After Certainty A History of Our Epistemic Ideals and Illusions Pasnau Full Chapter

After Certainty: A History of Our
Epistemic Ideals and Illusions Pasnau

Visit to download the full and correct content document:
https://ebookmass.com/product/after-certainty-a-history-of-our-epistemic-ideals-and-ill
usions-pasnau/
After Certainty
After Certainty
A History of Our Epistemic
Ideals and Illusions
ROBERT PASNAU
1
OUP CORRECTED PROOF – FINAL, 14/10/2017, SPi
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Robert Pasnau 2017
The moral rights of the author have been asserted
First Edition published in 2017
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2017942554
ISBN 978–0–19–880178–8
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
PREFACE
My earliest ambition as a philosopher, which goes back to my undergraduate days, was

to write about epistemology. Since the topic in its more recent manifestations looked (at
that time, to my youthful eyes) as if it had been quite exhaustingly studied already, it
seemed a sensible career move (at that time, to those youthful eyes) to go back to an era
where there was evidently much room for new research: the middle ages. When
I arrived there via upstate New York, I found, however, that I did not know how to
talk about medieval epistemology and could not even find much epistemology from that
era to talk about. Hence I settled for the theory of cognition.
Some years later I found myself aspiring to write a book on how metaphysics and
epistemology changed in the four centuries between Aquinas and Locke. It eventually
became clear, however, that metaphysics alone would put quite enough strain on
whatever readership there might be for such a book (to say nothing of the strain it
eventually put on the art of bookbinding). The research on epistemology never left my
spiral notebooks.
Still more recently I was invited to spend a term at Oxford delivering the Isaiah Berlin
Lectures on the History of Ideas. Thinking of those notebooks, I decided that I ought to
try, one more time, to write about the history of epistemology. This book is the result.
Since delivering those lectures in the spring of 2014, I have rewritten the main text
quite extensively, but I have endeavored, so as to preserve a sense of the occasion, to
retain the prose style I had adopted then for a public audience. I have also steadfastly
adhered both to the original number of lectures and to something like their original
length, so that the sprawling material of the notebooks might be distilled into a
concentrated argument of manageable size.
Yet, if brevity is the soul of wit, then either soulless or witless I must be, because
I found myself unable to resist cheating on these self-imposed constraints in that most
shameful and dreaded of ways: by adding endnotes. Very long endnotes. But, though
lacking in soul, I am not lacking in heart; hence I have written the main text so that it can
be read without the notes. Moreover, I have tried to write the notes so that they, too,
can be read continuously, without any need to refer back to the corresponding lecture.
Accordingly, readers should on no account attempt the maddening exercise of paging
back and forth between lecture and notes. Each lecture should be read without
interruption, and the notes for a given lecture should be consulted only after that lecture
has been read to the end, and only by readers who want to see more of the dense and
scraggly roots from which these lectures grew.
CONTENTS
Lecture One: The Epistemic Ideal 1

Lecture Two: Evident Certainties 21
Lecture Three: The Sensory Domain 46
Lecture Four: Ideas and Illusions 70
Lecture Five: The Privileged Now 94
Lecture Six: Deception and Hope 117
Notes 139
Acknowledgments 337
Bibliography 339
Index of Names 377
Subject Index 381
Lecture One
The Epistemic Ideal
Introduction
Over the weeks ahead I will sketch a series of chapters in the history of our thinking
about knowledge. Any serious attempt at such a history should confront, from the start,
the surprising fact that, of all the main branches of philosophy today, epistemology is the
most alienated from its history. In ethics, politics, metaphysics, and mind, even in logic
and language, philosophers pursue themes that go back to antiquity and that run almost
continuously over the subsequent centuries. Yet historical precedents are few and far
between for the sorts of discussions that have largely animated epistemology over the
last fifty years. Today the study of knowledge is one of the foundational subjects of
philosophy. But this has not always been so—indeed, for long periods of time, epistem-
ology can appear not to have been an important philosophical subject at all, let alone a
foundational one.
It is symptomatic of these discontinuities that the very term ‘epistemology’ goes back
only to the middle of the nineteenth century, before which time philosophers evidently
felt no need for a special label to talk about the study of knowledge. Moreover, the ways
in which philosophy has been divided over previous centuries have left no space for
epistemology as a distinct field of inquiry. For Aristotelians as well as for Stoics,
philosophy broke down into logic, physics, and ethics. (Aristotle’s metaphysics was
seen as furnishing a kind of appendix to the physical sciences.) Among Arabic philo-
sophers, following the traditions of late antiquity, theoretical philosophy standardly
divided into physics, mathematics, and metaphysics, logic being a further subject on
the side. In the seventeenth century things were much the same. Thomas Hobbes
divided the sciences that study natural bodies from those that study political bodies,
including among the former physics, ethics, poetry, and logic, but not the study of
knowledge itself.a John Locke distinguished “three great provinces of the intellectual
world”: the nature of things, moral philosophy, and the doctrine of signs.b He and his
contemporaries conceived of his great Essay as falling into that last category—a treatise
of logic. In none of these divisions is there any hint of epistemology—under any name—
as a special subject, let alone as a foundational philosophical subject.1
a b
Leviathan 9; see also De corpore 1.9. Essay IV.21.5.
2 Lecture One: The Epistemic Ideal
Still, it is by no means the case that premodern philosophy neglected epistemology.

There were debates over skepticism, of course, and over the relationship between sense
perception and knowledge; and there were extensive investigations into the cognitive
mechanisms that gave rise to sensation and belief. What was very unusual, however,
was to address at any length the problem that lies at the heart of modern epistemology:
how knowledge is to be defined. To be sure, there was Plato in the Meno and the
Theaetetus, and the casual follower of philosophy’s history might be forgiven for
supposing that his example carried forward more or less continuously until the present
day. In fact, however, Plato has always been more honored than imitated, and this is
particularly the case with respect to his interest in definitions. The Platonic dialogues,
especially the early ones, are interested in defining all sorts of things: knowledge, piety,
friendship, courage, justice, statesmanship, and so on. Most of these definitional projects
no longer interest us. Although philosophers still think sometimes about friendship and
courage and regularly about justice, it is rare to find attempts at definition. Knowledge is
the exception. It is only recently, however, that the quest to define the term has been
perceived as a central philosophical question. From Aristotle through the Middle Ages
and well beyond, philosophers took an interest in carefully circumscribing one or
another particular kind of cognitive grasp of reality—perception, imagination, assent,
deduction, and so on—but showed little interest in defining the broad category of
knowledge. That English contains this very general word of positive cognitive appraisal
did not strike philosophers, even those who worked in English, as calling for any special
definitional inquiry.
My aim over the course of these lectures is to consider the sorts of questions about
knowledge that philosophy has asked for most of its history and to examine how the
answers to those questions have changed. I will be particularly interested in what I take
to be one of the pivotal moments in the history of philosophy—the seventeenth-century
rejection of scholastic Aristotelianism—which I will explore by looking at both sides of
that divide. Lying behind the more famous innovations of that revolutionary century
are, I shall argue, a series of decisive transformations in the sorts of epistemic demands
we make on ourselves.2
I have been invited to give these lectures in the name of Isaiah Berlin, the great Oxford
historian of ideas. As fortune would have it, the theme that will run through the course
of these six weeks—of the ideals and illusions that beset our cognitive enterprise—is one
that permeates Berlin’s own thought. Over and over, Berlin warns against a certain sort
of idealizing tendency, the folly of supposing that “all the ideals of mankind” are
compatible. Within both politics and metaphysics, this tendency looms, with respect
to ideals like justice, truth, liberty, and progress. This faith—that we can achieve, all
together, all the various ideals that we aspire to—“is perhaps one of the least plausible
beliefs ever entertained by profound and influential thinkers.”a Even so, this “one belief,
more than any other, is responsible for the slaughter of individuals on the altars of the
great historical ideals.”b In these lectures I wish to trace the rise and fall of such ideals in
the domain of epistemology. Over the weeks to come we will look at how certain
a
Concepts and Categories, p. 198. b
“Two Concepts of Liberty,” in Four Essays, p. 167.
Aristotle’s Ideal Theory 3
ancient epistemic expectations coalesced in the Middle Ages around a comprehensive

theoretical framework, and how those ideals gradually dissolved, piece by piece, in the
seventeenth century. I will begin, this week and next, by characterizing these epistemic
ideals in a general way. In weeks three and four I will focus on perception and illusion;
and then, in the last two weeks, I will take up reason and its frailties, coming around
finally, inevitably, to talk about skepticism.
By the end of today I hope to have explained how epistemology came to stand as a
foundational topic in philosophy. To tell that story, I need to describe the rise of a
distinction between knowledge and science, a development that has its origins in the
breakdown of scholastic Aristotelian metaphysics. But before arriving at these famous
episodes at the dawn of modern science, we need to consider the framework in which
epistemology was pursued for most of its history. The story I wish to tell today and in
the weeks ahead turns on reconceptualizing what a theory of knowledge might look
like. If the history of epistemology looks strange to us, this is because we have lost sight
of the subject’s most prominent aim. Once that aim comes into clearer focus, we will be
able to make much better sense of a great many episodes from the history of this subject.
And, once we understand these episodes better, we will be in a position to ask whether
this alternative epistemic framework might be an improvement on more recent ways of
pursuing epistemology.
The framework I have in mind, and which I will argue dominated the history of our
theorizing about knowledge, is what I call an idealized epistemology. Rather than take as
its goal the analysis of our concept of knowledge, an idealized epistemology aspires, first,
to describe the epistemic ideal that human beings might hope to achieve and then,
second, to chart the various ways in which we commonly fall off from that ideal. As one
might expect, it turns out to be fairly easy to characterize what we would ideally like to
achieve in principle and quite hard to come to grips with what we might actually be able
to achieve in practice.
Aristotle’s Ideal Theory

To see an idealized epistemology in action, we should start by looking back to Aristotle,
the ancient inspiration for so much of what gets said about knowledge during the first
two millennia of philosophy in the Middle East and Europe. Like Plato, Aristotle devotes
an entire treatise to investigating the character of something they both call epistēmē.
But the results Aristotle arrives at, in his Posterior Analytics, are utterly different from
what Plato had suggested in his Theaetetus. Whereas it might plausibly be thought that
the Theaetetus has the same goal as that of many modern epistemologists—considering
our ordinary way of talking about knowledge and of pursuing necessary and sufficient
conditions for its satisfaction—it is clear from the very start of the Posterior Analytics
that this is not Aristotle’s aim. No conversation with an ordinary Athenian, no matter
how one-sided, could plausibly have elicited the result that knowledge concerns a
proposition that is necessary and universal, known on the basis of an affirmative
demonstration in the first syllogistic figure, the premises of which are necessary and
explanatory of the conclusion. This is not what even the most erudite Athenian could
have meant by epistēmē before Aristotle came along, and if this is what epistēmē is, then
we would have to conclude that it is something that hardly anyone has ever had, in
any domain.
But if the Posterior Analytics is not analyzing the meaning of ‘knowledge’, then what is
it doing? In what sense is this an epistemology at all? One line of answer to these
questions has been to find some other English word that better fits Aristotle’s project,
the most prominent such suggestion being that this is a theory of understanding. Clearly
this is a promising idea about how to translate epistēmē in the context of the Posterior
Analytics. The agent who comes to understand a proposition in the way Aristotle
describes goes well beyond simply knowing that proposition. One can come to know
quite well that vines shed their leaves, for example, simply by observation.a But
someone who grasps the general truth in the way Aristotle describes—on the basis of
necessary principles grounded in the vine’s essence—might plausibly be said to have a
better understanding of that truth. Even so, as helpful as this may be as a translation, it
does not go very far toward explaining what Aristotle is after. Is he simply engaged in his
own linguistic project, trying to understand a Greek word for which ‘understanding’ is
the closest English counterpart? Presumably there is something special about epistēmē as
the Posterior Analytics conceives of it, something that makes it worthy of being singled
out for special treatment. And understanding, to be sure, is eminently worthy of study.
But why study this, rather than knowledge? And why develop the details in the way
Aristotle does? Are the arcane requirements of his demonstrative method really intended
as necessary and sufficient conditions for understanding a thing?
The traditional reading of the Posterior Analytics takes it to be a theory of scientific
knowledge. This is another way in which epistēmē has been translated, and the treatise
itself is almost always described in these terms, as offering a theory of knowledge or
understanding in the domain of science. It is odd that this should be so, however, because it
is apparent even on a casual inspection that the treatise’s scope is much broader than
science as we now conceive of it. Although scientific examples figure prominently, they
are not its exclusive focus. The method is evidently meant to apply to mathematics too;
there are, indeed, as many mathematical examples as there are scientific ones. There is
also no reason to describe the method as scientific rather than philosophical, since it is
completely unclear how we would mark the divide between science and philosophy at
this early date. To be sure, various ancient authors use the plural form epistēmai to refer
more or less to what we now think of as ‘the sciences.’ But it is highly misleading to
describe the Posterior Analytics as a treatise on science, given how much more broadly the
theory is meant to apply. A theory that does not discriminate between science and
mathematics on one hand and between science and philosophy on the other is surely not
a theory of science in our sense at all. Scholars will perhaps defend themselves on this
point by insisting that, of course, they are using the term ‘science’ in the broad sense of
the Greek epistēmē. But, once that is said, it becomes clear that characterizing the treatise
as scientific in its concerns in fact explains nothing at all.3
To describe the Posterior Analytics as a theory of science is perhaps most charitably
regarded as shorthand for the more complex idea that it aims at an account of systematic
theoretical knowledge—the sort of thing one does in mathematics and philosophy just as
a
Post. An. II.16.
Aristotle’s Ideal Theory 5
much as in the sciences. One may speak of the project synecdochically as scientific, but
that is just because, as in many other prominent cases, we do not have in English the
right term for conveying what Aristotle is after. There is, however, more to be said here
than this. Commentators have almost unanimously latched onto the notion that the
Posterior Analytics offers a theory of science because they have not seen any other sort of
enterprise in the vicinity that the treatise could be concerned with. It is not just that we
lack a word for systematic theoretical knowledge, but that we lack any place in our
conceptual scheme for the study of such a thing. However, philosophers do of course
study the nature of science. Hence it has become an idée fixe in the recent literature that
this is what the Posterior Analytics does.
Regardless of how the topic of the Posterior Analytics is to be characterized, there is a
further puzzle concerning its methodological prescriptions: that the method described
seems both impractical and in fact unpracticed by Aristotle. If possessing epistēmē
requires grasping first principles and essences, then it seems unlikely that we have
achieved this condition in more than a few domains. (Mathematics would be the most
likely place to find such a methodology in place. But it is unclear whether mathematical
proofs satisfy the requirement that one know a proposition through principles that
explain the reason why it is so.) Perhaps unsurprisingly, Aristotle’s own writings,
including the Posterior Analytics themselves, contain no examples that satisfy all the
necessary requirements. To be sure, his many examples serve individually to illustrate
one or another dimension of the theory, but each seems incomplete in one regard or
another. The prescribed method, then, seems to be one that he himself is incapable of
fully putting into practice.4
All of these puzzles dissolve when one reads the Posterior Analytics as describing an
epistemic ideal. Aristotle characterizes his subject matter as epistēmē haplōs, uncondi-
tionally or unqualifiedly knowledge, in contrast to various lesser forms of knowledge,
which he is willing to count as epistēmai but which are in one way or another deficient.a
According to this distinction, these lesser kinds are simply the ordinary sorts of know-
ledge that human beings regularly do possess, and unqualified knowledge is the ideal
state that we should aspire to, even if its attainment is extremely difficult. This explains
why Aristotle elsewhere seems not to practice what he preaches—not because it is not
his goal, but because it is an idealized goal. For this reason, too, the Posterior Analytics
itself gives us little more than fragments of what an epistēmē taken haplōs is supposed to
be. This is not the perverse failing that it might seem, because Aristotle is describing the
ideal aim of inquiry rather than something that he himself is in a position to achieve.
Aristotle nowhere says explicitly that epistēmē haplōs is an ideal that he has not yet
realized. But he comes close in the Metaphysics, where he remarks:
The study of truth is difficult in one way, in another easy. A sign of this is that no one is able to
attain it completely, nor entirely misses it. But each individual says something concerning the
nature of things, so that while he may individually contribute little or nothing, from the
collaboration of all there comes a great amount. It is like the proverbial door: who can fail to
hit it? In this respect it is easy; but being able to grasp the whole and not only a part makes the
difficulty clear.b
a
E.g. Post. An. 71b10, 72b30, 73b17, 74a33. b
Meta. II.1, 993a30–b7.
Aristotle does not use the term epistēmē here, but it seems clear enough that the study
(theōria) he has in mind is just what he had described in rigorous detail in the Posterior
Analytics. For what is most distinctive about Aristotle’s conception of epistēmē is his
insistence that it involve a grasp not just of a single isolated proposition, but of the whole
causal and inferential network of propositions that lie behind it. Aristotle’s ideal theory
therefore requires a grasp of the whole door, not just a part, and what he tells us here is
that it is easy to make a contribution to epistēmē, but very hard to achieve the complete
ideal. Indeed, “no one” is able to do that.
As soon as one considers the possibility that Aristotle is offering not a theory of
‘knowledge’—even for a special refined domain of inquiry—but an account of the ideal
limit of human inquiry, it becomes easy to see why Aristotle would insist on the various
details of his account. According to the canonical definition of Posterior Analytics I.2,
epistēmē in the ideal, unqualified sense arises “when we think we know of the explan-
ation [aitia] because of which the object holds that it is its explanation, and also that it is
not possible for it to be otherwise.”a This imposes two requirements on epistēmē: that it
be grounded in an explanation; and that it concern what is necessary. Although, as we
will see today and next week, both of these conditions become controversial in the
seventeenth century, it is nevertheless easy to see why Aristotle would have taken each
one to be an element of the cognitive ideal. And, when subsequent chapters introduce
further conditions on epistēmē, these conditions are defended precisely as features of the
cognitive ideal. In I.24, for instance, he offers a lengthy series of arguments as to why
epistēmē should be of the universal rather than the particular. How do we choose
between these options? By determining which achievement is superior. Thus, “if you
know something universally, you know it better as it holds than if you know it
particularly. Hence universal demonstrations are better than particular demonstrations.”b
The next chapter argues in similar fashion about the reason why epistēmē should be
based on affirmative rather than negative premises: not because the latter fail to yield
knowledge or understanding or science, but because “the affirmative, being prior and
more familiar and more convincing, will be better.”c And so I.26 continues by showing
that positive arguments “are better” than arguments cast in the form of a reductio.
Obviously, arguments that are deficient in these respects can significantly increase our
understanding. Indeed, in other places Aristotle happily recognizes a wide variety of
cognitive states that fall short of the ideal described here, such as grasping particulars
and retaining them in memory, reaching conclusions in a nonexplanatory way (by way of
the fact ‘that’, hoti, rather than by way of the reason ‘why’, dia ti), and achieving practical
wisdom in action. We might reasonably describe all of these as kinds of knowledge. The
point of the Posterior Analytics is simply that they are not ideal.5
Normative Ideals
Among authors writing in Latin, from antiquity to the seventeenth century, epistēmē
becomes scientia, and scientia continues to be understood in terms of the cognitive ideal.
Jumping ahead to the high Middle Ages, we can find Albert the Great, in the prologue to
a
Post. An. I.2, 71b9–12. b
I.24, 85b13–15. c
I.25, 86b29.
Normative Ideals 7
his Posterior Analytics commentary (1261), remarking of scientia that “this is the end
and the most perfect and the sole unconditionally desirable thing among the logical
sciences.”a Thomas Aquinas, in his commentary, remarks that “to have scientia of
something is to cognize it perfectly.” From this principle he derives the twin features
of the canonical definition: that what is known in this way must be necessary, and that it
must be grasped through a grasp of its cause.b Describing Adam’s condition in the
Garden of Eden, he writes that “just as the first man was endowed in a perfect state
with respect to his body . . . so too he was endowed in a perfect state with respect to his
soul . . . And thus the first man was endowed by God so as to have scientia of all the
things about which a man is naturally suited to be instructed.”c Even after Aristotle’s
influence began to wane, authors continued to take for granted that scientia should be
understood as the cognitive ideal. Francisco Sanches, for instance, a late sixteenth-
century critic of Aristotelianism, puts in capital letters his definition of scientia as “the
perfect cognition of a thing.”d And René Descartes, in his early Rules for the Direction of
the Mind, conceives of these rules as a guide for the achievement of scientia, and writes
that they “will help us ascend to the peak of human cognition.”e Descartes’s case will be
considered in some detail next week.6
Once we begin to think of scientia (née epistēmē) as the ideal, the peak of perfection, it
becomes natural to worry that, however it is to be characterized, it will remain only an
ideal, unattainable to us. We noticed already how, in Aristotle, it is hard to find a single
conclusion that meets all the criteria of his ideal theory. And indeed, historically, it is
very common to worry about how close human beings might be able to come to the
cognitive ideal. Back in the second century, Ptolemy had remarked that, of the various
theoretical disciplines, “only mathematics can provide sure and unshakable knowledge
[eidēsis] to its devotees.” As for physics and metaphysics, “they should be called
conjecture rather than knowledge [katalēpsis epistēmonikē] . . . There is no hope that
philosophers will ever be agreed about them.”f Pietro Pomponazzi, in the Renaissance,
similarly remarked that “philosophy would be beautiful, if it were as certain as math-
ematics. For metaphysics and philosophy are conjectural, and on almost any subject one
may find different opinions, so that it is like playing with toys.”g (This is one of two
obstacles Pomponazzi describes as plaguing philosophy. The other, naturally, is that it
does not pay.)
Although Ptolemy and Pomponazzi put their complaints in strong terms, their
concern is one that nearly everyone has always shared about almost every area of
human inquiry. So, if epistemology is to be conceived in ideal terms, it might seem that
skepticism lies just around the corner. In fact, however, the strategy of idealization does
not require identifying some sort of absolute ideal and then holding it fixed regardless of
whether it can be obtained. The objective is not to identify standards that only a god
could achieve. The point instead is to define what sort of knowledge we might be able to
achieve, given the world we live in. Accordingly, if the highest cognitive ideal turns
out to be one that can be achieved only in certain limited contexts, then an idealized
epistemology had better be ready to make various retrenchments to the theory, as
a b
Comm. Post. an. I.1.1 (ed. Jammy, I: 514a). In Post. an. I.4.
c d e
Summa theol. 1a 94.3c. Quod nihil scitur, p. 200. Rules 2 (X: 364).
f
Almagest I.1, p. 36 trans.; see also Lecture Two, p. 27. g
See Perfetti, “Pietro Pomponazzi,” §5.
required to identify a level of cognitive excellence that ordinary people can meet in
ordinary circumstances. And this is in fact what we find happening, beginning with
Aristotle. Although the Posterior Analytics’ theory of demonstration sets a formidable
standard for epistēmē, Aristotle elsewhere describes in considerable detail an epistemol-
ogy of nonideal conditions, which he labels dialectic. The Topics, his treatise devoted to
this subject, explains what one should do in cases where one or another component of
demonstrative reasoning is not available, and considers what merit there is to arguments
that fall short of the rigor of epistēmē. Although dialectic does not rise to the level of the
ideal, it is appropriately deployed by certain kinds of people in certain kinds of situations
and is a worthy subject of philosophical investigation.7
For the next two millennia, epistemology largely wrapped itself around these two
frameworks, demonstrative and dialectical, and subsequent Aristotelians devoted con-
siderable effort to investigating the conditions under which one or another method was
most appropriately deployed. Accordingly, it became standard to register the different
senses in which one might speak of epistēmē or scientia. In Themistius’s fourth-century
commentary on the Posterior Analytics, for instance, we are told that ‘epistēmē’ is said in
two ways, broadly and strictly. The strict sense is that of the Posterior Analytics. Broadly,
in contrast, “we say that every apprehension [gnōsis] is knowledge [epistēmē], however it
comes about, whether it be of accidental things, through whatever method [tropos].”a
The first Latin commentary, that of Robert Grosseteste in the 1220s, distinguishes four
ways of speaking of scientia: he begins with Themistius’s very broad sense but arrives at
the strictest sense only after registering two intermediary levels, one for natural science
(where conclusions hold only for the most part) and one for mathematics (where
Grosseteste thinks we lack a grasp of the reason why).b In one form or another, this
four-fold division runs all through subsequent scholastic thought.8
If its goals are adjusted to fit the capacities of real agents and real circumstances, an
idealized epistemology might seem to embrace a purely descriptive account of human
cognitive activities. But in fact the theory is normative in its ambitions. In seeking to
establish the cognitive ideal, the theory aims at a question that lies at the heart of
epistemology: the question of when ordinary agents, in ordinary circumstances, should
believe the things they believe. This way of proceeding insists that a normative account
of our epistemic position, nonideal as it is, presupposes some conception of the
ideal. Such a methodology—ideal theory as foundational for real-world applications—
is familiar enough in other normative domains. In political philosophy, for instance, it is
common to frame a theory of justice around an account of what an ideally just state
would look like. The project is not, of course, to describe a form of government suitable
only for the gods. Nor is it supposed that the only just state would be one that perfectly
satisfies the ideal theory. The goal is an understanding of what a just state would be for
beings such as we are, in a world like this one. With some such picture of the ideal,
calibrated against what is actually possible, we are able to think about what sort of
political structures we might reasonably demand. This is precisely what we find in an
idealized epistemology: it begins with a conception of the human cognitive ideal, then
applies it to the question of what we ought to believe.9
a b
Paraphrasis Post. an. I.2 (ed. Wallies, p. 5; ed. O’Donnell, p. 247). Comm. Post. an. I.2, p. 99.
Normative Ideals 9
In many ways, this is a more promising approach than what one finds in epistemology
today. Rather than describe an ideal and then consider how close we might come to
achieving it, the modern epistemologist has tended to begin with questions of threshold:
exactly what divides knowledge from mere true belief ? It would be as if political
philosophers spent most of their time trying to define exactly where the borderline
falls between the just and the unjust state, or as if ethicists focused on just precisely how
good an act must be in order to count as praiseworthy. To be sure, there will be cases in
the moral or political domain where such questions of threshold have practical rele-
vance. But it would be odd to expect clear lines of demarcation, and odd to think that the
principal task of normative theory is to discover those lines. In epistemology, too,
boundary conditions clearly matter. We care about the theory of knowledge largely
because we care about what we ought to believe and what we ought to do, and we think
such questions of belief and agency are tied up with questions about what we know.
Next week we will see why, beginning in the seventeenth century, epistemology
becomes increasingly focused on whether our beliefs are justified. Once the issues are
so starkly normative it becomes natural, as we will see, to look for the boundaries
that demarcate knowledge. Even so, it is strange to embrace the widespread current
assumption that the way to investigate such questions runs through language—as if
finding necessary and sufficient conditions for the word ‘knowledge’ would show us
what the proper standards are for belief and action.a A more meaningful way to proceed
is to begin with an account of the epistemic ideals toward which we ought to aspire and
then to reflect on how much progress toward those ideals we should demand of
ourselves in one or another domain.
Like other normative disciplines, an idealized epistemology holds out the hope not
only of clarifying our actual practices, but of putting us in a position to critique those
practices. Mill’s utilitarianism describes an ethical ideal, maximizing happiness, which
does not immediately tell us exactly how much happiness an action must produce to
count as morally good. All the same, even if Mill’s theory does not mark the precise
boundaries of right action, it has tremendous potential to influence society’s conceptions
of where those boundaries should be drawn, simply by winning people over to a new
conception of the moral ideal. Something similar might happen in epistemology.
A pessimistic conception of our epistemic prospects—of the sort we will encounter in
Lecture Six—might encourage a more tolerant attitude toward belief. Instead of scorn-
ing those who hold various religious, ethical, and political views upon insubstantial
evidence, we might indulgently regard such naïfs as being on more or less the same
footing as everyone else. In contrast, optimism regarding the attainability of some
epistemic ideal might lead us toward higher expectations in everyday life. Indeed, this
may already have happened. William Whewell, in his nineteenth-century paean to the
progress of science, argues for the larger social consequences of the scientific revolution:
“an advance from the obscure to the clear, and from error to truth, may be traced in the
world at large.”b Whewell contends that rising standards of certainty and precision in
science caused society to elevate its cognitive expectations more generally—a clear case
of how an epistemic ideal, once seen to be realizable, might bring about a broader shift
a
Lecture Two discusses the linguistic tendencies of much recent epistemology.
b
History of Scientific Ideas I: 279.
in normative expectations. Even if Whewell’s enthusiasm is misplaced, the phenomenon

is not implausible. An idealized epistemology might be more than descriptive—it might
also have normative force. So, even if recent theorists of knowledge have only inter-
preted the word in various ways, the point might instead be to change it.10
The Breakdown of Aristotelian Essentialism

It is relatively easy to describe the absolute epistemic ideal—the sort of knowledge that a
god might have. The ideal would be to grasp, in a single glance, all of reality, with
complete certainty, and to understand all the explanatory connections between things.
Eventually, in my final lecture, I will consider whether there might be limits to this sort
of absolute ideal, even for a god. The more pressing questions of an idealized epistem-
ology, however, concern its application to this world. It is only here that the normative
dimension of the theory comes into play, since one can reasonably insist that human
beings ought to achieve a certain epistemic standing only if it is possible for us to do so.
This is what I mean by the normative epistemic ideal.
To see what sort of epistemic ideal is possible for us requires taking into account two
kinds of constraints: those imposed by our human limitations and those imposed by the
character of the world in which we live. This is perhaps part of the reason why the
theory of knowledge has not traditionally been conceived of as a distinct subject of
inquiry. Epistemology, once idealized, can be developed only as part of a much broader
theory. On one side, an account of the normative ideal must be embedded in a theory of
our cognitive capacities. Such constraints will take center stage in later lectures. Today,
however, I am concerned with those constraints that arise from the other side, from our
conception of the world around us. Such metaphysical questions have, traditionally,
gone hand in hand with epistemology. When Plato, for instance, discusses his theory of
knowledge in the Theaetetus and the Republic, he does so in the context of his theory of
Forms, which he regards as ideal objects of inquiry. Conceived of apart from the Forms,
Plato’s epistemology is unintelligible. Aristotle’s epistemology, as set out in the Posterior
Analytics, expressly sets itself against that Platonic conception of reality. In its place
Aristotle offers a more down-to-earth but still ideal object: the inner essences of things.
The details of Aristotle’s theory of knowledge depend on his essentialism just as much as
those of Plato’s depends upon his Forms.
It was, of course, the Aristotelian framework that became dominant after antiquity.
When this framework collapsed in the seventeenth century, it did so not mainly
because of doubts specific to Aristotelian epistemology, but because of doubts over
its metaphysical foundations. Specifically, seventeenth-century authors began to chal-
lenge the Aristotelian doctrine of essences. As long as that metaphysical doctrine had
thrived, so had the Posterior Analytics framework, along with its conception of the
cognitive ideal. But with the rise of the mechanical philosophy came a new metaphys-
ics, and with it a new conception of knowledge. Or, to be more precise, as we will see,
the seventeenth century gives rise to several distinct conceptions of the epistemic
ideal, engendering new meanings for both ‘knowledge’ and ‘science’ and paving the
way ultimately to our modern conception of epistemology as the foundational subject
of philosophy.
The Breakdown of Aristotelian Essentialism 11
For early chapters in this transformation, we can turn to Hobbes and Locke, authors
of the first two great philosophical treatises of the English language. To understand their
backstory, however, something should be said briefly about the history of English
epistemic vocabulary. A theory of knowledge ought not to be a theory of ‘knowledge.’
But we can perhaps disrupt epistemology’s preoccupation with language, at least for a
short while, by reflecting on the historical contingencies that led to our current patterns
of usage. When English began to take shape in its modern form, in the later Middle
Ages, various options were available for talking about our epistemic achievements.
There was our verb ‘to know,’ which is attested in Old English and is very common
throughout all periods, along with the noun ‘knowledge,’ which begins in Middle
English. Then, also going back to Old English, there is the verb ‘witen’ (from the
same root as the German wissen). Finally, beginning in the fourteenth century, English
starts to use the Latinate word ‘science.’ Each of these words might have come into
prominence as our principal way of talking about epistemic achievements. Indeed, it is
interesting to see that each of the three—‘knowledge,’ ‘witen,’ ‘science’—predominates
in one or another early translation of Boethius’s Consolation of Philosophy, which was the
first philosophical work to be rendered in English. By the early seventeenth century,
‘witen’ largely falls out of usage as a verb, whereas ‘knowledge’ and ‘science’ often seem
quite interchangeable. In some texts, we find ‘knowledge’ where we would expect to
find ‘science.’ Francis Bacon, for instance, in 1605, describes mathematics as “the most
abstracted of knowledges.”a Others use ‘science’ where now we would use knowledge,
and still others switch back and forth even within the same sentence. Joseph Glanvill, for
instance, in 1661, writes that “he is the greatest ignorant, that knows not that he is so: for
’tis a good degree of science, to be sensible that we want it.”b 11
Confronted with unsettled terminology and complex theoretical questions, philo-
sophers naturally set about constructing systematic accounts. But by the time this started
to happen in English, Aristotle’s philosophy lay under a dark cloud, disreputable even if
still enormously influential. Consider Hobbes’s Leviathan (1651). Writing in English,
Hobbes proposes that we think of “knowledge” as coming in two kinds: knowledge of
fact and knowledge of the connection between propositions. The first “is nothing else
but sense and memory.” The second is “the knowledge required in a philosopher,” and
this is what Hobbes says should be called science.c On its face, this looks entirely un-
Aristotelian, as one might expect, given that Hobbes has absolutely no sympathy for the
broader metaphysical framework of Aristotle’s epistemology. Moreover, quite unlike
Aristotle’s, Hobbes’s account seems to make science rather easy to acquire; we need
grasp only that one proposition entails another. But this turns out to be misleading,
because Hobbes goes on to express considerable doubt about whether we often manage
to achieve science. He remarks that geometry is the only science we have successfully
attainedd—ironically enough, coming from a man who would later spend years trying to
persuade the leading geometers of his day that the circle can be squared. As for the
natural sciences, they “cannot teach us our own nature, nor the nature of the smallest
creature living.”e The grounds for such pessimism are unclear from the Leviathan itself,
but can be grasped from the earlier De corpore. There, writing in Latin, Hobbes lays out a
a b
Advancement of Learning Bk. II (Major Works, p. 235). Vanity of Dogmatizing ch. 23, p. 225.
c d e
Lev. 9.1. 4.12. 21.33.
conception of scientia that turns out to be much more Aristotelian than one might have
guessed, inasmuch as science arises out of a series of demonstrations grounded in initial
definitions that establish “the causes and generation of things.”a Chains of inference that
do not bottom out in an understanding of ultimate causes do not count as science.12
Locke, the other great seventeenth-century fount of philosophical English, offers a
similar picture of the epistemic landscape. On the one hand, he has a very demanding
sense of ‘knowledge,’ which he associates with science and contrasts with a looser notion
of assent based on probabilities. (But, as we will see next week,b Locke is unlike Hobbes
in that he refuses to speak of such assent as knowledge.) Although Locke does not
expressly define ‘science’ as a distinct kind of knowledge, it seems clear that, like
Hobbes, he understands it in broadly Aristotelian terms, as requiring a grasp of the
necessary connection from causes to effects. Also like Hobbes, Locke despairs of our
being able to achieve any such thing, writing: “As to a perfect science of natural bodies
(not to mention spiritual beings) we are, I think, so far from being capable of any such
thing, that I conclude it lost labour to seek after it.”c Locke’s discussion of these matters
goes one step further, however, inasmuch as his skepticism about the prospects for
science results from a sophisticated refashioning of the Aristotelian doctrine of essences.
Whereas Hobbes had retained the causal requirement without an Aristotelian meta-
physics of essence, Locke maintains at least a vestige of that metaphysics, taking for
granted throughout his writings that there are some such “real essences” or “real
constitutions” that both define what a substance is and explain its various accidental
features. Thus, he says, “’tis past doubt, there must be some real constitution on which
any collection of simple ideas co-existing must depend.”d To have scientific knowledge
of the natural world, then, would be to grasp these ultimate causal principles. But Locke
completely rejects our ability to do any such thing: “we in vain pretend to range things
into sorts, and dispose them into certain classes, under names, by their real essences, that
are so far from our discovery or comprehension.”e Our conventional groupings of things
into kinds—what he calls their nominal essences—have little connection to the actual
distribution of real essences.
If Locke’s view of real essences had been merely skeptical, then not much would have
followed for the Aristotelian idea of scientia. After all, even for the Aristotelians, the
essences of things are merely an ideal goal of inquiry. But Locke’s account implies that
essences are a false ideal—that we do not live in a world that clusters so neatly into
kinds, and that what clusters there are may not reflect any deep metaphysical structure.
At this point Locke might have articulated some other epistemic ideal, which is in effect
what both Descartes and Newton did, albeit in very different ways, as we will see later
today and next week. Instead Locke despairs of our achieving anything that would count
as an epistemic ideal. Science, for him, is the epistemic ideal, and to possess it would be
to grasp the real essences of things and to understand why those essences give rise to the
various qualities and operations that we observe in the world around us. Yet, because of
the “darkness we are involved in,” such “connections and dependencies [are] not
discoverable in our ideas.” Accordingly, “we are so far from being able to comprehend
the whole nature of the universe, and all the things contained in it, that we are not
a b c
De corpore 6.13. Chapter Two, p. 41. Essay IV.3.29.
d e
III.3.15. III.6.9.
The Breakdown of Aristotelian Essentialism 13
capable of a philosophical knowledge of the bodies that are about us, and make a part
of us.”a In part he blames the world for not having a structure that is readily discernible,
and in part he blames our cognitive faculties: “the weakness of our faculties in this state
of mediocrity . . . makes me suspect that natural philosophy is not capable of being made
a science.”b Locke is enthusiastic about the achievements of a few contemporaries: he
describes himself as living “in an age that produces such masters as the great Huygenius,
and the incomparable Mr. Newton.”c But, by Locke’s standards, even work such as this
falls short of the ideal. We can see what that epistemic ideal is, Locke thinks, and we can
see its unattainability.13
Hobbes and Locke stand at the brink of post-scholastic epistemology. Having reject-
ing Aristotelian metaphysics, they are unable to engage in the epistemic program that
Aristotle had described. Yet, even so, they continue to accept that program as an ideal,
insofar as they accept that genuine knowledge or science would require a grasp of the
ultimate causes of things. That they hold such a view is really no surprise, because they
thus follow the virtually unanimous verdict of the philosophical tradition. We have seen
how the causal requirement appears in Aristotle’s canonical definition of epistēmē.
Indeed, Aristotle remarks that “study of the reason why is what reigns supreme in
knowledge.”d From antiquity through the Middle Ages and well into the seventeenth
century, this requirement gets taken for granted. Plato had spoken of the need to grasp
the “legitimate cause and reason” of natural phenomena.e From Peter Abelard to Pierre
Gassendi, authors invoke Vergil: “Happy is he who has been able to grasp the causes of
things.”f When Aquinas comments on Aristotle’s definition, he first insists that scientia is
perfect cognition, then adds that “one who has scientia, if he is cognizing perfectly, must
cognize the cause of the thing of which he has scientia.”g Throughout the sixteenth
century, and even among many of the most anti-Aristotelian seventeenth-century
authors, this ideal remained firmly in place. Francis Bacon offers the epigram that “to
know [scire] truly is to know through causes.”h On these grounds Descartes criticizes
Galileo’s crowning masterpiece, the Two New Sciences: “without having considered the
first causes of nature, he [Galileo] has searched only for the reasons beyond various
particular effects, and so has built without any foundation.”i 14
Hobbes and Locke both follow this traditional conception of the epistemic ideal, but
with the crucial difference that neither thinks that we can satisfy this ideal, not even in
principle. What drove them to this conclusion was not some sort of generalized skeptical
crisis or broad scruples over causality. Rather they came to have doubts of a much more
specific kind, regarding Aristotelian essentialism. That doubts over essence should lead
to despair over causal understanding may be surprising, but in fact it is a characteristic
feature of late scholastic Aristotelianism to think of causal explanation almost exclusively
in terms of grasping the essences of things. Strictly speaking, to be sure, the causal
demand applies to all four Aristotelian causes: not just to the “internal” causes, material
and formal, but also to the “external” causes, efficient and final. But scholastic philoso-
phy put little weight on efficient causes as explanatorily significant and treated matter as
a b c
IV.3.29. IV.12.10. Essay epistle.
d
Post. An. I.14, 79a24. e
Timaeus 28a.
f
Abelard, Logica “Nostrorum,” pp. 505–6; Gassendi, Syntagma II.I.4.1 (I: 283a)—both quoting Vergil, Georgics ii.490.
g h i
Comm. Post. an. I.4, n. 5. Novum organum II.2. To Mersenne, 1638 (II: 380).
simply the enduring background frame over which change occurred. Moreover, despite
the famous controversy that surrounds final causes, in truth these too were rarely given
much weight in scholastic natural philosophy. The governing program, instead, was to
understand a thing’s inner nature—its essential qualities and, above all, the substantial
form that gave a substance its coherence and enduring character. Thus Henry
Oldenburg, first secretary of the Royal Society, blamed substantial forms for having
single-handedly “stopped the progress of true philosophy.”a Even if formal causes were
just one of the four scholastic types of causes, they still dominated theoretical inquiry,
quite overshadowing material, efficient, and final explanations. Such an attitude endures
even in Pierre Gassendi, who betrays his basically scholastic outlook on explanation
when he remarks that “not much work needs to be expended on grasping the external
causes, which experience itself and the senses reveal at once. The difficulty lies with the
internal causes. Those who investigate these are rightly said to be searching deep into
the secrets of nature.”b For Gassendi, this characteristically scholastic program could still
be carried forward within the atomistic Epicurean framework, shorn of Aristotelian
forms. But when later seventeenth-century authors came to have fundamental doubts
about essential explanations, the whole program came to seem a false ideal.15
And here is where things get really interesting. Hobbes and Locke have a traditional
conception of the epistemic ideal and propose reserving the Latinate English word
‘science’ to pick this ideal out. If this usage had stood, it would have condemned “science”
to a marginal existence, as the sort of knowledge we might imagine in our dreams of the
life to come. We could praise the efforts of Newton and others for their practical benefits,
but not for having attained the level of science. As Locke puts it, “how far soever humane
industry may advance useful and experimental philosophy in physical things, scientifical
will still be out of our reach.”c What in fact happened, however, is that the great
seventeenth-century figures whom we now think of as scientists, rather than despairing
of success or reconciling themselves to mere practical benefit, articulated a new, post-
Aristotelian conception of the epistemic ideal, one that relinquished the goal of causal
understanding grounded in a grasp of essences. This is the next stage in our story.
Trading Depth for Precision

The history of modern science’s turn away from causal explanation has been told many
times and been much disputed, but it can be understood somewhat more clearly in light
of the historical background we have been surveying. A key transitional figure is Galileo.
In his third letter on sunspots (1612), he writes that “in our speculating we either seek to
penetrate the true and internal essence of natural substances, or content ourselves with a
knowledge of some of their properties [affezioni].” He goes on to set aside the ideal of
grasping essences, judging it impossible for us to have such knowledge, and in its place
he extols inquiry into the mere properties of both earthly and celestial phenomena:
“location, motion, shape, size, opacity, mutability, generation, and dissolution.”d On its
face, this is not so far from what we have seen Hobbes and Locke later say, but what is
a b
Correspondence III: 67. Syntagma II.1.4.1 (I: 284a).
c d
Essay IV.3.26. Discoveries, p. 123 (Opere V: 187–8).
Trading Depth for Precision 15
new here is Galileo’s enthusiasm for a method that simply lets go of the old ideal. It is
not that Galileo denies the desirability of understanding the causes of the phenomena
under investigation. Indeed, one can sometimes find him speaking of such knowledge as
his goal, and even boldly claiming to have achieved it. What motivates his new attitude,
then, is neither metaphysical scruples over causation nor any principled methodological
hostility toward causal explanation. Instead, his complaint is that philosophy has pitched
its ideal at the wrong level, not because causal explanation is not desirable but because,
in many domains, such explanations are not to be reasonably expected. As he remarks of
his theory of comets in The Assayer:
I should not be condemned for being unable to determine precisely the way in which comets are
produced, especially in view of the fact that I have never boasted that I could do this, knowing
that they may originate in some manner that is far beyond our power of imagination.a
Philosophy is damaged, Galileo thinks, by demanding a goal that, often, cannot be
achieved. Instead of pushing ourselves into speculation over causes we cannot under-
stand, we should celebrate our ability to grasp the “properties” of bodies and the rules
that govern them.16
In effect, Galileo is urging a recalibration of our epistemic ideal. This approach would
go on to find its most illustrious proponent in Isaac Newton. In a famous query from the
Opticks (Latin edition of 1706), he writes:
To tell us that every species of things is endowed with an occult specific quality by which it acts
and produces manifest effects is to tell us nothing. But to derive two or three general principles of
motion from phenomena, and afterwards to tell us how the properties and actions of all
corporeal things follow from those manifest principles, would be a very great step in philosophy,
though the causes of those principles were not yet discovered; and therefore I scruple not to
propose the principles of motion above mentioned, they being of very general extent, and leave
their causes to be found out.b
It would be hard to overstate the dramatic shift that this method represents, when
contrasted with the prevailing approach of the Aristotelians. Aristotle had begun his
Physics with the injunction to seek causes:
When the objects of an inquiry, in any department, have principles, causes, or elements, it is
through acquaintance with these that knowledge [eidenai] and understanding [epistasthai] is
attained. For we do not think that we know [gignōskein] a thing until we are acquainted with
its primary causes or first principles, and have carried our analysis as far as its elements.c
The contrast is striking. The expectation Aristotle had announced with the very first
words of his treatise would set the agenda in natural philosophy for two millennia. With
Newton, that agenda collapses. The “two or three general principles” that he offers
make no claim to take us all the way to Aristotle’s “primary causes or first principles”;
instead they merely describe “the phenomena.”
Do Galileo and Newton truly renounce the search for causes? Sometimes they talk in
this way, but elsewhere they seem to be still searching. Their attitudes become clearer
when we think in terms of epistemic ideals. Neither author believes that a grasp of
a b
Discoveries, p. 258 (Opere VI: 281). Opticks, query 31, pp. 401–2.
c
Physics I.1, 184a10–14.
causes is to be despised. Everyone would ideally wish to have such knowledge, if it were
available. Nor is their point merely that such knowledge is not forthcoming any time
soon. Everyone, even the most doctrinaire scholastics, acknowledges that much. Instead,
what is new here is an enthusiasm for a new research project, one that neither resorts to
fanciful causal speculation nor wallows in skeptical resignation. In part, what is new is
the insight that there is quite a lot of work to be done in framing general principles to
account for the phenomena. But of equal importance is the further thought that such
work is a worthy end of inquiry. Galileo’s Two New Sciences (1638), from the end of his
career, remarks:
The present does not seem to me to be an opportune time to enter into the investigation of the
causes of the acceleration of natural motion, concerning which various philosophers have
produced various opinions . . . Such fantasies, and others like them, would have to be examined
and resolved, with little gain. For the present it suffices our Author that we understand him to
want us to investigate and demonstrate some attributes [passioni] of a motion so accelerated
(whatever be the cause of its acceleration).a
What Galileo is accomplishing “suffices,” as he puts it, for present purposes. This is a
normative claim about what human beings ought to be doing—indeed, what their
“Author,” God, wants them to be doing—by way of understanding the world around
them. In the next life our aspirations can be higher. But, for now, in many domains, this
is the appropriate normative ideal. Newton speaks in strikingly similar ways in the
second edition of the Principia (1713). Famously abjuring the temptation to “feign
hypotheses,” he writes that “it is enough that gravity really exists and acts according
to the laws that we have set forth and is sufficient to explain all the motions of the
heavenly bodies and of our sea.”b “It is enough,” Newton says, not because there would
be no value in learning more, but because in this life, with respect to much of the world
around us, this is what we are capable of.17
For Newton in particular, an important part of what leads him to regard his approach as
worthy of the appellation ‘science’ is that he sets for himself an alternative epistemic ideal,
that of precision. In effect, what Newton gives up in speculative depth he compensates for
through the precise accuracy of his mathematical methods. His brief preface to the
Principia gives pride of place to this quest for accuracy: “Anyone who works with less
accuracy is a less perfect mechanic, and anyone who could work with perfect accuracy
would be the most perfect mechanic of all.”c As an ideal, this is not wholly new. Plato had
put considerable weight on the need for an epistēmē to be precise, and Aristotle, too, had
spoken of this desideratum in various places. But Aristotle had expected full precision to be
possible only in mathematics: “the precise reasoning of mathematics is not to be demand-
ed in all cases, but only in the case of things that have no matter. Therefore its method is
not that of natural science.”d For Newton, in contrast, the ideal is one that can be insisted
on across the sciences. If the natural sciences fail to be precise, “the errors belong not to the
art, but to its practitioners.”e Even in the formerly second-rate mechanical sciences,
Newton thinks, the highest ideal can be achieved. In place of philosophical depth, he
offers the precision of mathematics.
a b
Third Day, pp. 158–9 trans. (Opere VIII: 202). Principia, General Scholium, p. 943 trans.
c d
Principia, Preface, p. 381 trans. Metaph. II.3, 995a15–16. e
Principia, Preface, p. 381 trans.
Trading Depth for Precision 17
This transposition of ideals is not coincidental. Scholastic philosophers went deep into
identifying the substantial forms and elemental qualities that ground the natural world,
but in so doing they made precision impossible, because they had postulated the
existence of entities they were unable to characterize in any sort of accurate detail. In
this respect one might say that their method in natural philosophy recapitulated their
theology, where they likewise postulated ultimate entities (God and angels) about which
very little could be said with any precision. For Newton, as a natural philosopher, to
postulate such hidden principles is (as above) “to tell us nothing” (though famously,
when it came to theology, Newton took a much more indulgent view). Inevitably, there
is a trade-off here. We can seek precision about what lies close to the surface; or we can
aspire, inchoately, to the murky depths. For us, just as much as for Newton and his
scholastic predecessors, the goal of a fully precise account that goes all the way down
remains a distant, merely absolute ideal. When it comes to the sort of ideal that might
have normative force for our epistemic practices, we have to make a choice between the
competing ideals of depth and precision.18
By the early eighteenth century it was clear that Galileo and Newton would carry the
day, and the term ‘science’ would eventually come to be associated with the modes of
inquiry they pursued rather than with the remote absolute ideal to which Hobbes and
Locke had attached it. So, as the story goes, modern science is born. We rightly celebrate
this development for the way it saved science from the limitations of speculative
metaphysics, and we rightly see Galileo and Newton as its founding figures. If, in
comparison, one reads William Gilbert’s De magnete (1600), it is clear that something
importantly new is happening. In place of armchair speculation, there is a striking
emphasis on observation grounded in experiment. But Gilbert’s conclusions in the
end still strike a modern reader as something of an embarrassment, because he persists
in taking as his ideal an account of the underlying causal explanation for magnetism; and,
to satisfy that desideratum, he can do no better than to propose that magnets possess a
soul. Similar complaints might be made—and were constantly made—about many
stretches of Descartes’s scientific writing: even if his explanations are wholly mechanical,
still those explanations serve as the vehicle for speculative causal explanation. No
wonder that (as we have seen) Descartes found Galileo’s method so alien.
Yet, though we may admire the modern renunciation of causal depth and the
precision that such a renunciation makes possible, it must be said that it exacts a price.
For every Gilbert and Descartes whose work might have been improved as a result of
less metaphysical speculation and more attention to precise data, we can cite in contrast
a figure whose abjuration of explanation leads to absurdity. Consider Joseph Glanvill. In
his Vanity of Dogmatizing (1661), he takes up the great question of why anointing a
weapon helps cure the wound that the weapon had previously inflicted. Having learned
from his peers at the Royal Society how science is to be done, Glanvill magisterially
declines to frame hypotheses. “It is enough for me that de facto there is such an
intercourse between the magnetic unguent and the vulnerated body, and I need not
be solicitous of the cause.”a Here, one may protest, we have neither explanatory depth
nor precision. Sometimes the new method was clearly not “enough.”19
a
Vanity of Dogmatizing ch. 21, p. 208.
Even so, carried along by its successes, post-scholastic science made a virtue out of
resignation. Locke had thought that the lesson of post-scholastic philosophy was that
natural science is impossible, which is to say that, in the domain of nature, human beings
simply cannot achieve their epistemic ideal. Events would develop otherwise, however,
not by showing Locke to be wrong about what we are capable of, and not by reframing
science as a nonideal discipline, but instead by reconceiving the ideal. Inspired by the
brilliant examples of Galileo and Newton, natural philosophers by and large embraced
a diminution of ambitions. They did so, in general, without any sense of having
succumbed to skepticism, or even without any sense of having dethroned science
from its traditional status as the ultimate goal of inquiry. By the time of David
Hume’s Treatise of Human Nature (1739), ‘science’ and ‘knowledge’ had so parted ways
that, even while Hume judges knowledge to be largely unattainable, he takes science as
the governing ambition of the whole project: “In pretending therefore to explain the
principles of human nature, we in effect propose a complete system of the sciences, built
on a foundation almost entirely new, and the only one upon which they can stand with
any security.”a Within this new framework science could carry on, radically transformed
from the epistēmē of Aristotle or the scientia of the scholastics but still heir to those
notions, inasmuch as science remained the normative goal of inquiry, the epistemic ideal
to which human beings might aspire. So the scientific revolution begins from a
revolution in our cognitive expectations.20
Epistemology in the Ascendant

If this was to be the fate of science, then where does it leave philosophy? Next week we
will consider how, by the end of the seventeenth century, philosophers began to
reconceptualize knowledge in response to a rising interest in probability. But even
before looking into the changing fortunes of that concept, we are in a position here to
consider the fate of epistemology in the modern era. The rise of modern science stripped
philosophy of a large part of its traditional core, those fields once known as natural
philosophy but now rechristened as science. Shorn of the new sciences, philosophy faced a
choice. On the one hand, it could turn its back on the new scientific ideals, or at least
insist on its autonomy in relation to science. In that case philosophy might continue its
pursuit of deep causal explanations, not as a chapter of science but as metaphysics.
Leibniz is the shining early example of this sort of approach, and his efforts at a
metaphysics that could transcend natural philosophy would shape German philosophy
through Kant and beyond. On the other hand, philosophy might embrace the new
scientific conception of the epistemic ideal, and hence begin to assume a similar modesty
regarding conjectural causal explanations. The early champion of this approach, which
sets itself against the speculative metaphysics of the scholastics, was Locke. In one of his
most familiar passages, he pronounced it “ambition enough to be employed as an under-
labourer in clearing ground a little, and removing some of the rubbish that lies in the
way to knowledge.” Again, it is “enough” to do this, not because more would not be
better, but because this is the most we should suppose possible. The ambition to do
a
Treatise, Introduction, p. xvi. On Hume’s skepticism, see Lecture Six.
Epistemology in the Ascendant 19
more has made philosophy hitherto, in Locke’s biting words, “the sanctuary of vanity
and ignorance.”a
With this dilemma in mind, we can return to the question with which I began this
lecture. Why is it that epistemology—after so many centuries of not being a discrete
subject at all—has now become a foundational subject within philosophy? The answer is
that we have followed Locke’s path through the dilemma, inasmuch as philosophers in
the Anglo-American tradition have largely deferred to science on questions of what the
world is like and why. Thus the dominant philosophical tendency for three hundred
years, at least in the English language, has been to concentrate on those subjects that
once comprised logic in its broad traditional sense: the study of knowledge, language,
and patterns of inference. In place of explanatory depth, philosophy has come to prize,
above all else, precision. As the young Wittgenstein wrote, “[t]he aim of philosophy is
the logical clarification of thoughts . . . Philosophy should take the thoughts that are
otherwise cloudy and blurred, as it were, and make them clear, giving them sharp
boundaries.”b Once subservient to theology—the reigning medieval project of under-
standing the ultimate reasons why—modern philosophy has made itself handmaid,
underlaborer, actuarial clerk to the scientist. As Voltaire, under the sway of the new
English method, succinctly put it, “[p]hilosophy consists in stopping when the torch of
science fails us.”c So it has been, more or less, through Hume and Whewell, Mill and
Moore, Davidson, Dummett, and Quine.
Except that, in recent years, there have been signs of change. Although long accus-
tomed to the diminished Newtonian ideal of precision over explanatory force, philo-
sophers of the past several decades have been increasingly unwilling to embrace their
supporting role in that enterprise. Wittgenstein himself exemplifies one path of resist-
ance, when in his later works he calls into question his youthful aspirations to logical
precision: “we eliminate misunderstandings by making our expressions more exact; but
now it can look as if we are moving towards a particular state, a state of complete
exactness; and as if this were the real goal of our investigation.”d The main line of
Anglo-American development, however, has been not to challenge the ambition of
scientific precision, but rather to disavow the metaphysical resignation of earlier gener-
ations. And so in recent years there has been a recrudescence of speculative metaphysics,
which has again become the central philosophical topic of our day. In keeping with this
trend, historians of philosophy have increasingly found themselves drawn not to Locke
and Hume, but to the metaphysically more adventuresome work of Spinoza and
Leibniz, and even to the medieval metaphysics that Hume had long ago consigned to
the flames.
Predictably, as philosophers have returned to their speculative enthusiasms of old,
scientists have begun to complain in an increasingly acerbic voice. Stephen Hawking
declares that “philosophy is dead.”e Freeman Dyson, reviewing a book on that most
ambitiously philosophical question of all—Why is there something rather than nothing?—
calls philosophy today “a toothless relic of past glories.”f Why such belligerence? For as
long as philosophy played its subsidiary role, content to serve as handmaid to the
a b
Essay epistle, p. 10. Tractatus 4.112.
c
Philosophical Dictionary, “Soul,” VI: 168. d
Philosophical Investigations I.91.
e
Hawking and Mlodinow, Grand Design, p. 5. f
“What Can You Really Know?”
restrained ideals of modern science, the scientists have been willing to tolerate the
philosopher’s ancient pretensions to wisdom. But if the philosophers are not going to
abide by the old Newtonian code of resignation—if they are going to frame hypotheses
about reality in all its depths—then it becomes necessary for the real lovers of wisdom,
with their laboratories and government grants, to speak out against these pretenders.
The philosopher, having given birth to and fostered the various sciences, is now sent off
into the woods to die, yielding her lectureship and office space to worthier academic
enterprises.
But perhaps we need not rush quite yet to throw up barricades between the arts quad
and the science towers. The history of philosophy has its consolations, one of which is a
vivid awareness of the way in which, in one form or another, this power struggle
between science and philosophy is nothing new. Going back to antiquity, champions of
an austerely empirical conception of inquiry have battled against proponents of a
speculative rationalism. So it was between Plato and his unspeakable rivals, the Preso-
cratic atomists; so it was between Newton and Leibniz. At issue, I have been arguing, is
the nature of our epistemic ideals. What sorts of demands do we make of ourselves as
we investigate the world around us? An answer to this question depends in turn on still
deeper puzzles about what this world is like. If we suppose that all there is is particles in
motion, subject to forces of various sorts, then the great task of speculative inquiry must
be to measure those forces as accurately as possible. But what if there is more to the
world than this? What if we live in a world with beings whose agency runs beyond
anything hitherto imagined by science? What if these beings are sensitive to values
that cannot be weighed in any laboratory? Who, then, will measure these things, and
how much precision may we demand? Of course, these glimmers of transcendence
are perhaps just illusions, destined to be assimilated to the ever-growing reductive
empire of science. Yet who will decide when that reduction has succeeded? How will
we know?21
Lecture Two
Evident Certainties
Last week I described the project of an idealized epistemology, focused not on the
dividing line between what is and what is not knowledge, but on the ideal that agents
like us might hope to achieve in a world like this. Following the course of this project,
from Aristotle up to the seventeenth century, I described how modern science articu-
lated its own epistemic ideal, one that parted ways in crucial respects with the traditional
concerns of philosophy. Deferring to that scientific ideal, philosophers in modern times
have tended to conceive of their discipline as essentially actuarial, balancing the ledgers
as science rushes onward. On this new conception of philosophy, epistemology became
newly ascendant.
But what exactly is epistemology? As fundamental to philosophy as this subject has
become, it is surprisingly unclear what it is centrally about. Today I try to explain how
our modern conception of epistemology grew out of the decline of scholasticism.
Lexicology
For the past half-century, epistemology has primarily aimed at conceptual analysis, first
and foremost the analysis of knowledge itself. Its aim is conceptual analysis, but in truth its
principal concern has often seemed linguistic: under what conditions is it true to say, in
English, that someone knows something? Of course, one is free to choose between the
formal and the material mode of exposition, but the modern field of epistemology in
particular has often seemed to labor under the formal yoke, wishing to understand
the structure of our central epistemic concepts but settling for the conventions of our
ordinary epistemic vocabulary. Something about epistemology as it is currently con-
ceived of seems to lure its practitioners into the trap of lexicology.
Epistemology is prone to lexicology because it lacks other sorts of stable grounds on
which to build. Philosophers of mind do not belabor the ordinary English meaning of
‘consciousness’ but rather begin by disambiguating, and thereafter make no fuss over
the word itself. It is quite unclear, in contrast, whether ‘know’ is similarly ambiguous.
Theories of causation, likewise, do not care about the word ‘cause’—they take as their
starting point various paradigm cases and try to make those come out right. It is not so
clear, however, what counts as a paradigm of knowledge, inasmuch as the most familiar
cases, such as 2 + 2 = 4 and this is a hand, are notoriously problematic and heterodox in
22 Lecture Two: Evident Certainties
character. Unable to gain traction in these sorts of familiar ways, and yet looking for the
boundaries that demarcate the domain of knowledge, epistemology tends toward
lexicology. Few want such an outcome, but the modern history of the subject repeatedly
displays this pattern, and it is unclear what alternative there is.
In all of this, epistemology suffers from an unfamiliarity with its history. In taking as its
central mission the carving of boundaries for knowledge, epistemology cuts itself off from
the main ways in which past philosophers have conceived of the field. A better under-
standing of this history and of its orientation toward the epistemic ideal would offer at least
some stable ground from which epistemologists might escape from mere lexicology. So
I argued last week. But this history also points the way toward something more. For, as we
will see today, there are reasons why epistemology came to be concerned more with
defining the scope of knowledge than with the epistemic ideal. Indeed, in the end, we will
see that these reasons are not just a matter of historical accident but in fact help to
vindicate, at least in part, our modern preoccupation with boundary conditions.
Still, to understand all this properly, one needs to see how the history goes. Last week,
the main case study was Aristotle, and I argued that the Posterior Analytics’ theory of
epistēmē is not quite science, not quite understanding, and certainly not knowledge.
What it is, instead, is an account of our epistemic ideal. This week I begin with Descartes
and argue that he, too, should be read as offering not a theory of knowledge but an
idealized epistemology, with this dramatic difference: that certainty is now central to the
epistemic ideal. Not that Descartes marks the start of this tradition; on the contrary, he
comes near its end. Indeed, it is the failure of his approach that leads to the first
manifestations of epistemology in something like its modern sense. The crucial devel-
opments are the rejection of certainty as a normative epistemic ideal and a growing
tolerance for the merely probable. With this there emerges a new-found interest in the
question of just how much probability is good enough. Good enough, that is, to justify
us in our beliefs and thereby to give us knowledge.1
Descartes’s Ideal Theory

One hardly needs the background story of last week’s lecture to see that Descartes takes
scientia as the goal of inquiry and conceives of it in a way that is much indebted to the
Aristotelian framework for epistēmē. To be sure, he rejects large swaths of what one finds
in the Posterior Analytics, such as the syllogism and the restriction to universal premises,
to say nothing of his rejection of the inner essences that ground Aristotelian explanation.
But Descartes holds on to the general framework of an epistemic ideal at which
theoretical inquiry should aim. One can find this assumption in place from his very
earliest work. According to Rule 2 of his Rules for the Direction of the Mind, “we should
attend only to those objects of which our minds seem capable of having certain and
indubitable cognition.”a He then immediately remarks, in discussing this rule: “All
scientia is certain and evident cognition.” A few lines below, he adds this: “So, in
accordance with this Rule, we reject all such merely probable cognitions and resolve
to believe only what is perfectly cognized and what cannot be doubted.” These passages
a
Rules 2 (X: 362).
Descartes’s Ideal Theory 23
suggest that scientia is the kind of cognition that is certain, evident, and indubitable. He
goes on in this same section to characterize this and other rules as ones that “will help
us ascend to the peak of human cognition.”a It seems, then, that scientia is perfect
cognition—or at least as perfect as a human being can achieve.
This impression receives confirmation at the start of the Meditations, when Descartes
makes this famous pronouncement: “I realized it was necessary, once in the course of
my life, to demolish everything completely and start again right from the foundations if
I ever wanted to establish anything in the sciences that was stable and likely to last.”b The
italicized words announce Descartes’s goal: attaining stable and lasting results in the
sciences. Of course he does not here have in mind “the sciences” in our modern sense of
the term; he is referring instead to the acquisition of scientia, and his ambition is to show
how we can acquire scientia even in the hitherto murky domain of God and soul. His
view, indeed, turns out to be that scientia is possible especially in these domains. The
standard he holds himself to is remarkably high. In a letter to his disciple Regius written
in May 1640, just after he had completed the Meditations, Descartes distinguishes
between scientia and the conviction (persuasio) possessed by someone who cannot help
but assent to the clear and distinct perception of some self-evident truth. “I distinguish
the two as follows: there is [mere] conviction when there remains some reason that
could lead us to doubt; scientia is conviction based on a reason so strong that it can never
be shaken by any stronger reason.”c Conviction is a state in which one cannot refrain
from assenting to a proposition—in a purely subjective sense, the proposition is
indubitable. Presumably, such conviction will be based on some reason. But it will not
count as scientia unless it is based on unshakable reasons. This would seem to be
indubitability in a stronger sense: it is not just that one is not presently able to doubt
the proposition, but that there is no way in which one will ever be able to doubt the
proposition, given the reasons one has for it. As he puts it in the Second Replies, “no
cognition that can be rendered doubtful seems fit to be called scientia.”d
These passages point toward the two best known characteristics of Descartes’s
conception of scientia: its emphasis on certainty and its foundationalism. The theory is
foundationalist in the sense that, for a belief to count as scientia, its truth must either be
certain in its own right or else be grounded in something else that is certain in its own
right. What exactly it means for a belief to be certain is one of the central issues to be
explored today. Setting that issue aside for the time being, we should notice a third
feature of Descartes’s account: its internalism, which is to say that the possession of
certain, foundationally structured beliefs does not yield scientia unless the believer grasps
the reasons that show beyond doubt the truth of the beliefs. This last feature is
particularly clear in the following passage, where the requisite adequacy of reasons
gets expressed in interpersonal terms.
Whenever two persons make opposite judgments about the same thing, it is certain that at least
one of them is deceived, and it seems that neither has scientia. For if the reasoning of one of them
were certain and evident, he would be able to lay it before the other in such a way as eventually
to convince the other’s intellect as well.e
a b
Rules 2 (X: 364). Med. 1 (VII: 17), emphasis added.
c d e
III: 65. VII: 141. Rules 2 (X: 363).
As before, the evidentness that Descartes requires for scientia must be not just subject-
ively persuasive, but objectively good, one test of which is whether these reasons would
be able “eventually to convince” others. What we can now also say is that those reasons
must be possessed by the agent—they must be internal to her. It is not enough that she
be able to acquire reasons in principle, or even that she be able to understand them if
they were shown to her. Instead, she must have in hand her reasoning to such a degree
that, when she finds herself locked in disagreement, she can “lay it before the other” and
thus dissolve the conflict. Descartes regularly expresses the utmost confidence that his
views will pass just this test, once they are understood. Thus he tells his disciple Regius:
“I consider my opinions to be so certain and evident that whoever rightly understands
them will have no occasion to dispute them.”a
Unlike Aristotle, Descartes is not ordinarily read as advancing a theory of science.
Instead, though with no more plausibility, he is routinely thought to be advancing a
theory of knowledge, and hence made into the archetypical proponent of foundation-
alism, certainty, and internalism in epistemology. All of this is true enough, but only
with respect to Descartes’s ideal theory. As for what is commonly said about his theory
of knowledge, most of it is entirely wrong. If our subject is ‘knowledge’ as that word is
used today, then Descartes is neither a foundationalist nor an internalist, not even a strict
advocate of certainty. Indeed, if epistemology is conceived of in its usual modern guise,
then Descartes cannot be said to have a theory of knowledge at all. What he has is an
idealized epistemology, a theory of scientia.
Two sorts of considerations make this quite clear. First, the theory is so demanding
that virtually no one other than Descartes and his followers can be said to have achieved
scientia. Descartes in fact claims that, up until his time, the only scientia possessed by
anyone has been mathematical scientia. For instance, in explaining why we should attend
only to what we can cognize with certainty, he remarks that, “if my reckoning is correct,
out of all the sciences so far devised, we are restricted to just arithmetic and geometry if
we stick to this Rule.”b Years later, in The Search for Truth, he speaks of “the slight
progress we have made in the sciences whose first principles are certain and known to
all” and then adds:
In the other sciences, whose principles are obscure and uncertain, those who are willing to state
their view honestly must admit that, for all the time they have spent reading many a vast tome,
they have ended up realizing that they have scientia of nothing and have learned nothing.c
Of course, Descartes thinks that he himself has managed to push the bounds of scientia
quite a bit farther. But did he really believe that, up until the middle of the seventeenth
century, no one had knowledge of anything, except for a few claims in mathematics?
A skeptic might be happy with this result, but Descartes was no skeptic, or so it is always
supposed.
In fact Descartes goes even farther. He famously holds that “the certainty and truth of
all scientia depends on the one cognition of the true God, to such an extent that I was
incapable of perfect scientia about anything else until I recognized him.”d If scientia were
knowledge, then this would entail that the atheist lacks knowledge, as, apparently,
a b
Letter of July 1645 (IV: 248). Rules 2 (X: 363).
c d
Search for Truth (X: 526). Med. 5 (VII: 71).
Descartes’s Ideal Theory 25
would anyone who believes in the wrong God. Thus, in the Second Replies, in a passage
quoted in part already, he remarks: “I do not dispute that an atheist can clearly cognize
that the three angles of a triangle are equal to two right angles. I maintain only that his
cognition is not true scientia, since no cognition that can be rendered doubtful seems fit
to be called scientia.”a Even in geometry, then, scientia is available only to someone who
has the right religious beliefs and who uses those religious beliefs in just the right way to
ground that scientia. All things considered, it looks doubtful that anyone other than
Descartes (and his most devoted followers) has ever had scientia about anything. If we
think of Cartesian scientia as knowledge, then we should think of Descartes as a fairly
radical kind of skeptic. Insofar as that result seems obviously wrong, we should stop
thinking of scientia as knowledge.
The second reason for denying that Cartesian scientia is knowledge is that Descartes
denies that scientia should regulate belief. As observed last week, one of the reasons for
caring about boundary conditions in epistemology is that what we know seems closely
tied to what we ought to believe. For Descartes, however, the boundaries of scientia lie
very far from the boundaries that should delimit belief. In the First Meditation he
remarks that his habitual opinions “are doubtful in a way, but are nevertheless highly
probable, and are such that it is much more reasonable to believe than to deny them.”b
Admittedly, this passage compares only belief and denial: it does not say that it is more
reasonable to believe than to suspend judgment. But a passage from the Synopsis to the
Meditations goes farther, remarking that “the great benefit of these arguments is not, in
my view, that they prove what they establish—namely, that there really is a world, that
human beings have bodies, and so on—things that no sane person has ever seriously
doubted.”c It would be insane to doubt these matters; yet it is also the case, according to
Descartes, that no one has ever had a scientia of these things. Apparently, then, our lack
of scientia concerning some proposition has no direct relation to whether we ought to
believe it. This is quite alien to our modern conception of knowledge.
This aspect of Descartes’s account is closely related to the first aspect. The skeptics of
old had maintained that we ought to withhold assent. If we must act, we should act only
as if we have beliefs about the world. Because Descartes sees no such connection
between scientia and belief, his form of skepticism (if it should be called that at all) is
of a purely theoretical sort. Hardly anyone has scientia, but this makes no real difference
to anyone’s life. For Descartes as a philosopher and scientist, the rarity of scientia is a
depressing result, and one that he wants to change. But he has no expectation that his
methods will lead everyone to acquire scientia, and no real interest in seeing that happen.
Indeed, the preface to the Meditations explains that he wrote in Latin rather than in
French “lest weaker intellects might believe that they too ought to set out on this path.”d
Now, we have seen Descartes say that those of us who seek the truth should believe
only what we can grasp with certainty. But this advice applies strictly to those who are
pursuing scientia. Those who have no interest in that project, or lack the ability to pursue
it, or simply have not yet found the time to do so are certainly not supposed to give up
all their beliefs. That, as Descartes says, would be insanity. Even in the absence of
scientia, ordinary folk ought to go on believing what they do. Whether such beliefs
a b
Second Replies (VII: 141). Med. 1 (VII: 22).
c d
Med. synopsis (VII: 15–16). Med. preface (VII: 7).
should count as knowledge, in some more ordinary sense of the word, is not something
that Descartes shows any interest in.2
The Ideal of Certainty

Our normative epistemic ideal, says Descartes, encompasses foundationalism, internal-
ism, and certainty. All three of these notions have a long history, which might be
profitably explored. But the story that I want to tell, of how an idealized epistemology
began to concern itself with boundary marking, requires attending especially to the role
of certainty.
Familiar as the demand for certainty is in epistemology, its historical roots are not at
all easy to establish. Aristotle, surprisingly enough, says nothing about it in the Posterior
Analytics. Yet if one jumps ahead to later medieval commentaries, certainty is treated as
if it were an expectation rooted in the text itself. At the start of Albert the Great’s
commentary, for instance, he declares:
A human being ought to fill his soul not with what is [merely] plausible [probabile] and conjectural
[opinabile], because they do not yield a stable [stans] disposition in the soul, but with things that
are demonstrable and certain, which render the intellect certain and stable, because such things
are themselves certain and eternally stable.a
Thomas Aquinas inserts the certainty requirement directly into his gloss on Aristotle’s
canonical definition of epistēmē;b and, when John Duns Scotus defines scientia, the
certainty requirement appears as a fourth condition added on to the properly Aristote-
lian requirements that its object be (1) a necessary truth that is (2) evident through some
prior cause that (3) entails its conclusion syllogistically.c All these authors explicitly link
the certainty requirement with the notion that to achieve scientia is to achieve cognitive
perfection.3
So where, if not from Aristotle, does the demand for certainty arise? Given my claim
that premodern epistemology is largely idealized, and given how natural it is to build
certainty into such an ideal, it should be no surprise that the demand for certainty arises all
over the place. Without making any claim to completeness, let me here briefly describe
three streams of influence. A first one is the Stoic tradition of distinguishing between mere
impressions and the sort of cognition (katalēpsis) that is required for epistēmē, a distinction
Zeno memorably pictured as the difference between an open hand and a closed fist. To
have epistēmē is to have something beyond katalēpsis, something possible only for the wise
man. It is to have a closed fist with the other hand wrapped around it and squeezing
it tight. This requires a grasp of things that is “secure and firm and unchangeable by
reason.”d Moving into the Latin tradition, Cicero in his Academica first describes the Stoic
criterion and then strategically deploys it on behalf of the Academic skeptic. A similar
criterion appears in Augustine, who, despite writing Contra academicos, likewise starts from
a position quite favorable to Academic skepticism:
a
Comm. Post. an. I.1.1 (ed. Jammy, I: 514a).
b
Comm. Post. an. I.4 n. 5, glossing Post. an. I.2, 71b9–12, for which see Lecture One, p. 6.
c
Ordinatio prol. part 4 qq. 1–2 (Vatican edn. I, n. 208).
d
Sextus Empiricus, Against the Logicians I.153 (Loeb vol. 2).
The Ideal of Certainty 27
I do not call anything scientia where the person who professes it is sometimes mistaken. Scientia
does not consist merely in the matters that are apprehended. Instead, it consists in the fact that
they are apprehended in such a way that nobody should be in error about it or vacillate when
pressed by any opponents.a
That this notion comes from the Stoics is made clear when Augustine goes on to invoke
“certain philosophers” who think that only the wise man has scientia.4
A second source for the emphasis on certainty in epistemology is the Alexandrian
tradition of late antiquity. Last week we saw Ptolemy, near the head of that tradition,
dismiss two thirds of theoretical philosophy as being “guesswork rather than know-
ledge” and hold out only mathematics—which included astronomy—as providing “sure
and unshakable knowledge to its devotees.”b In that passage Ptolemy uses the phrase
katalēpsis epistēmonikē, which points back to Stoicism; but the very different context of
these remarks at the start of his famous Almagest warrants treating this passage as a
distinct line of influence on later discussions of certainty. And it would indeed be
influential, even many centuries later. The remark from Albert the Great quoted a
moment ago, for instance, is explicitly presented as a gloss on this remark of Ptolemy’s.
But, whereas for Ptolemy the lesson is that only the mathematical sciences yield
certainty, Albert expands this line of thought to embrace any conclusion that can be
produced through Aristotle’s demonstrative method. Albert thus looks not to mathem-
atics for certainty but to the Posterior Analytics, concluding that, of the various branches
of logic, “this alone is nobler and more excellent than the others on account of the
certitude of its proofs.”c In so doing, Albert is following later Alexandrian traditions,
which embraced the Stoic–Ptolemaic quest for certainty but transposed into an Aristo-
telian context. Consider the first surviving commentary on the Posterior Analytics—that
of Philoponus, from the early sixth century—the very first words of which are that “this
[subject] is the culmination [telos] of the study of logic.”d A little later he adds:
“Philosophers need demonstration as an instrument [organon] for the correct pursuit
[katorthōsis] of the parts of philosophy . . . It is by the standard [kanōn] of demonstration
that the philosopher distinguishes true from false in a theory and good from bad in
action.”e Later in that same Alexandrian tradition, Elias would distinguish the branch of
philosophy dealt with in the Posterior Analytics from the other parts of philosophy by the
fact that only it is comprised of “propositions that are true in all respects.”f By this point,
the sort of certainty that had seemed altogether impossible to the skeptics and that the
Stoics had reserved for the wise man and Ptolemy for mathematics has come to be
identified as the province of Aristotelian demonstration.
This late Alexandrian tradition was taken up directly into the foundations of early
Arabic philosophy, beginning in the ninth century, and thus we arrive at a third stream
of influence behind the later epistemic ideal of certainty. Al-Kindī, the fountainhead of
Arabic philosophy, took Ptolemy’s Almagest as his exemplar and sought to develop a
philosophy that would live up to the standards of certainty exhibited by mathematics.
Subsequently the association between knowledge and certainty would become virtually
a b
Contra academicos I.7.19. Almagest I.1 (p. 36 trans.); see Lecture One, p. 7.
c
Comm. Post. an. I.1.1 (ed. Jammy, I: 514a).
d
Comm. Post. an. (Commentaria in Aristotelem Graeca [= CAG] XIII.3): 1.5; p. 15 trans. McKirahan.
e f
2.24–3.1; p. 16 trans. Comm. Categorias proem (CAG XVIII): 117.2–3.
inevitable within the Arabic tradition, because Abuˉ Bishr Mattā’s standard tenth-century
Arabic translation of the Posterior Analytics employs ‘certainty’ (yaqīn) quite liberally
throughout the text, in places where Aristotle speaks simply of knowledge (ʿilm) or
demonstration (burhān). Al-Fārābī puts particular weight on the notion of certainty,
describing “certain philosophy” as the culmination of a process that gets an imperfect
start in sophistical and dialectical reasoning.a In his logical works he embraces the
Alexandrian picture of the Posterior Analytics as the sole basis for certitude, and hence
as the culmination of all reasoning. This picture subsequently runs through all of later
Islamic thought. According to al-Ghazālī, for instance, “true demonstration is what
provides necessary, perpetual and eternal certainty that cannot change.”b 5
Given how entrenched the notion of certainty had become in the later Middle Ages in
such a wide range of places, it is no wonder that this notion gets taken for granted by
Albert the Great and others, as an unquestioned element in their conception of the
epistemic ideal. No wonder, either, that Descartes takes it to be perfectly obvious that
scientia requires certainty. Taking up the torch of Ptolemy—though, to be sure, not
under Ptolemy’s now discredited name—he reconsiders the question of where outside
mathematics certainty may possibly be found. Rejecting the Aristotelian tradition that
ran from Alexandria through Baghdad and on to Paris, Descartes boldly replaces the
Posterior Analytics with his own approach. The method changes, but the epistemic ideal
remains the same.
Here, then, are the bare outlines of how the ideal of certainty has made its way across
the centuries. But now, with all this history on display, a question I have hitherto
avoided becomes inescapable. What exactly is this certainty that is so sought after? The
question is not easy to answer. If we think of the English word and its Latin origins, we
can say that ‘certainty’ has at its core the ideas of fixedness and stability. Applied to the
case of belief, we can understand such stability to hold with respect to either the belief ’s
subject or its object. Subjectively speaking, certainty requires a confidence so stable as
not to admit of any doubt on the believer’s part. Call this indubitability. Objectively
speaking, certainty requires that the thing believed be sufficiently fixed in its existence as
to be a stable object of knowledge. Call this necessity. Aristotle had insisted on this sort of
object-side certainty, but what characterizes later philosophy is the expectation of a
certainty that encompasses both of these components. In Albert the Great, for instance,
as quoted above, the proper objects of scientia “render the intellect certain and stable,”
and do so “because such things are themselves certain and eternally stable.” Within this
framework, the great question, then, arises of how to tie these subjective and objective
components together into a single, overarching expectation. Mere subjective confidence
might rightly be described as a kind of certainty, but it hardly captures the ideal. Thus
John Buridan begins his account of what it means to describe scientia as certain by
insisting that certainty requires truth. “No one is certain of something through a false
opinion, however firmly he may adhere to it.”c But of course it would hardly count as
ideal just to stumble upon a necessary truth and doggedly insist on it. Hence Albert says
that certainty on the subject side arises “because” of certainty on the object side. This
a
See e.g. his Book of Letters (Kitāb al-ḥuruˉ f ), n. 108 etc. (see Khalidi, Medieval Islamic Philosophical Writings, pp. 1 ff.).
b
Miʿyār al-ʿilm (McGinnis and Reisman, Classical Arabic Philosophy, p. 239).
c
Quaest. meta. I.3.
The Ideal of Certainty 29
requires some sort of method of discovery that reliably takes us back and forth between
truths in the world and belief in those truths: a bridge between objective and subjective
certainty. Aristotle offers one such method, Descartes another. In each case, the ideal is a
certainty that is more than merely subjective or objective, and indeed more than just the
simple conjunction of the two. What we want is to have indubitable beliefs that are
grounded in the truth in a way that makes those beliefs incapable of going wrong. Call
this infallible certainty.6
Over and over in the weeks ahead such infallibility will appear as a perennial ideal,
pursued but never attained. Part of what made its attainment particularly difficult is the
abiding notion that such certainty, if it is to count as genuine, must be achieved at the
ideal limit. Nicholas of Autrecourt, in the fourteenth century, makes this quite explicit:
“the certainty of evidentness has no degrees. Thus, if there are two conclusions, and
we are evidently certain of each, then we are no more certain of one than of the other.”a
Three centuries later, Thomas Hobbes would similarly conclude that “the certainty of all
scientiae is equal, for otherwise they would not be scientiae, since to have scientia does not
admit of more and less.”b Such absolutism is natural enough, given certainty’s role as a
cognitive ideal. But, given that no one could reasonably suppose that we often reach that
ideal, we should also expect authors to harp on such absolute certainty only when their
aims are destructive. Thus Hobbes’s remark comes at the start of a tract devoted to
attacking the geometers of his day for their incompetence. And Autrecourt stresses
that certainty is all or nothing along the way toward showing that his opponent grasps
almost nothing with certainty.7
Where doubts arise over the attainability of an epistemic ideal, the natural response
within the idealized framework is to reduce normative expectations. This is indeed what
one finds time and again throughout the history of philosophy, even while one
continues to find other authors—and sometimes the very same author—pursuing that
same epistemic dream. Consider again Descartes. He claims quite explicitly that we have
infallible certainty when we restrict ourselves to assenting to what we clearly and
distinctly perceive; in such cases “it clearly cannot happen that I err.”c But Mersenne
pushes him, in the Second Set of Objections, on whether the certainty of clear and
distinct perceptions can itself be supported by a noncircular certain argument. Without
that, the objection charges, we can have “no degree of certainty.”d Descartes responds,
predictably, by invoking divine goodness, but he sees that more is needed, and so his
next step is to shift the argument from infallibility to indubitability, “a conviction so firm
that it can in no way be removed, and such a conviction is clearly the same as the most
perfect certainty.”e Has Descartes made a concession or has he not? The inability to
doubt would seem to fall a long way short of the inability to be wrong, yet Descartes
carries on as if he has delivered exactly the sort of perfect certainty that he promised,
and subsequent works continue to make assertions of infallibility. Indeed, one might
say that the very essence of his method of radical doubt is to help us distinguish
the merely indubitable from the truly infallible and reframe our beliefs accordingly.
Lecture Five will return to consider exactly why Descartes is unable to achieve that goal;
a b
Second letter to Bernard, n. 6. De principiis et ratiocinatione geometrarum, p. 2.
c
Med. 4 (VII: 62); see also e.g. Med. 3 (VII: 35), Second Replies (VII: 144), Principles I.43.
d e
Second Objections (VII: 126). Second Replies (VII: 145).
but for now the point is just that infallibility is an ideal he can neither attain nor bring
himself to abandon.
Consider Galileo, too. As a young man, he had lectured on the Posterior Analytics, and
throughout his career he stressed the importance of achieving certainty. In a letter from
near the end of his life addressed to an old Aristotelian rival, he praises the Aristotelian
method of logic for providing “sureness of demonstration” and remarks: “up to this
point I am a Peripatetic.”a In his battles over the Copernican theory of the solar system,
he embraces a principle he found in Augustine, to the effect that, where human inquiry
is uncertain, we should follow biblical authority in reshaping it but, where human
inquiry achieves certainty, we should use those certain results as a guide to interpreting
the Bible. Hence his goal in astronomy has to be nothing less than a demonstration of the
heliocentric model. Otherwise the Bible would trump. But just how much certainty can
we expect? At the end of the first day of his Dialogue concerning the Two Great World
Systems (1632), Galileo has Salviati make a remarkable speech that compares our
epistemic perfection to God’s. How well we fair in such a comparison depends, Salviati
says, on whether we think about it extensively or intensively. Extensively, human beings
do poorly, because there are infinitely many things that God understands, only a very few
of which we understand. But intensively—that is, with respect to the certainty with which
we can grasp a single proposition—our minds can be equal in perfection to God’s. As
Salviati puts it, “with regard to those few propositions that the human intellect does
understand, I believe that its knowledge equals the divine in objective certainty, for here it
succeeds in understanding necessity, beyond which there can be no greater sureness.”b
“Very bold and daring,” Galileo has Simplicio reply. Bold enough that, when the
church began proceedings against Galileo a year later, this claim was considered among
his suspicious teachings.
This distinction between two dimensions of the epistemic ideal, intensive and
extensive, fits well with the diminished expectations we saw Galileo endorse last
week: his willingness to relax the causal requirement in order to permit a mode of
inquiry that describes the features of a given physical phenomenon, without entering
into speculation about why or how things work as they do.c Leaving such extensive
perfection to the metaphysician, Galileo contents himself with aiming at intensive
perfection in a few well-defined areas of natural philosophy. The ambition of God-like
certainty, he makes clear, can be humanly achieved only in one scientia—that of
mathematics. Still, with respect to his insistence on such certainty as a normative ideal
of epistemic inquiry, Galileo falls into line with a tradition we have seen stretch back to
the Stoics and Ptolemy, through the schools of Alexandria, Baghdad, and Paris. Indeed,
he and Descartes might be viewed as the last great champions of that tradition.8
At the end of today’s lecture I will make a suggestion about why the ideal of infallible
certainty has such a persistent hold on us, even while its attainability seems so obviously
doubtful. But first we should consider what comes next in our story: how the idealized
expectation of infallible certainty gave way to greater tolerance for mere probability,
which led ultimately to a new concern with finding the epistemic boundaries of
permissible belief.
a
To Liceti in 1640 (Opere XVIII: 248). b
Dialogue, first day (p. 118 trans.; Opere VII: 128–9).
c
See Lecture One, pp. 14–16.
Degrees of Evidentness 31
Degrees of Evidentness
Conventional wisdom has it that the era of Descartes and Galileo suffered from a strange
sort of epistemic malady, a crise pyrrhonienne. This notion is unfortunate in various
respects: it gives too much weight to the influence of ancient skepticism; it exaggerates
the degree to which there was anything approaching a skeptical “crisis” at the time; and
it distorts the motivation of Descartes and others, as if their main project were the
refutation of skepticism. Yet it is clear enough that the sixteenth and seventeenth
centuries display a growing concern over whether the traditional epistemic ideal of
certainty is achievable. Sometimes these doubts amount to an attack on philosophy
itself—which means, for that time, also an attack on science itself. Of the greatest
interest here, however, are those who responded to these currents by reframing the
epistemic ideal around probability rather than around certainty.9
A great deal has been written about the rise of interest in probabilistic reasoning
in seventeenth-century thought, but it is not easy to say exactly what is new here.
What is most clearly new is a sudden flourishing of mathematical treatments of
probability, running from the correspondence of Pascal and Fermat to the work of
Christiaan Huygens and on to Jacob Bernoulli at the end of the century. The
appearance of these and other works at this time coincides with a general rise in
the professionalization and sophistication of mathematics, but it also reflects a wider
cultural interest in the relationship between probability and certainty. Whereas, for
centuries, the Aristotelian tradition had privileged an infallible grasp of necessary
truth as the hallmark of scientia, the waning of scholasticism gave rise to an
increasing interest in the contingent and the merely probable. One sees this in a
wide array of domains, including law, medicine, ethics, religion, economics, and
natural philosophy.10
To understand these events in their proper historical context, one needs to go back
well before the seventeenth century. We have seen how medieval authors, both Latin
and Arabic, took for granted that certainty is a key part of the epistemic ideal. They did
so with little explicit encouragement from Aristotle. On the contrary, Aristotle had
famously warned, near the start of the Nicomachean Ethics, that one should resist the
temptation to apply the same standards to every domain: “precision [to akribes] is not to
be sought alike in all discussions.”a This passage was very regularly invoked by medieval
Aristotelians, but with an important difference. When Robert Grosseteste, in the 1240s,
produced the first Latin translation of the Ethics, he rendered the Greek akribes as certum.
This had a decisive impact on how the Ethics passage would be understood. Here, for
instance, is Aquinas:
According to the Philosopher in Ethics I, “certainty is not to be sought alike in all matters.” For in
the case of human actions . . . demonstrative certainty cannot be had, inasmuch as they concern
contingent and variable matters. Hence probable certainty suffices, which attains the truth for the
most part, even if in a few cases it falls short of the truth.b
a
NE I.3, 1094b13. b
Summa theol. 2a2ae 70.2c.
Aquinas and others regularly extended the point to natural philosophy, where one
likewise has to be satisfied with conclusions that hold only “for the most part.” Here
too, it was understood that a lesser kind of certainty is enough.11
Aquinas offers only a fairly coarse-grained distinction between demonstrative cer-
tainty and probable certainty. But when we reach John Buridan in the mid-fourteenth
century, we arrive at what would become the canonical three-level distinction between
absolute, natural, and moral certainty. Because Buridan’s discussion of these issues in his
Questions on the Metaphysics (1340s) is both sophisticated and the forerunner of how
epistemology would later develop, it is worth taking a little time to work through it.
Tellingly, the question Buridan asks is not whether we have knowledge but whether a
“comprehension of the truth” is possible for us. His initial response is that this is quite
easy, because to comprehend the truth is simply to assent to a true proposition, and we
all do this all the time. The hard question is whether we can do it with certainty. Here he
draws the usual distinction between objective and subjective certainty. On the object
side, what is needed is that the truth assented to be firm; and here, too, he thinks it
uncontroversial that this condition can be met. He gives the example of the proposition
that God exists, where such firmness holds simpliciter, “because in no case can it be
falsified.”a This last expression makes it clear that by “firmness” he means necessity—in
no possible world can the proposition come out false. Another class of examples exhibits
firmness ex suppositione—specifically, “on the supposition of the common course of
nature.” Here he gives various examples from natural science, which can be said to
display firmness of truth “notwithstanding the fact that God could make fire cold and so
falsify the proposition that all fire is hot.” This is what we now call natural necessity.
Matters get more involved when Buridan turns to the subject side of certainty and to
the need for what he calls “firmness of assent.” This is understood in purely subjective
terms, as “assenting to a proposition without any fear of the opposite.”b Again, he says,
there is no doubt that we can achieve this standard. Both Christians and heretics have
been known to be so confident about the truth of their beliefs that they were willing to
die for them. And we all achieve considerable subjective certainty all the time, in virtue
of what he calls “natural appearances.” But the final and hardest problem for Buridan is
whether we can achieve such subjective certainty in a way that is evident. Even if a true
belief is maximally certain both subjectively and objectively—as is the case, for instance,
with the faithful’s endorsement of church dogma—this will be mere faith, says Buridan,
and not genuine scientia, unless we have the right sort of evidentness.
What is it to be evident? For Buridan and his contemporaries, evidentness is the
bridge that connects the purely objective and the purely subjective senses of certainty.
Roughly speaking, it is the all-important quality that distinguishes scientia from mere
true belief. Indeed, although historians have paid it little systematic attention, evident-
ness is the central epistemic concept among both scholastic philosophers and their critics;
it features prominently first among Aristotelians, then in Descartes and throughout the
seventeenth century, and even up to the time of David Hume.
a
Quaest. meta. II.1, f. 8vb (p. 145 trans. Klima, Medieval Philosophy).
b
II.1, f. 8vb (p. 145 trans.). For “fear of the opposite,” see Lecture Six, p. 136.
As a start at an analysis, we can distinguish three entwined notions:

A. The evidentness of a cognitive object; that is, a thing’s being evident.
B. The evidentness of a cognition that grasps such an object; that is, an evident
cognition.
C. That which makes something be evident; that is, the evidence.
The last of these senses is most deeply entrenched in epistemology today. Moreover,
whether we are dealing with Latin (evidentia), French (évidence), or English (evidence),
modern readers find it natural to suppose that we are talking about type-C evidence. In
fact, however, it is not until the later eighteenth century that this third sense became
prevalent in philosophical texts. Before that time, the predominant senses were A and
B. In either sense, evidentness is widely understood to be what distinguishes scientia
from lesser cognitive states. Thus Aquinas: “the certainty of scientia and understanding
[intellectus] comes from the evidentness of the things that are said to be certain, whereas
the certainty of faith comes from a firm adherence to that which is believed.”a Skipping
ahead four hundred years, we can find Thomas Hobbes, writing in English, still taking
much the same view:
There are two things necessarily implied in this word knowledge; the one is truth, the other
evidence; for what is not true can never be known . . . Likewise, if the truth be not evident,
though a man holds it, yet is his knowledge of it no more than theirs that hold the contrary. For if
truth were enough to make it knowledge, all truths were known: which is not so.b
Translated into modern English, ‘evidence’ in the first line would best be rendered as
‘evidentness,’ to make it clear that Hobbes, like Aquinas, is speaking in terms of sense A.
That is, for both authors, and for a great many more in the centuries between them, the
most general characterization of scientia or knowledge is that it is a firm belief regarding
an evident truth.12
So it is that the most fundamental epistemic questions of these centuries concern what
evidentness is and under what circumstances a proposition counts as evident. In the
weeks ahead I will consider the second of these questions, looking at some of the ways in
which perception and reasoning yield evidentness. Focusing for now only on the
question of what it is, the start of an answer is to observe that things possess type-A
evidentness when they are apt to produce type-B evidentness. A mark of the latter, in
turn, is indubitability: someone who evidently cognizes a proposition cannot resist
assenting to that proposition. But here we reach a crucial question: is such indubitability
sufficient for evidentness, or is it a mere mark? The usual scholastic view seems to be that
evidentness requires more than a subjective inability to doubt—that it further requires
the proposition to be true, and indeed that we cannot go wrong about that truth. This is
to say that evidentness yields infallible certainty.13
For a particularly nuanced example of how such a story goes in detail, let us return to
John Buridan and his final and hardest problem: whether we grasp the truth with
evidentness. His solution begins by distinguishing among three degrees of being
evident. The first, unqualified degree (evidentia simpliciter) occurs “when, from the
nature of sense or intellect, a human being is compelled or necessitated to assent to a
a b
Sent. III.23.2.2.3c. Elements of Law I.6.2.
proposition in such a way that he cannot dissent.”a This applies to first principles and
their logical consequences. Next comes natural evidentness, which likewise compels the
intellect, but only on the supposition that “the common course of nature is observed.”b
So far, his account fits the analysis just offered. First, evidentness is grounded in things,
but is relative to our human faculties. Only certain sorts of propositions are evident, and
which ones count as evident depends on “the nature of sense or intellect” (as just
quoted). Second, the reason why evidentness is observer-relative is that the mark of
evidentness is cognitive compulsion. As is typical throughout this period, such evident-
ness is not characterized by concepts like justification and warrant. There is no suggestion
that we achieve this sort of evidentness and thereby earn the right to form a certain belief.
Quite to the contrary, Buridan—like others—describes evidentness as something that
compels assent. That which is evident is indubitable, whereas nonevident propositions
carry at least some measure of doubt. This is the sense in which evidentness yields
certainty on the subject side. Third, Buridan wants more from evidentness than mere
subjective compulsion; he wants a kind of objective infallibility, such that, as he puts it
elsewhere, “someone cannot be deceived.”c In cases of the first degree, the guarantee is
absolute. In cases of the second degree, it holds only conditionally, yet he insists that this
suffices for scientia regarding the natural world.
Buridan’s discussion takes a dramatic turn, however, when he introduces his third
degree of evidentness:
There is still another, weaker evidentness, which suffices for acting well morally. This goes as
follows: if someone, having seen and investigated all the attendant circumstances that one can
investigate with diligence, judges in accord with the demands of such circumstances, then that
judgment will be evident with an evidentness sufficient for acting well morally—even if that
judgment were false on account of invincible ignorance concerning some circumstance. For
instance, it would be possible for a judge to act well and meritoriously by hanging an innocent
man because through testimony and other documents it sufficiently appeared to him in accord
with his duty that that good man was a bad murderer.d
In contrast to the first two degrees, this sort of evidentness is fallible. Indeed, in the
example on offer, the judge’s judgment is evident but false. Furthermore, such moral
evidentness apparently does not compel assent. We are now instead in the domain of
moral appraisal, where someone who assents and acts on the basis of what is morally
evident is doing the right thing. With this, evidentness can play a role close to that of
justification. Given the degree to which things are evident, we can be said to be within
our rights to form such a belief.
This shift from compulsion to justification will loom large in what follows. In these
brief remarks, however, Buridan offers just the barest anticipation of the changes that
are to come. He says so little, indeed, that it is not even clear whether this weakest
category of evidentness is enough for scientia. Given that evidentness is supposed to be
what distinguishes scientia from mere faith, he is perhaps implicitly committed to
broadening scientia that far. Yet he does not explicitly say so; and, before moving on
a
Quaest. meta. II.1, f. 8vb (p. 145 trans. Klima, Medieval Philosophy).
b c
II.1, f. 8vb (p. 145 trans.). Quaest. Post. An. I.2c.
d
Quaest. meta. II.1, f. 9ra (p. 146 trans.).
to discuss later chapters in our story, it is worth pausing to reflect on why he does not.
As I have been stressing, it makes very little immediate difference, from within the
idealized framework, whether moral evidentness is sufficient to make a firmly held true
belief count as scientia. Buridan has told us what scientia in the strictest sense looks like,
and we can see how, if we want to be able to speak of scientia with regard to the natural
world, we need to weaken the requirements. Our having to presuppose the regularity
of nature, to achieve certainty in that domain, is not something that Buridan finds
worrisome—it is just a fact about our epistemic situation. Similarly, whether or not the
judge knows—or has scientia—that the accused is guilty is simply not an interesting
question from Buridan’s point of view. Elsewhere Buridan is perfectly happy to acknow-
ledge the scientia of wholly contingent propositions that are based on mere appearances:
“everybody speaks in this way: that ‘I know [scio] that this iron is hot, because I manifestly
sense that it is hot,’ and ‘I firmly know that Socrates was running yesterday, because I saw
him running.’”a Of course, from a more demanding point of view, such beliefs are
markedly uncertain. But what matters, in the idealized framework, is simply to map the
contours of our epistemic situation, not to draw arbitrary lines in the sand.
Again, then, we see early epistemology’s lack of interest in the question of exactly
where to place the boundaries between what is and is not knowledge. And this in turn
helps explain why authors from this era tend to be so little worried about skepticism.
For, without the sort of boundary mongering that is familiar today, it becomes unclear
how one is even to describe the alleged crisis of skepticism. To say that human beings
lack scientia in the strictest sense is virtually a truism, and hardly tantamount to
skepticism. So presumably the skeptical challenge has to be made at some lower level
of demandingness. But where? If one descends far enough—all the way, for instance, to
Buridan’s weakest form of evidentness—then skepticism seems to lose much of its force.
Yet where in between these two extremes do we find some privileged focal point?
Buridan himself attacks those who “speak very badly” by demanding absolute evident-
ness in all things,b but he does not fuss over the limits to scientia. And so it is that later
medieval authors in general are remarkably unconcerned by the threat of skepticism. It
is not just that their theological worldview gives them a ready response, since of course
the skeptic will challenge that worldview along with all the rest. Rather, skepticism looks
uninteresting from within an idealized framework, because it seems to collapse into
either the platitudinous complaint that we can never have complete certainty or the
unbelievable claim that we can never have even reasonable confidence.
Yet, alas, the moral of our story is not quite so simple and satisfying. With respect to
skepticism, we will see in Lecture Six how an adversary as determined as Hume can raise
deep skeptical problems even from within the idealized framework. And in what remains
of this lecture I want to consider how the demand for epistemic boundaries starts to loom
large even in the seventeenth century, as the notion of moral certainty, paired with an
increased interest in the merely probable, leads to an increasing concern with identifying
the point at which belief becomes warranted. Out of this is born the modern notion of
justification, understood not as an intellectual compulsion in the face of an evident truth,
but rather as an epistemic permission in view of what is sufficiently probable.14
a
Summulae VIII.4.4 (p. 710 trans.).
b
Quaest. meta. II.1, f. 9ra (p. 146 trans.); see also Lecture Six, p. 118.
Proportioned Belief
Although John Buridan did not use the phrase ‘moral certainty,’ his conception of a
lesser degree of certainty would start to appear regularly, under just this label, within a
few decades of his death. It seems to appear first in Jean Gerson and allied moral
theologians of the fifteenth century, and thereafter becomes commonplace throughout
scholastic discussions. These discussions continue into the early seventeenth century,
both on the continent and in England, especially in Protestant apologetics. William
Chillingworth, for instance, in his hugely influential Religion of Protestants (1638), defends
the textual reliability of the Bible by appealing to a lesser kind of certainty.
But [you say that] we cannot be certain in what language the scriptures remain uncorrupted. Not
so certain, I grant, as of that which we can demonstrate; but certain enough, morally certain, as
certain as the nature of the thing will bear. So certain we may be, and God requires no more.a
Subsequent religious controversialists, notably John Tillotson, took up this same line,
and from there it took on a prominent place in the scientific ideology of the Royal
Society, particularly through the influence of John Wilkins, one of the founding figures
of that society and also a keen participant in the religious controversies of his day. His
Principles and Duties of Natural Religion (1675) begins with a beautifully clear little chapter
on evidence and belief in which he draws the usual threefold distinction between what
he labels mathematical, physical, and moral certainty. Like others before him, Wilkins
urges that we should not expect greater certainty than the subject admits of. Similar
ideas are advanced at length by a great many others.15
Recent scholars, unaware of the long history of such distinctions, have often been
overly impressed with the novelty of these seventeenth-century developments. It
becomes harder to say what exactly is new here, however, once we see that the notion
of moral certainty can be found all throughout later scholastic discussions. The con-
stantly heard seventeenth-century injunction to look for only as much certainty as is
possible is nothing new either: as we have seen, one finds it all throughout the
Aristotelian tradition, and it has roots in Aristotle’s work itself. What changes most
dramatically in post-scholastic epistemology is that the possibility of infallible certainty
no longer gets taken for granted, even in the most ideal of cases. Thus, by the mid-
seventeenth century, talk of levels of certainty is routinely accompanied by the proviso
that even the highest sort of human certainty should not be regarded as absolute
infallibility. Wilkins, for instance, describes it as “a blasphemous arrogance” to charac-
terize even mathematical certainty as unerring. The best we can have is “a conditional
infallibility, that which supposes our faculties to be true, and that we do not neglect the
exerting of them.”b Even more strikingly new is the readiness of so many figures to give
up, in almost all cases, anything beyond moral certainty. Robert Boyle cautions, for
instance, that “there are I know not how many things in physics that men presume they
believe upon physical and cogent arguments, wherein they really have but a moral
assurance.”c Joseph Glanvill, in a similar spirit, writes:
a b
Religion of Protestants ch. 2, §55. Principles and Duties I.1, p. 9.
c
Excellence of Theology pt. II, §3 (Works VIII: 66).
Another random document with
no related content on Scribd:
Apollodorus’s shadow painting was not in any sense what we call
chiaroscuro and atmosphere, but sheer modelling in the round in the
sculptor’s sense; and of Zeuxis Aristotle says expressly that his work
lacked “Ethos.” Thus, this newer Classical painting with its
cleverness and human charm is the equivalent of our 18th-Century
work. Both lacked the inner greatness and both tried by force of
virtuosity to speak in the language of that single and final Art which
in each case stood for ornamentation in the higher sense. Hence
Polycletus and Phidias aline themselves with Bach and Händel; as
the Western masters liberated strict musical form from the executive
methods of the Painting, so the Greek masters finally delivered the
statue from the associations of the Relief.
And with this full plastic and this full music the two Cultures reach
their respective ends. A pure symbolism of mathematical rigour had
become possible. Polycletus could produce his “canon” of the
proportions of the human body, and his contemporary Bach the
“Kunst der Fuge” and “Wohltemperiertes Klavier.” In the two arts that
ensued, we have the last perfection of achievement that pure form
saturated with meaning can give. Compare the tone-body of
Faustian instrumental music, and within that system again the body
of the strings (in Bach, too, the virtual unity of the winds), with the
bodies of Attic statuary. Compare the meaning of the word “figure” to
Haydn with its meaning to Praxiteles. In the one case it is the figure
of a rhythmic motive in a web of voices, in the other the figure of an
athlete. But in both cases the notion comes from mathematics and it
is made plain that the aim thus finally attained is a union of the
artistic and the mathematical spirit, for analysis like music, and
Euclidean geometry like plastic, have both come to full
comprehension of their tasks and the ultimate meaning of their
respective number-languages. The mathematics of beauty and the
beauty of mathematics are henceforth inseparable. The unending
space of tone and the all-round body of marble or bronze are
immediate interpretations of the extended. They belong to number-
as-relation and to number-as-measure. In fresco and in oil-painting,
in the laws of proportion and those of perspective, the mathematical
is only indicated, but the two final arts are mathematics, and on
these peaks Apollinian art and Faustian art are seen entire.
With the exit of fresco and oil-painting, the great masters of
absolute plastic and absolute music file on to the stage, man after
man. Polycletus is followed by Phidias, Pæonius, Alcamenes,
Scopas, Praxiteles, Lysippus. Behind Bach and Händel come Gluck,
Stamitz, the younger Bachs, Haydn, Mozart, Beethoven—in their
hands an armoury of wonderful and now long-forgotten instruments,
a whole magician’s world created by the discovering and inventing
spirit of the West in the hope of getting more and more tones and
timbres for the service and enhancement of musical expression—in
their winds an abundance of grand, solemn, ornate, dainty, ironic,
laughing and sobbing forms of perfectly regular structure, forms that
no one now understands. In those days, in 18th-Century Germany
especially, there was actually and effectively a Culture of Music that
suffused all Life. Its type was Hoffmann’s Kapellmeister Kreisler. To-
day it is hardly even a memory.
And with the 18th Century, too, architecture died at last,
submerged and choked in the music of Rococo. On that last
wonderful fragile growth of the Western architecture criticism has
blown mercilessly, failing to realize that its origin is in the spirit of the
fugue and that its non-proportion and non-form, its evanescence and
instability and sparkle, its destruction of surface and visual order, are
nothing else than a victory of tones and melodies over lines and
walls, the triumph of pure space over material, of absolute becoming
over the become. They are no longer buildings, these abbeys and
castles and churches with their flowing façades and porches and
“gingerbread” courts and their splendid staircases, galleries, salons
and cabinets; they are sonatas, minuets, madrigals in stone,
chamber-music in stucco, marble, ivory and fine woods, cantilene of
volutes and cartouches, cadences of fliers and copings. The
Dresden Zwinger is the most completely musical piece in all the
world’s architecture, with an ornamentation like the tone of an old
violin, an allegro fugitivo for small orchestra.
Germany produced the great musicians and therefore also the
great architects of this century (Poppelmann, Schlüter, Bähr,
Naumann, Fischer von Erlach, Dinzenhofer). In oil-painting she
played no part at all: in instrumental music, on the contrary, hers was
the principal rôle.
VII
There is a word, “Impressionism,” which only came into general

use in Manet’s time (and then, originally, as a word of contempt like
Baroque and Rococo) but very happily summarizes the special
quality of the Faustian way of art that has evolved from oil-painting.
But, as we ordinarily speak of it, the idea has neither the width nor
the depth of meaning that it ought to have: we regard it as a sequel
to or derivative of the old age of an art which, in fact, belongs to it
entirely and from first to last. What is the imitation of an
"impression"? Something purely Western, something related to the
idea of Baroque and even to the unconscious purposes of Gothic
architecture and diametrically opposed to the deliberate aims of the
Renaissance. Does it not signify the tendency—the deeply-
necessary tendency of a waking consciousness to feel pure endless
space as the supreme and unqualified actuality, and all sense-
images as secondary and conditioned actualities "within it"? A
tendency that can manifest itself in artistic creations, but has a
thousand other outlets besides. Does not Kant’s formula "space as a
priori form of perception" sound like a slogan for the whole
movement that began with Leonardo? Impressionism is the inverse
of the Euclidean world-feeling. It tries to get as far as possible from
the language of plastic and as near as possible to that of music. The
effect that is made upon us by things that receive and reflect light is
made not because the things are there but as though they “in
themselves” are not there. The things are not even bodies, but light-
resistances in space, and their illusive density is to be unmasked by
the brush-stroke. What is received and rendered is the impression of
such resistances, which are tacitly evaluated as simple functions of a
transcendent extension. The artist’s inner eye penetrates the body,
breaks the spell of its material bounding surfaces and sacrifices it to
the majesty of Space. And with this impression, under its influence,
he feels an endless movement-quality in the sensuous element that
is in utter contrast to the statuesque “Ataraxia” of the fresco.
Therefore, there was not and could not be any Hellenic
impressionism; if there is one art that must exclude it on principle, it
is Classical sculpture.
Impressionism is the comprehensive expression of a world-feeling,
and it must obviously therefore permeate the whole physiognomy of
our “Late” Culture. There is an impressionistic mathematic, which
frankly and with intent transcends all optical limitations. It is Analysis,
as developed after Newton and Leibniz, and to it belong the
visionary images of number-“bodies,” aggregates, and the
multidimensional geometry. There is again an impressionistic
physics which “sees” in lieu of bodies systems of mass-points—units
that are evidently no more than constant relations between variable
efficients. There are impressionistic ethics, tragedy, and logic, and
even (in Pietism) an impressionistic Christianity.
Be the artist painter or musician, his art consists in creating with a
few strokes or spots or tones an image of inexhaustible content, a
microcosm meet for the eyes or ears of Faustian man; that is, in
laying the actuality of infinite space under enchantment by fleeting
and incorporeal indications of something objective which, so to say,
forces that actuality to become phenomenal. The daring of these arts
of moving the immobile has no parallel. Right from the later work of
Titian to Corot and Menzel, matter quivers and flows like a solution
under the mysterious pressure of brush-stroke and broken colours
and lights. It was in pursuit of the same object that Baroque music
became “thematic” instead of melodic and—reinforcing the “theme”
with every expedient of harmonic charm, instrumental colour, rhythm,
and tempo—developed the tone-picture from the imitative piece of
Titian’s day to the leitmotiv-fabric of Wagner, and captured a whole
new world of feeling and experience. When German music was at its
culmination, this art penetrated also into lyric poetry (German lyric,
that is, for in French it is impossible) and gave rise to a whole series
of tiny masterpieces, from Goethe’s “Urfaust” to Hölderlin’s last
poems—passages of a few lines apiece, which have never yet been
noticed, let alone collected, but include nevertheless whole worlds of
experience and feeling. On a small scale, it continually repeats the
achievements of Copernicus and Columbus. No other Culture
possesses an ornament-language of such dynamical impressiveness
relatively to the means it employs. Every point or stroke of colour,
every scarce-audible tone releases some surprising charm and
continually feeds the imagination with fresh elements of space-
creating energy. In Masaccio and Piero della Francesca we have
actual bodies bathed in air. Then Leonardo, the first, discovers the
transitions of atmospheric light and dark, the soft edges, the outlines
that merge in the depth, the domains of light and shade in which the
individual figures are inseparably involved. Finally, in Rembrandt,
objects dissolve into mere coloured impressions, and forms lose
their specific humanness and become collocations of strokes and
patches that tell as elements of a passionate depth-rhythm.
Distance, so treated, comes to signify Future, for what
Impressionism seizes and holds is by hypothesis a unique and
never-recurring instant, not a landscape in being but a fleeting
moment of the history thereof. Just as in a Rembrandt portrait it is
not the anatomical relief of the head that is rendered, but the second
visage in it that is confessed; just as the art of his brush-stroke
captures not the eye but the look, not the brow but the experience,
not the lips but the sensuousness; so also the impressionist picture
in general presents to the beholder not the Nature of the foreground
but again a second visage, the look and soul of the landscape.
Whether we take the Catholic-heroic landscape of Claude Lorrain,
the “paysage intime” of Corot, the sea and river-banks and villages
of Cuyp and Van Goyen, we find always a portrait in the
physiognomic sense, something uniquely-occurring, unforeseen,
brought to light for the first and last time. In this love of the character
and physiognomy in landscape—just the motive that was
unthinkable in fresco art and permanently barred to the Classical—
the art of portraiture widens from the immediately human to the
mediately human, to the representation of the world as a part of the
ego or the self-world in which the painter paints himself and the
beholder sees himself. For the expansion of Nature into Distance
reflects a Destiny. In this art of tragic, daemonic, laughing and
weeping landscapes there is something of which the man of another
Culture has no idea and for which he has no organ. Anyone who in
the presence of this form-world talks of Hellenistic illusion-painting
must be unable to distinguish between an ornamentation of the
highest order and a soulless imitation, an ape-mimicry of the
obvious. If Lysippus said (as Pliny tells us he said) that he
represented men as they appeared to him, his ambition was that of a
child, of a layman, of a savage, not that of an artist. The great style,
the meaning, the deep necessity, are absent; even the cave-dwellers
of the stone age painted thus. In reality, the Hellenistic painters could
do more when they chose. Even so late, the wall-paintings of
Pompeii and the “Odyssey” landscapes in Rome contain a symbol.
In each case it is a group of bodies that is rendered—rocks, trees,
even “the Sea” as a body among bodies! There is no depth, but only
superposition. Of course, of the objects represented one or several
had necessarily to be furthest away (or rather least near) but this is a
mere technical servitude without the remotest affinity to the illumined
supernal distances of Faustian art.
VIII
I have said that oil-painting faded out at the end of the 17th
Century, when one after another all its great masters died, and the
question will naturally, therefore, be asked—is Impressionism (in the
current narrow sense) a creation of the 19th Century? Has painting
lived, after all, two centuries more? Is it still existing? But we must
not be deceived by appearances. Not only was there a dead space
between Rembrandt and Delacroix or Constable—for when we think
of the living art of high symbolism that was Rembrandt’s the purely
decorative artists of the 18th Century do not count—but, further, that
which began with Delacroix and Constable was, notwithstanding all
technical continuity, something quite different from that which had
ended with Rembrandt. The new episode of painting that in the 19th
Century (i.e., beyond the 1800 frontier and in “Civilization”) has
succeeded in awakening some illusion of a great culture of painting,
has itself chosen the word Plein-air (Freilicht) to designate its special
characteristic. The very designation suffices to show the significance
of the fleeting phenomenon that it is. It implies the conscious,
intellectual, cold-blooded rejection of that for which a sudden wit
invented the name “brown sauce,” but which the great masters had,
as we know, regarded as the one truly metaphysical colour. On it had
been built the painting-culture of the schools, and especially the
Dutch school, that had vanished irretrievably in the Rococo. This
brown, the symbol of a spatial infinity, which had for Faustian
mankind created a spiritual something out of a mere canvas, now
came to be regarded, quite suddenly, as an offence to Nature. What
had happened? Was it not simply this, that the soul for which this
supernal colour was something religious, the sign of wistfulness, the
whole meaning of “Living Nature,” had quietly slipped away? The
materialism of a Western Cosmopolis blew into the ashes and
rekindled this curious brief flicker—a brief flicker of two generations,
for with the generation of Manet all was ended again. I have (as the
reader will recall) characterized the noble green of Grünewald and
Claude and Giorgione as the Catholic space-colour and the
transcendent brown of Rembrandt as the colour of the Protestant
world-feeling. On the other hand, Plein-air and its new colour scale
stand for irreligion.[364] From the spheres of Beethoven and the stellar
expanses of Kant, Impressionism has come down again to the crust
of the earth. Its space is cognized, not experienced, seen, not
contemplated; there is tunedness in it, but not Destiny. It is the
mechanical object of physics and not the felt world of the pastorale
that Courbet and Manet give us in their landscapes. Rousseau’s
tragically correct prophecy of a “return to Nature” fulfils itself in this
dying art—the senile, too, return to Nature day by day. The modern
artist is a workman, not a creator. He sets unbroken spectrum-
colours side by side. The subtle script, the dance of brush-strokes,
give way to crude commonplaces, pilings and mixings and daubings
of points, squares, broad inorganic masses. The whitewasher’s
brush and the trowel appear in the painter’s equipment; the oil-
priming of the canvas is brought into the scheme of execution and in
places left bare. It is a risky art, meticulous, cold, diseased—an art
for over-developed nerves, but scientific to the last degree, energetic
in everything that relates to the conquest of technical obstacles,
acutely assertive of programme. It is the “satyric pendant” of the
great age of oil-painting that stretches from Leonardo to Rembrandt;
it could only be at home in the Paris of Baudelaire. Corot’s silvern
landscapes, with their grey-greens and browns, dream still of the
spiritual of the Old Masters; but Courbet and Manet conquer bare
physical space, “factual” space. The meditative discoverer
represented by Leonardo gives way to the painting experimentalist.
Corot, the eternal child, French but not Parisian, finds his
transcendent landscapes anywhere and everywhere; Courbet,
Manet, Cézanne, portray over and over again, painfully, laboriously,
soullessly, the Forest of Fontainebleau, the bank of the Seine at
Argenteuil, or that remarkable valley near Arles. Rembrandt’s mighty
landscapes lie essentially in the universe, Manet’s near a railway
station. The plein-air painters, true megalopolitans, obtain as it were
specimens of the music of space from the least agitated sources of
Spain and Holland—from Velasquez, Goya, Hobbema, Franz Hals—
in order (with the aid of English landscapists and, later, the
Japanese, “highbrows” all) to restate it in empirical and scientific
terms. It is natural science as opposed to nature experience, head
against heart, knowledge in contrast to faith.
In Germany it was otherwise. Whereas in France it was a matter of
closing-off the great school, in Germany it was a case of catching up
with it. For in the picturesque style, as practised from Rottmann,
Wasmann, K. D. Friedrich and Runge to Marées and Leibl, an
unbroken evolution is the very basis of technique, and even a new-
style school requires a closed tradition behind it. Herein lies the
weakness and the strength of the last German painters. Whereas the
French possessed a continuous tradition of their own from early
Baroque to Chardin and Corot, whereas there was living connexion
between Claude Lorrain and Corot, Rubens and Delacroix, all the
great Germans of the 18th Century had been musicians. After
Beethoven this music, without change of inward essence, was
diverted (one of the modalities of the German Romantic movement)
back into painting. And it was in painting that it flowered longest and
bore its kindliest fruits, for the portraits and landscapes of these men
are suffused with a secret wistful music, and there is a breath of
Eichendorff and Mörike left even in Thoma and Böcklin. But a foreign
teacher had to be asked to supply that which was lacking in the
native tradition, and so these painters one and all went to Paris,
where they studied and copied the old masters of 1670. So also did
Manet and his circle. But there was this difference, that the
Frenchmen found in these studies only reminiscences of something
that had been in their art for many generations, whereas the
Germans received fresh and wholly different impressions. The result
was that, in the 19th Century, the German arts of form (other than
music) were a phenomenon out of season—hasty, anxious,
confused, puzzled as to both aim and means. There was indeed no
time to be lost. The level that German music or French painting had
taken centuries to attain had to be made good by German painting in
two generations. The expiring art demanded its last phase, and this
phase had to be reached by a vertiginous race through the whole
past. Hence the unsteadiness, in everything pertaining to form, of
high Faustian natures like Marées and Böcklin, an unsteadiness that
in German music with its sure tradition (think of Bruckner) would
have been impossible. The art of the French Impressionists was too
explicit in its programme and correspondingly too poor in soul to
expose them to such a tragedy. German literature, on the contrary,
was in the same condition as German painting; from Goethe’s time,
every major work was intended to found something and obliged to
conclude something. Just as Kleist felt in himself both Shakespeare
and Stendhal, and laboured desperately, altering and discarding
without end and without result, to forge two centuries of
psychological art into a unit; just as Hebbel tried to squeeze all the
problems from Hamlet to Rosmersholm into one dramatic type; so
Menzel, Leibl, and Marées sought to force the old and new models—
Rembrandt, Claude, Van Goyen, and Watteau, Delacroix, Courbet
and Manet—into a single form. While the little early interiors of
Menzel anticipated all the discoveries of the Manet circle and Leibl
not seldom succeeded where Courbet tried and failed, their pictures
renew the metaphysical browns and greens of the Old Masters and
are fully expressive of an inward experience. Menzel actually re-
experienced and reawakened something of Prussian Rococo,
Marées something of Rubens, Leibl in his “Frau Gedon” something
of Rembrandt’s portraiture. Moreover, the studio-brown of the 17th
Century had had by its side a second art, the intensely Faustian art
of etching. In this, as in the other, Rembrandt is the greatest master
of all time; this, like the other, has something Protestant in it that puts
it in a quite different category from the work of the Southern Catholic
painters of blue-green atmospheres and the Gobelin tapestries. And
Leibl, the last artist in the brown, was the last great etcher whose
plates possess that Rembrandtesque infinity that contains and
reveals secrets without end. In Marées, lastly, there was all the
mighty intention of the great Baroque style, but, though Guéricault
and Daumier were not too belated to capture it in positive form, he—
lacking just that strength that a tradition would have given him—was
unable to force it into the world of painter’s actuality.
IX
The last of the Faustian arts died in “Tristan.” This work is the giant
keystone of Western music. Painting achieved nothing like this as a
finale—on the contrary, the effect of Manet, Menzel and Leibl, with
their combination of “free light” and resurrected old-master styles, is
weak.
“Contemporaneously,” in our sense, Apollinian art came to its end
in Pergamene sculpture. Pergamum is the counterpart of Bayreuth.
The famous altar itself,[365] indeed, is later, and probably not the most
important work of the epoch at that; we have to assume a century
(330-220 B.C.) of development now lost in oblivion. Nevertheless, all
Nietzsche’s charges against Wagner and Bayreuth, the “Ring” and
“Parsifal”—decadence, theatricalness and the like—could have been
levelled in the same words at the Pergamene sculpture. A
masterpiece of this sculpture—a veritable “Ring”—has come down to
us in the Gigantomachia frieze of the great altar. Here is the same
theatrical note, the same use of motives from ancient discredited
mythology as points d’appui, the same ruthless bombardment of the
nerves, and also (though the lack of inner power cannot altogether
be concealed) the same fully self-conscious force and towering
greatness. To this art the Farnese Bull and the older model of the
Laocoön group certainly belong.
The symptom of decline in creative power is the fact that to
produce something round and complete the artist now requires to be
emancipated from form and proportion. Its most obvious, though not
its most significant, manifestation is the taste for the gigantic. Here
size is not, as in the Gothic and the Pyramid styles, the expression of
inward greatness, but the dissimulation of its absence. This
swaggering in specious dimensions is common to all nascent
Civilizations—we find it in the Zeus altar of Pergamum, the Helios of
Chares called the “Colossus of Rhodes,” the architecture of the
Roman Imperial Age, the New Empire work in Egypt, the American
skyscraper of to-day. But what is far more indicative is the
arbitrariness and immoderateness that tramples on and shatters the
conventions of centuries. In Bayreuth and in Pergamum, it was the
superpersonal Rule, the absolute mathematic of Form, the Destiny
immanent in the quietly-matured language of a great art, that was
found to be intolerable. The way from Polycletus to Lysippus and
from Lysippus to the sculptors of the groups of Gauls[366] is paralleled
by the way from Bach, by Beethoven, to Wagner. The earlier artists
felt themselves masters, the later uneasy slaves, of the great form.
While even Praxiteles and Haydn were able to speak freely and gaily
within the limits of the strictest canon, Lysippus and Beethoven could
only produce by straining their voices. The sign of all living art, the
pure harmony of “will,” “must” and “can,” the self-evidence of the
aim, the un-self-consciousness of the execution, the unity of the art
and the Culture—all that is past and gone. In Corot and Tiepolo,
Mozart and Cimarosa, there is still a real mastery of the mother-
tongue. After them, the process of mutilation begins, but no one is
conscious of it because no one now can speak it fluently. Once upon
a time, Freedom and Necessity were identical; but now what is
understood by freedom is in fact indiscipline. In the time of
Rembrandt or Bach the “failures” that we know only too well were
quite unthinkable. The Destiny of the form lay in the race or the
school, not in the private tendencies of the individual. Under the spell
of a great tradition full achievement is possible even to a minor artist,
because the living art brings him in touch with his task and the task
with him. To-day, these artists can no longer perform what they
intend, for intellectual operations are a poor substitute for the trained
instinct that has died out. All of them have experienced this. Marées
was unable to complete any of his great schemes. Leibl could not
bring himself to let his late pictures go, and worked over them again
and again to such an extent that they became cold and hard.
Cézanne and Renoir left work of the best quality unfinished because,
strive as they would, they could do no more. Manet was exhausted
after he had painted thirty pictures, and his “Shooting of the Emperor
Maximilian,” in spite of the immense care that is visible in every item
of the picture and the studies for it, hardly achieved as much as
Goya managed without effort in its prototype the “shootings of the
3rd of May.” Bach, Haydn, Mozart and a thousand obscure
musicians of the 18th Century could rapidly turn out the most
finished work as a matter of routine, but Wagner knew full well that
he could only reach the heights by concentrating all his energy upon
“getting the last ounce” out of the best moments of his artistic
endowment.
Between Wagner and Manet there is a deep relationship, which is
not, indeed, obvious to everyone but which Baudelaire with his
unerring flair for the decadent detected at once. For the
Impressionists, the end and the culmination of art was the conjuring
up of a world in space out of strokes and patches of colour, and this
was just what Wagner achieved with three bars. A whole world of
soul could crowd into these three bars. Colours of starry midnight, of
sweeping clouds, of autumn, of the day dawning in fear and sorrow,
sudden glimpses of sunlit distances, world-fear, impending doom,
despair and its fierce effort, hopeless hope—all these impressions
which no composer before him had thought it possible to catch, he
could paint with entire distinctness in the few tones of a motive. Here
the contrast of Western music with Greek plastic has reached its
maximum. Everything merges in bodiless infinity, no longer even
does a linear melody wrestle itself clear of the vague tone-masses
that in strange surgings challenge an imaginary space. The motive
comes up out of dark terrible deeps. It is flooded for an instant by a
flash of hard bright sun. Then, suddenly, it is so close upon us that
we shrink. It laughs, it coaxes, it threatens, and anon it vanishes into
the domain of the strings, only to return again out of endless
distances, faintly modified and in the voice of a single oboe, to pour
out a fresh cornucopia of spiritual colours. Whatever this is, it is
neither painting nor music, in any sense of these words that attaches
to previous work in the strict style. Rossini was asked once what he
thought of the music of the “Huguenots”; “Music?” he replied. “I
heard nothing resembling it.” Many a time must this judgment have
been passed at Athens on the new painting of the Asiatic and
Sicyonian schools, and opinions not very different must have been
current in Egyptian Thebes with regard to the art of Cnossus and
Tell-el-Amarna.
All that Nietzsche says of Wagner is applicable, also, to Manet.
Ostensibly a return to the elemental, to Nature, as against
contemplation-painting (Inhaltsmalerei) and abstract music, their art
really signifies a concession to the barbarism of the Megalopolis, the
beginning of dissolution sensibly manifested in a mixture of brutality
and refinement. As a step, it is necessarily the last step. An artificial
art has no further organic future, it is the mark of the end.
And the bitter conclusion is that it is all irretrievably over with the
arts of form of the West. The crisis of the 19th Century was the
death-struggle. Like the Apollinian, the Egyptian and every other, the
Faustian art dies of senility, having actualized its inward possibilities
and fulfilled its mission within the course of its Culture.
What is practised as art to-day—be it music after Wagner or
painting after Cézanne, Leibl and Menzel—is impotence and
falsehood. Look where one will, can one find the great personalities
that would justify the claim that there is still an art of determinate
necessity? Look where one will, can one find the self-evidently
necessary task that awaits such an artist? We go through all the
exhibitions, the concerts, the theatres, and find only industrious
cobblers and noisy fools, who delight to produce something for the
market, something that will “catch on” with a public for whom art and
music and drama have long ceased to be spiritual necessities. At
what a level of inward and outward dignity stand to-day that which is
called art and those who are called artists! In the shareholders’
meeting of any limited company, or in the technical staff of any first-
rate engineering works there is more intelligence, taste, character
and capacity than in the whole music and painting of present-day
Europe. There have always been, for one great artist, a hundred
superfluities who practised art, but so long as a great tradition (and
therefore great art) endured even these achieved something worthy.
We can forgive this hundred for existing, for in the ensemble of the
tradition they were the footing for the individual great man. But to-
day we have only these superfluities, and ten thousand of them,
working art “for a living” (as if that were a justification!). One thing is
quite certain, that to-day every single art-school could be shut down
without art being affected in the slightest. We can learn all we wish to
know about the art-clamour which a megalopolis sets up in order to
forget that its art is dead from the Alexandria of the year 200. There,
as here in our world-cities, we find a pursuit of illusions of artistic
progress, of personal peculiarity, of “the new style,” of “unsuspected
possibilities,” theoretical babble, pretentious fashionable artists,
weight-lifters with cardboard dumb-bells—the “Literary Man” in the
Poet’s place, the unabashed farce of Expressionism which the art-
trade has organized as a “phase of art-history,” thinking and feeling
and forming as industrial art. Alexandria, too, had problem-
dramatists and box-office artists whom it preferred to Sophocles, and
painters who invented new tendencies and successfully bluffed their
public. What do we possess to-day as "art"? A faked music, filled
with artificial noisiness of massed instruments; a faked painting, full
of idiotic, exotic and showcard effects, that every ten years or so
concocts out of the form-wealth of millennia some new “style” which
is in fact no style at all since everyone does as he pleases; a lying
plastic that steals from Assyria, Egypt and Mexico indifferently. Yet
this and only this, the taste of the “man of the world,” can be
accepted as the expression and sign of the age; everything else,
everything that “sticks to” old ideals, is for provincial consumption.
The grand Ornamentation of the past has become as truly a dead
language as Sanskrit or Church Latin.[367] Instead of its symbolism
being honoured and obeyed, its mummy, its legacies of perfected
forms, are put into the pot anyhow, and recast in wholly inorganic
forms. Every modern age holds change to be development, and puts
revivals and fusions of old styles in the place of real becoming.
Alexandria also had its Pre-Raphaelite comedians with their vases,
chairs, pictures and theories, its symbolists, naturalists and
expressionists. The fashion at Rome was now Græco-Asiatic, now
Græco-Egyptian, now (after Praxiteles) neo-Attic. The relief of the
XIXth Dynasty—the modern age in the Egyptian Culture—that
covered the monstrous, meaningless, inorganic walls, statues and
columns, seems like a sheer parody of the art of the Old Kingdom.
The Ptolemaic Horus-temple of Edfu is quite unsurpassed in the way
of vacuous eclecticism—so far, for we are only at the beginning of
our own development in this line, showy and assertive as the style of
our streets and squares already is.
In due course, even the strength to wish for change fades out.
Rameses the Great—so soon—appropriated to himself buildings of
his predecessors by cutting out their names and inserting his own in
the inscriptions. It was the same consciousness of artistic impotence
that led Constantine to adorn his triumphal arch in Rome with
sculptures taken from other buildings; but Classical craftsmanship
had set to work long before Constantine—as early, in fact, as 150—
on the business of copying old masterpieces, not because these
were understood and appreciated in the least, but because no one
was any longer capable of producing originals. It must not be
forgotten that these copyists were the artists of their time; their work
therefore (done in one style or another according to the moment’s
fashion) represent the maximum of creative power then available. All
the Roman portrait statues, male and female, go back for posture
and mien to a very few Hellenic types; these, copied more or less
true to style, served for torsos, while the heads were executed as
“Likenesses” by simple craftsmen who possessed the knack. The
famous statue of Augustus in armour, for example, is based on the
Spearman of Polycletus, just as—to name the first harbingers of the
same phase in our own world—Lenbach rests upon Rembrandt and
Makart upon Rubens. For 1500 years (Amasis I to Cleopatra)
Egypticism piled portrait on portrait in the same way. Instead of the
steady development that the great age had pursued through the Old
and Middle Kingdoms, we find fashions that change according to the
taste of this or that dynasty. Amongst the discoveries at Turfan are
relics of Indian dramas, contemporary with the birth of Christ, which
are similar in all respects to the Kalidasa of a later century. Chinese
painting as we know it shows not an evolution but an up-and-down of
fashions for more than a thousand years on end; and this
unsteadiness must have set in as early as the Han period. The final
result is that endless industrious repetition of a stock of fixed forms
which we see to-day in Indian, Chinese, and Arabian-Persian art.
Pictures and fabrics, verses and vessels, furniture, dramas and
musical compositions—all is patternwork.[368] We cease to be able to
date anything within centuries, let alone decades, by the language of
its ornamentation. So it has been in the Last Act of all Cultures.
CHAPTER IX
SOUL-IMAGE AND LIFE-FEELING
I
ON THE FORM OF THE SOUL
CHAPTER IX
SOUL-IMAGE AND LIFE-FEELING
I
ON THE FORM OF THE SOUL
I
Every professed philosopher is forced to believe, without serious

examination, in the existence of a Something that in his opinion is
capable of being handled by the reason, for his whole spiritual
existence depends on the possibility of such a Something. For every
logician and psychologist, therefore, however sceptical he may be,
there is a point at which criticism falls silent and faith begins, a point
at which even the strictest analytical thinker must cease to employ
his method—the point, namely, at which analysis is confronted with
itself and with the question of whether its problem is soluble or even
exists at all. The proposition “it is possible by thought to establish the
forms of thought” was not doubted by Kant, dubious as it may
appear to the unphilosophical. The proposition “there is a soul, the
structure of which is scientifically accessible; and that which I
determine, by critical dissection of conscious existence-acts into the
form of psychic elements, functions, and complexes, is my soul” is a
proposition that no psychologist has doubted hitherto. And yet it is
just here that his strongest doubts should have arisen. Is an abstract
science of the spiritual possible at all? Is that which one finds on this
path identical with that which one is seeking? Why has psychology—
meaning thereby not knowledge of men and experience of life but
scientific psychology—always been the shallowest and most
worthless of the disciplines of philosophy, a field so empty that it has
been left entirely to mediocre minds and barren systematists? The
reason is not far to seek. It is the misfortune of “experimental”
psychology that it does not even possess an object as the word is
understood in any and every scientific technique. Its searches and
solutions are fights with shadows and ghosts. What is it—the Soul? If
the mere reason could give an answer to that question, the science
would be ab initio unnecessary.
Of the thousands of psychologists of to-day not one can give an
actual analysis or definition of “the” Will—or of regret, anxiety,
jealousy, disposition, artistic intention. Naturally, since only the
systematic can be dissected, and we can only define notions by
notions. No subtleties of intellectual play with notional distinctions, no
plausible observations of connexions between sensuous-corporeal
states and “inward processes” touch that which is in question here.
Will—this is no notion, but a name, a prime-word like God, a sign for
something of which we have an immediate inward certainty but
which we are for ever unable to describe.
We are dealing here with something eternally inaccessible to
learned investigation. It is not for nothing that every language
presents a baffling complexity of labels for the spiritual, warning us
thereby that it is something not susceptible of theoretical synthesis or
systematic ordering. Here there is nothing for us to order. Critical
(i.e., literally, separating) methods apply only to the world-as-Nature.
It would be easier to break up a theme of Beethoven with dissecting-
knife or acid than to break up the soul by methods of abstract
thought. Nature-knowledge and man-knowledge have neither aims
nor ways in common. The primitive man experiences “soul,” first in
other men and then in himself, as a Numen, just as he knows
numina of the outer world, and develops his impressions in
mythological form. His words for these things are symbols, sounds,
not descriptive of the indescribable but indicative of it for him who
hath ears to hear. They evoke images, likenesses (in the sense of
Faust II)—the only language of spiritual intercourse that man has
discovered to this day. Rembrandt can reveal something of his soul,
to those who are in inward kinship with him, by way of a self-portrait
or a landscape, and to Goethe “a god gave it to say what he
suffered.” Certain ineffable stirrings of soul can be imparted by one
man to the sensibility of another man through a look, two bars of a
melody, an almost imperceptible movement. That is the real
language of souls, and it remains incomprehensible to the outsider.
The word as utterance, as poetic element, may establish the link, but
the word as notion, as element of scientific prose, never.
“Soul,” for the man who has advanced from mere living and feeling
to the alert and observant state, is an image derived from quite

After Certainty A History of Our Epistemic Ideals and Illusions Pasnau Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

After Certainty A History of Our Epistemic Ideals and Illusions Pasnau Full Chapter

Uploaded by

Copyright:

Available Formats

After Certainty: A History of Our

Epistemic Ideals and Illusions Pasnau

My earliest ambition as a philosopher, which goes back to my undergraduate days, was

Lecture One: The Epistemic Ideal 1

The Epistemic Ideal

Still, it is by no means the case that premodern philosophy neglected epistemology.

ancient epistemic expectations coalesced in the Middle Ages around a comprehensive

Aristotle’s Ideal Theory

in normative expectations. Even if Whewell’s enthusiasm is misplaced, the phenomenon

The Breakdown of Aristotelian Essentialism

Trading Depth for Precision

Epistemology in the Ascendant

Descartes’s Ideal Theory

The Ideal of Certainty

As a start at an analysis, we can distinguish three entwined notions:

There is a word, “Impressionism,” which only came into general

Every professed philosopher is forced to believe, without serious

You might also like