Medieval and Early Modern Epistemology (Proceedings of the Society for Medieval Logic and Metaphysic) [1 ed.] 1527542734, 9781527542730

Table of contents :
Table of Contents
Introduction
Overview of After Certainty
John Buridan Being After Certainty
Perception: Reductivism and Relationalism
Indirect Realism
Certainty: Last Champions and Beyond
Beyond the Ideal, the Social?
Some Thoughts after After Certainty
Appendix
Contributors

Citation preview

Medieval and Early Modern Epistemology

Also available in the series: The Immateriality of the Human Mind, the Semantics of Analogy, and the Conceivability of God Volume 1: Proceedings of the Society for Medieval Logic and Metaphysics Categories, and What Is Beyond Volume 2: Proceedings of the Society for Medieval Logic and Metaphysics Knowledge, Mental Language, and Free Will Volume 3: Proceedings of the Society for Medieval Logic and Metaphysics Mental Representation Volume 4: Proceedings of the Society for Medieval Logic and Metaphysics Universal Representation, and the Ontology of Individuation Volume 5: Proceedings of the Society for Medieval Logic and Metaphysics Medieval Skepticism, and the Claim to Metaphysical Knowledge Volume 6: Proceedings of the Society for Medieval Logic and Metaphysics Medieval Metaphysics; or Is It ‘‘Just Semantics’’? Volume 7: Proceedings of the Society for Medieval Logic and Metaphysics After God, with Reason Alone---Saikat Guha Commemorative Volume Volume 8: Proceedings of the Society for Medieval Logic and Metaphysics The Demonic Temptations of Medieval Nominalism Volume 9: Proceedings of the Society for Medieval Logic and Metaphysics Skepticism, Causality and Skepticism about Causality Volume 10: Proceedings of the Society for Medieval Logic and Metaphysics Metaphysical Themes, Medieval and Modern Volume 11: Proceedings of the Society for Medieval Logic and Metaphysics Maimonides on God and Duns Scotus on Logic and Metaphysics Volume 12: Proceedings of the Society for Medieval Logic and Metaphysics The Metaphysics of Personal Identity Volume 13: Proceedings of the Society for Medieval Logic and Metaphysics Consciousness and Self---Knowledge in Medieval Philosophy Volume 14: Proceedings of the Society for Medieval Logic and Metaphysics Hylomorphism and Mereology Volume 15: Proceedings of the Society for Medieval Logic and Metaphysics Being, Goodness and Truth Volume 16: Proceedings of the Society for Medieval Logic and Metaphysics

Medieval and Early Modern Epistemology: After Certainty (Volume 17: Proceedings of the Society for Medieval Logic and Metaphysics) Edited by

Alex Hall, Gyula Klima and Martin Klein

Medieval and Early Modern Epistemology: After Certainty (Volume 17: Proceedings of the Society for Medieval Logic and Metaphysics) Series: Proceedings of the Society for Medieval Logic and Metaphysics Edited by Alex Hall, Gyula Klima and Martin Klein This book first published 2020 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2020 by Alex Hall, Gyula Klima, Martin Klein and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-5275-4273-4 ISBN (13): 978-1-5275-4273-0



TABLE OF CONTENTS

Introduction ................................................................................................. 1 Martin Klein Overview of After Certainty ........................................................................ 5 Robert Pasnau John Buridan Being After Certainty .......................................................... 21 Martin Klein Perception: Reductivism and Relationalism .............................................. 37 Sebastian Bender Indirect Realism .............................................................................................. 47 Sabine van Enckevort and Han Thomas Adriaenssen Certainty: Last Champions and Beyond .................................................... 57 Stephan Schmid Beyond the Ideal, the Social? .................................................................... 71 Christophe Grellard Some Thoughts after After Certainty ......................................................... 87 Robert Pasnau Appendix ................................................................................................. 105 Contributors ............................................................................................. 107

INTRODUCTION MARTIN KLEIN

This volume had its origin in a workshop in Paris in 2018 which brought together Sebastian Bender, Christophe Grellard, Martin Klein, Stephan Schmid, and Jacob Schmutz in order to discuss with Robert Pasnau his most recent monograph, which had just appeared on the market: After Certainty: A History of Our Epistemic Ideals and Illusions (Oxford: Oxford University Press, 2017).1 Immediately after our meeting we considered the idea of a publication based on the symposium, and I am very much indebted to Gyula Klima and Alex Hall for this opportunity to present it in the Proceedings of the Society for Medieval Logic and Metaphysics. After Certainty tells the history of epistemology and its pitfalls from Aristotle to the present day. The main text of the book originated with the Isaiah Berlin Lectures which Pasnau was invited to give in Oxford in 2014. But the project expanded: in the book, the original six lectures are supported by long and detailed endnotes. How to navigate through this monumental work and what are its core theses is laid out for each lecture at the beginning of the present volume by the author himself. Pasnau’s overview of After Certainty is then followed by five comments which cover almost all of the lectures. Martin Klein concentrates in his comment on John Buridan, who plays a pivotal role in After Certainty in at least two respects. In Pasnau’s Lecture Two (“Evident Certainties”), Buridan is important in the transition from medieval to early modern epistemology for introducing the idea that evidentness and certainty can come in degrees. Klein thinks—more than Pasnau does—that, according to Buridan, what is later called “moral certainty” is enough for knowledge and is not too different from the evidentness required for the natural sciences. However, with regard to

 1

Referred to throughout this volume as After Certainty, followed by the relevant lecture and page numbers.

2

Introduction

Buridan’s understanding of the distinction between per se sensibles and proper sensibles and the extent to which he allows the senses to have a privileged access to sensible reality, Klein considers Buridan’s innovations less significant than Pasnau does in his Lecture Three (“The Sensory Domain”). Sebastian Bender’s “Reductivism and Relationalism” also concentrates on Lecture Three, scrutinizing Pasnau’s distinction between a reductivist and a relationist path in seventeenth-century theories of perception. Reductivists take it to be literally the case that we represent geometrickinetic patterns when we perceive a colored object—a quite implausible account, as Pasnau suggests in his lecture. Relationalists, on the other hand, seem to be better off in holding that the idea of a color does not represent geometric-kinetic patterns, but nevertheless represents something real in the material world, namely, a power. Bender objects that a closer look at Leibniz and Spinoza shows that the reductive account is less implausible than it initially seems, while the relationalist account seems less attractive, given that it depends on an arbitrary divine coordination of sensation with what is sensed. In “Indirect Realism,” Sabine van Enckevort and Han Thomas Adriaenssen comment on Lecture Four (“Ideas and Illusions”). Are the immediate objects of perception fundamentally mind-dependent entities? If they are, how does such a view relate to skeptical worries? If there is one figure before the seventeenth century who seems to have defended such a claim, it is Peter Auriol. Nevertheless, Auriol still tries to defend a kind of identity between the extramental real being of a cognized object and the “apparent being” the object takes on in being cognized. As Enckevort and Adriaenssen show, it is not clear how Auriol can defend this claim, since neither numerical nor specific identity seems to be available to him, even by his own theoretical standards. Regarding the second question, Enckevort and Adriaenssen reject the claim that “indirect realism somehow generates skeptical questions that direct realism does not” (p. 56). Rather, what leads to skepticism is not indirect realism, but the distinction between primary and secondary qualities (underlying both skepticism and indirect realism). In Lecture Six (“Deception and Hope”), Pasnau considers some of the lessons we can learn from the history of epistemology. If certainty can no longer be the ideal of knowledge, and yet we want to avoid falling into Humean epistemic defeatism but rather to take an optimistic stance, what

Martin Klein

3

could this optimistic stance be? Pasnau’s answer is that the best epistemic attitude we might find lies not on the rational side of the mind, but rather on the affective: he proposes that the right epistemic attitude is one of hope. To believe out of hope is distinguished from the three traditional epistemic attitudes of knowledge, opinion, and faith. It leaves behind the unattainable ideal of certainty, but this lack of certainty does not imply the fear that one’s belief is wrong, nor does it require cognitive confidence. With hope, as Pasnau puts it in his overview in this volume, “without elevating one’s credence through faith, one might simply stop fearing that one is wrong” (p. 19). Concentrating on Lecture Six, the last two comments in this volume ask how we are to understand Pasnau’s remedy of believing out of hope and whether there are other options for us to pursue beyond certainty. Stephan Schmid, first of all calling into question Pasnau’s narrative, is not so sure where to draw the line in identifying the last champions of certainty. While Pasnau opts for Descartes, Schmid advised us not to overlook the tradition of rationalism and German idealism. Schmid is also critical of Pasnau’s conception of epistemic hope, if this attitude does not require some sort of cognitive justification; if it does, however, then it is unclear to him how this position differs from Hume’s. Christophe Grellard takes up the idea of social epistemology. In “Beyond the Ideal, the Social?,” Grellard first emphasizes that skepticism should not be underestimated as a real problem in the history of epistemology, especially in the Middle Ages. Pasnau’s conception of hope might introduce new skeptical worries, since it seems that hope could introduce “other forms of error (in the form of self-deception for example)” (p. 78). More importantly, Grellard raises doubts about how the affective aspects of hope could be more helpful than the cognitive aspects which hope might share with faith. As an alternative to distinguishing hope from faith, Grellard suggests that we focus on the theory of faith in its connection with moral certainty. He suggests that late medieval theories of faith introduced “the idea that it is not by an individual process that I can secure my action, but by relying on a principle of trust that is warranted collectively” (p. 84). This volume concludes with “Some Thoughts after After Certainty,” in which Robert Pasnau replies to the comments individually.

4

Introduction

I had the pleasure of organizing the original meeting in Paris together with Christophe Grellard (EPHE) and Jacob Schmutz (Sorbonne Université) during my stay as a visiting research fellow at the Centre Pierre Abélard (Sorbonne Université) and the Laboratoire d’Études sur les Monothéismes (EPHE) with the financial support of my home institution at that time, the Topoi Excellence Cluster in Berlin. I am very grateful to Christophe and Jacob for their kind invitation to Paris, and in particular for giving me a free hand in the conception of the workshop while they took care of all the rest. The symposium became possible thanks to the generous funding of various Parisian institutions: Paris Sciences & Lettres, the École Pratique des Hautes Études, the Laboratoire d’Études sur les Monothéismes UMR 8584, the Sorbonne Université, and the Institut Universitaire de France. The Topoi Excellence Cluster provided funding for the publication of this volume, enabling it to benefit from the careful work of a language editor, Ian Drummond, who also translated Christophe Grellard’s contribution. Finally, I would like to thank both Robert Pasnau and all the commentators for participating in this discussion about the history of epistemology.

OVERVIEW OF AFTER CERTAINTY ROBERT PASNAU

Back in the mid-aughts, when I was doing the research for what would become my previous book,1 I thought that I was working on a single work which would encompass metaphysics and epistemology. But eventually the metaphysics overwhelmed everything else, and I set aside my research on epistemology. That changed when I was invited to give the Isaiah Berlin Lectures in Oxford in 2014. After Certainty, the subject of the symposium of which the present volume is a record, is a considerably augmented version of those lectures. The heart of After Certainty is six chapters, each of which is the descendant of one of the lectures I gave in Oxford. After the lectures comes a collection of endnotes of such length that they are longer than the main text of the book. In an attempt to make this structure more readerfriendly, I wrote the endnotes so that they could be read continuously, separately from the main lectures, with the idea that readers would first read the lecture itself without interruption, and then turn to the endnotes if they wanted further scholarly details. The range of the book is extremely wide, as will be evident from the breadth of the issues raised by the commentators in this volume. The book goes back to Aristotle, whom I discuss at some length in several of the chapters, and runs all the way forward to Hume, who also gets discussed at some length. Occasionally I make quite sweeping claims about the whole history of philosophy, for which some of the commentators take me to task. The book also differs from my earlier books in attempting to reach conclusions that are not just historical but also systematic. This is something I did not attempt in Metaphysical Themes, in part because I find myself lacking in many convictions about metaphysics. I find Aristotelianism tempting, but I also find Humean reductionism tempting.

 1

Robert Pasnau, Metaphysical Themes 1274–1671 (Oxford: Clarendon Press, 2011).

6

Overview of After Certainty

Topics in metaphysics strike me as fascinating but wholly bewildering. In epistemology by contrast, I am much more opinionated, and my opinions come forth at various points in the book, as I will try to indicate briefly in this overview. The book begins on an autobiographical note, with my long struggle to write about medieval epistemology. I, like many others, have written about medieval cognitive theory, and about medieval philosophy of mind. But what exactly is medieval epistemology? This query is connected to the observation I make at the very beginning of the book, with regard to the history of epistemology, that “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” (lect. 1, p. 1). The idea that finally made it possible for me to write systematically about the history of epistemology is the idea of an idealized epistemology. Rather than taking 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 second to chart the various ways in which we commonly fall short of that ideal. Although epistemology today does not conceive of itself in this way, it seems to me that we can understand much of the theorizing about knowledge that runs from antiquity into the modern era by thinking of it in these terms. Once we do, we can also see how, by and large, the history of philosophy is the record of the gradual diminishment of that ideal, as we reconciled ourselves over the centuries to lower and lower expectations. Lecture One begins with Aristotle’s project in the Posterior Analytics. On my account, it should be understood not as an attempt to characterize scientific method, but rather as an ideal theory, or what it would look like to frame a body of knowledge in the best possible way – best possible, that is, for beings such as us, in the kind of world we live in. This last qualification introduces an important theme of the book: that epistemology, pursued in this way, inevitably goes hand in hand with metaphysics, because one can hardly begin to think about the epistemic ideal without some view about what the ideal objects of knowledge are, which in turn requires a metaphysical decision about whether one is a Platonist, an Aristotelian, a nominalist, and so on. The Aristotelian concept of epistƝmƝ eventually makes its way into Latin as a theory of scientia, and dominates much of the history of philosophy up until the seventeenth century. Throughout, it is understood as a

Robert Pasnau

7

statement of ideal theory. As Albert the Great remarks of scientia in the prologue to his Posterior Analytics commentary (1261), “This is the end and the most perfect and the sole unconditionally desirable thing among the logical sciences.”2 A critical part of the notion of an idealized epistemology is that it attempts to describe not a perfect ideal, of the sort that a god might achieve, but rather an ideal calibrated to us, given the nature of our cognitive powers. An ideal of this sort might seem to be just a descriptive project, but it is also normative. In seeking to establish the cognitive ideal, the theory aims at answering a question that lies at the heart of epistemology: When should ordinary agents, in ordinary circumstances, believe the things they believe? This way of proceeding insists that a normative account of our epistemic position, non-ideal as it is, presupposes ideal theory. Thus, we can speak of a normative ideal. Here I make a systematic and not merely historical suggestion: that this way of engaging in epistemology is likely to be more fruitful than the project of finding the necessary and sufficient conditions for knowledge. Indeed, inquiry into the ideal might lead us to modify what we mean when we talk about “knowledge,” so that epistemology, rather than simply describing how we use our epistemic language, might be in a position to revise that usage. With this picture of an idealized epistemology in place, the rest of the book traces how various ideals established in antiquity and the Middle Ages were gradually abandoned in later centuries. In Lecture One, the focus is on the Aristotelian ideal that knowledge requires the grasp of a thing’s essence. This is a demand found in the Posterior Analytics itself, and it is very prominent among the scholastics, coming to lie at the heart of their causal theories. The theory gets challenged in the seventeenth century in various ways, of which two are most prominent. First, there is the Lockean challenge: that although there are real essences, they are undiscoverable and, moreover, they lack the explanatory generality that the Aristotelians had assumed. Second, there is the challenge to the very idea that the epistemic goal should be to grasp the underlying causes of things. First in Galileo and then in Newton, it came to be a hallmark of the new science to refuse to enter into metaphysical speculation about essences and other such deep causes.

 2

Albert the Great, Analytica posteriora I.1.1 (ed. P. Jammy, Beati Alberti Magni Opera [Lyon, 1651], vol. 1, p. 514a): “Est ergo finis et perfectissima et sola simpliciter desiderabilis inter logicas scientias.”

8

Overview of After Certainty

In place of this ideal, Galileo and especially Newton adopted the ideal of accuracy, an ambition described most famously in Newton’s preface to the Principia. This in turn has been immensely important to our modern conception of the epistemic ideal, particularly in those domains we now think of as sciences. Here is where science begins to form its identity as a discipline distinct from philosophy, defined in part in terms of a distinct ideal. It even becomes possible to think that science may be possible where knowledge is not, as in Hume, who takes himself to be constructing a science of human nature. So where does that leave philosophy? On the one hand, it could turn its back on these new scientific ideals, and continue its pursuit of deep causal explanations, not as a chapter of science but as metaphysics. Leibniz is the shining early example of this approach, and his efforts at articulating 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. Within this English tradition, epistemology comes to be perceived as a foundational topic in philosophy, as it has remained to this day. Lecture Two begins with Descartes, and argues that he fits into the longer history of figures who pursue an idealized epistemology. Here I argue that we can understand Descartes’s quest for scientia as the pursuit of an ideal, characterized by certainty, foundationalism, and internalism. This is what the epistemic ideal requires. But, I argue, it is a mistake to think of Descartes as having a theory of knowledge. One indication of this is that he thinks no one has ever had the kind of scientia that Descartes seeks. If scientia is knowledge, then the absurd implication would be that Descartes is a complete skeptic regarding the history of human thought up until his own time. Another indication is that he does not think that scientia, or its lack, should regulate our beliefs. That puts too much distance between his epistemic ideal and our conception of knowledge for it to be plausible that he is offering us anything like a theory of knowledge. Perhaps the best known feature of Descartes’s epistemology is his emphasis on certainty. This raises the question of where that emphasis comes from. Surprisingly, this is an aspect of the ideal that Aristotle did not stress; however, it emerged as a theme in quite a few different places,

Robert Pasnau

9

notably among the Stoics, within the Alexandrian commentary tradition on Aristotle, and in Arabic philosophy. Indeed, the Arabic translation of the Posterior Analytics repeatedly uses the Arabic word for certainty (yaqƯn) where there is no corresponding word for certainty in the Greek text. It is not at all easy to work out the various notions of certainty that play a role in epistemology over the centuries. One important distinction, however, is between subjective certainty and objective certainty, where the first is simply a subjective sense of confidence, and the second the fixedness or necessity of the object or proposition in question. Neither of these notions of certainty, on its own, can represent the epistemic ideal. Rather, the two need to be linked in some way, so that the things we are highly confident of are the things that necessarily obtain. Beginning in the Middle Ages, the theoretical concept that most often did this work was evidentia, or evidentness. What is it to be evident? For scholastic authors, 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, evidentness 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. We can speak of an evident cognition, or we can speak of things in the world being evident to an individual. (Evidence in our modern sense is an extended sense of the term, which refers to the information in virtue of which a thing is evident to someone.) But what exactly is such evidentness? One of its central features, historically, is that it delivers indubitability. Ideally, evidentness would also yield infallibility, and so one of the perennial questions becomes whether the ideal goal of infallible evidentness can ever be realized. The most remarkable development of a theory of evidentness is that of John Buridan in the mid-fourteenth century. Buridan distinguishes three levels of evidentness:3

 3

John Buridan, Quaestiones in Metaphysicam II.1 (ed. Jodocus Badius, In Metaphysicen Aristotelis Quaestiones argutissimae Magistri Ioannis Buridani in ultima praelectione [Paris: Prelum Ascensianum, 1518; repr. Frankfurt a.M.: Minerva, 1964], fols. 8v–9r).

Overview of After Certainty

10 x x x

evidentia simpliciter (first principles and their logical consequences) evidentia naturalis (on the supposition that the common course of nature is observed) evidentia moralis (suffices for acting well morally)

For Buridan, the last of these is fallible, and indeed is compatible with the proposition being false. Since Buridan talks about moral evidentness explicitly in only one passage, and not at much length, various important questions of interpretation arise with regard to his view. It is not clear, for instance, what the connection is between this weak form of evidentness and morality. It is also not clear whether evidentia moralis, when concerned with a true proposition, amounts to scientia. Buridan himself does not use the phrase “moral certainty,” but this idea, so labeled, becomes very important in later scholasticism, and the word “moral” in this context comes to lose all association with morality, meaning instead almost, but not quite. The rise of interest in moral certainty corresponds to a growing tendency to regard the ideal of certainty as too remote a goal to be a normative epistemic ideal. Accordingly, when we get to the seventeenth century, we find a great many authors interested in the status of beliefs that are merely probable. This is not a wholly new development, but there is a new emphasis at this time on probability, particularly in English authors such as William Chillingsworth, John Tillotson, Joseph Glanvill, John Wilkins, and Robert Boyle. To be sure, these developments did not go uncriticized. Thomas White, for instance, criticized Glanvill, and by extension the whole Royal Society, for wanting to “tear science itself out of the hands of the learned, and throw it into the dirt of probability.”4 The critical figure, ultimately, becomes John Locke, who articulates the principle of proportionality: that we should believe to the degree our evidence suggests that a proposition is likely. But Locke persists in thinking of knowledge as an ideal, one which we can only rarely achieve. To understand how our current conception of knowledge arises out of these discussions, I look in some detail at the less well-known case of John

 4

Thomas White, An Exclusion of Scepticks from All Title to Dispute: Being an Answer to the Vanity of Dogmatizing (London: J. Williams, 1665), p. 55. For an important new study of probability in the later scholastic tradition, see Rudolf Schuessler, The Debate on Probable Opinions in the Scholastic Tradition (Leiden: Brill, 2019).

Robert Pasnau

11

Wilkins, who associates knowledge not with the highest ideal of certainty, but with mere moral certainty. This requires violating the principle of proportionality, because for Wilkins we are in effect entitled to believe fully in things that are not fully certain. Lecture Three turns to a special case of evidentness, the evidentness that arises from sensory perception. The chapter begins with an observation about how common it is to find the assumption, in a wide variety of camps, that there is some “domain of epistemic privilege” where the senses are perfectly (or nearly perfectly) reliable. One way to achieve such a domain—the subject of Lecture Four—is to look within, at our own ideas. Lecture Three, by contrast, focuses on those who find some sort of domain of reliability outside of us. In Aristotle, this domain is the proper sensibles. I argue in some detail that we should understand Aristotle as defending a relational view of the sensibles. Thus, if one wants to know what a sound is, Aristotle’s answer is that it is a power (a potentiality) to produce the experience of hearing. In general, instead of treating a sensible quality as entirely independent of perceivers, Aristotle makes its nature depend on its perceivers, whose actualization is the actualization of the sensible quality. In my terminology, rather than treating sensible qualities absolutely, he treats them relationally. This kind of theory is a concession to Democritus, and his talk of how hot and cold, sweet and bitter, and colors exist by convention. What is right about this, Aristotle thinks, is that these sensible qualities can be defined only by reference to our sensory powers. Yet Aristotle’s sort of relationalism, I argue, is not the view of later Aristotelians. On that later view, the sensible qualities are real, absolute, mind-independent features of reality. They are natural kinds, fundamental features of reality with which the senses interact, and which get detected in virtue of that interaction. This is what is known today as Aristotelian realism. Various aspects of Aristotelian realism get challenged by some medieval authors, most notably Nicole Oresme and John Buridan. Oresme speculates that the object of sight is not color but light—or, equivalently, that color just is the reflection of light off of objects, a strikingly modern view. Buridan calls into question the very idea of a privileged domain of sensible qualities. For one thing, he argues, contrary to Aristotle’s text, that the senses are as reliable when it comes to the common sensibles as

12

Overview of After Certainty

they are when it comes to the proper sensibles. Even more fundamentally, he denies that the senses have any privileged access to the proper sensibles (as opposed to various other features of objects). On the standard medieval view, the proper sensibles are privileged in virtue of an especially tight causal relationship between sensory experience and a set of qualities in the world (color, sound, flavor, odor, heat). But Buridan claims, to the contrary, that sensation presents in a confused way all of the various external sensible objects, without having any sort of privileged access to any of them and without even representing any of them as distinct from any other. On this picture, the very distinction between per se and per accidens perception collapses. All these standard aspects of the medieval theory end up being true, for Buridan, only insofar as we impose such distinctions at the intellectual level. Of course, it is in the seventeenth century that such doubts fully take center stage. At this point the standard scholastic story about perception becomes radically inverted: the proper sensibles are no longer the privileged sensory domain, but have been demoted to mere “secondary qualities”; instead, what is most real are the kinetic-geometric features of the world, now called the primary qualities. These become the domain where the senses get things right, if they get anything right at all. I focus in particular on Descartes’s complex views about perception. On my story, he is tempted by three different kinds of views about sensible qualities: a relationalism akin to Aristotle’s, a reductionism that would treat color and the rest as simply particles in motion, and a subjectivism that would identify sensible qualities with internal sensations. Ultimately, he rejects the first two for the same fundamental reason—namely, that they fail to respect an ideal that he, and others, tacitly take for granted, which is that perception should be not just reliable but also faithful in its depictions of reality. By “faithful” I mean that the senses should show the world as it is, in “high fidelity.” Holding out for this ideal, Descartes tends toward subjectivism. Relationalism gets taken up later in the century by Boyle and Locke. But they too fail to embrace it fully, and instead end up mixing it with a sort of subjectivism because they lack a metaphysics that would be adequate to make sense of relationalism. Even though they propose that sensible qualities should be understood as powers, their metaphysics is too reductive to allow for an ontology of powers.

Robert Pasnau

13

Lecture Four, accordingly, turns inward, to the way of ideas. It begins with the remark that the seventeenth century is much more exceptional here than is generally realized, for this is the only period in the history of philosophy in which ideas are widely taken to be the immediate objects of perception. The central question of the lecture is why this is so. To make progress on this question, it is helpful to distinguish between two claims. The first is the claim that there are inner representational vehicles distinct from acts of perception; I call this a dual view. The second is the claim that these inner representations (that is, ideas) are the objects of perception; I call this a mediated view. Although there is little doubt that both views are characteristic of the era, there is a lot of controversy about whether one or another seventeenth-century author holds a dual view or a mediated view. Many scholars of the period seem quite eager to rid their heroes of one or both of these views. This strikes me as strange, because really we ought to celebrate this feature of the period as one of its most distinctive and intriguing characteristics. How can it be, as Malebranche writes, that “tout le monde” accepts that external objects are not seen and that ideas are instead the immediate objects of perception, when hardly any philosophers of other eras can be found who endorse these claims? The central conclusion of the lecture is that much again turns on the expectation of fidelity. Without this, it is easy to conclude that the senses have as their object something in the external world, because there is no need to find something in the world that the senses show us in high fidelity. But the desire to achieve fidelity, when paired with the rise of mechanism, drives seventeenth-century authors toward mediated theories of perception, because the mechanical philosophy makes it natural to conclude that there is nothing out in the world that corresponds to phenomenal sensory experience. So the way of ideas, on my telling of the story, is wrapped up with the theory of secondary qualities. Thus Malebranche insists that Augustine himself would have held a mediated theory of perception but for his being in the grip of “the prejudice that colors are in objects.”5 And thus it later seems to Hume that “the fundamental principle” of the “modern” philosophy is “the opinion concerning colours, sounds, tastes, smells, heat and cold; which it asserts to be nothing but impressions in the mind.”6

 5

Nicolas Malebranche, Oeuvres complètes, ed. A. Robinet, 20 vols. (Paris: Vrin, 1959–1966), 6: 68. 6 David Hume, A Treatise of Human Nature I.4.4, ed. L. A. Selby-Bigge, 2nd ed., rev. P. H. Nidditch (Oxford: Clarendon Press, 1978), p. 226.

14

Overview of After Certainty

This way of understanding the seventeenth-century debate should not lead us to a mistakenly crude interpretation of scholastic views. Contrary to what their critics suggest, the scholastics could not have maintained the absurd view that phenomenal experiences are out in the world. Still, they did think that those experiences track the basic physical structure of nature. For they supposed that sensible qualities—hot and cold, wet and dry, but also, secondarily, colors, tastes, and odors—are basic features of the natural world. Seventeenth-century philosophy could draw on advances in science to reject this sort of view. But if the senses are not tracking the basic physical structure of nature, then they are not faithfully capturing what is in the world. Since Descartes and his successors wanted to maintain fidelity, they looked inward for the objects of perception. In making that inward turn, these authors made a different kind of mistake, since what we now know is that the senses don’t faithfully capture what goes on inside the mind either. Descartes’s commitment to transparency is as bad a mistake as the scholastic commitment to realism about the secondary qualities. What we should conclude, then, is that the senses don’t show us anything in high fidelity: they don’t show us what goes on outside the mind, nor do they show us what is happening inside the mind, which is, after all, just patterns of neurons firing. Once we give up on fidelity, we can go back to the commonsensical view that the objects of perception are external. There is no longer any reason to find the way of ideas at all compelling. Lecture Five turns toward the epistemic ideal at the level of intellect. It begins with what I call, drawing on Anselm’s De casu diaboli, chapter 12, the “Anselmian glance” (lect. 5, pp. 94–95). This is a kind of epistemic ideal: that we seek to get the entirety of an argument in our head all at once, so that we can grasp as a whole every premise and the conclusion from which it follows. The desirability of doing this is connected to a certain sort of epistemic privilege: the way we privilege the self, and the way we privilege the now. Putting these together, I speak of the “privileged me now”. There is an interesting question here that is properly one for philosophy of mind: How much can we think all at once? This was a much debated issue in scholastic philosophy. Then there is a more properly epistemic question: Does the epistemic ideal of scientia require grasping a piece of demonstrative reasoning all at once, in a single Anselmian glance? There was a fairly extensive fourteenth-century debate about this, and it was not

Robert Pasnau

15

clear whether such a synoptic grasp is possible or whether, if it is possible, it is even desirable. There were also debates here about how reasoning works: whether it is fundamentally instantaneous, or whether it should be thought of as an event that takes time to be efficacious. But setting aside these difficult issues about how the mind works, I take it for granted in this lecture that, one way or another, there is fundamentally something ideal about coming as close to the Anselmian glance as possible. From here the lecture looks ahead to Descartes, who is likewise quite concerned with achieving something as close as possible to an Anselmian grasp of a whole argument. This is especially evident in the Meditations, where he stresses the importance of being able to grasp the whole course of his argument in a single Anselmian glance. Within the privileged now, the arguments of the first three meditations go through with the highest level of human certainty, without our having to presuppose the truth of their ultimate conclusion, which is that God exists. That ultimate conclusion is required, however, in order for these evident perceptions to be preserved beyond the privileged now as stable and certain dispositions within the mind. Such scientia endures beyond the limited time of philosophical achievement when one holds the whole argument in one’s head. And to have that, one has to be able to hold something else in one’s mind, namely, the conclusion that there is a benevolent God who would not let me go astray when I reason properly. One retains this conclusion even while not philosophizing, and someone who keeps it in mind can have the highest humanly possible level of scientia in an ongoing way regarding the nature of the soul, the body, and the world around us. All this is highly relevant to the puzzle of the Cartesian circle, and the degree of certainty that Descartes thinks his method affords us. From here, the lecture goes on to look at two contrasting cases. First, I look at Locke, who rejects Descartes’s privileging of our present reasoning. He thinks that it is enough that at some point in the past we reached a conclusion on the basis of valid reasoning. Even if we have entirely forgotten what that reasoning is, the historical fact that we did at one point carry the reasoning out is sufficient to make us justified in our present confidence.7 This is a much lower standard than Descartes insists on.

 7

John Locke, An Essay Concerning Human Understanding IV.1.9 (ed. P. H. Nidditch [Oxford: Clarendon Press, 1975], pp. 528–529).

16

Overview of After Certainty

Second, I look at Hume, who also puts great weight on the privileged now, but in a way precisely contrary to Descartes’s. Indeed, Descartes’s insistence on the “privileged me now” gets strangely transmogrified in Hume, embraced not as the foundation of reason but as the only escape from rational self-destruction. The first-person present is where rational arguments have their full force, but what those arguments show, on close inspection, is that reason “entirely subverts itself.”8 Rather than arriving at an ideal scientia of ourselves and of the world around us through the Anselmian glance, Hume thinks that such intense reflection leads us to recognize the groundlessness of all reasoning. Accordingly, whereas Descartes must take measures to shore up our accumulated certainties against the ravages of time, Hume positively welcomes “carelessness and inattention” as the sole “remedy” against skepticism.9 Escaping the destructive influence of critical reflection, Hume contends that the best we can do—the only ideal to which human beings can and ought to aspire—is to follow the sensitive part of our nature. Building on the case of Hume, Lecture Six turns toward skepticism. I begin with an inventory of various medieval skeptical scenarios, focusing in particular on the familiar worry about a deceiving God. This became a standard topic of discussion in the fourteenth century. One prominent line of reply is concessive: it consists in the claim that this possibility shows that we cannot have absolute certainty about anything, but only a conditional certainty—conditional on the assumption that God is not deceiving us. This is similar to Buridan’s position on merely natural certainty. A bolder line of reply, taken by John of Mirecourt in the midfourteenth century, holds that God’s ability to deceive is subject to a surprising limitation. For even if God can infuse thoughts directly into us, causing us to think certain things, there is a sense in which it is not we who are having that thought. Instead, the thought is something like an alien invasion: it is not our thought, but God’s thought, forced upon us. Mirecourt does allow that God can deceive us in another way: God can, for instance, simply create illusions in the world around us (patterns of light, say) and thereby cause us to believe that we are seeing things that are in fact not there. But if we focus on self-evident truths, for which we do not rely upon the senses, Mirecourt thinks that not even God can cause us to have a false belief. This seems not to have persuaded many that we are completely invulnerable to divine deception, even when it comes to

 8

David Hume, A Treatise of Human Nature I.4.7 (ed. L. A. Selby-Bigge, 2nd ed., rev. P. H. Nidditch [Oxford: Clarendon Press, 1978], p. 267). 9 Hume, Treatise I.4.2 (ed. Selby-Bigge and Nidditch, p. 218).

Robert Pasnau

17

apparently self-evident truths. One common way out, then, was to turn the tables on the divine-deception worry, and argue that, in view of God’s goodness, rather than worrying about the possibility that God is deceiving us, we should trust in God as a guarantor of our ability to grasp the truth, either through some kind of direct Augustinian illumination or (more commonly among late scholastics) simply as the creator of our reliable faculties. This leads into an extended speculative excursus, in which I contemplate the possibility that not even God can guarantee our infallibility, because not even God is infallible. The basic idea here is that if we try to imagine the experience of a being that seems omnipotent, eternal, all-knowing, and all the rest, we have to allow that this experience might be an illusion. Perhaps this supposed God is a mere god-in-a-vat, and not a perfect being at all. We can stipulate, of course, that if God is God, then God cannot be deceived. But the question is how anyone, even a perfect being, is capable of assuming a position outside itself from which it can confirm that things are as they seem to be. So I conclude that “not even the most perfect cognitive being is capable of perfect certainty” (lect. 6, p. 124). If this is right, then we can see the full absurdity of the demand for perfect certainty. As a matter of logic, no being could be perfectly and unqualifiedly certain. Does this entail that our worst fears of skepticism are true? Here I think it is important to be clear about the different things that might go under the label of skepticism. On the usual meaning of the term, a skeptic is someone who denies knowledge. But reflection on the case of God shows that the skeptic must take the word “knowledge” to be categorically inapplicable, not just to us as a species, but to any species. But this is an intolerable result: it is perfectly apt and useful to be able to say, in a wide range of circumstances, that people know one thing or another. It would be the most wildly absurd of overreactions to allow the bare logical possibility of error—grounded in nothing about the human condition—to shut down such ways of talking. Indeed, the verb “to know” is so indispensable that if on skeptical grounds we were not allowed to use it, we would have to go out and find some other word to take its place. Again, however, if we focus on the epistemic ideal, we can say that these reflections on the divine case merely highlight how far our epistemic situation falls short of the ideal. We might have thought that there are some things that we can grasp with ideally perfect certainty, but in fact it

18

Overview of After Certainty

seems very plausible that we can never achieve such certainty. Even worse, there is a case to be made that, ultimately, we can never have any positive evidence one way or another for any of our beliefs. This is the dismal verdict that I label “epistemic defeat.” It holds that ultimately there are no good non-circular, non–question-begging reasons for anything. I do not argue in the book for epistemic defeatism, or even claim that it is true. I claim only that this is a historically important thesis that is not at all an absurd view. It is the kind of view that the ancient skeptics held, and that Hume would later hold. The view is not well represented in the Middle Ages, but there is a sense in which it might be thought of as something of a platitude—that is, as an obvious view. Consider Aristotle’s famous dismissal of skepticism: “Their mistake is that […] they seek a reason for that for which no reason can be given; for the starting point of demonstration is not demonstration.”10 The idea that all arguments have to stop somewhere is not far from the idea of epistemic defeat. The threat of defeatism in epistemology parallels the recurring suspicion in other philosophical domains that we are unable to produce a robustly satisfying, realistic explanation of our most ordinary assumptions. Such doubts are indeed one of the hallmarks of modern thought across all the most basic philosophical questions, ranging over morality, freedom, perception, truth, and language. Consider in particular the moral domain. The moral realist, seeking to capture the assumptions embedded in our ordinary lives, thinks that the rightness of an action has some kind of objective ground that makes it morally good, independent of contingent facts about what human beings happen to care about. For the moral antirealist, by contrast, there is no such objective ground; if we can aptly speak of moral rightness at all, it is a function only of what we in fact happen to value. Epistemic defeatism poses an analogous challenge. Those who deny defeat hold that there are ultimate, objective evidential grounds that make some beliefs more rational than others. According to the epistemic defeatist, by contrast, if one can aptly speak of beliefs as being rational at all, those beliefs must ultimately take their rationality from subjective facts about what believers happen to think. In place of objective evidential grounds, the best we can do is make dogmatic assertions of privilege. Just as the moral antirealist despairs of any argument that runs from is to ought, so the epistemic defeatist despairs of our ability to go from seems to is.

 10

Aristotle, Metaphysics IV.6, 1011a12–13, translation from The Complete Works of Aristotle: The Revised Oxford Translation, ed. Jonathan Barnes, 2 vols. (Princeton: Princeton University Press, 1984).

Robert Pasnau

19

That brings me to Hume. Hume is very critical of skepticism in its ancient form for supposing that it involves the suspension of belief. But what Hume accepts is precisely epistemic defeat. It follows from his arguments that we are left with “not the lowest degree of evidence in any proposition, either in philosophy or common life.”11 So how does Hume justify our beliefs? He does so by rejecting Locke’s principle of proportionality—or more generally, by denying evidentialism. Hume’s way forward is to embrace a kind of quietism: he simply stops striving after such ideals and focuses his attention instead on how nature in fact operates. Abandoning the hope of grounding our beliefs in reason, he thinks that we need to begin by registering the fact that we do form certain sorts of beliefs, and that we cannot do otherwise. At this point I suggest that we can do better than this sort of antitheoretical quietism. But what other options are there? First, there is faith. To hold a proposition on faith, as I use the term, is to believe it firmly, and thus to attach high credence to it, even though one does not suppose that the evidence warrants such confidence. Believing on faith, so understood, directly clashes both with evidentialism and with Lockean proportionality, which is why it strikes so many as utterly disreputable. I suggest that it is better to believe on the basis of hope. To understand how I conceive of hope here, three background ideas are needed: first, a distinction between credence (level of confidence) and belief (commitment to truth); second, the idea that credence and belief can vary independently of one another; and third, the idea that belief requires freedom from an affective attitude of fear. This last builds on the scholastic idea that distinguishes mere opinion, which involves a fear of the contrary being true, from assent, which removes this fear of the contrary. One way to reach assent is through the evidentness that yields knowledge. Another way to reach assent is through faith. The third way is through hope: without elevating one’s credence through faith, one might simply stop fearing that one is wrong. Instead of fearing and hoping in equal measure, one would place one’s hopes entirely in being right, and hence one would believe. This is very far from the sort of ideal that philosophers historically had hoped to be able to achieve. But it may be the best we can do.

 11

Hume, Treatise I.4.7 (ed. Selby-Bigge and Nidditch, pp. 267–268).

JOHN BURIDAN BEING AFTER CERTAINTY MARTIN KLEIN

After Certainty is so rich that a short response like this can address only a few aspects of it. It might seem to be mere nitpicking if I concentrate on what Pasnau has to say about just one of the many authors whose theories and opinions he arranges so skillfully in presenting us with not just a panorama of views, but also a coherent overarching story, in both systematic and historical terms. However, if there is one particular author who plays a central role in such a project, a reply like this might be justified. I will therefore focus on the one whom Pasnau rightly calls “the most important philosopher at the most important university in the world for three decades in the mid-fourteenth century” (lect. 3, p. 59): John Buridan. Buridan plays a prominent role in Pasnau’s broader narrative as a thinker who significantly transformed traditional theories. With his innovations— which are significant in themselves—Buridan cleared the path for the dramatic steps taken later. Though Buridan did not go as far as thinkers of the seventeenth century would, he did, according to Pasnau, have sufficient theoretical resources to do so. In one case, he did not go far and seems to have been inhibited by his commitment to the philosophical tradition. In another case, he seems to have wanted to go further, but was prevented by the external pressure of threats from authority. The first case has to do with Buridan’s theory of knowledge. In Lecture Two (“Evident Certainties”) Pasnau argues that Buridan was the first scholastic philosopher who distinguished clearly between different levels of evidentness. Buridan maintained that absolute certainty is not necessary for something to count as knowledge, and that conditional certainty is sufficient, a view which became commonplace in the seventeenth

22

John Buridan Being After Certainty

century.1 But, according to Pasnau, “Buridan offers just the barest anticipation of the changes that are to come” (lect. 2, p. 34). The other case has to do with Buridan’s epistemological stance on perception as described in Lecture Three (“The Sensory Domain”). Here Pasnau claims that Buridan significantly undermines basic Aristotelian distinctions which were thought to secure sensory access to the world as it really is. In doing so, Buridan could already have reached the conclusion generally credited to early modern philosophers. In one of the very few historically counterfactual passages in his monograph, Pasnau states: If external conditions had been different, the “modern” revolution of the seventeenth century might well have happened in the mid-fourteenth century. If it had, the vanguard of that movement would have been the two philosophical giants of the era, John Buridan and Nicole Oresme. (lect. 3, p. 57)

However, Buridan and Oresme found themselves under threat of condemnation, and so were prevented from making an epistemological revolution. Nevertheless, Pasnau records some significant revolts already going on in Paris at the time. In my comments, I want to consider two questions. Regarding Buridan’s theory of knowledge, I will consider whether his position is in fact that far from the significant changes which Pasnau thinks are to come later. Regarding Buridan’s theory of perception, on the other hand, my question is, whether he was in fact as rebellious as Pasnau takes him to be.

Theory of Knowledge Let me start with the role Buridan plays in Lecture Two. Here Pasnau investigates how infallible certainty as the crucial ingredient of the normative ideal of epistemology was replaced by mere probability. Buridan is central to the story, as told in this lecture. Not only does Pasnau use him to explain the concept of certainty (certitudo) and evidentness (evidentia), he also credits him with changing significantly the view of what kind of certainty can be achieved.

 1

In Lecture Six (“Deception and Hope”) Pasnau underscores this contribution of Buridan; see pp. 118–119.

Martin Klein

23

Pasnau makes clear that what distinguishes knowledge from mere belief or opinion is certainty. Things can be believed with a high degree of subjective certainty, even if what is believed to be true is actually false. In order to distinguish knowledge from subjectively certain but objectively false beliefs, the concept of evidentness is crucial. Only if one is certain about a proposition (or about what is stated by the proposition) because of its evidentness can one be said to have knowledge. Thus, what distinguishes knowledge from mere opinion is not simply that we are subjectively unable to doubt what we believe, but rather that the proposition is assured to be true. As Pasnau summarizes very nicely: Throughout the Aristotelian tradition […] certainty comes in two flavors, objective necessity and subjective confidence, which get tied together by evidentness to yield infallibility. This picture of certainty, though commonplace, omits something crucial: the possibility of a certainty that is conditional upon the information we hold, or what is now known as epistemic probability. (lect. 2, p. 38)

Buridan’s transformation of the Aristotelian theory includes this conditionalization of certainty by distinguishing three degrees of evidentness. In the highest degree of certainty, which would later be called metaphysical certainty, it is simply impossible for us to go wrong. First principles, such as the principle of non-contradiction, are absolutely evident. Principles in natural philosophy, by contrast, are only conditionally evident, that is, on the assumption that the common course of nature is not disturbed by supernatural intervention; Buridan calls this natural evidentness. Finally, concerning our moral actions, it suffices to have what would later be called moral certainty, as when a judge sentences someone after having carefully taken everything into consideration even if, unbeknownst to the judge, the accused is in fact innocent.2



2 See After Certainty, lect. 2, p. 34, and Buridan, Quaestiones in Metaphysicam II.1 (In Metaphysicen Aristotelis quaestiones argutissimae Magistri Ioannis Buridani in ultima praelectione, ed. Jodocus Badius [Paris: Prelum Ascensianum, 1518; repr. Frankfurt a.M.: Minerva, 1964], fol. 9va): “Immo est adhuc debilior evidentia quae sufficit ad bene agendum moraliter, scilicet quando visis et inquisitis omnibus circumstantiis factis quas homo cum diligentia potest inquirere, si iudicet secundum exigentiam huiusmodi circumstantiarum, illud iudicium erit evidens evidentia sufficiente ad bene agendum moraliter, etiam licet iudicium sit falsum propter invincibilem ignorantiam alicuius circumstantiae. Verbi gratia, possibile est quod praepositus bene et meritorie ageret suspendendo unum sanctum

24

John Buridan Being After Certainty

Regarding this lowest degree of evidentness, Pasnau claims that we might take Buridan to have in mind a notion of evidentness which is close to the modern notion of justification: that is, the judge’s verdict is rightly decided, given all the evidence available to him.3 However, Pasnau does not see Buridan as having arrived there quite yet, since it is unclear whether he would accept that this kind of evidentness is enough for scientia. The crucial difference between Buridan and seventeenth-century philosophers, according to Pasnau, is that the latter think moral certainty is good enough even for natural philosophy, for which Buridan would require a stronger notion of evidentness. The type of evidentness which Buridan has in mind here does not quite line up with knowledge in the modern sense of warranted belief: this evidentia is less about being certain about something given the evidence one has, but more about being certain because the proposition, as well as the cognition of it, is evident.4 I want to propose a different reading of Buridan’s account by suggesting that we can understand assent with certainty to an evident proposition in terms of the proposition being justified. Thus, my answer to Pasnau’s question whether, according to Buridan, moral certainty suffices for scientia, will be yes, whereas Pasnau thinks that it is not quite clear. If I am correct, Buridan turns out to be a proponent of epistemic probability. What makes it difficult to answer the question whether Buridan maintains that moral evidentness is enough for something to count as knowledge is that he has so many different conceptions of knowledge that, according to Pasnau, it becomes unclear where “to place the boundaries between what is and is not knowledge” (lect. 2, p. 35). Unfortunately, says Pasnau, this is a question in which Buridan did not show much interest (see lect. 2, n. 15, p. 199). In fact, Buridan distinguishes among several kinds of knowledge, the first two of which are crucial for the distinction between natural and moral evidentness. Science is not only about what is necessarily

 hominem quia per testes et alia documenta secundum iura sufficienter apparet ipsi quod ille bonus homo esset malus homicida.” 3 For interpretations of Buridan’s theory of knowledge in terms of warranted or justified belief, see Jack Zupko, “On Certitude,” in J. M. M. H. Thijssen and Jack Zupko (eds.), The Metaphysics and Natural Philosophy of John Buridan (Leiden: Brill 2001), pp. 165–182, and Dominik Perler, Zweifel und Gewissheit: Skeptische Debatten im Mittelalter (Frankfurt a.M.: Klostermann, 2006), ch. 32, esp. at p. 368. 4 See lect. 2, p. 35. Pasnau accordingly distinguishes three different senses of evidentia (p. 33).

Martin Klein

25

demonstrated (the third type of knowledge); it also concerns the first principles which ground those scientific proofs. The knowledge of these principles falls under the first and second senses of knowledge.5 According to Buridan, knowledge of first principles is not innate, nor can we be examined about them, nor are they the result of demonstrations. Instead, we arrive at first principles by forming through induction a universal proposition from singular sense experiences. Buridan is well aware that induction is neither formally valid nor justified by the additional condition that something is the case not just for the single observed instances but for all single instances of the same type.6 It is therefore a question how the procedure of induction is to be defended. Buridan claims that those principles become evident on the basis of previous sensations, memory, and experience. I once touched a fire and realized that it is hot; I also remember that I once did so. Thus, faced with another fire I form the “experiential judgment” that this fire is hot without having touched it. What I assent to is still a singular, and hence not a common principle. Here we are in the realm of knowledge in the very basic sense of contingent propositions to which we assent with certainty and evidentness. Crucially, among these propositions Buridan counts prudential judgments which require moral certainty: And a human being can sense and remember such things so many times that the intellect by a natural inclination assents with certainty and evidentness to the fact that this piece of coal is hot and hurtful, although he never touched it. And this is an indemonstrable principle of prudence that yields the conclusion that it is not good to touch it.7

Experiential judgments concern natural and moral evidentness equally. Of course, in order to arrive at a common evident principle, I need to inductively infer that all fires are hot. This relates to Buridan’s second sense of knowledge as the intellectual cognition of a necessary

 5

See Buridan, Summulae VIII.5.2 (Summulae: De demonstrationibus, ed. L. M. de Rijk [Groningen: Ingenium, 2001], pp. 117–122. Translation: Summulae de Dialectica: An Annotated Translation with a Philosophical Introduction, trans. Gyula Klima [New Haven: Yale University Press, 2001], pp. 712–717). 6 See Buridan, Summulae VIII.5.4 ad 1–3 (ed. De Rijk, pp. 12–127). 7 Buridan, Summulae VIII.5.4 ad 4 (ed. De Rijk, p. 129; trans. Klima, p. 722): “Et totiens potuit homo talia sensisse et de eis memoriam habere quod intellectus ex naturali inclinatione assentit cum certitudine et evidentia quod ille carbo est calidus et laesivus, licet numquam tetigerit. Et hoc est indemonstrabile principium prudentiale ad concludendum quod non est bonum tangere ipsum.”

26

John Buridan Being After Certainty

proposition, and among necessary propositions Buridan counts first indemonstrable principles.8 Although this is no longer a prudential judgment, it is necessarily grounded in, or mediated by, experience as much as prudential judgments are. Moreover, pretty much the same condition of evidentness needed for moral certainty seems also to be at play in universal induction. As Buridan makes clear, when the intellect makes an induction from singular experiences, it is not simply compelled by the evidentness of some common and necessary truths; rather, it needs to carefully evaluate the circumstances under which the formation of a common principle is justified. Only after the intellect has perceived “no counterinstance in any of them, nor any reason why there should be a counterinstance in another,”9 is it in a position to make an induction. Knowledge about natural principles thus seems to be gained through a process in which one transitions from opinion to knowledge by collecting enough evidence to warrant a true belief. Just as a judge legitimately arrives at a verdict after having “investigated all relevant facts and circumstances that man can diligently investigate,”10 first principles are legitimately inferred in natural science if there is nothing in the cases considered that goes against the inference.11 Scientific investigations aimed at knowledge of first principles are prone to revision when they fall short of sufficient evidentness. If this happens, the principle is no longer known, and, as Buridan admits about the relation of our beliefs to first principles, it is possible to have knowledge and opinion successively about the same thing.12 Moreover, for natural evidentness—in a way analogous

 8

See Buridan, Summulae VIII.4.3 ad 5 (ed. De Rijk, pp. 109–110.). Buridan, Summulae VI.1.4 (Summulae: De locis dialecticis, ed. Niels Jørgen Green-Pedersen [Turnhout: Brepols 2013], p. 13; trans. Klima, p. 396): “[I]ntellectus ex eius inclinatione naturali ad veritatem in nullo percipiens instantiam aut etiam rationem, quare in aliquo alio debeat esse instantia, cogitur concedere non solum quod ita est in aliis, sed universalem propositionem.” See also Summulae VI.1.5 ad 1 (ed. Green-Pedersen, p. 17). 10 Buridan, Quaestiones in Metaphysicam II.1 (ed. Paris 1518, fol. 9ra), quoted above in note 6. Translation by Gyula Klima, in Gyula Klima with Fritz Allhoff and Anand J. Vaidya (eds.), Medieval Philosophy: Essential Readings with Commentary (Malden, MA: Wiley-Blackwell, 2007), p. 146. 11 See Buridan, Summulae VI.1.4 (ed. Green-Pedersen, p. 13): “[I]ntellectus ex eius inclinatione naturali ad veritatem in nullo percipiens instantiam aut etiam rationem, quare in aliquo alio debeat esse instantia, cogitur concedere non solum quod ita est in aliis, sed universalem propositionem, quae iam est principium indemonstrabile in arte vel scientia supponendum sine demonstratione.” 12 See Buridan, Summulae VIII.4.4 ad 3 (ed. De Rijk, p. 115). 9

Martin Klein

27

to the case of moral certainty—the intellect is not simply compelled externally to judge what is objectively given as true, but there is also an important aspect of actively accepting something as true: [O]f first principles one cannot have knowledge by demonstration […] but one can have opinion about them by dialectical argumentation. And this is true of many principles that at first had been doubted, until they were made evident by the senses, memory, and experience. For these can be taken to be false, and later opined, on the basis of insufficient evidence (evidentia), and finally evidently known, when experience has sufficiently been made complete.13

One could, however, argue that what separates natural evidentness from moral certainty for Buridan is that he allows evidentness in the latter to be the case even if that to which one assents is actually false, whereas in the former evidentness requires the actual truth of the proposition. This seems to be the reason why moral evidentness is weaker than natural evidentness, as Buridan claims, because usually a proposition cannot be evidently known if it is not true. Since knowledge is necessarily knowledge of the truth, moral certainty cannot be related to knowledge. As Pasnau argues, natural evidentness requires infallibility, unlike the lowest degree of evidentness. Of course, he also notes that natural evidentness is not absolutely infallible, since God could always interfere and falsify what is usually the case in nature; that is, the infallibility of natural evidentness holds only under the presumption that the common course of nature holds (see lect. 2, n. 13, pp. 194–195). But the question is how to understand the function of this presumption itself. As Buridan makes clear, we can simply not know whether the common course of nature holds or whether God is deceiving us, which implies that what has been known might have been “falsified” by God and thus turned into nonscientia.14 What we seem to get from the presumption itself is a call for



13 Buridan, Summulae VIII.4.4 ad 3 (ed. De Rijk, p. 115; trans. Klima, p. 710): “[P]rimorum principiorum non potest esse scientia per demonstrationem, quia dictum est prius quod per ‘primum’ et ‘immediatum’ intelligimus indemonstrabile. Sed ipsorum potest esse opinio per dialecticam argumentationem. Quod est verum de multis principiis quae a principio dubitata sunt, donec facta fuerint evidentia sensu, memoria et experientia. Haec enim possunt putari falsa esse et post, per insufficientem experientiam, possunt fieri opinata, et tandem evidenter scita, cum sufficienter fuerit completa experientia.” 14 See Buridan, Summulae VIII.4.4 ad 1 (ed. De Rijk, p. 113): “[P]ossibile est supernaturaliter scientiam meam, manentem eandem, verti in non-scientiam.”

28

John Buridan Being After Certainty

justification rather than an epistemic assurance of what can count as objectively certain. As Buridan says in one passage, the intellect assents to natural principles only to the best of its knowledge by “taking them to be evident (tamquam evidens).”15 Thus, what seems to make moral and natural evidentness prone to failure is that both the natural philosopher and the judge are to some extent subject to what Buridan calls, with regard to the latter, “invincible ignorance.”16 Consequently, what Buridan seems to admit for natural evidentness, just as he does for evidentness in moral affairs, is what, interestingly, Pasnau himself credits him with in Lecture Six, namely, the view that, rather than infallibility, which is unreachable anyway, “what we in fact mostly achieve is an indubitability that reaches the truth only given certain assumptions” (lect. 6, p. 132). At the same time, Buridan carefully draws a strict boundary between knowledge—in the most basic sense of the cognition of a proposition with certainty and evidentness (i.e., the first type of knowledge)—and opinion. For instance, he distinguishes prudential judgments, which fall under this

 Ibid., p. 114: “Alia est secundum quam virtus cognoscitiva determinatur ex natura sua ad assentiendum veritati seu propositioni verae, non possibili falsificati naturaliter, licet falsificari supernaturaliter.” 15 See Buridan, Quaestiones in Analytica Priora II.20 (Quaestiones in Analytica priora, ed. Hubert Hubien, unpublished typescript, ): “Cum enim saepe tu vidisti rheubarbarum purgare choleram et de hoc memoriam habuisti, et quia in multis circumstantiis diversis considerasti, numquam tamen invenisti instantiam, tunc intellectus, non propter necessariam consequentiam, sed solum ex naturali eius inclinatione ad verum, assentit universali principio et capit ipsum tamquam evidens principium per talem inductionem ‘hoc rheubarbarum purgabat choleram, et illud’, et sic de multis aliis, quae sensata fuerunt et de quibus memoria habetur; tunc intellectus supplet istam clausulam ‘et sic de singulis’, eo quod numquam vidit instantiam, licet consideravit in multis circumstantiis, nec apparet sibi ratio nec dissimilitudo quare debeat esse instantia, et tunc concludit universale principium” (my emphasis). 16 On ignorantia invincibilis as a technical term in the context of acting in a morally good way and for its connection with moral evidentness in Buridan’s commentary on the Ethics, see Christophe Grellard, “Probabilisme et approximation du vrai au XIVe siècle,” in Jean-Philippe Genet (ed.), La verité: Vérité et crédibilité; Construire la vérité dans le système de communication de l’Occident (XIIIe–XVIIe siècle) (Paris and Rome: Éditions de la Sorbonne / École française de Rome, 2015), pp. 65–79; , at § 13.

Martin Klein

29

type of knowledge, from opinion, in that prudence, just like knowledge, cannot be false, whereas opinion can be. While he acknowledges that prudential judgments can become false, Buridan insists that this is due to some change in external circumstances, and that once this happens the cognitive state is no longer prudence but opinion.17 (Once the judge realizes that the supposed culprit is in fact innocent, his conviction is false and thus no longer prudential; once we experience that a fire is cold, we cannot know that fire is hot.) Until then, we have to live with the presupposition that things are as we think we reliably perceive them to be, both in the realm of the natural and the moral.

Sense Perception In Lecture Three, Pasnau focuses on the sensory domain, where philosophers thought they were successful in their search for certainty. Shouldn’t our senses at least guarantee us that we perceive the world as it really is, at least in some respects? Lecture Three is about precisely the question of how to define this domain and its boundaries. Pasnau concentrates on the qualitative aspect of this question: What do the senses actually show us immediately about the world which would guarantee that things are in fact as they are presented to us? The narrative of Lecture Three is the following: Aristotle distinguishes most fundamentally between proper and common sensibles. He conceives proper sensibles, such as color, as relational; that is, they are sensible insofar as they are actually sensed. This provides a response, on the one hand, to naïve realism which takes them to be absolute qualities in the world, independently of whether actually sensed or not, and, on the other hand, to Democritean subjectivism, which takes them to be constructions of the mind, because what there really is, is just atoms. Medieval Aristotelianism neglected Aristotle’s relationalism and went for realism about absolute sensible qualities. Early modern philosophers sided with Democritean subjectivism about proper sensibles and declared them to be

 17

See Buridan, Summulae VIII.7.5 dubium (ed. De Rijk, pp. 153–155), esp. at p. 155: “Ideo concederem quod prudentiam possibile est esse falsam opinionem, sed impossibile est prudentiam esse opinionem. Et sic etiam videtur mihi quod in arte vel prudentia contingit falsum dicere, sed non contingit arti vel prudentiae falsum dicere.” It should be noted, however, that matters are apparently different in Buridan’s commentary on the Ethics, where he seems to deny that a prudential judgment is scientia; see Grellard, “Probabilisme et approximation du vrai au XIVe siècle,” §§ 13–15.

30

John Buridan Being After Certainty

only secondary qualities fabricated by the mind, as opposed to what is really fundamental, namely, the primary features of bodies, which are no longer taken to be the old proper sensibles, but rather the common sensibles. The geometric-kinetic properties of motion, shape, and size, however, were put into the category not of quality but of quantity. Crucially, in epistemological questions, early modern philosophers did not share Aristotle’s “ideal of fidelity,” which would have allowed them to continue to give proper sensibles an extramental reality. Hence, Aristotle’s balanced account got neglected and “history failed to repeat itself” (lect. 3, p. 69). To tell the story of the shift from Aristotelian realism to this quantitative turn in early modern philosophy, Pasnau decides to insert an intermezzo, in which he presents several theories within the Aristotelian tradition that redefine the various kinds of sensibles to such an extent that the Aristotelian project of considering “colors, sounds, and so on as fundamental to sensation, inasmuch as these are the physical agents that directly shape the contents of perceptual experience” (lect. 3, p. 59), becomes highly suspect. Here, Nicole Oresme and John Buridan enter the stage, the former for the obvious reason that he calls into question the real existence of colors as absolute qualities, identifying them rather with reflected light.18 The latter, however, enters for reasons which I would like to examine in what follows. Pasnau takes Buridan to challenge the standard theory even more radically than Oresme, “because he calls into question the very idea of a privileged qualitative domain” (lect. 3, p. 57). According to Buridan, “the senses themselves—prior to the conceptualization of intellect—offer us no special window into reality” (lect. 3, p. 59). Pasnau argues for this conclusion as follows. First, Buridan does not take Aristotle’s proper sensibles to provide us with any more certainty about reality than the common sensibles do: the latter are no less reliable than the former. Proper sensibles are said to give a high degree of sensory certainty only in contrast to per accidens sensibles.19 Be that as it may, even if some

 18

See After Certainty, lect. 3, p. 57. See After Certainty, lect. 3, p. 57, and n. 8, p. 235. Pasnau refers to Buridan, Quaestiones in libros Aristotelis De anima secundum tertiam lecturam II.11–15 (Questions on Aristotle’s On the Soul by John Buridan, ed. and trans. Gyula Klima et al. [Cham: Springer, forthcoming]). It is not clear to me why we cannot interpret Buridan as just admitting that in general the senses operate reliably under normal 19

Martin Klein

31

sensibles are not privileged over others in terms of infallibility, a different and more pressing question is the following: What would actually be the privileged domain of the senses in which they show us real features of the world? Pasnau concludes that for Buridan there is no “privileged domain of immediate sensory access” (lect. 3, p. 58). To defend his conclusion, Pasnau argues that, according to Buridan, the senses receive all of the various sensibles in a confused way and cannot distinguish among them. Thus, if they are all indistinguishably fused together in sensory experience, there is no privileged access to any of them. Pasnau takes this from a passage where Buridan claims that the external sense cannot distinguish the confusion of species of color, size, shape, and substance of a thing.20 On a closer look, however, it becomes clear that Buridan does not claim that the senses cannot make a distinction between the different sensibles, but rather that they cannot distinguish between each sensible quality and its underlying subject. According to Buridan, the senses do not cognize whiteness, because this would be the quality of white abstracted from its subject; rather, they cognize something white, that is, a particular instance of color inhering in a subject. The same is true for sensibles like size, which is decoded, as it were, by the internal sense. For as Buridan goes on to say, the internal sense cognizes confusedly in the same way.21 However, unlike the external senses, the internal sense is capable of distinguishing among the various sensibles, as Buridan explicitly claims in several passages.22 Moreover, the kind of confusion of species that Buridan talks

 circumstances. Buridan seems simply to broaden the scope of where the senses do their job successfully. 20 See After Certainty, lect. 3, p. 58, and Buridan, Quaestiones in De anima III.8.28 (ed. Klima et al.): “Quamvis igitur sensus exterior cognoscat Socratem vel albedinem vel album, tamen hoc non est nisi per speciem confuse et simul repraesentantem cum substantia et albedine et magnitudine et situ secundum quem apparet in prospectu cognoscentis. Et ille senus non potest distinguere illam confusionem si non potest abstrahere species substantiae et albedinis et magnitudinis et situs ab invicem. Ideo non potest percipere substantiam vel albedinem vel album nisi per modum existentis in prospectu eius.” 21 See Buridan, Quaestiones in De anima III.8.29 (ed. Klima et al.). 22 See Buridan, Quaestiones in De anima II.22.10 (ed. Klima et al.): “[P]onimus convenientiam et differentiam inter sensibilia diversorum sensuum exteriorum. Iudicamus enim hoc album esse dulce vel hoc rubeum non esse dulce sed amarum. Et canis, audiens dominum suum vocantem eum, iudicat vocantem esse dominum suum et illum quem videt, et per visum vadit ad eum. Vel si videt alium et non

32

John Buridan Being After Certainty

about seems not to rule out the proper sensibles being in some way prior to the common sensibles. In fact, Buridan speaks of a privileged domain of proper sensibles (over the common sensibles), and is faithful to Aristotelianism in taking proper sensibles to have the causal power of moving the senses first and foremost, thereby providing information about, for instance, size and shape: [The] sensation [of common sensibles] does not occur because of the reception of species other than the species of proper sensibles; rather, one must first imagine that sense first perceives, among the common sensibles, the location of the sensible relative to the sense, namely whether such and such sensible objects are to the right or to the left, above or below, in the front or behind.23

How exactly this is supposed to work I do not have the space to spell out here, but it seems to me that Buridan is to some extent indebted to the perspectivist tradition. Information about location, magnitude, and shape, for instance, gets detected by the senses because the species of proper sensibles are transmitted like rays from sensible objects to the perceiver. Depending on where they are received (on, say, the right or left side of the eye, or in the left ear rather than the right) the object appears accordingly in the cognizer’s “field of perception” (prospectus cognoscentis). Moreover, such proper sensibles create this “prospect” in the first place, which allows us to sense the common sensibles.24 Pasnau’s argument from indistinct sensory content is closely linked to his second argument from the intellectualization of sensory content. Here he claims that Buridan takes the difference between per se sensibles and per accidens sensibles to be a conceptual distinction, which Buridan would warn us not to impose on how the senses operate. Thus:

 dominum suum, iudicat vocantem non esse illum quem videt; ideo non vadit ad eum, sed quaerit alibi vocantem.” See also I.6.10, II.10.16 and II.17.19. 23 Buridan, Quaestiones in De anima II.13.14 (ed. and trans. Klima et al.): “[D]icendum est de modo per quem fiunt sensationes istorum sensibilium communium. Non enim fit sensatio eorum propter recipere alias species a speciebus sensibilium propriorum, sed oportet primo imaginari quod sensus percipit primo inter sensibilia communia situm sensibilis ad sensum, scilicet an talia sint ad dextram vel ad sinistram, superius vel inferius, ante vel retro […].” 24 See Buridan, Quaestiones in De anima II.13.11–16 (ed. Klima et al.). For differences between Buridan and the perspectivist tradition, see Maria Elena Reina, Hoc hic et nunc: Buridano, Marsilio di Inghen e la conoscenza del singolare (Florence: Olschki, 2002), pp. 203–209.

Martin Klein

33

[Buridan] thinks that we illicitly imported intellectual conceptualization into our account of perception, making it seem as if the senses have the conceptual resources to pull apart distinct aspects of what is given to them. (lect. 3, p. 58)

Here again, however, I would like to propose a different reading of the passage on which Pasnau bases his interpretation. To be sure, Buridan claims that the distinction between per se sensibles and per accidens sensibles is a conceptual distinction made by the intellect. But he then defines a per se sensible as that which is cognized by sense, and a per accidens sensible as that which is not: When a stone is white, we say that the white thing is sensed per se, that is, the white thing is sensed according to the concept according to which it is said to be white, and the stone is sensed accidentally, that is, it is not sensed according to the concept according to which it is said to be a stone.25

Again, what is cognized by sense is always a quality as it belongs to a substance, not quality or substance itself; thus, they are cognized exactly as they exist in the world, namely, as qualities inhering in substances. However, this by no means implies that sense would cognize white thing only if an intellect were to make the conceptual resources available to it. As Buridan makes clear in a number of passages, animals which lack intellectual powers do make sensory judgments such as this white or this big (which beings endowed with intellect would express linguistically according to the concepts they have). What they do not judge is “this thing” or “this stone,” since this would require a substantial concept formed intellectually by means of abstraction.26



25 Buridan, Quaestiones in De anima II.12.12 (ed. and trans. Klima et al.): “Verbi gratia, cum lapis sit albus dicitur quod album sentitur per se, id est, album sentitur secundum conceptum secundum quem dicitur album, et lapis sentitur per accidens, id est non sentitur secundum illum conceptum secundum quem dicitur lapis.” 26 See Summulae VIII. 2.4, dubium 2 (ed. De Rijk, p. 48): “[S]ensus bene iudicat substantiam secundum conceptum substantialem et accidens penes conceptum accidentalem, iudicando ‘hoc album’, aut ‘hoc modicum’ vel ‘[hoc] magnum’. Unde canis, qui non habet intellectum, si dominus suus vocat eum, canis, videndo eum, iudicabit de illo quod videt quod illud est vocans eum, et ibit ad ipsum; et si non videt dominum suum, sed alium, tunc de eo quem videt iudicabit quod non est vocans eum, nec ibit ad illum sed quaeret vocantem se, scilicet dominum.” See also QDA(3) I.6.9, II.12.8, and II.22.10 (ed. Klima et al.).

34

John Buridan Being After Certainty

Thus, it seems to me that rather than denying that there is such a privileged domain, Buridan uses the conceptual distinction between per se sensibles and per accidens sensibles to carefully redefine the boundaries of this domain. To claim that a sense is concerned only with color itself—that is, without the subject in which the color naturally inheres—would give us an illusory picture of the world, for colors do not in fact exist separately from their subject. Since qualities are in reality “fused together”27 with their substance, the senses would not give us information about the world as it is if they detected only colors without their subjects. Thus, one can take Buridan to highlight exactly what Pasnau attributes to Aristotle earlier in Lecture Three: that visual experience is about color in relation to its underlying subject (see lect. 3, pp. 53–54). In the end, I have difficulties in seeing either how Buridan’s theory of perception implies the view that what causes sensation might be entirely different from what the senses show us about the world, or how Pasnau can support the claim that it is “not held hostage to any doubtful story about how the sense powers interact with the physical world. Perhaps there are no such qualities as the senses allegedly detect” (lect. 3, p. 58). This seems exaggerated, given, for instance, Buridan’s painstaking account, in by far the longest quaestio (about fifty-six pages in the edition) in his Questions on De anima (II.18), of how the proper sensibles are causally linked to sensible objects and perceptive organs, and also given that Pasnau himself, later in Lecture Three, credits Buridan with holding the view that “each of the senses shows us something in high fidelity about its particular corner of outward reality—indeed, that each sense shows us the very nature of its proper objects” (lect. 3, p. 67).

Conclusion To conclude my modest qualms about just a snippet of the full and rich picture which Pasnau gives us of (as the subtitle of his book puts it) “the history of our epistemic ideals and illusions”: Pasnau makes it crystal clear how important a role Buridan plays in the history of epistemology. Buridan’s transformations in the way of conceiving standards of certainty, as well as what the senses tell us about the world, are important milestones. What I want to question is the exact role they play. I would

 27

Peter King, “John Buridan’s Solution to the Problem of Universals,” in J. M. M. H. Thijssen and Jack Zupko (eds.), The Metaphysics and Natural Philosophy of John Buridan (Leiden: Brill 2001), pp. 1–27, at p. 14.

Martin Klein

35

like to suggest that more emphasis should be put on Buridan’s lowering of the expectations for certainty such that epistemic probability is sufficient; however, I do not think that Buridan claims that perception gives only a blurred picture of the world that awaits further disentangling by the intellect. Interestingly, Buridan never went very far in questioning the distinction between qualities and their underlying substance, let alone the existence of the latter. He remains confident that at least our intellect distinctly comes to grip with the essence of things, although this would have been a straightforward candidate for showing how unsure we actually are about the world as it is, even under the presumption that the common course of nature holds.

PERCEPTION: REDUCTIVISM AND RELATIONALISM SEBASTIAN BENDER

Robert Pasnau’s After Certainty is a fantastically rich and hugely inspiring book. It subtly engages with a great number of vastly different philosophers, it spans a period of more than two millennia, and it deals with a huge array of interpretive details in a faithful and creative manner. Yet After Certainty never loses track of its highly original and innovative narrative: that the history of our thinking about knowledge was dominated by the framework (and eventual demise) of what Pasnau calls idealized epistemology.1 His book has taught me a great deal about how history of philosophy can and should be done, and about the history of philosophy itself. Last but not least, it’s a really fun read! Needless to say, I can discuss only a small part of Pasnau’s story in these comments. In what follows, I will concentrate on two topics from Lecture Three, on “The Sensory Domain.” My focus will be the reconstruction of seventeenth-century theories of perception that this lecture provides, in particular what Pasnau calls the “reductive path” and the “relational path.”2 As it stands, none of the comments I have to offer here will amount to an attack on the bigger picture that Pasnau paints in After Certainty, or even a minor quibble. All I intend to do is to raise some small-picture questions and to suggest that certain aspects be added to the big picture.



1 See lect. 1, p. 3, where Pasnau writes: “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.” As Pasnau explains, the main goal of “idealized epistemology” is not to provide a conceptual analysis of knowledge, but “to describe the epistemic ideal that human beings might hope to achieve and […] to chart the various ways in which we commonly fall off from that ideal” (ibid.). 2 These are two of the three “paths” introduced in lect. 3, p. 61.

38

Perception: Reductivism and Relationalism

1. The Three Paths To make my points about the “reductive path” and the “relational path” let me briefly recapitulate the three paths which, according to Pasnau, decisively shaped seventeenth-century theories of perception.3 Here is Pasnau’s story (obviously, I am glossing over a great number of details): the familiar early modern mechanistic picture of the material world as nothing but particles moving around (and occasionally bumping into each other) gives rise to a problem in modelling the processes that lead to perception. This problem can be put very simply: while on the mechanistic picture our senses get the world roughly right in the case of geometric and kinetic properties, it is different in the case of the proper sensibles (colors, smells, sounds, etc.). When I see a colored object, a blue shirt for example, I seem not to perceive what is really there—namely, particles in motion— but something quite different. It is not as if (as the scholastics had it) there is a quality “blue” out there in the material world. So what is really going on? Do we constantly misrepresent the world? If we do, that would be an unfortunate result. In short, the early modern mechanists need to give an account of the relation between the proper sensibles on the one hand, and the mechanistic processes giving rise to them on the other. Pasnau argues that early modern philosophers are confronted with a choice among three different options. He labels these three paths “subjective,” “reductive,” and “relational” (see lect. 3, p. 61). The subjective strategy— which Pasnau also calls the “way of ideas” (p. 64)—accounts for the proper sensibles in subjectivist terms and denies that they are representational at all. This strategy, Pasnau maintains, is prominent in Descartes (though on Pasnau’s view Descartes vacillates between different paths). If my blue sensation does not represent anything in the world, it also cannot misrepresent—which is a result that comes in handy, for pervasive misrepresentation can then be avoided. It is, after all, rather natural to take the standard seventeenth-century story of secondary qualities to entail that we misrepresent the world around us all the time. The subjective response to this is that we don’t, because there is no misrepresentation without representation! Pasnau puts this point as follows: “There is thus no problem of misrepresentation, when it comes to the secondary qualities, because there is no representation. The privileged qualitative domain is internal” (lect. 3, p. 64).

 3

For what follows, see the section “Descartes’s Three Paths” in lect. 3, pp. 61–65.

Sebastian Bender

39

The reductive account tells a different story. It treats the mechanistic processes that lead to, say, color sensations as the objects of such perceptions (see lect. 3, p. 64). If blue sensations are caused by globes of size s rotating clockwise with speed v, then when we see a blue shirt—so the reductive account has it—what we literally perceive are globes of size s rotating clockwise with speed v. As Pasnau notes, however, this account seems to have a severe and obvious problem: “Vision simply does not reveal to us such things as the microscopic rotation of particles, and so this sort of reductive explanatory level is not a plausible candidate to serve as the object of perception” (p. 64). On the face of it, it is indeed hard to believe that we actually perceive rotating globes when we have a blue sensation. Because… Well, we don’t!—or so the argument goes. Descartes definitely concurs. In the Principles he writes (in a passage also quoted by Pasnau): “[Sensations] may be clearly perceived provided we take great care in our judgments concerning them to include no more than what is strictly contained in our perception – no more than that of which we have an inner awareness.”4 As I will argue below, however, things are more complicated with other early modern authors. The third path, the relational one, goes (according to Pasnau) all the way back to Aristotle’s account in De anima. Relationalism basically takes our senses to be infallible by definition, for it claims that “the objects of perception are defined in terms of the response they produce in perceivers such as us” (lect. 3, p. 65). That is, we do indeed perceive and represent something in the world when we have a color sensation—of a blue shirt for example—namely, whatever it is in the shirt that causes our blue sensation; in this respect, relationalism differs from subjectivism.5 Unlike the reductive account, however, relationalism denies that we represent globes rotating clockwise, or anything of the sort. Instead, color sensations are more like signs for certain geometric configurations, or for whatever it is that causes the sensation.6 Pasnau argues that, while Descartes is



4 René Descartes, Principles I.66, in Oeuvres de Descartes, ed. C. Adam and P. Tannery (Paris: Cerf, 1897; repr. Paris: Vrin, 1996), vol. 9, p. 55. Translation from The Philosophical Writings of Descartes, 3 vols., ed. and trans. J. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny (Cambridge: Cambridge University Press, 1984–1991), vol. 1, p. 216. Quoted by Pasnau in lect. 3, p. 62. 5 For a reading of Locke that is in some respects similar, see Jennifer Marušiü, “Locke’s Simple Account of Sensitive Knowledge,” Philosophical Review 125, no. 2 (2016): 205–239. 6 Pasnau characterizes the relational path as follows: “Still, vision does represent something in the world, inasmuch as a certain color sensation is the sign of

40

Perception: Reductivism and Relationalism

sometimes sympathetic to the relational path but ultimately dismisses it, Boyle and Locke are the true heroes of relationalism (although they are somewhat tragic heroes according to Pasnau, for they do not really have a place for powers in their metaphysics). I find this framework of the three paths—subjective, reductive, and relational—extremely helpful for reconstructing early modern philosophy of perception, so I am going to accept and use it in the rest of my comments.

2. The Reductive Path It is noteworthy that while the subjective path (the “way of ideas”) and the relational path play important roles in Pasnau’s narrative, the reductive path does not feature as prominently as the other two. Recall that on the reductive account we represent a specific geometric and kinetic pattern (for example, globes rotating clockwise) when we perceive a blue shirt. This seems implausible, but initial implausibility does not always deter philosophers, either in the seventeenth century or today. This holds especially true for some of the great seventeenth-century rationalists, in particular Spinoza and Leibniz. My suspicion is that the reductive strategy might play a bigger role in the seventeenth century than Pasnau suggests in Lecture Three. Not all early modern philosophers took the reductive path to be as problematic as Descartes did. But let us consider an example. Leibniz is perhaps the clearest case. To begin with, in a letter to Sophie Charlotte, Leibniz explicitly distances himself from the subjective approach: [T]he external senses allow us to know their particular objects, which are colors, sounds, odors, flavors, and tactile qualities. But they do not allow us to know what these sensible qualities are, nor in what they consist. For example, whether red is the rotation of certain small globes that supposedly make up light.7

 something in objects that is the cause of such a sensation. What the nature of that something is—or even what it is roughly like—cannot be determined by the sensation itself. Yet the senses do have a kind of external domain: they represent that there is something there, even if their deliverances at this level are opaque” (After Certainty, lect. 3, p. 65). 7 Gottfried W. Leibniz, letter to Sophie Charlotte, in Die philosophischen Schriften, ed. C. I. Gerhardt, 7 vols. (Berlin: Weidmann, 1875–1890) [= G], IV, p.

Sebastian Bender

41

Leibniz here clearly accepts some aspects of the typical early modern mechanistic picture of secondary qualities. At the same time, he acknowledges, unlike Descartes, that there is a sense in which sensible qualities are out there in the world, even though we do not know what they are. It may be, for example, that red is nothing but “the rotation of certain small globes.” It seems to follow straightforwardly from this statement that whenever I have a red sensation I have a perception of rotating small globes (because red simply is a rotation of certain small globes, albeit unbeknownst to us).8 Leibniz makes this thought explicit in his Meditations on Knowledge, Truth, and Ideas, where he writes: [W]hen we perceive colors or smells, we certainly have no perception other than that of shapes and motions, though so very numerous and so very small that our mind cannot distinctly consider each individual one in this, its present state, and thus does not notice that its perception is composed of perceptions of minute shapes and motions alone, just as when we perceive the color green in a mixture of yellow and blue powder, we sense only yellow and blue finely mixed, even though we do not notice this, but rather fashion some new thing for ourselves.9

Leibniz here says that when we have color sensations we really do perceive the microscopic geometric and kinetic properties of the object. We just do not notice that we do. Why not? His idea is that while color sensations seem to be simple and structureless, this is not really the case. In fact, Leibniz argues, my blue sensation consists of many small perceptions (petites perceptions), which are too small to be detected. This of course goes directly against Descartes’s tenet that we are conscious of everything that is in our perceptions. Leibniz famously denies this Cartesian tenet, and once this principle is given up, the reductive path becomes a lot more plausible, which seems to be precisely the route Leibniz takes.

 499. Translation from Gottfried W. Leibniz, Philosophical Essays, ed. and trans. Roger Ariew and Dan Garber (Indianapolis: Hackett, 1989) [= AG], p. 186. 8 For a detailed analysis of Leibniz’s views on colors, see Stephen Puryear, “Leibniz on the Metaphysics of Color,” Philosophy and Phenomenological Research 86, no. 2 (2013): 319–346. 9 G IV, p. 426/AG, p. 27.

42

Perception: Reductivism and Relationalism

It is actually quite natural for Leibniz to arrive at such a result. For him it belongs to the very nature of ideas to be representational items.10 Representation in turn is spelled out in terms of expression, meaning some sort of structure preservation.11 It immediately follows from this that sensory ideas (e.g., color ideas), simply in virtue of being ideas, represent and express their objects. Leibniz makes this explicit in the New Essays: [T]hese “sensory ideas” depend on detail in the shapes and motions, which they precisely express, though the mechanical processes which act on our senses are too small and too great in number for us to sort out this detail within the confusion.12

On the Leibnizian model, then, the senses do not merely reliably detect or signal certain geometric and kinetic patterns in the world.13 They also do so with what Pasnau calls fidelity. Fidelity is given, in Pasnau’s words, when the senses “tell us something about the character of what we perceive; not just that a certain quality is present to us, but what that quality is like” (After Certainty, lect. 3, p. 66). To be sure, I cannot, on the Leibnizian model, access this sort of information in my blue sensation, but this does not mean that it is not there or that it is not part of the representational content. It is worth mentioning that Leibniz seems not to be an isolated case. Spinoza seems to be committed to a similar view. Consider Spinoza’s doctrine of parallelism, according to which each mode of thought, that is, each idea, corresponds to and is expressive of a mode of extension, that is, a body.14 The order of bodies (the world conceived under the attribute of

 10

See for example his short piece, “Quid sit idea?” (G VII, pp. 263–264). Descartes, in contrast, seems at times to be forced to accept non-representational ideas. 11 For the details of Leibniz’s theory of expression and structure preservation, see Chris Swoyer, “Leibnizian Expression,” Journal of the History of Philosophy 33, no. 1 (1995): 65–99. 12 Gottfried W. Leibniz, New Essays on Human Understanding, ed. and trans. P. Remnant and J. Bennett (Cambridge: Cambridge University Press, 1996), p. 403. 13 Here is how Pasnau understands reliability: “For a sense to exhibit reliability, as I will use the term, means that a sense’s signaling the presence of a quality closely correlates with that quality’s presence” (After Certainty, lect. 3, p. 66). 14 Spinoza formulates the parallelism doctrine in Ethics, book II, proposition 7: “The order and connection of ideas is the same as the order and connection of things.” See Baruch de Spinoza, The Complete Works of Spinoza, vol. 1, ed. and trans. Edwin Curley (Princeton: Princeton University Press, 1985), p. 451.

Sebastian Bender

43

extension) thus exhibits the same structure as the order of ideas (the world conceived under the attribute of thought). If this is so, then all ideas (not just ideas of shapes etc.) not only reliably signal the presence of certain geometric-kinematic properties but also do so with a high degree of fidelity. It thus appears to be the case that Spinoza and Leibniz both adhere to the ideal of fidelity without feeling the pressure to turn inward, as Descartes did. At the very end of Lecture Three, Pasnau sketches two distinct routes which are both found in early modern philosophy: Those who go down [the path of resignation] wholly abandon the ideal of fidelity […]. But there is another way. For, as we will see next week, even while giving up on perceiving the external world in high fidelity, the mainstream of seventeenth-century philosophy did not abandon the ideal of fidelity altogether, but turned instead to another domain where that ideal could be preserved, a domain within ourselves, where objects are seen even more properly and immediately. (lect. 3, p. 69)

The reductive path, which earlier in Lecture Three was introduced as one of the three options seventeenth-century philosophers could potentially take, has all but disappeared at this point. Now, if what I have just said about Leibniz and Spinoza is right, it would seem that the reductive path may have played a much more prominent role in seventeenth-century philosophy than Pasnau suggests in Lecture Three. To be sure, this would not change Pasnau’s story in a substantial way, but it would add an important aspect to it.

3. The Relational Path Let me now turn to relationalism, a thesis which Pasnau associates (correctly I believe) with Boyle and Locke.15 Pasnau takes relationalism to be a promising option, perhaps the most promising option available to early modern philosophy of perception. I am inclined to be a bit less optimistic about the prospects of relationalism, given certain more general features of the early modern framework. Let me explain. Here is an especially clear statement of the relational approach from Locke’s Essay (quoted by Pasnau in lect. 3, n. 15, p. 248):



15 Pasnau also localizes the roots of the relational path in Aristotle; this sounds right to me too, but I cannot provide a well-founded assessment of this issue here.

44

Perception: Reductivism and Relationalism [A]ll our simple ideas are adequate. Because being nothing but the effects of certain powers in things, fitted and ordained by God to produce such sensations in us, they cannot but be correspondent and adequate to those powers. For if sugar produce in us the ideas which we call whiteness and sweetness, we are sure there is a power in sugar to produce those ideas in our minds, or else they could not have been produced by it. And so each sensation answering the power that operates on any of our senses, the idea so produced is a real idea, (and not a fiction of the mind, which has no power to produce any simple idea) and cannot but be adequate, since it ought only to answer that power.16

The central idea of relationalism comes out quite succinctly in this passage. My idea of blue (more or less) reliably tracks something— namely, a power—in the material world. (In this respect relationalism differs from the subjective path.) However, my blue idea does not represent the geometric-kinetic basis of this power with fidelity. (In this respect relationalism differs from the reductive path.) Hence, what the relationalist gives up is fidelity.17 It is important to see that for authors like Locke the sensation on the one hand and the power on the other are correlated merely by divine decree or institution. The connection depends entirely upon God’s will and is thus arbitrary. Assume that God, as a matter of fact, has made it be the case that blue sensations are associated with globes rotating clockwise and red sensations with globes rotating counterclockwise. God could just as well have done it the other way around, for nothing hinges on what is correlated with what. Neither scenario is any better than the other, and God’s choice to do it one way rather than the other is wholly arbitrary. The only important thing is that the correlation be consistent, so that we can find our way around the world. It is quite natural for Locke to proceed this way, for he (unlike reductivist authors like Leibniz) denies that there is a representational relation between the idea and the power just in virtue of the idea’s and the power’s natures. Hence the need for God to establish such a representational relation by fiat.18



16 John Locke, An Essay Concerning Human Understanding II.31.2, ed. P. H. Nidditch (Oxford: Clarendon Press, 1975), p. 375. 17 See lect. 3, n. 15, pp. 247–249. For a careful analysis of Locke’s account of sensitive knowledge, see also Marušiü, “Locke’s Simple Account of Sensitive Knowledge.” 18 Put a bit more technically: for the reductivist the representational relation is an internal relation, while for the relationalist it is an external relation. For the

Sebastian Bender

45

You need not be an unflinching metaphysical rationalist to feel a bit uneasy about such large-scale arbitrariness.19 Spinoza and Leibniz (who are often thought to be unflinching metaphysical rationalists) would certainly find Locke’s appeal to groundless divine decrees unacceptable. Leibniz, for example, would presumably argue that since no divinely decreed correlation between a particular type of sensation and a particular geometric-kinetic pattern is better than any other, God’s choice would amount to a brute fact, which violates the Principle of Sufficient Reason.20 Therefore, the relationalist path cum divine institution is not very attractive to rationalist philosophers (and they consequently do not pursue it). Even apart from such strong rationalist commitments, however, the arbitrary divine coordination of sensations and powers suggested by Locke (and others) might seem to run counter to what was one of the motivations of the mechanistic project in the first place. One ambitious (and perhaps secret) hope was to tell a story that explains why a particular geometrickinetic pattern gives rise to a particular sensation. This explanatory project is obviously abandoned once the idea of arbitrary divine institution is accepted. Moreover, the mechanistic ideal also has it that perceptual processes are accounted for exclusively in mechanistic terms. Divine institution consequently introduces an element that is foreign to the mechanistic ideal. More generally, mechanistic explanations were supposed to make the world, and our place within it, more intelligible. However, appealing to arbitrary divine decrees does not contribute to such intelligibility, but instead reduces it. These might be reasons to find relationalism—or relationalism cum arbitrary divine institution, which is a frequent combination in the early modern period—less attractive an option than it might initially seem. Note that these reasons have nothing to do with fidelity. Hence, it is perhaps not just considerations of fidelity and reliability that drive developments in early modern philosophy of perception, but other issues may be of equal importance.

 distinction between internal and external relations, see David Lewis, On the Plurality of Worlds (Oxford: Blackwell, 1986), p. 62. 19 A “metaphysical rationalist” in the sense employed here is someone who subscribes to a robust version of the Principle of Sufficient Reason. On metaphysical rationalism in this sense, see Shamik Dasgupta, “Metaphysical Rationalism,” Nous 50, no. 2 (2016): 379–418. 20 This pattern of argument can be found, for example, in G VII, p. 374/AG, p. 329.

46

Perception: Reductivism and Relationalism

Let me mention in passing another potential downside of the relational path. It seems to me that this account has a rather counterintuitive implication. In its purest form, it states that anything that gives rise to, say, blue sensations in us just is blue. This is what renders the senses infallible, for “the objects of perception are defined in terms of the response they produce in perceivers such as us” (After Certainty, lect. 3, p. 65). Now, God surely can cause blue sensations in us without taking the “detour” through the material world—that is, God can surely cause blue sensations in us at will, without clockwise-rotating globes, or anything like them, being present. Hence, God himself is blue, or so the relationalist would have to say. I take it that most seventeenth-century philosophers would consider this an unwelcome theological result.

Conclusion This concludes my remarks on the reductive and relational paths. To reiterate, I am not suggesting (either explicitly or implicitly) that there is anything wrong with the story in Lecture Three of Pasnau’s book. In fact, I am a big fan of that story! All I have tried to do here is to add a few aspects to the story and perhaps to complicate it a bit. My twofold result can be summarized as follows: first, the reductive path might feature quite prominently in rationalist authors like Spinoza and Leibniz; second, there are some potential downsides to relationalism, which might explain why many early modern philosophers were less willing than Locke and Boyle to go down the relational path.

INDIRECT REALISM SABINE VAN ENCKEVORT AND HAN THOMAS ADRIAENSSEN

In Lecture Four of After Certainty, Robert Pasnau argues that the seventeenth century was the first period in which thinkers seriously engaged with the idea that the immediate objects of perception are internal to the mind rather than external. The notion that access to external reality is mediated by a mind-world interface was rejected outright by most medieval philosophers. According to Lecture Four, even Peter Auriol in the fourteenth century, an apparent exception to this rule, was far from accepting the early modern idea that the immediate objects of awareness are internal to the perceiver. In this commentary, we first take a closer look at Auriol’s claim that the objects of perception have a special kind of “apparent being” that is distinct from the “real being” of external bodies, and highlight a dilemma for this theory that deserves more attention than it has received. Second, we turn to the claim that it was only in the seventeenth century that thinkers could seriously engage with the idea that mental items are the immediate objects of perception. From the early modern period to the present day, this indirect realism has often been cast as a path leading to uncertainty about the relationship between reality and perception. We reject this picture, and argue that Lecture Four makes clear just how wrong a picture of early modern philosophy of mind it paints. At the same time, we suggest that the diagnosis offered in Lecture Four of why indirect realism became a viable option in the seventeenth century also helps us to better understand the by now traditional association between indirect realism and skeptical worries.

Auriol’s Apparent Being In distinction 3 of his Scriptum super primum Sententiarum, Auriol discusses a list of eight illusions, and argues that, when we fall prey to

48

Indirect Realism

such illusions, the object of perception has a special kind of “intentional” or “apparent” being. To see how he arrives at this conclusion, consider the second illusion he discusses: The second experience involves the rapid circular motion of a stick in the air. For it appears that some kind of circle is made in the air by a stick that is moved in this way. So it is asked what this circle is that appears to the perceiver. It cannot be something real that exists in the stick, for the stick is straight. Even less can it be something real that exists in the air, for a colored and determinate circle cannot be in the air. Neither can it be the act of vision itself, because then the act of vision would be seen. Moreover, the act of vision is not in the air, where the circle appears to be. For the same reasons, the circle that appears cannot be anywhere within the eye. Hence it remains that it is in the air, with intentional being, or in apparent, judged, or seen being.1

When you rapidly swirl a burning stick in the air, a fiery circle appears. But what kind of thing is this circle? Auriol goes over several alternatives. He first considers whether the circle is a real thing or an unreal thing. If it is something real, then it must somehow be reducible to one of the real things present in this scenario: the air, the stick, the eye, or the organ of vision. But Auriol assumes that the circle that appears to us can be nowhere other than where it appears to be. This is why Pasnau argues in Lecture Four that “in Auriol, objective and apparent beings are out there in the world, right where they seem to be” (lect. 4, p. 83).2 Hence the circle

 1

Peter Auriol, Scriptum 3.14 (Scriptum super primum Sententiarum, ed. Eligius M. Buytaert [St. Bonaventure, NY: Franciscan Institute, 1956], pp. 696–697): “Secunda experientia est in motu subito baculi et circulari in aere. Apparet enim quidam circulus in aere fieri ex baculo sic moto. Quaeritur ergo quid sit ille circulus qui apparet videnti; aut enim est aliquid reale existens in baculo, quod esse non potest cum sit rectus; aut in aere, quod minus esse potest, nam circulus coloratus et terminatus in aere esse non potest; nec potest esse ipsa visio, quia tunc visio videretur, et iterum visio non est in aere ubi circulus ille apparet; nec alicubi intra oculum esse potest propter easdem rationes. Et ideo relinquitur quod sit in aere habens esse intentionale sive in esse apparenti iudicato et viso” (our translation). 2 For similar statements see lect. 4, n. 14, pp. 275–276 and Robert Pasnau, Theories of Cognition in the Later Middle Ages (Cambridge: Cambridge University Press, 1997), pp. 72–73. That Auriol’s apparent beings exist in the external world is also confirmed by other passages. See, for instance, Scriptum 3.14 (ed. Buytaert, p. 697) and 27.2 (Scriptum super primum Sententiarum, ed. R. Friedman, C. Schabel, and L. Nielsen [The Electronic Scriptum,

Sabine van Enckevort and Han Thomas Adriaenssen

49

cannot exist in the stick, the eye, or the sensory organ, as these are simply not where it appears to be. Indeed if the circle must be where it appears to be, it has to exist in the air. But it cannot exist there in a real way, because there is no real circle in the air. Hence, Auriol concludes, the circle must have another, less than real kind of being. Auriol calls this existence “intentional,” “apparent,” or “objective” being. In fact, Auriol generalizes this conclusion, and argues that the objects of perception have apparent being not only when we fall prey to perceptual illusions, but whenever an object appears to us. Thus, even in veridical perception, the object that appears to us has apparent being. The difference is that with illusions like the one above, the fiery circle has apparent being only, whereas the objects of veridical perception have both apparent being in virtue of being the objects of perception, and real being as mindindependent entities. But if the veridical perception of an object involves an apparent as well as a real being, Auriol faces some serious epistemological difficulties. For if one supposes that the apparent thing exists alongside real things in the world, it seems natural to suppose that the apparent thing functions as an intermediary through which we perceive the world. This, at least, is how Auriol was read by contemporaries such as William of Ockham, Walter Chatton, and Adam Wodeham.3 But mediated accounts of perception are often taken to give rise to “veil of perception” problems: we do not see the real world, but some kind of interface instead, and because we are unable to look behind the interface to verify that it shows the world as it really is, the epistemic status of our beliefs about the world becomes shaky. As Wodeham put it in his discussion of Auriol on apparent being: “There will as it were be an intermediary veil here” that puts in jeopardy the possibility of certain knowledge about the world.4

 http://www.peterauriol.net], ll. 592–594; references to the Electronic Scriptum [= ES] cite line numbers in the edition published online). 3 Katherine Tachau opts for a similar interpretation and argues that apparent beings “have no extramental existence,” but rather are “mental similitudes.” See Katherine Tachau, Vision and Certitude in the Age of Ockham: Optics, Epistemology and the Foundations of Semantics, 1250–1345 (Leiden: Brill, 1988), pp. 89 and 103. 4 Adam Wodeham, Lectura secunda prol. 4.5 (Lectura secunda in librum primum Sententiarum, ed. Rega Wood [St. Bonaventure, NY: St. Bonaventure University, 1990], p. 92): “Et ita erit ibi quasi pallium et medium.” See also Walter Chatton, Reportatio et lectura, prol. 2.2 (Reportatio et lectura super Sententias: Collatio ad

50

Indirect Realism

However, this criticism in Wodeham and others ignores one very important passage in Auriol, in which he explicitly addresses the veil problem: The thing that is seen, along with the fact that it really exists, also has a judged and seen existence. This does not impose on that reality any variety or distinction or number with respect to anything absolute; rather, it adds an intrinsic and indistinguishable relation that is called an “objective appearance.” Accordingly, sight is not terminated at the thing that is its object through the mediation of anything absolute, as if there were some cloak or intermediary between the vision and the wall that is seen.5

This rather complex passage suggests that, in veridical perception, there is some kind of identity between the apparent and real: there is no “distinction” between them. And because there is an identity between the real and the apparent, there is nothing that stands between us and the world that we perceive. But how should we interpret the suggestion here that, in veridical perception, the apparent object and the real object of cognition coincide? On one reading of the above passage, Auriol thinks that the real object and the apparent object of perception are numerically the same. Indeed, he seems to point in this direction when he claims that the apparent being does not impose “any variety or distinction or number” on the real being. Dominik Perler seems to read Auriol along these lines when he argues that the apparent being is a mere “function” of the real object. According to Perler, the apparent being is not “something attached or supplemented to the thing” and therefore it does not have “an existence of its own.”6 On

 librum primum et prologus, ed. Joseph C. Wey [Toronto: Pontifical Institute of Mediaeval Studies, 1989], pp. 86–94), and William Ockham, Ordinatio 27.3 (Scriptum in librum primum Sententiarum: Ordinatio, ed. G. I. Etzorn and F. I. Kelley, Opera Theologica [= OT] IV [St. Bonaventure, NY: Franciscan Institute, 1979], pp. 230–258). 5 Peter Auriol, Scriptum 27.2 (ES, ll. 594–599) “[I]dcirco res quae videtur, cum hoc quod realiter existit, habet etiam esse iudicatum et visum, quod quidem non ponit varietatem aliquam aut distinctionem, vel numerum cum realitate illa quantum ad aliquid absolutum, sed addit respectum illum intrinsecum et indistinguibilem, qui dicitur ‘apparitio obiectiva.’ Non igitur terminatur visus ad rem obiectam mediante aliquo absoluto, quasi sit aliquod pallium vel medium inter visionem et parietem qui videtur.” 6 Dominik Perler, “What Am I Thinking About? John Duns Scotus and Peter Aureol on Intentional Objects,” Vivarium 33, no. 1 (1994): 72–89 at pp. 84, 85, and 84 respectively. Cf. Lukáš Liþka, “Perception and Objective Being: Peter

Sabine van Enckevort and Han Thomas Adriaenssen

51

this reading, the apparent thing just is the real thing insofar as it relates to the perceiving subject. It is not some kind of double that mediates between the mind and the external object of veridical perception. One problem with this reading, however, is that on Auriol’s own avowal, a real being and its apparent counterpart can have different histories. The real tree predates the apparent tree that comes into being when I become aware of the tree. And if God wills to destroy the tree while leaving the apparent tree intact, the apparent tree will postdate the real one. But if real beings and their apparent counterparts can have different histories in this way, surely they cannot be numerically identical. As Ockham puts it, “When two things are really the same, their generation and corruption coincide.”7 This is not the case for a real tree and an apparent tree, so no numerical identity obtains between real and apparent trees.8 Another problem with this reading is that it undermines Auriol’s account of illusions. According to Perler, the apparent being is a function of a real thing. But consider again the case in which God destroys the tree I am looking at while leaving the apparent tree intact. In this scenario, there is no real thing the apparent tree could be a function of. So it seems that in this case, what appears to me is an object of an altogether different kind than the object that appears to me in a case of veridical perception. In the latter case, the object that appears to me is a function or aspect of something real; in the former case, it is not. But Auriol rejects this kind of disjunctivism. In building his theory of perception, he starts off with a detailed account of perceptual illusion precisely because he thinks cases of perceptual error and illusion can teach us something about the object of perception in general. The assumption here is that in sensory error and veridical perception alike, we are dealing with the same kind of cognitive object. On a second reading of the identity between apparent and real beings, Auriol has in mind something like the identity that obtains between the several instantiations of a universal. When two objects instantiate the same universal, they are identical insofar as they share the same nature, but they differ numerically as distinct instantiations of that nature. In the same way,

 Auriol on Perceptual Acts and Their Objects,” American Catholic Philosophical Quarterly 90 (2016): 49–76. 7 Ockham, Ordinatio 27.3 (OT IV, p. 239): “Quando aliqua sunt idem realiter, illa sunt simul generatione et corruptione.” 8 See also lect. 4, n. 14, p. 275.

52

Indirect Realism

on this reading, the real whiteness and the apparent whiteness that are involved in the veridical perception of a white object are two instantiations of a universal whiteness, and therefore are identical insofar as they share the same nature, but differ numerically as distinct instantiations of that nature. On this reading, Auriol could say that the veridical perception of an object involves two beings, and that the real and apparent being involved in the process are identical in some robust sense. It seems then that he could have his cake and eat it too. Unfortunately, this reading seems to sit uneasily with a fundamental premise in Auriol. According to Auriol, everything that exists is singular in and of itself.9 In his own words: I ask what the principle of individuation is. […] I introduce three propositions. The first is that, if the question is what individuality adds to a specific nature, the question comes to nothing, because everything is singular insofar as it is.10

This means that there is no such thing as a universal whiteness. And this excludes the type of identity proposed above as a way of explaining the relation between the real and apparent whiteness involved in the veridical perception of a white object. In Lecture Four, Pasnau points out that for Nicholas of Autrecourt, the objects of perception are in fact universals capable of multiple instantiations. And at one point he suggests that Auriol might rely on a similar strategy here. It seems that Auriol did conceive of the real and apparent beings involved in an act of veridical perception as multiple instantiations of a single nature, thus treating the apparent being as “a special case not subject to the usual considerations of parsimony” (lect. 4, p. 86). This suggestion, however, seems to contradict Pasnau’s claim on

 9

On Auriol’s conceptualism, see Russell Friedman, “Peter Auriol on Intentions and Essential Predication,” in Sten Ebbesen and Russell Friedman (eds.), Medieval Analyses in Language and Cognition (Copenhagen: Royal Danish Academy of Sciences and Letters, 1999), pp. 415–430. 10 Peter Auriol, II Sent. 9.3.3 (Commentaria in primum–quartum librum Sententiarum [Rome: Aloysius Zannetti, 1596–1605], vol. 2, p. 114): “Quaero ergo quid sit principium individuationis. Et sine argumentis respondeo. Pono hic tres propositiones. Prima est quod realiter loquendo quaestio nulla est, cum quaeritur quid addit individuum ad rationem speciei, quoniam omnis res eo quod est, singulariter est.”

Sabine van Enckevort and Han Thomas Adriaenssen

53

the next page that objective beings for Auriol are “numerically the same” as their real counterparts (lect. 4, p. 87). So where does this leave Auriol? On the one hand, he seems to believe that in order to avoid a cloak or veil of perception, the real and apparent objects of veridical perception must be identical. But on the other hand, it seems he has no consistent way to account for such identity. He cannot appeal to numerical identity, as this would fly in the face of his conjunctivism, and make him vulnerable to Ockham’s criticism that no two objects can be identical and yet have different histories. But he cannot appeal to specific identity either, as this would go against his own nominalism. In the absence of a consistent account of the identity of real and apparent being, then, it seems that by Auriol’s own standards the veil problem still stands, leaving him in exactly the same position he set out to refute.

The Inward Turn For all of its problems, the idea that the objects of awareness have apparent or objective being was here to stay, and as Lecture Four points out, even early modern philosophers such as Descartes and Arnauld could not resist appealing to it. But if this makes them heirs to the scholastic tradition, there is a striking difference between their account of objective being and that of medieval thinkers like Auriol. For Auriol, objective or apparent beings are mind-dependent, but exist in the external world, at least if that is where they appear to be. Not so for early moderns like Descartes: they “turn inward and treat objective beings as creatures of the mind” (lect. 4, p. 83). But why? This question brings us to one of the core claims of Lecture Four—that the seventeenth century was “the first time in the history of philosophy when general doubts arose over whether there are suitable external objects of perception” (lect. 4, p. 92). Objects as they appear to us in perception are endowed with color, smells, and other sensory qualities. But according to thinkers like Descartes, external objects lack color and smell as perceived by us. Hence objects as they appear to us in perception are internal to the mind, where Descartes claims they have objective being.11 Objects as they

 11

See, for example, AT VII, p. 102/CSM II, p. 75 and AT VII, p. 161/CSM II, p. 114. We use the standard abbreviation AT to refer to the Oeuvres de Descartes, ed. C. Adam and P. Tannery, 11 vols. (Paris: Vrin, 1974–1983). CSM refers to the

54

Indirect Realism

appear to us in perception may have objective being for Descartes, as they did for Auriol. But once the external world has been stripped of colors and smells as we perceive them, the perception of this objective being demands that the mind “[turn] to itself, and [inspect] one of the ideas which are in it.”12 It has often been said that this inward turn brought with it an epistemic turn as well. If the immediate objects of perception are internal to the subject of perception, then it seems that we need to justify the trust we place in our cognitive abilities. Here is Locke: How shall the Mind, when it perceives nothing but its own Ideas know that they agree with the Things themselves?13

Locke thought he could answer this question, but others were not so confident. According to Thomas Reid in the eighteenth century, the position that the immediate object of perception is but an idea or image in the mind teaches men “to doubt even of those things that had been taken for first principles,” such as the reality of the things we perceive, and the general soundness of our cognitive make-up.14 Or as Hilary Putnam put the point some two centuries later: Our difficulty in seeing how our minds can be in genuine contact with the external world is, in large part, the product of a disastrous idea that has haunted Western philosophy since the seventeenth century, the idea that perception involves an interface between the mind and the “external” objects we perceive.15

Now, the grim picture that thinkers from Reid to Putnam have sketched of (early modern) indirect realism can be criticized on a number of counts. First of all, it is a mistake to think that skeptical questions about the reliability of representations are the result of any specific account of how

 Philosophical Writings of Descartes, ed. and trans. J. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny, 3 vols. (Cambridge: Cambridge University Press, 1984– 1991). 12 AT VII, p. 72/CSM II, p. 51. 13 John Locke, Essay 4.1.3 (An Essay Concerning Human Understanding, ed. P. H. Nidditch [Oxford: Clarendon Press, 1975], p. 563). 14 Thomas Reid, Essays on the Intellectual Powers of Man 2.10 (The Works of Thomas Reid, ed. W. Hamilton [Bristol: Thoemmes Press, 1994], pp. 285–286). 15 Hilary Putnam, The Threefold Cord: Mind, Body, and World (New York: Columbia University Press, 1999), p. 488.

Sabine van Enckevort and Han Thomas Adriaenssen

55

representation works. After all, every plausible theory of representation will have to allow for a gap between appearance and reality: the world is not always the way we represent it, regardless of precisely how representation works. But as soon as there is a distinction between appearance and reality, we can begin to ask of any given representation whether it represents its object as it really is. Now, to tell whether a given representation represents its object correctly, it seems we need access to the object independently of our representation of it. And such access seems hard to come by, whether one thinks of representations as pictures in the mind, as an indirect realist might, or as mere acts of apprehension, as a direct realist would. From Reid to Putnam, critics of the early modern theory of ideas have been wrong to see these theories as the source of epistemic troubles that could be avoided if only a better theory of perception were in place.16 Lecture Four suggests an even more fundamental way in which Reid and others got the early modern theory of ideas wrong. What Pasnau makes clear is that the inward turn of thinkers such as Descartes and Locke was the result of a growing awareness that the world is not the way we are naturally inclined to believe it is. It is not because philosophers like Descartes internalized the objects of perception that they subsequently began to doubt whether the world is the way we believe it is. Rather, it is the other way around: it is because they realized that colors as we perceive them are not features of external objects that they internalized the objects we perceive. So what was it that earned early modern indirect realism its reputation as a source of epistemic troubles? It might be argued that the notion of an interface between mind and world serves to make the gap between the world as it is and the world as it appears more vivid. But just by making the gap more vivid than it may be on other accounts, the theory of ideas hardly generated a problem that was not already there. A more promising answer was suggested some time ago by Monte Cook. According to Cook, early modern thinkers like Malebranche often spoke as if ideas were not just the immediate objects of perception, but the only ones. On such a view, all our claims about external bodies would be mere guesses about

 16

See also John Greco, Putting Skeptics in Their Place: The Nature of Skeptical Arguments and Their Role in Philosophical Inquiry (Cambridge: Cambridge University Press, 2000), ch. 4.

56

Indirect Realism

objects we have never perceived anyways. And surely such a view puts us in a worse position as knowers than its direct-realist competitors do.17 But as Cook himself points out, most early modern proponents of ideas were not committed to anything so extreme. Ideas present us with, among other things, the figures and shapes of bodies. But the figures and shapes that exist in the mind when we think of three-dimensional objects are also genuine features of external bodies. To see figures and shapes in our ideas is thus to be presented with genuine features of material substances. It may be the case that this access to external bodies is mediated, but to say that our access to material substances is mediated does not entail a commitment to a kind of idealism according to which all we ever perceive are ideas.18 So the initial question remains: What was it that earned the early modern inward turn its reputation as a source of epistemic troubles? Pasnau does not address this question in these terms in Lecture Four, but we believe that he does provide at least the outlines of an explanation for the perceived link between the way of ideas on the one hand, and skeptical worries about the reliability of our representations on the other. This link does not lie, as philosophers from Reid to Putnam have assumed, in the fact that indirect realism somehow generates skeptical questions that direct realism does not. Rather, it lies in the fact that indirect realism and skepticism about the appearance of external objects share a common root in the distinction between primary and secondary qualities. The first lesson that Descartes drew from the distinction between the two kinds of qualities was that the world often is not as we perceive it to be, and that we need a criterion to discern between features that can be ascribed to external bodies and those that cannot. A second lesson he drew from the distinction between primary and secondary qualities was that the world of colored bodies that we perceive is internal to the mind, not external to it. The inward turn to ideas and the need for a criterion to discern between true and false representations, then, are indeed related, but they are related as siblings, not as parent and offspring.

 17

Monte Cook, “Malebranche versus Arnauld,” Journal of the History of Philosophy 29 (1991): 183–199. 18 See also Pasnau, Theories of Cognition in the Later Middle Ages, pp. 209–213.



CERTAINTY: LAST CHAMPIONS AND BEYOND STEPHAN SCHMID

With his monumental Metaphysical Themes 1274–1671, Robert Pasnau surprised the philosophical community with an outstanding, philosophically highly informed history of metaphysics from the late Middle Ages up to the early modern period.1 This book, critics agreed, will “be a standard fixture on reading lists for many years to come and […] spark many a journal article.”2 Only six years later, Pasnau has raised the ante with After Certainty: A History of Our Epistemic Ideals and Illusions. His new book even goes beyond the aspirations of Metaphysical Themes in at least two respects: in its historical breadth and in its philosophical ambition. Based on his Isaiah Berlin Lectures on the History of Ideas, delivered at Oxford in the spring of 2014, After Certainty provides a breathtaking history of the development of modern epistemology as the search after the conditions under which a subject can be rightly described as knowing something. While his Metaphysical Themes already covered about four hundred years, After Certainty discusses more than two thousand years of the history of philosophy, from Aristotle to Hume. And while Metaphysical Themes already provided a highly philosophically informed history of philosophy, After Certainty pursues a distinct agenda. Apart from developing an extensive history of epistemology, Pasnau makes a powerful case for an “idealized epistemology,” that is, a conception of epistemology as an ideal theory which is supposed to describe the best kind of knowledge possible for us, on the basis of which we can reflect on the proper means to achieve it. (In this sense it is similar to theories of justice that study ideal scenarios of justice in order to make progress in improving justice in our utterly nonideal societies.) As Pasnau argues—and this constitutes the main line of

 1

Robert Pasnau, Metaphysical Themes 1274–1671 (Oxford: Oxford University Press, 2011). 2 Sydney Penner, review of Metaphysical Themes 1274–1671, by Robert Pasnau, The Philosophical Review 123, no. 1 (2014): 107–112, at p. 112.

58

Certainty: Last Champions and Beyond

the historical argument of the book—the conception of epistemology as an ideal theory is the original conception of it, which was abandoned only in the early modern period. In fact, Pasnau argues, virtually all authors from Aristotle to Descartes identified this ideal type of knowledge, which should set the goal of our epistemic activities, with certainty. According to these authors, we are after certainty when we seek knowledge—thus one facet of the book’s title. However, by defending epistemology as an ideal theory, Pasnau does not want simply to return to an ancient conception of epistemology; in light of his history, such a return would be naive indeed. For one important philosophical lesson that Pasnau gleans from this history—but also from an intriguing reflection on whether God is able to know about his or her (alleged) infallibility—is that certainty cannot be the ideal of knowledge. Thus, what Pasnau argues for on the philosophical side of his book is a form of idealized epistemology after certainty, that is, after having seen that certainty cannot play the role of the ideal of knowledge—thus another facet of the book’s title. According to his vision, the ideal to which we should gear our epistemic activities is the optimistic search after truth. But since Pasnau agrees with Hume that we can have no rational warrant that our search is ever successful, he recommends pursuing this ideal hopefully. As he puts it, we should “make the grown-up acknowledgment that our present, irresistible attitudes are not ones that can be sustained in the cold light of reason. Yet we still move ahead with life, hopefully, even while remembering the disappointing verdict of philosophical reflection on our epistemic predicament” (lect. 6, p. 138). As should already be clear from this broad outline, After Certainty is an immensely rich, profound, and thought-provoking contribution to both of the areas of philosophical research that it engages with—the history of philosophy on the one hand and present-day epistemology on the other. It should thus be equally clear that a brief commentary like this can address only a tiny fraction of Pasnau’s gigantic project. In order to do justice to Pasnau’s historical and philosophical areas of engagement, I want to focus on one claim each with respect to these two areas. The historical claim that I will discuss is Pasnau’s contention, put forward in Lecture Two, that Galileo and Descartes were the last champions of the ideal of certainty. The philosophical claim that I will address in my second set of remarks is the claim, put forward in Lecture Six, that once we have seen that certainty is out of reach we should accept what he calls “epistemic defeatism,” but instead of falling into skeptical despair we should assume an affective attitude of hope.

Stephan Schmid

59

The Last Champions of Certainty? As Pasnau aptly observes, modern epistemology is mainly concerned with the question of what we may firmly believe (lect. 2, pp. 43–44). According to his fascinating history, for this modern conception of epistemology to arise it was crucial that philosophers reject certainty as a normative ideal of epistemology. This is because certainty is so hard to achieve (if it is to be achieved at all) that it is disqualified as a plausible normative standard for firm beliefs. After all, our capacity to act, which is essential for us as living beings, depends on our having the capacity for holding firm beliefs – and it would be philosophically awkward to maintain that we are not entitled to do what we are forced to do. In Pasnau’s story of the gradual abandonment of the ideal of certainty, early modern philosophy plays a decisive role. As he explains, it was in this period that the ideal was entirely abandoned. In fact, Pasnau identifies Descartes and Galileo as the last defenders of the ideal of certainty: [W]ith 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. (lect. 2, p. 30)

I am inclined to disagree with Pasnau that the long tradition he describes came to an abrupt end with Descartes (1596–1650) as its “last great champion.” My reason for doing so consists in a counterexample, given by an early modern philosopher who is surprisingly absent in Pasnau’s history: Benedict de Spinoza (1632–1677). In his Ethics, Spinoza famously distinguishes three kinds of knowledge (cognitio).3 He refers to knowledge of the first kind as opinion or imagination (opinio vel imaginatio), knowledge of the second kind, which is concerned with general features of things, as reason (ratio), and knowledge of the third kind, which is concerned with the essences of singular things, as intuitive knowledge (scientia intuitiva). What is more, Spinoza identifies intuitive knowledge with the ideal that we should strive for, arguing that “blessedness is

 3

Spinoza, Ethics, Demonstrated in Geometric Order, 2p40s2, in The Collected Works of Spinoza, ed. and trans. Edwin Curley, vol. 1 (Princeton: Princeton University Press, 1985), pp. 477–488.

60

Certainty: Last Champions and Beyond

nothing but that satisfaction of mind that stems from the intuitive knowledge of God.”4 Also—and this is particularly important for my present concerns—Spinoza takes knowledge of the second and third kinds to be certain. In fact, these kinds of knowledge are certain according to both of the sides of certainty that Pasnau distinguishes: they are objectively as well as subjectively certain (see lect. 2, p. 28). On the objective side, knowledge of the second and third kinds is certain for Spinoza because “to knowledge of the second and third kinds pertain those [ideas] which are adequate; and so […] this knowledge is necessarily true.”5 On the subjective side, these ideas are certain because true ideas for Spinoza are “self-validating” in the sense that “[h]e who has a true idea at the same time knows that he has a true idea, and cannot doubt the truth of the thing.”6 It is of course an interesting and difficult question why Spinoza took himself to be entitled to the startling view that just by having a true idea one knows this idea to be true such that “truth is the standard both of itself and of the false,” as Spinoza puts it some lines later.7 But for the purposes of my remark, we can set this question aside. What is relevant for assessing Pasnau’s claim about Descartes and Galileo being the last champions of certainty is that Spinoza not only thought that true ideas are self-validating, but also held that we are in fact capable of having such self-validating (true) ideas. As Spinoza makes clear in a famous letter, he even thought that his philosophy expresses such a self-validating true idea.8 All this makes it plain that Descartes and Galileo were by no means the last defenders of the epistemological ideal of certainty. Spinoza was an adamant defender of this ideal as well. So what’s the big deal? Quibbles

 4

Spinoza, Ethics, 4app4, ed. Curley, p. 588. Spinoza, Ethics, 2p41d, ed. Curley, p. 478. 6 Spinoza, Ethics, 2p43, ed. Curley, p. 479. 7 Spinoza Ethics, 2p43s, ed. Curley, p. 479. For an intriguing attempt to make philosophical sense of Spinoza’s claim, see Michael Della Rocca, “Spinoza and the Metaphysics of Scepticism,” Mind 116 (2007): 851–874. 8 See Spinoza, letter 76, in The Collected Works of Spinoza, ed. and trans. Edwin Curley, vol. 2 (Princeton: Princeton University Press, 2016), p. 475: “I do not presume that I have discovered the best Philosophy; but I know that I understand the true one. Moreover, if you ask how I know this, I will reply: in the same way you know that the three Angles of a Triangle are equal to two right angles. No one will deny that this is enough […]. For the true is the indicator both of itself and of the false.” 5

Stephan Schmid

61

about who was the first or last adherent of a particular idea—such as the idea of certainty as an ideal of knowledge—are usually pretty fruitless and almost always completely uninteresting from a philosophical point of view. In this case, however, Pasnau’s claim that the ideal of certainty was abandoned after Descartes strikes me as telling for the historical narrative he provides. For Spinoza was by no means the last champion of the ideal of certainty either. Leibniz, for instance, argued that we can attain metaphysical certainty about contingent facts concerning our universe, provided that one of us “discovers the a priori origin of the world we see and pursues the question as to why things are the way they appear back to the ground of essence.” Though he conceded that “this would nearly approach the beatific vision and that it is difficult to aspire to this in our present state,” he didn’t take it to be impossible and cherished certainty as an epistemological ideal.9 Nor was this ideal abandoned after Leibniz. Kant was very much concerned with certainty as well, and so were his Idealist successors:10 Fichte and Reinhold, for instance, sought to ground their philosophical systems in an absolutely certain principle,11 while Hegel argued that the highest form of knowledge—i.e., “absolute knowledge”—consists in the coincidence of truth and certainty.12 It is interesting to note therefore that Pasnau’s grand history of epistemology omits an important strand of modern philosophy which continued to adhere to the epistemological ideal of certainty and was decisively inspired by Spinoza.13 This is not to say that are no good

 9

Gottfried Wilhelm Leibniz, letter to Foucher (1675), in Philosophical Essays, ed. and trans. Roger Ariew and Daniel Garber (Indianapolis: Hackett, 1989), p. 4. 10 For Kant, see his famous remark about it being “a scandal of philosophy and universal human reason that the existence of things outside us” cannot be proven (Critique of Pure Reason, B xxxix, ed. and trans. Paul Guyer and Allan Wood [Cambridge: Cambridge University Press, 1998], p. 121). 11 See Johann Gottlieb Fichte, “Rezension Aenesidemus,” in Gesamtausgabe, ed. Bayrische Akademie der Wissenschaften, vol. 1.2 (Stuttgart: Friedrich Frommann, 1965), pp. 33–67. 12 Georg Wilhelm Friedrich Hegel, Wissenschaft der Logik I, Werke V (Frankfurt am Main: Suhrkamp, 1986), p. 43. 13 For revealing discussions of Spinoza’s influence on the development of German Idealism, see Paul Franks, All or Nothing: Systematicity, Transcendental Arguments, and Skepticism in German Idealism (Cambridge, MA: Harvard University Press, 2005) and Eckart Förster and Yitzhak Y. Melamed (eds.), Spinoza and German Idealism (Cambridge: Cambridge University Press, 2012).

62

Certainty: Last Champions and Beyond

reasons for this lacuna in Pasnau’s history or for ending his study of the development of our modern conception of epistemology with the great British Empiricist David Hume. First of all, it was Hume much more than any other historical philosopher who inspired twentieth-century analytic philosophy, which in turn formed our contemporary conception of epistemology. In addition to this, the analytic tradition was to a great extent directed against the perhaps over-ambitious program of German idealism and its followers in Britain, whose philosophical project was often considered a philosophical dead end. For all that, it is important to note that there was an entire tradition of modern philosophy whose members were eager to defend the idea that true knowledge is certain— just like Aristotle, the Stoics, and Ptolemy, and Galileo and Descartes before them. Against this backdrop, Pasnau’s claim that Galileo and Descartes were the last champions of the epistemic ideal of certainty reveals a blind spot in his history. It omits the rationalist tradition that eventually inspired the development of German idealism.

Skepticism, Epistemic Defeatism, and Hope As mentioned above, in After Certainty Robert Pasnau also pursues a purely philosophical agenda. In Lecture Six he makes the case for a return to the original conception of epistemology as an ideal theory—after it has been recognized that certainty is untenable and therefore not a suitable ideal for knowledge. But is there any room for genuine knowledge, once we have seen that certainty is not to be had? As Pasnau rightly points out, after having realized that certainty lies beyond our reach, and perhaps even beyond God’s reach, our epistemic situation appears to be rather dismal. We might even be tempted to embrace skepticism, by acknowledging that there is nothing that we can really know. In Pasnau’s view, however, such a reaction would be premature. As he insinuates at one point, the denial of the possibility of our having knowledge due to our inability to know something for certain would be “puerile” (lect. 6, p. 125). This is because denying our capacity for knowledge on account of the impossibility of certain knowledge depends on a concept of knowledge that is categorically inapplicable (or at least this is what Pasnau thinks he has shown). But such a denial amounts to no more than a cheap triviality in Pasnau’s view, and as such it can hardly be problematic.

Stephan Schmid

63

Instead of skepticism, then, the right lesson to draw from our dismal epistemic situation is that we should adopt a combination of the position Pasnau calls “epistemic defeatism” and a practical attitude of hope. Let me briefly recapitulate these two elements of Pasnau’s lesson before raising some worries about it. In Pasnau’s view, the position of epistemic defeatism “has been so neglected as not even to have a name” (lect. 6, p. 128). Pasnau thus chooses the label for himself, and characterizes the position as follows: Epistemic defeatism holds that “ultimately” there is no good evidence in the sense that, for any subset of evidence that supports a proposition, additional evidence is always available that, in the end, wholly defeats the original evidence, entirely canceling its evidential weight.” (lect. 6, p. 128)

Given that certainty is not achievable, we should give up on the prospect of finding any epistemic reasons that are so strong that they could not be undermined by any other evidence. For such irrefutable reasons would provide us with precisely the kind of certainty which Pasnau has forcefully argued is unattainable. As already mentioned, Pasnau does not want epistemic defeatism to lead to skepticism. Nor should we try to refute skepticism by taking refuge in forms of idealism or naturalism. Rather, Pasnau recommends combining our acceptance of epistemic defeatism with hope: [We should] assume an attitude of hope with respect to our epistemic prospects, self-consciously granting just how nonideal our cognitive situation is, and yet expressing hope, all the same, that our beliefs hit their mark. (lect. 6, p. 134)

In light of epistemic defeatism, this attitude of hope cannot be warranted by epistemic reasons (since there are no epistemic warrants in the first place according to epistemic defeatism). In acknowledging this, Pasnau concedes frankly that “the constraints [sc. of this epistemic hope] will ultimately be grounded in something other than objective evidence” (lect. 6, p. 137). In fact, he goes so far as to suggest that such an attitude is not based on epistemic reasons at all: Hope, as I am thinking of it, requires no such cognitive confidence. It requires merely an optimism on the affective side—a cheerful willingness not to worry about the all-too-possible bad scenarios. (lect. 6, p. 138)

64

Certainty: Last Champions and Beyond

Given our dismal epistemic situation, after certainty has been recognized to be out of reach, there are not many epistemic ideals left. Pasnau, for one, promotes an ideal knowledge-seeker, whom he describes as a hopeful epistemic defeatist. This person, Pasnau explains, “exploits the affective aspect of belief formation, by letting her sentiments lead her to belief, even while her credences reflect the depressing limits on our overall evidence” (lect. 6, p. 138). So much for presentation. In the remainder of this section I will make some critical remarks on each of the three mentioned elements of Pasnau’s philosophical position in turn. (a) The puerility of skepticism According to Pasnau, embracing skepticism or generally denying our ability to know, on the grounds that certainty is in principle not possible, is puerile. But is it? To begin with, if knowledge requires something that is in principle not possible, the denial of our ability to know seems trivial. But trivialities are truths. So, if knowledge requires certainty, and if certainty is out of reach, then we are indeed doomed to skepticism. On Pasnau’s view, however, this only shows that we are using too strong a concept of knowledge and that we should operate instead with a weaker concept: The verb “to know” is so indispensable, indeed, that if on skeptical grounds we were not allowed to use it, we would have to go out and find some other word to take its place. (lect. 6, p. 125)

Certainly, the view that there are different standards of knowledge is a serious philosophical option—contextualists about knowledge have become famous for embracing it—and I am quite sympathetic to it. But this option cannot simply be taken for granted, but has to be defended against the (traditional) philosophical position that real knowledge does indeed require certainty. And simply pointing to the fact that real knowledge would be impossible if all knowledge had to be certain does not constitute a sufficient defense. This can be illustrated by appeal to the philosophical debate about freedom. Consider a libertarian who contends that being undetermined is a necessary condition for freedom. Suppose further that we have a powerful (Spinozist) argument to the effect that indeterminism is impossible. What

Stephan Schmid

65

follows from this? Well, for a libertarian it follows that we cannot be free. If we have (or believe that we have) independent reasons for believing that freedom requires indeterminism, it is not an option to adopt a weaker concept of freedom just because the adjective “free” is so indispensable that it cannot be given up. Instead, the staunch libertarian should defend an error theory about our widespread practice of taking ourselves to be free agents: she should explain this practice and our widespread conviction that attributions of freedom are often true, even though our practice is mistaken and such attributions are in fact false. Likewise, it seems to me, the defender of certain knowledge should adopt skepticism (in light of the fact that such knowledge is not attainable) and settle for an error theory of knowledge instead. But such a skeptic, it seems to me, is hardly puerile. (b) Epistemic defeatism Whatever evidence we have for our beliefs, there might always be further evidence that defeats our initial evidence and thereby destroys our alleged knowledge. This is the position that Pasnau recommends instead of skepticism, and for lack of a name he calls it epistemic defeatism. But is this position so unheard of? In fact, I think, I have heard of Pasnau’s epistemic defeatism under the name fallibilism—and it is very popular among contemporary epistemologists. Consider the two following (representative) quotations: The fallibilist holds […] the we are nowhere entirely immune from the possibility of error.14 [T]here is never a metaphysical guarantee to be had that such-and-such belief will never need revision.15

Clearly then, Pasnau’s epistemic defeatism is simply a variety of fallibilism. But it might be a hitherto unknown variety, for not every fallibilist would agree with the coherentist conception of justification or evidence that epistemic defeatism seems to involve for Pasnau. His

 14

Jonathan Dancy, An Introduction to Contemporary Epistemology (Oxford: Blackwell, 1985), p. 58. 15 Hilary Putnam, Words and Life (Cambridge: Cambridge University Press, 1994), p. 152.

Certainty: Last Champions and Beyond

66

epistemic defeatist could thus be characterized as a fallibilist who thinks that evidence only counts as evidence against the backdrop of many assumptions that might be questioned or even refuted. But is this position unheard of? Again, I doubt it. Consider the following passage from Wilfrid Sellars, where he criticizes what he identifies as the “Myth of the Given”: [T]he given, in epistemological tradition, is what is taken by these selfauthenticating episodes. These “takings” are, so to speak, the unmoved movers of empirical knowledge.16

Like Pasnau, Sellars agrees that there are “no good non-circular reasons for believing anything,” because such reasons would amount to untenable “unmoved movers” of knowledge. At the same time, Sellars is a staunch fallibilist, contending that with respect to empirical knowledge, and even science, one “can put any claim in jeopardy, though not all at once.”17 In fact, Sellars’s fallibilism is part and parcel of his rejection of the Myth of the Given in its general form. Pasnau’s epistemic defeatism might well be the right position to adopt once it is recognized that the ideal of certainty is untenable. But it seems to me that Pasnau is by no means the first to think so—though he might be the first to call the position “epistemic defeatism.” (c) Hope Accepting epistemic defeatism should not make us fall into skeptical despair. To the contrary, Pasnau recommends adopting the hope that, despite the real possibility of being mistaken, our firm beliefs actually hit the mark. I am sympathetic to this recommendation. I am also convinced that, in light of epistemic defeatism, this hope cannot be grounded in something like objective evidence (in particular, if “objective evidence” is supposed to yield “infallibility”; see lect. 6, p. 38). I wonder, however, whether it is not a bit too quick to dismiss any requirement for cognitive justification for our epistemic hope just because we cannot have any conclusive justification for it.

 16

Wilfrid Sellars, Empiricism and the Philosophy of Mind (Cambridge, MA: Harvard University Press, 1997), p. 77. 17 Ibid., p. 79.

Stephan Schmid

67

Pasnau’s dismissal of cognitive justification as a rational requirement for our epistemic hope strikes me as particularly premature in the face of other available options. Consider Sellars again. In light of his rejection of any non-circular justification, he writes about the rationality of empirical knowledge as a whole: [E]mpirical knowledge, like its sophisticated extension, science, is rational, not because it has a foundation but because it is a self-correcting enterprise which can put any claim in jeopardy, though not all at once.18

It seems to me that the fact that our search after truth constitutes a selfcorrecting enterprise provides a formidable (though not conclusive) cognitive justification for the hope with which we engage in this enterprise. We can be hopeful that our firm beliefs hit the mark because they have been formed in and through such a self-correcting enterprise, in which skeptical scenarios, though not impossible, seem rather unlikely. So why give up the idea that our epistemic hope can be cognitively justified, to hold instead that it “requires merely an optimism on the affective side— a cheerful willingness not to worry about the all-too-possible bad scenarios” (lect. 6, p. 138)? Pasnau’s claim that there is no need for any cognitive justification of our epistemic hope not only strikes me as premature; it also makes it hard for me to grasp the difference between his optimistic epistemic defeatism and Hume’s mitigated or “true” skepticism, which he takes pains to distance himself from. In the famous conclusion of Book 4 of his Treatise of Human Nature, Hume addresses the question whether our natural inclination to abide by our beliefs is irrational, given that these beliefs do not withstand rigorous philosophical scrutiny. His answer to this question is clear and simple: “No: If I must be a fool, as all those who reason or believe any thing certainly are, my follies shall at least be natural and agreeable.”19 In Hume’s view, then, we are entitled to abide by our firm beliefs not because they could be rationally justified, but because rational justification is no option in the first place. So, instead of making our entitlement to hold beliefs dependent on standards that cannot be met in principle, Hume suggests making it dependent on standards that can be met. And one of these standards is the standard of leading a good or pleasant life.

 18

Ibid. David Hume, A Treatise of Human Nature 1.4.7.10, ed. David Fate Norton and Mary J. Norton (Oxford: Oxford University Press, 2000), p. 175. 19

68

Certainty: Last Champions and Beyond

Pasnau, it seems to me, adopts a rather similar strategy when it comes to justifying his recommended attitude of epistemic hope: If the only thing in life that we value is maximizing our ratio of true to false beliefs, then evidentialism is undoubtedly rational and suspension of belief would seem to be the only rational response to epistemic defeat. But, inasmuch as we care about many things other than truth and falsity— inasmuch as we want to live rich, engaged lives—an attitude of hope recommends itself. (lect. 6, p. 138)

Like Hume’s resolve to stop worrying about the unattainability of certain knowledge, Pasnau’s attitude of epistemic hope is not justified by epistemic reasons, but merely by the practical fact that assuming such an attitude serves our desire to live rich and engaged lives. Yet Pasnau distances himself from Hume, arguing that according to his suggestion we overcome skepticism not “by intellectually refuting epistemic defeatism or by ignoring it in Hume’s manner, but simply by ceasing to worry about it” (lect. 6, p. 137). I am not sure, however, whether this captures Hume correctly. Hume ultimately opts for a position he calls “true skepticism”: the “true skeptic,” Hume explains, is “diffident of his philosophical doubts, as well as of his philosophical conviction; and will never refuse any innocent satisfaction, which offers itself, upon account of either of them.”20 So, the true skeptic for Hume does not simply ignore epistemic defeatism; rather, she stops being worried by our inability to provide conclusive justifications, and thereby gains the flexibility to enjoy any satisfaction arising either from rational considerations or from our natural instincts. My worries about Pasnau’s view on the kind of hope that he recommends assuming when we engage in epistemic activities can be put in the form of a dilemma: either there is a cognitive justification for this kind of hope or there is not. If there is—an option which I personally would prefer, and one which has been explored in the literature—then Pasnau is wrong when he writes that “[h]ope, as I am thinking of it, requires no such cognitive confidence. It requires merely an optimism on the affective side” (lect. 6, p. 138). If, on the other hand, Pasnau is right, and we can assume this hope only on practical grounds—perhaps on the grounds that “we want to live rich, engaged lives” (lect. 6, p. 138)—I fail to see the big difference

 20

Hume, Treatise 1.4.7.14, p. 177.

Stephan Schmid

69

between Pasnau’s position and Hume’s “true skepticism,” even though Pasnau presents his position as an alternative to skepticism.

Conclusion Rather than challenging the scholarly acumen of Robert Pasnau’s breathtaking journey through (as the book’s subtitle has it) the “history of our epistemic ideals and illusions,” I hope that my critical remarks reveal its historical and philosophical substance. In presenting a detailed examination of the history of epistemology in unprecedented depth and breadth, After Certainty considerably advances our understanding of this central philosophical subject. No prophetic gifts are needed to foresee that what has been said about his previous book applies to After Certainty as well: it will “be a standard fixture on reading lists for many years to come and […] spark many a journal article.”21 In its capacity to incite further research by providing us with a plethora of claims and observations that are as sweeping as they are controversial, Pasnau’s After Certainty—which is itself an exercise in the search after truth (about epistemology)—is in fact fully in line with what it describes as the ideal form of epistemic activity. In its agile prose it cheerfully ranges over two millennia of the history of philosophy, developing a wellfounded (though defeasible) argument that certainty, which was for a long time considered the ideal of our epistemic activities, is but an illusion. It is thus a powerful example of important epistemic values—like being thought-provoking, explanatorily powerful, and inspiring—well beyond certainty.22

 21

See note 2 above. I am grateful for the discussion with the audience and participants at the workshop on Pasnau’s After Certainty held in Paris in March 2018, organized by Martin Klein, Christophe Grellard, and Jacob Schmutz. I have particularly profited from comments by Robert Pasnau, Sebastian Bender, and Jean-Baptiste Guillon. In revising my remarks I have also profited from discussions with José María Sánchez de León Serrano about certainty in German Idealism, and from excellent suggestions by Ian Drummond concerning my English. 22



BEYOND THE IDEAL, THE SOCIAL? CHRISTOPHE GRELLARD

1. It is not easy to put forward a critical reading of a book whose thesis seems to me entirely correct. One solution that would ultimately be of little intellectual value would be to make a nitpicking list of my disagreements on various points of detail, even though they in no way call into question the general thesis. I have therefore decided on a different approach. I will mention one basic disagreement I have with Robert Pasnau’s book, but mainly I will try to address the distinction that Pasnau finally draws between “believing faithfully” and “believing hopefully,”1 to propose a third possible way, which rests on an approach that gets surprisingly brief treatment in his book, namely, that of social epistemology. I say “surprisingly brief,” since it was the main theme of an important article by Pasnau himself, in which he already introduced some of the issues presented in this book.2 What Pasnau’s book shows very well is that the epistemology that developed between the end of antiquity (beginning with a Stoic-influenced reading of Aristotle that appeared mainly in the school of Alexandria) and the thirteenth century, and which came to its full realization in Descartes, is an idealized epistemology centered on the search after certainty, understood as the search for infallibility, indubitability, fixity, and necessity.3 Its goal is to secure cognitive mental states absolutely so that one can be entirely immune to the risk of error. This search after certainty leads to the introduction of a set of dichotomies in the way cognitive acts

 1

See After Certainty, lect. 6, pp. 133–138, and nn. 15–19, pp. 328–336. Robert Pasnau, “Medieval Social Epistemology: Scientia for Mere Mortals,” Episteme 7 (2010): 25–41. 3 On the evolution of the idea of certainty in the Middle Ages, see the very precise and comprehensive account in Joël Biard, “Certitudo,” in Iñigo Atucha et al. (eds.), Mots médiévaux offerts à Ruedi Imbach (Turnhout: Brepols, 2011), pp. 153–162. 2

72

Beyond the Ideal, the Social?

are conceived. (1) First, adopting a view that goes back to a Platonic tradition, this epistemology establishes a difference between belief (opinio, doxa) and knowledge (scientia, epistƝmƝ) that is not just one of degree but of nature. (2) There is a corresponding difference, also of Platonic inspiration, at the level of objects between those that are necessary and immutable (i.e., essences), which are the object of science, and those that are contingent, which are the object of opinion. (3) Along with this twofold difference there is the privilege accorded to the logicalmathematical model of science. (4) Finally, in an innovation that came later and really culminated in Descartes, there is an epistemic privilege accorded to the individual subject, who is conscious and endowed with self-reflexivity, which is also the privilege accorded to intuition, inasmuch as knowledge is thought of on the model of vision; this is likewise a very Platonic way of thinking, and is well attested in Augustine with his theory of acies mentis.4 All these elements are found in exemplary fashion in Robert Grosseteste’s commentary on Aristotle’s Posterior Analytics, which will be central for all later medieval readings of this text.5 In a certain sense, the slow dismantling of this idealized epistemology consists in “making room for belief” (to borrow Kant’s famous phrase, in slightly modified form). For this, it was necessary first to introduce a quantitative (as opposed to qualitative) difference between believing and knowing, as is made possible by epistemic probabilism. It then had to be granted that contingent things can be objects of reliable knowledge, which presupposed both a critique of essences and the development of tools that would allow for the possibility of such knowledge. It was further necessary, by way of developments in ethical thought, to restore to the notion of akribƝs its proper meaning, which had been obscured by its translation as certitudo, already by James of Venice in his translation of the Posterior Analytics (1.27, 83a31–33, where the topic is akribestera epistƝmƝ, which James renders as certior scientia),6 and then by Robert Grosseteste in his

 4

See Frederick Van Fleteren, “Acies mentis,” in Allan D. Fitzgerald et al. (eds.), Augustine through the Ages: An Encyclopedia (Grand Rapids: Eerdmans, 1999), pp. 5–6. 5 Robert Grosseteste, Commentarius in Posteriorum analyticorum libros, ed. Pietro Rossi (Florence: Olschki, 1981). On this important work, see Eileen Serene, “Robert Grosseteste on Induction and Demonstrative Science,” Synthese 40 (1979): 97–115, and Amos Corbini, La teoria della scienza nel XIII secolo: I commenti agli Analitici secondi (Florence: SISMEL–Edizioni del Galluzzo, 2006). 6 On this point, see Sten Ebbesen et al., History of Philosophy in Reverse: Reading Aristotle through the Lenses of Scholars from the Twelfth to the Sixteenth

Christophe Grellard

73

translation of the Nicomachean Ethics. The tools for this change were found in the later Middle Ages in the notions of probable certainty and materia subiecta.7 Finally, it was necessary to put an end to the epistemic privilege of the subject (a point to which I will return).

2. As I have said, I do not have any fundamental disagreements with this vision of medieval epistemology and its early modern turn. However, there is one point on which I would be more cautious than Pasnau, and that is the role of skepticism. Of course, one should not overestimate the presence of skepticism in the Middle Ages: as a proportion of the number of pages filled by the theologians and philosophers of the era, it is tiny. But it is also necessary to move beyond the outmoded historiographical model of Richard Popkin, who connects skepticism, the Reformation, and modernity.8 In fact, even if no one in the Middle Ages was a skeptic— leaving aside a humanist Ciceronian tradition of “soft” skepticism found in John of Salisbury, Petrarch, and even Jean Gerson in certain respects9— there was a continuous discussion of skeptical themes, based on what was known of the Presocratics from Aristotle’s Metaphysics and on Augustine’s “refutation” of the Academici.10 The Augustinian model of a skeptical moment in knowing, as a mode of refuting empiricism and materialism, and preparing for Platonism, was widely adopted in the Franciscan tradition (including Bonaventure, Matthew of Aquasparta,

 Centuries (Copenhagen: The Royal Danish Academy of Sciences and Letters, 2014), pp. 148–165. 7 Attention was drawn to this important notion in the pioneering article of Zénon Kaluza, “Les sciences et leur langage: Note sur le statut du 29 décembre 1340 et le prétendu statut perdu contre Ockham,” in Filosofia e teologia nel Trecento: Studi in ricordi di Eugenio Randi, ed. Luca Bianchi (Louvain-la-Neuve: FIDEM, 1994), pp. 197–258. 8 Richard H. Popkin, The History of Scepticism from Erasmus to Descartes (Assen: Van Gorcum, 1964). 9 See Christophe Grellard, “Y a-t-il une tradition sceptique au Moyen Âge?,” in Dragos Calma and Zénon Kaluza (eds.), Regards sur les traditions philosophiques (XIIe–XIVe siècles) (Leuven: Leuven University Press, 2017), pp. 187–209. 10 See Christophe Grellard, “Comment peut-on se fier à l’expérience: Esquisse d’une typologie des réponses médiévales au scepticisme,” Quaestio 4 (2004): 113– 135; idem, “Academicus,” in Mots médiévaux offerts à Ruedi Imbach, pp. 5–16; idem, “Le problème sceptique au Moyen Âge de saint Augustin à Nicolas d’Autrecourt: Réception et transformation d’un problème philosophique,” Cahiers philosophiques 153 (2018): 55–78.

74

Beyond the Ideal, the Social?

James of Viterbo, Peter of John Olivi, and of course Henry of Ghent). Likewise, the question of the possibility of knowing soon makes its appearance in commentaries on the Metaphysics and the Posterior Analytics (from Siger of Brabant to John Buridan, including Radulphus Brito). There are multiple reasons for the presence of skepticism in late medieval discussions, but one of them probably had to do with the fact that it is the negative counterpart of idealized epistemology. When one imposes criteria for knowing that are so high that they are difficult to meet at all, one opens the door to skepticism in the form of a calling into question of the very possibility of knowing. Yet one can go further, and maintain that there was a true skeptical moment in the fourteenth century, between 1320 and 1380, that was tied to the development of the argument from a deceiving God. Contrary to what Pasnau asserts, divine deception is not “so obvious and natural that its constant reoccurrence in the history of philosophy requires no special explanation” (lect. 6, n. 1, pp. 309–310). If one is careful to distinguish demonic deception—which is simply a special form of sensible illusion, and an isolated and limited case—from divine deception, which presupposes a power to overturn all reality, it is clear that for the idea of divine deception in the proper sense to appear, a certain intellectual configuration had to be in place. Such a configuration is found in Duns Scotus with the conjunction of the radical contingency of creation and a conception of divine omnipotence that includes the power to suspend the laws of nature. This conjunction of theses belonging to what is called “post-Scotist theology”11 was taken up in the 1320s by the English Franciscans (e.g., William of Ockham, Walter Chatton, John Rodington, and Adam Wodeham), then by the Dominicans (e.g., William Crathorn and Robert Holcot), after which it spread more widely and prevailed on the continent. This systematic use of the idea of deception, which strikes at the very heart of knowing (its causal functioning) and which gives the deceptiveGod argument its skeptical force, can be found, in a reductio ad absurdum, in Nicholas of Autrecourt’s first letter to Bernard.12 Indeed, the hypothesis

 11

Luca Bianchi and Eugeno Randi, Vérités dissonantes: Aristote à la fin du Moyen Âge (Paris and Fribourg: Éditions du Cerf / Éditions universitaires de Fribourg, 1993), pp. 111–151. 12 Christophe Grellard, “Nicholas of Autrecourt’s Skepticism: The Ambivalence of Medieval Epistemology,” in Henrik Lagerlund (ed.), Rethinking the History of Skepticism: The Missing Medieval Background (Leiden: Brill, 2010), pp. 119–143.

Christophe Grellard

75

of a deceiving God opened the way for various ways of testing the limits of the existing model of cognition, which resulted in profound changes in the contours of skepticism, and made possible the transition to early modern skepticism. Recall how Nicholas makes his argument. Since Bernard posits a distinction between an intuitive cognition and its object, it is logically possible for one to exist without the other: Thus, it is clear, it seems to me, that from your claims it follows that you have to admit that you are not certain of the existence of the objects of the five senses. But what might be even harder to stomach: you must say that you are not certain of your own acts, for example, that you are seeing, or hearing, and what is worse, that you are not certain that anything is, or has been perceived by you. For in the passage cited above, the first book of the Sentences dist. 3, you say that our intellect does not have intuitive cognition of our actions.13

Thus, one can have a clear representation of an object, even when that object does not exist. Nicholas immediately deduces from this that it is impossible to have any certainty about the existence of external objects. The same goes for our cognitive mental states: according to Bernard, we do not have direct intuitive cognition of our own cognitive acts, but only abstractive cognition (which is indirect and so less clear). Nicholas is thus making an a fortiori argument, for since Bernard does not have certain cognition of the external objects that appear to him clearly, there is all the more reason why he cannot have certain cognition of the objects that appear to him less clearly, such as mental states: And thus, reviewing and summing up your position, it appears that you have to admit that you are not certain of those things which are outside of you. And so you do not know if you are in the sky or on earth, in fire or in water. And, consequently, you do not know whether today’s sky is the same one as yesterday’s, because you do not know whether or not there was any sky. Just as you do not know if the Chancellor, or the Pope, exists and if they exist, whether they are not, perchance, different persons in any given moment of time. Similarly, you do not know what things are in your

 13

Nicholas of Autrecourt, His Correspondence with Master Giles and Bernard of Arezzo, ed. and trans. L. M. de Rijk (Leiden: Brill, 1994), pp. 52–53: “Sic igitur clarum est ( michi videtur) quod ad dicta vestra sequitur quod vos habetis dicere quod vos non estis certus de existentia obiectorum quinque sensuum. Sed quod gravius sustineri posset: habetis dicere quod vos non estis certus de actibus vestris, utputa quod videatis, quod audiatis; ymo quod non estis certus quod aliquid appareat vobis vel apparuerit vobis. Nam (ubi supra, primo Sententiarum, dist. 3) dicitis quod intellectus noster de actibus nostris non habet intuitivam notitiam.”

76

Beyond the Ideal, the Social? direct surroundings, as whether you have a head, a beard, hair and the like. Hence it follows a fortiori that you are not certain of things which now belong to the past, as whether you have been reading, or seeing, or hearing.14

Nicholas can then conclude by trapping his adversary in complete solipsism: Bernard does not know whether the external world exists; he does not know whether he himself exists as a body; nor does he know whether he has mental states, and consequently cannot know whether he is in fact the one thinking. He is thus reduced to ignorance about his own existence as an intellect. One can thus see that this portrait of the skeptic has more a modern flavor than an ancient one. Someone is a skeptic, who, because of an improper usage of divine omnipotence, knows nothing and is therefore reduced to solipsism and inaction. Here we see a radical change in the conception of skepticism, the touchstone of which is solipsism. Recall what is said by Augustine, according to whom the Academic skeptics never dared to go so far as to call into question the existence of the external world: Your arguments were never able to refute the force of the senses such that you might convince us that nothing appears to us, and you never at any time dared to attempt to do so. But you made great efforts to persuade us that something can be other than what it appears to be.15

We have here a distinctive mark of the ancient and early modern forms of skepticism, and an indication of the fundamental changes that took place in the Middle Ages. For the argument from solipsism to work, the argument from a deceiving God was needed, to allow for the complete overturning of reality. These changes internal to skepticism were taking

 14

Nicholas of Autrecourt, His Correspondence, pp. 54–55: “Sic igitur, recolligendo dicta, apparet quod habetis dicere quod vos non estis certus de illis que sunt extra vos. Et ita nescitis si sitis in celo vel in terra, in igne vel in aqua. Et, per consequens, nescitis si hodie sit idem celum quod heri fuit, quoniam nec scitis si celum fuit vel . Sicut etiam nescitis si Cancellarius vel Papa sit, et, si isti sint, an sint alii et alii homines in quolibet momento temporis. Similiter nescitis que sunt intra vos, ut si habetis caput, barbam, capillos et cetera. Ex isto sequitur a multo fortiori quod vos non estis certus de hiis que transierunt in preteritum, ut si legistis, vidistis vel audistis.” 15 Augustine, Contra Academicos III.11.24: “Unde, inquit, scis esse istum mundum, si sensus falluntur? Nunquam rationes vestrae ita vim sensuum refellere potuerunt, ut convinceretis nobis nihil videri: nec omnino ausi estis aliquando ista tentare; sed posse aliud esse ac videtur, vehementer persuadere incubuistis.”

Christophe Grellard

77

place at the same time as a growing attention to contingent things as the object of cognition, and a greater epistemic privileging of the subject. Thus, skepticism developed in much the same way as the epistemological theories of which it is the negative counterpart. It is thus clear that skepticism remained as a problem beneath the surface in the Middle Ages and contributed to profound changes in late medieval epistemology.

3. I come now to my second point, namely, Pasnau’s ultimate proposal to leave behind idealized epistemology. This rests on the distinction between “believing on faith” and “believing on hope.” Believing on faith consists in (1) believing, that is, being committed to truth and being disposed to act; (2) according a high degree of confidence (“high credence”) to this belief; and (3) doing so despite the epistemic insufficiency of our warrants, or as Pasnau puts it, “even though one does not suppose that the evidence warrants such confidence” (lect. 6, p. 135; cf. n. 16, pp. 330– 331). In other words, it is a matter of forcing oneself to go beyond the insufficiency of our warrants by acting as if they were sufficient. One can readily recognize in this the medieval model of fides, especially as it appears in Thomas Aquinas, for whom this epistemic leap depends on the will, which he says can force the intellect to give its assent when the natural force of the arguments does not suffice.16 Aside from its voluntarist difficulties—How can one decide to believe in the absence of evidence, or even against the evidence?—Pasnau maintains that this position is incoherent because the cognitive agent would be maintaining a form of credence even while being aware that the available proof calls for a different credence (see lect. 6, p. 135). Here, however, it seems to me that he falls victim to one of the idols of idealized epistemology from which he has not succeeded in freeing himself, namely, the idea that awareness is transparent to itself and that one’s beliefs are necessarily mutually coherent. It seems to me rather that, as the “compartmentalization principle” (principe de coupure) in anthropology teaches, an individual’s beliefs form systems that are mutually incompatible, but which are never in play at the same time.17 Rather,

 16

On this issue, see Christophe Grellard, De la certitude volontaire: Débats nominalistes sur la foi à la fin du Moyen Âge (Paris: Publications de la Sorbonne, 2014). 17 See especially Roger Bastide, “Le principe de coupure et le comportement afrobrésilien,” in Anais do 31e Congresso Internacional de Americanistas (Sao Paolo:

78

Beyond the Ideal, the Social?

beliefs are brought into play according to the circumstances and external stimuli to which we must respond (since a belief is a disposition to action). But I leave this point aside, for it would take us too far afield; further on I shall propose a different way of defending “belief on faith.” The alternative solution that Pasnau proposes is “belief on hope.” It consists of (1) believing (and thus giving assent to the truth and being disposed to action), (2) without having great confidence in this truth, but without being impeded by this lack of confidence thanks to an affective factor: (3) a hope that has the function of suppressing fear (in the background of this lies the traditional medieval definition of opinion as assent accompanied by fear or hesitation; see After Certainty, lect. 6, p. 135). In other words, it consists in relying on the unavoidably affective component of our mental states in order to orient our assent. What is interesting about this solution is that it avoids the problem of voluntarism (since it is not the will that suppresses fear) and the decision to believe, and it offers a place for the affective dimension of cognition; in other words, it proposes to set epistemology within the more general framework of a theory of human nature. That said, it has to be conceded that concretely it is hard to see either (1) what exactly this affect of hope is and how it operates, or (2) how it will be that optimism does not introduce other forms of error (in the form of self-deception for example). Concerning the first point, it should be noted that the examples that Pasnau gives in note 18 on Lecture Six appeal mainly to hope considered as a theological virtue (see pp. 333–336). But if our cognition depends on this virtue, we fall into the problem of the atheist mathematician—or in this case, the atheist epistemologist—for whom such hope is impossible. Will we have to conclude then that certainty is denied to someone who has not received this theological virtue? Here we see the limits of a cognitive model that was developed within the context of scholastic theology, and the problems this raises. It seems to me therefore that there is an alternative way that can be proposed, which rests moreover on the idea of “belief on faith” and accepts certain medieval conceptions of fides in order to propose a more social orientation to cognition. Indeed, it seems to me that it is not in

 Anhembi, 1955), vol. 1, pp. 493–503. In a similar way, Paul Veyne defends the idea of a “Balkanization of minds” in Did the Greeks Believe in Their Myths? An Essay on the Constitutive Imagination, trans. Paula Wissing (Chicago: University of Chicago Press, 1988).

Christophe Grellard

79

medieval theories of scientia, which are part of idealized epistemology, that we will be able to find fruitful epistemological ideas, but in discussions of fides, in all senses of the term—both as the act of trusting that is produced by argument, and as the act of religious faith.18 What we are seeking in fact are the elements of a theory of sufficient assent (i.e., sufficient for action), without fear but having a minimal warrant. How can we secure both the suppression of fear and a minimal warrant? It seems to me that a solution can be found in a certain form of social epistemology. The best solution for reducing fear is to make the cognitive risks shared. Against the epistemic privilege of the self-transparent subject, we must restore to cognitive activity its participatory and social dimension: cognition is constructed in common and shared by the agents within a given community.19 Now, such a conception of knowledge began to arise at the end of the Middle Ages, not only in the context of theorizing about scientia, but also in theories of fides. It is all the more surprising that Pasnau barely alludes to it in his book, given that he himself emphasized this point in the article I have already mentioned.20

4. The idea that knowledge is produced collectively, and that it is produced by having trust in those with whom one lives, is an ancient idea. Augustine uses it in a text that was widely taken up in the Middle Ages, where he plays on the ambiguity of the term fides to show that it is central to every society: So I ask: if one should not believe that which one does not know, how will children be obedient to their parents, and love them with shared piety if they do not believe that they are their parents? For this is something that reason is powerless to make known. Regarding the father, one believes on the authority of the mother; regarding the mother herself, one usually relies not on the mother’s own testimony, but on that of midwives, nursemaids, and servants. For if someone’s son was stolen and replaced by another, could she not, having been deceived, deceive in turn? We believe her, however, and we believe her without hesitation, because we admit that we



18 On this polysemic concept, see Christophe Grellard, Philippe Hoffmann, and Laurent Lavaud (eds.), Genèses antiques et médiévales de la foi (Paris: Études Augustiniennes, 2019). 19 This idea was received quite late in Anglo-American epistemology. See the pioneering work of Alvin Goldman, Knowledge in a Social World (Oxford: Oxford University Press, 1999). 20 See note 1 above.

80

Beyond the Ideal, the Social? cannot know. Without this, would we not see filial piety, this sacred bond of society, be disdained and insulted by crime? […] One can very well give reasons that would prove that absolutely nothing in society remains standing if we have decided to believe nothing because we cannot have knowledge that is certain.21

Here Augustine lays out the impasse of idealized epistemology—in many cases “we admit that we cannot know,” that is, we cannot attain the level of infallible certainty—and proposes to reinforce the credit one accords to testimony. Trust in the beliefs held by others is indispensable for living in society, and in certain cases this trust can even translate into a high degree of certainty (“We believe without hesitation”). The value of testimony lies in taking into account the affective dimension of trust (one’s relationship with family and friends), as well as the rational dimension of the criteria of credibility. Indeed, testimony rests fundamentally on a principle of acceptance.22 This idea is found in William of Ockham, for example, for whom the agreement among a number of trustworthy men compels us to assent, unless there is stronger evidence to the contrary: In matters of fact, therefore, especially those which they imply they know directly, everyone is bound to believe them, unless he is certain of the contrary through those more worthy of belief.23

 21

Augustine, De utilitate credendi xii.26 (ed. J. Zycha, CSEL 25 [Vienna: Tempsky, 1891], p. 34): “Quaero enim, si quod nescitur, credendum non est, quomodo serviant parentibus liberi, eosque mutua pietate diligant, quos parentes suos esse non credant. Non enim ratione ullo pacto sciri potest: sed interposita matris auctoritate de patre creditur; de ipsa vero matre plerumque nec matri, sed obstetricibus, nutricibus, famulis. Nam cui furari filius potest, aliusque supponi, nonne potest decepta decipere? Credimus tamen, et sine ulla dubitatione credimus, quod scire non posse confitemur. Quis enim non videat pietatem, nisi ita sit, sanctissimum generis humani vinculum, superbissimo scelere violari? Nam quis vel insanus eum culpandum putet, qui eis officia debita impenderit quos parentes esse crediderit, etiamsi non essent? Quis contra non exterminandum iudicaverit, qui veros fortasse parentes minime dilexerit, dum ne falsos diligat metuit? Multa possunt afferri, quibus ostendatur nihil omnino humanae societatis incolume remanere, si nihil credere statuerimus, quod non possumus tenere perceptum.” 22 See for example H. H. Price, Belief, 1960 Gifford Lectures (London: Allen & Unwin, 1969), p. 116: “Accept what you are told, unless you see reason to doubt it.” 23 William of Ockham, Dialogus, pars 3, tract. 1, lib. 3, c. 23 (ed. John Kilcullen et al. [Oxford: Oxford University Press / British Academy, 2011], p. 288): “Et ideo in his quae facti sunt, praesertim de his quae per se ipsos se insinuant cognoscere, tenetur quilibet credere ipsis, nisi per fide digniores de contrario certus existat.”

Christophe Grellard

81

Thus, it is necessary to believe testimony unless one can prove that it is false. Of course, one must then give up on infallibility and be ready to correct one’s beliefs a posteriori. Ockham agrees that testimony does not give us infallible certainty, but mere credulitas, that is, a certainty that is possible and sufficient (certitudo possibilis et sufficiens).24 But one can go still further by defending the view that, socially and anthropologically, man is disposed to believe testimony.25 Robert Holcot argues for this by appealing to an epistemological fact: Any prudent man notices that he himself has frequently erred in judging and in taking false things for true, so that if he should believe some proposition with strong assent, and a multitude of learned people tell him affirmatively the opposite, then notwithstanding his strong assent and any prior evidentness, he will begin to doubt and to break down the assent caused by known propositions or the evidentness of the thing, and the opposite assent will come about. Thus, for the present, I adhere to the following proposition: if a straight line intersects another straight line it will cause two right angles or two that are equal to two right angles; and this seems to me evident. And yet, if a whole crowd of geometry students asserted that I had been deceived and that the opposite was true, I would begin to doubt, and even assent to the opposite of the proposition, solely on account of someone’s testimony.26

The existence of error—that is, the fallibility of individual judgment, which is something everyone can experience—reinforces the need to

 24

William of Ockham, Dialogus, pars 3, tract. 1, lib. 3, c. 24 (ed. Kilcullen et al., p. 289). 25 Holcot states explicitly a view that is very similar to Thomas Reid’s principle of credulity. See Robert Holcot, Quaestiones super Sententiarum, lib. 1, q. 1, art. 6 (Lyon: Josse Bade, 1497), fol. A-5v: “Naturaliter homo est natus credere testimonio aliorum in his quae ipse non novi.” 26 Ibid.: “Praeterea quilibet homo prudens advertit se ita frequenter errasse in iudicando et falsa probando per veris quod si credat aliquam propositionem vehementi assensu et aliqua multitudo homini scientificorum diceret assertive oppositum non obstante vehementi assensu et evidentia praecedente incipiet dubitare et corrumpere assensum causatum per propositiones notas vel per evidentiam rei et fieret assensus oppositus. Unde satis modo adhaereo isti propositioni: si linea recta supra lineam rectam steterit causabit duos angulos rectos vel aequales duobus rectis et apparet mihi evidens. Et tamen si tota multitudo studentium in geometria assereret me esse deceptum et oppositum esse verum inciperem dubitare immo assentire opposito illius propter solum testimonium alicuius.”

82

Beyond the Ideal, the Social?

confirm one’s own judgment by a shared higher-level warrant, especially about questions that are obscure and difficult. A warrant of this kind is provided in particular by the scientific community, and so knowledge cannot be the result of an exclusively individual process. Suppose, for example, that someone has a strong opinion about a mathematical theorem, and it is based on good arguments, though not on a scientific demonstration (as is often the case with people who are not experts in the discipline in question but have a passing but relatively superficial understanding of it). If this person sees that several or all of the experts in the field assert that her opinion is false, she will start to doubt, her assent to the proposition will weaken, and her intellect will become susceptible to accepting the opposite thesis. There is thus a communal or social aspect to the production of knowledge. The individual approach is prone to error and failure, while collaboration reduces this risk (without eliminating it entirely). The epistemic value of testimony is therefore rooted in the necessarily collective activity of knowing. If cognition can thus be socially constructed, what allows for the sharing of the risk of error is cognitive work that is shared socially. There is an example of this in John Major’s use of the idea of certitudo moralis:27 It is not evident to me that the body of Christ is contained in this Host. Proof: It can happen that the bread was not made of wheat, or that the priest did not intend to consecrate [the Host], or that he did not utter the words necessary for consecration, and so there was no consecration. Then it is argued on the same point: it is not certain to the priest that he consecrated it, and so it is not certain to me either. The inference holds, since he can discern this better than I can. The antecedent is clear: it can happen that he was not baptized, and so is not a priest, or else that a bishop did not ordain him; therefore etc.28

 27

Pasnau addresses the question of certitudo moralis in After Certainty, pp. 198– 201. I discuss John Major’s text in detail in Christophe Grellard, “Comment peuton croire à la transsubstantiation? La certitude morale et ses fondements sociaux chez Jean Mair,” Przeglad Tomistyczny 24 (2018): 627–650. 28 John Major, In quartum Sententiarum d. 11, q. 3 (Paris: J. Petit, 1519), fol. 59vb: “Non est mihi evidens quod corpus Christi continetur in hac hostia. Probatur. Stat panem non fuisse triticeum vel sacerdotem non intendisse consecrare, vel non protulisse verba necessaria consecrationi, et tunc non fuisset consecratio. Secundo arguitur ad idem: non est certum presbytero an consecrauerit, ergo non est mihi certum. Tenet consequentia, quoniam ipse potest hoc melius deprehendere quam ego. Antecedens patet. Stat eum non esse baptizatum, et sic non est sacerdos. Stat insuper episcopum non ordinasse eum, igitur.”

Christophe Grellard

83

The case described by John Major might seem like an idle example far removed from any serious epistemological investigation. But this is far from the case; rather, the situation he describes is a fundamental one in medieval society, since it has to do with the most important ritual at the heart of the Catholic Church, and it seems to lead necessarily to a form of skepticism. The issue is whether one can unconditionally venerate the Host at the moment of the sacrament of the Eucharist if one can never be certain that the priest has proceeded correctly in order for transubstantiation to take place (that is, he uttered the correct words with the necessary intention), or that the bread is made of wheat (a requirement reaffirmed at the Council of Florence in 1439).29 John Major concedes— and here we see implicitly the incursion of skeptical questions into unexpected domains—that one can never have infallible certainty that the consecration was performed correctly. This simple social fact thus endangers the whole of the epistemological edifice that had been patiently constructed from the thirteenth century onward. On the other hand, one can have a sufficient moral certainty, which is the product of a given social milieu: I have sufficient moral certainty that the body of Christ was transubstantiated; therefore under the species of bread there is the body of Christ. This inference is good, since I have moral certainty about the antecedent, and therefore I have it about the consequent. For if an inference is certain, and the antecedent is certain, the consequent is certain by the same certainty. But it is clear that I have moral certainty about the antecedent. We see that there is no difficulty about the baptism of one born among Christians and whose parents are known, either on the part of the parents and godparents, or on the part of the baptizer and the community. No difficulty arises about the intention of the bishop, nor about the priest’s intention to be ordained and to do what is fitting for an honorable man. For anyone should be presumed to be good until the opposite is shown; but the opposite is not shown about this priest, as is supposed [in the case adduced].30



29 On this point, see Roger Gaïse, Les signes sacramentels de l’Euchariste dans l’Église latine: Études théologiques et historiques (Fribourg: Éditions Universitaires de Fribourg, 2001), pp. 208–221. 30 John Major, In quartum Sententiarum, d. 11, q. 3, fol. 59vb: “Secunda conclusio: habeo certitudinem moralem sufficientem quod corpus Christi transubstantietur, ergo sub speciebus panis est corpus Christi. Haec consequentia est bona, quoniam certitudinem moralem habeo de antecedente, ergo talem habeo de consequente. Si enim aliqua consequentia sit certa, et antecedens certum, consequens eadem certitudine est certum. Sed certitudinem moralem habeo de antecedente quod patet. Videmus quod nulla est difficultas de baptismo nati inter

84

Beyond the Ideal, the Social?

There is no reason to doubt that the priest who consecrated the Host was in fact baptized, for there is a form of social control internal to the community (on the part of parents, godparents, neighbors, etc.), while the priest’s intention is guaranteed by a kind of “principle of credulity”: anyone is presumed to be good until there is proof to the contrary. There is a principle of trust that is collectively warranted. The social and collective dimension of faith thus makes it possible to establish a general principle of trust, such that any possible exceptions will involve a form of invincible ignorance. Moral certainty thus introduces the idea that it is not by an individual process that I can secure my action, but by relying on a principle of trust that is warranted collectively.

5. What I have tried to suggest in these brief remarks on skepticism and social epistemology is that in medieval theories of faith, and in the discussion of examples that might sometimes be surprising for a layperson, one can discern the beginnings of a theory of the social construction of knowledge, which offers an alternative to the two solutions proposed in the last lecture of After Certainty. Nevertheless, there is no doubt that Robert Pasnau’s book has identified a key element in the changes that took place in epistemology between the Middle Ages and the early modern era, namely, the construction of an idealized model of knowledge and its gradual weakening. The fact that this model culminated in Cartesian epistemology, which has extensively informed contemporary thought on knowledge in the English-speaking world both positively and negatively, clearly underlines all that can be learned from this book by philosophers, who are too often oblivious of their history. Indeed, as is revealed by the late and difficult entry of social epistemology into contemporary philosophy, and the all too frequent reduction of the social dimension to the intersubjectivity of two Cartesian subjects, idealized epistemology continues to have many effects. After Certainty is without question a major book on the history of philosophy, but it is also, and above all, indispensable reading for any philosopher who intends to engage in the field of epistemology today. For it is only with a knowledge

 christianos, et notos parentes, tum ex parte parentum et patrinorum, tum ex parte baptizantis et viciniae. De intentione episcopi non cadit difficultas, nec de intentione sacerdotis quin ordines suscipere intendit, et facere illa quae probum virum decent. Quilibet enim est praesumendus bonus donec de opposito doceatur, non docetur autem oppositum de hoc sacerdote, prout supponitur.”

Christophe Grellard

85

of the history that this book describes that one will be able to attain any degree of critical reflexivity on the practices of philosophers.



SOME THOUGHTS AFTER AFTER CERTAINTY ROBERT PASNAU

I am much indebted to my commentators for their extremely interesting remarks. The cumulative impression I take away from this symposium is that there is still a great deal of work left to do on this material, both in terms of enlarging the scope of the discussion beyond where I took it in After Certainty and in refining the details of those matters that I did manage to discuss. Although one might suppose that the history of epistemology would be ground as well trodden as any in philosophy, I think one can see from this discussion that there remains tremendous scope for research, even with regard to canonical figures. I will be glad enough if my book helps to motivate some of that future work. Here, moving one by one through the commentators, I will try to make a few suggestions of my own about how the argument of my book might be further developed.

Reply to Martin Klein I regard moral certainty as one of the most important developments in epistemology from the fourteenth century onward, and I take John Buridan to be the originator of the concept (although the notion has a prehistory that goes back to Aristotle). As I tell the story, a pivotal moment comes when seventeenth-century authors begin to accept that moral certainty is sufficient for knowledge (lect. 2, p. 43). But if this does not happen until the seventeenth century, then of course the question arises of how moral certainty relates to knowledge back in the Middle Ages, and particularly in the work of Buridan. I cannot see that Buridan clearly tells us what the relationship is between moral certainty and knowledge (that is, scientia). To be sure, moral certainty cannot be sufficient for knowledge, because one can be morally certain about a belief that is false. Martin thinks we have good reason to think that morally certain true beliefs do count as knowledge for Buridan, because they fit into his larger story about how evidentness is what

88

Some Thoughts after After Certainty

distinguishes mere true opinion from knowledge. As Martin puts it, “Buridan carefully draws a strict boundary between knowledge—in the most basic sense of the cognition of a proposition with certainty and evidentness (i.e., the first type of knowledge)—and opinion” (p. 28). Now, to be sure, Buridan offers moral certainty as a kind of evidentness. And I am inclined to agree with Martin that he probably thinks of it as good enough for a kind of scientia. But I do not think the boundary is at all “strict.” Nor do I think it is helpful to assimilate the case of moral certainty to the case of induction, as Martin does, remarking that “pretty much the same condition of evidentness” is involved in both. Buridan thinks that knowledge through induction meets a higher standard of evidentness, which can be characterized as conditional on the assumption that the regularities of nature remain constant, unimpeded by God or by anything else. I would be happy to think of moral certainty on this model if there were a story to be told about what condition needs to be presupposed for someone to count as being morally certain. What we would need is a proposition such that when that proposition is granted, the conclusion follows as a matter of strict necessity. But one of the striking and puzzling features of moral certainty is that there is no such condition. What Buridan tells us, in the only passage that expressly considers the matter, is this: “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.”1 This seems to amount to little more than an injunction to investigate diligently and believe accordingly, which has little in common with his treatment of induction. Perhaps what these distinct treatments do have in common is that they invoke evidentness as an essential requirement for scientia. I think this is Martin’s picture, and it is my picture too. As I say, “Roughly speaking, it [evidentness] is the all-important quality that distinguishes scientia from mere true belief” (lect. 2, p. 32). I think that Martin and I are also in agreement that the key feature of evidentness is that it compels assent. But I do not think that Buridan is attempting to give us a careful or strict boundary between knowledge and mere opinion. And for me this is symptomatic of the prevailing medieval attitudes to epistemology, which have very little concern for identifying a strict boundary that marks where

 1

Quaestiones in Metaphysicam II.1, (ed. Paris 1518, fol. 9ra), translated in Gyula Klima, Fritz Allhoff, and Anand Jayprakash Vaidya (eds.), Medieval Philosophy: Essential Readings with Commentary (Malden, MA: Wiley-Blackwell, 2007), p. 145.

Robert Pasnau

89

knowledge begins. Instead, they tend to think of our belief states as running along something more like a continuum, more or less approaching an ideal, and more or less adequate for various different purposes in different contexts. Martin’s other topic is sense perception. He questions whether Buridan is as innovative and distinctive as I claim he is. One way in which Buridan clearly is distinctive is that he simply refuses to treat the proper sensibles as a domain of distinctive privilege. There’s a long history here of attempts to defend Aristotle’s doctrine that the senses never (or very rarely?) go wrong with regard to the proper sensibles. Various constraints were put on the theory to make it come out right: one needs the proper medium, the proper distance, the proper condition of the organ, and so on. Buridan’s view is that once all these necessary constraints are put in place, we have to conclude that the senses are just as infallible with respect to the common sensibles as they are with respect to the proper sensibles. What I take to be the most interesting and revisionary feature of Buridan’s view, however, is the way he seems to attack the very distinction between the per se and per accidens sensibles. Here let me right away draw a distinction that is perhaps not sufficiently clear in the book. One way to draw the per se–per accidens distinction is at the causal level. One might say that the per se sensibles are the ones that make a direct causal impact on the senses—that is, the qualities that are the causal agents for a specific perceptual modality. This is a question of physics. Officially, Buridan retains the standard physics: he thinks that sight, for instance, works by colors making their way through an illuminated medium, from object into eye. In virtue of receiving the species of such colors, sight sees everything that it sees. Martin stresses this point, and adduces some texts, and I do not deny it. (I do think Buridan is suspicious of this story, because he has doubts about the usual story about real qualities, but that’s a different issue.) The issue I mean to focus on is epistemological. For even if the colors are real qualities and are the vehicle of visual perception, this is still not to say that the senses are capable of grasping this. On the contrary, Buridan thinks that the epistemic situation at the level of the senses is wholly confused. Here’s the passage I quote in the book: Although an external sense cognizes Socrates, or whiteness, or something white, still this happens only through a species represented confusedly and all at once along with the substance, the whiteness, the size, and the

90

Some Thoughts after After Certainty location, in accord with how it appears in the cognizer’s field of vision (in prospectu). The sense cannot distinguish the confusion. That is, if it cannot abstract from one another the species of substance, whiteness, size, and location, then it cannot perceive substance, whiteness, or something white, except in the way it exists in its field of vision.2

Accordingly, whatever the mechanism of vision is, it is epistemically hidden from us at the sensory level. This is not to say, however, that we are incapable of understanding the sensory process, or that we are incapable of understanding the nature of the sensible qualities that make an impression on the senses. Buridan thinks that we can reach philosophical conclusions on these subjects. Indeed, as Martin stresses, Buridan goes on at some length about how these stories go. But my point is that he does not think that sensation itself delivers these verdicts. At the sensory level, our perceptions are confused in such a way that it is misleading to speak of there being proper sense objects at all. Our mistake is to read our philosophical conclusions about the sensory process back into the sensory level, and to suppose that sensory experience itself delivers all the various conceptual distinctions between sense objects that we associate with Aristotelianism.

Reply to Sebastian Bender I am glad to have the opportunity, in responding to Sebastian’s comments, to reflect at least briefly on Leibniz, who does not play much of a role in the book. Sebastian holds Leibniz out as an example of an author whose view is (on my taxonomy) reductive, while still satisfying the demand for fidelity. It was the lack of fidelity, I claimed, that made reductivism unappealing to Descartes. But Leibniz tries to have it both ways. He gets fidelity because he distinguishes between what our sensory ideas “precisely express” and what we can “notice” or “distinctly consider” (drawing on the passages quoted above on pp. 41-42). As Sebastian points out, Descartes’s doctrine of transparency precludes his adopting this sort of view: the content of our ideas, for Descartes, must at least be noticeable. I agree that this presents us with a view that complicates the taxonomy I offered, inasmuch as it opens up the possibility of fully achieving both reduction and fidelity. But I would point out that Leibniz is able to get this result by pulling his theory of mental representation apart from his theory of knowledge. It is at the level of mental representation that Leibniz gets

 2

Quaestiones de anima III.8, quoted in lect. 3, p. 58.

Robert Pasnau

91

reductivism and fidelity: the objects of perception are the microscopic mechanical features of things, in all their details, and in this sense, that is what we perceive. But at the level of epistemology “in this, [the mind’s] present state” (as quoted above on p. 41), such fine-grained representations do us no good. What the senses reveal to us is too “confused” (the technical term Leibniz uses in these passages) to deliver reality itself. As a result we “fashion some new thing for ourselves” (ibid.). That is to say, we project (to use more modern parlance) some reality out onto the world; to use Leibniz’s analogy, we project green where there is really only blue and yellow. So when it comes to the epistemic situation of the senses, unaided by Leibniz’s fancy theoretical apparatus, the senses fail dismally when it comes to fidelity. Nor can we use the Cartesian method of careful scrutiny and get down to the level where fidelity obtains. For because transparency fails, we are not capable of doing this through the senses themselves. Philosophical theory is required to reach such conclusions. Leibniz’s view, it seems to me, is something like Buridan’s, at least at the phenomenal level. For, as I discussed in my response to Martin Klein, Buridan similarly thinks that the more complex causal story about sensory information, distinguishing between the per se and the per accidens sensibles, is not available at the sensory level itself, because of the confusion of the sensory representation. We have to reconstruct the process intellectually, in light of philosophical inquiry into how the reductive story must go. For Buridan, as for Leibniz, the contents of sensation are not transparently available. Sebastian goes on to show how the example of Leibniz sheds light on the case of Locke. Locke, for me, is a paradigm of the relational view, when he famously argues that color in objects is nothing but a power to produce visual sensations in us. But, I say, he cannot develop such a view to its full potential, because he is unwilling to endorse a metaphysics of powers: he is too much of a mechanist for that. And medieval Aristotelianism offered no help to him here—if we can even imagine him being interested in it— because the medieval Aristotelians neglected the strand in Aristotle that pointed toward this sort of relational view of the sensible qualities. That’s my story. Sebastian highlights another aspect of Locke’s view, however—the part of the view that expresses doubts about how the power relates to the sensation. In particular, Sebastian wants to focus on the way in which Locke seems committed to the idea that it is purely a matter of divine decree that such-and-such mechanistic events in objects give rise to

92

Some Thoughts after After Certainty

such-and-such sensations within the perceiver. What difference does this make? The passage that Sebastian quotes on p. 44 above shows Locke’s own strategy for minimizing these difficulties. For he says there that sensations are “adequate” representations of what is out in the world. Here “adequate” pretty clearly just means reliable. It has been “fitted and ordained by God” that our sensible ideas will correspond to the things in the world. Our ideas will be, as he goes on to say, “marks of distinctions in things.” But this again will be reliability without fidelity. Sebastian’s point, however, is that Locke’s theory has drawbacks that go beyond mere infidelity. For it is not just that, at an epistemic level, we are incapable of using the senses to grasp what is going on within objects, but also that, from a physical point of view, the causal story that goes from object to perceiver is entirely arbitrary or random—in other words, it is inexplicable. To make this clearer: suppose that the perception of a determinate shade of red is caused by the rotation of particles at a certain frequency—2200 rotations per second, let’s say. As Locke apparently thinks of it, there cannot be an explanation even in principle for why just that frequency, 2200 rps, causes just that determinate visual experience. It could just as well be 2300 rps. Why is it 2200? Because God arbitrarily picked that number. The issues here are metaphysical more than epistemological, and so in a sense they go beyond the matters the book is concerned with. I agree with Sebastian, however, that it certainly is interesting. And it is interestingly contrasted with cases we have considered already. It is quite different from Buridan, for instance, who, as Martin Klein stresses, is willing to accept the orthodox causal explanation of his day. On Aristotelian orthodoxy, there is nothing like Locke’s mystery. Even better, consider Leibniz, who thinks not only that the causal explanation is intelligible in principle, but that we can actually say something in detail about how it works. For he thinks that, whatever the microlevel explanation on the object side is, that is what the senses represent, albeit at a subconscious level. So, as one might expect from Leibniz, we get the complete opposite of arbitrariness. Rather, as Sebastian’s earlier discussion highlighted, we get intelligibility and even fidelity of representation. Now, perhaps, Leibniz is simply moving the proverbial bump in the carpet, because he may now face a question about intelligibility at another level, namely: Why do microperceptions of a certain kind give rise to the specific macroperceptions that we are introspectively aware of? If Leibniz has no good way to answer that question, then ultimately he is in no better shape

Robert Pasnau

93

than Locke on that issue. But I will leave adjudication of that question for others who are better informed on this topic. Sebastian suggests, at the end of his discussion, that early modern relationalists, given their religious commitments, must think that God can be the cause of our sensations. In that case, sensible qualities—the color blue, for instance—might, at least on some occasions, be God, and so God would turn out to be blue. I think that this is more or less correct for those authors who suppose that perception does not have the material world as its object. Berkeley is the most famous case, but one might also mention Malebranche. To make this seem less strange, these authors distinguish God from God’s ideas (although if God is simple then it is not easy to know how to work out the theological details). Leaving these cases aside, I do not think standard views like Boyle’s and Locke’s face the same result, because they want to identify the colors with those powers that ordinarily and naturally give rise to color sensations. (Too much gin may leave one seeing pink flashing stars, but the gin is not pink.) Working this out in detail would require a well-developed ontology of powers, to make sure that the colors are identified with powers at the correct explanatory level. But part of the reason Boyle’s and Locke’s picture is so vulnerable to attack is that they do not have the metaphysical courage of their convictions. For as I have discussed in an earlier work,3 neither Boyle nor Locke is willing to accept the ontological profligacy of a powers ontology. This leaves their relationalism in a precarious position, vulnerable to mechanistic reduction on the one hand, and idealism on the other.

Reply to Sabine van Enckevort and Han Thomas Adriaenssen For some decades now, historians of medieval philosophy have recognized that cognitive theory—as opposed to the more familiar categories of epistemology and philosophy of mind—represents one of the chief achievements of medieval philosophy, running back to Augustine’s De Trinitate and all the way forward to the many rival accounts developed in the fourteenth century. Telling the story in detail has proved a challenge, however, because many of the most important texts have not been available even in reliable Latin editions, let alone in modern translations. Peter Auriol is one of the most challenging such cases, because his work is



3 See Robert Pasnau, Metaphysical Themes 1274–1671 (Oxford: Clarendon Press, 2011), ch. 23.

94

Some Thoughts after After Certainty

indisputably important to later medieval cognitive theory and yet his texts are voluminous, difficult, and still not available in critical editions. Auriol’s position is particularly interesting in the context of my book, because he seems on the face of things to provide an example of a medieval author who treats the objects of perception as mind-dependent entities, a view that I claim is virtually never found before the seventeenth century. On close inspection, however, Auriol turns out not to be a counterexample to my historical generalization. For although he does appeal to cases of illusion to postulate “objective” (or “apparent”) beings as the objects of perception, he does not take the familiar route of locating these apparent beings within the mind. Rather, they are out in the world, exactly where they appear to be. Ultimately, I take Sabine and Han Thomas’s remarks on Auriol to be consistent with my own. For they run through the various interpretive constraints on what Auriol might mean, and reach similar conclusions: Auriol takes apparent beings to be out in the world, as individual entities; and he takes them to be somehow the same as physical objects, yet they must be somehow distinct too. I criticize other interpretations of Auriol for failing to respect one or another of these constraints, but I do not arrive at a consistent and satisfying interpretation. If Sabine and Han Thomas differ from me in this regard, it is only because they are more explicit in their verdict that Auriol fails to arrive at a satisfactory account. From here, Sabine and Han Thomas turn to the seventeenth century. They accept, I think, my central idea that there is something special about that century that led philosophers to treat ideas as the immediate objects of perception. Building on that, they want to consider why this sort of picture led to such notorious and intense problems regarding skepticism. After all, as they argue (and here I entirely agree with them), even the most directly realist conception of perception is vulnerable to familiar worries about dreams, illusions, and divine deception. The suggestion they arrive at is that skepticism seemed like such a threat to the theory of ideas because of a tendency to suppose that ideas are not only the immediate objects of perception but also the only objects of perception. If we perceive only ideas, then, to be sure, worries about the veil of perception become quite serious. There is, however, no need for the theory of ideas to make this further claim, and indeed, “most early modern proponents of ideas were not committed to anything so extreme” (p. 56).

Robert Pasnau

95

Reflecting on this line of thought, I would have liked to see Sabine and Han Thomas put more weight on several other aspects of my story. I claim that there is a general principle in the philosophy of perception that influences the way these discussions develop, which I call the Fidelity Constraint: Where the senses reliably track the presence of two or more items along a single causal chain running from external object to perceiver, the item that is tracked with greater fidelity is that which is most properly the object of perception.4

This principle provides guidance on how to identify the object of perception in a case where the causal story runs through many intermediaries (as of course it always will for anyone who accepts the usual story about our minds and the world). Look to find the place in that causal chain whose character is most fully revealed by the content of the perception: this is what we perceive. Now, in principle, this constraint could admit of ties, and in such cases we might be inclined to say that we perceive both one thing and another, which might be distinguished only in terms of immediacy. One might thus want to say that we perceive both the television screen and the event we are watching on the screen. But this is definitely not the sort of situation that seventeenth-century philosophers took themselves to be facing. For they took the new science to have shown that colors and other secondary qualities, to the extent they are out in the world at all, are not at all like the contents of our perceptions. And, at the same time, they found it natural to think that our mind contains ideas that are exactly like what we seem to perceive. Thus, Locke writes that an idea “can be no other but such as the mind perceives it to be.”5 Accordingly, it made perfect sense for these authors to adopt a very extreme theory of ideas, on which ideas are, properly speaking, the only things we perceive. To be sure, careful discussions of perception would register that, in a broader sense, we perceive ordinary objects in the world. But the logic of the theory of ideas consistently pushed authors toward the more radical thought that, if we are speaking strictly, then we should say that the only things we perceive are ideas. No wonder the skeptical dangers of such a claim seemed so manifest.

 4

After Certainty, lect. 4, p. 90. John Locke, Essay 2.29.5 (An Essay Concerning Human Understanding, ed. P. H. Nidditch [Oxford: Clarendon Press, 1975], p. 364), quoted in lect. 4, p. 92.

5

96

Some Thoughts after After Certainty

Response to Stephan Schmid Stephan devotes the first part of his comments to considering my remark that Galileo and Descartes might be viewed as “the last great champions” of a tradition, running back to antiquity, of setting certainty as our epistemic goal (lect. 2, p. 30). No doubt he is right that this claim ought to have been more carefully qualified. I certainly did not mean to suggest that one cannot find later figures who put great weight on the demand for certainty. So ultimately I must concede much of the force of Stephan’s remarks. But let me draw a few further distinctions that may partly redeem the book’s claim. Let us distinguish a first camp of authors who think that certainty is an ideal of knowledge that can be achieved (like Galileo and Descartes), from a second camp of those who think that certainty is an ideal of knowledge that cannot be achieved (like Hume, say, or, in recent times, Peter Unger). These two camps share a conception of the ideal, and have a shared understanding about where we ought to apply the term “knowledge.” They can be contrasted with a third camp of authors who think that we should not treat certainty as a normative ideal of any sort—that is, that certainty is not something we should even aspire to, and (relatedly) that it is not part of what it is to know. This is now the mainstream view in epistemology, and I claim that it can be found in John Wilkins before the end of the seventeenth century. Now let us single out those in the first camp who are part of the tradition I describe that goes back to antiquity. Part of what I meant to claim in singling out Galileo and Descartes is that they are part of a continuous history of concern with the ideal. I think there is a fairly straightforward historical justification for this, inasmuch as their thinking is deeply informed by earlier scholastic traditions. What I meant to suggest was that, after Galileo and Descartes, it becomes harder to see that sort of continuity. Even authors who persist in seeking certainty are not really continuous with that longer historical story. So what of Spinoza, who is Stephan’s first counterexample? Here I am happy to say that Stephan is basically right—that we should count Spinoza, much like Descartes, as committed to the ideal of certainty in a way that is continuous with the older tradition. But, at the risk of splitting hairs, I might demur regarding whether this is a very serious objection to my claim, because I think that in this context Spinoza is simply another

Robert Pasnau

97

Cartesian. One could go through various later seventeenth-century Cartesians and doubtless find others committed to the Cartesian ideal of certainty, and again, I would not think of these figures as significant counterexamples. But this is by no means to deny the value of investigating what these later Cartesians have to say about certainty, and no doubt all sorts of important differences would emerge between them and their master. I am less willing to accept Stephan’s remarks about Leibniz. To be sure, Leibniz assigns a robust role in his philosophy to necessary truths as known through the Principle of Non-Contradiction, and he even has a story about how contingent propositions can, in principle, be known to be certain through a priori reasoning, where “a priori” has the older meaning of being grasped from the natures of things. But for me, Leibniz is an important part of the new, post-scholastic story of epistemology in virtue of the attention he gives to probability. Surely part of the reason he devoted such attention to Locke’s Essay is that he was impressed by Locke’s interest in probability. As he remarks there: I have said more than once that we need a new kind of logic, concerned with degrees of probability, since Aristotle in his Topics could not have been further from it: he was content to set out certain familiar rules […] without taking the trouble to provide us with the scale that is needed to weigh the appearances and to arrive at sound judgments regarding them.6

Although it was Locke who, among his contemporaries, was perhaps more famous for these views, it was Leibniz who got the attention of the cognoscenti. Hume, for instance, in the abstract to his Treatise, credits Leibniz, rather than Locke, with promoting the significance of probability in our epistemic lives. As usual, Leibniz shows himself to be a keen historian of philosophy, when he compares his own interest in probability with Aristotle’s Topics. For it would be here, if anywhere, that the ancient tradition offers a precedent for exploring epistemology in non-ideal conditions—that is, under conditions of uncertainty. But although the Topics was by no means

 6

Leibniz, New Essays 4.16 (Gottfried Wilhelm Leibniz, New Essays on Human Understanding, trans. P. Remnant and J. Bennett [Cambridge: Cambridge University Press, 1996], p. 466), quoted in After Certainty, lect. 2, n. 20, p. 212.

98

Some Thoughts after After Certainty

ignored by later Aristotelians,7 it is fair to say that this tradition was hardly a flourishing philosophical success. As Leibniz says, the Topics offered nothing more than piecemeal treatments of “familiar rules,” and so did not provide a path to a systematic treatment of reasoning outside the mainstream demonstrative framework of the Posterior Analytics. Stephan goes on to say that my neglect of Spinoza and Leibniz is symptomatic of my neglect of the whole German tradition of philosophy that runs through Kant and German idealism. To be sure, I do neglect that whole tradition, and I wish that I had been able to take it into account. If only this book had been as long as my last one! What held me back, I must confess, was not so much length as the limits of my own expertise. Still, it seems to me doubtful whether the later Kantian tradition is a good example of those who fall into the first camp above—namely, those who hold that certainty is an ideal that we can achieve. Or rather, one might say that, yes, they do hold onto that ideal, but only by taking refuge in one form or another of idealism. So perhaps the thing for me to say is not that the ideal of certainty disappears in these later figures, but that we have entered into such radically different terrain here—having completely given up anything like the abiding realism of the Aristotelians and Cartesians— that we can no longer speak of the continuous tradition on which I focused. Let me now turn to Stephan’s second main line of discussion, which concerns Lecture Six. There I articulate the form of skepticism that I find most plausible, which I call epistemic defeat, and which I define as “the view that, in the final analysis, we have no good evidence for the truth of any proposition” (lect. 6, p. 128). On my account, epistemic defeat should not be linked to the question of whether we have knowledge. This is all part of my resistance to lexicology. Let’s not worry about the word, I say, because the word will take care of itself, and once we figure out what ideal we can achieve, the words we use will adjust themselves accordingly. The point of an idealized epistemology is to worry first about the ideal, rather than to start with linguistic analysis.

 7

For a good sense of what this later tradition looks like, one can consult Eleonore Stump’s two book-length translations of Boethius’s studies of the Topics: De topicis differentiis, trans. Eleonore Stump (Ithaca: Cornell University Press, 1978), and In Ciceronis Topica, trans. Eleonore Stump (Ithaca: Cornell University Press, 1988).

Robert Pasnau

99

But Stephan thinks I am too quick to dismiss what he calls “the (traditional) philosophical position that real knowledge does indeed require certainty” (p. 64). Here I want to push back much harder than I did against his previous remarks. For I think it’s simply a historical mistake to talk this way. He speaks here of “real knowledge”; but what does that even mean? I cannot readily recall a historical text in any language that uses a version of that phrase. Now, it is common enough to talk about knowledge in the strict sense (epistƝmƝ haplǀs in Greek, or scientia proprie dicta in Latin), and it is part of my point to stress that knowledge, so conceived, requires certainty. But that is just a description of the ideal; and that, to my mind, is very different from supposing that there is some thing called knowledge that has some kind of real definition that requires that it be certain. Rather, what one sees over and over is a willingness to allow that “knowledge” is said in many ways. And it is very hard to find texts—in English, Greek, Latin, or Arabic—insisting that knowledge in these looser senses of the term is not really knowledge. It is that way of thinking that breeds our modern preoccupation with skepticism. But it is because this is such a wholly uncompelling way of thinking about the issues that there is basically no one today who endorses skepticism. Stephan goes on to draw a comparison with the way we think of freedom, and I too find this helpful, because it illustrates so nicely the way I see these issues. One certainly can speak of an ideal sort of freedom, and try to sketch what that would be like. But, notoriously, that ideal of freedom is conceptually problematic for all sorts of reasons, and so a lot of people, once they see the difficulties involved, retreat to some lesser form of freedom, such as compatibilism. At this point the libertarian—the proponent of some more ideal form of freedom—might point to the word “freedom” and complain that What you’ve given us is not real freedom! But again, I think this debate is just pointless. What is at issue is not the word, but rather what sort of control of our actions is possible to us in the real world. We figure that out, and then we use the word accordingly. Setting aside the case of freedom, and returning to the case of knowledge, Stephan wants to push the idea that if we decide epistemic defeatism is correct then the proper conclusion to reach is that we were wrong about knowledge all this time. In his words, we should “settle for an error theory of knowledge” (p. 65). But, again, I would resist this characterization of the situation. Our usual application of the concept of knowledge is extremely variable and flexible. This point is not very clearly recognized in many classic texts from the seventeenth and eighteenth centuries, which

100

Some Thoughts after After Certainty

often give the impression that if we cannot have perfect certainty then we do not have knowledge at all. But scholastic authors do a much better job here, since they flag from the start that knowledge is said in many ways. At the low end of the scale, we speak of knowledge where we have nothing more than a firmly held true belief (no justification required). Is that not real knowledge? I do not think this is a helpful question to ask. Let’s stop worrying about the word, and focus instead on the facts on the ground about what we can actually achieve. Stephan next asks whether epistemic defeatism is, as I say, “so neglected as not even to have a name” (lect. 6, p. 128). On my view, it is very far from a novel view; rather, it is precisely the heart of ancient skepticism. But my point is that the word “skepticism” gets attached to so many other doctrines that the issue of epistemic defeat tends to get neglected. So I suggest we give it a special name. According to Stephan, it already has a familiar name, and that name is fallibilism. But I think the term “fallibilism” is better reserved for the view that we cannot have perfect certainty. And that’s different. It is one thing to say that we have pretty strong reasons for our beliefs, but not perfectly certain reasons—that is the sort of picture you get from the skeptical stage of Descartes’s inquiry, where he wants to persuade us, provisionally, that there is enough doubt for us to abandon our beliefs. The fallibilist is someone who might simply accept that possibility of doubt and shrug, saying that one can believe, and even have knowledge, without ruling out such doubts. The epistemic defeatist goes much, much farther, and says that we do not even have “pretty strong” evidence, but that, to the contrary, there is ultimately no stronger evidence on one side or the other. Stephan also sees epistemic defeatism in the sort of coherentist, antifoundationalism that one finds in Wilfrid Sellars. But this too strikes me as a different sort of view. One way to argue for epistemic defeatism is to argue against foundationalism—that is, to argue that there are no basic principles that are somehow intrinsically justified and through which everything else is justified. Sellars wants to reject such basic principles and instead embrace a holism on which everything is in principle revisable. But he also wants to embrace a coherentism on which the theory as a whole takes on a justified status, and has some level of fallible evidential support from its internal coherence. The epistemic defeatist is going to claim that coherence does not provide any such justification. So again it is a different view.

Robert Pasnau

101

Stephan’s final concern is that the view I ultimately take in Lecture Six sounds much like Hume. I don’t entirely mind this complaint; it is right enough that my way of thinking about these issues is deeply influenced by Hume’s way of proceeding. But in the end I don’t think Hume offers much help when it comes to figuring out where to go in the face of epistemic defeat. Here’s a representative example from the Treatise: I know not what ought to be done in the present case. I can only observe what is commonly done; which is, that this difficulty is seldom or never thought of; and even where it has once been present to the mind, is quickly forgot.8

As I understand this, it turns in part on the issues I discuss in Lecture Five, where I discuss the historical tendency for each of us to give special epistemic weight to how things seem to ourselves as individuals right now. I call this the “privileged now.” Ironically, Hume is as much dependent on the privileged now as Descartes is. But whereas Descartes builds his whole theory around a way of securing, over time, the certainty that we can properly have only within the privileged now, Hume builds his theory around his confidence that, outside the privileged now, his skeptical arguments will lose their force. His view is ultimately quietist and antitheoretical. He doesn’t think there is a way within philosophy to respond to skepticism. (Admittedly, this is a controversial reading of Hume.) My own response to epistemic defeat is not quietist, but instead invokes the theoretical resources of philosophy. Like Hume, it is not a response from within the domain of reason, offering evidence to hold off epistemic defeat. Rather, it is a response from the affective side, appealing to the optimistic attitude of hope as a way of engendering belief even in the absence of evidence. Appealing to the affective side certainly sounds very Humean, so again, I think Stephan is not wrong to see my position as having real affinities with Hume. But I do not think you can find in Hume anything that is developed in the way I develop this position.

Response to Christophe Grellard Christophe begins his comments with a plea for taking seriously the importance of skepticism in the Middle Ages. I don’t exactly disagree, but

 8

Hume, Treatise, 1.4.7 (David Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge, 2nd ed., rev. P. H. Nidditch [Oxford: Clarendon Press, 1978], p. 268), quoted in After Certainty, lect. 6, p. 130.

102

Some Thoughts after After Certainty

it seems to me scholars could do a better job distinguishing between various views that get classified as skeptical. Lecture Six argues that skepticism, as we now conceive of it, contains three different threads: first, the denial of knowledge; second, the suspension of belief; and third, the repudiation of evidence. The last of these, in its most radical form, is the view I call epistemic defeatism. If this sort of repudiation of evidence is sufficient for skepticism, then one finds ample instances of skepticism, beginning in the sixteenth century, and it would not be unreasonable to see this as a hallmark of “modern” thought. But it is harder to find this sort of view among medieval authors. Nicholas of Autrecourt is perhaps the best example, but in a way he is not a good example at all. In his famous letters, his skeptical claims are entirely ad hominem. It is Bernard of Arezzo who is, allegedly, committed to these skeptical consequences. And the force of the letters as an attack on Bernard depends on the idea that of course these consequences are absurd. Moreover, in Autrecourt’s principal philosophical work, his Exigit ordo, or Tractatus as I prefer to call it, he concerns himself at great length with establishing an alternative epistemology that avoids these skeptical results. So, as Christophe himself has written about elsewhere, Nicholas is not a skeptic in any straightforward sense. But Christophe’s point is not so much that anyone is persuaded by skepticism. It is rather that certain kinds of skepticism—in particular the deceiving-God hypothesis—are deeply important to the progress of philosophy. But the idea is quite old—it is found, for instance, in Cicero’s Academica—and I myself find it hard to imagine an era when anyone worthy of being called a philosopher would have been unable to arrive at a skeptical thought in effectively the same way. Yet although the core ideas that give rise to skepticism are far from subtle, it is hard to find medieval authors who are self-professed skeptics. Even when one breaks the notion of skepticism down into the three elements I suggest above, it is very unusual to find medieval authors defending even one of those skeptical theses. And even when we conceive of knowledge in the lofty demonstrative sense of the Posterior Analytics, it is hard to find a scholastic author who denies that we can, at least sometimes, have such scientia.

Robert Pasnau

103

Let me turn to the second principal focus of Christophe’s remarks, which concerns my suggestion that the way forward, in light of the threat of epistemic defeatism, is to appeal to hope. Christophe grants, at least for the purposes of this discussion, the skeptical point of view that inspires epistemic defeatism. But he thinks the proper approach is to consider the social context of belief—that is, the way we believe not as isolated individuals but as members of a group. In effect, he wants to move toward a social epistemology. This seems to me a really interesting suggestion, and I want to embrace it. The prominent role that Augustine gives to testimony makes it natural to think that belief, in the later Middle Ages, will have a strongly social aspect. Moreover, Christophe has identified several very promising scholastic sources for this approach, in Robert Holcot and in William of Ockham’s political writings. We might contrast this situation with the seventeenth century, which tilts much more heavily toward the cult of the isolated, iconoclastic genius at war with the broader community: consider Descartes, Hobbes, or Spinoza. I do not think, however, that Christophe’s suggestion requires me to retreat from any of the claims that I make in the book. For even if we agree that there is a social aspect to belief, the question still remains: How will we fundamentally be able to reach a state of belief in light of the reality of epistemic defeat? The basic problem is that evidentialism in general, and the principle of proportionality in particular, look extremely difficult to reject. It is not even clear that we can take the attitude of believing a proposition while recognizing that the evidence is insufficient. It seems to me that Christophe’s appeal to social epistemology is neutral between various different ways of explaining the origins of our beliefs. One way in which the larger role of society might get involved is to persuade us to reject epistemic defeat. We might be persuaded by the beliefs of our culture that there must really be good evidence for all the things that the people around us believe. This is perhaps what ordinarily happens, but it is hardly an epistemically ideal situation, or anything close to it. So what else? The strategy of Lecture Six is to contrast three options that remain, if epistemic defeatism is true. The first option is to embrace Hume’s quietism, which is to say that we should simply stop thinking about the issues, and let the influence of time do its work. The second, historically very prominent option is to believe through faith, which, as I understand it, is flatly to reject evidentialism, and let one’s confidence run

104

Some Thoughts after After Certainty

free of the evidence. The final option is to believe through hope, which is to let one’s affective optimism cause one to stop worrying about being wrong, and so to believe even though one’s objective credences are low. I do not mean to suggest that these are the only options, but of the three, it seems to me that believing on hope is epistemically the most honest. It allows us to believe, which is essential; but even so, it minimizes the degree to which we have to disregard the badness of our epistemic state. It requires us to make adjustments on the affective side rather than on the cognitive side. So although I am eager to embrace Christophe’s suggestions about the social dimension of belief, I am inclined to think that it can play that role within the context of hope.



APPENDIX

Volume 17, 2020 The Proceedings of the Society for Medieval Logic and Metaphysics (P.S.M.L.M.) is the publication of the Society for Medieval Logic and Metaphysics, collecting original materials presented at sessions sponsored by the Society. Publication in the Proceedings constitutes prepublication, leaving the authors’ right to publish (a possibly modified version of) their materials elsewhere unaffected. The Society for Medieval Logic and Metaphysics (S.M.L.M.) is a network of scholars founded with the aim of fostering collaboration and research based on the recognition that ™ recovering the profound metaphysical insights of medieval thinkers for our own philosophical thought is highly desirable, and, despite the vast conceptual changes in the intervening period, is still possible; but ™ this recovery is only possible if we carefully reflect on the logical framework in which those insights were articulated, given the paradigmatic differences between medieval and modern logical theories. The Society’s web site (http://faculty.fordham.edu/klima/SMLM/) is designed to serve the purpose of keeping each other up-to-date on our current projects, sharing recent results, discussing scholarly questions, and organizing meetings. If you are interested in joining, please contact Gyula Klima (Philosophy, Fordham University) by e-mail at: [email protected] © Society for Medieval Logic and Metaphysics, 2020



CONTRIBUTORS

Martin Klein, Universität Würzburg Robert Pasnau, University of Colorado, Boulder Sebastian Bender, Humboldt-Universität zu Berlin Sabine van Enckevort, Rijksuniversiteit Groningen Han Thomas Adriaenssen, Rijksuniversiteit Groningen Stephan Schmid, Universität Hamburg Christophe Grellard, École Pratique des Hautes Études, Paris