and of The Subject as Freedom in particular, emerges

from extended conversations with a number of close colleagues, each of whom has contributed a
great deal to my thoughts about these matters. First and foremost, I owe an enormous debt of
gratitude to Nalini Bhushan, with whom I have read and taught this text, and with whom I have
discussed these ideas in detail. Her impact on my thinking is evident throughout. I also owe a
great deal to participants in the Yale-NUS/NUS/Kyoto University joint faculty seminar on Asian
Engagements with Kantian and Post-Kantian European Thought in 2013–14. I especially
acknowledge the contributions of Ben Blumson, Taran Kang, Neil Mehta, Nico Silins, Neil
Sinhababu, Saranindranath Tagore, and Matt Walker. I also thank Ben Blumson, Neil Mehta,
Saranindranath Tagore, and Ryo Tanaka for helpful comments on an earlier draft and Nalini
Bhushan for an invaluable close reading and critique of an earlier draft.
1. See Balslev (2013, pp. 127–136) for a reading of Bhattacharyya as providing a modern Advaita
Vedānta account of the self and consciousness.
2. Through a reading of Vedānta inflected by Bhattacharyya’s reading of Husserl’s Ideas, and a heavy
dose of thinking about the body deriving from tantric traditions.
3. One might try to save Kant here by suggesting that all that he precludes is any knowledge of the
specific nature of the self, or direct perceptual acquaintance with it. But that won’t work. He
denies that we can make any assertions whatsoever about it, in virtue of the inapplicability of the
categories to noumena, and so the inability to deploy the copula. See Priest (2003) for more on
this.
4. All references to The Subject as Freedom are from the edition reprinted in Burch (1975).
5. This distinction is drawn in the first paragraph of The Subject as Freedom:
1. Object is what is meant, including the object of sense-perception and all contents that have
necessary reference to it. Object as the meant is distinguished form the subject or the subjective of
which there is some awareness other than meaning-awareness. The subjective cannot be a
meaningless word: to be distinguished from it, it must be a significant speakable and yet if it be a
meant content, it would be but object. It can thus be neither asserted nor denied to be a meant
content and what cannot be denied need not be assertable. Apparently, the significant speakable is
wider than the meanable: a content to be communicated and understood need not be meant. (87;
emphasis in the original)
6. See Balslev (2013, pp. 131–132) for a similar reading.
7. There is a nice parallel here to the problematic with which Wittgenstein wrestles in the Tractatus,
a book with which Bhattacharyya would not have been familiar.
8. Compare to Wittgenstein’s discussion of discourse about inner states in Philosophical
Investigations.
9. As John Perry (1979) was famously to point out. So, I might erroneously believe myself to be John
Perry. I would then misidentify John Perry as the person thinking this thought. I cannot, however
be wrong about the fact that I am thinking this thought.
10. There is another dimension to Bhattacharyya’s project that we cannot ignore, and that is the
political dimension. As anyone who has read Bhattacharyya’s powerful essay “Svaraj in Ideas”
knows, Bhattacharyya was deeply concerned with the intellectual impact of colonization on Indian
philosophical thought. He worried that the imposition of a European framework on the Indian
academy not only marginalized Indian philosophy, but set the European tradition up as the
subjective standpoint from which philosophical thought itself was to be exercised, with Indian
traditions relegated at best to objects of contemplation from that standpoint. The Subject as
Freedom can be seen as a determined reversal of that direction of gaze. Here the European
tradition is interrogated from an Indian standpoint.
11. It is important to note that when Bhattacharyya uses the term “belief” he nearly always has in
mind the sense of believing in as opposed to believing that.
12. This is but one of many cases in which Bhattacharyya delights in pointing out that prima facie
symmetrical relations are in fact surprisingly asymmetrical.
13. It is likely that Bhattacharyya is thinking of the Sanskrit term “ākāra” here, often translated as
“image,” though more often these days as “representation,” a term that would have the semantic
range he is here attaching to “image.” Balslev (2013, pp. 129, 131) also observes that
Bhattacharyya often has Sanskrit in mind when he writes in English.
14. The distinction Bhattacharyya draws and the ways in which he draws it closely track the
distinction between pratyakṣa/svalakṣaṇa (perception/particular) and anumana/sāmānyalakṣaṇa
(inference/universal) as these are drawn in Nyāya-Vaiśeṣika and Buddhist epistemology.
15. Balslev (2013, pp. 136–137) also notes the anticipation of Anscombe.

References
Anscombe, G. E. M. (1975), “The first person,” in S. Guttenplan (ed.), Mind and Language. Oxford:
Clarendon Press, 1975, pp. 45–65.
Balslev, A. N. (2013), Aham: I, The Enigma of Self-Consciousness. New Delhi: Oxford University Press.
Bhattacharyya, K. C. (1923), The Subject as Freedom. Rpt in G. C. Burch (ed.) (1975), The Search for
the Absolute in Neo-Vedānta. Honolulu: University of Hawai’i Press.
Bhattacharrya, K. C. (1928), “Svaraj in ideas.” Rpt in N. Bhushan and J. Garfield (eds.) (2011), Indian
Philosophy in English from Renaissance to Independence. New York: Oxford University Press, pp.
101–111.
Bhattacharyya, K. C. (1930), “The concept of Rasa.” Rpt in N. Bhushan and J. Garfield (eds.) (2011),
Indian Philosophy in English from Renaissance to Independence. New York: Oxford University Press,
pp. 193–206.
Husserl, E. (2012), Ideas: A General Introduction to Pure Phenomenology. London: Routledge.
Merleau-Ponty, M. (2002), The Phenomenology of Perception. London: Routledge.
Perry. J. (1979), “The Problem of the Essential Indexical.” Nous, 13 (1), 3–21.
Priest, G. (2003), Beyond the Limits of Thought, 2nd ed. New York: Oxford University Press.
Schopenhauer, A. (1999), On the Fourfold Root of the Principle of Sufficient Reason. Trans. E. F. J.
Payne. La Salle: Open Court.
Sellars, W. (1963), “Empiricism and the philosophy of mind,” in W. Sellars, Science, Perception and
Reality. London: Routledge and Kegan Paul.
Wittgenstein, L. (1956), Philosophical Investigations. New York: Macmillan.
 Part Five

Concepts and Cognitions

 16
Nyāya Theory of Concepts
Keya Maitra

This chapter focuses on Nyāya theory of concepts.1 While Nyāya texts seldom address the
topic of concepts directly, it is my contention that the Nyāya system can accommodate a
sustained notion of concepts and clarifying its parameters can help us with a debate in
Nyāya epistemology of perception that has been popular in contemporary discussions of
Indian philosophy. Indeed, a number of participants in this debate do touch on the notion
of concepts in the Nyāya system without offering a thorough treatment of concepts. Thus,
Arindam Chakrabarti (1998, 2000, 2004) in a number of his publications on this topic has
questioned the viability of nirvikalpaka or indeterminate perception based on his claim
that nonconceptual cognition is a mere myth (1998, p. 315). Again, Monima Chadha
(2001) in her discussion of what can be perceived in perceptual experiences argues from
the premise that all cognition is intentional and therefore conceptual. My goal is to
explore the nature of concepts within the Nyāya framework especially keeping in mind the
role concepts play in the Nyāya theory of perception. In general the recent discussion has
focused on the viability of nirvikalpaka perception within the Nyāya system. The general
consensus is that if one maintains that all perceptual cognition is conceptual, then that
excludes the possibility of nirvikalpaka. My aim is to consider how a Nyāya theory of
concept might help us respond to that question. The spirit of my endeavor is inspired by
Chakrabarti’s (2004, p. 366) recommendation for investigations that make “use of the
resources of contemporary analytic philosophers and cognitive scientists along with the
insights and detailed arguments and counterarguments of classical Indian philosophers.” I
will be using a recent theory of concept that has been offered by Ruth Millikan (1998,
2000) as a template for a Nyāya theory of concepts. My working conclusion is that there is
a way to develop a Nyāya theory of concepts which need not require the elimination of
what Chakrabarti calls “immaculate”—nirvikalpaka—perception. In other words, what I
will be suggesting is a way of understanding the notion of concepts that can accommodate
the possibility of nirvikalpaka perception.
While it is true that concepts have been understood in a number of different and
confusing ways within Western philosophy, there are a couple of distinctions that have
become prominent in the literature on concepts both in philosophy and psychology. One
such distinction is between concepts and conceptions and a related distinction is between
epistemological and metaphysical approaches to concepts. According to the first, the
concept of a thing, say either of the cow Bessie or the natural kind cow, is the thing itself,
while a conception of it might be a means that one uses to identify it. Taking the example
of the concept cow,2 while you and I have the same concept, our conceptions can differ
especially if we have different beliefs about cows or use different ways of identifying
cows. This is an important distinction that any theory of concept would want to be able to
make. One of the goals of my project here is to articulate how such a distinction can be
maintained within Nyāya.
A related distinction that has been made in the literature is between the
epistemological and metaphysical accounts of concepts (Smith, 1989). While metaphysical
issues focus on “what is what,” epistemological issues focus on “how we know what’s
what” (Rey, 1985, p. 297). In the context of concepts a metaphysical account considers
“What makes an entity an instance of a particular kind,” and an epistemological account
considers “how an agent decides whether the entity is of a particular kind” (Smith, 1989,
p. 57). Elaborating on why one should care about the metaphysical account of concepts,
Smith (1989, 58) notes that only such an account “provides ‘identity’ conditions for
concepts, i.e., conditions for deciding whether two concepts are the same or different.”
The intuition motivating this distinction draws from the Twin Earth kind of scenarios—
first introduced by Hilary Putnam in his seminal article “The Meaning of ‘Meaning’”—
where it is argued that one’s terms and concepts are causally tied to the stuff one
interacts with in one’s environment and not to something that might be
phenomenologically indistinguishable. This argument would seem especially pertinent
within the Nyāya realist context. I highlight these distinctions not only to draw our
attention to how the recent discussion of Nyāya views of concept has failed to appreciate
this distinction clearly, but also to suggest that this pair can serve as two virtues that a
Nyāya theory of concepts can cultivate.
The rest of the chapter will be divided into three sections. The first section will
introduce the distinction between nirvikalpaka and savikalpaka perception and also recap
some of the highlights of the discussion on the viability of nirvikalpaka perception. I
conclude this section by reviewing how this discussion intersects with a discussion of
concepts. The second section first introduces Millikan’s theory of concepts and then
highlights why this account can serve as a template for the Nyāya theory of concepts. The
third section develops the specifics of a Nyāya theory of concept especially in the context
of its theory of perception. In conclusion I consider a couple of worries.

The nirvikalpaka versus the savikalpaka in the Nyāya and Chakrabarti’s

worries about the nirvikalpaka
The “problem of perception” has been traditionally articulated in terms of the need for a
standard that clearly demarcates veridical perception from illusion and hallucination. In
the Western tradition we have philosophers arguing for some form of the sense data
theory as providing this standard of demarcation. In a loosely similar vein, the classical
Indian philosophical move has been to introduce a distinction between two kinds of
perception: nirvikalpaka (most often translated as indeterminate or pre-predicated or non-
qualificative) and savikalpaka (translated as determinate or predicated or qualificative)
perception. While the nirvikalpaka perception is the perception of an object as pure, and
property-less, the savikalpaka perception is the perception of an object as qualified by
some property or other. Thus, while the nirvikalpaka perception cannot be expressed in
language, savikalpaka perception can be so captured.
While most major schools of classical Indian philosophy incorporate this distinction,
there is much disagreement about their exact nature and respective validity among the
various schools. For example, the Yogācāra-Sautrāntika school of classical Indian
Buddhism, in a fashion similar to most sense data theorists, argues that while nirvikalpaka
perception is always veridical, no such illusion-freeness could belong to savikalpaka
perception. Nyāya and a number of other realist schools, however, would like savikalpaka
perception to be veridical as well. Further, there is also disagreement about what is
perceived in these two perceptions. Thus, for the Buddhists, pure simples (svalakṣaṇas) of
experience are sensed in the nirvikalpaka, while objects as determined by the conceptions
of name, class, relation, attributes, and so on are perceived in the savikalpaka. For the
Nyāya, however, given its direct realism, “any perception, nirvikalpaka or savikalpaka, is a
direct cognition of the real individual which is a unity of the universal and the particular”
(Chatterjee, 1965, p. 193). According to the Nyāya, therefore, the same object is cognized
in the savikalpaka as well as the nirvikalpaka, thereby both perceptions having the same
content. Where they differ precisely is how the object is cognized. While in the
savikalpaka the object (the qualificand) and its attribute (the qualifier) stand in the
subject-predicate relation, in the nirvikalpaka they are not so related; instead the qualifier
is directly and simply cognized. A perception of cow at the nirvikalpaka stage would be
just cowhood followed by the savikalpaka stage where one makes a predicative claim,
“this is a cow”—that is, the particular cow in front of one qualified by cowhood. Thus, the
difference between them is only in point of predication. Before moving on to outlining
Chakrabarti’s worries about the mischiefs that this distinction causes for Nyāya realism,
let me comment briefly on how this distinction comes to be made within the history of
Nyāya philosophy.
Most philosophers writing on this topic acknowledge that this distinction is a later
addition made by the ninth century ceNyāya commentator Vācaspati Miśra in his
commentary on Gautama’s original definition of perception in Nyāya-Sūtra 1.1.4.
Gautama’s definition of perception is as follows: “perception is the knowledge which
arises from the contact of a sense organ with its object and which is avyapadeśyam
(variously translated as non-linguistic, non-verbal etc.), non-deviant with regard to its
object and of the nature of certainty” (Nyāya-Sūtra 1.1.4). As this definition stands, there
is no indication that there are two kinds of perception. In accommodating this distinction
under the influence of the Buddhists, according to Chakrabarti (2004, p. 367), Vācaspati
is assuming that all predicative or concept-enriched awareness is linguistic, and thereby
precluding the possibilities of understanding concepts as “pre-linguistic recognitional
capacities” (see also Chakrabarti, 2000, p. 5). So Vācaspati is accounting for the
appearance of the term “avyapadeśya” (nonlinguistic) in the definition of perception by
arguing that Gautama’s definition also indicates that there are two kinds or stages of
perception and while one of the defining characteristics of the definition apply to one
kind, others apply to the second kind. As Chatterjee (1965, p. 195) confirms, “Vācaspati
Miśra, ... in his Tātparyatīkā (p. 125) traces the distinction to Nyāya-Sūtra, 1.1.4.
Following his teacher, Trilocana, he takes the words avyapadeśyam and vyavasāyātmakam
contained in this sūtra to mean respectively nirvikalpaka and savikalpaka perceptions.”
Most Naiyāyikas after Vācaspati have accepted this distinction including the Navya-Nyāya
school. Indeed, Gangeśa, an eminent Navya-Nyāya author of fourteenth ce, offers one of
the clearest defenses for the nirvikalpaka perception in terms of a two-step argument.
Quoting Gangeśa, Phillips (2001, p. 105) presents this argument in the following fashion:

[nirvikalpaka perception] is posited by force of the following inference as the first step of a two-
step argument. “The perceptual cognition, ‘A cow’ (for example), is generated by a cognition of the
qualifier, since it is a cognition of an entity as qualified (by that qualifier appearing), like an
inference.” The second step takes a person’s first perception of an individual (Bessie let us say) as
a cow (i.e., as having some such property) as the perceptual cognition figuring as the inference’s
subject (or pakṣa) such that the cognizer’s memory not informed by previous cow experience could
not possibly provide the qualifier, cowhood. The qualifier has to be available, and the best
candidate seems to be its perception in the raw, a qualifier (cowhood), that is to say, not (as some
are wont to misinterpret the point) as divorced from its qualificandum (Bessie) but rather as
neither divorced nor joined, and furthermore, not as qualified by another qualifier (such as being-a-
heifer) but rather just the plain, unadorned entity. In the particular example, the entity is the
universal, cowhood, or being-a-cow, although, again, it would not be grasped as a universal or
anything except itself.

As this above argument makes clear, within the Nyāya context nirvikalpaka perceptual
cognition is not available for apperception or the awareness that arises immediately
following every mental state and therefore no direct evidence is there for its existence.
Further, nirvikalpaka perception, especially under the Navya-Nyāya understanding,
cannot even be considered veridical since it is not even in a propositional form (see, e.g.,
Chakrabarti, 1998, p. 322). These aspects of nirvikalpaka perception highlight what
Matilal (1986, p. 345) calls the “Nyāya ambivalence” about “a pure, pre-linguistic,
conception-free sensory grasp of the object in its theory of perception.” The basic
argument in its favor seems to come down to this: “Unless I know what ‘blue’ or being
blue is, I cannot judge something to be blue” (p. 344). Such a prior conception of the
qualifier becomes especially necessary in the case of one’s experiencing an object for the
first time.
Chakrabarti (2000) provocatively argues that what this inclusion of the “pure”
perception does for Nyāya is far more than mere textual violation; rather, the price is an
awkward conclusion about the impossibility of self-consciousness of our nirvikalpaka
perceptual states thereby creating an ad hoc exception for the general Nyāya rule that
every awareness is immediately available for apperception. An even deeper problem
surfaces once one explores what possible role can nirvikalpaka perception play especially
given Nyāya direct realism. Chakrabarti considers Mohanty’s (1992, p. 174) claim that
“only such an awareness opens an access to things as they really are.” Against this
argument Chakrabarti (2000, p. 4) writes, “Since things in the world are qualified in
themselves, a perception that shows them to be unqualified and disjointed would not give
us access to them as they really are.” Chakrabarti’s retort here revolves around the
nature of Nyāya direct realism. According to this realism “things really are qualified” and
their qualificative structure lies “out there in the world.” “Greenness will inhere in an
emerald even when nobody will look at it ... Since we do not contribute the featuredness
(vaiśiṣtya) onto the object, why do we need to see it in a causally layered way, first the
qualifier alone [in the nirvikalpaka perception] and then the qualified complex [in the
savikalpaka that supposedly follows]?” (pp. 6–7; emphases in the original).
Chakrabarti’s (2000, p. 4) conclusion here can be summarized by his point that
nirvikalpaka perception is not a kind of perception at all. Because “in being pre-
predicative it is also pre-cognitive and nothing pre-cognitive is a perception.” In his
revised articulation of Nyāya, Chakrabarti wants to jettison the nirvikalpaka from the
Nyāya fold and to understand Gautama’s use of the term “avyapadeśya” (nonlinguistic) in
terms of employing nonlinguistic concepts understood as deployment of capacities of
“demarcating, classifying, recognizing and identifying” (Chakrabarti, 1998, p. 318).
Chakrabarti (1998) has argued that it is the Buddhist prejudice that employment of
concepts has to presuppose language, that is, our conceptual abilities needing to rely on
language that pressured Vācaspati to accommodate the nirvikalpaka perception to justify
Gautama’s use of the term “avyapadeśya” (nonlinguistic) in his definition of perception.
But Chakrabarti questions the viability of conceptuality’s dependence on language and
therefore “nonlinguistic” being understood as “nonconceptual.” The culprit, Chakrabarti
argues, is the presupposition that conceptuality necessarily depends on linguistic ability
that forces some of the later Naiyāyikas to embrace this distinction and make room for
nirvikalpaka perception.
Engaging with this debate from a slightly different angle Chadha (2001) asks how
nirvikalpaka perception of particular-as-such is possible the way Yogācāra-Sautrāntika
Buddhists require. In arguing for the incoherence of that possibility one of the premises
she uses maintains that all cognition requires the possibility of recognition and therefore
applies concepts. The main premises are that “cognition requires the possibility of
recognition” (p. 203); and that possibility of recognition requires application of concepts
to sense impression in a way where sense-faculties and mind come to “cooperate
(immediately) in the act of perception” (p. 202). Concepts for her come to play a role in
our perceptual experience through “imaginative construction” which she also equates
with “conceptualization.” The primary role of concepts for her is that they allow us to
impose an order on our otherwise fleeting, discrete sense impressions. Since Buddhists
argue that perception of pure particulars is possible, Chadha’s main question is how does
such a state qualify as a cognition especially since by the Buddhists’ own admission no
concepts are involved at that level. Chadha contrasts this “conception-free awareness” of
the Buddhists with what she takes to be the Nyāya position: “The Naiyāyika holds that
causal interaction between sense-faculties and an object results in sensory impression
that is no more than a mere physiological change” (p. 200). Thus her interpretation of the
Nyāya process of perception does align it very closely to Kant’s well-known expression,
“intuitions without concepts are blind.” Due to the unavailability of conceptualization,
Chadha argues that Buddhist nirvikalpakaperception of the pure particulars becomes
incoherent. In the process of making that case she develops what she calls a Nyāya-
Kantian model of cognition that maintains that all cognition is concept-laden. The
relevance of this discussion for my project here is where this leaves us in terms of the
nature of a concept for a Naiyāyika.
Chadha (2001, p. 207) argues that she is working with a “minimalist claim” about
concepts; for her concepts are “vehicles to structure the input from sense-object contact”
(emphasis in the original). Chadha’s Kantian interpretation of Nyāya is clear and evident
here. If concepts are viewed as “vehicles to structure input,” then it remains unclear what
role they play in enabling and shaping that input. It seems that concepts come into play
after the input has been registered. So what kind of ability would that make a concept?
Definitely not an ability to reidentify and recognize an object but rather a sense-
impression, a “mere physiological change.” As such it would be more of a classifying and
categorizing ability. Or, as Chadha says, to differentiate “by unifying or synthesizing
discrete sensory impressions” (p. 203). This leads us to ask this question: where do we
use concepts in the Nyāya realist framework? In our interaction with the world? Or, in our
thinking about the interaction with the world? Chadha seems to think that it is at the
latter level where concepts become operational and thus elicits the worry from a number
of commentators that her Nyāya-Kantian model, due to its internalist assumptions cannot
be an appropriate model for Nyāya externalism. To her defense Chadha has clarified that
concepts are not necessary and sufficient conditions that a subject has to know before
applying them; rather for her, concepts are an “ability to identify it [the given] as a
distinct, unique particular” (p. 201).
In insisting on a concept-enriched character of our experience, Chakrabarti and Chadha
seem to be making a shared point. However, while for Chadha concepts are vehicles to
order discrete sense impressions, for Chakrabarti (1998, p. 318) concepts are things that
have “anchorage in direct sense-experience.” Further, Chadha does not highlight the
possibility of pre-linguistic concepts; on the contrary, her use of the model of the Kantian
conceptual schema does not reflect any such possibility. Chakrabarti emphasizes such a
possibility. He also seems to align concepts far more directly with the world and not just
with some systematizing grid that the mind imposes on the world. As he writes, “In order
to fit words, concepts need not have a wordy origin. A common external world can be the
mother of both perceived properties and meanings of common nouns and adjectives.
There could be a deep connection between senses as the avenues through which the
world impinges on our consciousness and sense as that which is grasped when a word is
understood” (Chakrabarti, 1998, p. 321). Thus of the two virtues noted in the introduction
it seems that Chadha clearly fails to draw the distinction between concepts and
conceptions. While Chakrabarti might be able to draw this distinction, it remains unclear
whether his account can draw the distinction between epistemological and metaphysical
accounts of concepts. Even though he makes room for pre-linguistic concepts, he seems to
be still considering concepts only in terms of epistemological concerns.
One could apply the important gap between linguistic abilities and conceptual abilities
that Chakrabarti articulates to the indeterminate-determinate perception debate and
argue that while determinate perception is linguistic, indeterminate perception is pre-
linguistic (avyapadeśyam) but not nonconceptual. However, Chakrabarti (2007) derives a
somewhat different conclusion in arguing for the position that nirvikalpaka or
indeterminate perception is incoherent. Now as we have noted above he offers multiple
reasons—seven to be exact—for this and it might very well be that some of his reasons
might be truly damaging for indeterminate perception by themselves. I do not wish to
pursue that line of reasoning here. What I want to argue instead is that once we develop a
Nyāya theory of concept, we are able to accommodate both nirvikalpaka and savikalpaka
perception within the Nyāya fold.

Millikan’s account of concepts

What we have noted above is that both Chadha and Chakrabarti focus on our
recognitional abilities/capacities as foundational to the Nyāya view on concepts. Further
Chakrabarti has also emphasized that possession of a concept is not the same as knowing
the meaning of a word. It seems then that an abilities view of concepts—where concepts
are conceived in terms of various abilities—might be useful in the Nyāya context. This is
what brings us to Millikan’s account. She develops it as a response to the dominant trend
in the literature on concepts, namely, the descriptionist view where concepts are not only
taken as items within our cognitive system—basic elements required for any kind of
thinking—but also as items identical to descriptions of properties of things.
Descriptionism maintains that descriptions associated with a concept determine what it is
a concept of. Thus descriptionism would argue that the extension of the concept cow is
determined by various descriptions/conceptions we have about cows, like four-legged
animal, maker of “moo” sounds, and believed to be worshipped by Hindus, and so on.
Arguing against this descriptionist assumption, Millikan contends that extensions of our
concepts are determined directly by the thing or stuff in the external world and not by any
description—macro-level or micro-level—that we might have about that thing or stuff. So
right away the affinity of this account to Nyāya becomes apparent. However, that might
not provide reason enough for using Millikan’s account as a template for the Nyāya
theory. Other realist and externalist philosophers have also maintained similar positions.
Strawson (1998, p. 324), for example, seems to make a similar case when he writes that
our grasp on many of our concepts is “forced upon us in the course of perceptual
experience of our common world.” What makes Millikan’s account especially relevant in
my opinion is how it enables a number of ways to capture some salient aspects of the
Nyāya sensibility.
Let me highlight two aspects of Millikan’s account of substance3 concepts: first, the
clear articulation of the role of concepts in maintaining the integrity of our cognitive
systems; and second, having a concept is not necessarily the same as having some
linguistic abilities. Given that the central job of a cognitive system “is the exceedingly
difficult task of reidentifying individuals, properties, kinds, and so forth, through diverse
media and under diverse conditions” (Millikan, 2000, p. xi), concepts become fundamental
to Millikan’s account since at the basic empirical level they “just are abilities to
reidentify” (Ryder, Kingsbury, and Williford, 2013, p. 9; emphasis in the original). The
concept of a thing for Millikan (1998, p. 58) “is the capacity to represent the substance in
thought for the purpose of information gathering and storage, inference, and ultimately
the guidance of action.” Having concepts of these different substances thus means having
the abilities to reidentify these substances in our different encounters with them under
different conditions and through diverse media. Millikan further claims that early
substance concepts in children are not ‘predicate concepts applied to prior subject
concepts” (p. 59). She argues, “The very point of having the concept mouse would seem to
be that using it, one does not distinguish Amos from Amos’s brother; one conceives them
as the same”—as more mouse (p. 58, emphasis in the original). Thus the structure of early
substance concepts is not given through any descriptions but in terms of children’s ability
to just see more of the same thing. In the next section I will explore how this point might
allow us to model the Nyāya understanding of nirvikalpaka perception, but for now let me
note a few other salient points of Millikan’s account of concepts.
Millikan’s clarification that early substance concepts do not have the predicative
structure or content allows her to highlight the distinction between the ability to classify
and the ability to reidentify. As she indicates, in much of basic level concepts ability to
reidentify precedes the ability to classify since one has “to be able to identify a substance
under diverse circumstances in order to come to know its properties,” before classifying
those properties on subsequent encounters (Millikan, 1998, p. 62). Drawing from her
reasoning here, one could maintain that while classification has “the structure of subject-
predicate judgment,” reidentification does not have a predicative content (p. 62). Thus to
apply a substance concept that a child develops early, say the concept mouse, or the
concept cow, when in the presence of a mouse or a cow, is not to attribute properties,
relations or causes to it. Citing Putnam, Millikan writes,

To call a thing “gold” or “mouse” is not to describe it. Neither concept consists of a representation
of properties. Rather, the extensions of “gold” and “mouse,” ..., are natural units in nature, units to
which the concepts gold and mouse do something like “pointing,” and to which they can continue
to point despite large changes in the properties the thinker represents them as having. (p. 56,
emphases in the original)

This also highlights the clear distinction that Millikan is able to draw between concepts
and conceptions.
As a corollary to this above point, Millikan (1998, p. 56) also clarifies that most of the
early substance concepts do not depend on any mastery of language as human infants and
any animal that collect “practical knowledge over time of how to relate to specific stuffs,
individuals and real kinds, must have concepts of them.” In her account Millikan
characterizes the concept of a thing in terms of our abilities to reidentify that thing with
some regularity in various encounters with it. Millikan also highlights the role of
representation in the success of this ability. She clarifies that having an ability to
reidentify is not the same as having a disposition because however good one’s abilities are
they are still fallible, at least in principle, as “there is no such thing as a way of identifying
a substance that works with necessity” (Millikan, 2000, pp. 7–8). Thus for Millikan using a
concept, that is, reidentifying an object as the same again, is not by any means like
uttering mental identity statements containing two modes of presenting the same thing.
One might argue that this notion of concept is what Chakrabarti (1998, p. 318) has in
mind when he writes, “When a lion has many memories of individual oxen, it has a
concept of an ox although it is not going to be able to give reasons for calling something
an ox or ascribe possession of this concept to itself.”

Millikan’s account as a template for a Nyāya theory of concepts?

The first point that makes Millikan’s account applicable to the Nyāya context is her strong
commitment to realism and argument against descriptionism which maintains that
extensions of concepts, that is, what our concepts are about are determined by
descriptions. The lack of clear distinction between concepts and conceptions might hint at
this bias as well. Given Millikan’s (1998) externalism—the idea that extensions of
concepts are determined by the things in the world (or “natural units in nature” as she (p.
56) puts it) and not by some description in the mind makes her view an attractive
candidate to model a Nyāya view on, given their shared commitments to both realism and
externalism. I think that Millikan’s account is useful in this context also because she offers
us a couple of distinctions that will be useful for drawing out the Nyāya theory.
 One of the distinctions that I have in mind is between identification and classification. I
think it is also useful how she highlights that especially in relation to substance concepts
what a child focuses on is not the specific properties of a thing but more of the general
nature—commonalities of things that allow for reidentification of the thing in different
encounters. This seems to me to come rather close to the Nyāya view of indeterminate
perception. Further, in offering a clear way to theorize about pre-linguistic concepts, her
account provides a way to articulate exact relations between cognition, concept
application, and mastery of language.
Millikan’s account also offers us clear ways to think about the relations between
properties, universals, and concepts. In the Nyāya literature concepts are often equated
with properties and universals. Strawson (1998) worried that concepts cannot be the
same things as properties or universals. Simply speaking, only cognitive systems can have
concepts while almost anything, including the vase on my table, can have properties and
fall under the associated universals. Using Millikan’s account as a template allows us to
articulate this distinction in the Nyāya context clearly. While properties are typically
treated as descriptions thereby taking us to the extensions of concepts, it is really the
thing itself that determines the concept. So in addition to Strawson’s worry we can add
that while properties act as conceptions, concepts are different in being directly
extensionally connected with the thing.
Most importantly it seems to me that the two factors that feature in Millikan’s account
of concepts—namely, our ability to reidentify an object as the same again under multiple
encounters with it and the ability to track that process representationally in a way that
enables mostly successful interaction with the environment, allow us to capture the
nirvikalpaka and the savikalpaka stages of perception for Nyāya. What makes this work is
the interesting way Millikan expresses the structure of the ability to reidentify: as she
points out, at the heart of this ability in a child is her ability to grasp the common aspect
of a thing, that is, “more mouse” rather than being able to differentiate “Amos the mouse”
from his brother. Interestingly, this is what is grasped during indeterminate perception
according to Nyāya.4 Millikan’s account also allows us to articulate how representation
plays a role in savikalpaka perception and the role of language. In this relation we need to
take into account how both Chadha and Chakrabarti (and for that matter Siderits [2004])
connect intentionality with recognitional abilities. Millikan’s account being a
representational account helps articulate this aspect of the Nyāya view as well. Above all,
Millikan’s account allows the Nyāya theory to adhere to the two important distinctions
outlined in the introduction.

A Nyāya theory of concepts

As a first step to developing the Nyāya account of concept by using the template of
Millikan’s account, I want to focus on the topic of recognition in the Nyāya context. As
noted above, both Chadha and Chakrabarti seem to focus on our recognitional abilities as
playing an important role in concept deployment. But it will be useful to explore how
these abilities relate to concepts. It is also important to clarify the notion of recognition or
pratyabhijñā as it is used in the Nyāya context as pointed out by Phillips (2004). While
discussing recognition, Chatterjee (1965) differentiates between two senses of
“recognition.” In the first sense, recognition can simply mean understanding the nature or
character of a thing. “In this sense, to recognise a thing is to know it as such-and-such, as
when I know that the animal before me is a cow” (p. 205). He contrasts this “wide sense”
with the “narrow sense” in which “recognition means knowing a thing as that which was
known before. To recognise thus means to cognise once again that which we are aware of
having cognized before.” The Nyāya notion of pratyabhijñā is recognition in the latter
sense. Thus when I say “this is that Devadatta,” or “this is the same jar that I saw,” there
is a “conscious reference of a past and a present cognition to the same object” (p. 205).
Both these two senses of recognition that Chatterjee mentions involve the use of
language and thus renders them into kinds of savikalpaka perception. If we restrict
recognition to these two senses, then of course Millikan’s account cannot provide a
template for Nyāya recognition that can be used to unpack the structure of nirvikalpaka
perception. However, I want to argue that there is another sense of recognition where it
means reidentification and where the thing is not identified as a cow in front of me but
rather as “more cow” or “cowness.” When we allow this sense of recognition then there is
a way to understand the ability to recognize that is not dependent on our abilities to use
language. What is also important to note in this regard is that when Chakrabarti says a
concept can be understood in nonlinguistic terms and also in terms of an organism’s
ability to recognize, I think he would like for us to open up this third sense of recognition
that is not reliant on language use.
The first aspect of Nyāya theory of concepts is highlighted in Phillips’s (2004, p. 397)
following claim when drawing from Gangeśa’s characterization of nirvikalpaka perception,
he writes, “There are perceptual instances where all the information presented comes
from the object perceived, that nothing comes from the side of the subject, such that the
view of cognition as ‘having form of itself’ (sākāra-vāda) is clearly wrong” (emphasis in
the original). So clearly at least some concepts are completely determined by information
presented during our interaction with an object. Moreover, countering the common
criticism that in nirvikalpaka perception we encounter the pure qualifier as “divorced”
from the object, Phillips clarifies, we encounter the qualifier “as neither divorced nor
joined, ..., not as qualified by another qualifier (such as being-a-heifer) but rather just the
plain, unadorned entity. In the particular example [of seeing a cow], the entity is the
universal, cowhood, or being-a-cow” (pp. 397–398; emphasis in the original). Finally, an
important aspect of Nyāya theory of concept is that at the basic level (i.e., at the
nirvikalpaka level) of the deployment of many concepts there is no predicative content;
rather simple reidentification of the commonalities. Nyāya theory of concept, being a
realist view, also acknowledges that things have multiple properties many of which go
unnoticed on any given occasion of experience. Therefore, the way concept deployment
happens is by, at the basic level, directly perceiving the stuff or the kind. Phillips writes:

If I can touch what I saw, for example, then when I am only touching the thing, I normally will not
be aware of the thing’s color perceptually. If the ontological layering of things and their qualifiers
were not reflected in the causal ordering that has the qualificandum known through knowledge of
one or more of its properties, properties that are already known, then perception of a
qualificandum should entail that the “thick” particular be presented, the thing with all of its
properties, and, as Gangeśa points out in the section on inherence, a blind person in touching a
yellow piece of cloth would know its yellow color. (p. 398)

One could reason from the above discussion that nirvikalpaka perception is like the
perception of infants and small children. Indeed Chatterjee (1965, pp. 194 and 198) likens
nirvikalpaka perception to the perception of children as it is a simple apprehension of the
object. So even though it cannot be expressed in words, or embodied in propositions, it is
cognition nonetheless, since there is a pick of information and deployment of concepts.
The Nyāya theory of concepts developed here maintains a clear distinction between
concepts and conceptions. The distinction between concept and conception is not alien to
the Indian philosophical discussion either. Matilal seems to allude to this distinction while
discussing Udayana’s argument against the Buddhist view of sakāra-vijñāna-naya or
“awareness-with-a-form.” As Matilal (1986, p. 341) puts it,

The problem before the Buddhist is to make sure at first that the concept firehood is not different
from, but rather an integral part of, the conception-loaded awareness, and at the same time make
it possible for it to be externalized (in fact to make it seem to belong to the external). For in that
way he can maintain his thesis that the conception-loaded awareness is both dubious and
corrigible, while pure perception is not so.

It seems to me that Matilal is clearly indicating the need for distinguishing between
concepts and conceptions especially if one wants to maintain commitment to some form of
external reality.

Conclusion: considering a couple of worries

The first worry that needs addressing is how viable is the possibility of accommodating
Millikan’s framework which is clearly evolutionary and naturalist within the scope of
Nyāya. I think it is a good question but one that need not paralyze us. I want to make two
points in relation to this worry: first, one of the ways of understanding Millikan’s
evolutionary framework is where the human mind is viewed as having evolved from
primitive forms of mentality. Now in a framework of Cartesian dualism this evolutionary
understanding of human mind might be problematic. But the Nyāya ascribes to no such
dualism. The Nyāya ontology recognizes nine kinds of substances and self or ātman is
accepted as one of the nine. Since ātman or self is the only noncorporeal and immaterial
substance Nyāya accepts, it has been argued that a kind of dualism emerges in Nyāya as
well (see, e.g., Chakrabarti and Chakrabarti [1991]). However, this dualism between the
physical world and the immaterial self is very different from the Cartesian mind-body
dualism and I believe need not concern us in our discussion here. What is of relevance
here is that the Nyāya views the human mind just as another organ (“inner organ” or
manas)5 having physical form, and thus it seems at least plausible for a Nyāya philosopher
to argue that just as the human eyes evolved from more primitive visual systems, the
human mind too evolved from primitive mentalities. Second, the thorough externalism of
the Nyāya in articulating its mental semantics and its direct realism also seem, at least in
principle, compatible with the evolutionary naturalist framework. In this regard, let me
also mention the overlaps that I detect in what I would like to call Millikan’s and Nyāya’s
“naturalist attitude”—where their respective realisms are firmly situated in our world and
no argument is made by bringing in some mere logical possibilities offered by possible
worlds.
Another worry can be sensed in the following scenario that Matilal (1986, p. 353)
presents:

Suppose I am used to drawing the curtains every morning and seeing a black lamp-post outside.
After a while I become so used to this fact that I expect to see the black lamp-post immediately
after opening the curtain (my memory presents me with the required notion of the qualifiers).
Hence in the first instance I see something as a black lamp-post or see that it is a black lamp-post,
without the intervention of a sensory, non-constructive, awareness of black colour, lamp-post etc.
In a recurring or continuous perception, the perceptual awareness that arises after the first
moment would likewise need no prior presentation of the qualifier by a conception-free awareness.
For the qualifier here would be presented by the judgmental awareness of the first moment.

It might appear that this above scenario presents a problem for using Millikan’s account
of concept as a template for the Nyāya theory of concept. However, since Millikan’s
account only suggests that this is how some basic substance concepts generally come to
have their structure, it is quite consistent with the above scenario of recurring or
continuous perception. However, another observation of Matilal might present a more
serious challenge for my project here:

It may be plausible to argue along the line of Udayana (who was influenced by Bhartṛhari) that the
so-called sensory grasp of an infant (to the extent it is indistinguishable from the reception of the
photo lens of a camera) does not even amount to awareness. For, as it is emphasized, thought,
concept, or implicit language or even manaskāra must penetrate the sense-given to develop into an
awareness event. What is called sensory experience, or ālocana, in this theory will refer to such an
awareness-event. It is contended here that such an awareness-event cannot be totally
unconceptualizable since it is, though very modestly, conception-loaded. It may lack full-blown
concepts but then it is only unconceptualized in this sense, not unconceptualizable.” (pp. 353–354)

What might be problematic for my account is Matilal’s statement that an infant’s

sensory grasp does not amount to cognition or awareness. But the fact that Matilal likens
an infant’s sensory grasp to the reception of photo lens of a camera, therein lies the
biggest difference. I agree that such cognitive event cannot happen without the help of
some concept, where Millikan would differ would be in maintaining that no such abilities
to reidentify in a very general sense are available to the infant.
Finally, pointing to the recent works on concept in the contemporary philosophy,
Chakrabarti (2004, p. 366) laments,

I have for a long time felt that not just Kant’s but just about every Western philosopher’s concept of
a concept is regrettably unclear. Notwithstanding his book A Study of Concepts even Peacocke’s
notion of a concept does not yield obvious answers to such simple queries as: “Can two people
possess the same concept?” or “When I use a concept that I possess to process a perceived
content, do I make the concept itself an object of my perception?”

I believe that the Millikan-inspired Nyāya theory of concepts outlined here can respond to
both these questions. Thus two people can possess the same concept in so far as they
have the ability to reidentify whatever thing or stuff that the concept represents. Two
people share the concept cow when they are able to reidentify a cow as being a cow in
their different encounters, say, in a cattle ranch in Arizona or on a road in New Delhi.
What undergirds this ability in both occasions is not their possession of some necessary
linguistic description but rather their ability to see the cowness when presented with a
cow with some regularity. Similarly, Naiyāyikas can respond to Chakrabarti’s second
question by maintaining that in using a concept to process a perceived content we do not
make the concept an object; rather concepts are abilities that enable us to process—to
reidentify—that content.

Notes
1. Earlier versions of this chapter were presented at the Seventeenth International Vedanta
Conference, Oxford, Ohio, and the 2006 APA Eastern Division Meetings.
2. I will follow the convention in the literature of using italics to indicate concepts.
3. Millikan develops her account of concepts for what she calls “substances” that include real kinds,
individuals, natural kinds, and stuff. A substance for Millikan is something about which enduring
knowledge can be acquired from one or a few encounters that then can be carried over as
numerous means of identification to possible future encounters with it. Millikan’s paradigmatic
examples include individuals like the Empire State Building, stuff like milk and gold, and natural
kinds like mouse. Since these include what Chakrabarti called particular and non-particular
individuals that form the object of most common perceptions in Nyāya, I think this account can be
used as a template for the Nyāya context.
4. It can be argued that the Nyāya constraint that the content of nirvikalpaka perception can never
become accessible to introspection might not be shared by Millikan but I do not think that
necessarily takes away from the striking similarity that exists in terms of the structure of the
ability underlying reidentification.
5. Nyāya-Sūtra 1.1.16 provides an argument, namely, the absence of the possibility of our having two
simultaneous cognitions, for positing the manas as the organ governing our selective attention. As
Ganeri (2012, p. 262) elaborates,
Each self is asserted [by the Nyāya] to have an accompanying but distinct “unconscious mind”
(manas), responsible for a range of unconscious executive functions including the regulation of
respiration and the relay of signals between the body and consciousness, and in general for the
mediation of all aspects of the relationship between the body’s sensory systems, memory stores,
and the conscious states of the self.

References
Chadha, M. (2001), “Perceptual cognition: a Nyāya-Kantian approach.” Philosophy East and West, 51,
197–209.
Chakrabarti, A. (1998), “Experience, concept-possession, and knowledge of a language,” in L. E. Hahn
(ed.), The Philosophy of P. F. Strawson. Chicago: Open Court, pp. 315–323.
Chakrabarti, A. (2000), “Against immaculate perception: seven reasons for eliminating nirvikalpaka
perception from Nyāya.” Philosophy East and West, 50, 1–8.
Chakrabarti, A. (2003), “Perception, apperception and non-conceptual content,” in A. Chatterjee (ed.),
Perspectives on Consciousness. New Delhi: Munshiram Manoharlal, pp. 89–107.
Chakrabarti, A. (2004), “Seeing without recognizing? More on denuding perceptual content.”
Philosophy East and West, 54, 365–367.
Chakrabarti, K. K., and Chakrabarti, C. (1991), “Toward dualism: the Nyāya-Vaiśeṣika way”, Philosophy
East and West, 41, 477–491.
Chatterjee, S. C. (1965), The Nyāya Theory of Knowledge. Calcutta: University of Calcutta Press.
Ganeri, J. (2012), The Self: Naturalism, Consciousness, & the First-Person Stance. Oxford: Oxford
University Press.
Gangeśa, Tattvacintāmaṇi. Edited by S. Nyāyapodhyaya (1906–1908) with Dīdhiti and Jagadīśi of
Jagadīśa; Varanasi: Chowkambha.
Gangopadhyaya, M. (1982), Nyāya: Gautama’s Nyāya-Sūtra with Vātsyāyana’s Commentary. Calcutta:
Indian Studies.
Matilal, B. K. (1986), Perception: An Essay on Classical Indian Theories of Knowledge. Oxford:
Clarendon Press.
Millikan, R. G. (1998), “A common structure for concepts for individuals, stuffs and real kinds: more
mama, more milk and more mouse.” Behavioral and Brain Sciences, 21, 55–65.
Millikan, R. G. (2000), On Clear and Confused Ideas: An Essay about Substance Concepts. Cambridge:
Cambridge University Press.
Mohanty, J. N. (1992), Reason and Tradition in Indian Thought: An Essay on the Nature of Indian
Philosophical Thinking. New York: Oxford University Press.
Phillips, S. (2001), “There’s nothing wrong with raw perception: a response to Chakrabarti’s attack on
Nyāya’s nirvikalpaka pratyakṣa.” Philosophy East and West, 51, 104–113.
Phillips, S. (2004), “Perceiving particulars blindly: remarks on a Nyāya-Buddhist controversy.”
 Philosophy East and West, 54, 389–403.
Putnam, H. (1975), “The meaning of ‘meaning,’” in Mind, Language and Reality. Philosophical Papers,
Vol. 2. Cambridge: Cambridge University Press.
Ryder, D., Kingsbury, J., and Williford, K. (eds.) (2013), Millikan and Her Critics. London: Wiley-
Blackwell.
Strawson, P. (1998), “Reply to Arindam Chakrabarti,” in L. E. Hahn (ed.), The Philosophy of P. F.
Strawson. Chicago: Open Court, pp. 323–325.
 17
Vasubandhu’s Arguments for the Cognition of Nonexistent
Objects
Zhihua Yao

Introduction
In recent years, I have been working on the general philosophical concept of nonbeing,
especially the epistemological issue of “how to know what there is not.” Within the
Buddhist tradition, the elaborate discussion on the Dharmakīrtian concept of non-
cognition (anupalabdhi) is definitely the most relevant source. I had attempted to trace
the development of this concept back to some sources that explicitly advocate a third
pramāṇa non-cognition (apramāṇa), which can be possibly associated with Dharmakīrti’s
teacher Īśvarasena (Yao, 2011). Further traces for this concept of negative cognition or
cognition of negative facts can be found in the extensive texts before the beginning of the
Buddhist Epistemological School of Dignāga.
I paid particular attention to the concept of the cognition of nonexistent objects
(asadālambanavijñāna). In one of my studies, I traced the origin of this concept to the
earliest available Buddhist sources, to the earliest Buddhist School of Mahāsāṃghika and
its subschools (Yao, 2008). In yet another study, I examined some early texts of the
Yogācāra School that argue for the validity of this concept (Yao, 2014). In between
Mahāsāṃghika and Yogācāra, there was the well-known controversy on this very issue
between two sectarian schools of Sarvāstivāda and Dārṣṭāntika-Sautrāntika. It has been
studied in some details by La Vallée Poussin (1936–7), Sakamoto (1981, pp. 135–156),
Yoshimoto (1982, pp. 146–156), Cox (1988), and Dhammajoti (2007, pp. 44–48). Most of
these studies, however, treat the numerous texts and figures concerned under the general
title of “Dārṣṭāntika-Sautrāntika,” without examining carefully the development of this
concept within a span of at least five hundred years.
In my own study, I attempt to examine the different layers of the evolvement of this
concept by analyzing carefully the various arguments proposed by different Dārṣṭāntika-
Sautrāntika masters. The current chapter on Vasubandhu is an example of such an
attempt. As we know, Vasubandhu, with his numerous writings, is a key figure in the
progress or formation of several important Buddhist philosophical schools, including
Sarvāstivāda, Dārṣṭāntika-Sautrāntika, and Yogācāra. By examining Vasubandhu’s
arguments for the cognition of nonexistent objects, we will have a better idea of this
central Dārṣṭāntika-Sautrāntika concept and its relationship with the Mahāsāṃghika and
Yogācāra concepts. Hopefully my study can contribute to a somewhat better
understanding of the relationship between Dārṣṭāntika, Sautrāntika, and Yogācāra, an
issue that has always been interesting, but has attracted more attention in recent years.1

Background: the Vaibhāṣika arguments for the existence of the past and the
future
Unlike the cases of the Mahāsāṃghikas and early Dārṣṭāntikas, in studying Vasubandhu’s
concept of the cognition of nonexistent objects, we do not have to rely on scattered
sources. Nor do we have to learn his views only through the eyes of his opponents.
Vasubandhu left us extensive writings, in which we can find rather systematic arguments
for this concept. These arguments were further refuted, equally extensively, by his critic
Saṃghabhadra. With these rich sources, we can learn how this concept is conceived in
Vasubandhu’s own right as well as in the eyes of his opponent. Moreover, we have the
advantage of identifying the background and context in which the concept was developed.
A general and less explicit context is the discussion on latent defilements (anuśaya).
The main section of the text under discussion appears, curiously, in the fifth chapter on
latent defilements in the Abhidharmakośabhāṣya (hereafter AKBh). However, this is not
surprising if we recall that the concept of the cognition of nonexistent objects originated
in the Mahāsāṃghika discussions on latent defilements and the nonexistence of their
objects (Yao, 2008, pp. 80–85). The early Dārṣṭāntikas in the Mahāvibhāṣa also argued for
this concept on the basis of the unreal objects of conjunction (saṃyoga). Although the
Vaibhāṣikas, the orthodox Sarvāstivādins, refuted their views, the issue is somehow
associated with the theme of latent defilements. As a matter of fact, the Vaibhāṣikas do
not distinguish between latent defilements (anuśaya) and defilements (kleśa) in general
for they hold that both are bound to their objects. By contrast, the Mahāsāṃghikas hold
that anuśaya, being a latent mental state, has no objects, and it only possesses an object
when becoming a manifested defilement (paryavasthāna). For the Vaibhāṣikas, a person is
bound to objects of the past, present, and future by defilements, either latent or
manifested. Then a natural question arises: Do things in the past and future exist?
The question leads to the specific and explicit context that the cognition of nonexistent
objects is disputed. If things in the past and future do not exist, then it is impossible to say
that a person is bound to these objects by defilements, or that he can be liberated from
them. If things in the past and future really exist, then it means that they always exist and
are thus eternal. But this goes against the basic Buddhist teaching of impermanence. The
Vaibhāṣikas maintain that past and future phenomena really exist, but they are not eternal
for they are endowed with the characteristics (lakṣaṇa) of conditioned things.
The Vaibhāṣika opponent in AKBh outlines four arguments to support this view. Two of
them resort to scriptural authority. In a sūtra passage, the Buddha speaks about the
existence of the past and future material form (rūpa).2 In another sūtra passage, the
Buddha speaks more generally about the conditions for the arising of cognition or
consciousness: “Consciousness is produced by reason of two. What are these two? The
organ of sight and a visible thing ... the mental organ (manas) and dharmas.”3 Mental
consciousness (manovijñāna) is believed to be able to cognize the past and future
dharmas. Now if they do not exist, mental consciousness would not arise because of the
lack of one condition.
This discussion on conditions of cognition leads to the first argument from reasoning:
“[The consciousness] has an existent object.”4 The Vaibhāṣika further explains this point:
“A consciousness can arise if an object (viṣaya) exists (sati), but not if it does not exist
(asati). If past and future things do not exist, there would be consciousness of a
nonexistent object (asadālambanaṃvijñānaṃ). Thus in the absence of object
(ālambanābhāvāt), consciousness itself would not exist.”5 This crucial statement is known
as the epistemological argument for the existence of the past and the future. There are
many issues involved with this dense argument, such as the exact meaning of existence
(sati) or nonexistence (asati), the difference between consciousness with a nonexistent
object and consciousness in the absence of object. We shall treat them in due course
below.
The final Vaibhāṣika argument has to do with the central Buddhist idea of karma or
action. A good or bad action can give forth a result. At the moment when the result is
produced, the retributive cause is past. If the past does not exist, then action cannot
produce any result. Note that this argument focuses only on the past. Some
Vibhajyavādins, for example, Kāśyapīyas, probably convinced by this argument, concede
that the past actions that have not yet given result can also be said to exist.6
With these four arguments from scripture and reasoning, the Vaibhāṣika concludes that
both the past and the future exist, and only those who affirm the existence of the dharmas
of all three times are qualified as a sarvāstivāda (a believer in the existence of all).

Argument one: conditions of cognition

After presenting the Vaibhāṣika arguments, Vasubandhu starts to formulate his own
arguments against them. Following Puguang and Fabao, two Chinese commentators of
AKBh, La Vallée Poussin attributed Vasubandhu’s arguments to the Sautrāntika by
interpolating the phrase “the Sautrāntika criticizes.”7 This is based on the understanding
that Vasubandhu, in the prose part of AKBh, subscribes to the views of Sautrāntika,
especially when he is arguing against the Vaibhāṣika orthodoxy. But the relationship
between Vasubandhu and the Sautrāntika is controversial. Some hold that Vasubandhu
himself created the label of “Sautrāntika” to designate his real Yogācāra stance, and
Vasubandhu was the very first Sautrāntika, but in the sense of a disguised Yogācāra.8 I
would rather hold a more traditional view that takes Sautrāntika as a movement of a great
variety of loosely connected groups against the Vaibhāṣika orthodoxy. This movement had
started with the early Dārṣṭāntika masters, and joined force with Kumāralāta,
Harivarman, Śrīlāta, and Vasubandhu. With regard to some key issues, for example,
simultaneous causation (sahabhūhetu), Vasubandhu holds different views from other
Sautrāntika masters like Śrīlāta (Dhammajoti, 2007, p. 26; Katō, 1989, pp. 218–220).
Therefore I will distinguish Vasubandhu’s own views from those of the Sautrāntika he
quotes. In Saṃghabhadra’s criticism of Vasubandhu, he distinguishes the views of the
Dārṣṭāntikas, Kośakāra (the author of AKBh, i.e., Vasubandhu), and the Sthavira (Śrīlāta).
Actually, we will see Vasubandhu’s arguments on the cognition of nonexistent objects in
some aspects are quite distinctive from those of early Dārṣṭāntikas and other Sautrāntika
masters such as Harivarman and Śrīlāta.
Vasubandhu’s first argument aims at the Vaibhāṣika’s reference to the sūtra passage on
conditions of cognition that we cited above (see note 3). According to this passage,
consciousness, whether sensory or mental, arises on the basis of two conditions: senses
and their respective objects. In the case of mental consciousness, its sense is manas or the
mental organ, which is not an “organ” in its proper sense. Rather it is taken to be the
previous continuous consciousness that gives rise to the consciousness of the next
moment.9 The object of mental consciousness is dharma, or, more precisely, all dharmas.
Since what is under dispute is exactly what dharma means or comprises, for now we
understand dharma simply as that which is knowable of mental consciousness, which
virtually covers everything.
With regard to the two conditions, Vasubandhu insists that they play different roles: the
mental organ is a generating condition (janaka-pratyaya), whereas dharmas are
conditions in a quality of object (ālambana-pratyaya). So when mental consciousness
arises, he asks, “Are these dharmas, similar to the mental organ, acting as its generating
condition, or are they only [conditions] in a quality of object (ālambanamātra)?”10
Apparently he is in favor of the latter option.
Vasubandhu uses the example of future dharmas to support his view. Evidently future
dharmas, which will exist after thousands of years or which will never exist, cannot be the
generating condition of a present mental consciousness. To understand this example, we
need to clarify two basic features of the generating condition. First, as his commentator
Yaśomitra puts it, “The posterior cause of a previous effect is unreasonable.”11 In other
words, a generating cause or condition cannot be temporally posterior to its effect. Now if
future dharmas act as the generating condition of the present mental consciousness, then
it would mean that the cause is temporarily posterior to its effect, which is taken to be
unreasonable by Vasubandhu and Yaśomitra. Whether Vasubandhu here also rejects
simultaneous causation is unclear, but elsewhere in AKBh, he does endorse this type of
causation.
The second feature of generating condition is made explicit by Saṃghabhadra,
Vasubandhu’s critic, who says: “If you (=Vasubandhu) say that the mental organ and the
consciousness that arises on the basis of it are of the same continuum without intervals,
and that [the former] that gives rise to [the latter] can be called a generating [condition],
but dharmas cannot be thus called, then ...”12 In other words, if x gives rise to y, then x
and y must be of the same continuum without intervals. Dharmas, by definition, are not of
the same mental continuum as mental consciousness and the mental organ. Hence
dharmas cannot be the generating condition of mental consciousness. If there is an
interval of thousands of years between dharmas and mental consciousness, it would be
even more impossible for dharmas to play this role.
At this juncture, Saṃghabhadra criticizes Vasubandhu for being inconsistent. For
instance, if we apply Vasubandhu’s analysis to the case of visual consciousness arising on
the basis of eyes and material form, then the visual organ (i.e., eyes) should be the
generating condition of visual consciousness, but they are not of the same continuum at
all, one being material, the other mental.13 Moreover, Vasubandhu denies that dharmas of
the far future (“after thousands of years”) can be the generating condition of the present
mental consciousness, but how about dharmas of the near future? For Saṃghabhadra,
dharmas of the far or near future are of the same nature, serving as the objective
condition of mental consciousness. “The mental organ, being the basis (āśraya) of mental
consciousness, acts as the generating condition; dharmas, being the object [of mental
consciousness], give rise to mental consciousness. Although a basis is different from an
object, they share the same meaning of generating condition.”14
Vasubandhu uses nirvāṇa as a further example to argue for the view that dharmas can
only be objects of cognition, but not its generating condition. Nirvāṇa, being an
unconditioned dharma, is defined as “the cessation of all arising.”15 Since it is
contradictory to all arising, nirvāṇa cannot give rise to mental consciousness, and cannot
be its generating condition. As Yaśomitra comments: “Nirvāṇa, because of its cessation,
cannot give rise to [mental] consciousness.”16
With the support of both cases of future dharmas and nirvāṇa, Vasubandhu concludes
that dharmas can be merely the object of cognition without acting as its generating
condition. On the contrary, Saṃghabhadra, taking a Vaibhāṣika point of view, insists that
both the object and the basis of consciousness are the generating condition of a
consciousness, hence it is necessary for a consciousness to arise with two conditions: its
basis and object.
There are some other proposals to bypass this Vaibhāṣika dogma of two conditions of
cognition. Harivarman proposes that “only for the purpose of rejecting the view of self
(ātman), it is thus said that all consciousnesses arise on the basis of two conditions, but
not four.”17 The Vaiśeṣikas believe that a combination of self, mind (manas), sense
(indriya), and object produces a consciousness. The Buddha rejects the roles of self and
mind, and hence speaks of two conditions instead of four. This teaching of the Buddha can
only be seen as a conventional, but not a necessary truth.18 Kamalaśīla, a Yogācāra-
Madhyamaka, holds that “there are two types of consciousness: one with objects and the
other without objects (anālambana). With regard to consciousness with objects, the World-
honored One speaks of ‘consciousness arising on the basis of two conditions.’”19
Kamalaśīla’s opinion makes sense in limiting the scope of this teaching of the Buddha to
cognition with objects. However, the cases where this teaching does not apply include not
only cognition without objects (anālambana), but also cognition with nonexistent objects
(asadālambana), which is exactly what Vasubandhu is arguing for. For him, a crucial step
to establish the cognition with nonexistent objects is that objects of cognition do not have
to act as the generating condition of cognition, that is, the condition that gives rise to
cognition, but are merely its objective condition, that is, the condition in a quality of
object. Then nonexistents such as the past and the future would possibly become objects
of cognition. Vasubandhu thus concludes his first argument: “If dharmas can be merely
objects of cognition (ālambana), then I claim that the past and the future are also objects
of cognition,”20 which leads to his next argument.

Argument two: speaking of the past and the future

As compared to the earlier argument from the past and the future developed among the
Mahāsāṃghika subgroups and early Yogācāras, Vasubandhu’s argument is more elaborate
and sophisticated. As Saṃghabhadra’s detailed refutation of his argument is also
preserved, we are going to examine their debate on the existence of the past and the
future in three aspects: linguistically speaking of the past and the future,
epistemologically knowing them, and their ontological status.
In speaking of the past and the future, we encounter some puzzling features of the
existential verb in Sanskrit. For instance, the Buddha did speak of “asty atītam asty
anāgatam” (AKBh, 299,6–7), literally meaning “the past exists (asti) and the future exists
(asti).” But according to Vasubandhu, here “the word ‘asti’ is irregular.”21 Alternatively, we
can translate the Buddha’s saying as “There is the past and there is the future.” The
subtle difference between the two English expressions becomes more significant in
understanding the following Sanskrit sentence: “asti dīpasya prāgabhāvo ’sti
paścādabhāva iti” (AKBh, 299,7–8). If we translate “asti” into “to exist,” it runs like this:
“The previous nonexistence of the lamp exists; the posterior nonexistence of the lamp
exists.” We have a self-contradictory statement: the lamp is nonexistent, and yet it exists.
Alternatively, we can translate the sentence as follows: “There is (asti) previous
nonexistence of the lamp; there is (asti) posterior nonexistence of the lamp.” This way the
tension between nonexistence and existence is eased as the phrase “there is” is broader
than the term “to exist.” Vasubandhu provides us with another example: “asti niruddhaḥsa
dīpo na tu mayā nirodhita iti” (AKBh, 299,8–9). This sentence is unintelligible if we
translate “asti” as “to exist”: “The lamp exists extinction, but it [was] not extinguished by
me.” There is an apparent contradiction between the lamp’s existence and extinction. The
word “asti” here should be understood as a copula rather than an existential verb, hence
the sentence means: “The lamp is (asti) extinguished, but it [was] not extinguished by
me.”
The source of the problem goes back to the overlapping or confusion between copula
and existential verb in almost all Indo-European languages. In contrast, the two types of
words are clearly distinguished in some non-Indo-European languages such as Chinese
and Arabic (Graham, 1965). English, though an Indo-European language, has a tendency
to distinguish “to be” from “to exist,” hence its copula and existential verb are less easily
confused. That is why we can make better sense of these sentences in their English
translation than in their Sanskrit original.
Without the convenience of distinguishing the copula from the existential verb in
Sanskrit, Vasubandhu came up with an alternative solution to the problem. He says:

We also say that there is (asti) the past and there is (asti) the future.22 The past is that which
existed previously (bhūta pūrva). The future is that which, given its cause exists (sati), will exist
(bhaviṣyati). It is in this sense that we say that there is (asti) [the past, as well as the future].23 But
they are not substantial entities (dravyatas).24

Substantial entities refer to existents in the present. For Vasubandhu, neither the past
nor the future can exist as does the present, rather they only existed previously or will
exist. In other words, he would admit that there are the past and the future, but they do
not exist now. In this sense, the term “to be” (asti) not only has its normal meaning of “to
exist,” but also can mean “not to exist.” Xuanzang’s Chinese translation states explicitly:
“The word ‘is’ (asti) is applied to what exists, as well as to what does not exist.”25
To argue against this interpretation, Vasubandhu’s Vaibhāṣika interlocutor refers to
another sūtra passage: “Past action (karma) which has been destroyed, which has
perished, and which has ceased, does exist (asti).”26 This passage has more to do with the
Vaibhāṣika argument from karma for the existence of the past and the future. Given the
general Buddhist soteriological framework, it seems difficult to hold that past action
merely existed previously. For it implies the denial of the karmic causation.
Vasubandhu responds that when the Buddha said that past action exists, he had in view
its power of giving forth a result, a power which was placed in the series of the agent
through action which has now passed away. “To understand otherwise, that is, if [past
action exists] at the present in and of itself (svena bhāvena), it cannot be considered as
past.”27 To support his position, Vasubandhu quotes from the Paramārthaśūnyatāsūtra:
“The eye, arising, does not come from any place; perishing, it does not remain in any
place. In this way, Oh Bhikṣus, the eye exists (bhavati) after having been nonexistent
(abhūtvā), after having existed (bhūtvā), disappears.”28 Though a short passage, it
provides Vasubandhu with enough scriptural authority to challenge the Vaibhāṣika view
on the existence of the past and the future.29
As we will discuss more below, if the past and the future exist, it would be difficult to
account for becoming and change. The foundational Buddhist teaching of impermanence
(anitya) is, however, committed to becoming and change, and requires one to account for
it. In this sūtra, the eye is arising and disappearing by going through a sequence of
nonexistence–existence–nonexistence, which corresponds respectively to its future,
present, and past states. If a future eye existed, the Buddha would not have said that the
eye exists only after having been nonexistent (abhūtvā bhavati).
So far it seems that Vasubandhu has made a convincing case for the nonexistence of the
past and the future. According to him, “There is (asti) the past and there is (asti) the
future,” but it does not necessarily mean that “the past and the future exist (asti).” In a
linguistic tradition where copula and existential verb are to a great extent confused,
Vasubandhu has made a significant contribution in distinguishing “to be” from “to exist,”
hence we can meaningfully speak of the past and the future without being committed to
their existence.

Argument three: cognition of the past and the future

If the past and the future are nonexistent, then can we know them? And how can we know
them? To be known, they have to be objects of cognition. However, “if they do not exist,
how can they be objects of cognition (ālambana)?”30 Vasubandhu says: “They are (asti) in
the manner in which they are taken as objects of cognition.”31 The past is taken as object
with the mark of the past, that is, having existed (abhūt); while the future with the mark
of the future, that is, coming into existence (bhaviṣyati).
Vasubandhu asks us to reflect our experiences of memory and anticipation. While
recollecting a certain thing or feeling of the past, we do not perceive it as vividly as it
exists (asti) at the present. Instead we remember the way it has existed (abhūt). Similarly,
we can only anticipate a future thing or sensation in the way it will exist (bhaviṣyati), but
not in the way that it exists at the present.
Saṃghabhadra, however, thinks that Vasubandhu misses an important point in his
phenomenological descriptions. That is, any experience of memory or anticipation has to
be based on perception at the present. Without such a basis, it is impossible to distinguish
these experiences from purely illusory cognitions such as perceiving a person in place of a
pillar, or a dove in place of a piece of wood. An object that has existed or will exist has its
basis in the object that exists now. Otherwise these past or future objects would become
pure illusion.32
Vasubandhu seems to have anticipated this objection. He agrees that the present object
is involved in the experiences of memory and anticipation. Only if we experience a thing
at the present can we then recall this object at a later time. “The way that a past [object]
is remembered is like the way a present material form is experienced.”33 When
anticipating a certain object in the future, we also imagine that it will be like the present
object, being grasped by a cognition. “If this [object of memory or anticipation] exists
(asti) just like [the present object], then a present [object] is attained. If it does not exist
(nāsti), [but has existed or will exist], then it is established that there is [the cognition] of
nonexistent objects.”34 In other words, perception of a present object is the cognition of
existent objects, whereas memory or anticipation of a past or future object is the
cognition of nonexistent objects.
For Saṃghabhadra, however, Vasubandhu is contradicting himself. If he said that we do
not perceive a past object as vividly as it exists at the present, but only remember the way
it has existed, then he should not say that the way that a past object is remembered is like
the way that a present object is experienced. It is evident that when we perceive an
existing object, we are not recollecting a past object. Similarly, when we remember an
object that has existed, we do not perceive a present object. Perception and memory,
existing and having existed, should be sharply distinguished.35
I think Vasubandhu would admit such a distinction too. The reason for him to mention
the similarities between memory, anticipation, and perception is to distance them from
illusory experiences. A variety of illusions are used by the Dārṣṭāntikas as supportive
cases for the cognition of nonexistent objects.36 In his arguments for the same concept,
Vasubandhu seems never to refer to a single case of illusion. However, when he treats
objects of memory and anticipation as nonexistent objects, the same status as illusory
objects, their distinction from illusions is blurred. For Saṃghabhadra, at least in the case
of memory, it would make more sense to admit the existence of its objects. He says:

If one remembers a past [object] like the way of [perceiving] a present existent, and holds that the
past object is (asti, alternatively, exists) in the manner in which it is taken as object of cognition
(ālambana), then it is established that the past really exists. Because we remember the past
[object] as existence in the manner of experiencing a real existent at the present. If you admit its
existence as recollected, how is it possible that the past is not really existent?37

Although Saṃghabhadra’s Vaibhāṣika view makes a bit more sense in the case of the past
object, and hence convinced some Vibhajyavādins, it is more difficult to argue for the
existence of the future object. Vasubandhu brings up this issue with the example of sound.
The issue at stake is: “What is the object of [the cognition] which takes as object the
previous nonexistence of sound?”38 This question itself can be credited to the
Dārṣṭāntikas,39 but Vasubandhu elaborates it in further details.
Vasubandhu’s Vaibhāṣika interlocutor answers that the object of this cognition is “the
sound itself” (śabda eva). Vasubandhu thinks this is unreasonable, since it would mean
that “anyone who is in quest of the [previous] nonexistence of sound should make a
noise.”40 The Vaibhāṣika further clarifies that he means “the future state” (anāgatāvastha)
of the sound itself. But it still does not make sense, because, according to the Vaibhāṣika
theory, the future state of sound is existent (sati), how can it be understood as nonexistent
(nāsti), that is, the previous nonexistence (prāgabhāva) of sound? If the Vaibhāṣika holds
that “nonexistence” here means that “it does not presently exist” (vartamāno nāsti),
Vasubandhu thinks that the Vaibhāṣika cannot legitimately speak of the difference
between the future and the present in this manner because he takes the past, present,
and future as of the same nature (ekatvāt). If, however, he would like to acknowledge the
difference between the future and present sounds in terms of nonexistence versus
existence, then it establishes Vasubandhu’s own theory of “existence after not having
existed” (abhūtvābhāva).
In this debate, Vasubandhu is attacking a weak point of the Vaibhāṣikas. On the one
hand, they hold that all the three times exist; on the other hand, they try to account for
various types of nonexistence (abhāva) as developed among Indian philosophers, Buddhist
and non-Buddhist. Hence they would have to endorse a self-contradictory statement: A
future—therefore existent—sound is the previous nonexistence of sound. To resolve this
self-contradiction, Vasubandhu would hold that the previous nonexistence of sound is the
future nonexistent sound. Therefore, the short answer to the question under discussion is:
“The object of the cognition which takes as object the previous nonexistence of sound is
nonexistence.” This way Vasubandhu has proved that “both existence (bhāva) and
nonexistence (abhāva) can be an object of consciousness.”41
In defense of the Vaibhāṣika position, Saṃgabhadra provides an interesting account of
the issue. He revises the Vaibhāṣika view by making the following proposal:

Regarding the consciousness that takes as object the previous nonexistence of sound, its object is
not the sound itself, but the supporting condition (adhiṣṭhāna) where the sound will be. This means
that its object is only the various things on which the sound will be located. Since they are in a
state that the sound has not yet occurred, they are taken to be the nonexistence of sound.42

In this proposal, Saṃghabhadra insists on the Vaibhāṣika view that the object of cognition
has to be something existent, rather than nonexistent. If the sound itself fails the test,
then try the supporting condition of sound. In the location where the sound will be, there
surely exist various things, so the supporting condition is existent. Meanwhile, since the
sound has not yet occurred in this location, it is the nonexistence of sound. The earlier
contradiction between existence and nonexistence is thus resolved.
A key point in his resolution is to reduce previous nonexistence (prāgabhāva) to mutual
nonexistence (anyonyābhāva). Together with posterior nonexistence (dhvaṃsa) and
absolute nonexistence (atyantābhāva), they are the four types of nonexistence generally
accepted among major schools of Indian philosophy including Buddhism. But there were
attempts to reduce all the four types to mutual nonexistence, which can then be analyzed
into the mutual exclusion of two beings, as Saṃghabhadra points out: “All the mutual
nonexistents must be taken as being based upon existents.”43 This way he has successfully
defended the existent status of the future sound, and reduced the cognition of the
previous nonexistence of sound to a cognition of existent objects, that is, the supporting
condition where the sound will be. As I have shown elsewhere, the same strategy in
dealing with negative judgments is found in the later Buddhist philosopher Dharmakīrti
and contemporary philosopher Edmund Husserl (Yao, 2007).

Argument four: the ontological status of the past and the future
In their debates on the linguistic and epistemological issues, both Vasubandhu and his
Vaibhāṣika opponent have indicated their respective ontological commitments. Now let us
examine in more details their different views on the ontological status of the past and the
future. One example is illustrative. Saṃghabhadra is questioned by an interlocutor
without clear school affiliation: “Whether the future lamp has lighted or not?” Then a
dilemma is formulated: if it has lighted, then no difference is found between the present
and future lamps; “if it does not light, then it cannot in itself be a lamp.”44 In face of this
dilemma, Saṃghabhadra explains that “the past and the future in themselves (svabhāva,
ti 體) exist, but their function (kāritra, yong 用) does not exist.”45 This means that the past
and future lamps have no function of lighting, but they themselves exist. Why? “The past
and the future themselves (svabhāva) are taken to be existent because they, being
dharmas to be known, have the nature of knowable (jñeyatva).”46 Saṃghabhadra seems to
have adopted a Vaiśeṣika view that takes the knowable to be identical to existence. This is
also his rationale for developing an epistemological argument for the existence of the past
and the future:
Whatever is knowable is existent.
The past and the future are knowable.
Therefore, the past and the future exist.
Examining this argument more carefully, however, one will find that it is more or less a
circular argument: an epistemological argument is based on an epistemological definition
of existence itself, which never touches its ontological grounds.
Vasubandhu’s Vaibhāṣika interlocutor takes a different approach by holding that past
and future objects are nothing other than atoms in a state of dispersion (vikīrṇa). In
contrast, present objects are atoms in a state of combination (saṃcaya). Objects across all
the three times exist; their difference lies only in their atoms being in a state of dispersion
or combination.
Vasubandhu comes up with several reasons to refute this Vaibhāṣika claim. First, both
parties agree that past and future objects are knowable, but for Vasubandhu, “they are
never grasped in the form of dispersion.”47 Hence the claim is not vindicated
epistemologically. Second, if all the material things are explained in terms of the
dispersion or combination of atoms, then the atoms would become eternal. “There would
be no any [atom] arising or ceasing.”48 This way it would fail to account for any change,
violate the foundational Buddhist teaching of impermanence, and fall into the heresy of
Ājīvikas, one of the major rivals of early Buddhism. If we follow the view as found in the
Paramārthaśūnyatāsūtra instead, impermanence and change can be explained as
“existence after having been nonexistence and disappearance after having existed”
(abhūtvā bhavati bhūtvā ca prativigacchati) (cited above in note 28). Third, mental factors
such as feeling (vedanā) are not made of atoms. How can they be interpreted as the
dispersion or combination of atoms?
Having refuted the Vaibhāṣika theory on the ontological status of the past and the
future, Vasubandhu concludes this argument with a further proof for the cognition of
nonexistent objects. With regard to the feeling that is just mentioned, if we remember a
past feeling as it was experienced when it was present, then this feeling would be, like
atoms, eternal, which is not admissible. Alternatively, “if [the past feeling that is
remembered] is not like [the feeling that is experienced at the present], then it is proved
that there is [the cognition] of nonexistent objects.”49

Argument five: absolute nonexistence

Having discussed the past and the future, which correspond respectively to posterior
nonexistence and previous nonexistence, Vasubandhu moves on to a third type of
nonexistence, that is, absolute nonexistence (atyantābhāva). Taken to be the most extreme
type of nonexistence, in an Indian context it is usually exemplified by the hair of a turtle,
the horn of a hare, or the son of a barren woman. Vasubandhu here uses the example of
the thirteenth sense-sphere (āyatana).
A popular way to characterize the reality among Buddhist philosophers is to classify it
into twelve sense-spheres, which comprise six senses (eye, ear, nose, tongue, body, and
mental organ) and their respective objects (visible form, sound, smell, taste, the tangible,
and dharma). The category of twelve sense-spheres is believed to cover all that is existent,
so there is not a thirteenth sense-sphere. Exactly for this reason, the thirteenth sense-
sphere becomes an example of absolute nonexistence.
Vasubandhu’s Vaibhāṣika interlocutor challenges his concept of the cognition of
nonexistent objects with this example of absolute nonexistence: “If nonexistence (asad)
can be the object of cognition, then a thirteenth sense-sphere also could be [the object of
cognition].”50 For the Vaibhāṣika, “nonexistence,” in its proper sense, means absolute
nonexistence such as the thirteenth sense-sphere. A nonexistent object of cognition is
impossible. If it is an object of cognition, then it is existent. If, however, it is nonexistent,
then it cannot be an object of cognition. And in this case, the cognition can be considered
as having no object rather than having a nonexistent object. At this juncture,
Saṃghabhadra fiercely criticizes Vasubandhu for upholding a self-contradictory position:

It is self-contradictory for him (=Vasubandhu) to say that there is the cognition (buddhi) of
nonexistent objects. If cognition has objects, then he should not say that these objects are
nonexistent; if objects are nonexistent, then he should not say that this cognition has objects. Since
nonexistence means nothing altogether. If what he means is that the objects of this cognition
themselves are nothing altogether, he should explicitly say that this cognition has no objects. Why
does he act like a coward and deceive us by saying that there is cognition that takes nonexistents
 as objects? Therefore, there is definitely no cognition that takes nonexistents as objects.51

Vasubandhu’s approach is rather clever. He asks his interlocutor: “Then what is the object
of a consciousness (vijñāna) which consists in saying, ‘A thirteenth sense-sphere does not
exist (nāsti).’”52 By saying “a thirteenth sense-sphere does not exist,” it is presupposed
that we know that a thirteenth sense-sphere does not exist. This cognition or
consciousness must have an object, then what is it?
The Vaibhāṣika answers: “The object of this [consciousness] is just its name [‘thirteenth
sense-sphere’].”53 According to the Vaibhāṣika theory, name (nāman) is classified as one of
the conditionings dissociated from mind (cittaviprayuktasaṃskāra), which is existent. So
the cognition that takes this name as object is a cognition of existent rather than
nonexistent objects. As a matter of fact, one scheme of classifying different modes of
existence among the Vaibhāṣikas lists nominal existence (*nāmasat) as one of the five
types of existence, and the examples singled out include the hair of a turtle, the horn of a
hare, and sky-flower.54 This way the Vaibhāṣikas have expelled nonexistence from our
scope of knowledge, as the ancient Greek philosopher Parmenides had proposed.
Whatever we say about or think of is at least a nominal existence, and we can never
countenance nonexistence.
For Vasubandhu, however, if the object of this cognition is the name “thirteenth sense-
sphere,” then the earlier proposition can be reformulated as “the name ‘thirteenth sense-
sphere’ does not exist.” But this is contradictory to the Vaibhāṣika’s own theory that
names are existent. Since the Vaibhāṣika’s position is self-defeating, Vasubandhu
establishes the concept of the cognition of nonexistent objects once again in the case of
absolute nonexistence.

Argument six: doubts

Vasubandhu’s final argument has to do with doubts (vimarśa). His Vaibhāṣika opponent
quotes a sūtra passage for support: “It is an impossible state that I know, that I see that
which does not exist (nāsti) in this world.”55 The same passage is quoted by Harivarman’s
opponent in the Janakaparamopadeśa and Īśvara in the Abhidharmadīpa to support the
Vaibhāṣika position.56 Its exact parallel is not found in Chinese Āgamas or Pāli Nikāyas,
but a similar passage is included in the Madhyama Āgama, where the Buddha shared his
experience of seeing light during meditation with his disciples. He had doubts with regard
to this light: “If there is no such [light] in the world, can I see or know it?”57 Because of
the doubt, he lost concentration in meditation and his vision of light also disappeared.
In Vasubandhu’s understanding, this passage does not support the Vaibhāṣika view that
negates the cognition of nonexistent objects. Instead, the Buddha is meant to say: “I am
not like other prideful (ābhimānika) ascetics who perceive a nonexistent light (avabhāsa)
as existent. As for me, I only perceive as existing that which exists (asti).”58 In other
words, the Buddha did not deny the possibility for the ascetics to cognize a nonexistent
light, but only disapproved their claim that the light exists.
Moreover, if the Vaibhāṣika thesis that “all cognitions have existent objects
(sadālambana)”59 is correct, then there is no longer any place for doubts (vimarśa). For
Vasubandhu, doubts are possible only if one has to decide between existent and
nonexistent objects. If all objects of cognition were existent, then all their cognitions
would be assertive and true. But this is counterintuitive.
Saṃghabhadra disagrees with Vasubandhu’s interpretation. First of all, a nonexistent
light cannot appear as existent for the ascetics to perceive. If the nonexistent light can be
perceived, then a thirteenth sense-sphere can also be perceived. As we discussed above,
the thirteenth sense-sphere is absolute nonexistent. So the sūtra passage is meant to say
that other prideful ascetics claim to perceive the light that has not yet appeared, but the
Buddha only perceives the existent light as light. It indicates a difference between the
wrong knowledge of ascetics and the right knowledge of the Buddha. For Saṃghabhadra,
“because all cognitions have existent objects, one can have doubts about these objects,
wondering whether he has right or wrong knowledge about the perceived objects.”60 The
difference between right and wrong can only be found within existent objects, and doubts
arise in the face of the uncertainty between right and wrong. Nonexistence, unlike
existence, cannot be differentiated, hence no difference of right or wrong, good or bad
can be found within nonexistence itself, or between nonexistence and existence.
Therefore, the Vaibhāṣika account of doubts is more reasonable.
Vasubandhu quotes another sūtra passage to support his own position:

Now, Bhikṣus! My disciple, being instructed by me in the morning, makes progress by the evening,
being instructed in the evening, makes progress by the morning. He will know that which exists as
existence and that which does not exist as nonexistence (sac ca satto jñāsyaty asac cāsattaḥ), that
which is not the highest as not the highest, that which is the highest (anuttara) as the highest.61

The same passage is quoted in the Yogācārabhūmi to support the Yogācāra arguments
for the cognition of nonexistent objects.62 This passage indicates that the Buddha would
possibly agree with Vasubandhu’s concept of the cognition of nonexistent objects as he
explicitly mentioned the knowledge of that which does not exist as nonexistence. Hence
Vasubandhu concludes that the Vaibhāṣika argument for the existence of the past and the
future is an unsound argument (ahetu) since it relies on an unestablished premise: “All
cognitions have existent objects.”
Still, Saṃghabhadra has another way to defend the Vaibhāṣika position. He argues that
the word sat in this passage does not mean “existence,” rather it means “good.”
Accordingly asat means “bad” rather than “nonexistence.”63 So the passage has nothing to
do with the cognition of existent or nonexistent objects, rather it is meant to encourage
the disciple to strive for the good and the highest, that is, nirvāṇa.

Conclusion
I have examined Vasubandhu’s theory of the cognition of nonexistent objects by
organizing them into six main arguments. All of them are aimed to refute the Vaibhāṣika
view that the past and the future exist. In particular, the Vaibhāṣikas developed an
epistemological argument for their view:

1. Whatever is knowable is existent.

2. The past and the future are knowable.
3. Therefore, the past and the future exist.
To substantiate premise (1), the Vaibhāṣikas hold that a cognition is possible with two
conditions, a basis (i.e., sense) and an object (i.e., the knowable), and both act as
generating conditions (janaka-pratyaya). Since a nonexistent cannot produce anything, it
cannot be a generating condition of cognition and hence not an object of cognition.
Therefore, whatever is knowable or an object of cognition, it must be existent.
Vasubandhu’s Argument One focuses on the conditions of cognition. He agrees that
both conditions are necessary, but they function differently: the sense is a generating
condition, while the knowable is a condition in a quality of object (ālambana-pratyaya). He
uses the examples of future dharmas and nirvāṇa to illustrate this point. Both are no
doubt knowable, but they cannot give rise to any cognition. Because a cause, that is,
future dharmas, cannot be temporally posterior to its effect, that is, their cognition at the
present. Whereas nirvāṇa is, by definition, the cessation of all arising. If these objects do
not have to act as the generating condition of their cognition, but are merely its objective
condition, then nonexistents, though lacking any generating power, can still be objects of
cognition. Therefore, it is established that there is the cognition of nonexistent objects
and premise (1) is falsified.
With regard to premise (2), the two parties are not disputing on whether the past and
the future are knowable, but they disagree on how they are known, that is, whether they
are known as existents or nonexistents. In Argument Two, Vasubandhu starts with
linguistic issues. In Sanskrit, when one says “asty atītam asty anāgatam,” it can mean
“there is (asti) the past and there is (asti) the future,” or “the past exists (asti) and the
future exists (asti).” This is because copula and existential verb are to a great extent
confused in Sanskrit and other Indo-European languages. Vasubandhu thinks that the
word asti is applied to what exists, as well as to what does not exist. The past does not
exist, but it “existed previously”; the future does not exist, but it “will exist.” Therefore,
we can say “there are the past and the future,” but it does not necessarily mean that “the
past and the future exist.”
In Argument Three, Vasubandhu addresses the question: “If the past and the future do
not exist, how can they be objects of cognition?” The short answer is that the past is taken
as object with the mark of the past, that is, having existed, while the future with the mark
of the future, that is, coming into existence. He further illustrates this view with the
experience of memory and the cognition of the previous nonexistence of sound. In both
cases, these past and future objects have to be nonexistents rather than existents.
Argument Four is concerned with the ontological status of the past and the future,
Vasubandhu does not agree with the Vaibhāṣika view that past and future objects are
nothing other than atoms in a state of dispersion. He thinks that this would make atoms
eternal entities and violate the foundational Buddhist teaching of impermanence.
All the above arguments have proved that the past and the future, that is, posterior
nonexistence and previous nonexistence, can be objects of cognition. In Argument Five,
Vasubandhu moves on to the most extreme type of nonexistence—absolute nonexistence—
and discusses the cognition which consists in saying, “A thirteenth sense-sphere does not
exist.” Here “a thirteenth sense-sphere” is an example of absolute nonexistence. The
Vaibhāṣikas hold that even in this case the cognition is taking as object the name
“thirteenth sense-sphere,” which is a nominal existent. But Vasubandhu thinks that the
Vaibhāṣika position is self-defeating, since the earlier proposition can be reformulated as
“the name ‘thirteenth sense-sphere’ does not exist,” which would mean that an existent is
nonexistent. The final argument from doubts is concerned more with premise (1) again. If
the Vaibhāṣika thesis “all cognitions have existent objects” is correct, then there is no
longer any place for doubts. Doubts are possible only if one has to decide between
existent and nonexistent objects.
Vasubandhu’s arguments for the cognition of nonexistent objects mark the final stage of
the development of this concept before it was taking over by Buddhist logicians in a new
direction and with different terminologies. We can find traces of influence by early schools
in many of his arguments. In particular, three of his arguments focus on the cognition of
the past and the future. As I have studied elsewhere, this type of argument was most
probably originated in the Mahāsāṃghika subgroup Uttarāpathaka and was accepted and
further developed among the Vibhajyavādins including the Mahīśāsakas and the
Dharmaguptakas (Yao, 2008, p. 90). Early Yogācāras also took it as one of their main
arguments (Yao, 2014, pp. 135–137). As I have acknowledged earlier, the specific
argument from the cognition of the previous nonexistence of sound is credited to the
Dārṣṭāntikas. The argument from absolute nonexistence can also be traced back to the
Dārṣṭāntikas (Cox, 1988, pp. 55–59) and, less explicitly, to early Yogācāras (Yao, 2014, p.
145). And the argument from doubts is similar to the Dārṣṭāntika argument from cognitive
error (Cox, 1988, pp. 49–51) and the Yogācāra argument from heretical views (Yao, 2014,
pp. 144–145). As compared to his predecessors, Vasubandhu is innovative in his argument
from conditions of cognition, especially his distinction between the generating condition
and the condition in a quality of object. If we take this central Dārṣṭāntika-Sautrāntika
concept as a testing case for the relationship between various Buddhist philosophical
schools, then I would not quickly conclude that Vasubandhu or the Sautrāntika in AKBh is
a disguised Yogācāra, because, as we see, the influences are from almost all directions. So
I would rather think that Vasubandhu is a great syncretic.
In this chapter, I mainly study the argument for the cognition of nonexistent objects,
and cannot deal with the counterarguments of the Vaibhāṣikas and Saṃghabhadra in
details. Actually, they deserve much credit for bravely launching onto an intellectual path
that leads to the view that everything exists, thus becoming the common enemy for many
Buddhist and non-Buddhist philosophical schools. Although holding a more or less unusual
view, they find a companion in the ancient Greek philosopher Parmenides. And, according
to Priest (2009), Parmenideanism is the orthodox view among contemporary philosophers
in the twentieth century. So I would not pretend to show that Vasubandhu had defeated
his Vaibhāṣika opponents, rather it is still an open issue regarding whether everything
exists or some things exist and some things do not. The cognition of nonexistent objects
seems to hold a key to resolving this controversial issue.64

Notes
1. See especially Krizer (2005) and Dhammajoti (2007, pp. 5–40).
2. See AKBh, 295,10–13. This sūtra is included as No. 79 of the Saṃyukta Āgama, see T2:99.20a10–
24.
3. AKBh, 295,16–17: dvayaṃpratītya vijñānasyotpāda iti uktaṃdvayaṃkatamat | cakṣū rūpāṇi yāvat
mano dharmā iti | Similar passages can be found in No. 214 of the Saṃyukta Āgama, T2:99.54a22-
b1 and the Saṃyutta Nikāya, 35,93.
 4. AK 5.25b: sadviṣayāt.
5. AKBh, 295,20–21: sati viṣaye vijñānaṃpravartate nāsati | yadi cātītānāgataṃna syād,
asadālambanaṃvijñānaṃsyāt | tato vijñānam eva na syād ālambanābhāvāt |
6. Vibhajyavāda is a general title assigned to those who argued against the Sarvāstivāda view of the
existence of the past and the future. For the Kāśyapīya view, see Vasumitra’s
Samayabhedoparacanacakra, T49:2031.17a27-b2.
7. La Vallée Poussin (1971, p. 55): “Le Sautrāntika critique ...” See Puguang’s commentary in
T41:1821.311b17ff and Fabao’s in T41:1822.704a7ff.
8. See, for instance, Kritzer (2005, pp. xxvi–xxx).
9. AK 1.17ab: ṣaṇṇām anantarātītaṃvijñānaṃyad dhi tan manaḥ | “Of the six consciousnesses, the
one which continually passes away, is the manas.”
10. AKBh, 299,20–21: kiṃtasya yathā mano janakaḥpratyaya evaṃdharmā āhosvid
ālambanamātraṃdharmā iti |
11. AKVy, 474,13: na hi pūrvakālīnasya phalasya paścātkālīno hetur yujyata iti.
12. NA, T29:1562.628a9-10: 若謂”意根與所生識，一類相續，無間引生，可名能生。 法不爾”者⋯⋯
13. See NA, T29:1562.628a10-12.
14. NA, T29:1562.628a5-7: 意為意識所依生緣， 法為所緣能生意識。 所依緣別，生緣義同。
15. AKBh, 299,23: sarvapravṛttinirodhā[t].
16. AKVy, 474,14: nirvāṇaṃhi vijñānaṃniruddhān na janayet.
17. Janakaparamopadeśa, T32:1646.364a18-19: 但遮計神，故如是說： 若諸識生， 皆由此二，非四因緣。
18. Janakaparamopadeśa, T32:1646.254b19-20: 佛破神我， 故說二法因緣生識， 非盡然也。“For the
purpose of refuting self, the Buddha speaks of a consciousness arising on the basis of two
conditions. But it is not necessarily true.”
19. Tattvasaṃgrahapañjikā, 630,16–18: dvividhaṃhi vijñānan sālambanam anālambanam ca | yat
sālambanaṃtad abhisandhāya dvyāśrayavijñānadeśanā bhagavataḥ ||
20. AKBh, 299,23–24: atha ālambanamātraṃdharmā bhavanti | atītānāgatam apy ālambanaṃbhavatīti
brūmaḥ |
21. AKBh, 299,7: astiśabdasya nipātatvāt |
22. Alternatively, the past exists (asti) and the future exists (asti).
23. Alternatively, [the past and the future] exist (asti).
24. AKBh, 299,1–3: vayam api brūmo ’sty atītānāgatam iti | atītaṃtu yad bhūtapūrvam | anāgataṃyat
sati hetau bhaviṣyati | evaṃca kṛtvā ’stīty ucyate na tu punar dravyataḥ |
25. Xuanzang’s translation of AKBh, T29:1558.105b12: 有聲通顯有無法。
26. AKBh, 299,11: yat karmābhyatītaṃkṣīṇaṃniruddhaṃvigataṃvipariṇataṃtad asti. Original source
cannot be identified. Yaśomitra refers to the Saṃyukta Āgama when discussing this passage; see
AKVy, 473,16.
27. AKBh, 299,13: anyathā hi svena bhāvena vidyamānam atītaṃna sidhyet |
28. AKBh, 299,14–16: cakṣur utpadyamānaṃna kutaścid āgacchati, nirudhyamānaṃna kvacit
saṃnicayaṃgacchati | iti hi bhikṣavaś cakṣur abhūtvā bhavati bhūtvā ca prativigacchatīti
[Pradhan: pratigacchatīti]| (My corrections of Pradhan’s edition of AKBh follow Odani and Honjō
(2007, Appendix 15–19), except for a few occasions.) An abridged Chinese translation of the
Paramārthaśūnyatāsūtra is included as No. 335 of the Saṃyukta Āgama (T2:99.92c12-26), and the
quoted passage is found in T2:99.92c16-17. The complete Chinese translation of this short sūtra
was made available later (T15:655.806c23-807a20).
29. For the importance of this sūtra to Vasubandhu’s AKBh, see Miyashita (1986).
30. AKBh, 299,24: yadi nāsti katham ālambanam |
31. AKBh, 299,25: yathā[Pradhan: yadā] tadālambanaṃtathāsti.
32. See NA, T29:1562.628b11-14.
33. AKBh, 299,26–27: yathā khalv api vart[Pradhan: tt]amānaṃrūpam anubhūtaṃtathā tad
atītaṃsmaryate |
34. AKBh, 299,28–29: yadi ca tat tathā eva asti vartmānaṃprāpnoti | atha nāsti | asad apy
ālambanaṃbhavatīti siddham |
35. See NA, T29:1562.628b14-20.
36. See especially NA, T29:1562.622a16-27. For contemporary studies, see Sakamoto (1981, pp. 142–
150) and Cox (1988, pp. 49–55).
37. NA, T29:1562.628b20-24: 若如現有追憶過去，而說彼有如成所緣, 是則極成過去實有。以如現在領實有相，如
是追憶過去為有。既許彼有如所追憶，如何過去體非實有？
38. AKBh, 300,9–10: yaś ca śabdasya prāgabhāvam ālambate, kiṃtasyālambanam |
39. See NA, T29:1562.622a25–26.
40. AKBh, 300,10–11: yaḥśabdābhāvaṃprārthayate, tasya śabda eva kartavyaḥsyāt |
41. AKBh, 300,13: ubhayaṃvijñānasyālambanaṃbhāvaś cābhāvaś ca |
42. NA, T29:1562.624b18-20: 此中緣聲先非有識，緣聲依處，非即緣聲。謂但緣聲所依眾具，未發聲位，為聲非
有。
43. NA, T29:1562.624b15: 諸互非有，定依有說。See also T29:1562.431b23.
44. NA, T29:1562.0636a22: 若不然者，應體非燈。
45. NA, T29:1562.0636a23: 謂去來世，體有用無。
46. NA, T29:1562.0636a23-24: 體謂去來所知法性，有所知性，故說為有。
47. AKBh, 300,1: vikīrṇasyāgrahaṇāt |
48. AKBh, 300,3: na tu kiṃcid utpadyate nāpi nirudhyata ity.
49. AKBh, 300,6–7: atha na santi | asad apy ālambanam iti siddham |
50. AKBh, 300,7–8: yady asad apy ālambanaṃsyāt, trayodaśam apy āyatanaṃsyāt |
51. NA, T29:1562.0623b02-07: 又彼所言自相違害。謂說有覺非有為境，若覺有境，則不應言此境非有；若境非
有，則不應言此覺有境。以非有者是都無故，若謂此覺境體都無，則應直言此覺無境。何所怯怖懷諂詐心，矯說有
覺非有為境？是故定無緣非有覺。
52. AKBh, 300,8: atha trayodaśam āyatanaṃnāstīty asya vijñānasya kim ālambanam.
53. AKBh, 300,9: etad eva nāmālambanam |
54. See Mahāvibhāṣa, T27:1545.42a29. The other four types of existence are real (*dravyasat, shiyou
實有), conventional (*prajñaptisat or *saṃvṛtisat, jiayou 假有), composite (*saṃghātasat or
*sāmagrīsat, heheyou 和合有), and reciprocal existences (*anyonyasat or *apekṣāsat, xiangdaiyou 相
待有).
55. AKBh, 300,14–15: yal[Pradhan: yat tat] loke nāsti tad ahaṃjñāsyāmi vā drakṣyāmi vā
nedaṃsthānaṃvidyata iti |
56. See Janakaparamopadeśa, T32:1646.254a20 and Abhidharmadīpa, 269,7. For the authorship of
the Abhidharmadīpa, see Li (2013).
57. Madhyama Āgama, T1:26.0536c27-28: 若世中無是，我可見可知彼耶？ Its parallel is not found in the
Pāli Upakkilesa Sutta (No. 128 of the Majjhima Nikāya), which corresponds to this Chinese sūtra.
58. AKBh, 300,15–16: apare ābhimānikā bhavanty asantam apy avabhāsaṃsantaṃpaśyanti | ahaṃtu
santam evāstīti paśyāmi.
59. AKBh, 300,16–17: sarvabuddhīnāṃsadālambanatve.
60. NA, T29:1562.0622c22-24: 故一切覺皆緣有境，由此於境得有猶豫，謂我於此所見境中，為是正知，為是顛
倒。
61. AKBh, 300,18–21: etu[Pradhan: etat] bhikṣur mama śrāvako yāvat sa mayā kalyam[Pradhan:
kalpam] avoditaḥsāyaṃviśeṣāya paraiṣyati | sāyam avoditaḥkalyaṃ[Pradhan: kalpaṃ] viśeṣāya
paraiṣyati | sac ca satto jñāsyaty asac cāsattaḥ[Pradhan: sacca sato jñāsyati asaccāsataḥ],
sottaraṃca sottarataḥ, anuttaraṃcānuttarata[Pradhan: cānurattarata] iti | See No. 703 of the
Saṃyukta Āgama, T2:99.189a20-b9 and the Aṅguttara Nikāya, 10,22.
62. See Yao (2014, p. 139). Its Chinese translation is found in T30:1579.305a16-18 and Tibetan
translation in D4035:dzi64a3.
63. See NA, T29:1562.623a3-4.
64. This work was supported by the Academy of Korean Studies (KSPS) Grant funded by the Korean
government (MOE) (AKS-2012-AAZ-104). My thanks to Joerg Tuske and Achim Bayer for their
comments and corrections.

References
Primary sources
Abhidharmadīpa with Vibhāṣāprabhāvṛtti of Īśvara. Jaini, P. S. (ed.) (1977), Abhidharmadīpa with
Vibhāṣāprabhāvṛtti. 2nd ed. Patna: K. P. Jayaswal Research Institute. Tibetan Sanskrit Works Series
IV.
[AK(Bh)] Abhidharmakośa (AK) and -bhāṣya (AKBh) of Vasubandhu. Pradhan, P. (ed.) (1975),
Abhidharmakośabhāṣya of Vasubandhu. 2nd edition. Patna: K. P. Jayaswal Research Institute. Tibetan
Sanskrit Works Series VIII.
[AKVy] Sphuṭārthā Abhidharmakośavyākhyā of Yaśomitra. Wogihara, U. (ed.) (1936), Sphuṭārthā
Abhidharmakośavyākhyā: The Work of Yaśomitra. Tokyo: The Publishing Association of
Abhidharmakośavyākhyā. Sankibo Buddhist Book Store, 1989 rpt.
Aṅguttara Nikāya. Morris, R. (1955–76), The Aṅguttara-Nikāya. 6 vols. London: Pali Text Society.
Janakaparamopadeśa of Harivarman. Cheng shi lun 成實論, trans. Kumārajīva. T32:1646.
Jushe lun ji 俱舍論記 of Puguang 普光. T41:1821.
Jushe lun shu俱舍論疏of Fabao 法寶. T41:1822.
Madhyama Āgama. Zhong ahan jing 中阿含經, trans. Saṃghadeva. T1:26.
Majjhima Nikāya. Trenchner, V., Chalmers, R., and Rhys Davids, C. A. F. (eds.) (1888–1925) 1977–93.
The Majjhima-Nikāya. 4 vols. London: Pali Text Society.
Mahāvibhāṣa ascribed to 500 arhats. Apidamo da piposha lun 阿毘達磨大毘婆沙論, trans. Xuanzang 玄奘.
T27:1545.
[NA] Nyāyānusāra of Saṃghabhadra. Apidamo shun zhengli lun 阿毘達磨順正理論, trans. Xuanzang.
T29:1562.
Samayabhedoparacanacakra of Vasumitra. Yibu zonglun lun 異部宗輪論, trans. Xuanzang. T49:2031
Saṃyukta Āgama. Za ahan jing 雜阿含經, trans. Guṇabhadra. T2:99.
Saṃyutta Nikāya. Feer, L. (ed.) (1884–1904) 1975–91. The Saṃyutta-Nikāya. 6 vols. London: Pali Text
Society.
Tattvasaṃgrahapañjikā of Kamalaśīla. Dwarikadas, S. S. (ed.) (1968), Tattvasaṅgraha of ācārya
Shāntarakṣita with the commentary “Pañjikā: of Shrī Kamalashīla. 2 Vols. Varanasi: Bauddha Bharati
Series 1. 1981 rpt.
Yogācārabhūmi ascribed to Maitreyanātha or Asaṅga. Yujiashi di lun 瑜伽師地論, trans. Xuanzang.
T30:1579; rnal ’byor spyod pa’i sa, D4035, sems tsam, tshi 1a1-283a7.

Secondary sources
Cox, C. (1988), “On the possibility of a nonexistent object of consciousness: Sarvāstivādin and
Dārṣṭāntika theories.” Journal of the International Association of Buddhist Studies, 11 (1), 31–87.
Dhammajoti, K. L. (2007), Abhidharma Doctrine and Controversy on Perception. 3rd ed. Hong Kong:
Centre of Buddhist Studies, The University of Hong Kong.
Graham, A. C. (1965), “‘Being’ in linguistics and philosophy: a preliminary inquiry.” Foundations of
Language, 1 (3), 223–231.
Katō, Junshō 加藤純章. (1989), Kyōryōbu no kenkyū 經量部の研究. Tokyo: Shunjūsha.
Kritzer, R. (2005), Vasubandhu and the Yogācārabhūmi: Yogācāra Elements in the
Abhidharmakośabhāṣya. Tokyo: International Institute for Buddhist Studies of the International
College for Postgraduate Buddhist Studies.
La Vallée Poussin, L. de. (1936–37), “Documents d’abhidharma: la controverse du temps.” Mélanges
chinois et bouddhiques, 5, 25–128.
La Vallée Poussin, L. de. (1971), L’Abhidharmakośa de Vasubandhu. Tome IV. Institut Belge des Hautes
Études Chinoises.
Li, Xuezhu 李学竹. (2013), “Abhidharmadīpa no jobun nitsuite” Abhidharmadīpa の序文について.
Indogaku Bukkyōgaku Kenkyū, 65 (1), 379–373.
Miyashita, Seiki 宮下晴輝. (1986), “Kusharon ni okeru hon mu kon u ron no haikei: Shōgi kūsho kyō no
kaishaku o megutte” 『倶舎論』における本無今有論の背景: 『勝義空性経』の解釈をめぐって. Bukkyōgaku
seminā 仏教学セミナー, 44, 7–37.
Odani, Nobuchiyo 小谷信千代 and Honjō, Yoshifumi 本庄良文. (2007), Kusharon no genten kenkyū:
zuiminbon 俱舎論の原典研究: 随眠品. Tokyo: Daizō Shuppan.
Priest, G. (2009). “Not to be,” in Robin Le Poidevin, Peter Simons, Andrew McGonigal, and Ross P.
Cameron (eds.), The Routledge Companion to Metaphysics. London and New York: Routledge, pp.
234–245.
Sakamoto, Yukio 坂本幸男. (1981), Abidatsuma no kenkyū 阿毘達磨の硏究. Tokyo: Daitō shuppansha.
Yao, Zhihua (2007), “Dharmakīrti and Husserl on negative judgments.” Phenomenology 2005, Vol. I:
Selected Essays from Asia, edited by Cheung Chan-fai and Yu Chung-chi. Bucharest: Zeta Books, pp.
731–746.
Yao, Zhihua (2008), “Some Mahāsāṃghika arguments for the cognition of nonexistent objects.” Journal
of Indian Council of Philosophical Research, 25 (3), 79–96.
Yao, Zhihua (2011), “Noncognition and the third Pramāṇa,” in Helmut Krasser, Horst Lasic, Eli Franco,
and Birgit Kellner (eds.), Religion and Logic in Buddhist Philosophical Analysis. Wien: Verlag der
Österreichischen Akademie der Wissenschaften, pp. 477–490.
Yao, Zhihua (2014), “The cognition of nonexistent objects: five Yogācāra arguments,” in JeeLoo Liu and
Douglas L. Berger (eds.), Nothingness in Asian Philosophy. New York and London: Routledge, pp.
133–147.
Yoshimoto, Shingyō 吉元信行. (1982), Abidaruma shisō アビダルマ思想. Kyoto: Hōzōkan.
 Glossary of Frequently Used Sanskrit Terms

ābhāsa appearance
abhāva absence, nonexistence
advaitācāra nondualistic
āgama revelation, scripture
aitihya tradition
ajīva non-soul
ākāra form, appearance, mental images, representations
ākāśa ether, space
ākṛti shape, configuration
ālambana objective support, ground
alaukika extraordinary
ālaya-vijñāna storehouse consciousness
anātman no-self
anekānta vāda multiplexity of reality
anityatva impermanence
anubhava experience
anumāna inference
anupalabdhi non-perception, non-cognition
anuśaya latent defilement
anuvyavasāya higher order cognition
anvaya positive correlation
apoha exclusion
artha wealth, object
arthāpatti circumstantial evidence
ārya-satya noble truth
asadālambanajñāna cognition of nonexistent objects
asati nonexistence
āśraya substrate, basis
asura demon
ātman self (sometimes translated as “soul”), agent
avaktavyatva inexplicability
avasthā state
avayavin a whole
avidyā ignorance
avyapadeśyam nonlinguistic, prelinguistic
bādhaka defeater
bhakti devotion
bhāva existence, essence
bhāvanā conditioning
bhrānti illusion
bhūta element
bīja seed
brahman the whole cosmic entity
buddhi understanding, cognition
caitanya consciousness
caryā practice of what is prescribed and knowledge of what is
forbidden
cetanā intention
cit consciousness
darśana philosophical system, view
dehātmavāda doctrine that the body is the self, psychological materialism
deśa parts
deva god
dharma property, elementary constituent of reality, teaching
dharmin property bearer
dravya substance
dṛṣṭānta example, instance
duḥkha suffering
grahaṇa cognition
guṇa quality, property, attribute
hetu inferential sign/property, reason, probans
hetv-ābhāsa fallacy, pseudo-prover
indriya sense
īśvara God
jaḍa inanimate
jāti class, genus, universal, essence
jīva soul
jñāna awareness, knowledge, cognition
kalpanā mentally constructed
kāma pleasure
kāraka-sāmagrī causal factors
kāraṇa cause, instrument
karma action
kleśa defilement
kṣaṇikatva momentariness
kṣaṇika-vāda doctrine of momentariness
lakṣaṇa characteristic
liṅga inferential sign
loka world
lokaprasiddha inference which proves something within the framework of “this-
anumāna worldly” way of life
manas mind
manaskāra attention
manovijñāna “mental” consciousness
moha bewilderment
mokṣa liberation
naraka hell
navya new
naya-vāda theory of viewpoints
nikṣepa standpoints
nirvikalpaka indeterminate or undifferentiated perception
pratyakṣa
niścaya certainty, certain knowledge
nyāsa standpoints
pakṣa place of inference
para-artha- inference for others
anumāna
parāmarśa realization
paraprakāśa other-illumination
paryāya mode
prācīna old
pradhāna primary nature, universal soul
pradhvaṃsa destruction
prakāśa light
prakāśatā luminosity
prakṛti primordial nature
pramā knowledge
pramāṇa source or instrument of knowledge
prāṇa vital function, breath
prasaṅga indirect proof, consequence
pratibimba optical reflection
pratijñā thesis
pratiṣedha negation
pratītya samutpāda conditioned arising, dependent origination
pratyabhijñā recognition
pratyakṣa perception
prayatna effort, will
preta ghost

pudgala self, person

puruṣa self, soul
pūrvapakṣa established position
rajas the atmosphere, one of the three qualities in Sāṃkhya
philosophy, responsible for activity, lightness
rasa distinctive flavor, sentiment, or feeling prevailing in an artistic
work
rūpa form
śabda language, word, scripture
sādhana proof
sādhya object or property to be inferred, probandum
sākṣin witness
śakti capacity, power, disposition
sal-lakṣaṇa true characteristic
samādhi meditation, concentration
sāmānya universal
samavāya inherence
sambhava inclusion
saṃjñā recognition
saṃsāra cycle of rebirth
saṃskāra memory trace
saṃvedana awareness
santāna continuum, series
sapakṣa known places that contain the hetu and sādhya
sapta-bhaṅgī seven-figured model (alternative name for syād-vāda)
sarga creation
sārūpya resemblance
śāśvata-vāda doctrine of eternalism
sati existence
sattā existence
sattva existence, essence, living being, one of the three qualities in
Sāṃkhya philosophy, responsible for making a person good,
pure, true, honest, wise
savikalpaka determinate or differentiated perception
pratyakṣa
siddhānta established/settled worldview
skandha psychophysical aggregate of living beings
smṛti memory
sparśa contact
sukha happiness
sva-artha-anumāna inference for oneself
svabhāva self-nature
svalakṣaṇa particular
svaprakāśa self-illumination
svarga heaven
svasaṃvedana self-apprehension, reflexive awareness, self-awareness
syād-vāda sevenfold modal description of Jainism (literally, the theory
[vāda] of “somehow” [syāt])
tamas darkness, one of the three qualities in Sāṃkhya philosophy,
responsible for heaviness, anger, lust, illusion
tarka reasoning
tattva element, entity
trairūpya three characteristics (of a valid reason)
uccheda annihilation, destruction
upādhi inferential undercutter, qualifier
upamāna comparison, analogy
utpanna-pratīti inference of a previously experienced object
anumāna

vāda debate aimed at discovering the truth

vāsanā impression
vastu real thing, object
vedanā feeling
vijñapti mental representation
vijñaptimātratā mere-cognition, that is, the view that reality consists of mental
 representation only
vikriyā transformation
vimarśa realization, doubt
vipakṣa known places that lack the hetu and sādhya
virodha opposition, contradiction
viṣaya object
viśeṣa individuator, qualifier
viśiṣṭa qualified
vyāpti pervasion, invariable concomitance
vyatireka negative correlation
vyavasāya first-order cognition
yukti reason
 Chronological Table of Main Indian Thinkers and Texts
Mentioned in This Volume

Seventh to first Upaniṣads

century bce
Sixth century bce Pārśva (Jain)
Fifth century bce Kundakunda (Jain, author of numerous works, incl.
Pavayaṇasāra [Pravacanasāra])
Vardhamāna Mahāvīra (Jain)
Sixth to fourth Gautama (the “historical” Buddha)
century bce
Second century Mahābhārata (epic)
bceto second
century ce
First century ce Vaiśeṣikasūtra (attributed to Kaṇāda, Vaiśeṣika)
Second century ce Nyāyasūtra (attributed to Gautama Akṣapada, Naiyāyika)
Ṣaṣṭitantra (attributed to Vārṣagaṇya, Sāṃkhya; possibly as
early as first century bce)
Nāgārjuna (Buddhist, author of numerous works, incl.
Mūlamadhyamakakārikā and Vigrahavyāvartanī)
Fourth century ce Asaṅga (Abhidharma Buddhist, author of several works, incl.
Yogācārabhūmi (also ascribed to Maitreyanātha)
Vasubandhu (Abhidharma Buddhist, author of numerous works,
incl. Vādavidhi, Abhidharmakośabhāṣya, and Viṃśatikā)
Fourth to fifth Īśvarakrṛṣṇa (Sāṃkhya, author of Sāṃkhyakārikā)
century ce Vasumitra (Buddhist, author of Samayabhedoparacanacakra)
Fifth century ce Bhartṛhari (Grammarian, author of several works, incl.
Vākyapadīya)
Patañjali (author of Yogasūtra)
Umāsvāti (Jain, author of Tattvārthasūtra and bhāṣya)
Vātsyāyana (Naiyāyika, author of Nyāyabhāṣya)
Sixth century ce Diṅnāga (Buddhist, author of numerous works, incl.
Nyāyamukha and Pramāṇasamuccaya)
Gauḍapāda (Sāṃkhya, author of Sāṃkhyakārikābhāṣya)
Praśastapāda (Vaiśeṣika, author of Praśastapādabhāṣya
[Padārthadharmasaṅgraha])
Pūjyapāda Devanandin (Jain, author of Sarvārthasiddhi)
Śaṅkarasvāmin (Buddhist, author or Nyāyapraveśa)
Sthiramati (Buddhist, author of numerous works, incl.
Madhyāntavibhāgaṭīkā)
Yaśomitra (Buddhist, author of Sphuṭārthā
Abhidharmakośavyākhyā)
Seventh century ce Candrakīrti (Buddhist, author of Prasannapadā)
Devendrabuddhi (Buddhist, author of Pramāṇavārttikapañjikā)
Dharmakīrti (Buddhist, author of numerous works, incl.
Pramāṇavārttika and Vādanyāya)
Kumārila Bhaṭṭa (Mīmāṃsaka, author of several works, incl.
Ślokavārttika)
Uddyotakara (Naiyāyika, author of Nyāyavārttika)
Samantabhadra (Jain, author of Āptamīmāṃsā and
Yuktyanuśāsana)
Śaṅkara (Vedāntin, author of numerous works, incl.
Brahmasūtrabhāṣya)
Yuktidīpikā (Sāṃkhya text, author unknown)
Eighth century ce Akalaṅka (Jain, author of numerous works, incl.
Tattvārthavārttika [Rājavārttika])
Dharmottara (Buddhist, author of numerous works, incl.
Kṣaṇabhaṅgasiddhi)
 Kamalaśīla (Buddhist, author of Tattvasaṅgrahapañjikā)
Maṇḍanamiśra (Vedāntin, author of Brahmasiddhi)
Prajñākaragupta (Buddhist, author of Pramāṇavārtikabhāshyam
or Vārtikālaṅkāraḥ)
Śāntarakṣita (Buddhist, author of Tattvasaṅgraha)
Ninth century ce Jayanta Bhaṭṭa (Naiyāyika, author of Nyāyamañjarī)
Jayarāśi Bhaṭṭa (sceptic affiliated with the Lokāyata
(materialist) school, author of Tattvopaplavasiṃha)
Tenth century ce Abhinavagupta (Śaiva, author of Īśvarapratyabhijñāvimarśinī,
Tantrasāra, and Tantrāloka)
Jñānaśrīmitra (Buddhist, author Vyāpticarcā)
Bhaṭṭa Rāmakaṇṭha (Śaiva, author of numerous works, incl.
Kiraṇavṛtti)
Siddharṣigaṇin (Jain, author of Nyāyāvatāravivṛti)
Somānanda (Śaiva, author of Śivadṛṣṭi)
Śrīdhara (Vaiśeṣika, author of Nyāyakandalī)
Utpaladeva (Śaiva, author of several works, incl. Śivadṛṣṭivṛtti,
Īśvarapratyabhijñākārikā and Ajaḍapramātṛsiddhi)
Vāmanadatta (Śaiva, author of Saṃvitprakāśa and Ātmasaptati)
Vācaspati Miśra I (Naiyāyika, author of
Nyāyavārttikatātparyaṭīkā)
Vyomaśivācārya (Vaiśeṣika, author of Vyomavatī)
Eleventh century ce Kṣemarāja (Śaiva, author of Śivasūtravimarśinī and
Spandanirṇaya)
Ratnakīrti (Buddhist, author of numerous works, incl.
Kṣaṇabhaṅgasiddhi Anvayātmika)
Manorathanandin (Buddhist, author of Pramāṇavārttikavṛtti)
Udayana (Naiyāyika, author of numerous works, incl.
Ātmatattvaviveka)
Twelfth century ce Mokṣākaragupta (Buddhist, author of Tarkabhāṣā)
Thirteenth century Rājānaka Jayaratha (Śaiva, author of Tantrālokaviveka)
ce Keśava Miśra (Naiyāyika, author of Tarkabhāṣā)
Malliṣeṇa (Jain, author of Syādvādamañjarī)
Maṇikaṇṭha Miśra (Naiyāyika, author of Nyāyaratna)
Fourteenth century Gaṅgeśa (Navya-Naiyāyika, author of Tattvacintāmaṇi)
ce Maheśvarānanda (Śaiva, author of Mahārthamañjarīparimala)
Sixteenth century ce Raghunātha Śiromaṇi (Navya-Naiyāyika, author of numerous
works, incl. Tattvacintāmaṇidīdhiti)
Vijñānabhikṣu (Sāṃkhya, author of Sāṁkhyapravacanabhāṣya)
Seventeenth century Mathurānātha Tarkavāgīśa (Navya-Naiyāyika, author of
ce numerous works, incl. Tattvacintāmaṇidīdhitimāthurī)
Viśvanātha Nyāyasiddhānta Pañcānana (Navya-Naiyāyika,
author of numerous works, incl. Nyāyasūtravṛtti)
Yaśovijaya (1624–88, Jain, author of Jainatarkabhāṣā)
Nineteenth to
Krishnachandra Bhattacharyya (1875–1949)
twentieth century ce
 Index

ābhāsa here, here, here

and consciousness here
Abhidharmakośabhāṣya here, here, here, here
Abhinavagupta here, here, here, here, here, here, here, here, here, here, here, here, here,
here, here, here
absolute consciousness here
absolute freedom here
absolute idealism 208, here
absolute nonexistence 406, here, here
absolute subjectivity here
abstract entity here, here
active powers (karma-indriya) here, here
Advaita Vedānta here, here, here, here, here, here
on reflection here, here
aesthetic experience here, here
Ajita Keśakambalī here
āṇava impurity here
annihilationism (uccheda-vāda) here
anthropology here
Aristotle here
Armstrong, D. M. here, here
arthāpatti here
Asaṅga here, here
Abhidharmasamuccaya here
Mahāyānasaṃgraha here
asatkāryavāda (effect non-inherent in its cause) here
asiddhi (unestablished) here
ātman here, here, see also puruṣa; soul
auxiliary causality here, here
awareness here, here, here, here
of absence here, here
and soul here

bādha/bādhita (defeated) here

Bandyopadhyay, Nandita here
Bhartṛhari here
Bhāsarvañja here, here
Bhattacharya, Ramkrishna here
Bhattacharyya, Krishnachandra here
awareness of absence here, here
bodily subjectivity here, here, here
characterization of introspection here
distinction between the speakable and the meanable here
on freedom here
multilayered self here
psychic subjectivity here, here
self-knowledge model here
spiritual subjectivity here, here
Subject as Freedom, The here, here, here, here
on subjectivity and freedom here
Bhaṭṭa Mīmāṃsā
higher-order representation (HOR) here
on meta-cognition here
on other-luminosity here
unchanging self and momentariness here
bhūtas here
bhūtavāda here
bodily subjectivity here, here
felt body here, here
immediate knowledge here
perceived body here, here
physical grade here
body is the self here
Brahman here
Bṛhadāraṇyaka Upaniṣad here
Bṛhaspati’s aphorism here, here
Bṛhaspati-sūtras here, here
Buddha (Gautama) here
Buddhism here, here, here, here, here, here
critique of the Brahmanical position on reflection here
on destruction here
dreams here
existence here
five skandhas here
inference theory, Jayarāśi’s critique here
insubstantialism here
mere-cognition (vijñaptimātra(tā)) here
momentariness here, here, here
pratītya samutpāda here
reality concept here
refutation of external objects here
self
as agent here, here
as substance here, here
as unitary essence here
self-illumination here
substance reductionism here
unchanging self and momentariness here
Buddhist cosmology here
Buddhist philosophy here
epistemology here, here
inference (anumāna) here, here
logic here, here
metaphysics here, here, here
reasoning here
svabhāvapratibandha here
trairūpya here, here
Buddhist reflexivism here, here

Candrakīrti here, here, here

capacity
and incapacity, incompatibility of here
multiple capacities here
and production here
space and time here
time-indexing of here, here
ultimate level here
Cārvāka here, here, here
self as unitary essence here
Cārvāka materialism
argument from karma and rebirth here
argument from linguistic usage here
argument from personal identity here
commonsense empiricism (lokaprasiddha anumāna) here
cosmological materialism of here
epistemological ground of here
nature of here
psychological materialism (dehātmavāda) here
verifiabilism (utpanna-pratīti anumāna) here
causal connection here
causality here
causal relation, non-ascertainment of here
cause and effect here, here, here, here, here, here, here, here
and inference here
certification here, here
Chadha, Monima 381
on nirvikalpaka perception here
Chakrabarti, Arindam here
on nirvikalpaka and savikalpaka here
Chatterjee, Satischandra 390
Chattopadhyaya, Debiprasad here
Citsukha here, here
cognition here, here, here, here
citta-caitta model here
dual-aspect view of here
and external objects here
of nonexistent objects here, here, here
cognitive subjectivity here
cognizedness here
collective viewpoint (saṅgraha) here, here
commonsense empiricism (lokaprasiddha anumāna) here
comparison (upamāna) here
comprehensive viewpoint (naigama) here, here
conception 382, here, here
concepts
Chadha’s account of here
Chakrabartis’s account of here
epistemological approaches 382
metaphysical approaches 382
Millikan’s account of here
substance concepts here
concepts, Nyāya theory of here, here
nirvikalpaka perception 381
versus savikalpaka here
conceptual distinction (vikalpa) here
conscious mental states here
consciousness here, here, here
cognizance of here
first-person ontology here
luminosity of here
and reflection here, here, here
reflexivist theory of here
and soul here
three absences here
transitivity principle here, here
conscious non-perception here
contradiction of existence and negation here
cycle of rebirths, liberation from here

darśana here, here

Dārṣṭāntika-Sautrāntika here
Descartes, René here
descriptionism here
destruction
and discontinuation of existence here
Devadatta here
Devendrabuddhi here
dharma here
Dharmakīrti here, here, here, here, here, here, here, here, here, here, here, here, here
non-cognition (anupalabdhi) here
Pramāṇavārttika here
refutation of external objects here
nature of cognition here
Sautrāntika theory of perception here
Vāda-nyāya here
Dignāga here, here, here, here, here, here
on cognition here
reflexivist theory of consciousness here, here
direct viewpoint (ṛju-sūtra) here, here
discontinuation of existenceand destruction here
distinguishing features here
doubts here
dreams
actions in here
and mere-cognition here
dualism here, here, here

egoity here
empirical viewpoint (vyavahāra) here, here
epistemic idealism here, here, here, here
epistemology here
essences here
eternalism (śāśvata-vāda) here
ethical experience here
etymological viewpoint (samabhirūḍha) here, here
exclusion (apoha) here, here, here, here
existence of other worlds here
existence of soul here
external objects, refutation of
Dharmakīrti’s here
Vasubandhu’s here

factual viewpoints (evaṁ-bhūta) here, here

first-order cognition here, here, here, here
five-membered syllogism here
four standpoints theory (nikṣepa, nyāsa) here
freedom
and introspection here
and subjectivity here
fully competent cause here
function, activity (vyāpāra) here

Gaṅgeśa here, here, here, here, here

on hetv-ābhāsa here
on nirvikalpaka perception , here, here
on tarka here, here, here
Tattva-cintā-maṇi here
Gautama Akṣapāda 3, here, here, here, see also Nyāya-sūtra
general acceptability of inductive examples here
Gödel, Kurt here
Grammarians (Vyākaraṇa) here
guṇaguṇibheda here

hell/hell guardians here

hetv-ābhāsa (pseudo-provers) here
asiddhi (unestablished) here
bādha/bādhita (defeated) here
sat-pratipakṣa (counterinference) here
savyabhicāra (deviation) here
viruddha (contradictory) here
higher-order representation (HOR) here, here
Hume, David here
hyperdualism here

idealism here
and realism here
immaterial consciousness here
impermanence (anityatva) here
impressions of speech (abhilāpavāsanā) here
incompatible properties (virūddhadharma) here
Indian philosophical schools here
inference (anumāna) here, here
Buddhist theory of here
from cause to effect here
and connections here
and consciousness here
from existence here
of impermanence (anityatva) here
Jayarāśi’s critique
causation and invariable relation here
impossibility of grasping the invariable relation here
Nyāya theory here
object of here
from own nature here
of the soul (ātman) here
inferential fallacy here
inferential reasoning here, here
inferential reflection here
inferential subject (antarvyāpti) here
instrument (karaṇa) here
egoity here
external instrument (bāhya karaṇa) here
mind (manas) here
understanding (buddhi) here
insubstantialism here
intentionality here
interpretation here
intersubjectivity here, here, here, here
intrinsic validity (svataḥ prāmāṇyam) here, here
introspection
and self-knowledge here
introspective awareness here
Īśvarakṛṣṅa here, here, here, here, here, here, here
Iwata, Takashi here

Jainism here, here, here, here, here

actual condition (bhāva-nikṣepa) here
annihilationism (uccheda-vāda) here
change and identity here, here
communication and interpretation here
complete account (sakalâdeśa) here
context in here
eternalism (śāśvata-vāda) here
existence here
four standpoints (nikṣepa, nyāsa) here, here
incomplete account (vikalâdeśa) here
individuators here
language, ambiguity here
material representation (sthāpanā-nikṣepa) here
multiplexity of reality (anekānta-vāda) here, here
name (nāma) here
seven-figured model (sapta-bhaṅgī) here
sevenfold modal description (syād-vāda) here, here
substance (dravya-nikṣepa) here
theory of view-points (naya-vāda) here, here
transient occurrences (vivarta, vartanā) here
unchanging self and momentariness here
jāti (fallacious rejoinders) here
Jayamiśra here
Jayantabhaṭṭa here
Jayarāśi Bhaṭṭa here, here, here
causal relation here
destruction and discontinuation of existence here
 inference
Buddhist theory here
causation and invariable relation here
from cause to effect here
impossibility of grasping the invariable relation here
Nyāya theory here
object of here
of impermanence (anityatva) here
means of knowledge here
negation and contradiction of existence here
Jayaratha here
Viveka here
Jñānaśrīmitra here
Kṣaṇabhaṅgādhyāya here

Kamalaśīla here, here, here

Kant, Immanuel here
conception of knowledge here
Critique of Pure Reason here
on self here, here
karma here
Cārvāka response to here, here
kārma impurity here
knowledge here, see also pramāṇa
means of
circularity of here
true characteristic of here
Kṣaṇikavādins here
Kṣemarāja here
Netratantra-uddyota here
Kumāra-Kassapa here
Kumārila here, here, here, here, here, here
on cognition here
unchanging self and momentariness here
Kundakunda here

language, ambiguity of here, here

latent defilements (anuśaya) here
learned Cārvāka here
liberation here
light (prakāśa) here
liṅga here
linguistic usages
Cārvāka response to here
logic here, here
logical relation here
logical truth here
Lokāyata/Cārvāka-darśana here
materialism in, see Cārvāka materialism
Lokāyata materialism here, here, here, see also Cārvāka materialism
luminosity (prakāśatā) here, here
other-illumination (paraprakāśa) here
self-illumination (svaprakāśa) here

Madhyamaka here, here, here, here

Mahāsāṃghika here, here
Maheśvarānanda here
Mālinīvijayottara-tantra here
manas (mind) here
Manorathanandin here
Mataṅgapārameśvara here
materialism here, here, here
Ajita Keśakambalī here
in Lokāyata/Cārvāka-darśana, see Cārvāka materialism
 negative and positive elements here, here
Prajāpati here
Virocana here
Matilal here
māyā here, here, here
māyīya impurity here
McGinn, Colin here, here
memory here
mental consciousness (manovijñāna) here
mentalist causal framework here
mental representations (vijñapti-mātra) here
mere-cognition (vijñaptimātra(tā)) here, here, here
mere imagination here
metaphysical idealism here, here, here
Millikan, Ruth here
account of concepts here
Mīmāṃsā here, here
inference of soul here
Mīmāṃsaka here, here, here
cognition here
Mīmāṃsā philosophy here
Mīmāṃsa-Vedānta here
mind (manas) here
mind/consciousness dualism here
Mipham, Jamgon here
mirror images here
ontology here
mode-expressive (paryāyârthika) viewpoints here, here
modus tollendo tollens here
momentariess here
inference from existence here
Ratnakīrti, see Ratnakīrti
momentariness (kṣaṇikatva) here, here, here, here, here
Mṛgendra-tantra here
multiplexity of reality (anekānta-vāda) here

Nāgārjuna here, here

Naiyāyikas here, here, here, here, here, here
Nārāyaṇakaṇṭha here
Navya Nyāya here, here, see also Prācīna Nyāya; Nyāya
absences/absentee here
epistemology here
hetv-ābhāsa (pseudo-provers) here
inference (anumāna) in here, here
jāti (fallacious rejoinders) here
nigraha-sthāna (occasions for censure) here
tarka (suppositional reasoning) here, here, here
upādhi here, here
negation here, here
and contradiction of existence here
negative correlation (vyatireka) here, here
nigraha-sthāna (occasions for censure) here
nirvāṇa here
nirvikalpaka perception here
versus savikalpaka here
non-apprehension (anupalabdhi) here
nondualism here, here
nonexistent objects, cognition of here
non-perception here
Nyāya here, here, here, see also Navya Nyāya; Prācīna Nyāya
Gautama Akṣapāda here
nirvikalpaka perception here, here
other-illumination (paraprakāśa) here
realism here
 savikalpaka perception here
self
as agent here
as substance here
as unitary essence here
syllogism here
theory of concepts, see concepts, Nyāya theory of
view of causality here
Nyāya-Sūtra here, here
Nyāya-Vaiśeṣika here
on destruction here, here
metaphysics here

object-bound viewpoints here, here

object-support here
optical reflections (pratibimba) here
other-illumination (paraprakāśa) here
causal argument here
intentionality argument here
Nyāya’ account of here
phenomenological argument here
other-illumination (paraprakāśa) here

Paesi/Pāyāsi here, here

Pakṣilasvāmin here
Pāli here
Pāñcarātrika Vaiṣṇavas here
Pañcaśikha here
Pāṇinian here
Pārthasārathimiśra here
Pārśva here
past and future
cognition of here
ontological status of here
speaking of here
Patañjali
Mahā-Bhāṣya here
perception (pratyakṣa) here, here, here
with concepts (savikalpaka) here
without concepts (nirvikalpaka) here
and inference here
nirvikalpaka versus savikalpaka here
simultaneous perception here
personal identity here
Cārvāka response to here
perspectivism here
pervasion (vyāpti) here
Phillips, Stephen here
Plato here, here, here, here
Phaedo here, here
pleasure
and soul here
pluralism here
Popkin, Richard here
positive correlation (anvaya) here
Prācīna Nyāya here, here, see also Navya Nyāya; Nyāya
epistemology here, here
inference (anumāna) here, here
logic here, here
metaphysics here, here, here
reasoning here
svabhāvapratibandha here
trairūpya here, here
Prajāpati here
Prajñākaragupta here
pramāṇa here, here, here, here, here
and pramā here
prasaṅga-reasoning here
Praśastapādabhāṣya here
Pratyabhijñā here, here, here
pretas here, here
psyche here
psychic subjectivity here, here
psychological materialism (dehātmavāda) here
puruṣa (soul) here

Raghunātha here
rajas here
Rāmakaṇṭha here
view on self here
rational inquiry here
Ratnakīrti here, here, here
account of causality here
denial of unrealized capacity here
incompatibility of capacity and incapacity here
inference from existence here
Kṣaṇabhaṅgasiddhi here
realism here, here, here, here
and idealism here
reasoning here, here, here, here, here, here
rebirth
Cārvāka response to here, here
recognition (pratyabhijñā) here
reflections
and consciousness here
Mīmāṃsakas on here
Naiyāyikas on here
objective support of here
on psychic subjectivity here
as reflected object here
as reflecting entity here
Śaivas on here
Śāntarakṣita on here
Vasubandhu on here
reflections-manifestations here
reflexive awareness here, here
reflexivism here, here
relational property here
relations here
resemblance here
Ṛgveda here

Sadānanda, Vedānta-Sāra here

Saiddhāntic scriptures here
Śaiva nondualism here, here
and reflection here
Īśa here
light here
Mantra here
Mantreśvara here
nirākāravāda here
prakāśa-vimarśa here
Pratyabhijñā philosophy here
sākāravāda here
Vijñānākala here
Śaiva Siddhānta here, here
mirror analogy here
self in
 against Buddhism here
critique of an unchanging self here
diagnosis here
against Nyāya here
between Nyāya and Buddhism here
substance reductionism here
Śaivism here, here
samanantarapratyaya-argument here
Saṃghabhadra here, here
Sāṁkhya here, here, here, here
active powers (karma-indriya) here
cosmogony and psychology here
egoity here
elements here
emanation theory here
immaterial consciousness here
insubstantialism here
Īśvarakṛṣṅa’s version here
manas (mind) here
materiality here, here
nature (prakṛti) here
qualities (guṇas) here, here
rajas here, here
sattva here, here
senses here
sensibilia here, here
soul as pure consciousness here
substance reductionism here, here
tamas here, here
twenty-five tattvas here
understanding (buddhi) here, here
unmanifest (avyakta) here, here
Sāṁkhya-Kārikā here, here, here
and Buddha here
Sāṃkhya-Yoga here
meditation form here
saṃvedana-argument here
Śaṅkara here
on consciousness here
Sāṅkhya here, here
Sanskrit here
Śāntarakṣita here, here
Madhyamakālaṃkāra here
sarga here
Sartre, Jean-Paul here, here, here
Sarvasiddhāntasaṇgraha here
Sarvāstivāda here
Ṣaṣṭitantra here, here
satkāryavāda (effect inherent in its cause) here
sat-pratipakṣa (counterinference) here
sattva here
Sautrāntika realism here, here
Sautrāntika theory of perception here
savikalpaka perception
versus nirvikalpaka here
savyabhicāra (deviation) here
second-order cognition here, here
self here, here
as agent here
Buddhist arguments against, here
Nyāya versus Buddhism here
Śaiva Siddhānta here
against Buddhism here
 critique of an unchanging self here
diagnosis here
against Nyāya here
between Nyāya and Buddhism here
as substance here
Buddhist arguments against, here
unchanging self and momentariness Bhāṭṭa Mīmāṃsā and Jainism here
Buddhism here
unknowability here
as unitary essence here
self-awareness (svasaṃvedana) here, here, here, here
self-consciousness here
self-illumination (svaprakāśa) here, here, here
intentionality argument against here
memory argument here
meta-cognition argument here
self-knowledge here, here
Bhattacharyya’s view on, see Bhattacharyya, Krishnachandra
and introspection here
subjectivity and freedom here
self-reference here
Self-Śiva here
sense-faculty here
sense-object contact (indriyārthasannikarṣa) here
senses here
seven-figured model (sapta-bhaṇgī) here
sevenfold modal description (syād-vāda) here, here
shared experiences here
simultaneous perception of objects here
Śiva here
absolute I of here
kḷpti and kalpanā here
and universe here, here, here
Śivadṛṣṭi here, here
Śiva-nature here
Śivasūtra-vimarśinī here
skepticism here, here, here
Ślokavārttika here
Ākṛtivāda here, here
Anumānapariccheda here
Apohavāda here, here
Pratyakṣasūtra here
Vanavāda here, here
Socrates here
solipsism here
Somānanda here, here, here, here, here, here
soul here, here, see also ātman; puruṣa
and awareness here
existence of soul here
and inference here
and pleasure here
soul as pure consciousness here
soul-body identity here
space and time here
Spanda-kārikā here
spatial and temporal determinacy here
speech-bound viewpoints here, here
spiritual subjectivity here, here
Śrī Harṣa here
storehouse consciousness (ālaya-vijñāna) here
subjective idealism here, here
subjectivity here
bodily subjectivity here, here, here
 cognitive subjectivity here
and freedom here
psychic subjectivity here, here
spiritual subjectivity here, here
substance (dravya) here
substance-expressive (dravyârthika) viewpoints here, here
substance reductionism here, here
suffering (duḥkha) here
supreme consciousness here
sūtra (foundational text) here
svabhāvahetu here

tajjīvataccharīravāda here
tamas here
Tantric Śaiva here
Tantric scripture ritualcaryāpāda here, here
kriyāpāda here
vidyāpāda here
yogapāda here, here
tarka (suppositional reasoning) here, here, here
Tattva-Samāsa here
Tattvopaplavasiṃha here
trairūpya here
transient occurrences (vivarta, vartanā) here
transitivity principle here, here
Trikaśāsana here
Trika scriptures here
truth, criteria of
absence of sublation here
efficiency of activity here
intrinsic validity here
production by non-defective causal factors here
truthfulness here
Twin Earth here

Udayana here
Uddālaka Āruṇi here
Uddyotakara here, here, here
unchanging self and momentarinessBhāṭṭa Mīmāṃsā and Jainism here
Buddhism here
understanding (buddhi) here, here
unique particular (svalakṣaṇa) here
universal characteristics (sāmānyalakṣaṇa) here
universals here, here
as essences here
Kumārila’s views here
Western philosophy here
unmanifest (avyakta) here, here
unmanifest materiality here
upādhi here, here
Upaniṣads here, here, here, here
Upāya-hṛdaya here
Utpaladeva here, here, here, here, here, here
Ajaḍapramātṛsiddhi here
Vṛtti here, here, here

Vācaspati Miśra here, here

Vādavidhi here
Vaibhāṣika arguments for existence of the past and the future here
Vaiśeṣika here, here, here, here, here, here
self as agent here
self as substance here
Vaiśeṣika (old) here
Vākyapadīya here
Vārṣagaṇya hereVasubandhu here, here, here, here, here, here, here, here, see also
Yogācāra
argument of absolute nonexistence here
argument of cognition of the past and the future here
argument of conditions of cognition here
argument of doubts here
argument of ontological status of the past and the future here
argument of speaking of the past and the future here
reflection thesis here
refutation of external objects here
Vātsyāyana here
Vedānta here
Vedas here, here, here
verbal testimony (śabda) here
verbal viewpoint (sāmprata-śabda-naya) here, here
verifiabilism (utpanna-pratīti anumāna) here
view-points theory (naya-vāda) here
collective viewpoint (saṅgraha) here, here
comprehensive viewpoint (naigama) here, here
direct viewpoint (ṛju-sūtra) here, here
empirical viewpoint (vyavahāra) here, here
etymological viewpoint (samabhirūḍha) here, here
factual viewpoints (evaṁ-bhūta) here, here
verbal viewpoint (sāmprata-śabda-naya) here, here
Vijñānabhairava here
Vijñānabhikṣu here
Vijñānavāda here, here, here, here, here
Vīrāvalī here
Virocana here
viruddha (contradictory) here

Yājñavalkya here, here

Yogācāra here, here, here, here, here, here, here
idealism here
view on self-luminosity here
Yogācāra-Sautrāntika school
nirvikalpaka versus savikalpaka perception here
Yukti-Dīpikā here, here
 Bloomsbury Academic
An imprint of Bloomsbury Publishing Plc
50 Bedford Square London WC1B 3DP UK
1385 Broadway New York NY 10018 USA
www.bloomsbury.com
BLOOMSBURY and the Diana logo are trademarks of Bloomsbury Publishing Plc
First published 2017
© Joerg Tuske and Contributors, 2017
Joerg Tuske has asserted his right under the Copyright, Designs and Patents Act, 1988, to be identified
as the Editor of this work.
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, recording, or any information storage or
retrieval system, without prior permission in writing from the publishers.
No responsibility for loss caused to any individual or organization acting on or refraining from action
as a result of the material in this publication can be accepted by Bloomsbury or the editor.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
ISBN: HB: 978-1-4725-2953-4
ePDF: 978-1-4725-2930-5
ePub: 978-1-4725-3447-7
Library of Congress Cataloging-in-Publication Data
Names: Tuske, Joerg, editor.
Title: Indian epistemology and metaphysics / edited by Joerg Tuske.
Description: New York : Bloomsbury Academic, 2017. |
Includes bibliographical references and index.
Identifiers: LCCN 2017009504 | ISBN 9781472529534 (hb) |
ISBN 9781472529305 (epdf)
Subjects: LCSH: Indian philosophy. | Knowledge, Theory of–India. | Metaphysics.
Classification: LCC B5131 .I47 2017 | DDC 181/.4–dc23
LC record available at https://lccn.loc.gov/2017009504
Cover design: Irene Martinez Costa
Cover image © V & A Images/Alamy
Typeset by Newgen Knowledge Works (P) Ltd., Chennai, India.
Printed and bound in Great Britain
To find out more about our authors and books visit www.bloomsbury.com. Here you will find extracts,
author interviews, details of forthcoming events and the option to sign up for our newsletters.