Professional Documents
Culture Documents
Indian Epistemology and Metaphysics
Indian Epistemology and Metaphysics
Indian Epistemology and Metaphysics
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
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.
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
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).
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.
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
3. (∃x) (Hx & ¬Sx). (There is something that has the prover without having the
probandum.)
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.
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.
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
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.
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.
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.
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.
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.
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 >
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.
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.
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,
∀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:
“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).
“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).
“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).
“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.
2. ¬Pπ2ε1x.
“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).
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
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
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.
[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
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.
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.
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.
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
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?
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.
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
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
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.
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.
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.
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
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
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
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
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
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
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
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
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.
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 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
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
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
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)
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
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:
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:
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.
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
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.
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.
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.
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.
12.10
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.
* * *
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:
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.
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)
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:
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:
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
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.
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)
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.
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.
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)
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)
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
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.
[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.
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.”
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.
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)
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).
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.
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
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.
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:
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
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
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
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
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