Possibilities of Perception 9780199678440, 0199678448

Table of contents :
Cover
Contents
Acknowledgments
List of Figures
Introduction
Chapter 1 Perception and the Experience of Objectivity
1. Perceiving and Knowing
2. Immediacy and its Problems
3. Self-evidence and the Experience of Objectivity
4. A Sensory Requirement?
Chapter 2 The Role of the Imagination
1. The Need for Active Imagining
2. The Evidence for Imagining
3. Imagining and Imagery
4. The Validation of Imagination
5. Conclusion
Chapter 3 Perceiving Reasons
1. Reasons and Explanatory Relations
2. Perceiving Causes
3. Perceiving Constituents as Reasons
4. Perceiving Justificatory Reasons
5. The “Space of Reasons”
6. The Advantages of Perceiving Reasons
7. Conclusion
Chapter 4 The Further Reaches of Perception
1. Perceiving Remote States of Affairs
2. Perceiving Future States of Affairs
3. Perceiving Abstract States of Affairs
4. Rational Perception—Lessons from Plato and Descartes
5. Conclusion
Chapter 5 Moral Perception
1. Standard Commitments of Perceptual Theories
2. Perceiving Persons as Persons
3. Uncanniness and Ambivalence
4. Perceiving What to Do
5. Back to Kant
6. Conclusion
Chapter 6 Aesthetic Perception
1. The Objects of Aesthetic Perception
2. The Process of Aesthetic Perception
3. The Double Consciousness of Aesthetic Perception
4. Aestheticizing Morality and Moralizing Aesthetics
Conclusion
References
Index
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
R
S
T
U
W
Y

Citation preview

Possibilities of Perception

This page intentionally left blank

Possibilities of Perception Jennifer Church

1

3

Great Clarendon Street, Oxford, ox2 6dp, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries # Jennifer Church 2013 The moral rights of the author have been asserted First Edition published in 2013 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer British Library Cataloguing in Publication Data Data available ISBN 978–0–19–967844–0 Printed and bound in Great Britain by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Contents Acknowledgments List of Figures Introduction Chapter 1 Perception and the Experience of Objectivity 1. 2. 3. 4.

Perceiving and Knowing Immediacy and its Problems Self-evidence and the Experience of Objectivity A Sensory Requirement?

Chapter 2 The Role of the Imagination 1. 2. 3. 4. 5.

The Need for Active Imagining The Evidence for Imagining Imagining and Imagery The Validation of Imagination Conclusion

Chapter 3 Perceiving Reasons 1. 2. 3. 4. 5. 6. 7.

Reasons and Explanatory Relations Perceiving Causes Perceiving Constituents as Reasons Perceiving Justificatory Reasons The “Space of Reasons” The Advantages of Perceiving Reasons Conclusion

Chapter 4 The Further Reaches of Perception 1. 2. 3. 4. 5.

Perceiving Remote States of Affairs Perceiving Future States of Affairs Perceiving Abstract States of Affairs Rational Perception—Lessons from Plato and Descartes Conclusion

Chapter 5 Moral Perception 1. Standard Commitments of Perceptual Theories 2. Perceiving Persons as Persons

vii ix 1 11 12 18 25 38 47 48 58 63 78 83 87 88 90 101 106 112 123 136 139 140 147 151 163 185 187 189 197

vi

CONTENTS

3. 4. 5. 6.

Uncanniness and Ambivalence Perceiving What to Do Back to Kant Conclusion

Chapter 6 Aesthetic Perception 1. 2. 3. 4.

The Objects of Aesthetic Perception The Process of Aesthetic Perception The Double Consciousness of Aesthetic Perception Aestheticizing Morality and Moralizing Aesthetics

204 209 216 224 227 229 240 246 258

Conclusion

265

References Index

267 281

Acknowledgments I am grateful to the American Council of Learned Societies for the grant that launched me on this book project, and to Vassar College for the research leave that enabled me to bring the project to completion. Over the past several years I have presented my work on the topic of perception in many different forums—conferences, classes, colloquia, workshops—receiving valuable criticism and encouragement from numerous people. Few have understood how it is all supposed to hold together (that is why it needs to be brought together in a book), but many have offered useful comments and suggestions. Among the voices that have been especially influential in my thinking about different sorts of perception are those of Ned Block, Paul Boghossian, John Campbell, Quassim Cassam, Martin Davies, Christel Fricke, Jesse Kalin, Amy Kind, John McDowell, Mitchell Miller, Uma Narayan, Amelie Rorty, David Rosenthal, Jeff Seid´ sta Sveinsdo¨ttir, Doug Winblad, and two anonymous readers for man, A Oxford University Press. David Velleman’s steadfast support and clear-headed advice kept me on track at several crucial junctures. Daniel Friedman’s unfailing curiosity and generosity played an essential role in sustaining my own enthusiasm. Grant Magruder did excellent work as my research assistant. Peter Momtchiloff encouraged me with his early interest in the project and his calm persistence through the many steps leading to publication. An earlier version of some of the arguments of this book appeared in an article entitled “Seeing Reasons,” published in Philosophy and Phenomenological Research in 2010. Much of Section 5 of Chapter 3 began as a conference contribution that appeared under the title “Locating the ‘Space’ of Reasons,” in Teorema in 2006. The Frances Lehman Loeb Art Center, Vassar College, Poughkeepsie, New York, generously allowed me to reproduce images of three items from their collection: Arshile Gorky’s The Horns of the Landscape, 1944 (Oil on canvas, 30  34”, gift from the collection of Katherine Sanford Deutsch, class of 1940, 2000.6); Georgia O’Keeffe’s Two Figs, 1923 (Oil on board, 73/4  53/4”, bequest of Mrs Arthur Schwab (Edna Bryner, class of

viii

ACKNOWLEDGMENTS

1907), 1967.31.12); Sanford Robinson Gifford’s Tappan Zee, 1880 (Oil on canvas, 11 3/8  18 3/8”, bequest of Robert Miller Walker and Alice Smith Walker, class of 1933, 1994.18.1). The Arshile Gorky Foundation gave permission to use the image on this book’s cover: Gorky’s The Horns of the Landscape # 2012 The Arshile Gorky Foundation / The Artists Rights Society (ARS), New York. The Georgia O’Keeffe Museum gave permission to use an image in Chapter 6: O’Keeffe’s Two Figs # 2012 Georgia O’Keeffe Museum / Artists Rights Society (ARS), New York.

List of Figures “Fern Shadows” (photo by Daniel Friedman) “Rensselaerville Barn” (photo by Daniel Friedman) “Storm on Breakwater” (photo by Amy Church) “TNT Crystallizing” (photo by Daniel Friedman) “Sled Ride Doorways” (photo by Daniel Friedman) Georgia O’Keeffe, Two Figs, (# 2012 Georgia O’Keeffe Museum/ Artists Rights Society (ARS), New York) 6.2 Sanford Robinson Gifford, Tappan Zee 1.1 2.1 3.1 4.1 5.1 6.1

10 46 86 138 186 226 233

This page intentionally left blank

Introduction We speak of seeing, or hearing, or feeling—more generally, of perceiving— many different things: the flight of a crow, the brightness of a star, the reporter on the evening news, the humor in a string quartet, the meaning behind a mother’s words, the cause of an accident, the inevitable outcome of a dare, the reason to leave a meeting, the personality of a child, the mood of a painting. But which, if any, of these things are truly perceived is a topic of ongoing philosophical dispute. At one extreme, it has been claimed that the only things that are literally perceived are spatial arrays of colors and textures, and temporal arrays of sound—not clouds or barns, words or melodies. At the other extreme, it has been claimed that psychics can see the future, that mathematicians can perceive the truth of an unproven mathematical theorem, and that enlightened souls can perceive the form of the Good. Even with respect to ordinary objects such as birds and trees, philosophers continue to disagree about whether we can perceive a goldfish or an oak as such, and whether we perceive an entire bird or an entire tree despite the fact that only one side faces us. Why does it matter what counts as perception? Why insist on drawing the line in one place rather than another? Why not accept, indeed revel in, the variety of ways in which we can acquire knowledge of the world— sometimes through our eyes, sometimes through our ears, but sometimes also through our hearts and our minds? The answer to these questions has two sides. The first has to do with knowledge and its justification. Perceptual knowledge is foundational knowledge insofar as it can justify further knowledge without itself needing to be justified. Just what this means, and how it can be true, require further elaboration (which is one of the things this book attempts to do), but the epistemic significance of this claim should be clear: when we perceive something its reality is self-evident in a way that the reality of things we

2

INTRODUCTION

merely believe is not. The other side of the answer to the question of why we should care about what counts as perception has to do with the phenomenology of our perceptual experiences. Perceiving a crow or a melody or the face of a friend has a richness and an immediacy that merely thinking of a crow or a melody or the face of a friend lacks. This too requires further elaboration (which is also something this book aims to provide), but it promises to tell us something important about the structure of our minds and the structure of our values. Many accounts of perception address just one side of the question—the epistemic importance of perception, or the phenomenological importance of perception—but not both. Epistemic accounts that view perceptual states as especially reliable indicators of objective states of affairs (having been evolutionarily designed for such a purpose, for example) have trouble connecting their reliability to their subjective character, thereby losing sight of the special phenomenology of perception. On the other hand, phenomenological accounts of perception that focus on its special phenomenological character or content (sensual content, or non-conceptual content, for example) have trouble connecting that character or content to the contents of an objective world, thereby losing sight of the special epistemic role of perception. Still others bridge the epistemic-phenomenological divide only insofar as they suppose that the objects of perceptual knowledge are constituted, at least in part, by our experience of them (so-called secondary properties, or response-dependent properties). This book defends an account of perception that brings the two sides into closer harmony, showing how a distinctive sort of phenomenology (the phenomenology of objectivity, which is closely related to the phenomenology of spatiality) serves a distinctive epistemic role (the role of selfevidence or self-justification). Offered first as an account of what it takes to perceive ordinary objects such as birds and trees, it is then extended to show how it is also possible to perceive such things as causal relations, justificatory relations, distant states of affairs, and abstract entities. In each of these cases, I argue, our perception can be literal (not merely figurative or metaphorical) and substantive (not merely formal or deflationary). The account also helps to explain the advantages of perceptual versus nonperceptual knowledge, and helps to make sense of some key texts in the history of philosophy (Plato’s cave, Descartes’ defense of intuition, and Kant’s Categorical Imperative).

INTRODUCTION

3

The overall argument of the book can be understood as running in two directions. On the one hand, starting with an account of what ordinary perception requires, the book proceeds to show how these requirements can be met in the more puzzling and controversial cases of rational perception, moral perception, and aesthetic perception. To the extent that the proffered account is compelling, the possibility of these further sorts of perception is supported. On the other hand, starting with the assumption that rational perception, moral perception, and aesthetic perception are phenomenologically real and epistemically relevant phenomena, the book provides an account of perception that encompasses these different types of perception, showing what they have in common with more ordinary cases. To the extent that one already finds the possibility of rational perception, moral perception, and aesthetic perception compelling, the presented account of perception is supported. (There may be other accounts of what these different cases have in common, of course. The account I offer is supported only insofar as it is the best account of what these different sorts of perception have in common.) I have long shared the Kantian conviction that empiricist and rationalist understandings of perception are both problematic, and that each shortchanges the insights of the other. The first draft of this book began with a discussion of several historical views of perception—not to show what was problematic with each so much as to show what remains constant across these contrasting accounts of perception. Those discussions have been shortened, and have been moved to later parts of the text (mainly Section 4 of Chapter 3, plus references to Kant throughout). They remain indicative, however, of the two-directional intention of my argument: from an account of perceiving nearby physical objects to an account of perceiving reasons, and from an account of perceiving reasons to an account of perceiving nearby physical objects. What links the epistemology of perception and the phenomenology of perception is the possibility of experiencing things as objective—the possibility of experiencing an object’s independence from our experience from within a given experience. Paradoxical as this may sound (for how can x’s independence from y be evident from a consideration of y alone?), it is precisely what needs explicating in order to understand what is so distinctive, and what is so important, about perception as such. Experiencing something as objective is no guarantee that it does in fact exist objectively.

4

INTRODUCTION

There can be illusions of objectivity. We often turn to others to confirm or disconfirm the accuracy of our experiences, and we frequently favor the pronouncements of theories that provide us with a coherent worldview over the more immediate verdicts of experience. The distinctive epistemic role of perceptions is not to provide us with certainty about any particular fact; it is, rather, to enable us to experience the world as an objective world, a world that persists independently of our particular point of view at a particular time, a world that we can either succeed or fail to know. The phenomenology of perception, then, does not secure any particular piece of knowledge so much as it secures the possibility of reference and thus the possibility of knowing anything at all. How is it possible, though, to recognize the objectivity of something from within one’s experience? How is it possible to experience something as existing independently of any given experience without oneself being able to step outside that experience? The answer to this crucial question, I argue, turns on the possibility of combining multiple avenues of access within a single experience. These different avenues are sometimes sensual and sometimes not. When our two hands converge on the same object, we can experience its presence in objective space, and when we access an object’s edges with eyes as well as hands, we can experience the objectivity of its shape. Similarly, when we find several different methods of adding two numbers converging on the same sum, we can experience the objectivity of a mathematical truth; and when a lie appears unacceptable from many different points of view, we can experience the objectivity of a moral truth. Although it is sometimes possible to occupy different perspectives or pursue different avenues of access at the same time (as when we see the top of a cup and feel its bottom), and it is sometimes possible to recall some relevant alternatives in memory (as when we remember the feel of an object we are looking at), most experiences of objectivity draw heavily on the projections of our imagination (what Kant would call “productive” as opposed to “reproductive” imagination). In order to experience something as real, we must imagine the possible. To experience the objectivity of a distant barn, we must imagine what it looks like from other points of view; we must use our imaginations to fill in information that is not currently present to our senses. Likewise, to experience the objectivity of a causal relation, we must imagine sequences that we don’t actually observe; and to experience the objectivity of an obligation, we must imagine points of view

INTRODUCTION

5

that we don’t actually occupy. Actively imagining and synthesizing such alternatives is what enables us to experience an object, or a fact, as having an independent existence; and to the extent that our imagining is successful at tracking the truth, the resulting experience will be a successful case of perception. Some of the above claims will seem more natural than others, and different readers will come to this book with different sympathies. (The initial sympathies of philosophers, especially philosophers who have developed their own views on a topic, tend to be especially idiosyncratic and stubborn.) While I hope that the detailed arguments of each chapter, taken together, will prove convincing, I expect many readers to focus on sections of particular interest to them. It is important, therefore, to be clear about the argumentative structure of this book as a whole—what depends on what, and what can stand alone. Chapter 1 considers our perception of nearby midsized objects and defends a distinctive account of what is required for perceptual knowledge of such objects—an account that emphasizes the experience of objectivity and de-emphasizes the importance of sensory content. This chapter should interest those who are intrigued by different understandings of epistemic “immediacy” and by the very idea of truths that are “self-evident.” It should also give pause to those who emphasize the sensory aspect of perception. Chapter 2 details the role of the imagination in perceptual experience. Experiencing something as objective depends on imagining alternative perspectives on, or alternative ways of accessing, that thing. The relevant sort of imagining must be spatial but it needn’t be visual, and it must be active but it needn’t be (and in fact often can’t be) deliberate. I offer a transcendental argument for these claims, but also review some empirical evidence that makes the claims more plausible than they initially appear. Quite apart from its bearing on the topic of perception, this chapter may be of value to those who are interested in the nature of imagining. Chapters 3 and 4 extend the account of perception defended in Chapters 1 and 2 to a series of more controversial cases—to the perception of causal relations between objects, the perception of dependency relations of various sorts, and the perception of distant and abstract states of affairs. A skeptical reader may be convinced of some of these extensions but not others; or she may accept them as legitimate implications of the basic account and, given

6

INTRODUCTION

their implausibility, conclude that there must be a flaw in account defended in Chapters 1 and 2. Alternatively, a reader who is already sympathetic to the possibility of perceiving such things as causes, reasons, distant galaxies, or mathematical relations might regard these implications as counting in favor of my account. In either case, this chapter can be read as elucidating the insights of some traditional defenders of “rational” perception. (Some key texts of Plato and Descartes are discussed in the final section of Chapter 4.) Chapter 5 addresses the possibility of moral perception. Accounts of moral perception tend to be offered by ethicists without much concern for how moral perception is or is not like the perception of nearby midsized objects—a tendency that fuels the suspicion that moral “perception” isn’t really perception after all. After clarifying what is at stake in the domain of ethics, I show how my basic account of perception can be applied to the case of moral perception, explaining the possibility of perceiving moral “objects” (i.e. persons) and moral facts (about what we ought to do). To the extent that our understanding of moral perception can be brought into alignment with our understanding of perceiving tables and chairs, both sides can benefit. Those who are persuaded by my account of perceiving nearby midsize objects might find themselves more sympathetic to the possibility of moral perception, while those who are already convinced of the possibility of moral perception might find themselves more attracted to my account of perceiving physical objects. In addition to its defense of the possibility of moral perception, Chapter 5 offers a reading of Kant’s Categorical Imperative that places it on the side of moral perception, and it makes some important claims about what must remain hidden when we perceive persons as persons. (These secondary claims may be of interest, and may be accepted, independently of one’s interest in, or acceptance of, the possibility of moral perception.) Chapter 6 draws on the advocated account of perception to help make sense of the “double objects” and the “double attitudes” of aesthetic perception—the physical object and what it reveals, our simultaneous engagement and “distance” from what is presented. This chapter attempts to show what aesthetic perception has in common with moral perception and with “depth perception” more generally. By emphasizing what aesthetic perception and moral perception have in common, however, it runs certain risks—the risk of moralizing aesthetics and the risk of aestheticizing

INTRODUCTION

7

morality. That is the topic of a final section that can also be appreciated independently of whether or not one is convinced of the book’s overall account of perception. Perception of things such as causes and reasons, persons and artworks, is not only possible; it matters. But why? I have already said something about why perception in general is necessary for knowledge: perception enables us to experience the world as an objective world, something we can be right or wrong about. But why should it matter whether our knowledge of any particular thing is perceptual? Once we are assured of the objective existence of causes or persons, for example, why should it matter whether we perceive a snake as the cause of a person’s anxiety—as opposed to figuring this out in some other way? Philosophers and mathematicians, psychologists and artists, have often been at pains to teach us how we can learn to perceive more— how we can extricate ourselves from the cave, how we can move past our reliance on formulas and achieve true insight, how we can enhance our moral perception. But if we allow that perception is just one way of knowing, and an imperfect one at that, it is appropriate to ask why we should care whether something is perceived rather than known in some other way. My answer to this question (which is fleshed out in Section 6 of Chapter 3, and in Sections 2 and 4 of Chapter 5) has several parts. First, because perception requires us to imagine alternative points of view simultaneously, it has the epistemic advantage of making inconsistencies more apparent (which, under most conditions, will increase one’s knowledge). Second, because perception depends on the spatialization of alternative perspectives and implications, it helps to expose gaps in one’s knowledge and suggests ways of filling those gaps (thereby facilitating new discoveries). Third, because the active imagining that supports perception also activates knowledge that is necessary for action, perception triggers action more immediately and automatically than other forms of knowledge (ensuring that it has a kind of practical effectiveness that it might otherwise lack). Finally, in the case of ethical knowledge, the imagining that goes into interpersonal perceptions represents a level of engagement and concern that is itself a sign of moral recognition (which is something we value independently of any particular outcome such recognition may have). As we shall see, each of these advantages is correlated with special risks—risks

8

INTRODUCTION

of overconfidence, risks of impulsiveness, and risks of added cruelty, for example. But those risks are seldom enough to outweigh the advantages that perception bestows. Offering an accurate account of the nature and the possibilities of perception, which is what this book aims to do, is thus valuable not only for its potential to unify a wide range of otherwise disparate claims but also for its ability to encourage and facilitate the extension of our perceptual capacities into some of the areas where it matters most.

This page intentionally left blank

Figure 1.1. “Fern Shadows” (photo by Daniel Friedman)

Chapter 1 Perception and the Experience of Objectivity This chapter introduces and defends a distinctive understanding of perception—an understanding that places the experience of objectivity at the center of both the phenomenology and the epistemology of perception. To start with, the focus will be on rather mundane instances of perception, such as seeing a crow or hearing a buzzing sound; subsequent chapters then extend the conclusions of this chapter to more controversial cases of perception. Although much of this book will be devoted to a discussion of just what sorts of things we can perceive, they all fall within the broad category of states of affairs. While we often speak of perceiving objects alone (a bird, a bush, a cloud, a star) or properties alone (smoothness, roughness, darkness, brightness)—depending on the focus of our attention or the purpose of our reporting—it does not seem possible to perceive objects without perceiving at least some of their properties, or to perceive properties without perceiving at least some instantiation of those properties. When we suppose that Chloe¨ saw a crow, for example, we suppose that Chloe¨ saw at least some of its properties; and when we note that Aaron heard a buzzing sound, we imply that he heard a particular instance of a buzzing sound.1 So even where the focus is on objects or properties alone, I will be assuming that perception

1 The assumption that seeing an object requires one to see at least some of its properties may seem to overlook empirical findings that indicate a separation between our capacity to recognize where something is and our capacity to recognize what something is (Milner and Goodale (2006)); but even when we are unable to identify a perceived object as a bird or as a bush, we still note its roundness, or smallness, or brownness, or fuzziness—i.e. we still register what we see as falling within some category or other. Likewise, even if we are unable to assign a specific location to a perceived sound, we still hear it as far away rather than nearby, or all around rather than off to one side. (For further discussion of perception’s dependence on space, see Section 3 below.)

12

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

is always perception of a state of affairs, where states of affairs are particular instantiations of properties. (This includes relational properties, which will be the focus of future chapters.)

1. Perceiving and Knowing We want to understand what is special about perceiving a state of affairs—as opposed to thinking about it, or receiving reports about it, or inferring its existence. For this inquiry, we do not need to take a stand on whether perception constitutes a special sort of knowledge or merely contributes to knowledge in a special way. Nevertheless, as a step towards answering the question of what is special about perception, it will be useful to situate perception in relation to some traditional requirements for knowledge. If one adopts the traditional view that knowledge requires justified, true belief, then perceiving will constitute knowing only insofar as perceptions are true, insofar as perceptions include some sort of belief, and insofar as perceptions are justified in some way.2 In this section, I describe wider and narrower understandings of belief, justification, and truth—indicating how perceptions can qualify as true justified beliefs if one adopts sufficiently broad understandings of these notions. Even if one adheres to a narrower understanding of these notions, however—an understanding according to which perceptions do not have the propositional structure required for truth or falsity, do not include an attitude of belief, or do not come with any justification—one may prefer to abandon the traditional requirements on knowledge rather than deny that perception constitutes knowledge. Indeed, it has been argued that perception must qualify as knowledge in order to serve as the justification for any other knowledge.3 If perception fails to meet the traditional requirements for knowledge then, it might be argued, so much the worse for those requirements. 2 Since Gettier (1963), problems with the analysis of knowledge as true, justified belief have been widely acknowledged. But most philosophers have responded by either (a) adding further requirements, (b) modifying traditional understandings of justification, or (c) abandoning the attempt to provide sufficient conditions, accepting that these are at least necessary constituents of knowledge. Williamson (2002) is an apparent exception, though I am inclined to view his account as an instance of strategy (b). 3 This line of argument is elaborated in McDowell (1994) and in Williamson (2002), for example. Russell (1987) further argues that cognitive development is only intelligible insofar as perception provides us with knowledge.

1.

PERCEIVING AND KNOWING

13

Must perceiving include believing? Does seeing a crow flying by require one to believe that a crow is flying by? There are at least three considerations that suggest it does not. First, there are cases of what Fred Dretske calls “non-epistemic seeing”—cases in which our senses pick up information from a given state of affairs without our noticing that state of affairs—as, for example, when our gaze sweeps across the objects on a table without our noticing the cufflinks we are looking for.4 It is hard to say just what is seen in such cases, though (cufflinks on a table? a shiny patch in the midst of dull brownness? a variation in the energy field?); and it is natural to report that we never actually saw the cufflinks (that our eyes may have registered it, but that we never saw it). Whether or not one is inclined to speak of “seeing” (or hearing, or feeling, etc.) in cases like these, however, the received information does not contribute to the phenomenology of our experience since “what it is like” to look at the table without noticing the cufflinks is no different from “what it is like” to look at the table without the cufflinks on it.5 And if the registration of information makes no difference to the phenomenology of our experience, neither will it make a difference to our beliefs. Insofar as we are interested in what is phenomenologically and epistemically distinctive about perception, then, what Dretske calls “non-epistemic seeing” is irrelevant. A second, more significant reason for doubting that perceiving requires believing stems from the conviction that believing requires us to employ concepts while perceiving does not—that believing that a crow is approaching requires one to apply the concept of a crow, while seeing that a crow is approaching does not. Much will depend, however, on what one takes concepts to be. If concepts are simply ways of organizing

4

Fred Dretske (1969), p. 18. For further discussion of the notion of non-epistemic seeing, see Daryl Close (1976) and John Heil (1982), who argue against its intelligibility, and F. N. Sibley (1971) and Edmond Wright (1981), who defend the notion. 5 This observation might also be put in term of consciousness: there is no conscious difference between looking at the table without noticing the cufflinks on it and looking at the table without any cufflinks on it. But consciousness is itself a much-contested notion, and there is the potential for confusion between “being conscious of a difference” and “being a difference in consciousness.” It is possible to be too inattentive, or too distracted, to be aware of what one is in fact perceiving. I want to allow for phenomenological differences of which we are unaware—cases in which looking at the table with the cufflinks prompts an experience with a distinctive phenomenology even though I am not aware of its distinctness. Evidence for the unnoticed change in phenomenology might include its effects on subsequent memories (“Now I remember where I saw them!”) or its effects on a current belief (“I’m sure they are somewhere in this room.”)—in which case such seeing is not epistemically inert after all.

14

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

and using the information we receive—as reflected in similarities versus differences in the way we respond to such information, for example, then perception must employ concepts for the simple reason that perceiving depends on pattern recognition. We won’t perceive a crow approaching unless we group various bits of information together in one way rather than another, assimilating the movement of this bird to the movements of other birds, for example, and distinguishing it from other bits of information we receive from the environment. On the other hand, if employing a concept depends on possessing a linguistic term for that concept, then it is certainly possible to perceive things for which one lacks a concept. I can distinguish crows from the trees they perch in and from many other sorts of animals without having names for any of them. There are many understandings of concepts that fall somewhere between these two extremes, of course: differences in responses that are under our conscious control may indicate conceptual versus non-conceptual modes of organization; concepts may be globally dependent on language without requiring one-to-one correspondence between individual concepts and individual words; and so on.6 Finally, even if one accepts that perception involves the employment of concepts, broadly construed, one might still object to the claim that perceiving requires believing on the grounds that belief requires a commitment that is not yet present in perception. Belief might be thought to require the acceptance or endorsement of a certain presentation of how things are, while perception merely provides a presentation that may or may not be endorsed. I may see a crow approaching but not yet believe my eyes—if the presence of crows in this vicinity seems too unlikely, for example; I actually see the crow but I withhold belief in its presence. Defenders of the view that perceiving requires believing tend to regard such cases as instances in which an initial and largely automatic (perceptual) belief is checked by a more considered (cognitive) belief—rather than a case in which there is no initial belief. Again, there are wider versus narrower

6

See Bermudez (1995, 2007) and Byrne (2005) for nuanced discussions of different arguments and possibilities regarding “non-conceptual content.” Much will turn on whether one thinks that perception and cognition are two different capacities, however much they may influence each other, versus two different points on a continuum that leads from input to output (with perception much closer to the input end of things).

1.

PERCEIVING AND KNOWING

15

conceptions of belief at play here, and deciding between these different conceptions depends on a broad range of theoretical considerations.7 Are perceptions true? If it is only propositions that can be true, if propositions must employ concepts, and if perceptions do not employ concepts, then perceptions cannot be true.8 On a broader understanding of truth, though, a state is true if it gets things right, and it gets things right if it presents things (propositionally or otherwise, using concepts or not) accurately or veridically, as they actually are. Now, it is widely accepted that perception must be veridical—that a perceived state of affairs must be an actual state of affairs. For Chloe¨ to see a crow approaching her, there must be a crow approaching her; if there is no crow approaching her, then she can’t see a crow approaching her.9 (She can, of course, seem to see a crow where there is nothing, or seem to see it approaching when it is standing still; but these would be apparent perceptions, not actual perceptions.) There are at least two different ways to understand this veridicality requirement on perception. The first way regards it as an external addition to the content of an experience—the experiential content (the appearance of an approaching crow) remaining the same whether or not it presents the world as it actually is, whether or not it deserves to be called a “perception.” On this view, the content (or an important part of the content) of a perception is representational content, and the veridicality of a representation will depend on how well its features “match” those of the represented state of affairs. While, traditionally, such matches were described in terms of resemblance between intrinsic properties (the blue of my representation of the sky resembling the blue of the actual sky), it is now more common to understand the relevant matches in terms of functional fit—the ability of an inner representation to co-vary with the represented outer state, for example, or the ability of a inner marker to lead us to the represented outer state. (Note how this more 7 As above, there are complex questions about whether beliefs must be conscious, whether beliefs must be determinate, and whether beliefs can be determinate in the absence of language. Burge (2010) engages with many of these issues in his criticisms of Quine and Davidson. 8 Note how, on this understanding of truth, it is not beliefs but propositions that are true. The traditional definition of knowledge should then be revised to: justified belief in true propositions. 9 A certain amount of imprecision is inevitable when it comes to deciding whether a given experience is veridical enough to count as a perception. How accurate must your experience of a crow’s shape be in order to qualify as perceiving its shape? If you experience a crow as taller than it is wide when in fact it is wider than it is tall, do you still perceive its shape? But this sort of vagueness is perfectly appropriate to the topic; after all, a certain amount of imprecision is also inevitable when it comes to deciding whether a given sentence is true.

16

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

contemporary understanding makes veridicality, or inner–outer fit, into an internal requirement on the possibility of representational content in general without, however, requiring that each particular representation be veridical. More exactly, veridicality becomes a precondition for assigning content to experience types, but not to each instantiation of that type.10) The second way of understanding the veridicality requirement regards the reality of what is perceived as an internal condition on that particular experience having the content that it does. On this view, if it turns out that there is no crow where I thought I saw one, I am not only mistaken in calling it a perception, I am also mistaken in thinking that the content of my experience was that of a crow. Either it was an experience with some other content—the experience of a dark shadow, perhaps—or it was not an experience of anything at all. (Note how this approach is committed to the possibility that two states with the same phenomenology may not share any contents, or to the possibility that two states with different phenomenologies may not be distinguishable to the subject of those states.11) Whether one embraces the veridicality of perception as an external requirement or an internal requirement, however, it means that perception will meet the truth requirement on knowledge—as long as truth is understood quite broadly as applying to any state that gets things right.12 The final requirement on knowledge, as traditionally analyzed, is justification. Are perceptions justified? If perception does not involve a belief in the accuracy of perceptual content, then it is hard to see how perceptions can be justified—since justification gives us reason to take things to be a certain way, and if we don’t take them to be a certain way, then there is nothing to justify. But even if perceptions do involve beliefs, they may seem 10 This view has been defended in many different forms, including the holism of Davidson (2001) and the evolutionary functionalism of Millikan (2000). 11 This is a very brief statement of the so-called “disjunctivist” option—actually a group of options— that has attracted many followers and many critics since defended by Snowdon (1979, 1990) and McDowell (1982). See a collection of papers on disjunctivist positions in Byrne, et al (2009). Note how this position takes a step beyond the Sellarsian (1956) insight, originating in Kant, that “success verbs” such as “see” or “hear” are more rather than less fundamental than “non-committal verbs” such as “appears” or “seems” (since appearances only acquire content against a backdrop of actual sightings, and we can only seem to hear the kinds of sounds that can actually be heard). It insists that there is no content that is common to actual seeing and mere seemings. 12 Adherents to the second view, above, could object to talk of veridicality, or getting things right, on the grounds that it doesn’t make sense to attribute veridicality to states that must, necessarily, be veridical. Siegel (2010) responds to this objection by noting how we can attribute veridicality to our beliefs in necessary truths even though the objects of those beliefs are necessarily true.

1.

PERCEIVING AND KNOWING

17

too foundational to admit of justification; for they are the beliefs by which other beliefs are justified, and justification must come to an end somewhere. There are at least two lines of response to the latter concern about perception and justification. First, if one embraces holism rather than foundationalism with respect to justification, there is no need for justification to come to an end—let alone to come to an end with perceptual beliefs. (Any finite activity of justification must come to an end, of course, but holism insists that there is no particular location that justification must come to an end, that where it comes to an end mostly depends on the context of inquiry.) Chloe¨’s belief that a crow is approaching may be justified by her confidence in her birdwatching skills, by her belief about the prevalence of crows in the area, by her trust in the report of the person standing next to her, and so on—where each of the things on this list may be used in turn to justify each other. For the holist, then, there is no need to bring justification to an end; perceptual beliefs, like all other beliefs, will be justified insofar as they are sufficiently supported by the other constituents of one’s overall web of belief.13 Second, even if one is not a holist about justification, one might suppose that justification does not come to an end with our perceptual beliefs but extends out into the world itself, with our relations to the perceived states of affairs themselves providing the ultimate justification of our perceptions. On this view, my perception that a crow is approaching is not justified by any beliefs I (or others) may have, but rather by its causal connections to the actual approach of a crow. One may prefer to speak of these causal connections as providing “entitlement” rather than “justification” for a perceptual belief—since the perceiver may not know anything about the relevant connections—but their role in transforming merely true belief into knowledge is like justification insofar as lawlike regularities are invoked in order to rule out beliefs whose truth is merely accidental. Variations on this idea include the view that perceptual beliefs are justified (entitled) as long as 13 The phrase “web of belief ” is from Quine and Ullian (1970), one of the earliest and most influential twentieth-century defense of holism. A thoroughgoing holism must treat rules of thought (such as the principle of non-contradiction) as themselves susceptible of falsification. Note that holism as such is neutral about (a) whether the relevant web of belief must be internal to the perceiver, and (b) whether the perceiver must be aware of the support offered by other parts of the relevant web. To require both of these things makes one an “internalist” about justification; to deny either makes one a certain sort of “externalist”—one that finds justification outside of an individual consciousness, without, however, stepping outside of mental states as such. (I say more about this distinction below.)

18

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

those types of beliefs are reliable indicators of the relevant states of affairs (with different possible ways of individuating “types”: same phenomenology, same sensory mode, same conceptual content), as well as the view that a perceptual belief is justified (entitled) only so long as that particular belief is a reliable indicator of that particular state of affairs. What is common to these views is a willingness to locate justification outside the realm of belief (and mentality more generally). So even animals that have no notion of reliability or causality can, nevertheless, be justified (entitled) in accepting the evidence of their senses if their senses are indeed reliably accurate, or appropriately caused. In what follows, I shall talk of perception and perceptual knowledge interchangeably—assuming that perceptions do meet justified true belief requirements, broadly understood. If a narrower interpretation that excludes perception from justified true belief is preferred (and I have recounted some of the reasons for such a preference), the reader can either reject these traditional requirements for knowledge while agreeing that perceiving is a form of knowing, or the reader can reinterpret my claims about perceptual knowledge to be claims about perceptual experience (that can contribute to knowledge) instead. Either of these options will be compatible with my account of what is special about perceptual versus non-perceptual ways of registering the approach of crows or the presence of a buzzing sound. As we will see, it is possible to understand what is special about perceptual versus non-perceptual experience of things, regardless of whether such experience constitutes knowledge or merely contributes to knowledge.

2. Immediacy and its Problems One of the most distinctive things about perceptual knowledge is its immediacy. When I perceive a bird in the tree outside my window, I gain more immediate knowledge of the bird than I would if I merely heard you report on its presence, or if I inferred its presence from the fact that seeds are falling from that tree. In these latter cases, your perceptions and your words, or my perception of falling seeds and my background beliefs about the likely source of these seeds, mediate between the bird’s

2.

IMMEDIACY AND ITS PROBLEMS

19

presence and my knowledge of its presence. This much seems clear; and it seems important for understanding what is distinctive about perception. What is not clear is just what sort of immediacy is important for a belief to be a perceptual belief, or for an experience to be a perceptual experience. In this section, I consider several different proposals concerning the sort of immediacy that is characteristic of perception. Each of these proposals is shown to be flawed, but their respective flaws are instructive. A clearer understanding of what perceptual immediacy is not sets the stage for a better account of what perceptual immediacy is—the account that is introduced in the next section. Let us consider, first, the suggestion that the causal relation between perceptions and their objects is more immediate than the causal relation between non-perceptual beliefs and their objects.14 Given the wealth of physical events leading from the presence of a crow to a person’s perception of that crow—bouncing light rays, contracting irises, changing retinal surfaces, and so on—it may not be possible to do without any causal intermediaries, but perhaps perceptions differ from reflective beliefs in having fewer such intermediaries. It is not hard to construct a scenario, however, where the causal pathway that connects a crow to my vision of the crow involves more steps than the causal pathway that connects the crow to my thought of a crow. Imagine, for example, a case in which I must look through complex computer-glasses in order to see the crow, while a transmitter attached to the crow emits an electrical pulse that activates a particular area of my brain, directly prompting the thought of a crow. The computer-glasses may involve more causal intermediaries than the transmitter device, but the result might be a perception nonetheless. We might try restricting the relevant causal intermediaries to mental intermediaries (or, perhaps equivalently, we could require causes to be specified at a suitable level of generality15)—allowing mental states to count as perceptions just in case the causal path leading from the perceived state of affairs to the perceptual state does not pass through any other state that qualifies as mental. But the empirical difficulties of identifying relevant causal pathways 14

It is interesting to note how Quine regards perceptions as the causal foundations for knowledge even though he denies that they provide justificatory foundations; for he identifies perceptions (or “observations”) as the beliefs that lie at the “periphery” of a web of belief—i.e. at the location where the world “impinges” most directly on our beliefs. 15 Yablo (2003) defends the importance of specifying causes at relevant levels of generality.

20

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

in an exceedingly complex brain, and the philosophical complexities of deciding just what constitutes a mental state, seem both daunting and beside the point. Surely we do not need to argue against Marr’s representational intermediaries, for example, or to insist on the non-mental character of the representations he describes, in order to know that vision has the immediacy that is characteristic of perception. (I assume that if vision lacks a particular type of immediacy, then that cannot be the type that is required for perception—since vision is the epitome of perception.) Insofar as the immediacy of perception is something we recognize independently of any particular theory about how information is processed by our brains, perhaps it is only conscious mental intermediaries that must be lacking in order for our knowledge to be perceptual. Harman (1973) argues that perception relies on inferences, but that the relevant inferences remain unconscious in the case of perception.16 Likewise, Burge (2010) maintains that the principles governing the transformation of information in perceptual systems are unlike the principles governing the transformation of information in thought insofar as we must remain unconscious of the former.17 The immediacy of perception thus becomes less a matter of causal immediacy than of experiential immediacy or immediacy in awareness. But if it is the experiential immediacy of our knowledge that is the mark of perception— the fact that we are not aware of any mental states mediating between the object of our knowledge and our knowledge of it—then any knowledge that seems to arrive “out of the blue” or “in a flash” will have the requisite sort of immediacy. If a transmitter attached to a crow sends an electrical signal directly to my brain, prompting me to think that there is a crow nearby, without my being aware of any mediating thoughts or memories, my thought will have just as much experiential immediacy as a perception—yet it remains a thought rather than a perception. We have not yet claimed that such experiential immediacy suffices for perception, so it could be that the difference between my “in a flash” thought of a crow’s presence and my perception of a crow’s presence is 16 Harman acknowledges that his view extends the use of the term “inference” into the unconscious, but rejects the suggestion that this extension amounts to a change in the meaning of inference (p. 178). 17 Burge does not insist that thinkers are always conscious of the principles governing their thought, but he does insist that the principles governing thought, unlike the principles governing perception, can be represented (and thus made accessible to conscious). I say more about this aspect of Burge’s view in Chapter 2, Section 1.

2.

IMMEDIACY AND ITS PROBLEMS

21

not due to different sorts of immediacy but to something else entirely—to the presence of sensory qualia, for example. (We will return to this possibility in Section 4.) Even if we thought that additional requirements could be identified, though, equating perceptual immediacy with experiential immediacy encounters a problem from another direction—from the possibility of perceiving something while fully aware of the intermediary role played by other mental states. I can perceive a bird in the tree in full awareness of the mediating role of a mentor’s advice (“Look for movement that is different from the movement of the leaves,” for example, or “Pay special attention to the branches that are most horizontal”), or in full awareness of the mediating role of my own memories (of similar birds in similar trees). Awareness of these mediators seems irrelevant to the status of one’s knowledge as genuinely perceptual. Put another way: these particular mediators—whether or not we are aware of them—do not threaten the sort of immediacy that matters to perception.18 A related problem with understanding the immediacy of perception to be experiential immediacy is the way that it makes perceptions disappear with certain gains in self-consciousness and increase with certain losses of selfconsciousness. If the relevant immediacy is experiential immediacy, then the less attentive we are to the memories and associations that play a causal role in producing our experiences of birds the more likely we are to perceive the birds. There is a sense, of course, in which being too attentive to mental intermediaries in the perceptual process can prevent one from seeing the bird that is in front of one’s eyes. But this is because of the limitations of a divided attention rather than an incompatibility between attending to the object of a perception and attending to factors that play a causal role in producing to that perception. There is a sense in which very good birdwatchers cease to be conscious of the many considerations and 18 Millikan (2000) maintains that “There are two things that distinguish [causally] direct perception quite sharply from the acquisition of information through language [which she also calls ‘perception’], but neither implies a difference in [experiential] immediacy. In direct perception, the spatial and temporal relation of the perceiver to the object perceived is given, whereas it is not normally given through language . . . The second feature that distinguishes perception is its near infallibility. It is remarkably difficult to deceive people about what they are actually seeing or hearing. This is why ‘seeing is believing’.” (p. 85) The second feature suggests that Millikan views “direct” perception as having a kind of justificatory immediacy (being near infallible, leading immediately to belief ). The first feature concerns spatial placement. I agree with Millikan about these distinguishing features, but think that both need more explication than she offers. Furthermore, as will become clear below, I think that they have a common source.

22

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

deliberations that have made their perceptions possible, but very good birdwatchers do not cease to perceive the birds whenever those considerations and deliberations return to consciousness.19 There is no reason, then, for thinking that self-consciousness about the process of perception is incompatible with perception as such. One way to circumvent these problems (and to highlight the crucial epistemic role of perception) is to focus on justificatory immediacy rather than causal immediacy or experiential immediacy. For even if my companion’s remarks played a important mediating role in transforming the play of light on my retina into my perception of the crow, and even if I am quite conscious of that mediating role, once I perceive the crow for myself I do not rely on her remarks to justify my belief that there is a crow in the tree. Contrast this with a case in which I confidently echo my companion’s report that there a crow in the tree without myself perceiving it, in which case I will need to appeal to her remarks in order to justify my belief. On this proposal, then, the sort of immediacy that is relevant to perception is justificatory immediacy: perceptual belief, or perceptual knowledge, is immediate in the sense that it does not rely on any other beliefs, or any other knowledge, for its justification.20 Currently, this is the dominant way of understanding the immediacy of perception. But it too is problematic. Consider two different interpretations of justificatory immediacy. On the first interpretation, my knowledge of x has justificatory immediacy just in case it cannot be justified by reference to any other knowledge. On the second interpretation, my knowledge of x has justificatory immediacy just it case it need not be justified by reference to any other knowledge. Does perceptual knowledge have either sort of justificatory immediacy? On the one hand, it seems that all knowledge can gain support from other knowledge. If I see a bird in a tree (and thus know, without further justification, that there is a bird in that tree), I could always provide further justification for my belief by appealing to other perceptual judgments (about color and movement, about falling debris, etc.), or by appealing to the judgments of other people (with better vantage points, more expertise, etc.). On the 19

A similar example concerning Brownian motion and the perception of mu mesons is discussed by Brandom (1994). 20 If one holds that justificatory mediation is always inferential, and that inferences are always conscious—a thesis defended by Searle (1992), for example, then justificatory mediation will entail experiential mediation (but not vice versa).

2.

IMMEDIACY AND ITS PROBLEMS

23

other hand, the need to invoke other knowledge to justify a belief seems to depend on contextual matters and not on the perceptual versus nonperceptual character of one’s knowledge. In certain contexts (where squirrels are much more common, where the light is dim), I may need to appeal to other knowledge in order to justify my judgment that there is a bird in the tree (knowledge that I am an experienced birdwatcher, knowledge that I have excellent night vision). In most contexts, though, I do not need to invoke to any other knowledge for justification even when my judgment is not perceptual (my knowledge that I was born in Michigan, for example, or my knowledge that the Spanish word for fig is “higo”). Would it help to distinguish entitlement from justification, allowing that perceptual judgments need entitlement but not justification? While justification supports the truth of a judgment by providing reasons in favor of its particular content, entitlement supports the truth of a judgment by establishing the presence of favorable conditions for judgments of that type. The fact that a bird is black and the fact that it makes a harsh cawing sound can justify my claim that it is a crow, while the fact that I have normal eyesight and the fact that I have sufficient familiarity with birds can entitle me to my perceptual judgments about birds in general.21 There is a rich and recent literature that defends this distinction, and just as there are both internalist and externalist understandings of justification, there are internalist and externalist understandings of entitlement. (According to the internalist about entitlement, perceptual knowledge requires knowledge that relevant background conditions obtain; according to the externalist about entitlement, it is enough that such conditions obtain, whether recognized or not.22) Perceptual judgments are not unique in needing entitlement without justification, however; we also need entitlement but not justification for judgments that come from memory or from hearsay, for example. So the notion of entitlement does not serve to identify what is especially immediate about perception.

21 One could view entitlement as a particular form of justification, having the following form: S’s judgment that p constitutes knowledge under conditions C. Conditions C obtain. Therefore, S knows p. Therefore, p. (Note that there is nothing that prevents C from including unconscious inferences on the part of S, and there is nothing in the scheme that specifies whether or not S must know that conditions C obtain.) 22 Brandom (1994) is an internalist about entitlement; Burge (2003) is an externalist about entitlement.

24

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

Still another approach to justificatory immediacy would distinguish between judgments for which one need not be able to offer further justification in order for them to qualify as knowledge and judgments for which one must be able to offer further justification in order for them to qualify as knowledge. Analyses that rely on capacities are rightly suspect, given the difficulty of determining just when one does or does not have the requisite capacity. (Am I capable of offering further justification if I tend to get confused when challenged? Am I capable of offering further justification if my memory is failing?) But even if we are willing to invoke capacities in this way, needing versus not needing a capacity for justification marks the divide between internalists and externalists about justification rather than the divide between non-perceptual and perceptual forms of knowledge. For the internalist insists that a belief is justified only if the justification is available to the believer, while the externalist will allow that beliefs are justified even when the justification is unavailable to the believer. (It is possible to be an externalist about the justification of perceptual beliefs and an internalist about the justification of non-perceptual beliefs, but splitting one’s allegiance in this way would need to be independently motivated in order to provide an explanation of what is special about perceptual knowledge.) Finally, one could simply accept that the need for justification depends on context, and maintain that the perceptual versus non-perceptual character of our knowledge varies accordingly. When no further justification is needed, our knowledge is perceptual; when further justification is needed, it is not perceptual. This would result in what I would call a “deflationary” understanding of perception. A deflationary understanding of a category like perception explicates what the category is by detailing how and why we use it as we do, resisting the urge to explain our usage by appeal to intrinsic features of an underlying referent. (The deflationist does not deny that there is a referent, but does “deflate” its role, insisting that it cannot explain— versus merely express, or reflect—our use.23) So if the purpose of categorizing certain experiences as perceptions is simply to discharge the need for any further justification, then it is not surprising that states with a particular phenomenology could count as perceptions in one context but not another; 23

Put another way: deflationists describe the rules we follow but reject any attempt to establish foundations or justifications for following those rules. See, for example, McDowell (1984).

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

25

for the call for further justification will be appropriate in some contexts but not others. In the absence of a more substantive account of perceptual immediacy— one that helps to explain what is special about the phenomenology of perception, a deflationary position can be tempting. But embracing a deflationist position means abandoning the attempt to connect the phenomenology of perception to its distinctive justificatory role. Philosophers’ accounts of the normative role of perception become effectively disjoined from their descriptions of its phenomenological character. But there is something exceedingly odd about this disconnect; surely there is a connection between what it is like to perceive a crow (the phenomenology of perceptual experience) and what sort of justification is provided by such a perception (its epistemic role). In any case, despairing of such a connection seems premature and, other things equal, accounts that can make sense of the connection are preferable to accounts that do not.

3. Self-evidence and the Experience of Objectivity In this section I introduce an account of perception that shows how the justificatory immediacy of perception and the experiential immediacy of perception are linked, an account that not only describes the normative role of what qualifies as “perception” but seeks to ground that role in the phenomenological structure of perceptual experience. This account is not fully filled out until Chapter 2, but its basic structure and its basic appeal—especially against the backdrop of the flawed accounts discussed above—should be evident by the end of this section. Perceptual knowledge or experience, I suggest, has justificatory immediacy not because it is incapable of or unneedful of further justification but, rather, because it is self-justifying. The veridicality of perceptual experiences is signaled from within the experience itself. Self-justifying experiences can still be doubted, and further justification is sometimes needed; but the selfjustifying character of perceptual knowledge is what distinguishes it from non-perceptual knowledge. The claim that perception supplies us with self-evident truths is not new, of course; but satisfactory explanations of self-evidence or self-justification

26

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

have proven elusive. To begin to understand what the self-justifying character of perceptions might amount to, consider the following case: Five-year old Sam sometimes becomes convinced that his neighbor’s pet spider has entered his room. He has no idea what this particular spider looks like, or how it entered, but he confidently reports its presence and its location, and he is almost always right. Asked whether there was any special sound or sensation that led to his conviction, he answers “no.” Asked to give a reason for his conviction, Sam says “I don’t know; I just thought it was here.” A series of experiments is conducted, and they establish Sam’s reliability without, however, suggesting any explanation for it: Sam’s location relative to the spider seems irrelevant (he can be sitting on the chair under which the spider is crawling, or he can be facing in the opposite direction); covering Sam’s eyes and ears makes no difference; and he reliably reports the presence of the spider even when it is enclosed in a lead box. Eventually, everyone (including Sam) is convinced that he does register, in some mysterious way, the presence of this spider, and they trust his reports.

This seems to be a case of a belief that has experiential immediacy and justificatory immediacy without, however, being a case of perception. Sam’s conviction comes to him directly, without relying (consciously, anyway) on any other beliefs, and insofar as Sam is a reliable reporter of the spider, there is no need to justify his claims by appeal to any other facts. Yet Sam’s belief comes to him in the form of a “blind” conviction rather than in the form of a perception.24 Cases similar to that of Sam are sometimes invoked in order to discredit externalist theories of justification. Bonjour (1985), for example, argues that although a clairvoyant may be entirely reliable in his judgments, he is not thereby justified in his beliefs; and Lehrer (1990) makes a similar point about someone who, unbeknownst to himself, has a brain implant that ensures that his temperature judgments are reliable.25 In Sam’s case, though, Sam 24 This hypothetical case of Sam is different than actual cases of so-called “blindsight” in several ways: (1) Sam has normal eyesight, and could see the spider clearly if he looked at it, (2) Sam offers his reports quite freely and confidently, not reluctantly and doubtfully, and (3) Sam does not report an accompanying feeling of movement. The case of Sam is more like that of the chicken sexer who can distinguish between male and female chicks without knowing how. McDowell (2002a), responding to Brandom, allows that chicken sexers have non-inferential knowledge, but insists that it is “non-observational”: “The chicken sexers find themselves with inclinations to say things, which they can supply with justification only by adducing considerations external to what is available to them in their present angle on reality—considerations about the skill that they know they have acquired.” (p. 280) 25 These and other cases are discussed in Sosa (2007), Kitcher (1992), and Kornblith (2002), all of whom defend, in one way or another, reliabilist accounts of justification. Lyons, though, maintains that

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

27

may know that his judgments in this domain are reliable, in which case they are justified on internalist as well as externalist conceptions of justification. But that doesn’t make them into perceptual judgments. (It might even be the case that Sam knows that his judgments are reliable even though he cannot cite any evidence in support of his claim; he may not remember how he learned to trust himself on this matter.26) Sam’s case, then, does not challenge a reliabilist understanding of justification (whether internal or external); what it does challenge is some standard accounts of the immediacy of perception—accounts that focus on experiential immediacy on the one hand, and accounts that focus on justificatory immediacy on the other. What is missing from Sam’s experience, I suggest—what perception has that Sam’s knowledge lacks—is an indication of veridicality, a sign of objectivity, from within the phenomenology of that experience. (Whether this is the only thing that prevents Sam’s experience from being a perception is a question I return to in Section 4 of this chapter.) Sam can know that a particular spider is nearby, and he can know that he knows, but unlike a case in which he actually sees the spider there is nothing about the phenomenological character of his blind conviction that supports its objectivity. He may be justified in assuming the accuracy of his belief, but his justification is based on features that are external to the phenomenological character of the experience by which he acquires his belief. It would be a mistake to suppose that the phenomenological character of an experience could guarantee its accuracy (the phenomenology of a genuine perception and the phenomenology of an illusion may sometimes be the same), but it would also be a mistake to suppose that our reasons for trusting perceptual experiences are entirely independent of the phenomenological character of those experiences (that the phenomenology of seeing a crow could just as well have turned out to be like the phenomenology of scratching a mosquito bite). reliability is only sufficient for the justification of beliefs acquired through perceptual systems, where perceptual systems must meet several other conditions—including energy transduction through sensory organs and an extended learning process. 26 Brewer (1999), pp. 242–5, uses the phrase “non-demonstrative perceptual knowledge” to cover cases where one knows one has a reliable, automatic epistemic skill (a) though one is incapable of recounting features that were previous bases of inference (as in “Tony is furious”) or (b) simply on the basis of trial and error feedback (“That is the Tristan chord”). These are cases of knowledge about spatiotemporal particulars, but not knowledge of their spatiotemporal properties; one can point to the object but not to the property in question. How similar they are to my case of Sam will depend on just how the relevant property is experienced: is it more like a case of clairvoyance or more like a case of an ability that one can adapt to variations in circumstances? (See modifications of the Sam story below.)

28

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

Just how it is possible for an object’s independence from experience to be evident from within experience is a puzzle we shall turn to shortly. It should be clear, however, that to the extent that the veracity of an experience is indicated from within the phenomenology of that experience, it is a selfjustifying experience. Justification is no guarantee of truth, and self-justifying experiences are not always veridical, but if an experience is self-justifying then the very character of that experience gives us reason (inconclusive as it may be) to trust its accuracy. Contrast this with the case of believing that Nashville is the capital of Tennessee. Most of us are entitled to trust this belief even if we are incapable of providing any evidence on its behalf (it is the sort of belief about which we are reasonably reliable whether or not we understand our own reliability). But there is nothing in my experience that makes the truth of this judgment self-evident. It is not enough, then, for Sam’s knowledge of the spider to have experiential immediacy—to come “in a flash,” without any conscious reliance on inference. Nor is it enough for Sam’s knowledge to have justificatory immediacy—for Sam to be entitled to his conviction without the need for further justification. To be a case of perception, the objectivity of the spider must be supported by the phenomenology of Sam’s experience; the justification for his judgment must be integral to the experience itself. Michael Ayers (1993), elaborating on John Locke’s understanding of perception, considers some imaginary cases that are similar to our case of Sam. He contrasts standard cases in which “when I believe my eyes, there is no mystery about the source and basis of my belief ” with a case in which a person merely feels that someone is watching him from behind, where his feelings are reliable indicators of such an observer.27 Although this person may indeed know the source of his belief, Ayers maintains that his sensation “presents something other than the state of affairs in question and so is not appropriate to the belief which it causes,” with the result that there is still a “subjective mysteriousness” to his knowledge. I do not think that this is the best way to put the point, since there is nothing “inappropriate” about forming a belief about one thing on the basis of one’s experience of another (assuming that there is a reliable correlation between the two), and as long a person understands enough about the causal path that leads from the outof-sight watcher to the feeling-of-being-watched she will not find her 27

Ayers (1993), pp. 171–2.

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

29

knowledge “mysterious.” But Ayers is right to emphasize the difference between experiences whose sources are evident from within those very experiences—that is, perceptual experiences, and experiences that fail to display their sources in this way—for example, indicative sensations.28 Even if we were to remove the mystery of Sam’s accuracy, discovering that his judgment is based on odors that are dispersed by a particular spider and registered by Sam’s unusually sensitive skin, there would be nothing in the phenomenology of Sam’s experience to indicate the spider as its source. For an object to be the source of an experience, it must exist independently of how (and whether) it is experienced.29 More generally, for a state of affairs to be objective, it must exist or obtain independently of how (and whether) it is experienced.30 To experience a state of affairs as objective, then, must be to experience it as existing independently of how and whether it is experienced. But how is this possible? It is easy to understand how we might experience one state of affairs as independent of various other states of affairs if we observe that it remains constant despite changes in those other states of affairs. And it is easy to see how we might experience one experience as independent of various other experiences if we notice that it remains constant despite changes in those other experiences. But how can we separate the object of our experience and the experience itself sufficiently to recognize the constancy of the first despite changes in the second? We can, of course, stand back from our experiences enough to notice that the

28 Ayers (1993), p. 172, Ayers claims that it is the “intrinsic intentionality of sensation” that is needed in order for sense experience to play an epistemic role. The important thing here, though, is not whether a given sensation has intrinsic intentionality (a tingling in one’s thumbs may be the intrinsic object of a sensation) or whether that experience can play an epistemic role (the thumb tingling may be very good evidence for the presence of someone watching from behind). What is important, rather, is that the object of a perceptual belief is revealed as the source of one’s experience from within that experience itself. 29 There are various ways in which one object or state of affairs can be the “source” of another. Whether sources are always causes depends, in part, on one’s understanding of causation. Paul Snowdon (1979, 1990), for example, has argued against a causal requirement on perception, but his approach depends on endorsing certain counterfactual relations between perceptual judgments and the objects of those judgments—counterfactual relations that, according to many, commit one to causal relations. Campbell (2002) defends a view according to which perception relies on an “intuitive physics” that enables us to understand the perceived object as the source of our experience. 30 Burge (2010), pp. 46–50, considers various definitions of objectivity and points out that equating objective states of affairs with mind-independent states of affairs rules out artifacts and animals with minds from being objective. But if one equates an objective state of affairs with a state of affairs that exists independently of any representation of that state of affairs, then artifacts and animals are objective since, although they must themselves engage in representation, they need not themselves be represented.

30

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

accuracy of some experiences is inconsistent with the accuracy of others; and assuming that the world is not inconsistent, we may rightly conclude that the world is independent of our experiences. But this will, at best, tell us that there is some independent state of affairs; it won’t give us the experience of a particular state of affairs as independent. Furthermore, knowledge of independence or objectivity that is reflective in this way seems already to presuppose a less reflective way of recognizing objectivity—perceptual knowledge, which is precisely what we are seeking to understand. Alva Noe¨’s account of perception offers a revealing response to this challenge. He distinguishes two aspects or components of perception, and argues that experiencing the dependence of one aspect on the other amounts to an experience of objectivity. According to Noe¨, the contents of perceptual experience include a factual dimension and a perspectival dimension, an aspect that indicates how things are and an aspect that indicates how things appear. We experience not only how things are, but also how they look from here. We experience that the plate is round and that it looks elliptical from here. Its elliptical look from here is a genuine property of the plate—we see the shape and we see the perspectival shape from here—but it is also a relational property, one that depends on where ‘here’ is. Perceptual content . . . is two-dimensional. It can vary along a factual dimension, in regard to how things are. And it can vary along a perspectival dimension, in regard to how things look from the vantage point of the perceiver. Visual experience always has both these dimensions of content.31

Noe¨ treats instances of immediate but non-perceptual knowledge (as in our case of Sam) as cases in which our experience has factual content (the spider 31 Noe¨ (2003), pp. 2–3. Elsewhere, Noe¨ (2002), p. 74, makes a similar distinction between the representational and sensory factors that make up the qualitative character of experience:

“The qualitative character of experience, as we have seen, depends on two factors. First, it depends on the qualities that we experience (e.g. looks, sounds, etc). This is a representational feature. Second, it depends on the character of the activity in which the temporally extended activity may consist. So, for example, the fact that we do not make eye movements when we explore the environment haptically makes a difference to what it is like to touch. These differences in the sensorimotor contingencies governing the different sensory modalities are differences in the qualitative character of experience that do not correspond, directly at least, to differences in what is perceived.” It is not clear (at least not to me) whether Noe¨ intends the second, non-representational aspect of experience to count as part of its content—akin to the perspectival contents discussed above. The term “content” doesn’t matter, perhaps, but the ability to recognize the dependence of subjective phenomenology on objective fact, through the dependence of one factor on another, does.

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

31

is in the room) but not perspectival content (how it looks from here), contents which are dependent on the facts they pick out but not on the perspective from which those facts are known. Perception, he argues, requires both sorts of content. And whereas factual content’s dependence on the world (its truth) must be discovered from an external point of view, we can discover the dependence of perspectival content on factual content by moving through space and noticing how the way things look depends both on how they are and on where we are.32 Indeed, we only become capable of distinguishing looks from facts insofar as we become capable of tracking such dependencies. While I agree with the suggestion that we experience the objectivity of a state of affairs when we experience constancy across perspectival change, Noe¨’s insistence that we experience appearances (e.g. the elliptical look of a plate) as well as facts (e.g. its roundness) creates more contents (and more puzzles) than are needed.33 When we look at a plate from the side, we do not see an elliptical appearance in addition to the round shape of the plate; we either see the round plate or we mistake the round plate for something elliptical. Likewise, when we experience the shapes around us as if they were flattened onto a screen before our eyes, we do not simply attend to an already present aspect of experience (as Noe¨ suggests); rather, we shift from seeing three-dimensional objects to imagining their two-dimensional counterparts. While we do need to undergo changes in perspective in order to discover what remains invariant across changes in perspective, and while changes in perspective may involve changes in the content of experience (different sides of an object come into view, and different parts of its surroundings appear), this does not mean that the content of our percep-

32

The question of whether these movements must be actively initiated or whether mere movement with respect to an object (which may be initiated by the object rather than the viewer) is left open. 33 Noe¨’s talk of two different “dimensions” of content is less problematic than his talk of two different types of experiential content—“looks” as well as facts, for example perspectival shape as well as objective shape. He is clearly aware of the dangers of invoking sense data as the true object of perception, but he is determined to include appearances, understood as relational properties (between the perceiver and the perceived) as part of the contents of perception. I agree that experiencing alternative perspectives is necessary for experiencing something as objective, but I reject the view that every perspectival experience has, as part of its content, a perspectival property. This point is related to one made by Williams (1973) about the possibility of imagining different perspectives without thereby imagining anything about oneself. I can imagine another perspective on the plate without imagining a perspectival property of the plate. Noe¨ (2008) has recently replied to several commentators who criticize his two-aspect view.

32

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

tions can be divided between factual and perspectival components or dimensions.34 The phenomenology of objectivity depends on the convergence of information gained from different points of view, not the experienced dependence of some experiential contents on others. In short, I agree with Noe¨’s reliance on the coordination of different perspectives to explain the experience of objectivity, but I disagree with his invocation of a special type of content, or a second aspect of content, that is perspectival. Consider the perception of a table tilting relative to the floor. I may acquire information about this particular state of affairs both by looking at it and by touching it, by looking at it from one side and from another, by looking at it from close up and from a distance. Insofar as I am able to coordinate the information that is acquired through these different perspectives and modalities, to track what is constant and to use one perspective to check another, then I will experience the tilting table as an objective state of affairs. And, equally, insofar as I experience the tilting table as an objective state of affairs, I will expect a certain constancy across these different sources of information. The relationship between using one perspective to confirm another on the one hand, and taking both perspectives to be perspectives on something objective on the other, is one of mutual dependence; neither one has causal or explanatory priority over the other. That a table’s tilting can be known in any number of ways ensures that it is independent of any particular method by which we acquire our knowledge; and the independence of the table’s tilting ensures that it can be known in several different ways.35 There are important differences between the sorts of information that are acquired through different sensory modalities. Sight but not hearing gives us information about colors, hearing but not sight gives us information about sounds, smell but not touch gives us information about odors, and so on. While some properties are discerned through more than one sense organ (e.g. shape being discerned through both sight and touch), there are other

34 This is part of what it means for experience to be “transparent.” In the words of Martin (2002): “there does not seem to be some private entity corresponding to each object of perception, or a subjective quality to correspond to each perceived feature of such objects.” 35 The importance of mastering multiple perspectives in order to fix the content of an experience is familiar from the work of Dretske (1981), Millikan (2000), Davidson (2001), Burge (2010), and others. What is distinctive about Noe¨’s analysis is his attempt to use the phenomenology of converging perspectives to explain the phenomenology of objective reference.

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

33

properties that we learn through one sense only. There are also important differences in the degree to which informational convergence can be achieved within a given modality, the extent to which different modalities allow for different perspectives on a single property. While it is widely recognized that vision allows for different perspectives on the same shape or the same color, many are doubtful about the possibility of different perspectives on the same flavor or the same odor—maintaining, instead, that each apparent difference in flavor or odor is an actual difference in flavor or odor, that flavor and odor alter with the subject’s “perspective.” What is important for our purposes, however, is a recognition of the way that perspectival differences and perspectival convergence, whether across modalities or within a given modality, are aligned with spatial differences. For when different perspectives converge to give us the experience of an object’s independence, they also converge to show us an object’s location in space. This seems obvious in the case of vision, where different lines of sight must be juxtaposed in order to determine the spatial position of what is seen, but it is true within other sensory modalities as well. We discover the location of a sound just as we discover its objectivity—by tilting our heads from side to side, or by walking to and fro, accumulating different auditory perspectives on a single state of affairs. We may also imagine hearing the sound from a closer location, or from a location farther away from various distracting noises; and these imaginings can give us a more vivid sense of the objective nature and location of the sound. Likewise with touching and tasting and smelling; by moving through different spatial positions (both actually and imaginatively) while tracking the same texture or flavor or odor, we come to experience both the objective character and the particular location of those qualities.36

36

Cassam (2007), Chapter 3, describes a case of hearing two unseen people talking, and uses this example to argue against the claim that we need to perceive objects in space in order to differentiate them from one another, and thus to perceive them at all. He also describes a case of hearing oneself speaking to argue against the claim that we must differentiate what is perceived from oneself, and must therefore experience it in (outer) space. But he defends the claim that being able to perceive objects in space (by touch or by sight) is necessary for being able to perceive them at all. My position is situated between his negative and his positive claims, for I maintain that multiple perspectives (and thus space) must be actually imagined in order to perceive something; it is the imagining of these alternative perspectives that makes my experience one of objectivity and thus, on my view, perception. (I defend the need for actual imagining in Chapter 2, below.)

34

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

Cross-modal convergences are also experienced as spatial convergences insofar as modal distinctions depend on spatial distinctions. Not only is it the case that different sensory organs (eyes, mouths, ears) are located in different places so must receive information along different spatial pathways; it is also the case that different properties (colors, tastes, sounds) are distributed differently in space (on surfaces, underneath surfaces, beyond surfaces) so distribute their information from different locations. There is no one-toone correlation between sensory qualities and sensory organs; for each of our sense organs gives us knowledge of many different sensory qualities (we see color and brightness and sheen, we taste sweetness and bitterness and sourness, we hear pitch and timbre and volume); and many sensory qualities are perceived through more than one sense organ (we can both see and touch the smoothness of a tomato, we can both feel and hear the beat of a drum, we can both taste and smell the sourness of a lemon).37 Different sensory modalities rely on different spatial pathways to connect the distribution of information and its reception, however. Part of what distinguishes one sensory modality from another is the differences in causal pathways that support each sense—pathways that also reflect differences in the enabling conditions for one sense versus another. Sight but not hearing requires intervening light; touch but not smell requires an absence of intervening surfaces; taste but not hearing requires close proximity; and so on. Our bodily recognition of these differences is evident in the ways we negotiate our environment to improve our sight, our hearing, our touch, our smell. When encountering a new environment or trying to learn more about one that we are already familiar with, we may move to the side in order to hear better, move forward to smell better, move away in order to get a better view; different sensory modalities demand different spatial maneuvering.38 We distinguish between sensory modalities, then, not only on the basis of differences in what is sensed or on the basis of differences in the bodily

37

These observations count against a traditional, Aristotelian view that each sense has its own “object.” Sean Kelly (2005) has nicely documented the way in which perception implicitly relies on the recognition of ideal viewing positions, and movements that approach that ideal. Brian O’Shaughnessy (1990) makes interesting distinctions between visual properties that “attach” to objects versus sounds and smells, e.g., that traverse or pervade a space. Dominic McIver Lopes (2000) insists that different sensory modalities have different phenomenologies, and his descriptions of these differences are largely dependent on their different spatial functioning. 38

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

35

locations of the sensing organs but also on the basis of the different causal pathways—through space—that connect the two. Insofar as the same property can be experienced through different sensory modalities, then, those modalities offer different spatial perspectives on that property, and insofar as a property is perceived through one modality only, that modality must allow for a multiplicity of spatial perspectives. For if information can be received from one perspective only, the distinction between perspective and object disappears, and the experience is no longer one of self-evident objectivity or perception of something as located in space. Returning, then, to the case of Sam: at least part of the reason that Sam does not perceive the spider in the room (nor does he perceive the presence of special emissions given off by the spider—even if we determine that his reliable reports are caused by his special sensitivity to such emissions) is that Sam does not experience the objectivity of the spider (or the emissions). He may come to know that certain convictions (and, perhaps, certain sensations) are reliable indicators of an objective state of affairs, but their objectivity is not evident from within the experience itself. In order for Sam to experience the objectivity of the spider, he would have to experience it through a variety of perspectives or modalities. Suppose, for example, that the strength of Sam’s conviction about the presence or absence of the spider increases as his skin exposure increases (perhaps the emissions pass through the skin, with more skin exposure allowing better pick up of emissions) and as the amount of air in his lungs increases (more air carrying more emissions). Sam might then learn to adjust his clothing and his breathing in order to arrive at more confident judgments about the presence and the location of the spider. Wondering about the spider’s presence, Sam could try exposing more skin and he could try taking deeper breaths, and if both of these strategies together yielded a stronger conviction of the spider’s presence he would have greater reason to believe in the spider’s objectivity, reason for regarding the object of his experience to be independent of any particular way he has of knowing it. Whether or not he made such adjustments consciously, and whether or not he could explain their efficacy, his actions could demonstrate a familiarity with different means of confirming the presence (or absence) of the spider. Sam’s methods would constitute different ways by which to acquire knowledge of the spider, and insofar as the verdicts of these different ways converged within a single

36

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

experience, the veridicality of his belief could become self-evident. Sam could still not see the spider (his knowledge would not be visual) but he would now experience the spider’s objectivity; its objectivity could be evident from within the phenomenology of his experience. We allowed that Sam’s initial convictions on their own could include specifications of where the spider is located: “the spider is under the bed,” “the spider is on the edge of the windowsill.”39 But it is only when Sam was able to coordinate different movements to provide supporting evidence for these convictions that he was able to experience the objectivity of the spider. Without the convergence of these different methods his knowledge would not be self-justifying; his convictions could give him knowledge of the spider’s location but they would not meet the requirements of perception. To what extent is the plurality of ways in which we can know a state of affairs, and thus the independence of that state of affairs, something that is evident from within a given experience at a given time? If the experiences that are used to check one another occur separately, how can the commonality of their object be revealed by any one perceptual experience?40 First, it is worth noting that while it is true that we can experience only a limited number of ways of knowing at a time, it is certainly possible simultaneously to touch and to look, to touch from two different angles, or to look from the slightly different angles of two separate eyes. Furthermore, it is also possible to imagine several different perspectives and modalities simultaneously; these will include perspectives that we do not now occupy, and modalities that we do not now engage. The coordination of those different perspectives and modalities in imagination can also create an experience of objectivity. As I look at the tilted table from here, I can also imagine what it looks like from over there; and I can imagine what it would feel like to rest my hands on the table, or to lie on its surface; and so on. Each of these 39 Derivative knowledge of particulars is always possible—via testimony, for example. See Evans (1982) and Snowdon (1990) on the distinction between basic and derivative demonstrative thought. It is fine to emphasize the importance of demonstratives for perception, but this will not really be explanatory unless we can give an explanation of our grasp of spatiotemporal locations. 40 Noe¨ (2006) claims that different perspectival contents occur at different times, not all at once; “experience is a temporally extended phenomenon; it is an activity of skillful probing” (p. 430) and “experiential presence is virtual all the way in . . . The rear side is present virtually, but the present side is present simpliciter. Notice, however, that you do not, as a matter of fact, have the whole of the facing side of the tomatoes in consciousness all at once.” (p. 427)

3.

SELF-EVIDENCE AND THE EXPERIENCE OF OBJECTIVITY

37

combinations can serve to secure our experience of the table as something objective. There is not a sharp line between what is experienced as objective and what is not experienced as objective. This is true for a number of reasons. First, the number of different perspectives or modalities that we can experience—whether actually or imaginatively—varies with object and with circumstances. There are many different ways to know the shape of a lemon, but few ways to know its taste. I easily imagine multiple perspectives on the desk before me, but I struggle to imagine even one other perspective on the dark fleck that just crossed my visual field. As a result, we experience the lemon’s shape as more objective than the lemon’s taste, and we experience the desk as more objective than the fast-moving fleck. Second, our ability to coordinate the different perspectives we remember or imagine also varies quite widely. If you show me photos of a bird that crossed my visual field—photos that enable me to imagine what it looks like from many different perspectives, this will only enhance my experience of its objectivity insofar as I am able to imagine how those other perspectives fit together with my present perspective; I must have a sense of how different perspectives converge on a single bird. The more coordinated the presented alternatives, the more evident the objectivity. Third, there are differences in the extent to which we can take on (in imagination or in reality) different perspectives simultaneously. Even if I understand how my current perspective on the bird is related to the perspectives of the photos you show me, it may be harder for me to conjure up those other images while observing the bird than it is for me to imagine other perspectives while observing my desk.41 Insofar as the phenomenology of perception is a phenomenology of objectivity, and insofar as the synthesizing that supports a phenomenology 41 Experiences of objects can also be more or less veridical, depending on how accurately we imagine the alternative perspectives. This is another way in which perception can be a matter of degree, but one that pertains to its status as knowledge, not to its status as perceptual knowledge. O’Callaghan (2007) elaborates on this point in response to Noe¨ and others: “Noe¨ and Cohen and Miskin are right that ordinary experiences frequently include veridical perspectival content, but ordinary and not-so-ordinary cases of perception may also be deficient in veridical perspectival content. Seeing with mirrors, periscopes, microscopes (including optical and, for instance, scanning electron microscopes), telescopes, and bifocals involves illusions of perspective on the objects experienced. So does hearing with ear trumpets, parabolic sound disks, two cans attached by string, reflecting surfaces, and other acoustic appliances. Seeing in low light and hearing in noisy environments also degrades perspectival content. It would be dogmatic to deny that perceptual connections exist in these cases. Perceiving is a matter of degree. Experiences can be more or less veridical, more or less illusory.” (p. 157)

38

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

of objectivity comes in degrees, some experiences will be more perceptual than others. As long as our intuitions about clear versus unclear cases accord with the predictions of this analysis, the resultant fuzziness counts for rather than against the analysis offered here. When we wonder about the perceptual capacities of an animal, for example, it makes sense to ask whether it experiences things as objective— as persisting in space independently of any particular position the animal assumes or any particular sensations the animal has. If the animal can receive information from the object in one way only (as an amoeba discerns the concentration of salt in water in one way only), we may wonder whether it has any sensations but we do not regard it as having perceptual experience. If the animal can access the object in a number of different ways and can use those different ways to guide its actions (as a squirrel looks and smells, moves side to side, and finds its way back to a hidden acorn), we don’t hesitate to say that it perceives the object. And if we are unclear about the multidimensionality of an animal’s information processing (as when bees take new pathways back to their hive, or when they discern location from another bee’s “dance”), we are equally unclear about whether (and what) they perceive. Furthermore, even when we are clear about the few ways an animal acquires information from an object (as when a frog tracks a fly), we may consider it a borderline case of perception—sort of seeing and sort of just conditioning. What I hope to have established is an appropriate criterion for the existence of perception, not a determinate verdict about which animals perceive what.

4. A Sensory Requirement? So far, I have tried to show how the phenomenology of perception is a phenomenology of objectivity. But is it something more besides? Isn’t it possible to experience something as objective—the number four, or the reality of God, for example—without one’s experience having the phenomenology of a perception? Does the phenomenology of perception require a sensory component as well? It is common to suppose that perceptions must be acquired through our senses—that states of affairs that do not affect our sense organs, or do not

4.

A SENSORY REQUIREMENT?

39

prompt sensations of some sort, cannot be perceived. At least in part, this is due to the physicalist assumption that the world is a wholly physical world and the empiricist assumption that our knowledge of that world must be acquired through our senses. It is not very clear what constitutes a sense organ or a sensation, however, and it is not clear how the involvement of our senses enters into the phenomenology of perception. Goldie (2007a), for example, defines perceptual belief as “(i) a belief that something has a certain property, from the way it appears, relative to one or more sense modality [my italics]; and (ii) this belief arises in a way that is phenomenologically immediate; in other words it must not be the product of a conscious process of inference.” Mathematical knowledge, according to Goldie, can meet the second requirement but not the first, but it is not clear what is meant by “sense modality” such that mathematical knowledge could never appear “relative to” a sense modality. Brandom (1994) focuses on the epistemic role of perception and defends the view that perceptual beliefs are beliefs that have justificatory immediacy—more precisely, they are able to justify other beliefs without themselves being justified; and then he adds (without any explanation) that the relevant information must be received through sensory organs. It is not clear, though, whether this addition is necessary for the epistemic standing of perceptual beliefs (necessary for the justificatory immediacy of our beliefs) or necessary for their phenomenological character, or both.42 Burge (2010) equates perception with our most basic knowledge of the physical world (versus our most basic objective knowledge of the world of mathematics, for example), and relies on empirical studies of animal psychology to determine how such knowledge is possible. It is no surprise, then, that he regards the involvement of sense organs as a necessary part of perception. He is careful to distinguish objective representation (which may be unconscious) from consciousness (which may not be representational), however, so it is not clear how he views the role of sense organs vis-a`-vis the phenomenology of perception. And another sort of unclarity surrounds the position of Byrne (2007), who claims that sensations form the “core” of perception (and imagination) without, however, being strictly 42

Brandom does not usually concern himself with the phenomenological character of belief, perceptual or otherwise. So when he contrasts the scientist who can perceive the mu mesons in a cloud chamber versus the scientist who can infer but not perceive the mu mesons, it is not clear whether the phenomenologies of these different scientists differ, or whether such differences as there may be have any bearing on their status as perceptions.

40

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

necessary. “There is a continuum of perceptual cases. And—as one might expect—there is a corresponding continuum for the imagination, starting with cases where the ‘sensory core’ of the imagining, in some intuitive sense, determines the imagined content, and proceeding through cases where the sensory core becomes less and less important.”43 Surely it can’t be definitive of perception that only beings with our means of perceiving can count as perceiving the world around them. As Paul Snowden points out, “We have the strong conviction that it is possible for there to be perceptual mechanisms which rely on physical properties quite unlike those around us as earth-bound animals.”44 If a sensory requirement for perception amounts to the requirement that a perceiver must receive and process information from the perceived state of affairs (in whatever way may be possible given the nature of the perceiver and the nature of the state of affairs in question), then it doesn’t amount to anything more than the requirement that a perceiver’s experience must be causally dependent on what is perceived. While there are many controversies surrounding the nature of causal dependence, it would be a mistake to equate the broad notion of causal dependence with the narrower notion of sensory dependence. Put another way: if that is all that is meant by a sensory requirement on perception, then it applies to all kinds of knowledge and cannot help to distinguish between perceptual and non-perceptual experiences of objectivity. What about the suggestion that the phenomenology of perception, unlike the phenomenology of mathematical experience, say, must include sensory content? The focus then shifts to the notion of sensory content. Sensory content must not be equated with bodily sensations since we can experience the presence of a bird on the windowsill, for example, without also experiencing correlated changes in our own bodies. When I see the bird’s black feathers I do not experience “color sensations” or “shape sensations” in my eyes; and when I hear its harsh call I do not experience “tone sensations” in my ears. (It is only when the light is too bright or the sound is too loud that seeing or hearing prompts bodily sensations.) 43

In the online version p. 19. Perhaps the “core” should be viewed as criterial rather than necessary. But that, of course, opens a host of other questions about the nature of criteria. 44 Snowdon (1990), p. 140. He is more willing to view the concept of seeing (versus the concept of perceiving) as tied to the organs and capacities of our particular bodies—to treat it as a “natural kind notion” whose proper explication depends on empirical acts.

4.

A SENSORY REQUIREMENT?

41

Furthermore, even if I were to experience distinctive sensations at the top of my head whenever I contemplate the properties of numbers, this would not have any bearing on whether or not my knowledge of numbers is perceptual. Still, one might insist, we must have some sense of our own position in space in order to experience another object as located in space. When I see a bird on my windowsill I needn’t undergo any distinctive bodily sensations, but I do need to have a sense of where I am located relative to the bird. A sense of my own location might depend on my ability to move through space, and my ability to move through space might depend on bodily sensations (the sensations of proprioception, if nothing else). I am not convinced that we do, in fact, need to experience bodily sensations in order to experience ourselves as moving through space (a sense of continuously changing perspective seems sufficient); but even if bodily sensations were required, proprioceptive or otherwise, these would not bestow sensory content on our perceptions. Rather, they would establish sensory experience of one’s own body moving through space as a precondition for perception of objects outside one’s body. Sensory content might be thought of instead as picking out properties that depend on relations between subjects and objects for their very existence—so-called “secondary” properties. Color properties, for example, are often thought of as existing only in virtue of relevant relations between certain animals and certain light reflectancies; apart from such animals, there would be no color properties. Sound properties, likewise, are often thought of as existing only in virtue of relations that exist between certain animals and certain pressure waves. The phenomenology of perception often involves such properties, but not always: I can discern the shape of a bowl with my hands, for example, without picking up on any of its secondary properties. So if sensory content is content that is responsive to secondary properties, then sensory content is not required for perception. (Neither is sensory content, so understood, unique to perception, for there are many experiences whose phenomenal character tracks our relations to our environment—emotional experiences, for example—without thereby being perceptual.) Consider, finally, the suggestion that an experience has sensory content just in case it is the experience of something (object or property or both) as extended in space. This would allow the experience of a bowl’s shape, as

42

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

described above, to have sensory content even though the bowl’s shape is not a secondary, observer-dependent property. It would allow mathematical contemplation of triangles and number lines to have sensory content as well, however, so the difference between self-evident perceptual experiences and self-evident mathematical experiences could not be the presence of sensory content in the case of perception but, rather, the accuracy or veridicality of that content in the case of perception; the mathematician may imagine a triangle but unless that triangle actually exists it can’t count as a perception. The deeper problem with this understanding of sensory content, though, is that the experience of something as extended in space is already built into our account of the phenomenology of objectivity. If an experience has sensory content just in case it is experienced as extended in space, then self-evident experiences of objectivity are already guaranteed to have sensory content. Recall, from Section 3 above, how our recognition of different sense modalities and our recognition of objectivity are really two sides of the same coin. There are different modalities of experience only insofar as the same states of affairs can be encountered in different ways—and it is precisely our ability to encounter the same state of affairs in different ways that enables us to experience it as an objective state of affairs. Whether a given modality constitutes a sense modality will depend, in turn, on whether we are able to explain its reliability as a source of knowledge by reference to the functioning of distinct parts of one’s body. If we discover that the center of Sam’s brain is sensitive to a particular emission from the nervous system of a spider, that would not constitute a sense modality unless (a) that is just one of several ways in which Sam is able to register the spider’s presence, and (b) other ways rely on other pathways to or through Sam’s body. Even if Sam’s conviction that the spider is present were reliably accompanied by a distinctive sensation in the middle of his head, it would only count as a way of sensing the spider’s presence if Sam could experience it as just one of several ways in which he could track the spider’s presence. On the other hand, as noted above, Sam could recognize certain organs (like eyes) to be sense organs—organs with particular perceptual functions—without experiencing any sensations “in” those organs. The additional of a sensory component can seem like the only way to distinguish the phenomenology of a perception from the phenomenology of a brute conviction. But we have already addressed that difference by

4.

A SENSORY REQUIREMENT?

43

incorporating the experience of objectivity into perception, and by attributing that experience to the active imagining of alternative perspectives and modalities. (The nature and role of such imagining is further elucidated in Chapter 2.) So we do not also need to add a sensory component in order to give perception a distinctive phenomenology. This is good news, because some alternate ways of unpacking the notion of a sensory requirement were problematic—reliance on sensory organs proving to be too minimal, and involvement of sensation or secondary properties proving to be too rich. More importantly, however, it is good news insofar as it offers a unified account of the phenomenology and the epistemology of perception; equating perception with experiences of self-evident objectivity means that the phenomenology of perception is inextricable from its special epistemic function. Other things being equal, a theory that shows how the distinctive phenomenological character of perception gives it its distinctive epistemic role is preferable to theories that treat the phenomenology of perception as a merely contingent “add on.”45 It remains possible, of course, to distinguish between different sorts of perception—some with, and some without, a sensory component. Husserl, for example, insists on widening the concepts of perception and imagination to cover “categorical intuition” as well as “sensuous intuition,” for both play a founding role in our acquisition of knowledge, and both depend on the unification of experience into a single presentation.46 Yet he acknowledges some differences between the two sorts of perception: the objects of sensuous intuition, or sensuous perception, are said to be sensuous objects of “a lower level;” and the acts of synthesis that contribute to a sensuous intuition are said to be less deliberate and less complex than those that 45 As noted in Section 1, above, I assume that perceptual knowledge (like all knowledge) must include true belief; so while I am claiming that the experience of objectivity accounts for what is distinctive about perceptual knowledge, I do not claim that it suffices for perceptual knowledge. I am, however, claiming that experiences of objectivity that are veridical (whether such veridicality is an external addition to, or an internal condition on, the content of those perceptions) suffice for perception; nothing more, and nothing less, is required. 46 Husserl (1913/2001), Section 45, p. 348: “The essential homogeneity of the function of fulfillment, as of all the ideal relationships necessarily bound up with it, obliges us to give the name ‘perception’ to each fulfilling act of confirmatory self-presentation, to each act whatever the name of an “intuition,” and to its intentional correlate the name of ‘object’. . . Plainly the connection between the wider and narrower, the supersenuous (i.e. raised above sense, or categorical) and sensuous concept of perception, is no external or contingent matter, but one rooted in the whole business on hand. It falls within the great class of acts whose peculiarity it is that in them something appears as ‘actual,’ as ‘self-given’ . . . It is clear, in any case, that the concept of imagination must be widened in correspondence with the concept of perception.”

44

PERCEPTION AND THE EXPERIENCE OF OBJECTIVITY

contribute to a categorical intuition.47 These are notable differences— differences that I will be returning to in ensuing chapters—but, as Husserl emphasized, they can best be understood as differences within the category of perception as such. For those who remain convinced that the phenomenology of objectivity must be supplemented by an additional sensory requirement in order to yield perception, the onus is on them to explain just what that requirement amounts to and just how it is related to the epistemic role of perception. Even if those challenges can be met and an additional sensory requirement is embraced, this chapter can be read as providing an account of a crucial element of perception—namely, the phenomenology of objectivity—and that account will help to explain how many different sorts of knowledge can share in this aspect perception—whether or not one agrees to extend the label “perception” on that basis alone.

47 Husserl (1913/2001), Section 46, p. 350: “Sensuous or real objects can in fact be characterized as objects of the lowest level of possible intuition, categorical or ideal objects as objects of higher levels . . . Sensuous objects are present in perception at a single act-level: they do not need to be constituted in many-rayed fashion in acts of higher level, whose objects are set up for them by way of other objects, already constituted in other acts.”

This page intentionally left blank

Figure 2.1. “Rensselaerville Barn” (photo by Daniel Friedman)

Chapter 2 The Role of the Imagination Chapter 1 argued for a certain conception of what perception is—namely, the experience of a thing’s objectivity; and it began to explain how the objectivity of a thing might be evident from within our experience of that thing—namely, through the inclusion of alternative perspectives within a single experience. This chapter elaborates on the process by which alternative perspectives can combine into a single experience, defending the claim that perceiving requires the actual imagining of relevant alternatives. Section 1 offers a (somewhat qualified) “transcendental” argument on behalf of this claim—an argument to the effect that the only (or best) way to explain the phenomenology of objectivity is through the contributions of active imagining. Section 2 addresses the fact that we are largely unaware of such imagining by recounting empirical evidence in favor of its presence, and by highlighting some of the differences between phenomenal consciousness and reflective consciousness. Given the extent to which we use terms like “perception,” “imagination,” and “consciousness” in different, not always compatible ways, I do not think that empirical observations alone vindicate any particular theory of perception, any particular theory of imagination, or any particular theory of consciousness. I do think it is important, however, to explain how a given account fits together with empirical evidence, and that is what Section 2 attempts to do. Section 3 elaborates on the relevant notion of imagining, emphasizing the role of spatial imagery and responding to some traditional worries about overly imagistic conceptions of imagining. In Section 4, I address several concerns regarding the validating role of imagining—concerns as to whether mere imagining can fulfill the justificatory role I attribute to it, concerns about the determination of relevant alternatives, and concerns about the seeming circularity of validating perceptions by appeal to imaginings and validating imaginings by appeal to perceptions. By the end of the chapter, I hope to

48

THE ROLE OF THE IMAGINATION

have established both the need for imagining in perception and the validity of imagining in perception.

1. The Need for Active Imagining Many people can agree that our perception of an object as an object, or a state of affairs as an objective state of affairs, depends on our ability to occupy or imagine alternative perspectives on that object—without supposing that alternative perspectives must actually be occupied or imagined each time we perceive an object.1 Bill Brewer, for example, argues for a close relation between perceiving an object and imagining alternative perspectives on that object, but he describes the requisite imagining as something we must be capable of rather than something we must actually do when we perceive. [The perceiver] is therefore at least in a position to simulate, to some extent, the monadic contents which would be associated with his taking up different points of view upon the same range of particular things, in the particular places they occupy. That is to say, he has the materials to construct, in imagination, the systematically varying monadic contents which he would arrive at by immersing himself in various alternative, possible but non-actual, perspectives upon the same mindindependent things, just where they are around him.2 [italics added] [The perceiver] has the potential, at least, to trip to and fro between a fixed relational conception of where a certain thing is relative to him, and both his actual, immersed monadic impression of its location, and various non-actual possible alternatives to it had he perceived that thing from a different point of view, simulated in imagination. Although his skill in this regard need not be at all well 1 A related claim about how concepts can be present in experience is made by Coates (2007): “ . . . the exercise of concepts in experience has a dispositional aspect. I am prepared for certain kinds of transformation to the sense-image-model I am currently experiencing. The sense in which experience of other members of the same kind are ‘alive’ in my present experience consists in this preparedness . . . I have implicit expectations about the future transformations of my phenomenal experience while I am perceiving the object . . . This is to understand the role of the imagination in a more dispositional way than described by Sellars.” (pp. 177–8) Patricia Kitcher (1990) offers a functionalist interpretation of Kant whereby perceiving an object depends (both causally and logically) on having the capacity to imagine various other representations of that same house—without, however, requiring that such imagining actually be activated while perceiving any particular object (see pp. 148–62). But her analysis fails to take the convergence requirement seriously enough, losing the crucial distinction between “seeing as” and “thinking as” (since a functionalist account of thought is equally insistent on our capacity to think other, related thoughts at other times). 2 Brewer (1998), pp. 23–4.

1.

THE NEED FOR ACTIVE IMAGINING

49

developed, this must at least be possible for him, in the sense that his relational egocentric spatial Ideas are the essential ground for whatever such imaginative routines are actually engaged. [italics added]3

Brewer clearly endorses the claim that experiencing something as an objective state of affairs requires us to recognize the possibility of other perspectives on that state of affairs, and that recognizing these alternative perspectives depends on being able to imagine them. Brewer does not require us to imagine these alternatives when we perceive the state of affairs in question, however; he does not require us to imagine relevant alternatives at the very same time that we enjoy our present perspective on things. But that is precisely what is required for the phenomenology of objectivity (versus an idea or a thought of objectivity). For the independence of an object to be evident from within that experience, the possibility of alternative perspectives must also be evident from within that experience. This section is devoted to a defense of this claim.4 It could be argued that we do not need to imagine alternative perspectives when we perceive because perception already includes multiple perspectives. This could be true in either of two ways, which I shall consider in turn. A perceiver could occupy more than one perspective at the same time—as when I look through two differently placed eyes, from the different perspectives of a constantly shifting gaze, and from the different perspectives of a slightly shifting body. Slight as these differences may be, their importance to the perception of depth is well established and, it might be supposed, they are sufficient to give us the experience of objects in space. This space may not be “well rounded” insofar as it does not give us any sense that an object has a backside, and it may not be a truly “objective” space insofar as it does not give us any sense of perspectives other than our own, but it may suffice

3

Brewer (1998), p. 31. According to Brewer (1998), what results is said to be the “recognition” of objects as such. But “recognition” is an interestingly neutral term; it does not distinguish between the thought of, or belief in, objectivity and the experience of objectivity. I claim that it is only by actively imagining alternatives that an experience of objectivity is possible. This claim does not depend on the flawed assumption (criticized by Millikan (2000) and Dennett (1991) among others) that representing x and y as simultaneous requires simultaneous representation of x and y. Simultaneous aspects of an event may be described in a temporally extended story, or a series of delayed flashbacks, for example. My claim is rather that both x and y (perspectives) must be experienced simultaneously (and thus both actively, not merely potentially, represented) in order to produce the phenomenology of objectivity. 4

50

THE ROLE OF THE IMAGINATION

for the experience of an object “out there” which is accessed from several locations that are “here”—from one’s right eye as well as one’s left, by one’s right hand as well as one’s left, and so on. This seems to be the idea behind the notion of “egocentric space”—a notion invoked by John Campbell and others (including Brewer, above) to describe the way we can situate objects in relation to ourselves without yet having a grasp of their independent relations to each other or to other observers; when we experience space egocentrically, we experience that object as the one that I can touch by extending my hand in this way, the one that I can smell by lifting my head in this way, and so on.5 We are not always occupying multiple perspectives nor attempting multiple modes of access at the same time, however. Reaching out in the dark, I may touch a hard surface and register its location without attempting any other way of accessing the same object. But experiencing its location in space—that is, experiencing the hardness as the hardness of a surface that is independent of me (as opposed to an odd bodily sensation that occurs when I extend my arm too far in a particular direction) depends on recognizing the possibility of alternate perspectives or modes of access. I may not recognize every possible mode of action but I need to recognize some, and that recognition must somehow be present in my current experience in order for me to experience something as objective. Furthermore, even in the case of sight (or hearing) where our two eyes (or ears) present us with two different perspectives on the same object, our egocentric sense of space seems to be much more well-rounded than these slight differences would allow. The bird that I am watching (or hearing) at a distance is not just experienced as existing at a particular distance away from my eyes (or ears); rather, it is experienced as existing at the nexus of a very broad range of movements that would enable me to see it better, hear it better, touch it with my hand or my face, and so on—possible rather than actual movements (on the part of the bird and on the part of myself ). Second, multiple perspectives could be built into our current experience insofar as our perceptual systems are designed (through evolution or through individual training) in such a way as to automatically coordinate 5 See Campbell (1994). Egocentric space is supposed to characterize the way that non-human animals experience space, for example. It is supported by empirical findings and by the supposition that grasping objective space requires us to imagine ourselves from positions “outside” ourselves—something that comes later in human development.

1.

THE NEED FOR ACTIVE IMAGINING

51

different perspectives and different modalities—giving us an experience of objectivity and a spatial independence without those different perspectives and modalities themselves being represented to us in any way. This is the position defended, in detail, by Burge.6 According to Burge it is not we who coordinate different imagined perspectives to produce an experience of independent states of affairs but subpersonal information processing systems that, in response to the demands of the environment, coordinate with each other to produce an experience of independent states of affairs. One thing to note in response to Burge’s position is the difference between tracking things through objective space and experiencing things as existing in objective space. It is certainly possible to design a robot that is able to track things through space without the need for internal representations of multiple perspectives on that thing; but does that suffice for an experience of objects as existing in space? Burge is careful to distinguish perception from consciousness—maintaining that perception is possible without consciousness, and consciousness is possible without perception. His account is an account of perception rather than consciousness, so even if the tracking robot is not conscious of objects in space it might succeed in perceiving objects in space (if the relevant transformations and coordinations of information are in effect).7 Appropriate responsiveness to objective states of affairs may also suffice for perceptual knowledge, according to Burge, as long as it provides the (externalist) entitlement that his account of knowledge requires. Since Chapter 1, however, our concern has been the selfevident (versus externally justified) character of perceptual knowledge, and the phenomenology (versus functionality) of perceptual experience. So even if Burge is right in his characterization of what psychologists mean by “perception,” it does not explain the experience of self-evident objectivity. Another thing to note about Burge’s account of perception is his precarious distinction between information that must not be represented by an individual and information that may or may not be represented by an individual. Burge insists that the rules that govern perspectival shifts—that is, that govern the co-ordination between subsystems—are unlike the rules that govern inference, insofar as they cannot be represented by the perceiver. 6

Burge (2010), see especially pp. 396–416. Burge requires “representation” rather than tracking, but representation is given a functional rather than a phenomenological analysis. 7

52

THE ROLE OF THE IMAGINATION

An important feature of empirical accounts of perception is that the general principles, law, or operations determining transformations among informational and perceptual states need not be represented in the perceptual system (or by the individual) in any way . . . It is worth comparing and contrasting principles governing transformation in perceptual systems with principles and inference rules governing inferential transformation in thought. [In the later case] the individual must have the conceptual capacity to think the principle (have the principle as the representational content of a thought) or specify the rule.8

While it is certainly true that reflection on what we are doing—on how we are managing to see, or how we are managing to think—can get in the way of doing that thing, I see no reason to think that we can’t have introspective knowledge, however imperfect, of at least some of the processing that goes into perception, just as we can have introspective knowledge, however imperfect, of at least some of the processes that go into inference. Indeed, in the next section I offer some suggestions about how such introspective knowledge might be elicited. Another way to explain our experience of the multidimensionality of space without appeal to our imagining of alternative modes of access is this: we believe that our perspective is just one of many, and that belief gives us a sense of spatiality without actually conjuring up any other perspective. Our belief in alternate perspectives may be based on our past experiences of other perspectives, but it may not depend on actively imagining those experiences in order to give us the experience of independent objects in space. The knowledge that there are alternative perspectives can underwrite our experience of objects in space even when none of those alternatives is actually under consideration. Beliefs affect our experience in various ways and in various degrees, of course. There is a difference, however, between a change that is merely intellectual and one that alters the way in which we experience the world— changing not only what we think of the world but the way the world appears to us. My belief that the lights in the night sky come from huge and distant stars may or may not transform the phenomenology of my experience as I gaze out at those lights; it may make the lights look much farther away, and it may make me feel very small, or it may be merely an “add on” belief—a belief that supplements my perception of the night sky without altering it in 8

Burge (2010), pp. 404–5.

1.

THE NEED FOR ACTIVE IMAGINING

53

any significant way.9 Clearly, if our belief that there are alternative perspectives on an object is responsible for our experience of that object as positioned in space, that belief must be more than an “add on” belief; it must infuse the experience itself, giving it a character that it would not otherwise have. The question thus becomes: can the belief in alternative perspectives infuse10 our experience in this way without relying on the imagination to actively imagine alternative perspectives? Some beliefs transform our experience of the world by infusing it with a new feeling or quality. My belief that I am loved, or my belief that death is near, can transform my experience of the world at large by making it seem more inviting, more vibrant, or more precious, for example. Putting aside the question of whether these qualities are really perceived in the world or merely projected onto the world, they alter the way the world appears to me; they are not simply “add on” beliefs. But it is hard to see how a belief in the possibility of alternate perspectives could bestow a distinctive feeling or quality on our experience of the world. First, the belief that there are alternative perspectives on the objects around us does not seem to be correlated with any felt quality. But second, and more fundamentally, it is not possible to believe something of an object without already taking it to be an object, and taking something to be an object already depends on recognizing the possibility of multiple perspectives on that thing. We cannot look to infusions of belief to give us the experience of objectivity, or of spatiality, since beliefs themselves presuppose such experiences (indirectly if not directly). The belief in multiple perspectives cannot precede the experience of objects in space, so cannot be used to explain the distinctive phenomenology of perception. Sean Kelly, developing some themes from Husserl and Merleau-Ponty, has emphasized the role of expectations in our experience of space and time. When we experience objects in space, we expect to move around them in various ways and to make contact with them in various ways; likewise, when we experience events as spread across time, we expect to gain a perceptual grip on some happenings and we expect to lose a perceptual grip on others. These expectations are not expectations that are explicitly 9

For further discussion of this and related distinctions, see Church (2002). Strawson (1974) emphasizes the need for an “infusion” of what is imagined into what is perceived, while acknowledging that the metaphor of “infusion” stands in need of further elucidation. 10

54

THE ROLE OF THE IMAGINATION

entertained so much as they are felt, according to Kelly. Following MerleauPonty, he speaks of a “tension in the experience that feels as though it is about to resolve itself.” 11 Felt expectations or anticipations seem to depend, however, on actively imagining what is to come—imagining that may be either vague or vivid, visual or tactile, full of detail or rather sparse. So the feeling of expectation is not an alternative to imagination so much as its correlate. And the expectation of a resolution across space or time must be the expectation of a change in spatial or temporal point of view. Actively—that is actually, and not merely potentially—imagining alternative points of view thus seems to be necessary for our experience of objects in space. Our present point of view is usually too impoverished to provide a sense of genuine independence and well-rounded space; and the belief in alternative points of view, even if accompanied by a distinctive feeling, cannot give us the experience of objects in space (and time) without the help of an active imagination. It is not possible, of course, actively to imagine every other perspective on a given state of affairs. There are an infinite number of such possibilities, and although the number of possibilities that we are capable of imagining may be infinite, the possibilities that we can actually imagine are surely limited in number. We do not need to imagine all sides of an object, however, in order to recognize it as accessible from many points of view. It is important to keep in mind what is and what is not being explained by appeal to the active imagining of alternative perspectives and alternative modalities. When we imagine multiple perspectives on a single state of affairs, we imagine the convergence of those perspectives on a single location in space, and that gives us the experience of something that exists independently of any one perspective. It does not guarantee the existence of anything, let alone the thing we perceive, at that location; it does not even guarantee a sense of certainty about what exists at that location. The objectivity of what we experience finds support from within that very experience, but the distinction between available justification and truth remains. Indeed, the persistence of such a distinction is what it means for a state of affairs to be objective. The active imagining of alternatives is invoked to explain the phenomenology of perception, not to guarantee its success. 11

Kelly (2005), p. 233. See also Kelly (2003) and Campbell (1997) on “sustainable expectations.”

1.

THE NEED FOR ACTIVE IMAGINING

55

When Wittgenstein emphasizes the impossibility of carrying out (in one’s imagination or otherwise) all possible continuations of an addition function, he is not casting doubt on our ability to experience that function as having an objectively determined continuation—a continuation that may not always correspond to what we conclude when we try to apply it. Nor is he denying that our ability to carry out multiple instances of addition (in our imagination or otherwise) is essential to our knowledge of that function. Rather, he seeks to highlight the indeterminacy of reference that remains no matter how far we continue, given the finiteness of our capacities. Wittgenstein’s concern is not just knowledge of mathematics, of course; it is all knowledge that requires conceptualization, which (for him, anyway) means knowledge of anything at all. So the parallels with our perceptual knowledge of cats and hats, tables and chairs, should not be surprising. We cannot imagine all possible perspectives on such objects nor all possible ways in which they might manifest themselves, but our ability to imagine many such perspectives and many such ways can help to explain our confidence in a thing’s objectivity. Furthermore, I am suggesting, when the relevant possibilities are imagined simultaneously they can help to explain how the objectivity of what we experience can be evident from within a single experience. When we imagine alternative perspectives on an object in space, often (though not always) we also imagine smooth transitions between those perspectives—continuous changes in perspective to match continuous changes in space.12 The possibilities that we imagine can be more like movie clips than like snapshots—imaginings in which one perspective flows into another, and the way that one perspective gives way to another. So even though our active imagining is finite, it can include imagining of continuities that, as such, cover an infinite number of positions and possibilities. This is especially true when our imagining includes the imagining of different actions relative to an object—reaching out to grasp it, lifting it slowly, leaning against it, and so on.13 Clearly, we cannot imagine the infinite number of actions that could be performed on a given state of 12 Note that shifts between perspectives, like shifts between modalities, do not need to be continuous in order to support experiences of objectivity. Note also that if current physics is to be believed, space may not in fact be continuous, however continuous we imagine it to be. 13 Unlike Hurley (1998) and others, I do not maintain that perceiving an object requires that one act towards that object. I do, however, agree that the variations in perspective that are afforded by actions are an important source for the imagining that is required for perception.

56

THE ROLE OF THE IMAGINATION

affairs, but when we do imagine possible actions, we tend to imagine their continuity across space. Husserl distinguishes between different degrees of “adequacy” or selfevidence in our experience of a thing depending on, among other things,14 how completely we are able to represent and to synthesize the different aspects of that thing as revealed from different possible perspectives. If we imaginatively envisage an object turning itself to every side, our sequence of images is constantly linked by synthesis of fulfillment in respect of its partial intentions . . . [T]he whole synthesis of the series of imaginings or percepts represent an increase in fullness in comparison with an act singled out from the series: the imperfection of the one-sided representation is, relatively speaking, overcome in the all-sided one. We say ‘relatively speaking’, since the all-sided representation is not achieved in such a synthetic manifold in the single flash which the ideal of adequation requires.15

This passage is notable for its articulation of three different factors that contribute to the adequacy, and the self-evident truth, of an experience: (1) the number of different perspectives that are envisioned, (2) the degree to which different perspectives are synthesized, and (3) the degree to which the synthesis is compressed into a single moment. Imagining three different perspectives on the table in front of me might be sufficient for me to experience it as an object in space, but imagining ten perspectives adds “fullness” to my experience and adds to the internal evidence of its objectivity. Imagining ten different perspectives flowing into one another (as the object turns, or as we move around the object) effects a synthesis that a series of “snapshots” lacks (even if one knows how each of the snapshots is positioned relative to one another). Other things being equal,16 the greater the synthesis the greater the adequacy, and the greater the sense of objectivity. Put another way: envisioning the transitions between different perspectives—which is not the same as envisioning each intervening perspective—gives the object of our perception a completeness that further 14 Husserl (1913/2001), Investigation VI, Chapter 3, section 23. The two other factors that Husserl cites are the “liveliness” of our representation, which he explicates in terms of their “resemblances” to “the corresponding moment of the object,” and the “reality-level” of the fullness, which is connected with the number of “strictly presented contents.” 15 Husserl (1913/2001), Investigation VI, Chapter 3, section 16, p. 312. 16 The three cited factors may pull in different directions and, as noted in footnote 14, above, there are other relevant factors also.

1.

THE NEED FOR ACTIVE IMAGINING

57

secures our sense of its objectivity. Finally, insofar as a series of perspectives can be synthesized in a given moment—a moment in which we imagine ourselves as being like an ideal observer, seeing all sides at once rather than in sequence—our experience of that thing will be more adequate, its presence more self-evident.17 I think that Husserl is right to think that each of these factors comes in degrees, that each must be present to some degree in order for there to be the experience (and not just the thought) of an objective existence, and that an increase in any one of them will, other things being equal, add to the selfevidence of our knowledge. Let me end this section with a few remarks about how we experience the unity of the different sides of a thing, and how we synthesize different perspectives on a thing. The above passage from Husserl suggests three different types of synthesis, corresponding to the three listed contributions to adequacy. There is the convergence of different sides of an object, or different views of an object, at a single location in space; there is the continuous transitioning between sides, or views, across space; and there is the simultaneous existence of different sides and different views at a single time. It is, of course, possible to believe in each of these unities without actually experiencing them as unities: to believe that three different views are views of the same thing even though we can’t see how they fit together; to believe that there is a smooth transition from hearing a trumpet up close to hearing it far away without being able to imagine the intervening perspectives; or to believe that the many walls of a labyrinth exist simultaneously without being able to imagine them existing simultaneously. Experiencing the unity of different sides, and synthesizing different perspectives, depends on experiencing them as fitting together, simultaneously, within a single space. The synthesis of different sides of an object, or different perspectives on an object, is inextricable from the experience of objects as unified across space and time. The experience of objects as unified across space and time depends on the unification of various perspectives on those objects, and vice versa. The dependence is mutual, with the fact of objective unity contributing to an (metaphysical) explanation of how subjective syntheses are possible, and

17

Husserl (1913/2001), Investigation II, Chapter 4, section 24, p. 159.

58

THE ROLE OF THE IMAGINATION

the fact of subjective syntheses contributes to an (epistemic) explanation of how it is that we can have knowledge of objective unities.18

2. The Evidence for Imagining Still, it will be argued, we are not usually aware of imagining other perspectives as we perceive an object before us. Does it even make sense to speak of unconscious imagining—as opposed to unconscious believing, say, or unconscious information processing?19 Introspection is not the only source of information about our imagining, and it may even be quite unreliable, but it is a good place to start. Is it true, then, that we are unaware of imagining other perspectives when we perceive objects? Certainly it is true that we are not usually focused on the structure of our own experiences when we are attending to an object in space; our focus is on the object of perception—the pitcher on the table before us, or the crack in the side of the pitcher. But a bit of introspection often shows that we are in fact imagining more than what is immediately present to our senses. As we reach towards the pitcher, we are already imagining how it feels, and we are already imagining its backside as being symmetrical with its front. As we contemplate the crack, we are already imagining that it extends through to the inside. One way to prompt an awareness of imagining in the midst of perception is to ask: “Are you imagining the backside as round or square?” “Are you imagining the milk leaking through the crack?” The answers, in most cases, will not be: “I can’t say, since I am not really imagining anything at all.” Nor will the usual answer be: “I wasn’t imagining it either way before, but now that you have prompted me I am imagining it as having a round back, and I am imagining the milk leaking through.” Rather, the answer will come quickly and confidently: “I was imagining the backside as round, like the front” and “I was imagining the milk beginning to ooze

18 This is Kant, without his transcendental idealism—a Kant who doesn’t attempt to construct an objective world out of subjective processes but, rather, seeks to show how the unity of subjective experience is only intelligible on the assumption that there is an objective unity—hence that it is not possible to explain one without assuming the other. 19 For related discussion of the nature and possibility of unconscious imagining, in contrast to both unconscious believing and unconscious information processing, especially in the context of psychotherapy, see Church (2008).

2.

THE EVIDENCE FOR IMAGINING

59

through the crack.” (Our sense that we were imagining one thing rather than another before being asked may be mistaken, of course; but here the burden of proof lies with those who doubt our introspective evidence, not with those who accept it.20) Yet another way to prompt introspective awareness of the imagining that accompanies our perceptions is to present an observer with a surprise—a pitcher that is round on the front and square on the backside, or a crack that is only on the surface after all. Introspection following the surprise will indicate that it was not only the belief in a round backside or the belief in a deep crack that was thwarted; the backside and the crack were imagined as having a particular shape, and their failure to conform to one’s imagining is a source of surprise. So too in the classic example of watching one billiard ball hit another: as the one ball approaches, we actively imagine the different courses the second ball might take, depending on whether it is hit dead center, slightly to the right of center, far to the right of center, and so on. For if our expectations are thwarted—if it follows an unexpected trajectory—we can specify just where we pictured it going instead.21 (The possibility of perceiving causal relations is addressed in more detail in Chapter 3.) Introspection also reveals the way that unavailable perspectives are automatically imagined when certain parts of a scene are obscured from view, or when certain parts of a sequence of sounds are blocked from hearing. When a pillar stands between our seat and the stage, for example, we use our imagination to fill in the obscured events; and when a radio crackles as we listen to music we use our imagination to fill in the missing notes.22 Likewise, when we are shown (in a film, for example) a ball following an impossible trajectory we tend to use our imagination to “correct” our vision; and when we are presented with a nonsensical series of vocalizations, 20 Dennett (1991), takes the opposite approach, placing the burden of proof on those who think there is a difference between the realist versus constructivist interpretations of such things, for he argues that there is nothing to decide between these two interpretations. The dispute then turns on whether the prima facie evidence of introspection can be explained away, not simply ignored. 21 For evidence that infants are often imagining possible outcomes, see Butterworth and Cochran (1980). 22 Note the difference between these cases (where I imaginatively insert an obscured embrace, or a missing note) and the case of blind spots or visual and aural “blinks,” where it may be more plausible to suppose that an absence of information is simply ignored. (For more on the difference between overlooking a gap in perception and filling it in, see Noe¨ (2004), pp. 47–67; he summarizes Dennett’s “perceptual presence” account and offers his “implicit understanding” alternative.)

60

THE ROLE OF THE IMAGINATION

we tend to use our imagination to “correct” our hearing. Frequently, we experience a film as showing us more than it actually does, or we hear a speaker as making more sense than he actually does; we remain unaware of the role of the imagination in shaping our experiences. The contributions of the imagination in such cases are quite familiar to actors and painters and musicians, however, and they are contributions that are relatively easy for anyone to observe introspectively. Is our imaginative “filling out” or “filling in” of a scene something that occurs consciously or unconsciously? Certainly the process of “filling out” or “filling in” obscured sides of an object or obscured details of a scene is seldom a process that is undertaken consciously; it is not the result of a conscious intention. In this sense, the relevant imagining is unconscious imagining. The contents of such imagining, however—the information that is filled in to round out an object or to complete a scene—must be part of the phenomenology of one’s experience. Otherwise it will not be able to explain the phenomenology of objectivity. If one equates phenomenological content with conscious content, then, the contents of our filling-in must be conscious (or, depending on one’s preferred idiom, they must be present to consciousness), whether or not we attend to them—that is, whether or not they are objects of reflective consciousness or introspective access. For our purposes, it doesn’t much matter whether one equates phenomenal states with conscious states, defends the possibility of unconscious phenomenology, or thinks that there are two distinct sorts of consciousness.23 What matters is the contribution of active imagining to the phenomenology of perception. As long as the contents of phenomenology and the contents of introspection are distinguished, introspection will be only one source of knowledge about the phenomenology of experience. In addition to the various sorts of introspective evidence cited above, there is also a considerable amount of non-introspective evidence that suggests that imagination is actively involved in the perception of objects. For example, the ease with which we recognize previously seen objects, even when they are now viewed from alternative points of view, suggests 23 In Church (2008), I defend the possibility of unconscious phenomenal states, noting how, in the case of imagining, this can’t be due to the splitting off of feeling from representational content that both Lear (1998) and Gardner (1993) appeal to in their defense of unconscious emotions. In Church (1998) I argue against Block’s claim that there are two different sorts of consciousness.

2.

THE EVIDENCE FOR IMAGINING

61

that we already imagined (something close to) these other points of view when we first viewed those objects. The alternative hypothesis—that we coordinate different points of view only later, in an effort to match the objects before us with those seen previously—is not only implausible in most cases, given the speed of our recognition; it simply shifts the time of our imagining, without questioning the claim that most of the perceiving that we do—pretty much all perceiving that involves recognition—depends on actively imagining alternative points of view. Another example of nonintrospective evidence for unconscious imagining comes from the observation that even babies lose interest in alternate perspectives on previously presented objects. Insofar as a drop in interest is indicative of previous exposure (a common assumption in studies of infants), these finding suggests that babies have already been exposed to alternate perspectives through their imagining—that they imagined different points of view when the object was first presented, making subsequent occupation of such perspectives less interesting because more familiar. There are many situations in which people’s actions convince us that they are imagining certain things even when they lack introspective access to those imaginings. A small child hides by covering her eyes, clearly imagining that she can no longer be seen, but she can’t explain her action (and is not likely to assent to the suggestion that she believes she can’t be seen). A parent searches through his son’s clothing (anxiously, furtively, feeling along the seams of his clothing), showing that he is imagining a hidden stash of drugs despite his honest (albeit self-deceived) claims to the contrary: “I was just getting his things ready for the laundry.” A friend takes an odd route to the library, and we suppose that she is avoiding the more obvious route because she is imagining an unwelcome encounter along that route— even if she is unable to recall any such imagining. These are all cases where the best explanation of a person’s action seems to require the attribution of unconscious or unreflective imagining. Introspective denials provide evidence against such attributions but not conclusive evidence; the verdicts of observation must sometimes override the verdicts of introspection. Another interesting source of indirect evidence for the role of the imagination in ordinary perception comes from studies of the not-soordinary perception of depressed people. Depressed people not only report that their world has become “flatter;” they actually see three-dimensional

62

THE ROLE OF THE IMAGINATION

objects as flatter than the rest of us.24 Putting this together with studies that show that depressed people generate imagery more slowly,25 and less vividly,26 than non-depressed people, it is plausible to suppose that it is failures of the imagination that are responsible for the relevant failures of perception. If so, these studies provide further evidence of the importance of active imagining for the perception of the world as a fully objective threedimensional world. Finally, the neurology of perceiving a given object and the neurology of imagining or remembering that object from other points of view have so much in common that it is hard to escape the conclusion that perceiving always involves a good deal of imagining as well. The activation patterns specific to any one perceptual encounter are tiny compared with the activation patterns that are shared across multiple encounters with the same object and, more importantly, the activation patterns that are present when merely imagining that object. When I see a bird (or hear its song), for example, much of the resulting neural activity is more closely correlated with past experiences and with imagined experiences than with the present encounter. The sensory input of perception accounts for a minority of synaptic contacts in the cortex while the majority is described as “re-entrant activity”—activity that was originally prompted by other perceptual input, and activity that is associated with memories and imaginings where the occasioning input is no longer present.27 It is usually hard to say, on the basis of neurological evidence alone, just what is being remembered or imagined, but it is a well-established fact that the regular association of one stimulus with another leads to the automatic triggering of one response by the other. (This is the basic idea behind stimulus-response conditioning.) And it is clear that the stimuli resulting from different points of view on an object are regularly associated stimuli; so an encounter from just one point of view is bound to trigger many of the responses that are correlated with associated points of view. More generally, it is clear that each perception we have activates neural activity that is correlated with various associated perceptions. There is no such thing as a perception that is “uncontaminated” by 24

Kunzendorf (2011). Cocude et al. (1997). 26 Sacco and Ruggieri (1997). 27 Llinas and Ribary (1994), p. 114. “[A] large part of the thalamocortical connectivity is organized in what is presently known as reentrant activity . . . or previously viewed as reverberating activity.” 25

3.

IMAGINING AND IMAGERY

63

associations. The neurological evidence, then, only strengthens the claim that the imagination is active whenever there is perception. Let me close this section with a few remarks about the distinction between the “personal” and the “subpersonal.” These terms have been used to distinguish between representations (or non-representational contents) and explanations that are available to the subject of an experience, on the one hand, and representations (or non-representational contents) and explanations that are not available to the subject of an experience. I am not very fond of distinctions that depend on what is possible to explain what is actually experienced—especially given the power of avoidance or repression of various sorts, the sophistication that is often required for accurate introspection, and the still untapped possibilities of biofeedback, for example. Nor am I fond of views that insist on an unbridgeable gap between the personal and the subpersonal. On the other hand, it is clear that explanations that appeal to patterns of neural firing operate at a different level than explanations that appeal to memories, desires, and suppositions, for example. The imagining that I am invoking to explain our experiences of objectivity creates contents that pervade the phenomenology of experience and, as such, they exist at the level of the “personal.” Insofar as such contents are unavailable to introspection, I maintain that not everything that is phenomenologically present to us is introspectively available to us. And insofar as subpersonal, neurological facts provide evidence of imagining while perceiving, they provide evidence merely by virtue of confirming the relevant correlations.

3. Imagining and Imagery More needs to be said about just what imagining is—how (mere) imagining differs from perception and how it differs from (mere) supposition.28 On the one hand, no one thinks that imagining is a matter of perceiving objects that exist inside our heads. If we look inside the head of someone who is imagining a building on fire, there is no building or fire. On the other hand, there seems to be a clear difference between imagining a building 28

For helpful overviews of the history of philosophical debate about imagining, see Russow (1978), Warnock (1976), and Thomas (2003).

64

THE ROLE OF THE IMAGINATION

and merely positing its existence, between imagining a fire and simply supposing that there was a fire. Posits and suppositions seem to lack something of the perceptual character that imagination has.29 The challenge for any account of the imagination, thus, is to make sense of this “perceptual character” of imagining without supposing that the objects and events that we imagine actually exist in some special place—for example, in the head— where they are in fact perceived. (We must also be careful about the potential circularity of explaining perceiving by appeal to imagining while also explaining imagining by appeal to perceiving. I return to this concern in Section 4 below.) What about pictures or images in the head? Could it be that when we imagine buildings and fires we create inner pictures of buildings and inner images of fires? It all depends on what one means by a picture or an image. To the extent that a picture or an image must itself possess the properties that it represents—a blue surface to represent a blue surface, a sharp angle to represent a sharp angle, a rising pitch to represent a rising pitch, and so on— everyone can agree that most of what I imagine cannot be pictured in the head; there are no inner images of that sort.30 But pictures or images can also represent properties they don’t themselves have—as when the picture of a ballerina represents movement without itself moving, or when a curving line represents a rising pitch without itself making any sound. Oftentimes, we take one property to represent another (sometimes through visceral association, sometimes through convention). The real problem with inner pictures, then, does not lie with the fact that brain pictures cannot share in the colors of what is pictured, cannot replicate the shapes of what is pictures, and so on; the real problem is that, in the absence of such similarities, the attribution of brain images doesn’t do any explanatory work. If there were blue squares in my brain whenever I imagined blue squares, and if those blue squares stimulated my optical nerves in much the way that 29 Nagel (1974), in a footnote, further distinguishes between perceptual imagining and sympathetic imagining, as follows: “To imagine something perceptually, we put ourselves in a conscious state resembling the state we would be in if we perceived it. To imagine something sympathetically, we put ourselves in a conscious state resembling the thing itself.” Here we will be concerned with what he called perceptual imagining—without accepting his requirement of consciousness and agency, however. 30 There is some evidence that neural patterns in the visual cortex (as well as neural patterns on the retina) replicate the shape of what is imagined. But this similarity seems to fall far short of what Byrne (2007) has in mind when he claims that imagining, unlike conceiving, has a phenomenology that is intrinsically tied to its object via a mental image with the same properties as its object.

3.

IMAGINING AND IMAGERY

65

actual blue squares stimulate my optical nerves, then the presence of inner blue squares might explain the perceptual character of my imagining. But given the fact that there are no such blue squares in my brain, there is nothing to be gained by positing some other sort of inner picture, for then it is no longer the properties of the picture that explain our response but, rather, the properties of our response that explain the picture. An inner picture account of imagining is no better off than an inner object account of imagining. It would be a mistake, though, and an overreaction, to simply abandon the distinction between pictures and sentences, or between picturing and describing. Block (1981) rightly emphasizes that whether imagining relies on images or on sentences all depends on specifying the relevant similarities and differences. I have already noted the essentially spatial character of the imagining that supports perception, but more needs to be said to distinguish it from inner pictures on the one hand and from mere description on the other. The notion of psychological “schemas” or “models” is sometimes used to characterize mental representations that have some but not all of the properties of mental images. N. J. T. Thomas, drawing on the work of Neisser and others, defines a schema as “a data structure, implemented in the brain, that functions to govern perceptual exploration of the world so that appropriate perceptual tests are applied at appropriate times and places, and that is continuously modified or updated by the results returned by those tests so as to be able to govern perceptual exploration more efficiently in the future.”31 JohnsonLaird, preferring the term “models,” notes that models, like propositions and images, is a higher-order category than the string of “symbols” by which the brain operates. Models, he claims, differ from propositions in having a structure that is analogous to the represented state of affairs. But unlike images, models do not correspond to any particular view of things; a model underlies an image; it must be specific, but can be used to represent a general class if one treats the model as no more than a representative sample from a larger set.32 If we are not aware of any mental images, perhaps it is schema rather than images that are active when we are perceiving objects? 31 Thomas (2003) and Neisser (1976). This seems to be the idea behind Campbell’s (1997) use of the term as well. 32 Elaboration of this point can be found in Johnson-Laird (1983), pp. 157–8.

66

THE ROLE OF THE IMAGINATION

No one can quarrel with the suggestion that there are data structures in the brain that guide our imagining; and offering higher order descriptions of these structures that fall somewhere between images and propositions seems entirely reasonable. Insofar as schema, or models, do not correspond to any particular view of things, however, and insofar as the specific state of affairs that they represent is not the state of affairs that one actually confronts (but, rather, a representative sample from a larger set), then it is not clear how the possession of a relevant schema or model can give us the experience of this object as having other sides. Put another way, it is only by actually generating the images of alternative perspectives on this object that schema or models can enable us to experience objects as existing in space. Thus, while the attribution of unconscious schema may be less contentious than the attribution of unconscious imagining, we need to appeal to images and not just schema in order to explain the phenomenology of perception. Before offering something better, a couple of clarifications are in order. First, it should not be assumed that all imagining is visual imagining— imagining in the modality of sight. We can imagine the taste and the smell of honey, we can imagine the feel of a granite countertop, and we can imagine the sound of fireworks at a distance. To the extent that it makes sense to speak of mental images of a rising sun or an absent friend, it also makes sense to speak of mental images of a sweet smell or a crackling sound; and to the extent that positing inner images is misleading in the case of visual imaging, it is misleading in the case of other forms of imagining as well. For our purposes, though, it will be useful to focus on visual imagining since it is usually easier to imagine alternative perspectives on what we see than on what we taste or what we hear. It is easier to imagine what the backside of an apple looks like than to imagine what the apple would taste like on different parts of one’s tongue, and it is easier to imagine what a train would look like from farther away than to imagine what it would sound like from farther away.33 33 To the extent that this is right, we will experience the objects of vision as more objective than those of taste or hearing. The deeper issue here is the issue of spatial location—whether the properties that are heard or smelled can themselves exist (and persist) at a distance from (and in the absence of ) the perceiver. Robert Hopkins (1998), p. 183, offers the following insightful observations: “What has emerged thus far is most clearly encapsulated in the following claims about hearing, touch, and vision. All three senses represent things as located, and represent those locations by their relation to some point from which they are presented. But in vision the locations of the objects seen are necessarily represented as distinct from the point of presentation, the point of view. In touch objects are necessarily represented as located in places identical with the point, or points, from which they are presented. In hearing either may occur: objects

3.

IMAGINING AND IMAGERY

67

(Whether it even makes sense to speak of different perspectives on the same taste or the same sound is controversial. According to many, as we move food across our tongue we pick up new tastes rather than new perspectives on the same taste, and as we listen to a distant train we hear new sounds rather than a different perspective on the same sound. I think that we experience tastes and sounds as objective properties, knowable from different perspectives, but that assumption is not crucial to the account imagination offered below.34) It is also worth noting that the contents of our imagining are usually determined by cognitive stipulation as well as phenomenal qualities. This is true whenever I decide to imagine a particular individual, and it is true in many other cases as well. When I imagine my neighbor’s dog Digby, for example, the phenomenal qualities of my imagining do not suffice for fixing its content. Even if I have a vivid memory for detail, and even if I imagine the dog acting in some distinctive ways, other dogs can, and probably do, look and act the same way. A purely cognitive stipulation—“It’s Digby,” or “It’s my neighbor’s dog”—must be added to establish this as a case of imagining Digby rather than simply imagining a short, big-eared dog wagging its tail. (This is not a point about the impossibility of imagining a particular individual as opposed to imagining a type since imagining can portray an object as located in a particular spatiotemporal location relative to oneself, and spatiotemporal particularity is usually enough to ensure the particularity of the object in that location. It is, rather, a point about establishing the identity of an imagined particular.) Cognitive stipulation pervades our imagining more generally, however. When I imagine a dab of blue paint on an office door, the fact that it is paint rather than ink and the fact that it is an office door rather than a bedroom door is not usually given by the phenomenal qualities of our imagining; it is stipulated. Likewise, when we imagine a flock of snowy egrets gathering, their identity as snowy egrets rather than white ibises is usually determined by cognitive stipulation rather than phenomenal details.35

(sounds) heard may be represented as sharing the location from which they are presented, or may be represented as located elsewhere.” 34 It is possible to argue different sides of this question; see Kelly (2003), for example. And it is possible to resist the claim that knowledge must be of independent objects; see Cassam (2007), for example. For empirical studies that tend to support the position I am advocating, see Arditi et al (1988), pp. 1–12 and Sacks (2003). 35 Stenning (2002), p. 62, argues that we need semantics for memory. Kind (2001) argues that all imagining involves an image but not one-to-one correlation between an image and what is imagined.

68

THE ROLE OF THE IMAGINATION

Insofar as imagination does involve experiences with distinctive phenomenal content, and not merely stipulative content,36 there must be a way to make sense of such content without appealing to images “in the head.” Current attempts to meet this challenge tend to share the following basic insight: it is possible to simulate or replicate various experiences without simulating or replicating the objects of those experiences (without creating simulacra or replicas “in the head”).37 Just as I can simulate climbing ladder, or brushing a child’s hair, without the presence of an actual ladder, or an actual brush and child, so too I can simulate looking at a neighbor’s dog or looking at an office door without the presence of any dog or any door. One sort of simulation—acting without props—is carried out in external behavior while another sort is carried out through internal processes only, but the idea of simulation in the absence of appropriate objects is common to both. The short answer to the question about how imagination acquires its phenomenal content without inner images, then, is this: imagination simulates actual encounters with objects and, through such simulation, acquires (many of ) the phenomenal qualities of such encounters. We are able to recreate or generate processes that are similar to the processes that result from encounters with physical objects—similar enough to produce experiences with many of the same phenomenal qualities as those produced by actual encounters. Let us now consider the phenomenal similarities between actual visual encounters and imagined visual encounters. Both produce an immediate and unavoidable awareness of an object’s spatial properties. If I (visually) imagine a lizard in my room (versus merely supposing that there is a lizard in my room), for example, I immediately and unavoidably imagine it as having a certain shape, I imagine its appearance from a particular angle, and I imagine its location relative to the furniture in the room. These spatial

Kung (2010) defends the claim that the phenomenal content of imagining is epistemically useful while stipulative content is not. 36 It is fine if one wants to use the word “imagination” to cover stipulative or “propositional imaginings” as well. See Currie and Ravenscroft (2002); also McGinn (2004). The sort I am interested in, however, is the sort that they call “perceptual imagining.” 37 Interestingly enough, this basic claim is common to Ryle (1949) and to Currie and Ravenscroft (2002)—theorists with otherwise quite divergent views. See Currie and Ravenscroft (2002), pp. 31–3, for their discussion of Ryle. Simulationist accounts of imagining have sometimes been presented as part of an “adjectival theory” of mental events, but the basic insight is independent of any particular linguistic theory.

3.

IMAGINING AND IMAGERY

69

properties may be imagined only vaguely and only partially, of course, but they may also be perceived only vaguely and only partially (if the light is dim, if the lizard is moving quickly, if I only saw it out of the corner of my eye, if I am distracted by fear, and so on). Much has been made of the supposed indeterminacy of the objects of imagination versus the objects of perception—the fact that the number of spots, or the tail thickness, or the eye color of an imagined lizard remains indeterminate, while these details are always determinate in the case of a perceived lizard.38 But this contrast only emphasizes the obvious fact that imagining an object doesn’t suffice to make it real; it does not rule out the existence of mental images (since images, like pictures, will leave many details indeterminate as well), and it certainly doesn’t demonstrate a defining difference between the phenomenology of imagining and the phenomenology of perception. The number of spots, the tail thickness, and the eye color of an actual lizard (versus a merely imagined or pictured lizard) are of course determinate, but our perception of an actual lizard seldom gives us this determinate information.39 (Pictures and imagining can even, at times, be determinate about more details than we get from perception—especially when the object is extremely small or extremely far away, for example.) Do we imagine objects as having spatial properties without imagining them as appearing in any particular location? Colin McGinn has claimed that the phenomenology of imagining is different than the phenomenology of perceiving because (among other things) imagining is not subject to restrictions on the visual field: “boundaries” that are “imposed by the constraints of optics and retinal anatomy;” and because “the imagined object is not presented as in some definite spatial relation to the perceiver—that is, typically, in front of him. The content of the image does not inform us as to where the object is in perceiver-centered space.”40 These are two separate claims and ought not to be conflated. I am doubtful about the first claim, which concerns visual field boundaries, since empirical studies (as well as my own introspection) have shown that imagining does mimic something like a visual field, with some objects situated closer to 38

A related claim, defended by Brewer (1999), for example, is that perceptions have more “informational richness” and more “cross-modal integration.” 39 This is not just a point about what makes its way into consciousness; much of the information that is available to our senses is not registered at all. 40 McGinn (2004), pp. 22–3.

70

THE ROLE OF THE IMAGINATION

center “stage” and others so far off to the side, or so far in the distance, as to be barely discernible.41 When I imagine my study, for example, certain pieces of furniture are peripheral to what I am focusing on, closer objects seem to obscure more distant objects, and so on. The fact that my imagining is not restricted by blinking eyelids or a blind spot (two details cited by McGinn) is surely irrelevant to the phenomenology of imagining versus perceiving since (a) my eyelids or my blind spot are seldom part of the content of my perceiving, and (b) it would be quite easy to include the presence of eyelids and blind spots in my imagining as well. I find the second claim, which concerns the supposed lack of a relation to a perceiver in imagining, even odder. Surely this is not always the case, as I can imagine a lizard now sitting just to the left of my knee, for example. More importantly, though, McGinn accepts that we imagine scenes as having a certain depth and as including foreground/background distinctions, and it is not clear how this is possible without “the content of the image [informing] us as to where the object is in perceiver-centered space.” In a footnote, he adds “Of course, I may have beliefs about the location of objects I am imagining, but the image does not intrinsically specify any such location. The visual field, by contrast, indicates precisely where in (egocentric) space the objects of vision are located; it needs no belief supplementation to contain such information.” But if we assume that the visual field of perception shows us where objects are relative to our perception of them precisely by displaying such things as depth and foreground/background relations, why should imagining require supplementary beliefs to do the same?42

41 A Steve Kosslyn (1994) study shows that an imagined object disappears around the edges of an imagined visual field when we imagine ourselves at about the distance where it would disappear around the actual visual field. Currie and Ravenscroft, (2002), p. 71: “[T]here are distinctively visual ways of representing and transforming information and visual imagery [that] enable us to represent and transform in just those visual ways. To get more or different information about something while looking at it, we may have to move towards it, or rotate it, or view it in a different part of the visual field. Visually imagining things is partly characterized by being a form of imagining that mirrors these modes of transformation.” And on p. 73: “Modes of perceptual access like sight have characteristic patterns of facilitation and constraint; the success with which information about an object can be obtained by looking at is substantially determined by the object’s size, shape and orientation, and by the distance and point of view of the observer. The same patterns often show up in visual imagery.” 42 The fact that imagining can be mistaken for perceiving (in the case of psychotics, for example), and the fact that perceiving can be mistaken for imagining (when confronted with a sufficiently faint image of what one is asked to imagine, for example) offer further evidence of their phenomenological similarity. See Segal (1970), cited in Currie and Ravenscroft (2002), p. 73.

3.

IMAGINING AND IMAGERY

71

So far, then, although it may be typical to accept more vagueness and indeterminacy in our imagining than in our perceiving, this is not always the case, and we have not identified any essential difference between the phenomenology of imagining and the phenomenology of perception.43 There may be significant differences between what we do (and what we can do) with imagination versus perception, though—differences that might affect the phenomenology of these two experiences. Elaborating on some suggestive passages in Wittgenstein’s Zettel,44 McGinn (2004) writes: “Perhaps the most obvious difference between images and percepts is that images can be willed but percepts cannot.”45 We can control our imagining in ways that we cannot control our perceptions. Clearly, since the objects of perception exist independently of us, we do not have much control over whether and where they exist; and insofar as the object of a perception must exist, and (in many cases) must exist nearby, we cannot control what we see. This does not yet amount to a difference in the experience of perceiving versus the experience of imagining, however; it could, after all, be viewed merely as a difference in the external conditions under which we call a given experience a case of perception versus a case of imagining. It becomes an experiential difference only insofar as we engage with the contents of an experience (whether real or unreal) in an active versus a passive way. Put another way: whether we are actually perceiving a tree or merely hallucinating a tree depends on worldly facts that are outside of our control—a fact of which we may remain quite oblivious; but whether we are controlling our experiences or merely undergoing them is something of which we are (normally) aware. I do not doubt that there is a phenomenological difference between actively creating and passively undergoing an experience. I do not think that this difference is at all neatly aligned with the difference between

43

I have not considered Hume’s simple suggestion that images are fainter than percepts, since my arguments against the indeterminacy claim, if successful, apply to Hume’s claim as well. 44 Wittgenstein (1967) writes: “Images are subject to the will” (ä621), “We do not ‘banish’ visual impressions, as we do images” (ä633), “The concept of imagining is rather like one of doing than receiving”(ä637). 45 McGinn (2004), p. 12. I find his reifying talk of “images” and “percepts” misleading, but the point is easily enough translated into one that contrasts the experiences of imagining versus perceiving rather than the objects of imagining versus perceiving. See Hacking’s (2005) discussion of this terminology. Hacking, like Braun (1991), raises some good questions about McGinn’s claim that we can discover things about the objects of perception but not about the “objects” of imagining.

72

THE ROLE OF THE IMAGINATION

imagining and perceiving, however. I am often quite active in determining which features of an object I perceive (without thereby thinking that those features depend on me), and I am often quite passive in discovering the features of an imagined object (without thereby thinking that the object must be real). Looking across at my bookshelf, I can direct my attention to notice the different shades of blue, to discern the publisher’s symbol on the spine of a book, or to compare the thickness of books on the left with that of books on the right. The phenomenology of such active looking is quite distinct from the phenomenology of passive perceiving—not only in the feeling of control but also in the clarity with which things are seen. Without such active direction on my part, I seldom perceive the different shades of blue (though they certainly have somewhat different effects on my retina), I don’t perceive the publishers’ symbols, and I don’t distinguish the different books sufficiently to perceive their different thicknesses. These contrasts are precisely those that characterize active versus passive imagining as well. Passive imagining does not cease to be imagining any more than active perceiving ceases to be perceiving. If I actively direct my imagining of my mother’s face I have a greater feeling of involvement, and I imagine more details than when my imagining is more passive—the thinness of her nose, the arch of her eyebrows, the white hairs in her eyebrows. Still, one might insist, there is a difference between the feeling of discovering the new details and the feeling of creating the new details—the former being characteristic of the experience of perceiving while the latter is characteristic of the experience of imagining. But when I imagine the shape of my mother’s eyebrows, it feels much more like a discovery than a creation. Certainly, we do not create the objects that we perceive (they are already there), but neither do we create the objects that we imagine (they either exist or not, regardless of our imagining); rather, in both cases, we can exert a certain amount of control over the experiences we have—sometimes more, sometimes less. I do not have total control over my perceptions or over my imagining, and it is not clear how the feeling of controlling one is any different than the feeling of controlling the other—how the feeling of actively searching for a particular book title differs from the feeling of actively imagining that title, for example. I have been arguing against various attempts at distinguishing between the phenomenology of (visual) imagining and the phenomenology of (visual) perception. This doesn’t mean that we (normally) have trouble

3.

IMAGINING AND IMAGERY

73

telling which is which, however.46 There may be differences that are typical, even if not necessary—characteristics that are indicative of one or the other without being exclusive to either.47 And whether we believe a given experience to be a case of perceiving or a case of imagining will have significant implications for how that experience functions within our overall psychology. The most obvious functional difference between perceiving and imagining concerns the difference in their role vis-a`-vis action. Roughly: our perceptions guide our actions, our imaginings do not. The explanation of this difference is not that imagining allows too much indeterminacy for action (for we always act in the face of considerable indeterminacy) nor is it that imagining fails to locate its objects in relation to one’s body (for it is not possible to imagine something in space without locating the position from which it is viewed); the difference is, rather, due to our taking the things that we perceive to be real, and thus capable of being acted on. Things are actually a bit more complicated than this, since we may also assume the reality of some of the things that we imagine, and our imagining may even help to guide our actions with respect to those things. I may imagine the inside of a dark theater, may assume that my imagining is more or less correct, and may direct my steps accordingly; or I may imagine where my cat is hiding, assume that I am correct, and reach out for it. In these cases, the connection between imagining and acting is less direct than the connection between perceiving and acting, for imagining needs the addition of a belief concerning the reality of what is imagined (and where it is imagined to be) before issuing in action, whereas perception needs no such addition; it already “contains” that conviction, as it were. It is sometimes said that imagining (and simulations more generally) run “off line”— severing the usual connections between experience and action. This is right insofar as uncertainty about the reality of what is imagined leads us to hesitate in acting on our imaginings, but it is wrong insofar as it suggests

46 As mentioned above, though, it does sometimes happen that imagining is mistaken for perceiving and perceiving for imagining. 47 This is analogous to the claim that imaginary animals are more likely than real animals to come in large sizes, to be extremely loving or extremely vicious, to have eyes on the backs of their heads, and so on. These are reasonably reliable signs of their imaginary status, but not constitutive (or even criterial) of that status. My concern is not skepticism, as it was for Descartes, but his suggestion about the relative continuity of perceiving versus imagining is another example of an indicative but not defining difference.

74

THE ROLE OF THE IMAGINATION

that we cannot use our imagining to guide our actions with just as much confidence as we use our perceptions to guide our actions. The relevant difference concerns the directness of the connection to action, not the confidence with which the connection is made.48 We are now in a position to say a bit more about what makes imaginings into simulations of perceptions. One requirement for a simulation is that it be relevantly similar to what it simulates, and insofar as the phenomenology of perception is relevant, we have argued that imagining is indeed relevantly similar. Similarity is a symmetrical relation, however, and simulation is not. For in order for one thing to be a simulation of another, it must function as a substitute or stand-in for what it simulates, and while two similar things can substitute for each other on different occasions, or with respect to different tasks, they cannot simultaneously and with respect to the same task substitute for one another. Substitutions serve many different purposes, of course, and one thing can substitute for another in many different respects. Just which similarities are relevant, therefore, will depend then on the purpose(s) of a particular simulation. If we are simulating future weather conditions in order to give residents plenty of notice of impending hurricanes, recreating the feel of the wind is not important; but if we are simulating another person’s state of mind in order to be a more sympathetic friend, feeling what she feels may indeed matter. Simulations typically function as stand-ins for states of affairs or events that are not themselves accessible—because those states of affairs or events do not (yet) exist, as when future weather conditions are simulated on a computer; because they are not (directly) available to us, as when a detective simulates the movements of a murderer; or because dealing with the real thing is too risky, as when medical students simulate a major operation. Simulations give us information, or teach us skills, that are otherwise too difficult or dangerous to obtain. Imagining, accordingly, functions as a stand-in when perception is not possible, when perception is too difficult, or when perception is too dangerous. An athlete might imagine a difficult jump, for example, in order to discover where she is most likely to encounter problems, in order to improve her ability to make the jump, or in order to provide a safer version of the thrill that would accompany the actual jump. 48

Gendler (2006) discusses the emotional effects of imagining.

3.

IMAGINING AND IMAGERY

75

Simulations can be more or less reliable in fulfilling these functions. (If they are too unreliable they will cease to be simulations at all.) Currie and Ravenscroft (2002) propose the following two conditions on the success of imaginatively simulating a mental process: 1. the component states of imagining are counterparts to the component states of the mental process that is being simulated; 2. after the input stage, the imagining only relies on information that already governs the process being simulated.49 Much will depend, of course, on what makes one component a “counterpart” of another; and much will depend on what information in included in the “input stage.” Can parts of the left hemisphere of our brains function as the counterparts to parts of the right hemisphere? Does the memory of a past event (a particular surgery, or a particular jump, for example) enter into our imagining before or after the “input” stage? Does it “already” govern the process (another surgery, another jump) being simulated? Currie and Ravenscroft claim that the above conditions can be met when we are imagining another person’s (propositional) reasoning, but “given our current uncertainty about the extent to which vision and visual imagery share mechanisms, we cannot say with any confidence that visual and other forms of perceptual imagining pass the relevant test . . . that they are cognitively conservative past the input stage.”50 They do not deny the relevance of perceptual imagining for our knowledge of what it would be like to experience situations other than our own, but they are reluctant to call this simulative imagining. I, on the other hand, see no reason to doubt that perceptual imagining can meet the above requirements, properly understood; and the very fact that perceptual imagining can give us knowledge of situations that we can’t perceive indicates that they do indeed function as successful simulations. Let me close this section by considering one more objection to the claim that perception includes the active imagining of alternatives: Why doesn’t our imagining of alternatives interfere with perceiving what is actually before us? J. Michael Young (1988), for example, considers Kant’s view of 49

Currie and Ravenscroft (2002), p. 94. Currie and Ravenscroft (2002), p. 94. They go on to say that perceptual imagining can be nonsimulative yet still important to learning about someone else’s experiences, predicting likely actions, and so on. It is not clear to me, however, why the cases they cite (such as envisioning a descent into a narrower and narrower cave) should not also be considered simulative. 50

76

THE ROLE OF THE IMAGINATION

the imagination and states: “In perceiving a house we do not in fact find ourselves presented with mental images of it as it might appear under various conditions; if we did, the result would most likely be mere confusion.”51 And Wittgenstein makes fun of the suggestion that “someone whose eyes are open sees something that is not there before him, and at the same time also sees what is before him, and the two visual objects don’t get in each others’ way?”52 There is something right as well as something wrong about the idea that perception and imagination interfere with one another. What is wrong, of course, is the idea that the objects of perception and imagination must have distinct locations in order for us to tell them apart. For, clearly, I can imagine a dragon in the very place I perceive a tree to be (with its same outline, even) without thereby confusing the two. (Apart from differences in the vividness and stability of the two “images,” there will be differences in what I take to be their causal origins, and thus their reality.) What is right, however, about the suggestion that the objects of perception and the objects of imagination must not get in each other’s way is the idea that because the objects of imagination seem to inhabit the space before us, imagination will not facilitate perception (rather than interfere with perception, however weak or innocuous the interference may be) unless our imaginings serve to reinforce rather than compete with perception. The imagining we are interested in, then, is the sort that rounds out our perception rather than the sort that undermines it.53 When we see a tree, we imagine distinct perspectives on distinct parts of that tree, and it is the combination of those distinctive perspectives and parts that gives us the experience of the tree as a well-rounded object in space.

51

This worry leads him to emphasize the idea of alternatives rather than the imagining of alternatives (or a capacity for imagining rather than actual imagining?): “To perceive, I have proposed, is to interpret one’s sensible state as the awareness of something that might also appear in other ways and on other occasions. It is thus to link or unite one’s current sensible state with other such states, i.e., with other ways—past or merely possible—in which the thing might appear.” (p. 145) I worry that his notions of “interpretation” and “link” are either too intellectual or they are not explanatory. 52 Wittgenstein (1980), ä1014. This passage is about “seeing-as,” which is different from normal perception in a number of ways. (See Chapter 6, and Church (2000) for further discussion of this difference.) It is also worth noting that Wittgenstein may be mocking a two-image explanation without rejecting a two-image description. 53 See de Vega (1996) for data on perception-imagery interference versus facilitation; also for a reply to Pylyshyn’s argument against the cognitive penetrability of images, p. 212. Currie and Ravenscroft (2002), p. 79, offer a clear summary of some of these findings.

3.

IMAGINING AND IMAGERY

77

Doesn’t it remain the case, however, that alternative perspectives on this tree are incompatible with my actual perspective, so actively imagining those alternatives is bound to interfere with my perception of the tree from this perspective? To prevent such interference and the resultant confusion, it would seem important to know which perspective is my actual perspective, and it would seem important for this knowledge itself to be a part of the phenomenology of my experience, not purely intellectual. Or, put another way: insofar as perception must register the spatiotemporal independence of its objects, mustn’t it also register the spatiotemporal position of the subject who perceives? If registering one’s own subject position meant perceiving oneself as an object in space-time, then we would embark on an infinite regress: perceiving oneself from some further space-time position, itself perceived from yet another space-time position, ad infinitum. On the other hand, if registering one’s own subject position were a matter of thinking that one is perceiving a tree from about four feet beyond its longest branch, say, then perception would seem to involve too much thinking and too little looking.54 A third option—one that has received considerable attention recently,55 suggests that we register knowledge of our own position insofar as we are capable of effectively interacting with what we see; according to this view of the matter, our actions are not the result of registering our position, they are the registration of our position. Surely it is correct to emphasize the close ties between perception and action: it is both an empirical fact and, I would argue, a conceptual fact that the locations from which we perceive and the locations from which we act tend to coincide. (If I could only see the tree from this side, but could only act on it from the other side, I would eventually begin to see (through to) the other side and/or to act on this side after all. That is to say, what began as indirect knowledge would be transformed into direct knowledge.56) It might also be true that we could not distinguish between different perspectives without

54 This is one example of what Burge (2010) criticizes as hyper-intellectualized accounts of perceptual objectivity. 55 Recent attention to this option is largely due to the work of Hurley (1998) and Noe¨ (2004). Two issues of Psyche (2008), vol. 12, issues 1 and 2, were devoted to it, and numerous other discussions have recently appeared in print; see, for example, a long article by Briscoe (2008). 56 Note Brandom’s (1994), pp. 223–4, discussion of scientists’ perception of mu-mesons. Hurley (1998) also describes a number of interesting cases in which such transformations occur.

78

THE ROLE OF THE IMAGINATION

having the ability to manipulate such perspectives with our actions. This does not require us to be acting from any particular perspective at the time of perceiving something, however; we can perceive a cup, even as we can imagine a cup, without acting on it. Nor does it rule out the possibility of actions that are more closely tied to imagined perspectives than to an actually occupied perspective—as when our imagining of the underside of a chair, not our view of its top side, keeps us from sitting on it. In sum, it would be wrong to assume that actively imagining alternative perspectives interferes with receiving information from our actual perspective or acting effectively from our actual location. Both a priori considerations and empirical considerations indicate that imagining other possibilities is not only an essential part of perception and action, but an increase in such imagining often enhances the depth and precision of both perception and action.

4. The Validation of Imagination Imagining alternative perspectives, I have argued, is what accounts for the phenomenology of perception, which is the phenomenology of spatiality. For experiencing an object as located in space depends on an active recognition of the different sides of that object, the different ways it would look from different points in space. Imagining alternative perspectives also accounts for the distinctive epistemic role of perception, which is the presentation of states of affairs as self-evident. For the only way to validate an experience from within that experience is to include different evidential avenues within that very experience. The spatial character of perceptual experience and the self-evidence of perceptual experience can be understood as deriving, then, from the active imagining of alternative perspectives or alternative modes of access. Indeed, the ability to explain both the phenomenology and the epistemology of perception is a large part of what this account has going for it. In Sections 2 and 3 above I responded to some doubts about imagination’s contribution to the phenomenal character of perception. In this section I respond to some doubts about imagination’s contribution to the epistemic value of perception. These doubts can be expressed as questions:

4.

THE VALIDATION OF IMAGINATION

79

(a) How can perspectives that are merely imagined help to establish the objectivity of what one experiences? After all, we can imagine alternative perspectives on a ghost without thereby adding to the epistemic standing of our purported experiences of ghosts. (b) How do we know which alternatives, perspectives, and modalities are the relevant ones to imagine? Why is the view of the chair from the other side of the room relevant, while the view from the moon is not; and why is looking at the chair relevant while listening for the chair is not? Neither looking at the chair from the moon nor listening to the chair seem to be relevant alternatives to imagine, but why? (c) When we imagine alternative perspectives on a thing, aren’t we imagining alternative perceptions of that thing? And doesn’t that lead to a problematic circularity, where perception is explained by reference to imagining, and the relevant imagining is explained by reference to perception? If we assume, for the moment, that one thing counts as evidence for another insofar as it is a reliable indicator of the other, and it is a reliable indicator of the other insofar as there is a reliable correlation between the two,57 then question (a) can be usefully divided into the following two questions: (a-1) Why think that there is a reliable correlation between the way that we imagine alternative perspectives on a thing and the reality of such perspectives? (a-2) How can the reliability of our imagining be evident from within a particular perceptual experience?

To answer (a-1), we must recall the largely automatic nature of the relevant imagining, and we must recognize the extent to which our automatic imagining is directed by past experience. It is not always the most reliable aspects of past experience that stick with us, of course, and correlations that have been reliable across past situations may not be reliable in our current situation. Nevertheless, it is generally true that the hidden sides, the obscured

57 I do not assume that a reliable correlation between two things is the only basis on which one can count as evidence of the other. (X may be evidence for Y because Y is the most likely explanation of X in this particular case, even though X is not usually correlated with Y.) Nor do I assume that a reliable correlation is sufficient for an evidential relation between two things. (The correlation between X and Y may be reliable in such a narrowly circumscribed context as to undermine the evidential value of the one with respect to the other.) I do not think such qualifications disqualify what follows, however.

80

THE ROLE OF THE IMAGINATION

object, and the receiving space that we automatically imagine as we look around us are accurate projections from past experience; and it is generally true that past experience is a good predictor of the future.58 When someone asks me “How are you imagining the other side of that car?” the images that come to me automatically are, for the most part, images that replicate my experience of similar cars from similar angles. By the same token, the less automatic my imagining, and the less extensive my past experience of cars, the less we can rely on the accuracy of the imagining that accompanies my perception of the car. Before trying to answer (a-2), we might wonder whether it needs answering. Why isn’t the actual reliability of automatically imagined alternatives enough to bestow the necessary epistemic standing—to enable the imagined alternatives to constitute for evidence of objectivity. We can’t expect to justify every impression and conviction we have, after all, on pain of falling into an infinite regress. (If our current impressions and convictions are justified by appeal to imagined impressions and convictions, and imagined impressions and convictions are justified by appeal to past impressions and convictions, how are those impressions and convictions to be justified? I return to this challenge below, in my reply to question (c).) Still, if we want to explain the self-evident character of perceptual knowledge by appeal to the alternatives that are imagined as part of the perceptual experience, then the epistemic relevance of that imagining must indeed be evident from within that perceptual experience. Suppose I look out my window and seem to see an oddly shaped blob floating in the distance. To perceive it as a blob already requires some imagining of its backside, and to perceive it as floating in the distance already requires some imagining of a view from closer up. But given that I have never seen such a blob at such a distance, why should I accord any evidential value to the images that occur to me, automatically or not? What is missing from this description is an acknowledgment of how multiple images fit together in a perceptual experience, and how the experience of their coordination or convergence in space becomes evidence of their epistemic relevance. Overall coherence, like past correlations, is no guarantee of truth, of course; but like past correlations, the overall coherence of 58

Whether induction can itself be justified or not, most people would agree that past regularities count as evidence for more of the same.

4.

THE VALIDATION OF IMAGINATION

81

our experience does, generally speaking, constitute evidence for its accuracy. Imagined alternatives thus appear as evidence of objectivity precisely when, and to the extent that, they help to fill out a unified view of things. Putting aside the actual existence or non-existence of ghosts, and putting aside the presence or absence of a belief in ghosts, most of us are unlikely to have perception-like experiences of ghosts because our past experiences are not likely to prompt the automatic imagining that is required for an experience of a ghost as a spatially located and spatially extended object. Furthermore, even if past exposure to ghost stories and ghost movies were sufficient to trigger the requisite imagining, we would probably find it difficult to fit together that imagining and other automatically prompted perspectives on the world around us.59 (Does the ghost block our view of what lies behind, or can we see through ghosts? Does the ghost have definite edges and can they be felt? And so on.) The above answers to question (a) suggest answers to question (b) as well. The determination of relevant alternatives is largely automatic, and is guided by past experience of similar objects. We discover objectivity by recognizing what is invariant across different perspectives and different modalities, and discovering what is invariant is largely a matter of trial and error. The invariance, and the objectivity, of a cloth’s surface is evident from some distances and some angles of vision but not others, from some types of touch and not others, under some lighting and not others. The invariance and the objectivity of a mold spore is only evident through the perspective of a microscope, through contact with an appropriate stain, and under appropriately intense light. As we move around the world, and as we pursue specialized investigations, we learn where invariance can be found and also how invariance can be found. Such learning may or may not be deliberate or reflective; in most cases, though, it leads to shifts between perspectives and between modalities that happen pretty much automatically. We automatically tilt the cloth and rub it a bit harder in order better to perceive its surface; we automatically focus up and down and adjust the light aperture on the microscope in order better to perceive a mold spore. The imagining

59 The presence or absence of overall integration in our imagining, and the possibility of partial or imbedded coherence, is also relevant to understanding the difference between perceiving objects in our immediate surroundings and perceiving objects on a movie screen. See Chapter 4, Section 1 for a discussion of perception via movies, and Chapter 6 for a discussion of seeing movies as fictional.

82

THE ROLE OF THE IMAGINATION

of relevant alternatives follows suit; our automatic imagining conjures up the same perspectives and the same modalities that we have found, in practice, to reveal invariants. Our actual investigative practices, and our imaginative imitations of those practices, are not always optimal. I may be in the habit of inspecting cloth at too great a distance, or rubbing it with too heavy a touch. You may be in the habit of inspecting mold spores at too low a magnification or without proper light adjustment. If so, the alternative perspectives that we automatically imagine when perceiving the surface of a cloth, or the shape of a spore, will most likely be suboptimal as well. Still, they will be perspectives that are relevant to the experience of invariance in that particular object. We do not launch our imagining from scratch, then, but rather from a wealth of past experience—in particular, experience of finding invariants across different perspectives and modalities. Given my past experience with chairs, imagining how my chair looks from three feet off to the side is something that happens pretty much automatically. Imagining how it looks from the perspective of the moon, or how it sounds, is not. If I were to imagine these perspectives—to imagine them automatically, even—they would not add to my experience of the chair as an objective entity because they would not produce any information that could “fill in” or “fill out” my “picture” of the chair. Unlike a being who could see objects a million miles distant, and unlike a being who could discern the shape of things by listening to the shifts in sound that it produces, if I were to look towards my chair from the moon, I would see nothing, and if I were to try to discern my chair’s shape by listening I would fail.60 We turn finally to question (c). If the contributions of the imagination are valid because of their similarity to past perceptions, and past perceptions are self-evident because of the contributions of the imagination, aren’t we caught in a vicious circle with imagination invoked to explain perception and perception invoked to explain imagining? There are two different considerations that ought to diffuse worries about this circularity. The first consideration concerns the possibility of mutual dependencies. It may not be possible to perceive an object without having 60 Under some circumstances, adopting a perspective in which nothing is seen, or nothing is heard, might provide indirect evidence of a thing’s properties—evidence that it is not huge, and is not spinning at high speed, for example—but it is hard to see how the imagining of such perspectives could help to “fill out” our experience of a chair, for example.

5.

CONCLUSION

83

the ability to imagine what it looks like from other points of view, and it may not be possible to imagine what it looks like from other points of view without having the ability to perceive it—because the two skills develop in tandem and neither can be fully actualized without the other. I do not think that first we learn to perceive and later we learn to imagine. Rather, as we become more and more aware of alternative perspectives on the same thing—by imagining and synthesizing perspectives that we do not presently occupy, we are able to experience that thing as something spatial, and something objective—that is, we are able to perceive it. Conversely, as we perceive more and more things in objective space, we become able to imagine more and more perspectives on that thing. The other consideration concerns the distinction between actual and possible perceptions. Actual perceptions, on my account, depend on actual imaginings, but actual imaginings only depend on possible perceptions. To see a tower requires us actually to imagine the backside of the tower, but imagining its backside does not depend on actually perceiving the backside of that tower—or any other tower, for that matter; it is enough to have seen some relevantly similar structures. (Nor does it depend on imagining, of oneself, that one is perceiving the backside of the tower. One can imagine how things look from another spatial position without imagining oneself as occupying that position.61) While any given perception implies the actual occurrence of assorted imaginings, those imaginings do not, in turn, imply the occurrence of still more perceptions—ad infinitum. The regression is stopped by the fact that our imagination is productive, not merely reproductive. Imagination draws on past perceptions and from past perspectives, and then goes on to generate new perceptions and new perspectives.

5. Conclusion This chapter has shown why imagining is necessary for perceptual experience, what the relevant sort of imagining is, and how such imagining can play

61 Or, as Burge (2010) might say: representation of objects as such may depend on a convergence of different perspectives, but those different perspectives do not need to be further represented as perspectives. Williams (1973) and Velleman (2001) are helpful in clarifying the possibility (contra Berkeley) of imagining a situation without imagining oneself (or any other observer) in that situation.

84

THE ROLE OF THE IMAGINATION

a validating role in our experience. With these key parts of the account in place, we are now in a position to move beyond the perception of nearby, medium-sized objects, to see how imagining can play precisely the same role in other domains, enabling us to perceive causes and other sorts of modal relations (Chapter 3), enabling us to perceive distant objects and abstract objects (Chapter 4), and enabling us to see persons as such and artworks as such (Chapters 5 and 6). As we move beyond the most obvious and ordinary instances of perception, the two-directional nature of this book’s argument is again worth emphasizing. On the one hand, I will use the account of ordinary perception that I have defended over the last two chapters to defend the possibility of the less ordinary sorts of perception I consider in the next four chapters. On the other hand, the fact that this account of perception is able to make sense of perception in these further domains (and is able to do so in a more integrated way than competing accounts) counts in favor of it as an account of ordinary perception as well. Thus even if one is not yet convinced by the account as it applies to the perception of ordinary objects, one may become convinced by the account insofar as it succeeds across a wider selection of perceptual experiences.

This page intentionally left blank

Figure 3.1. “Storm on Breakwater” (photo by Amy Church)

Chapter 3 Perceiving Reasons The experience of objectivity was at the heart of the account of perception developed over the last two chapters. So far I have focused on the perception of states of affairs such as approaching crows, buzzing sounds, and nearby spiders.1 In this chapter, I defend the more controversial possibility of perceiving explanatory relations, or reason-giving relations between states of affairs: perceiving that a falling box caused the chair to break, for example; perceiving that the crystalline structure of ice makes it hard; perceiving that a friend smiled because of my remark; or perceiving a child’s suffering as a reason to intervene. Understanding how we can perceive such reason-giving relations turns on understanding how the dependency of one state of affairs on another can be evident from within a single experience, how the dependence of one state on another can enter the phenomenology of one’s experience. Imagining relevant alternatives will again be crucial, but here the relevant alternatives will pertain to different arrangements of objects and their properties rather than different perspectives on the same object. Whereas perceiving an approaching crow requires us to imagine alternative perspectives on that crow, perceiving that a falling box caused the chair to break requires us to imagine alternative arrangements of the box and the chair, and perceiving my remark as the reason for a friend’s smile requires one to imagine alternative utterances and alternative responses.

1 I did not take a stand on just which properties are perceived in each of these cases—in particular, on whether the perceptual content includes natural kinds or properties such as “crow” and “spider.” This is one of the questions pursued by Siegel (2010), who argues that introspective distinctions plus an absence of possible defeaters supports the view that such properties can be part of the phenomenal content of perception. I agree that there is a phenomenal difference between seeing a crow as a crow and seeing it as a big black bird; on my account, this difference is due to the different imagining that infuses each experience.

88

PERCEIVING REASONS

1. Reasons and Explanatory Relations It is sometimes said that X is a reason for Y if X answers a question of the form “Why Y?” If I ask you why the chair broke and you tell me that Jan threw it against the wall, you are citing a reason. If I ask you why the cement is soft and you tell me it is still wet, you are offering a reason. If I ask you why Jill went up the hill and you tell me that she needed water, you are giving a reason. And if I ask you why I should tell the truth and you say that it makes life easier in the long run, you are providing me with a reason. As anyone who spends time around a curious child knows, we can ask “why?” about pretty much anything, and the answer that we give will depend, in part, on the context of inquiry: why Y as opposed to what? In one context it is the fact that the chair was thrown against the wall that is relevant, but in another context (e.g. where we have just watched several chairs being thrown against the wall and only this one has broken) it is the fact that it was poorly made; in one context it is the fact that Jill needed water that is relevant, but in another context (e.g. where we all know that the only well in town is at the top of the hill but have never before seen Jill fetch the water) it is the fact that Jack is sick that is relevant. It would be a mistake, however, to suppose that reasons exist only in the context of a “why?” (or “why not?”) question, and that only those factors that are relevant to the context of inquiry count as reasons. There are reasons for events and actions that no one thinks about (reasons for the falling of an unobserved tree, or reasons for stopping at a particular intersection, for example), and there are reasons that remain irrelevant to the context of inquiry (reasons for smiling, for example, that are not currently of interest). A more defensible view is the view that reasons supply answers to possible “why?” questions in possible contexts of inquiry. The tie between reasons and “why?” questions is non-explanatory, however, since it could equally well be said that “why?” questions are requests for reasons. What we really want to know is: what sorts of relations are we interested in when we ask for a reason? Michael Strevens (2008) argues that a complete theory of explanation has two parts: a criterion of explanatory relevance (which will depend on the context of inquiry) and a domain-specific dependency relation.2 In the 2 Strevens (2008), p. 5. Less common approaches derive dependency relations from explanation relations (Philip Kitcher) or maintain that all dependency relations are explanatorily relevant (Peter Railton). Strevens discusses some of these alternatives on pp. 36–60 of his book.

1.

REASONS AND EXPLANATORY RELATIONS

89

domain of mathematics, for example, the dependency relation might be logical dependency, in the domain of physics it might be causal dependency, and in the domain of biology it might be functional dependency. X is a reason for Y, then, if Y is dependent on X in some domain-appropriate way. There is a long history of distinguishing between different sorts of reasons, or different sorts of dependency. Aristotle, famously, distinguished between four different types of reasons or causes: material, formal, efficient, and final. In the modern period, philosophers stopped referring to all of these relations as “causes” (reserving that term for Aristotle’s efficient causes only), and there continue to be lively disputes about just which sorts of reasons, or which sorts of dependency relations, are fundamental and which sorts derivative. Still, the importance of these different sorts of reasons continues to be recognized in standard typologies of explanation and in explanations that appeal to constituent parts, essences, or purposes rather than causes.3 Furthermore, while it is common to distinguish justifications from explanations, it is not clear just how this distinction should be made, whether the same thing can be both a justification and an explanation, or how justificatory reasons are related to what Aristotle called “final causes.” Since there is no general agreement about restricting the term “reason” to just one of these types of dependency, and since my interest is not in the typology of reasons so much as the epistemology of reasons—in particular, the possibility of perceiving reasons, I will consider three important types of reasons without worrying about whether they exhaust the field of reasons, whether one is more fundamental than another, or whether they all deserve to be called “reasons.”4 The three types of reasons I shall discuss are causes, constituents, and justifications. If I can show that all three of these types of reasons can be perceived (at least sometimes, under some conditions), I will have shown that a wide range of relations that we call “reasons” can be perceived.

3 Nancy Cartwright (1999) provides a typology of explanation that has some interesting similarities to that of Aristotle. Her category of “pattern-subsumption,” for example, echos Aristotle’s “formal cause.” 4 There is an influential school of thought following in the footsteps of Peter Winch, for example, that insists on a sharp distinction between reasons and causes, but their claims remain controversial and, in any case, can be accommodated by distinguishing between causal reasons and justificatory reasons, as I do below.

90

PERCEIVING REASONS

2. Perceiving Causes One of the most common forms of explanation is, of course, a causal explanation. Assuming that causal relations are real rather than illusory, and objective rather than subjective or response-dependent,5 we can still wonder whether causal relations are the kind of relations that can be perceived. Clearly, we can perceive at least some spatial relations, as when we perceive that the cat is on the mat; and we can perceive at least some temporal relations, as when we perceive that the cat stepped onto the mat before it lay down.6 Can we perceive the fact that the cat moved the mat or that its movements influenced the mat’s position? As in the case of Sam’s judgments about a spider (from Chapter 1, Section 2), it is not enough that our judgments about causality are phenomenologically immediate and epistemically reliable; in order to perceive a causal relation, its epistemic reliability itself must be phenomenologically immediate; the objective, or independent reality of the causal relation must get support from within the phenomenology of our experience. But how is that possible? In Chapter 1, we saw how the objectivity of ordinary objects can be experienced through the juxtaposition of multiple spatial perspectives on that object, but multiple perspectives on the moving cat and the moving mat can only show us their objectivity, not the objectivity of the causal relation between them.7 More needs to be said, though, about what a causal relation is before we can decide whether causal relations are perceivable.

5 If causal relations are illusory, then they cannot be perceived. In Michotte’s (1963) famous experiments with moving shapes, one projected shape appears to cause the other to move, even though it is not in fact the cause; and viewers seem to see a causal relation between the two, even though they do not. If causal relations are subjective, as Hume is often read as claiming, then we do not perceive causal relations between objects, but only (at best) within ourselves. See Costa (1989) for a nice review of the current state of debate on this aspect of Hume’s view. Menzies and Price (1993) defend a response-dependent account of perception that treats causal properties as analogous to color properties (causal relations being those relations that tend to elicit certain active responses, just as color properties tend to elicit certain sensory experiences). 6 How we can perceive change across time is itself a topic of philosophical dispute, of course—from Zeno’s paradoxes of motion, to McTaggert’s arguments against the reality of time, to Soteriou’s (2010) suggestion that disjunctive theories of perception are best able to make sense of the perception of time. It is enough for our purposes, though, that perception of temporal change is possible. 7 Dretske (1969) introduces a notion of “secondary seeing” to cover cases where seeing p (in the primary sense of seeing, where things would not look the way they do unless they were so) plus a belief that p entails q leads one to believe q without, however, inferring q from p. This notion is supposed to cover such cases seeing that a screwdriver is magnetized, or seeing that a cake is not yet done given the toothpick result. These are not cases of seeing a causal relation, though—of seeing that the magnetization causes the screw to jump, or seeing that the underdone cake causes the toothpick to stick.

2.

PERCEIVING CAUSES

91

Consider three different analyses of what a causal relation is:8 1. An event C causes an event E just in case C-type events and E-type events are regularly conjoined in space and time, with C-type events preceding E-type events. 2. C causes E just in case there is an appropriate transfer of energy from C to E.9 3. C causes E just in case E would not have happened if C had not happened.10 Not all causal relations must be of the same type, of course, and not all causal relations must be perceivable;11 but all of the above analyses raise challenges for the possibility of perceiving causality. On analysis (1), the possibility of perceiving causal relations depends on perceiving regularities, but regularities extend across long periods of time while perception occurs within a short interval of time. So regularities cannot be perceived in a single moment. On analysis (2), perceiving a causal relation depends on perceiving a transfer of energy, but the transfer of energy occurs at the level of subatomic particles and forces that can’t be perceived through the senses. And on analysis (3),

8 I assume that the relata of causal relations are events; but nothing turns on this. The relata could also be states or facts, for example. The following list is just one way of grouping the many different accounts on offer. The recently issued Oxford Handbook of Causation, edited by Beebee (2009), lists five “standard approaches”: regularity theories, counterfactual theories, probabilistic theories, causal process theories, and agency and interventionist theories. For our purposes, what I say about counterfactual theories should carry over quite easily to probabilistic theories by replacing the imagining of necessary correlations or outcomes with the imagining of likely correlations or outcomes. Agency theories can be set aside insofar as they make causation subjective rather than objective; if agency theories of cause are correct, then causal relations cannot be perceived because they are not objective relations. 9 This is another traditional account, defended by Dowe (2000) and discussed in Collins et al. (2004), pp. 12–15. One of the attractions of this view is that it does not suffer from the over-determination problem of other accounts. A disadvantage is the apparent circularity of explaining cause by appeal to energy transfer and energy transfer by appeal by appeal to cause. 10 This sort of analysis currently dominates the field, largely owing to the work of David Lewis (1986). One of the main problems with counterfactual analyses is the possibility of over-determination—the possibility that an effect would have occurred without the cause because some other cause would have sufficed; Paul (2009) suggests that the effect would actually have been different in such cases, but also points out that over-determination is only a problem for reductive analyses. The relevant counterfactual, as I state it, concerns individual events. But it could also be stated as a counterfactual about types of events (which would make it into a law-based notion of cause). Also popular are accounts according to which C causes E just in case the occurrence of C makes E more probable. 11 Nancy Cartwright has been an influential proponent of the idea that many different causal relations are relevant to different sorts of scientific inquiry. Ned Hall (2006) argues that we operate with two distinctly different notions of cause—one tracking dependence, the other tracking influence.

92

PERCEIVING REASONS

perceiving a causal relation seems to depend on perceiving a non-actual possibility, whereas we can only perceive what is actual. I will briefly mention some ways of responding to the challenges raised by analyses (1) and (2), then devote the bulk of this section to the challenges raised by analysis (3). Helen Beebee (2003) has argued that regularity analyses of causation (type (1) above) do not preclude the possibility of perceiving causes because our current perceptions are often shaped by regularities in our past experiences (or by the reports of others). We now see the cat causing the rug to move— we see the rug’s movement as following with a kind of inevitability— because we have seen so many similar sequences in the past. Furthermore, the fact that we now experience the sequence as inevitable gives us good reason to believe that there are similar sequences of events both in the past and in the future.12 On behalf of energy transfer analyses of causation (type (2) above), it has been argued that we perceive our own power to put our intentions into action as a kind of transfer of energy,13 and/or that we can feel (through our fingers, for example) the transfer of energy from one object to another. If this is right, and if a background of such experiences also shapes our observation of more distant movements, then we may succeed in perceiving causation in these cases as well. I do not think that either of these accounts accurately describes most experiences of causation, however. Furthermore, given the prevalence of counterfactual analyses of causation (type (3) above), it is important to determine whether causation, so understood, could be perceived or not. Recall that the problem with perceiving causes, analyzed counterfactually, was the nonexistence of the relevant counterfactual possibilities. If C causes E just in case E would not have occurred if C had not occurred, perceiving C as the cause of E seems to require that we see the non-actual case in which C does not occur; and that seems absurd. This argument 12 Note that Beebee does not attempt to defeat the skeptic about induction. She considers Peacocke’s (1986) claim that we can’t perceive a cause because a single perception cannot meet the rationality requirement for knowledge—i.e. acquisition via a method that is appropriate to “canonical acceptance conditions,” where “the canonical acceptance conditions for the holding of a causal relation . . . have to do, ultimately, with laws.” As I understand her response, she views the background conditions that enable us to experience one event as causing another as fulfilling the canonical acceptance conditions. I am less clear about her claim that a single experience—e.g. of pain upon touching a fire—can provide a rational basis for a causal judgment, on a regularity analysis of cause. 13 See Reid (1792/2001).

2.

PERCEIVING CAUSES

93

overlooks the crucial role of imagination in perception, however. Earlier I argued that perceiving a state of affairs as objective requires us to imagine multiple perspectives on that thing; we must imagine how it looks from different angles in space or how its properties may be accessed in different ways. Otherwise, we won’t experience it as an object in space at all. But most of these different perspectives and different modes of access are nonactual for the perceiver at the time of perception: if I am looking at a table from a distant corner of the room, I can only imagine what would be seen straight on and close up; and I can only imagine what its shape would feel like. And yet, I have argued, my experience must be informed and “infused” 14 by such merely imagined alternatives in order for me to experience the table and its properties as objective—and thus in order for me to perceive them at all. Likewise, I suggest, the non-actual alternatives that transform spatiotemporal juxtapositions into causal dependencies are alternatives that, when actively imagined, can inform and infuse our current experience in such a way as to make causality perceivable. Consider, for example, what happens as we watch a paddle move through the water. The paddle moves back and the water curls to the side; the paddle moves deeper and a small funnel is formed in the water; the paddle lifts up and drops of water follow it into the air; as it reenters the water, there is a sharp slapping sound; and as it is pulled back, there is a lower sucking sound. We can observe these things as a disconnected array of events but, more typically, we observe them as causally connected—the movement of the paddle causing the swirling of the water, the re-entry of the paddle causing the slapping sound, and so on. Typically, what enables us to see the causal connections, I suggest, is our simultaneous imagining of relevant alternatives: alternatives in which the paddle remains still, alternatives in which the paddle moves forward rather than back, alternatives in which the paddle moves more quickly, and so on. These alternatives show us what depends on what by showing us what varies with what.15 14 Strawson (1974) describes the contents of our imaginings as “infusing” the contents of our perceptions. Wittgenstein (1980), Section 483, writes: “It is as if one had brought a concept to what one sees, and one now uses the concept along with the thing. It is itself hardly visible and yet it spreads an ordering veil over the objects.” 15 Siegel (2009) develops a similar line of argument in defense of perceiving causation: “Gibson suggested that we sometimes perceive ‘affordances’ of things, where these are possibilities for interaction with them—such as the rollability of a ball, or that a flat solid surface would support us . . . If we can perceive affordances, then there will be cases where we can perceive possible—and sometimes merely

94

PERCEIVING REASONS

The imagining that is required in order to perceive a causal relation is exactly analogous to the imagining that is required in order to perceive a physical object. In order to perceive a paddle as a physical object, we must imagine the paddle’s appearance from different perspectives; we must imagine shifts in appearance responding to shifts in space. Likewise, in order to perceive the paddle’s movement as the cause of the water’s swirling, we must imagine how the water’s movements co-vary with the movements of the paddle; we must imagine shifts in one movement in response to shifts in another. In both cases, by imagining relevant variations (variations of perspectives in the one case, variations of movement in the other), the features that are invariant across those changes (the shape of a paddle in the one case, the water curling around the moving paddle in the other case) can be experienced as objective. For it is those features that seem to continue unchanged despite variations in the imagined alternatives. As in the case of perceiving objects, we are sometimes exposed to relevant variations without needing to rely on the imagination. Just as we can reach out and touch a table from two different directions at once in order to locate it in space, we can drop boxes on a table from different directions, and we can drop boxes with different weights from different heights with different levels of force, in order to determine the cause of its damage. Most cases of perceiving causes, however, like most cases of perceiving objects, are not like this; we see that the ice was the cause of a friend’s fall without actually witnessing the relevant alternatives, and we see that one person’s expression is the cause of another’s distress without simultaneously being exposed to other pairings of expressions and reactions. That we have witnessed relevant alternatives in the past—other pairings of ice and falls, and other pairings of facial expressions and feelings—plays a large role in determining just what we imagine and whether our imagining possible—continuations of what we actually see . . . More exactly, we can experientially represent that the ball would have continued in the direction it was moving, had its path not been interrupted . . . If these counterfactuals can be represented in visual experience, then McGinn goes too far if he suggests that no counterfactuals can be so represented.” (p. 532) Siegel goes on to say that perceivable counterfactuals are probably restricted to those that describe “natural continuations or movements of scenes that the subject sees,” and that the projected possibilities must be “closely enough connected” to what the subject actually sees. With this restriction, Siegel’s claim becomes considerably more cautious than my claim—namely, that in perceiving causal relations we imagine (or “experientially represent”) not only one but several possible trajectories of the involved objects. I claim this is required in order to experience the relevant relation as not just possible but necessary (or highly probable).

2.

PERCEIVING CAUSES

95

is relatively accurate (see Chapter 2, Section 4); but those other alternatives are no longer part of our present phenomenology unless they reenter through our present imagining. As in the case of perceiving objects, beliefs or suppositions or entertained hypotheses concerning relevant alternatives can have an impact on our phenomenology; there is something it is like to believe or to suppose or to entertain a hypothesis. Such believing or supposing cannot produce the phenomenology of a perception, however—the phenomenology of an experience in which the objectivity of a state of affairs is self-evident— without activating images of relevant alternatives.16 I might believe the doctor who tells me that a drop in barometric pressure is the reason for the ache in my bones, and that might make the ache seem less painful, but I don’t thereby come to perceive the causal relation between the pressure drop and the ache in my bones. Even if a belief about a causal relation created a distinctive feeling in us (a sensation of force, perhaps17), and even if we were to project that feeling of force onto the objects in question (the dropping box and the table, for example) that would not suffice to give us the experience of an objective dependency. If there is a sense of forcefulness that infuses our perception of causation—the perception of a dropped bomb as the cause of an explosion, for example—it is, rather, a sense of the force by which one event is objectively bound to another, a bond that would hold across various changes in the surrounding circumstances, a sense of the inescapability of one event following upon the other.18 What are the non-actual alternatives that must be imagined when we perceive one thing as a cause of the other? There are alternatives in which the water swirls in a similar way even though the paddle stays fixed—because there is an undercurrent, or because there is a propeller nearby—but these are not the alternatives that we imagine when we perceive the moving paddle as

16

There is certainly a usage of “imagination” which applies to cases in which we merely suppose or entertain a particular hypothesis (“I imagine that you would have enjoyed the concert” or “Imagine what would happen if you refused to pay”). Imagining without images will not fulfill the phenomenological role I am interesting in explaining, however; as explained in Chapter 2 above, the imagining that is required for perception must be imagistic (though not all images need be visual). 17 Hume describes an “impression of reflection” that is a feeling of necessitation (in response to regular sequences of the sort Hume describes), but there is nothing in the phenomenology of a feeling to convey the modal, or counterfactual, character of a necessity. 18 This is not to say that we can’t imagine a dropped bomb that fails to explode, only that we find it difficult given the normal context for such imagining.

96

PERCEIVING REASONS

the cause. The alternatives that must be imagined when perceiving one event as a cause of another are alternatives in which various background conditions remain constant; if the likelihood of an undercurrent or a nearby propeller is slight, then the imagined alternatives should not include these. (Ceteris paribus clauses added to causal laws indicate the need to fix conditions in this way, but it is never possible to provide an exhaustive list of what the relevant conditions are.19) Insofar as many different factors contribute to a given event, we can perceive the causal power of each only by imagining alternatives in which most other factors remain constant. What is perceived as the cause will depend, at least in part, on the particular interests around which our imagination turns.20 If we are interested in predicting forest fires we might imagine variations in the weather, or variations in the ground cover; if we are interested in holding someone responsible we are more likely to imagine variations in the behavior of campers. And insofar as the background conditions are already sufficient for the event in question (i.e. insofar as there is causal over-determination), the factor that we vary in our imagination will not be perceived as a cause. When we perceive one event as the cause of another, the distinction between background conditions and cause is not always simple. When we perceive the swing of the hammer as the cause of the moving nail, not only do we imagine what would happen if the hammer were not swung; we also imagine what would happen if the hammer were swung in some other way—with less force, at a different angle, and so on. Similarly, we do not imagine everything except the hammer as fixed; for as we imagine the hammer swinging with greater force, we also imagine swinging it further back, and holding it a bit differently, and so on. The imagining that supports the perception of causes requires us to entertain many different possibilities, more or less simultaneously; for only then do the relevant dependencies 19 In the words of Yablo (2004), C is the cause of E if and only if had C not occurred “and had suitably chosen other factors been held fixed,” then E could not have happened. 20 Menzies (2009) notes the normativity inherent in determining relevant background conditions, and thus in determining a cause. “[I]t is not unusual for the determination of what counts as the normal course of events or the default values of the variable to be determined on the basis of normative considerations. This may violate the strictures of causal naturalism, but commonsense attributions of causation are steeped in normativity of all kinds . . . For the concept of a default value of a variable is really just a formalization of the concept of a background condition that is part of the normal course of events. All the same, the distinction between cause and background condition falls neatly into place as part of the general conception of a cause as an intervention that makes the difference to the normal course of events.” (pp. 364–5)

2.

PERCEIVING CAUSES

97

become salient. (Again, the situation is exactly analogous to the imagining of multiple perspectives that enables us to experience a state of affairs as objective.) Furthermore, there will be cases in which we fix the background conditions and discover that although the presence of one event is not necessary for another it makes it much more probable—in which case, if it is perceived at all, the causal relation will appear as a somewhat weaker force or tendency, the effectiveness of which may depend on background conditions of which we remain unaware.21 Ideally, we fix background factors fully enough to reveal the cause as both necessary and sufficient for the effect; but, of course, this ideal is never wholly realized.22 Given the amount of work that is being attributed to the imagination in this account, and given the fact that we are often unaware of imagining the alternatives in question, it is worth reviewing some of the considerations in favor of this account. (The relevant arguments were presented in Chapter 2, Sections 1 and 2; what follows is merely a useful review.) The role of imagination in the perception of both objects and causal relations can be argued in a couple of ways—an empirical way and a transcendental way. Introspectively, we can transform our ordinary experiences of objects into experiences of patches of light and color (as though spread across a canvas, for example) by ceasing to imagine alternative perspectives (the backside of the house, the close-up view of a tree, and so on); and we can transform our experiences of causal relations by blocking the infusion of non-actual alternatives into our experience (alternatives in which the paddle moves in some other way, or alternatives in which the paddle is entirely absent). Such exercises help to demonstrate the essential role of the imagination in ordinary perception.23 There is also third-person empirical evidence that we are pretty much always imagining alternative perspectives and alternative 21 This may be the phenomenological counterpart of seeing causes on a probabilistic account of causal relations. Suppes (1970) is a classic defense of probabilistic theories of causation. Glymour (1998) and Reed (1999) are informative on the complex relations between causation and probability. 22 Causal dependencies are often said to be asymmetrical: the effect depends on the cause but not vice versa. If X is (ceteris paribus) necessary and sufficient for Y, however, then Y is necessary and sufficient for X as well. The relevant asymmetry seems to concern the direction of time, and the possibility of effective intervention across time, rather than asymmetry of dependency as such. 23 Some have used this sort of evidence, along with evidence of extensive “filling in” during visual saccades, to argue that we actually see very little of the world, that most of what we see is virtual rather than actual. This is a pervasive theme in Akins (1996). This conclusion leads to some very counterintuitive results, however. Rather than claiming that most of what we see is merely imagined, I claim that most of our seeing is heavily dependent on imagining.

98

PERCEIVING REASONS

arrangements of what we see insofar as we are clearly primed for some developments and surprised by others. We are surprised to discover that what we took to be a building is merely the backdrop for a movie set because we were already imagining alternative perspectives; and we are primed to intervene in some ways rather than others because we have already imagined alternative possibilities that show us what depends on what.24 The central argument of Chapter 2, Section 1, was, however, a transcendental argument to the effect that it is only by imagining alternative perspectives and ways of information gathering that we could perceive states of affairs, because only then could the objectivity of what is seen be evident from within perception. Assuming that this is right, and assuming that general skepticism about the objectivity of our experiences is not a tenable position, it follows that we must be imagining such alternatives. The counterpart of that argument here, then, is an argument to the effect that it is only by imagining alternative states of affairs that we could perceive one event as the cause of another, because only then could the relevant dependencies be evident from within perception. Is it also true that general skepticism about the perceivability of causal relations is untenable? Kant certainly thought so. In the Critique of Pure Reason, he argues that the unity of consciousness depends on experiencing the unity of diverse properties within a single substance; and he argues that experiencing a unified substance depends on experiencing substances as causally connected. For the binding together of properties in a single substance depends on recognizing their common source—recognizing that a single thing (burning gas, for example) is responsible for the blue light, the hissing sound, and the sweet smell. In order to be experienced as properties of a single substance, then, different properties must be experienced as having a common origin or source. Kant extends this argument a step further to show that the unity and continuity of consciousness across time depends on the recognition of a unified and continuous substance from which not only different properties 24 Much empirical work in psychology assumes that (a) quicker responses indicate that certain representations are already formed and unconsciously “present,” and (b) the experience of surprise also indicates that certain representations are already formed and unconsciously “present.” Other thirdperson empirical evidence for the role of imagination in perception comes from brain scans, which indicate that activation of the visual cortex must be added to retinal activation to produce perception. Again, see Chapter 1, Section 4, for a fuller discussion of these findings.

2.

PERCEIVING CAUSES

99

but also different objects derive. The diversity of individual objects across space and time must be experienced as originating in a single substance, or matter, out of which each emerges. Otherwise, as objects appear and disappear, there would not be the unity and continuity in the objects of experience that is necessary to sustain the unity and continuity of a single consciousness.25 Science’s confidence in the conservation of matter, and science’s never-ending search for the most basic constituents of matter, are thereby portrayed by Kant as prerequisites for the very possibility of conscious experience.26 The unity of consciousness thus depends on evident unity in the world of which we are conscious, and that in turn depends on our recognition of a single substance that is the source of the diverse properties and the diverse objects that we observe across time. Continuing for a moment with Kant: From the conservation of matter and the fact of change across time, it is a short step to the conclusion that earlier events cause later events. For if there is no possibility of matter or substance emerging from nothing or passing away into nothing, and likewise no possibility of individual parts of matter or substance emerging from nothing or passing away into nothing, then every new object or state of affairs must be regarded as a transformation of some previous object or state of affairs. Since there is nothing outside the totality of substance with which to effect a transformation of substance, that transformation must emerge from within substance itself—later objects or state of affairs always originating in previous objects or states of affairs. This is enough to establish a kind of temporal dependency between different states of affairs—the dependency of later events on earlier events such that the later events would not have occurred without the earlier events. At best, this Kantian argument establishes the need for causal relations in general, not the need for any particular causal relation;27 and it shows how causal dependencies are necessary for the unity of consciousness across time—not for a world without consciousness. It still leaves plenty of room 25 As P. F. Strawson (1966) points out, however, an overlapping of objects—none of which is itself permanent—seems to be enough to ensure the experience of continuity. 26 With respect to the conservation of matter, see Kant’s (1787) First Analogy of experience. With respect to the impossibility of arriving at the smallest constituent of matter, see the Second Antinomy of pure reason. 27 Note P. F. Strawson’s (1966) objection to inferring the necessity of a given sequence from the necessity that there be some objective sequence; but see Ralph Walker (1978) on alternative readings of Kant’s argument that could validate this inference.

100

PERCEIVING REASONS

for skepticism about the perceivability of any particular cause, and about the objectivity of causality in general. On the other hand, insofar as we can’t but experience the unity of consciousness across time, Kant’s argument insists that we can’t but experience events in the world as causally related. Concretely, this means that we cannot perceive the appearance of a new flower in the garden without also perceiving it as a transformation within some underlying substance (even if we lack any clear idea of the components and powers of that underlying substance); and we cannot perceive a boat occupying a series of positions on a river without perceiving its later position as in some way due to its earlier position (even if our understanding of the relevant causality is quite limited). I am not willing to endorse Kant’s argument to its full extent since I think it is possible to experience a flower as appearing de novo, from nothing. I do think that consciousness depends on experiencing the temporal continuity of at least some objects, however, and that experiencing the temporal continuity of an object requires one to experience its later manifestations as causally dependent on its earlier manifestations. Thus I think there is a transcendental argument (along generally Kantian lines) for the conclusion that we cannot experience the world as objective without also having experienced at least some events in the world as causally related. Even if one rejects such a transcendental argument, however, there are good empirical reasons to think that we must experience some relations as causal. (Certainly, the studies of Michotte have convinced most psychologists that, under certain conditions, we can’t help but experience one thing as the cause of another.28) And since there are also good empirical reasons to think that causal relations are objective relations (the predictive power of causal laws, for example), it would be surprising if none of our purported perceptions of causal relations actually got things right.29 At the beginning of this section we noted three different analyses of the causal relation—the regularity analysis, the energy transformation analysis, and the counterfactual analysis. Adopting a counterfactual analysis of causation, we then offered an account of how imagining counterfactuals enables us to perceive causes. It should be clear, however, that the imagination 28

Michotte (1963). Siegel (2009) adopts a more cautious position, offering a defense of perception-like experiences of cause without making any commitments about the objective accuracy of such experiences. 29

3.

PERCEIVING CONSTITUENTS AS REASONS

101

could play a key role in the perception of causation even if one favored the regularity analysis or the energy transformation analysis of causation. For actually experiencing (versus merely being influenced by, or merely believing in) the regularity of a correlation across time seems to depend on reproducing past events in one’s present imagining (i.e. iconic or episodic versus discursive memory).30 And experiencing something as possessing or containing energy (versus merely being energized, or merely believing that something has energy) seems to depend on imagining at least some of its possible effects.

3. Perceiving Constituents as Reasons Not all explanatory reasons are causes, of course. Some explanations invoke the constituents rather than the antecedents of a state as the reasons for that state—as when we explain the volume of a cube by appeal to the dimensions of its faces, when we explain the shape and hardness of a crystal by describing its atomic structure, or when we explain the difficulty of a conversation by noting its component parts. Our knowledge of a state’s constituents is sometimes a priori—as when we know that a cube has six faces, sometimes a posteriori—as when we know that a diamond is composed of carbon atoms, and sometimes a combination of the two—as when we know a priori that complex conversations tend to be more difficult conversations and we know a posteriori that conversations about religious disagreements tend to be more difficult. Knowing the constituents of a state of affairs does not suffice, however, for knowing the constituents as reasons for that state of affairs; for we can know that diamonds are composed of carbon atoms without knowing that it is the structure of these atoms that explains how a diamond looks and acts, and we can know that a cube is composed of six equal faces without knowing that a cube’s volume is determined by the dimensions of these

30

For discussion of iconic or episodic memory, see Wollheim (1984), pp. 62–127, and Soteriou (2008). It is important to distinguish between the mistaken claim that the representation of two events as simultaneous (or as not simultaneous) requires us to have simultaneous (or not simultaneous) representations of each—versus the correct claim that a present experience of past events requires us to experience both of those events in the present.

102

PERCEIVING REASONS

faces. And just as our knowledge of a state’s constituents may be either a priori or a posteriori, our knowledge of its constituents as reasons may be either a priori or a posteriori. On the one hand, if I know a posteriori a lot about the nature of molecules and their arrangements, I may be able to figure out, a priori, how the properties of diamonds can be derived from the properties of their constituent atoms. On the other hand, I may know a priori that a cube has six faces, but only come to know that the dimensions of the faces determine the volume of the cube by making a long list of measurements and observing the correlations, a posteriori. Like causal explanations, constitutive explanations explain one fact in terms of another by indicating the dependency of one on the other—the dependence of a diamond’s shape on its constituent carbon atoms, the dependency of a cube’s volume on the dimensions of its faces, the dependence of a conversation’s difficulty on its religious topic. Dependencies may be stronger or weaker—tracking necessary connections or merely probable connections, and they may be more or less circumscribed—obtaining under all conditions or merely those that are common to the domain at issue. The relation between the dimensions of a cube’s faces and its volume is necessary rather than probable, for example, and it is circumscribed only to the extent that the applicability of Euclidean geometry is circumscribed. There is a probabilistic rather than a necessary relation between a conversation’s difficulty and its religious topic, however, and that relation only holds under restricted conditions (where the conversational partners are heavily invested in their views, where the religious topic is fairly complex, and so on). Whether strong or weak, whether more or less restricted, it remains the case that Y’s dependency on X is what makes X a reason for Y. Thus, perceiving the constituents of an object as the reason for that object possessing a particular property will require us to perceive that property as dependent on those constituents. I have already argued, in the case of causal dependencies, that we can perceive dependencies to the extent that we can simultaneously imagine the counterfactual alternatives that demonstrate the dependency at issue. To perceive the diamond’s shape and hardness as dependent on its atomic constituents, we must accurately imagine alternative correlations of atomic structures and crystalline shapes; and to perceive a cube’s volume as dependent on the dimensions of its sides, we must accurately imagine cubes with

3.

PERCEIVING CONSTITUENTS AS REASONS

103

different dimensions and different volumes. Not all constitutive dependencies are easy to perceive, of course—even when the dependencies in question are geometrical dependencies. The shape of a diamond crystal is constitutively dependent on particular alignments of carbon atoms, but this is not something that is easily perceived. We must look at diagrams and read explanations and become familiar with various aspects of chemistry more generally before we can envision the chemical structure of a diamond crystal in sufficient detail, across sufficient variations, and with sufficient reliability to be able to perceive how the diamond’s shape and the diamond’s hardness depend on the bonds that are formed between its constituent carbon atoms. The carbon atoms and their bonds are too small to be perceived in isolation, of course, but it is possible (at least for some) to get to the point where the chemical structure of a diamond is not just inferred from its shape, and not just imagined as one might imagine the appropriate diagram of its internal structure—where, instead, one sees through the geometrical planes of a diamond’s surface to the underlying chemistry on which they depend. (For a fuller defense of the possibility of perceiving very small or very large objects, see Chapter 4.) There is a difference between visualizing constituent atoms in a way that reveals their contribution to a diamond’s shape and visualizing constituent atoms in a way that accords with what one might see if viewing them through an electron microscope, for example. The first is most likely a projection based on instructive models or diagrams, while the second is a result of the highly sophisticated interventions and transformations of a specialized machine. Neither results in a perception (versus an instructive image) unless the viewer engages in relevant imagining of relevant alternatives. In both cases, properties of the prompting image (colors, smooth edges, continuous lines)—whether an image from a textbook or the image projected on a computer screen—will not be accurate to the atoms at issue. But in both cases it is (at least for some people, some of the time) possible to employ or to interpret the presented image in such a way that genuine perception of the relevant spatial relations is possible. Note the phenomenological difference between these “seeing-in” or “seeing-through” experiences and the experience of merely believing that certain atomic alignments are responsible for certain shapes and properties, or merely imagining the relevant geometrical structures while contemplating

104

PERCEIVING REASONS

a diamond, without any synthesis of the imagining and the viewing. Likewise, there is a phenomenological difference between merely believing that a cube with two-foot edges will enclose eight cubes with one-foot edges, or merely imagining eight cubes with one-foot edges while looking at a cube with two-foot edges, versus perceiving that eight one-foot cubes fit into the cube with two-foot edges. Seeing into the structure of a diamond, like seeing into the volume of a cube, is not a matter of figuring out, or picturing, something hidden so much as it is a matter of imagining relevant alternatives to the point where the invariants are revealed. So far, we have treated causal reasons and constitutive reasons as two different sorts of reasons—the first relating two independent events across time, the second relating properties of a whole and properties of its constituent parts at the same time. Some reasons, though, are both causal and constitutive. Bart’s desire to greet a friend may cause him to wave or to call out, but being caused by such a desire is also part of what constitutes the wave as a wave (versus a stretch or a tic, for example), or the call as a call (versus a cry or a hiccup, for example).31 Whether all mental explanations are both causal and constitutive, as some functionalists claim, is not important for our purposes. What is important is that the account of perceiving reasons defended above can be extended to cover a variety of mental reasons as well: it is possible to perceive the reason for someone’s gesture because it is possible to simultaneously imagine relevant counterfactual alternatives. In the context of a party, for example, I might imagine how your behavior would have been different if you were uninterested in the people around you, I might imagine how you would have behaved if the person you are greeting were either closer or farther away, or if your vision had been blocked, or if you wanted to warn them away rather than greet them; it is because of the way that your hand movement varies with your interests, as I imagine them across a range of counterfactual possibilities, that I come to see your desire to greet someone as the reason for the gesture you use. Or, to take another case: we may come to see why someone avoids certain kinds of conversation by imagining what she would have done if the conversation had taken some other turn, what she would have said if she were less worried about being rejected, 31

Note how the dependency can work in the other direction as well: the desire to greet someone must cause one to act in certain ways, under certain circumstances; otherwise it will not count as a desire.

3.

PERCEIVING CONSTITUENTS AS REASONS

105

how she would respond if we continued to press a point, or if we pressed it in a different way, and so on. For it is only by imagining such alternatives that we come to see just what depends on what. The accuracy of our imagining, and thus the possibility of actually perceiving (as opposed to merely seeming to perceive) the reason for someone’s behavior may arise out of past observations of others, or it may arise out of reliable projections of one’s own behavior in a variety of circumstances, or both. For one’s knowledge to be perceptual rather than inferential, though, the relevant alternatives must be actively imagined in such a way as to infuse the very look of the gesture (as jaunty and inviting rather than slow and absentminded, as it would be if it were merely a stretch) and the very sound of the conversation (as uneasy and evasive rather than impatient and restless, as it would be if the speaker were merely bored). Indeed, in order to see an action as an action of any sort (versus merely a bodily movement), we must see it as both caused and constituted by a particular desire or fear or hope or plan. We must perceive the gesturer’s want and the avoider’s insecurity as palpable powers in a field of forces that is both causal and constitutive. Beliefs and desires, hopes and fears, may also provide justificatory reasons for a particular action. The desire to greet a friend not only causes a wave, and helps to constitute it as a wave; it also serves to justify the wave. It is not simply the case that (other things being equal) one tends to wave when one wants to greet someone, and (other things being equal) waves tend to be caused by the desire to greet someone; it is also the case that (other things being equal) one ought to wave when one wants to greet someone. Mental reasons, thus, are often causal, constitutive and justificatory.32 Justificatory reasons present some special challenges for the notion that reasons can be perceived, however—challenges to which we now turn.

32 Note how Aristotle’s final causes (like evolutionary explanations) also combine causation and justification. Aristotle thought that all natural states of affairs have a final cause, but it is now more common to claim that only conscious beings have purposes and aims, so explanations that appeal to reasons of this narrower sort are only appropriate when the state to be explained belongs to a conscious being. The underlying thought, here, is something like the following: a future aim can only be efficacious insofar as it is contemplated and acted on in the present, and only conscious beings are capable of such contemplation and action. Or, put another way: it is not the outcome as such, but rather our representation or anticipation of an outcome that is causally effective and therefore explanatory.

106

PERCEIVING REASONS

4. Perceiving Justificatory Reasons Justificatory reasons are reasons for how things ought to be—whether or not they actually are that way. Justificatory reasons could be distinguished either by the type of reasons that they provide or by the type of thing for which they give reasons. Conceived of in the first way, when X justifies Y, the relation between X and Y is a justificatory as opposed to an explanatory relation; it is a relation of warrant rather than cause. Conceived of in the second way, X explains Y and Y is a state of affairs that consists in something being good or right,33 it is not a special sort of relation but a special sort of property that makes it into a justificatory reason. So, for example, if your suffering justifies my intervention, we can think of your suffering as standing in a special sort of reason-giving relation to my intervention or we can think of your suffering as standing in reason-giving relation to a special kind of object—a state of affairs that consists of my intervention having the property of being good or right.34 There is considerable disagreement about which of these conceptions is primary and which derivative but, fortunately, we do not need to take a stand on this question since it is easy to effect translations between the two. I shall, for the most part, adhere to the first conception in which it is the relation rather than the object that makes the justificatory relation distinctive, but everything I say could be quite easily accommodated by the competing conception. Is it possible to perceive relations of this sort?35 Surely it is possible to have immediate and reliable knowledge about what justifies what; many, perhaps even most people, do not rely on reflective inferences in order to 33 More exactly: X explains the fact that Y has more goodness or more rightness than it would have in the absence of X. Y need not be the best possible state of affairs, and it may even be a bad state of affairs, everything considered, but the existence of X must make it better, or more right, that it would be otherwise. 34 Which conception one prefers will depend, largely, on what one takes the property of goodness, or the property of being right, to consist in. If adherence to the justificatory practices of a community makes something right or good, then justification will be more fundamental than rightness or goodness; but if justification only qualifies as justification insofar as it successfully tracks what is right or good, then rightness or goodness will be more fundamental than justification. 35 There is an important, if minority, view that insists that all moral knowledge is perceptual—that no other sort of knowledge works for the sort of thing that morality is. (I discuss this view in Chapter 5, which is devoted to the topic of moral perception.) Moral justification is just one sort of justification, however, and as far as I know, no one claims that all justificatory knowledge is perceptual. The more basic question, addressed here, is the question of whether knowledge of justificatory reasons can be perceptual.

4.

PERCEIVING JUSTIFICATORY REASONS

107

conclude that a child’s suffering is a reason for a parent’s intervention, or that a patient’s suffering is a reason for a doctor’s intervention. Perception requires more than reliable non-inferential belief, however. There are cases in which we automatically, and reliably, know that one thing justifies another—without ever perceiving a dependency between the two. A welltrained doctor may know, immediately and reliably, that a patient’s breathing difficulties justify the administration of oxygen, but that knowledge may lack the phenomenological character of a perception. Consider the difference between these two cases: The Case of Dr K: Dr K has been trained to distinguish the sounds of different breathing abnormalities—type A, B, C, and D (for example: short and fast, with long pauses between inhale and exhale, gasping, and even but faint)—and she has been trained to administer oxygen (ceteris paribus) when and only when the abnormality is of type B or type D. Having been in practice for many years, she makes her judgments about when to administer oxygen automatically, without reflection (perhaps having forgotten the details of her training, and perhaps having strayed somewhat from that training). Asked to justify her judgment, all she can say is “I’ve been a doctor for a long time now, and others recognize me as an expert on this matter. I know that this type warrants the administration of oxygen (and that type does not). I don’t remember just how I learned which sounds warrant oxygen and which don’t, and I don’t remember why, but I know that I am good at making the right decision.” The Case of Dr P : Dr P experiences different types of breathing as having different weights and shapes—heavy versus light, jagged versus round, solid versus gappy, and so on—and she pictures different interventions, including the administration of oxygen, as affecting those weights and shapes in various ways—spreading them out more evenly, smoothing their edges, filling in the gaps, and so on. She recognizes the weight and shape of normal breathing and she is committed (ceteris paribus) to normalizing her patient’s breathing. So when she hears abnormal breathing, she also visualizes its shape and she imagines how its shape would be affected by various interventions; and this imagining shows her that oxygen is what is needed to restore normal breathing.

Dr P’s training may be much like Dr K’s, and the reliability of her judgment may be no better and no worse than that of Dr K. The phenomenology of her judgments is quite different, however. Dr P, unlike Dr K, experiences the relation between certain sorts of breathing and certain sorts of intervention as objective dependencies or necessities. The phenomenology of her

108

PERCEIVING REASONS

experience is the phenomenology of perception. (Whether it also meets the veridicality requirement for perception will depend on just how the details of her phenomenology map onto the physical properties of breathing and its responsiveness to oxygen. Are there actually gaps in the breathing? Does her experience of “weight” map onto the compression of air in the lungs?) Here, as in the case of perceiving objects (Chapter 1), and as in the case of perceiving causes (Section 2, above), the phenomenological character of perception, which is also the phenomenology of objectivity, arises out of the active imagining of relevant alternatives. The case of justificatory relations may seem to present some special problems, however. First, the correlation between an act and its justification often seems less reliable than the correlation between an event and its cause. We can have reasons to act in a particular way (to administer oxygen, for example) yet fail to act on these reasons (because distracted, or because other patients needed it more, for example). And we can act (kindly or meanly) in the absence of any justification (because we have been conditioned to act a certain way, or because our physiology is making us irritable). So, it can be objected, it is not on the basis of observed regularities that we come to see an act and its justification as related; indeed, if people always did what they should do there would be little need for justifications. Second, whereas our interest in causal explanations typically (though not always) concerns either past or present events—events that have already been observed, our interest in justifications typically (though not always) concerns future possibilities— events that are not yet observable. But if the state that is justified is not observed, how is it possible to perceive a relation between it and the state of affairs that provides its justification? I address these two challenges in turn. Causal dependencies, it will be recalled, always assume certain background conditions (ceteris paribus conditions) against which one state of affairs will not occur without the other; and the alternatives that we imagine in coming to perceive causal dependencies are alternatives in which the background conditions remain fixed. The nail would not move without being hit by the hammer on the assumption that it is not surrounded by an unusual magnetic field, and being hit by the hammer necessitates the nail’s movement on the assumption that it is not welded to the surrounding surfaces, and so on. When we perceive the hammer blow

4.

PERCEIVING JUSTIFICATORY REASONS

109

as causing the nail to move, we imagine other possible trajectories of hammer and nail—without, however, imagining variations in the surrounding forces and surfaces. Likewise, one act justifies another only against the background of certain assumed conditions. A doctor assumes that the goal is restored health, and that ministering to this patient is not endangering others. A child’s suffering justifies intervention on the assumption that such intervention will help rather than hurt the child. And so on for innumerable other background assumptions. When we perceive a child’s suffering as a reason to intervene in a particular case, we imagine how the child’s suffering would proceed without our intervention, and how different sorts of interventions would affect the suffering differently; we do not imagine variations in the child’s physiology or the presence of still more needy children, or a threat to be shot if we help. For justificatory reasons, like causal and constitutive reasons, imply the truth of certain counterfactuals (if X is a justificatory reason for Y, then if X were to occur Y should occur as well); and there are ethical necessities just as surely as there are physical and mathematical necessities.36 In each case, though, the relevant counterfactuals and associated necessities only hold ceteris paribus. Even so, the objector may insist that, unlike the causal case where ceteris paribus the hammer swing does ensure that the nail moves, in the ethical case there is no such guarantee that ceteris paribus a threat on a child’s life will ensure intervention. We are not aware of exceptions to physical laws but we are aware of innumerable cases in which people fail to do what they should do. I would like to suggest, however, that facts about what we should do— ethical necessities—are also facts about what we would do under appropriate conditions. After all, it is reasonable to suppose that what we should do must be something we can do—and, indeed, something we would do under

36

See Fine (2002) on the formal parallels between moral necessities and other sorts of necessities. Whether the relevant counterfactuals for X being a justificatory reason for Y include “If not-X, then not-Y should occur (ceteris paribus)” or “If X does not obtain, then it is not the case that Y should occur (ceteris paribus)” is an interesting and complex question. If I lose my justification for a certain action, does that mean I should abstain from that action, or merely that the action is no longer required? These questions are not different in kind, however, from analogous questions about causal relations. As in the case of causal relations, the answer will depend on what is included in the relevant ceteris paribus clauses. Do they rule out competing and defeating reasons, or competing and defeating causes? Do they rule out over-determining reasons, or over-determining causes? Many different options are compatible with what is said here about the perceivability of justificatory relations.

110

PERCEIVING REASONS

the right circumstances. The relevant question thus becomes a question about what circumstances are required in order for us to do the right thing—that is, what conditions would make our actions fully responsive to justificatory reasons for or against those actions. Some will say we would do what we should do if we were fully informed about our own desires and about relevant worldly facts. Others will say that we do what we should do as long as we are properly educated by our community or as long as we adopt appropriately “thick” ethical concepts. Still others will say we would do what we should if our “ears” were open to the word of God, if we were responsive to the claims of reason, if our will were strong, or if we could successfully resolve our Oedipal complex, and so on. As in the case of causal explanations, the relevant background conditions needn’t be normal in the sense of usually obtaining; rather, they must specify the “norm” in the sense of identifying an ideal or standard within which the relevant dependencies do hold and against which deviations can be registered as such. My point, then, is just this: ethical reasons, like causal or constitutive reasons, only support counterfactuals against the backdrop of certain fixed conditions, and as long as the imagined alternatives conform to those fixed conditions, the relevant dependencies can be perceived. As long as I restrict my imagining to cases where my will is strong and my knowledge is complete, for example, I will perceive the threat to a child’s life as necessitating intervention.37 The other worry about the possibility of perceiving justificatory reasons as opposed to explanatory reasons concerns the future status of the justified event. It should be noted that while justifications are more likely to concern not-yet-observed events, and explanations are more likely to concern already-observed events, this is not always the case. We frequently justify our past actions—to our friends, or before the law, for example; and we

37

Brandt (1996) was a prominent defender of basing ethics on the imagining of ideal conditions, where ideal conditions are conditions in which one has full information and complete consistency. Others (e.g. McDowell) have emphasized ideal training, thick concepts, and integration in an ethical community. There is an interesting recent literature, especially Gendler (2000) exploring what a moral person cannot or will not imagine. See also Goldie (2007a), who develops an analogy with chess, maintaining that, with enough practice at imagining and eliminating failing strategies, we can get to the point that only one option remains present in our imagining. But seeing only one option as available is different from seeing a justificatory relation—from experiencing the necessity of a particular relation, which is what I seek to explain.

4.

PERCEIVING JUSTIFICATORY REASONS

111

frequently offer an explanation of something that has not yet happened—an upcoming eclipse, for example. If seeing reasons depended on the past versus future status of the event in question, this would not rule out the possibility of seeing justificatory reasons for events that have already occurred and have been observed. It would not rule out the possibility of me seeing why you are helping your mother up the stairs, for example— because she is in pain, because you care for her, because she needs you (these are all things that can be seen). Finally, though, insofar as perceiving reasons is the result of imagining relevant counterfactual arrangements, it is not clear why we can’t perceive dependency relations between future states of affairs. We often say that we can see what is going to happen, and why—that the teacher is going to slip because of the oil on the floor, that the sun is about to appear because the clouds are shifting, that the pineapple will be ready to eat tomorrow because it is soft now, that the roof will collapse because the rafters are broken—and these reports are often based on experiences that have both phenomenological and justificatory immediacy. On the account I have been defending, as long as such foresight includes the present imagining of relevant dependencies, then our knowledge of these dependencies may indeed be perceptual. Whether this means that our perception of a given state of affairs need not be caused by that state of affairs,38 that future events can be the cause of present perceptions,39 or that future states of affairs can be a constituent of present relations40 is a topic that I return to in Chapter 4 (which specifically addresses the possibility of perceiving future states of affairs). For the moment, it is enough to note that not all justificatory relations concern future events, and that the perception of future events is compatible with a variety of positions regarding causation and perception.

38

See Dretske (1969), pp. 108–9, for example. Dummett (1954) defends the possibility of what he calls “quasi-cause” (“quasi” because not tied up with agency in the way that cause is). He argues for the intelligibility of precognition. Flew’s (1957) response to Dummett argues that the possibility of precognition is incompatible with the presence of memory. 40 It might be argued, for example, that a baby’s future as a human being is part of what is already present and, as such, it is something that can be not only anticipated but perceived. 39

112

PERCEIVING REASONS

5. The “Space of Reasons” For much of the twentieth century, philosophers insisted on a sharp divide between explanation and justification, between is and ought, and between cause and reason. I have argued that it is possible to perceive justificatory as well as explanatory dependencies, and in the previous section I cast doubt on some ideas about what is special about justification. In this section, I want to highlight how, when it comes to characterizing the objects of perception, there are difficulties that plague any attempt to keep justificatory relations and causal relations distinct. After reviewing two recent and influential accounts of what it means for facts to inhabit a “space of reasons” (that of Robert Brandom, and that of John McDowell), I want to suggest a third way—one that acknowledges the normativity of the contents of perception without undermining their independence (as Brandom seems to do) and without making their independence mysterious (as McDowell seems to do). I side with McDowell in locating the objects of reason in the external world (rather than an inner world of merely psychological states of affairs), but I argue against his “two world” solution to the problem of relating reasons and laws, phenomenology and physics. Readers who are not interested in these particular philosophers or these particular disputes may want to skip ahead to the next section. It was Sellars who introduced the notion of a “space of reasons” defined by justificatory relations as opposed to causal relations. The basic metaphor is clear enough: the space of reasons is a “space” in which things that can enter into rational or justificatory relations are situated, and it is a “space” whose contours are determined by the many intersecting paths that justification may take—not by the actual spatial relations between things.41 As a metaphor for the complex and interrelated claims of justifications, the image is apt, and uncontroversial. It becomes controversial, however, insofar as the states of affairs that inhabit this “space” are supposed to be different in kind from the states of affairs that populate the non-metaphorical space that surrounds us.42

41

The picture is one of different justifications emerging from or converging on the same point, with more or less direct ways to justify one thing by appeal to another, and so on. 42 An important source of this insistence can be found in Kant’s distinction between descriptive and regulative uses of reason, where the first requires spatial substances and temporal causes while the second requires us to adopt a standpoint “outside” or “beyond” these restrictions. Sellars, Brandom, and

5.

THE

“SPACE

OF REASONS”

113

Sellars introduces the notion of a space of reasons in the context of his discussion of observational reports such as “This is green.” Such a report can only express a state of knowledge, he argues, if the speaker is able to justify the report by appeal to some further, and more general, knowledge about the reliability of such reports. The essential point is that in characterizing an episode or a state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logical space of reasons, of justifying and being able to justify what one says.43

Sellars here makes a distinction between two different ways of characterizing an episode or a state—a way that indicates its empirical properties, and a way that indicates its availability for justification. An utterance of the sentence “This is green,” for example, may be described as a sequence of sounds resulting from certain bodily changes or it may be described as expressing a state of knowledge. The first characterization locates it within the empirical world while the second “places” it within the “world” of justification—that is, the space of reasons. Brandom’s explication of Sellars’s notion of a space of reasons extends the relevant characterizations beyond the laudatory category of “knowing.” [Sellars] could as well have said that in characterizing an episode or state as one of believing, or applying concepts, or grasping propositional contents we are not giving an empirical description . . . but placing it in the logical space of reasons . . . 44

Brandom’s point is that in the very act of employing concepts we open ourselves to questions of justification and thereby enter into the normative space of reasons. Observational reports such as “This is green” belong to the space of reasons not merely because we regard them as expressions of knowledge but, more fundamentally, because using concepts to make claims (versus merely uttering words) depends on having the ability to make appropriate inferences to and from those claims. [W]hat distinguishes concept-using creatures from others is that we know our way around the space of reasons. Grasping or understanding a concept just is being able

McDowell differ in their interpretations of Kant, but each traces his central insights to Kant. As will become clearer in Chapter 5, my own reading of Kant points in a different direction. 43 44

Sellars (1956/1997), Section 36. Sellars (1956/1997), p. 160.

114

PERCEIVING REASONS

practically to place it in a network of inferential relations: to know what is evidence for or against its being properly applied to a particular case, and what its proper applicability to a particular case counts as evidence for or against.45

The suggestion is that our capacity to recognize one claim as justifying another depends on our capacity to recognize justificatory relations between each of these claims and any number of other claims as well. This is logically (and, I would say, phenomenologically) similar to the way that our capacity to recognize how one object is situated with respect to another in space depends on our capacity to situate both objects in relation to number of other objects as well. McDowell extends the notion of a space of reasons still further, to encompass not only intentional states (beliefs, claims, etc.) but the very contents of those states—the facts that they are about. My belief that the grass is tall gives me reason to believe that it is summer, but the reason for me believing that the grass is tall is not yet another mental state; it is the mind-independent fact of the grass being tall—a fact that I access in perception. The facts that we perceive have a place within the space of reasons whether or not we are aware of those facts. Perception, according to McDowell, is not a process of conceptualizing input from a world that stands outside the space of reasons; rather, it is a receptivity to a world that is already conceptual and already within the space of reasons. That things are thus and so is the content of the experience, and it can also be the content of a judgement . . . So it is conceptual content. But that things are thus and so is also, if one is not misled, an aspect of the layout of the world: it is how things are. Thus the idea of conceptually structured operations of receptivity puts us in a position to speak of experience as openness to the layout of reality. Experience enables the layout of reality itself to exert a rational influence on what a subject thinks.46

As McDowell understands it, it is not the recognition that we are presented with a fact that gives us a reason to believe. The relation is more direct: in being presented with a fact, we are given a reason to believe—whether or not we acknowledge it as a reason (and whether or not we believe what we perceive).47 Thus McDowell adheres to something of a pre-modern 45

Robert Brandom (2001), p. 203. McDowell (1994), p. 26. In replying to Barry Stroud’s comments, for example, McDowell (2002a), p. 278, writes “I think we need an attitude of perception as something in which there is no attitude of acceptance or 46 47

5.

THE

“SPACE

OF REASONS”

115

sensibility in which reasons do not inhabit our minds so much as our minds inhabit a world of reasons. We do not construct or acquire reasons for belief or action, we discover them (and we do not discover them within ourselves, we discover them in the world around us). Consider how McDowell arrives at this expanded understanding of the space of reasons as including facts, whether or not we notice them. He begins his book Mind and World by endorsing Kant’s famous statement that “thoughts without content are empty, intuitions without concepts are blind.”(A51/B75) But McDowell argues for this double-sided claim in a rather un-Kantian way, and his targets are not the eighteenth-century philosophers Hume and Leibniz so much as the twentieth-century philosophers Quine and Davidson. Quine is accused of relying on unconceptualized intuitions (his “tribunal of experience”) to legitimate our web of belief, overlooking the fact that unconceptualized input cannot serve as a reason for anything. If experience were the reception of unconceptualized inputs, it would be “blind” and could not function as a rational constraint on our beliefs (“a bare presence cannot be a ground for anything”).48 Davidson, according to McDowell, corrects for Quine’s mistake by retreating to an equally unsatisfactory position whereby beliefs can only be justified by other beliefs, never by the world itself. But if experience were already a state of belief, it could not provide us with independent grounds for belief; our beliefs would be answerable to the demands of coherence but not to an independent world. (“[I]f spontaneity is not subject to rational constraint from the outside . . . then we cannot make it intelligible to ourselves how exercises of spontaneity can represent the world at all.”49) The challenge, then, for both Kant and McDowell, is to show how the contents of perceptual experience can be both conceptual content (which is the only sort of content that can serve as the content of a belief ) and also objective content, existing independently of our beliefs. McDowell responds to this challenge by showing how the contents of our most basic perceptions are able to secure a toehold in the network of

endorsement at all, but only, as I put it, an invitation to adopt such an attitude, which, in the best of cases, consists in a fact’s making itself manifest to one.” 48 McDowell (1994), p. 19. Not everyone agrees with this argument, of course. It was criticized by Jerry Fodor (1995), and it has been criticized more recently at greater length by Richard Heck (2000). 49 McDowell (1994), p. 17.

116

PERCEIVING REASONS

inferential relations while also ensuring contact with a world outside belief. The contents of our perceptions are conceptual as long as we are able to retain a memory of what is perceived and as long as we are able to recognize other things as being of the same type; and that is all that is required in order for the content of my experience to be capable of entering into inferences of various sorts (and, hence, to belong to the space of reasons). [W]hat ensures that it is a concept—what ensures that thoughts that exploit it have the necessary distance from what would determine them to be true—is that the associated capacity can persist into the future, if only for a short time, and that, having persisted, it can be used also in thoughts about what is by then the past, if only the recent past.50

Perceptual content ensures contact with a world that is independent of perception insofar as it contains a demonstrative this, whose content is precisely that object to which perception is receptive. Without such receptivity there would be no demonstrative content, and without demonstrative content there would be no perception. The objectivity of its content is, as it were, built into the very nature of perception. To say that an experience is not blind is to say that it is intelligible to its subject as purporting to be awareness of a feature of objective reality: a seeming glimpse of the world . . . that can be so only against the background of an understanding of how perception and reality are related, something sufficient to sustain the idea that the world reveals itself to a perceiving subject in different regions and aspects, in a way that depends on the subject’s movement through the world.51

Various objections have been raised against McDowell’s account of perceptual content. Many commentators have argued that demonstrative contents are simply too thin to count as conceptual in any interesting sense, and many have worried that McDowell’s account fails to make room for a content that is shared by perception and misperception alike. (McDowell has carefully articulated responses to both of these criticisms.52 It is, in any case, incumbent on such critics to demonstrate that these supposed drawbacks are more

50

McDowell (1994), p. 57. McDowell (1994), p. 56. In particular, he draws on Fregean resources to explicate the concept of that; and he allows for shared content between perceptions and misperceptions through the notion of how things seem to be. 51 52

5.

THE

“SPACE

OF REASONS”

117

serious than the drawbacks of the two alternatives McDowell is maneuvering between, or to show that there is still another alternative to be had.) If we follow McDowell’s lead, however, we are returned to a world of reasons that are already “out there” to be discovered. We simply open our “eyes” to find conceptualized contents in the world around us, and those conceptualized contents given us reason to believe what we do. In seeing a flame, for example, we see something outside ourselves that justifies our belief in the presence of a flame; the flame appears as something independent of us and its appearing—which is equally our perceiving—provides a reason (not just a cause) for our belief.53 By embracing this picture of things, we regain something of the “innocence” of the ancients, freed from the need to engage in the sort of “constructive” philosophy that (fruitlessly) aims to refute the skeptic. (The project of restoring such “innocence” by dissolving the need to respond to the skeptic is a project in which McDowell is more closely aligned with Wittgenstein than with Kant—though he differs from both of them in seeking to restore an ancient innocence about values as well.) So far, so good. As the space of reasons expands to include all perceivable facts, it seems to include almost all physical facts as well. At least on the face of it, the facts that we perceive also belong to the realm of natural law, however, and McDowell, like Sellars, wants to draw a sharp distinction between the space of reasons and the realm of law. [W]e can say that the way our lives are shaped by reason is natural, even while we deny that the structure of the space of reasons can be integrated into the layout of the realm of law.54

Sellars, it will be recalled, distinguished between empirical characterizations of episodes or states and normative characterizations of those same episodes or states; the utterance “this is green,” for example, can be characterized in two different ways—one way indicating its physical relations to various other states, the other indicating its rational or justificatory relations to other states. Davidson’s anomalous monism is much like Sellars’ position in this

53 There is no need to first ascertain the reliability of such appearings. McDowell (2002b) defends this position against what he sees as Brandom’s misleading appropriation (or, perhaps, Sellarsization) of his ideas. 54 McDowell (1994), pp. 88–9.

118

PERCEIVING REASONS

respect; both emphasize contrasting modes of conceptualization—one normative, the other descriptive. McDowell, on the other hand, insists that the objects of our perceptions and our thoughts are already individuated by the concepts that we use; there is no such thing as concept-independent content for our perceptions, and therefore no possibility of perceiving or thinking about the same event through different concepts.55 But how, then, are we to think of the relation between facts in the space of reasons and facts in the realm of law? 56 McDowell insists that both sorts of facts are parts of nature, broadly construed. He insists that we need not fall into “the supernaturalism of rampant Platonism” 57 (or the unintelligibility of Kant’s things-in-themselves) as long as we are willing to at least partially “re-enchant” nature, recognizing the irreducible reality of facts that are constituted by their normative relations alongside facts that are not so constituted. A “link” between the two sorts of fact is to be found, says McDowell, in the process through which certain modes of teaching or upbringing (ways described by Aristotle, for example) will produce a “second nature”—that is, a rational nature. In this way, normativity is said to have a causal “foothold” in the realm of law, but the resulting normative facts cannot be derived from, or reconstructed from, the practices that create them. Many have objected to a lack of explanation at this point, unhappy with the minimal accounting that McDowell provides.58 I share the desire for a fuller explanation of just how the space of reasons emerges from the space of natural law—a topic that has puzzled McDowell’s sympathizers and critics alike. The worry I want to pursue here, however, concerns the “location” of the contents of our perceptions. For the content of many of our perceptions seems to require a place within both the space of reasons and within the realm of law. When I perceive that a flame is growing, the fact that a flame is

55

McDowell’s rejection of Davidson’s ontological claim—i.e. his monism—is discussed most explicitly in McDowell (1998), p. 339, where he insists that the cause of a brain state has nothing to do with the reasons for a mental state; see, also McDowell (1994), pp. 74–6. 56 Since Mind and World (1994) McDowell (2006) has described the relevant contrast as a contrast between space-of-reasons intelligibility and rational-scientific intelligibility (rather than subsumptionunder-laws intelligibility). 57 McDowell (1994), p. 78. 58 See Simon Blackburn (2001); Robert Brandom (1995); Charles Larmore (2002). There is a familiar standoff between realists and anti-realists here—the anti-realists accusing realists of supplying nonexplanations while realists accuse anti-realists of providing explanations where none are needed.

5.

THE

“SPACE

OF REASONS”

119

growing is a constitutive part of my perception such that it must also belong within the space of reasons. But the fact that a flame is growing is also a part of the realm of law, standing in lawful relations to facts concerning heat, oxygen, and so on. Indeed, it is hard to see how our perceptions of flames could justify our belief in various laws concerning flames, heat, oxygen, and so on unless the subject matter of our perceptions were also the subject matter of those laws. At times, it seems that the phrase “space of reasons” is used in two different ways. One way (more prominent in early sections of Mind and World ) has everything that is perceivable belonging, necessarily, within the space of reasons. The other way (more prominent later on) has only the activities of rational and ethical beings—beings which are capable of spontaneity as well as receptivity—belonging within the space of reasons.59 Noting the two different uses, Michael Friedman accuses McDowell of conflating Kant’s account of the operations of the understanding (which infuse our perceptions, and our world, with conceptual content) with Kant’s account of the operations of autonomous reason (which determine how things ought to be, and free us from the constraints of causal law).60 On the one hand, the understanding (which synthesizes intuitions in accordance with rules) gives causal laws a necessary place within the space of reasons, for it is the causal laws that ensure objectivity. On the other hand, the spontaneity of moral action creates a realm of reason that cannot be reconciled with the necessity of causal laws. In trying to keep these two “worlds” apart while placing them both within nature, Friedman thinks that McDowell eventually falls into just the sort of post-Kantian idealism that he seeks to avoid—an idealism whereby the realm of law and the realm of morality are both created through by the operations of the understanding. And this, of course, is an outcome that McDowell cannot accept. Whatever else he is, he is a realist; we are not responsible for creating an objective world but are, instead, responsive to it. There are deep reasons for McDowell’s (and Kant’s) alignment of conceptual capacities and ethical capacities, however—reasons that McDowell 59

McDowell claims that non-rational animals exist in an environment, not in a world (which is also a space of reasons), while we rational animals exist in both. This suggests that what we come to know through perception—for example, that this flame is growing—is something different from what animals come to know through perception. McDowell (2002b) elaborates on his view of animal knowledge. 60 Michael Friedman (1996), pp. 427–67.

120

PERCEIVING REASONS

articulates most clearly, perhaps, in a paper entitled “Two Sorts of Naturalism.”61 There he maintains that the ability to apply concepts presupposes an ability to evaluate those applications, recognizing the possibility of error and standing willing to revise judgments as necessary; this, indeed, is what makes the making of judgments a normative enterprise. Furthermore, in acquiring the ability to evaluate judgments, one simultaneously acquires the ability to evaluate actions. We cannot make sense of a creature’s acquiring reason unless it has genuinely alternative possibilities of action, over which its thought can play. We cannot intelligibly restrict the exercise of conceptual powers to merely theoretical thinking . . . we need to make room not only for conceptual states that aim to represent how the world anyway is, but also for conceptual states that issue in interventions directed towards making the world conform to their content . . . This is to represent freedom of action as inextricably connected with a freedom that is essential to conceptual thought . . . We cannot allow ourselves to suppose that God, say, might confer reason on wolves, but stop short of his giving them the materials to step back and frame the question “Why should I do this?”62

I quote this passage at length in order to show how the possibility of reflection on one’s beliefs and actions underwrites a kind of freedom for both understanding and morality. McDowell is not conflating Kant’s account of the understanding and his account of morality so much as highlighting their common reliance on the free-play of reflection. If this is right, then it is our capacity for reflection that enables us to move beyond informed responses to the world (the sort of responses that other animals make as well) to the making of judgments about the world and judgments about ourselves. “Second nature” is not just ethical nature, it is reflective nature, and reflective nature is what enables us to recognize both the independence of the world we perceive and its capacity to provide us with reasons for our beliefs and actions. [W]e arrive at the notion of having one’s eyes opened to reasons at large by acquiring a second nature.63

61 62 63

McDowell (1998), pp. 167–97. McDowell (1998), pp. 170–1. McDowell (1994), p. 84.

5.

THE

“SPACE

OF REASONS”

121

By McDowell’s own lights, this independent reason-giving world includes both the fact that a flame is growing and the fact that hurting someone is wrong—both of which are normative in the sense that both warrant the acquisition of some beliefs rather than others, and the performance of some actions rather than others. The ethical is a domain of rational requirements, which are there in any case, whether or not we are responsive to them. We are alerted to these demands by acquiring appropriate conceptual capacities. When a decent upbringing initiates us into the relevant way of thinking, our eyes are opened to the very existence of this tract of the space of reasons. (my italics)64

Ethical facts, then, are just one part of the space of reasons; physical facts are another. My worry, then, is not that of Friedman, who fears that too much is being assimilated within the space of reasons. I worry, rather, that the sharp divide between the realm of law and the space of reasons begins to break down once we realize that most if not all facts—all facts that can be the objects of our intentional states—must belong in both domains. Put another way, instead of supposing that there is a realm of law, containing one kind of thing, that exists alongside a space of reasons, containing another kind of thing, we should recognize that the very same facts exist both within the space of reasons (on account of their rational relations to our beliefs and actions) and within the realm of law (on account of their lawful relations to each other). Assuming, further, that the identity of a fact or event is constituted by its place in justificatory practices and explanatory practices, then there will be no such thing as normative facts versus descriptive facts. Consider a fairly simple case of what I have in mind: the case of blushing. Blushing does not count as blushing unless it is lawfully related to the flow of blood through the skin and rationally related to relevant judgments about others’ perceptions of oneself. Similarly, wincing will not count as wincing unless it stands in lawful relations to the movements of facial muscles and in normative relations to judgments about an error, an insult, or a pain, for example. Any attempt to separate a blush or a wince into the lawful part and the rational part seems misguided. There are not two separate states in two separate 64

McDowell (1994), p. 82.

122

PERCEIVING REASONS

realms here; there is, rather, a type of entity that must secure its identity through ties of two different sorts. I suggest that we take this same approach to the contents of perception to accommodate much of what seems right about McDowell’s position. The perception of a growing flame, for example, will only count as the perception of a growing flame if there are both lawful relations connecting the contents of this perception to the contents of certain other perceptions concerning heat, oxygen, and so on, and rational relations connecting the contents of this perception to the contents of a belief about a growing flame. I do not need to know that there is a lawful relation between the growing flame and the increasing heat in order to perceive that there is a growing flame; nor do I need to know that there is a rational relation between the growing flame and my belief that there is a growing flame in order to perceive the growing flame. But both sorts of relation must indeed hold in order for the content of my perception to be that there is a growing flame. By insisting that the contents of perception—the fact that the flame is growing—are conceptual contents, McDowell has placed objective facts within the space of reasons. [T]he deliverances of our receptivity . . . can innocently be taken to belong together with our world-views in the space of reasons, since they are already in the space of concepts.65

But growing flames and increasing heat are the sorts of things that exist in the realm of law as well—with the result that the identity of perceivable facts must be constituted by both normative relations and lawful relations. Nothing will count as the fact that a flame is growing unless it can justify various beliefs and actions and unless it can enter into lawful relations with facts about heat, light, oxygen, and so on. Future beliefs and theories or future developments in science may transform the relevant rational relations or the relevant lawful relations, respectively, but that doesn’t mean that rational relations and lawful relations are not both constitutive of the growing flame as such. Chapter 5 extends this approach to the quintessentially normative domain of moral facts and our perception thereof.

65

McDowell (1994), p. 141.

6.

THE ADVANTAGES OF PERCEIVING REASONS

123

6. The Advantages of Perceiving Reasons We are now in a better position to consider why the ability to perceive reasons matters. There is, I take it, no puzzle about why having accurate beliefs about reasons is important: having reasons enables us to discover connections, formulate unities, make predictions, and increase our power to change things. But none of these advantages seems to depend on perceiving reasons—knowing them directly and imagistically rather than knowing about them in a more indirect or discursive way. Indeed, the attempt to imagine counterfactual possibilities in an attempt to see why things happened as they did or why things should happen in a particular way may seem more misleading than helpful; relying on our imaginations can cause us to overlook important distinctions, to be distracted by extraneous details, and to confuse what is spatial with what is not.66 These are all legitimate worries—worries that should temper any preference we might have for perceptual rather than non-perceptual ways of knowing reasons. By clarifying the process involved, though, the account offered above has clarified some of the advantages of perceiving reasons, and these advantages are worth highlighting. The distinguishing feature of perceptual versus non-perceptual knowledge—one that was noted by Descartes, and mentioned above—is that the justification for a perceptual belief is experienced at the time of belief. I do not need to remember (or simply trust) that I had a justification for my belief if the justification is presented to me at the time of belief. When I believe that overhead clouds are the cause of the current rain, for example, I believe that the presence of such rain depends on the presence of such clouds. And when I perceive this dependence—by actively imagining a variety of counterfactual situations in which the clouds or the rain were different, I experience the justification for my belief. In contrast, when I simply remember that clouds cause rain without experiencing the dependency of rain on clouds, I do not experience the justification for my belief. Why, though, is it important to experience the justification for a belief at the time of belief (a belief that is about reasons—or anything else for that 66 There is a large, and growing, literature on the reliability of imagination as a source of knowledge about what is (and what is not) possible. On my account, imagination must be a fairly good indicator of possibility (though not all possibilities, and not without the occasional mistake) in order to underwrite genuine perceptions.

124

PERCEIVING REASONS

matter)? Descartes seems to have thought that only then could knowledge be completely secure from the vagaries of memory; but there is no guarantee that what we experience in the present is sufficient justification for a given belief (a problem that Descartes tried to fix by restricting his confidence to beliefs that consisted of “clear and distinct” ideas, and by arguing for the existence of a God who guarantees the truth of such ideas). Furthermore, many of our most reliable beliefs (beliefs about one’s name, for example, or the belief that Venus is a planet rather than a star) are not perceptual beliefs since they are not accompanied by any experience of their justification. If experiencing the justification for a belief merely makes us feel more secure, without in fact being more secure, then it is not an advantage after all. The fact that perceptual beliefs are not entirely secure, and the fact that some non-perceptual beliefs are more secure than some perceptual beliefs, is entirely compatible with the claim that perceptual beliefs, or beliefs with self-evident justifications, ought to be favored over non-perceptual in general. In the complex process of acquiring and revising our beliefs, we may often choose to discard a perceptual belief whose justification is selfevident—in favor of a non-perceptual belief whose justification is less direct but more extensive; but insofar as indirect justifications must ultimately rely on directly experienced justifications of some sort or another, then perceptual beliefs will as a group have a privileged position. There are a number of different ways to understand this privilege. For example, it has been argued that a perceptual belief ought to be overridden by a non-perceptual belief only if the latter preserves, or is derived from, a greater number of perceptual beliefs. (Davidson has elaborated his so-called “principle of charity” along these lines, which raises the problem of just how to count beliefs.) But however one understands the epistemic privilege of perceptual beliefs in general, it is still a fact that for any particular belief that p the epistemic privilege that accrues to perceptual beliefs in general will not prevent that particular belief from being overridden by a non-perceptual belief that not-p.67 So the fact that I can, through the exercise of my imagination,

67 Furthermore, the justification for not-p can be stronger than the justification for any one of the observations on which it is based if not-p is the only possible explanation of those observations and if the probability that at least one of those observations is accurate is greater than the probable accuracy of any particular one of them.

6.

THE ADVANTAGES OF PERCEIVING REASONS

125

come to perceive clouds as the reason for rain does not seem to bestow any epistemic advantage on that particular belief.68 The epistemic advantages of perceiving that one thing is the reason for another, as opposed to knowing this in some non-perceptual way, lie elsewhere. One such advantage derives from the way perception requires one to activate a number of different expectations all at once—something that can function to ensure the overall consistency of those expectations. In the case of perceiving an object—a nearby oak tree, for example—one must imagine other perspectives on the tree, and it will only be possible to imagine these different perspectives all at once if they are consistent with one another. If the tree has moss one side only, then it will appear on the right from one perspective, on the left from another, and from still another perspective it will be invisible; if the leaves are thickest near the top, then more of the tree will be visible from beneath than from above; and so on. I can imagine (or draw) these different views separately in ways that are not consistent—imagining the moss as appearing to the right from perspectives, or imagining the sky blocked when looking up from under the tree yet a clear view of the ground from above the tree—but if I try to imagine them simultaneously, the inconsistency will become evident and the associated beliefs (about the tree’s surfaces or its leaves) will be automatically discarded or adjusted. This contrasts with the case of non-perceptual beliefs that, even if held simultaneously, do not reveal their inconsistencies so easily. For it is much easier to believe that a patch of moss can be seen from all angles if we don’t try to imagine the view from these different angles simultaneously; and it is much easier to believe that one cannot see the sky from below but can see the ground from above if we don’t actively imagine both views at once. Because non-perceptual beliefs do not require us to experience the world as we believe it to be, inconsistencies in the contents of our beliefs need not be correlated with inconsistencies in our experience—with the result that inconsistencies are more likely to be overlooked when our beliefs are not perceptual beliefs. This does not mean that we can possess concepts without having some relevant perceptual experiences, nor that we can have

68 People are often advised to attach images to words in order better remember them. It is not clear, however, that associating images with words is any more helpful as an aid in remembering words than attaching words to images is an aid in remembering images. More modes of access improve memory, whatever those modes may be.

126

PERCEIVING REASONS

perceptual experiences without employing some concepts69—only that concepts can sometimes be applied and beliefs can sometimes be entertained in the absence of the associated perceptions, and that activating the associated perceptions can correct for inconsistencies that might otherwise be overlooked. How does this work in the case of perceiving reasons? How does perceiving a dependency eliminate inconsistencies that would likely go undetected if our knowledge were non-perceptual? Whereas in the case of perceiving the moss or the leaves of a tree, it is different perspectives on the tree that must be simultaneously imagined, in the case of perceiving clouds as the reason for rain, it is different atmospheric conditions that must be simultaneously imagined and coordinated. It is relatively easy to believe that rain requires clouds and that a change in temperature is enough to cause the moisture in air to precipitate out as long we do not simultaneously imagine the possibilities implied by each—as long as we do not try to imagine rain disappearing whenever clouds disappear while also imagining rain falling when uniformly moist air is cooled. When we imagine both of these things, we discover an inconsistency in our beliefs that we are impelled to eliminate (by discarding the belief about clouds, or by adding a belief about the way that moist air groups itself before releasing its moisture). Perception enables us to experience a number of different dependencies simultaneously rather than sequentially; and the simultaneous consideration of multiple dependencies provides some check on the accuracy of each. If I make a series of calculations to determine the fastest way to climb up a particular mountain, for example, and I write down all of these calculations (about distances, relative steepness, available water, etc.), I will have reasons for concluding that the fastest way up is, say, from the east. Apart from increasing my concentration and repeating my calculations, however, I will have no protection against mistakes in my numbers or my calculations. And assuming that the calculations are fairly complex, I will not be capable of keeping all of my calculations in mind at once. In contrast, if I am able to visualize the different distances, the relative steepness of different routes, and the locations of various water sources, actively imagining the way that the speed of a climb depends on all these factors, then the coherence or lack of 69

As noted in Chapter 1, Section 1, the plausibility of this claim about concepts will depend, in large part, on what one takes concepts to be.

6.

THE ADVANTAGES OF PERCEIVING REASONS

127

coherence between these considerations can provide an extra check on the accuracy of my calculations and conclusions. Likewise, if I ask you a number of questions in an effort to measure your courage, your articulateness, and your quickness, and I compare your answers with the answers of others I have studied, concluding that you will thrive under pressure, I may not notice a certain mismatch between the sort of courage you manifest and the way in which you are articulate—making it unlikely that you would be both courageous and articulate at the appropriate juncture. Repeating my questions and redoing the comparisons seems to be the only way I can strengthen my confidence in the reasons I have acquired in this way. In contrast, if I can perceive your courage, your articulateness, and your quickness, and I am able to actively imagine how a combination of those traits would affect your behavior under pressure, then my prediction that you will thrive under pressure will not be vulnerable to the same errors and oversights. Perceiving the relevant dependencies is not the only way to recognize and correct for inconsistencies in one’s beliefs, of course, but the active imagining that perception requires does tend to ensure greater consistency in our beliefs by ensuring a more consistent set of experiences. This suggests, also, that to the extent that it is possible to use our imagination to visualize the referents of our words, we will be better able to recognize inconsistencies in what is said, and more likely to discover unforeseen implications of what is said. This prediction is borne out by studies in which people become much better at reasoning with complex conditionals once they are able to visualize their referents as concrete states of affairs rather than mere numbers.70 (Visualizing referents can also be distracting or misleading, of course, as extraneous details can command one’s attention, and as the visualization of what is normal prevents one from contemplating something abnormal.71) It is often said that some people are “visual people” while others are “verbal people,” and that some problems are spatial problems while others 70 See the Watson card task studies on this. Ruth Byrne (2007) has an interesting discussion of variations on this task. 71 Keith Stenning (2002), p. 134ff, details studies showing that some students (those with low GRE scores) improve their reasoning by learning syllogistic logic via sentences and syntactic rules, while others (those with high GRE scores) improve their reasoning by learning logic via graphic representations (and that teaching the wrong method to the wrong group can actually damage performance). He attributes this to the greater ability of students with high GRE scores to discern invariance across multiple cases, and to abstract away from irrelevant aspects of graphics.

128

PERCEIVING REASONS

are conceptual problems. Both of these observations are undoubtedly true, but they do not change the fact that simultaneous imagining of multiple implications—when it is possible—has a better chance of discovering and correcting for inconsistencies than does sequential contemplation of a more verbal nature. Some people are better at visualizing mathematical relations while others are better at providing formulas that describe such relations, and some mathematical relations may not be visualizable by anyone, but it is quite clear that spatial presentations allow more relations to be contemplated simultaneously and that inconsistencies are more likely to be detected when the inconsistent elements are considered at the same time. Perceiving reasons, thus, has the epistemic advantage of revealing inconsistencies that might not otherwise be noticed; and the feeling of security we get from perceiving reasons as opposed to merely recounting reasons is valid insofar as perception gives us added protection against inconsistencies in our beliefs.72 It is important to be clear about what I am, and what I am not, claiming here. I am claiming that the phenomenology of perception, because it depends on the simultaneous imagining of multiple possibilities, is epistemically significant; it is not an inert add-on, an epiphenomenal flourish. Imagining various possibilities that are implied by a given belief, and imagining them simultaneously, provides an extra check on consistency—a check that is absent when the relevant imagining is absent. (Consistency does not guarantee truth, of course, and inconsistent sets of beliefs may sometimes contain more truth than consistent sets; but on the assumption that true beliefs must be consistent with one another, consistency remains an epistemic virtue.) I am not, however, claiming that the feeling of security or certainty that characterizes perceptual beliefs helps to justify those beliefs; the feeling is, rather, a manifestation of the added security that perceptual beliefs already have. A second advantage of perceiving reasons, related to the above reflections on consistency, derives from the way that perception tends to “fill in the gaps” in what is presented to us. We are more likely to imagine the parts of a house that are blocked by trees than we are to imagine the parts of a story that were drowned out by the noise; and we are more likely to imagine 72 I am assuming that revealing and correcting inconsistencies is epistemically advantageous—even if some ways of avoiding or preventing inconsistencies (by keeping one’s eyes closed, or by being a dogmatist, for example) are epistemically disadvantageous.

6.

THE ADVANTAGES OF PERCEIVING REASONS

129

causal connections between a series of events that we observe than between a series of events that we read about. Our imagining is often guided by memories of past experiences—memories of other houses, memories of other stories—but our imagining does not merely reproduce these past experiences (or “cut and paste” them into the presented scene or story); our imagining is also productive in generating new material for new circumstances—continuation of a color that we have never seen before, extensions of a window design that is new to us, and so on.73 This productive use of imagination also serves to generate new knowledge insofar as its forays into new domains are accurate and justified. If we imagine the molecular structure of ice, for example, we can also imagine how an increase in molecular movement—i.e. an increase in temperature— would break down that structure and result in the ice melting. Before visualizing the molecular structure of ice, we might infer (given our belief about the molecular structure of ice and our belief that a rise in temperature causes ice to melt) that a temperature rise that causes ice to melt must cause a change in its molecular structure, but we wouldn’t yet know why; thus we wouldn’t yet know that the transition from ice to water is abrupt, or that changing ice to thirty-two degree water requires more heat than changing thirty-two degree water to thirty-three degree water.74 These are facts that 73 In order to unify the information I receive from different senses at different times (green, brown, rough, smooth), I must experience them as belonging to a single substance (a tree), and according to Kant that is only possible insofar as my present observations serve to reactivate certain past experiences (of the same or similar trees)—imaginatively reproducing the brown of the trunk even as I now focus on the green of the leaves, and imaginatively reproducing the smoothness of the leaves even as I now simply look at their color. As we have seen above, though, mere convergences of properties in space and time, imaginative or otherwise, are not enough to secure the unity that consciousness requires. In addition to retrieving and simultaneously reproducing assorted sensations from the past, the imagination must also “fill in the blanks” to supply a connective “fabric” that show us how what is observed fits together with what is not observed. This is imagination in its “productive” rather than merely “reproductive” role. To experience the tree as a fully unified entity, I must imagine its properties as they are during times I am not observing them and I must imagine various features that I have never observed (the length of its top branches, the color of its roots, etc.). This requires us to make various projections—from actual experience to possible experience—in a way that seems objective rather than merely subjective; and that depends on grasping and following rules that not only order our experience but also serve to explain and guide our experience. (Here Kant differs from Hume in (a) insisting that adherence to rules is a prerequisite of experience itself, and (b) treating the “productive” imagination as genuinely informative rather than merely habitual.) Imagination thus does the work that sustains a unified consciousness by ensuring that the objects of consciousness are presented as belonging to an independently unified world. 74 Some other examples: When we see the molecular structure of crystals, we can also see that electricity will be released when pressure is applied along the axis of symmetry. If I see the fear in a friend’s behavior (or in a painting), I am usually better able to interpret certain other aspects of her

130

PERCEIVING REASONS

become evident once we picture the internal process involved, and otherwise not. Although these facts may be derived (rather laboriously) from various propositions about temperature and molecular structure, they might not be discovered unless the relevant relations are imagined simultaneously. While it is true that a picture can’t show one how to go on unless one already knows how to interpret the picture, it is also true that seeing a picture often leads one to new discoveries precisely because it activates the dispositions that constitute one’s interpreting skills. Indeed, the ability to follow through on a given interpretation (the knowledge of how to go on) seems to depend on having an “in a flash” experience because it depends on the simultaneous activation of various dispositions associated with a given interpretation. Some applications of a concept and some expectations that come with a concept may themselves be present in the “in a flash” experience, but many other appropriate applications and expectations will occur only later. These later judgments are logically related to the initial judgments, however:75 if these are the duck’s wings, then it must be getting ready to fly; and if this is a happy smile, then the ensuing tone of voice will be happy also. The instantaneous seeing of a duck “in” the drawing depends on imagining this as an eye and that as a shoulder, and vice versa. Seeing it as a duck in this way enables one to make further discoveries, such as a discovery that the duck is worried, that it is about to fly away, and so on. Likewise, it is the instantaneous seeing of happiness “in” a friend’s face that enables us to go on to discover the invitation in her raised eyebrow, the selfbemusement expressed in the curl at the corner of her mouth, or the incipient laughter. Merely being told that a drawing is a drawing of a duck, or that a face expresses happiness, is not sufficient to guide our interpretations in this way. The duck must be seen as a duck, and the happiness must be seen as happiness. As Wittgenstein writes: If you see the drawing as such-and-such an animal, what I expect from you will be pretty different from what I expect when you merely know what it is meant to be.76 behavior (or other features of a painting)—aspects that were previously either unnoticed or unintelligible to me. 75 Like Wittgenstein, I am using “logically” in a very broad sense to include conceptual connections of all kinds, not just those dictated by the logical constants. 76 Wittgenstein (1953), Part II, p. 205. See also p. 203: “What does it mean to say that I ‘see the sphere floating in the air’ in a picture? Is it enough that this description is the first to hand, is the matter-of-

6.

THE ADVANTAGES OF PERCEIVING REASONS

131

If someone sees a smile and does not know it for a smile, does not understand it as such, does he see it differently from someone who understands it?—He mimics it differently, for instance.77

Others may use a few choice words to point us in the right direction, but we won’t activate the full set of dispositions needed for successful interpretations without a phenomenological experience that enables us to see it for ourselves. It is the “in a flash” experiences, and the imagining that goes into them, that govern how we go on—directing and not merely accompanying our subsequent discoveries.78 Not all discoveries are made in this way, of course. In fact, given the difficulty (perhaps even impossibility) of perceiving certain relations, some discoveries are more likely to be made through non-perceptual calculations. We discover that ducks are carriers of a virus by testing their blood rather than by perceiving the virus in their blood; and we discover that a friend is fifty years old by doing the calculations rather than by perceiving the time interval between her birth and the present. But because perception, when it is possible, requires an activation of the imagination in a way that is not required by non-perceptual modes of knowledge, perception provides an immediate impetus towards imaginative extension that has the epistemic advantage of more or less automatically adding to our knowledge. If I can accurately imagine the virus in relation to its surroundings, I will immediately see some things about where it can or cannot fit, how it might damage a cell, etc.; and if I can accurately imagine the interval between a person’s birth and their present, I will immediately see some things about what that person experienced when.79

course one? No, for it might be so for various reasons. This might, for instance, simply be the conventional description.” 77

Wittgenstein (1953), Part II, p. 198. It is not just the fact that we have used a particular concept, or exercised a particular ability, in effecting an “in a flash” experience. It is, rather, the fact that that concept, or the exercise of that ability has been so successful in unifying different aspects of our experience; that is what justifies our continuing to exercise the same ability in the future. There may be several different concepts or interpretations that are equally successful, of course. The importance of such indeterminacy in perceiving people as people is discussed in Chapter 5. 79 For many people, with respect to many domains, calculations are more reliable than imaginings. My point remains, however, that if one can perceive a given relation (thereby imagining the relevant possibilities accurately), one already has a head start on extending one’s knowledge to further, implied possibilities. 78

132

PERCEIVING REASONS

Christopher Peacocke maintains that an animal’s ability to discover new routes between familiar objects indicates that it is relying on a mental map that shows familiar objects in familiar relations but also, in virtue of being a spatial map, shows a multiplicity of unfamiliar relations between these same objects. When a bee, moving from flower to flower, finally has enough pollen, it does not retrace its steps but, rather, makes a beeline for home, regardless of the fact that it has never followed that particular pathway before.80 And when a dog, exploring the neighborhood, returns to check on the bone it buried, it knows how to go there directly even though the direct route is not a route it has ever taken before. Peacocke argues, quite plausibly, that the ability depends on an animal’s ability to represent (or imagine) different locations in a common space, some parts of which one has not yet encountered. Again, one could calculate a shorter route home based on the path taken (replacing each three steps east and each four steps north with five steps southwest, and so on), but such calculations are unnecessarily complex given our ability to simply see (by engaging our imagination) a way to “cut the corner.” The other way in which visualization can lead to new discoveries derives from the fact that our image of an object usually carries far more information than we have put into words.81 (This is consistent with the claim that words also carry extra information, not present in images.) By imagining both flying and non-flying birds, for example, we may discover that flying birds have proportionately bigger wings than non-flying birds—something that might be figured out on the basis of aerodynamic calculations but something that we can also simply see by imagining (based on remembering) eagles and crows and penguins and ostriches. Or, to take another example, by imagining a friend’s face upon hearing bad news, we might discover that bad news excites her, or frightens her (whereas it dulls or depresses someone else); while we have never drawn this inference before, we have registered the brightening of her eyes in such situations, and the incorporation of that information in our current imagination leads to a discovery that we would 80 For alternative readings of this behavior, see Gattis (2001), p. 20, where she discusses Peacocke on animal maps. 81 What I am calling visual versus verbal information need not correspond to the distinction between non-conceptual versus conceptual information. I am inclined to think that all content is conceptual— indeed, must be in order to play an evidential role. There are good analyses of this by Elisabeth Camp (2004, 2007). But that dispute (and the many different definitions of “conceptual” that come into play) needn’t engage us here. It is enough to maintain that visual memories (and imaginings) often carry more information than has been explicitly verbalized.

6.

THE ADVANTAGES OF PERCEIVING REASONS

133

not have make otherwise; even though nothing in our previous descriptions of our friend would suggest such a conclusion. As noted above, the presence of extra information can also be an epistemic liability. We can be distracted by the colors of flying versus nonflying birds and come to believe that that is where the crucial difference lies. Or we can think that a characteristic smile is more significant than it actually is. Often, it is the superficial features of things that grab our attention, that stand out as most vivid, and that become the focus of subsequent attempts at explanation.82 What appears to be a discovery may in fact be a red herring, and the fact that certain patterns can be seen rather than merely inferred can create confidence in their importance—confidence that may actually impede the correction of our mistakes. My claim is not about the likely benefits versus risks of visualization, however; it is about a certain type of discovery that can come with actually imagining the subject matter under discussion. Whether these potential benefits outweigh the potential harms is probably not something that can be determined in general; it will depend, in large part, on what is at stake if we succeed versus what is at stake if we fail. A third advantage of perceiving reasons is that perceptual beliefs usually motivate us more quickly and more reliably than non-perceptual beliefs. This is particularly important in moral contexts where perceiving another’s suffering can move us to act when merely believing that another is suffering does not. (There are also cases where it is important for deliberative reasons to be able to override perceived reasons. It is not possible to perceive everything we need to know, morally or otherwise.) Perceiving reasons can also be valuable in non-ethical contexts where immediate action is called for; visualizing the constitutive reasons for melting ice can strengthen one’s motivation to retreat from a thawing lake quickly—a retreat that may not feel so urgent otherwise. Here the important thing is not the more secure or more generative epistemic standing of perception but its visceral urgency.83 Many of our movements arise quite forcefully and automatically

82 Foucault (1970) documents historical forms of explanation that were more closely tied to the appearances of things. Learning and doing logic via diagrams versus formulas makes certain inferences more automatic, but it also invites confusion about relevant versus irrelevant features of diagrams. 83 Moran (1994), p. 100, claims that much of philosophy and literature does not aim at altered beliefs but, rather, (1) changed associations and comparisons, (2) more vivid or “felt” appreciation of something already known, or (3) changes in habits of attention and senses of what is important versus trifling.

134

PERCEIVING REASONS

in response to things that we see; we dodge fast-approaching objects without even having the time to think about it, we reach out to catch a falling vase before remembering that we hoped it would break, and we make our way through a crowded room thinking only of the person at the other end. Without perception of the relevant objects, these actions are often possible but usually less likely and less effective. It is a welldocumented fact that we respond to a perceived opportunity, or a perceived danger, more quickly and more reliably than we respond to a reported opportunity or danger—not merely because it takes us longer to process the information received in a report but (also) because perception is linked to action in ways that are relatively unmediated by verbal thought; when we see something attractive (or dangerous) our bodies are drawn toward it (or repelled away from it) before we even have time to think about what we are doing and whether it actually serves our purposes.84 Likewise, when we can see what follows from what, or see what a situation requires, our actions tend to fall in line automatically; no additional resolve or will power is needed. Emotionally also, we are more responsive to things that we see than we are to things that we merely believe: seeing that a painting depicts war affects us with a force that merely believing that it depicts war does not, and seeing a child’s delight moves us in a way that merely believing that it is delighted does not. Insofar as increased responsiveness to the world and to others is desirable, then seeing reasons will be preferable to understanding that remains non-perceptual. Jonathan Cohen has articulated a distinction between belief and acceptance, where belief involves a feeling of commitment while acceptance requires our reflective endorsement; and because belief involves feeling, it moves us more directly to action. Whatever labels we choose to use, we can certainly agree that our conviction of another person’s guilt, for example, can be deeply felt rather than well-reasoned (or vice versa), and while the latter sort of conviction may rightly guide a jury’s verdict, it is the first sort that is more likely to guide our interactions with that person. My contention, then, amounts to the claim that visualizing an event or a character or a 84 A famous passage of the Confucian Mengzi (1998) 2A6, describes how the proper response to a baby running towards a well is automatic, not reflective; and that is presented as a model for ethical action in general.

6.

THE ADVANTAGES OF PERCEIVING REASONS

135

series of implications tends to produce a felt conviction (Cohen’s “belief ”) in what is visualized and, as a result, tends also to have a more direct impact on action. When our images and convictions are mistaken, or when acting on them is unwise, a more immediate translation into action might be harmful; but when our beliefs are true, having automatic inclinations to respond to that truth (versus a need to rally one’s will power) is usually a good thing.85 Even when immediate action is ill-advised, retaining (and perhaps redirecting) the impulse to act is preferable to thought which remains affectively dissociated from action. For impulses held in check nonetheless prepare us to act when the occasion arises; and they help to ensure a kind of emotional engagement (which is essential to the feeling of living) that is otherwise absent.86 If we assume that “in a flash” perceptual experiences are themselves the result of activating some dispositions rather than others, then the very fact of having such experiences shows us to be disposed towards certain behaviors rather than others. The phenomenological experience of seeing a duck in a drawing, or seeing happiness in a face, already implements a certain way of organizing our present experience into a unified whole, and that way guides our future responses as well. Wittgenstein’s remarks on the connection between seeing and doing nicely highlight the way that our inclinations already “shape” our seeing, and the way that our “seeing” enables us to act more “freely”: Why does it seem so hard here to separate doing and experiencing? It’s as if doing and the impression didn’t happen side by side, but as if doing shaped the impression.87

85

Moran (1994), p. 90, suggests that part of the power of a metaphor is due to its production of images that control our thinking “at a level beneath that of deliberation or volition.” And pp. 101–2: “A picture can be used to get a point across without incurring the risks and responsibilities of asserting that point . . . On the other hand, when it comes to reproducing a picture, there doesn’t seem to be anything like the distinction between mention and use.” Johnson (1990) also emphasizes the special power of images because of their ability to tap more primitive, automatic, and emotional parts of the brain. 86 Blackburn (1990) writes: “When we think of categorical grounds, we are apt to think of a spatial configuration of things . . . Categoricity in fact comes with the subjective view: there is nothing dispositional, to the subjects, in the onset of a pain or a flash in the visual field . . . But the problem remains that this gives us no help in understanding what, except counterfactuals, is true of the objective order of nature . . . It almost seems that carelessness and inattention alone afford a remedy—the remedy of course of allowing ourselves to have any idea at all of what could fill in space.” 87 Wittgenstein (1982–92), Part I, ä585–6.

136

PERCEIVING REASONS

One kind of aspect might be called “aspects of organization.” When the aspect changes parts of the picture go together which before did not . . . “Now he’s seeing it like this,” and “now like that” would only be said of someone capable of making certain applications of the figure quite freely. The substratum of this experience is the mastery of a technique.88

What the final line suggests is an explanation of why seeing and doing are so closely related: mastery of a technique (i.e. establishment of an appropriate set of dispositions) forms the “substratum” (i.e. the constituent material) for perceptual experience. No wonder, then, that the ability to perceive a particular reason will be especially well correlated with a tendency to act in accordance with that reason. Psychologists are well aware of the way that implicit beliefs guide our behavior more quickly and more reliably than explicit beliefs.89 Recent work by Daniel Kahneman and others further suggests that we are able to perceive causes and intentions precisely when our beliefs in the relevant counterfactuals are implicit rather than explicit (or “automatic” rather than “elaborative”), and that imagining versus calculating future consequences has the advantage of being informed by a wealth of otherwise inaccessible implicit beliefs.90 It is the implicit nature of the knowledge used in mental simulation that makes such simulations potentially instructive, when their outcomes conflict with explicit beliefs . . . mental simulation draws on sources of knowledge that cannot be accessed otherwise and simulation exercises can therefore be truly instructive.91

From an empirical standpoint, then, it should not be surprising that insofar as perceiving reasons relies on implicit beliefs about alternative possibilities, reasons that are perceived will be especially effective in guiding our action when confronted with those possibilities.

7. Conclusion The account of perception offered in Chapters 1 and 2 has now been extended to make sense of the claim that we can perceive reasons of various sorts—causal 88 89 90 91

Wittgenstein (1953), Part II, p. 208. See discussion of some of this data in Church (2002). Kahneman (1995). Kahneman (1995), p. 380.

7.

CONCLUSION

137

reasons, constitutive reasons, and justificatory reasons. In each case I have attempted to show how the phenomenology of perception is the phenomenology of objectivity, and how the phenomenology of objectivity depends on the imagining of relevant alternatives. But whereas, in the case of perceiving a table, it is alternative perspectives on that table that must be imagined, in the case of perceiving a broken leg as the cause of a table’s tilt, it is alternative arrangements of the table’s parts and its surroundings that must be imagined. The basic idea here is that we can come to see the dependence of one state or event on another (in a given context) by actively imagining relevant counterfactual possibilities. Despite some important differences between explanatory reasons and justificatory reasons, it was possible to give a similar account of their perceivability. Finally, I described three respects in which perceiving reasons is preferable to knowing reasons in a non-perceptual way: it is easier to check the consistency of our beliefs when they are presented perceptually, it is easier to generate new knowledge when the relations between existing knowledge can be considered simultaneously, and our actions are more likely to accord with our reasons when those reasons are perceived.

Figure 4.1. “TNT Crystallizing” (photo by Daniel Friedman)

Chapter 4 The Further Reaches of Perception The previous two chapters focused on the perception of nearby states of affairs and the perception of dependencies between such states of affairs. This chapter considers a still wider range of things that might be perceived. Section 1 addresses the possibility of perceiving states of affairs that are located in remote regions of space and time—people in faraway lands or happenings in distant galaxies, for example. Section 2 addresses the possibility of perceiving future states of affairs—perceiving the future course of a hurricane, for example. Section 3 addresses the possibility of perceiving abstract states of affairs—the relation between the number three and the number six, or the relation between power and justice. In all of these cases, there is an understandable temptation to think that while our knowledge may be accompanied by images (visual or otherwise), claims to actually perceive such things are metaphorical at best. If I am right, though, appropriate sorts of imagining can make our knowledge genuinely perceptual, and much of what is discounted as merely metaphorical perception ought to qualify as literal perception. In Section 4, I comment on some key passages in the writing of Plato and in the writing of Descartes, to show how my account of “rational perception” accords with theirs. In order to perceive an object or an event, we must experience it as situated in space. (This may seem obvious, but it is not; I offered a detailed defense in Chapter 1, Section 3. Briefly: I argued that perceiving an object depends on experiencing its objectivity, that experiencing its objectivity depends on imagining alternative perspectives or modes of access, and that imagining alternative perspectives or modes of access requires that we imagine the convergence of different trajectories through space.) Likewise, to perceive a relation between two objects or events, we must experience both as situated

140

THE FURTHER REACHES OF PERCEPTION

in space. There is a difference, however, between experiencing an object across the room and experiencing an object as if it were across the room; and it is something else again to experience something non-spatial as if it occupied a space across the room. For example: I may experience my mother standing across the room, or I may experience her (as we talk on the phone, say) as if she were across the room; or again I may experience her concern as if it were a cloud hovering over the room. In investigating the possibility of perceiving distant or abstract states of affairs, we need to be careful to distinguish genuine perception from various sorts of perceptionlike experiences, and we need to consider what sorts of objects and what sorts of properties are in fact perceivable. In this chapter, I try to show that spatial limitations are less restrictive than they might seem to be, and that experiencing states of affairs as if they were at a place where they are not need not disqualify them from being genuine perceptions.

1. Perceiving Remote States of Affairs On the face of it, it might seem impossible for us to perceive events that occur on the other side of the globe. Our vision doesn’t extend that far; it doesn’t turn corners or pass through solid rock. On the other hand, we speak to our friends on the other side of the globe by phone and we watch disasters unfold via television. Are these not cases of perception also—cases of actually hearing a friend’s voice, and of actually seeing a house go up in flames? Given the fact that all perception involves processes that mediate between the perceived object or event and our perception of it, it would be hard to discount the possibility of perceiving via phones and cameras on the grounds that these instruments are causal intermediaries between the source of our information and our reception of that information. (I have already argued against understanding perceptual immediacy as physical or causal immediacy.1) And given how automatically and unconsciously we interpret the data we receive from phones and cameras, it would be hard to argue that 1 See Chapter 1, Section 2. Note, too, that given how normal phones and cameras (let alone eyeglasses and hearing aids) have become, dismissing these causal pathways as non-standard or abnormal seems arbitrary at best.

1.

PERCEIVING REMOTE STATES OF AFFAIRS

141

the knowledge they bestow is not perceptual because it is inferential. We do not first have the impression of a small house burning inside our television, and then infer that a distant house is responsible for this impression; rather, we see the distant house via the television.2 Still, there seems to be an important difference between coming to know what someone said via a radio recording and coming to know what someone said via a friend’s recounting, or between knowing the extent of a fire from images beamed from an overhead satellite and knowing the extent of a fire from a pencil drawing—even though all of these ways of knowing seem to occur without any reliance on inference. Is there some other criterion, apart from causal or inferential directness, by which to determine when and if distant states of affairs are perceived? Gareth Evans considers this question, and notes that “in the ordinary perceptual situation, not only will there be an information-link between subject and object, but also the subject will know, or will be able to discover, upon the basis of that link, where the object is.”3 For perception (or what Evans calls “standard cases of demonstrative identification”), it is not enough that a house is the source of my information; nor is it enough that I know that a house is the source of my information—knowledge that would enable me to think of a particular house under the description “the house that caused this visual experience.”4 The information I am receiving must enable me to locate that house in space. There are several different ways in which we might be able to locate an object in space on the basis of an information-link, however; and not all of them are relevant to perception. Consider the following cases.

2 There are various ways of distinguishing inferential and non-inferential knowledge, of course, and not everyone agrees that inference must be conscious. For a fuller discussion of the possibilities and their problems, see Chapter 1, Section 2. 3 Evans (1982), p. 170. Similar statements are found on pp. 150 and 173. An externalist such as Burge (2010) or Millikan (1984) maintains that the existence of an appropriate information-link suffices for perception, whether or not the subject has any knowledge of such a link. But appropriate information links for perceptual content are typically understood to be those that enable an organism to keep track of the object or the property in question. So the discussion below of a locatability requirement is relevant to externalists as well. 4 Evans (1982), Chapter 6, distinguishes between fundamental and non-fundamental demonstrative identification, with non-fundamental cases identifying a particular via demonstrative reference to a fundamental demonstrative experience of that particular. My Case Four, below, would give rise to a nonfundamental demonstrative thought along the lines of “the house that is at the other end of the information-link responsible for my present experience.”

142

THE FURTHER REACHES OF PERCEPTION

Case One: We might know enough about television technology to be able to trace the causal path that leads back from the screen image—through cables, across airwaves, into the studio, through the original camera, and so on—to the photographed house. This could amount to a practical ability to locate the house in space on the basis of an information-link that connects us to the house, but our locating ability would be based on the character of the link rather than on the character of the information conveyed through that link. Evans, presumably, would insist that perception enables us to locate its object on the basis of the information transferred through the link (the content of the resulting experience) rather than on the basis of auxiliary information about the link itself.5 Case Two: Sitting on our living room couch miles away, we do not know the location of the televised house. Still, as the camera zooms in, we get a vivid idea of what the house looks like and what its surroundings look like—good enough that we would recognize it if we came upon it. As a result, if we wanted to go to the trouble and the effort of finding that house, we could; for we could systematically search all areas with that sort of shoreline, or that sort of architecture, until we came upon this particular house. This would be a case of locating the house on the basis of the information received through a particular information link, but this ability to succeed in trial-and-error matching is not the sort of locating ability that Evans has in mind. When we perceive a house, the information received is supposed to tell us something about the house’s features and something about its position relative to us—its position in what he calls “egocentric space.” Case Three: We might see a familiar landscape behind the televised house—a landscape that tells us, immediately, just where the featured house is located relative to our present position. We recognize a distinctive outcropping in the Hudson River, for example, and are able to locate the house relative to that. In this case, as in Case Two, our ability to locate the house is based on the information that is conveyed through the television link, rather than knowledge of the link itself; but unlike Case Two, the information received enables us, immediately, to find our way to the house in question. 5 Another case of object location based on the link but not on the information conveyed through the link would be a case where one wrote to the television reporter to ask for the location of the televised house.

1.

PERCEIVING REMOTE STATES OF AFFAIRS

143

It seems odd, however, to suggest that the person who is familiar with a river in the background perceives the televised house while the person unfamiliar with the background does not. Evans could accept that the person who recognizes the outcropping has demonstrative knowledge of the house while the person unfamiliar with the outcropping does not— without insisting that the one perceives the house while the other does not. But since (fundamental) demonstrative identification was supposed to be precisely the thing that perception enables us to do, this response would undermine his attempt to make sense of perceptual knowledge by reference to demonstrative identification. Alternately, he could deny that the person who recognizes the house has demonstrative knowledge of the house, since the information conveyed through the television link does not suffice for locating the house; the television watcher must also have information about the location of the television relative to the house. But then he must explain (without appealing to perception, which is also being explained) why my knowledge of my position relative to the television is more immediate than my knowledge of my position relative to my house, relative to my street, relative to my town, and so on. Case Four: There might be an interactive information-link that enables a subject situated in a “control room” to scan a house and its immediate surroundings at will (changing the angle or the location of the camera, for example), and to affect the house in various ways (beaming a bright light at it, or using a robot to open and shut doors, for example). A worker hired to monitor security cameras mounted at an unspecified (possibly overseas) location has little knowledge of the house’s location in relation to herself or in relation to anything outside its immediate surroundings, but she has quite nuanced knowledge of how its various parts are located relative to each other and relative to its immediate surroundings. And she is able to effectively interact with the house in question. Insofar as this worker is not able to locate the house in egocentric space, she would seem to lack precisely that skill that is definitive of demonstrative identification (and, presumably, of perception) according to Evans. And yet, in many ways, her knowledge of the house seems to be more intimate and more reliable than that of the television watchers of Case One, Two, and Three. Where does this leave us with respect to the possibility of perceiving distant states of affairs via television? Following Evans’ suggestion that perception gives us the ability to locate its objects in space, we considered

144

THE FURTHER REACHES OF PERCEPTION

various interpretations of this ability and found them all problematic. The ability described in Case One is the sort that would enable one to discover the location of the televised house with the greatest reliability, for it is only by tracing out a particular causal chain that we can eliminate duplicates (other houses in other locations may look quite similar, after all). But the ability to actually trace the causal chain leading from the recording camera to my television screen is wholly independent of the experience of watching a house on television (and vice versa). Tracing out the causal chain may be the best way to ascertain the truth of my experience (and thus its status as a perception rather than an illusion), but it isn’t an ability that is given to me by my experience of the house. The locating ability described in Case Two, which relied on a search-and-match mission, was also too external to the experience of watching a television screen to have a bearing on its status as perception. The locating ability described in Case Three turned on the familiarity of the house’s surroundings; and familiarity with a house’s surroundings seems irrelevant to whether or not that house is perceived. Finally, Case Four involved locating abilities that are very much internal to the experience of perception, reliably mimicking the network of experiences that we are likely to have when we are close to a house,6 yet these abilities were the least likely to enable one to discover the location of a particular house. Evans, as we have seen, puts a premium on perception’s ability to use information to locate objects in space.7 This, for him, is a prerequisite for knowledge of objects as objects—which, in turn, is a prerequisite for all knowledge whatsoever, since that is the only way to ensure that our judgments have a referent (of which they may be either true or false). I agree

6

Hurley (1998) has used psychological experiments and neurological evidence to argue for the importance of the linkage of perception to action. Milner and Goodale (2006), on the other hand, document a distinction between the “ventral stream” system that serves “vision for perception” and a “dorsal stream” system that serves “vision for action.” Although these two systems are normally collaborating, they can also be dissociated—without, it seems, losing perceptual capacity. But would dorsal stream vision be unconscious and thus not perception? Hopkins (1998), p. 186, claims that in vision as opposed to visualizing one’s spatial frame of reference controls action immediately. 7 Evans (1982), pp. 149–50, emphasizes the embodied nature of perception—our reliance on our own sense of up versus down, for example, in arriving knowledge of objective space. And he suggests that “primitive people” who thought the voice on the radio came from a person inside the radio failed to make a (fundamental) demonstrative identification because they could not accept that that man was located at a distant rather than a nearby place.

1.

PERCEIVING REMOTE STATES OF AFFAIRS

145

with him about the importance of locating objects in space if we are to have knowledge at all, but I do not think it is an object’s location relative to us (in egocentric space) that matters to perception so much as an object’s location relative to various other positions in objective space (in allocentric space). Generally, of course, we are able to locate an object’s position in space by locating it in relation to the locations that we ourselves occupy (and vice versa: we are able to locate ourselves by locating ourselves in relation to the locations that various objects occupy); and, generally, we are able to identify our own location at a given time with the point of view from which an object is experienced at that time. Cameras produce images (on paper, on television screens, and on movie screens) that enable us to experience and to coordinate points of view that are quite distant from our present location, however; and they enable us to locate objects in objective space when we are unable able to locate them in egocentric space. I can experience a house from the point of view of an on-site camera rather than the point of view of my seat on my couch, or my surveillance room, and those experiences can enable me to experience the house’s spatiality even more accurately, and with even more confidence, than were I to experience it from the point of view of my present location. What matters to perception, then, is not a subject’s ability to experience an object in egocentric space but, rather, her ability to experience an object in objective space; and that depends on the coordination of multiple points of view, whether or not the location of those points of view is the same as the location of the viewing subject.8 Pursuing these conclusions a step further, it should be evident that, on this understanding of perception, a scientist who views distant galaxies through computers connected to an automated observatory could perceive these galaxies—even if she has no idea of where those galaxies are relative to her current position.9 (Perhaps she has lost track of the hour, perhaps she was introduced to this project without ever having learned just where the studied galaxy is located relative to the galaxy we live in, or perhaps the time lag between the emission of light from that galaxy and its reception on 8 Things get more complicated if one thinks that subjects are necessarily located at the point of view from which they experience things—that to the extent that we can experience points of view at locations other than that occupied by our body, we, as subjects, are located in those positions. 9 Peacocke (1983), p. 153, and Burge (2010), p. 200, introduce examples in which a perceiver may not know the location of an object because atmospheric refraction, mirrors, or prisms intervene. I push things further by introducing the intervention of a computer.

146

THE FURTHER REACHES OF PERCEPTION

earth is still unclear.) What she must have is an immediate, non-inferential ability to locate various features of the distant galaxy in relation to their surroundings rather than in relation to her own—an ability that will depend on her capacity to coordinate a variety of imagined perspectives or modes of accessing those features.10 Locating objects or events in time is similar. The television watcher or the astronomer in front of her computer will usually be able to recognize the temporal location of a burning house or a planet’s position relative to other events viewed via the screen—thus giving her an experience of events in objective time as well as objective space—even if she is unable to accurately locate those events relative to her present position in time. We see the house burning after the explosion and before the smoldering ashes, but we may not know whether this series of events took place five minutes ago or fifty years ago. The astronomer sees the planet passing to one side of the star, then moving away from that star, but she may not know whether these events took place five years ago or five hundred years ago. If the television or computer screen shows us an unchanging house on an empty street, or an unmoving planet—preventing us from recognizing its temporal relation to other objects and events—we can still perceive it (not just think about it) as located in objective time, insofar as we effect a synthesis of different imagined temporal perspectives. We experience the house, or the planet, as persisting across time because we actively imagine the view at time-1, time-2, time-3, and so on (without needing to know when those times are located relative to now). The indistinguishability of these different temporal perspectives is no different from the indistinguishability of different perspectives on a perfectly symmetrical sphere; it is the variations in imagined perspectives, not the variations in the imagined properties that fulfill the requirements for experiencing objectivity.11 10 This is a variation on Robert Brandom’s (1994) discussion of a scientists’ ability to see mu mesons in a cloud chamber. Like Brandom, I assume that the scientist in question does not infer locations from the received data; rather, it is directly evident. Brewer (1999), p. 250, citing the case of watching an electrical current via a meter, calls such cases “inferential perceptual knowledge.” Like mu mesons and electrical currents, distant galaxies cannot be seen with the naked eye, but with technical assistance (and training) they can be seen. Unlike seeing mu mesons in a cloud chamber, however, the location of distant galaxies relative to the viewer is not usually known. 11 Photographs do not usually encourage the imagining of different temporal perspectives; but there are exceptions (photographs of a bomb hitting a building, for example; or the photographs of Jean-Luc Mylayne). In any case, the experience of objectivity depends on synthesizing multiple perspectives, without any need for these different perspectives to be distributed across time.

2.

PERCEIVING FUTURE STATES OF AFFAIRS

147

Insofar as it is possible to perceive distant objects, it will also be possible to perceive various relations that exist between those objects. If we can see a house via television, then we can also see its nearness to an explosion; and, drawing on the arguments of Chapter 3, Section 2, we can also see that the explosion caused the house to burn. Likewise, if the astronomer can perceive a distant star and a nearby planet via a computer, she might also perceive that the star is causing the planet to follow a particular path. We can also see via TV how the constituent parts of a house explain its ability to remain standing; and an astronomer might see how the constituent parts of a galaxy explain its shape. To the extent that one accepts the possibility of perceiving objects or states of affairs that are distant in space and time, one should also accept the possibility of perceiving causal relations and constituent relations between objects or states of affairs that are distant in space and time.

2. Perceiving Future States of Affairs What about perceiving objects or events that are distant in the future rather than the past? Initially, this sounds implausible since it seems to require an effect to precede its cause. Moreover, it seems to require there to be an information link between subject and object before the object even exists.12 We do, of course, speak of seeing what will happen—when we watch someone trip on the steps and anticipate just how they will fall, when we hear a comment and immediately know what the response will be, or when we track the path of a hurricane in order to predict where it will land. Are these all cases in which it is as if we perceive rather than cases in which we do in fact perceive—cases in which we merely imagine (quite accurately) what is about to happen rather than perceive it? The answer to this question will depend on whether perceptual experiences must be caused by their objects and, if so, just what notion of causality is appropriate. In this section, I consider reasons for requiring perceptual experiences to be caused by their objects; whether, in light of those reasons, a counterfactual relation between

12

I restrict myself here to future events that actually do occur. In Chapter 5 I turn to the possibility of seeing that something ought to occur (whether or not it actually occurs).

148

THE FURTHER REACHES OF PERCEPTION

perceptions and their objects suffices;13 and, finally, whether present events can be counterfactually dependent on future events (in the relevant way). One reason for supposing that perceptions must be caused by their objects concerns the supposed connection between knowledge and reliability on the one hand and reliability and causality on the other. If someone’s impression of a red stripe on the ground before her is caused by her disfunctioning brain rather than by an actual red stripe on the ground, then we are disinclined to call it a case of perception (or knowledge)— even if there is a red stripe just where it appears to be, and even if she is unaware of the relevant brain disfunction. Her belief is true but it is not justified insofar as justification depends on acquiring one’s belief in a way that is reliably truth-preserving. Unless one’s sensory experiences are reliably accurate in a given domain, they cannot count as knowledge and they cannot count as perceptions. (Indeed, the very content of an experience—a red stripe—may depend on the reliability with which experiences of that type are correlated with actual red stripes.)14 It requires another step, though, to get from the need for reliable accuracy to the need for a causal connection to the object of one’s experience. Dretske, for one, wants to separate the need for reliability from the need for a causal connection, insisting that perception only requires reliability, not causality. More precisely, he maintains that S perceiving that b is P requires conditions to be such that things would not look as they do to S unless b were P, regardless of why this might be the case.15 In short, he requires a perception to be counterfactually dependent on its object, but he does not assume that this requires the object to be the cause of the perception. Many might claim that such reliable correlations—between what is experienced and what is in fact the case—are themselves indicative of a causal connection between the two since reliability depends on counterfactuals 13 Chapter 3 canvassed several different understandings of causality, and though the possibility of perceiving causal relations did not seem to depend on any particular understanding of cause, a counterfactual analysis of cause was favored. 14 Spelling out the relevant notion of reliability is a challenging enterprise. Sosa (2007) has recently presented an interesting version of a reliability theory. Williamson (2002) defends a more radical alternative. 15 Dretske (1969), see especially pp. 52–3 and p. 183. Snowdon (1979) also argues against the need for a causal connection between perception and its objects. Bonjour (1998) requires “responsiveness” and “a relation of influence” that does not operate through space and is not therefore causal. His preliminary account of how such influence works in the case of rational insight draws on medieval notions of form instantiated in matter.

2.

PERCEIVING FUTURE STATES OF AFFAIRS

149

whose truth is best explained by the presence of some causal connections. Alternatively, one might treat the counterfactuals as not indicative of but constitutive of causal connections. This could be spelled out in a number of different ways—by supposing that x causes y just in case the occurrence of x (under conditions C) sufficiently raises the probability of y, or just in case the non-occurrence of x (under conditions C) guarantees the non-occurrence of y, and so on. The details do not matter for our purposes, however, since there is nothing to prevent a present (or past) event from being counterfactually dependent on a future event.16 If I am a reliable predictor of rain, then the following counterfactual will be true: If it were going to rain tomorrow, then I would have predicted rain; and if it were not going to rain tomorrow, then I would not have predicted rain. I do not need to be a perfect predictor in order for this counterfactual to be true, for the truth of any counterfactual assumes relevant background conditions (arrived at by standardization, or idealization, for example), and my reliability vis-a`-vis future rain is reliability under conditions that do not always obtain. The more distant a perceived state of affairs is (whether this be distance in space or distance in time), the greater the opportunity for intervening states of affairs to disrupt the causal chain that links perception to its objects. There are many ways to disrupt the causal pathway that leads from a distant galaxy via a telescope via a computer image to our perception of that galaxy, for example; and each possible disruption adds to the unreliability of our observations. The relevant counterfactuals (e.g. that we would not have had the experience we had unless there were such a galaxy in that position) become weaker and weaker (less and less probable) or more and more qualified (assuming a more and more extensive set of background conditions in order to be true). Similarly, when an event is more distant in the future, our present predictions will be less reliable on account of the increased likelihood of intervening events that diminish the reliability of our current evidence. Both past and future events become less likely to be perceived as they become more distant, then, because the way in which we are linked to them (causal or not) becomes less and less reliable.17 16

See Chapter 3, Section 2, for a fuller discussion of the options. One could add a temporal order requirement to a counterfactual analysis of causation, of course, but there would have to be an independent reason for such an addition—or it would be question-begging (in this context, anyway). 17 Yablo (2003) maintains that spatial contiguity enters to avoid mere correlation, but that counterfactual analyses help to rule out mere correlation—as long as scope of cause is proportional (not

150

THE FURTHER REACHES OF PERCEPTION

By focusing on reliability rather than causation, we bypass various controversies surrounding the nature of causation, and we also bypass any need to identify spatiotemporal pathways running between a perception and its objects. It is not clear that we can simply ignore such pathways, however. Given our earlier insistence that perception is an experience of the objectivity of a state of affairs, and that the objectivity of a state of affairs depends on imagining multiple ways of accessing that state of affairs, isn’t it still necessary to be able to identify spatiotemporal pathways to and from the objects of perception? In Section 1, above, we argued against Evans’ suggestion that perceived objects must be locatable relative to oneself, in egocentric space; I do not need to know where a televised house is relative to me in order to perceive it. We agreed, however, on the importance of experiencing the objects of perception as located in an objective space, and that depends on actively imagining different perspectives on that object in space. We must be careful here to distinguish between the need to imagine different perspectives on an object—something that I have argued is essential to all perception—and the need to imagine a causal chain stretching between a viewer and the object viewed. Imagining different perspectives requires us to imagine different spatial positions relative to the object in question, and that will immediately suggest different pathways across the intervening space. But spatial pathways need not be causal pathways, and the imagining of spatial relations does not depend on the imagining of causal relations. (Causal relations, furthermore, need not be experienced as following a particular pathway through space; we do not need to be convinced that there is a continuous series of spatiotemporal events linking viewer and viewed in order to be convinced that our perceptions are caused by their objects.18) In sum, there is nothing in a counterfactual understanding of causality that excludes backwards causation, and the further one diverges from a counterfactual account of causation, the weaker the connection between causality and reliability and the less motivated the requirement that

egregiously, pointlessly weak or strong—as with parasitic disjuncts and conjuncts) and natural. “Wide causes contain irrelevancies because all causes do.” (p. 318) 18

The possibility of action at a distance has been in dispute at least since the time of Newton. It would be a mistake, therefore, to rely on an understanding of causation that excludes this possibility.

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

151

perception be caused by its object. My own adherence to a counterfactual analysis, as elaborated in Chapter 3, thus allows for backward causation. But even if I were to reject that account of causation in favor of an account that disallowed backward causation, I would not automatically discredit the possibility of perceiving future events, since my understanding of perception is more committed to counterfactuals than to causation. Once we allow for the possibility of perceiving future events, we also allow for the possibility of perceiving relations between future events. If I can see how a child will fall forward in the near future, there is nothing to prevent me from also seeing how its fall will bring about various other events—from the breaking of a dish to an interrupted conversation, to an onlooker jumping up, and so on. Showing how the perception of future events is theoretically possible does not establish its actuality, of course. But given that some of our judgments about the future (the judgment that a child is about to fall, for example) have the same epistemic standing as perceptual judgments about the present—being non-inferential judgments to which we are entitled without further justification—and given the fact that they have the same phenomenal character as perceptual judgments about the present—presenting the relevant state of affairs as occurring in objective space, the elimination of theoretical hurdles leaves us without any principled reason to deny their status as genuine perceptions.19

3. Perceiving Abstract States of Affairs We have now defended the possibility of perceiving a wide range of relations (including causal relations, constitutive relations, and justificatory relations) that obtain between spatiotemporal objects—including objects that are spatiotemporally distant or not yet existent. Even if we think of these relations as abstract—either because we consider all properties and relations to be abstract (since many different individuals are capable of 19

The neurology/physiology of short-term anticipation is also very like the neurology/physiology of short-term memory. With increasing temporal distance, whether past or future, we tend to rely less on imagining and more on narrative descriptions. (See Berntsen and Bohn (2010).) As we have seen in the case of astronomical observations, though, it is not temporal distance per se that makes the difference; what matters is the phenomenal structure of our experience.

152

THE FURTHER REACHES OF PERCEPTION

having the same relation to one another) or because we consider counterfactual relations in particular to be abstract (since they invoke various non-actual possibilities)20—the objects that we perceive as standing in these relations have not been abstract; they have been spatiotemporal particulars. But, of course, properties and relations can themselves be the “objects” of further properties and relations. For example: the greater-than relation has the property of transitivity, the property of being a circle is an instance of the property of being a closed figure, legal ownership has the property of being regulated by the state, and the property of being a mother is a subset of the property of being a parent. Whether or not we speak of the relevant properties and relations as “abstract objects,” it will be convenient to refer to the state of affairs in which a property or relation possesses some other property, or stands in some other relation, as an abstract state of affairs. The question before us now, then, is whether we can perceive abstract states of affairs. Insofar as we can defend an affirmative answer, we will have taken a significant step towards vindicating the historically important notion of “rational perception.” Bonjour (1998) has argued that we must have perceptual knowledge of various principles, including principles of logic, since we cannot acquire knowledge through the evidence of our senses without relying on such principles to determine what counts as confirming evidence (and what counts as disconfirming evidence). Principles such as the Law of NonContradiction cannot support our inferences unless they are themselves known, they cannot be known through inference (without prompting an infinite regress), so they must be self-evident—that is, they must be perceived. Bonjour also appeals to phenomenological considerations to support his claim that we have perceptual knowledge of logical and other abstract facts, but he acknowledges that “no one has managed to explain sufficiently clearly just what an act of rational insight could amount to, ‘how some cognitive act, of a sort that we might plausibly enjoy, is able to yield immediate knowledge of the modal properties of properties’.” 21 Some of Bonjour’s critics, despairing of such an account, have defended the importance of knowledge that is “blind”;22 others have claimed that because 20 The term “abstract” can be used to characterize (a) features that are characteristic of many different objects or situations or (b) features that exist independently of whether or not they are instantiated at all. 21 Bonjour (2001), p. 673; embedded quote is from comments by Boghossian (2001). 22 E.g. Boghossian (2001).

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

153

justification is holistic, there is no need for the truths of logic to be selfevident.23 I share Bonjour’s instincts and aspirations, however, and think that the account of perception I have been defending can go a long way towards meeting the challenge he cites. Once one agrees that we can perceive dependencies that exist between spatiotemporal states of affairs by imagining appropriate counterfactual possibilities, the “cognitive act” by which we might also perceive dependencies between abstract states of affairs might seem obvious. If we perceive that the movement of a segment of water depends on the speed and angle of a particular paddle by imagining relevant counterfactual possibilities and noting what remains constant, why can’t we also perceive that the movement of liquids (in general) depends on the speed and angles of intervening surfaces (in general)—also by imagining an appropriate range of possibilities and noting what remains constant? If we can perceive that a friend’s suffering warrants a comforting word by noting certain constancies across alternate possibilities, why can’t we also perceive that suffering (in general) warrants a comforting response (in general)—again, by imagining a relevant range of alternatives and noting what remains constant? It is certainly possible to imagine a range of cases involving different particulars, and to notice what is common to all of those cases. I imagine different sorts of liquid affected by different intervening objects, and I recognize the role of pressure in all such cases; or I imagine ideal responses to different kinds of suffering and realize that they are all cases of providing comfort. Sensitivity to commonalities (in this case, high-level commonalities) across a range of cases is not enough for perception, however; for the phenomenology of perception, I have argued, is the phenomenology of objectivity, and discerning what is common across a range of imagined cases, while important, is not enough to ensure that the commonalities are experienced as independent unities rather than convenient groupings of overlapping properties, for example. In order to experience a higher-order property such as pressure or comfort as objective properties, different instances of that property—the pressure of a paddle, the pressure of a hand, the pressure of a boat, and so on—must be experienced as different occurrences or manifestations of a single thing, not just a series of forces 23

E.g. Gendler (2000).

154

THE FURTHER REACHES OF PERCEPTION

with notable similarities.24 But, as I argued in Chapter 1, perceiving the unity of something requires us to experience it as located in space, for only then is its objectivity self-evident. The deep difficulty, then, with perceiving abstract states of affairs can be put in the form of a dilemma: either the object of one’s knowledge is experienced as situated in space, in which case it is not experienced as abstract; or the object of one’s knowledge is not experienced as situated in space, in which case it is not a perceptual experience. Or, put another way: either one experiences an abstract object as a spatial object, in which case one’s experience lacks the veridicality required for perception; or one experiences an abstract object as a non-spatial object, in which case one’s experience lacks the phenomenology required for perception. Avoiding the horns of this dilemma depends on showing how an experience can present non-spatial things as spatial yet still get things right, how the requisite spatiality can reveal the objectivity of a state of affairs without spatiality being attributed to that state of affairs. Below, I explore three different ways in which abstract objects can be experienced as spatial, showing how the second and third circumvent the above dilemma, but only the third can disclose versus merely display the objectivity of abstract objects. One way in which abstract objects might be experienced as spatial is exemplified by synesthetes who experience sounds as having shapes.25 The sound of a trumpet, for example, might be experienced as saw-toothed, or a low E might be experienced as rectangular. While some synesthetes have thought that sounds actually do have shapes (and colors, and tastes),26 most are quick to agree that in fact sounds do not have shapes, and that the shapes are merely imagined. There are a couple of different ways to understand such disclaimers. On the one hand, they might express a synesthete’s readiness to let various background beliefs override appearances 24 Whether this is the right way to think of higher-order properties is a different matter. The metaphysics of properties, and of types, is an important and unsettled issue. I am not taking a stand on whether properties and types are objective unities—only on what it takes to experience them as objective. If such experiences are mistaken, they won’t qualify as genuine perceptions. My account of what makes them perception-like would then be a contribution to understanding an epistemic mistake rather than an epistemic success. 25 What makes the synesthete different from others who also associate certain shapes with certain sounds is the vividness and unavoidability of the synesthete’s associations. This gives the synesthete’s associations more reliability, which is relevant to their potential role in acquiring knowledge. 26 Messiaen, for example, was a Catholic mystic who believed that his experiences of sounds as having shapes and colors gave him insight into the deeper character, and the deeper unities, of the world.

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

155

(as when I say “it looks like a crow but it isn’t”); the experiences of synesthetes would then count as misperceptions—thereby falling prey to the first horn of the above dilemma (the experience must be veridical in order to be perceptual). On the other hand, the synesthete may be claiming that, in her experience, trumpet sounds are always accompanied by sawtoothed images and low Es are always accompanied by imagined rectangles. In that case, though, while there is no misperception concerning the sounds, there is also no presentation of sound as spatial (and hence objective)—thereby falling prey to the second horn of the dilemma. This does not rule out the possibility of a synesthete using spatial associations to acquire knowledge of abstract states of affairs—of relations that hold between different sound properties, for example. Listening to a wide range of pitches, one might notice that the saw-tooth shape only appears in the middle range, and one might (rightly, and justifiably) conclude that a trumpet sound is not compatible with very low and very high pitches. Such knowledge would not be perceptual, however, since the spatial image would be serving as a reliable sign only; there would be nothing internal to the image to attest to the truth of this conclusion. Let us turn, next, to a case where a spatial image serves not merely as a sign of an abstract state of affairs but as an expression or translation of that state of affairs. A graph that depicts fluctuations in the GNP over time will typically present a greater GNP as farther up rather than farther down on the page, and will typically portray the GNP at an earlier time as farther left rather than farther right on the page. In reading such a graph, we do not perceive the GNP as further up or down, left or right on the page. But we do take spatial properties of the graph to be expressive of properties of the GNP. Given our background familiarity with the conventions of graphs, and given our trust in its creator, the GNP fluctuations become immediately apparent to us; no (conscious) deliberations or calculations are necessary. (This example is like the case of perceiving electrical current through an (analog) meter, which Dretske calls “secondary seeing” and Brewer calls “inferential perception.”27 The perceived electricity, however, is a particular spatiotemporal event, not an abstract property.) 27 Dretske does not consider this to be inferential because the implication is drawn so automatically, and unreflectively; Brewer, on the other hand, allows inferences to be automatic and unreflective (eventually, anyway, even if not initially).

156

THE FURTHER REACHES OF PERCEPTION

Unlike the above case of synesthetic associations, there is a structural alignment between spatial properties and non-spatial properties that allows us to explore different perspectives on a past GNP by exploring different perspectives on the graph. We might approach the GNP in a given year from a perspective prior to that year (orienting ourselves from a position farther left on the graph), or we might approach that same GNP from a present perspective (orienting ourselves from a position farther right on the graph). Viewed from the left perspective the GNP may increase, while from the right it may decrease.28 Spatial renditions of non-spatial properties structured so as to give us simultaneous access to multiple perspectives are thus able to generate an experience of the objectivity of non-spatial properties. Most spatial expressions of non-spatial properties are less rigorous and less quantitative. The psychological dynamics of a classroom, for example, can be expressed as a kind of force field in which some people send out more powerful signals while the signals of others carry farther. Some forces pull people towards one another while others repel, the air is sometimes thick with electric sparks, and other times a vacuum. Experiencing a classroom in this way is not merely a matter of finding that certain images are reliably correlated with certain classroom dynamics. (That would make it like the synesthete’s case, discussed above.) We come to experience the classroom dynamics through the image, not alongside the image.29 The image allows us to coordinate the information we’ve received into an intelligible whole, and it guides our intervention quickly and automatically in ways that deliberation could not. We sense disagreements between students as filling the air with generative energy, as drawing everyone into a spiral, as creating a point of tension that is about to shift, and so on; and we act accordingly, depending on our position in the force field, our aims, and the surrounding conditions. Our interventions, in turn, can serve to reinforce or revise the picture we have of the classroom dynamic. Nothing in this image is as precise or as purely quantitative as that of the GNP graph, but it is equally

28

Note that, as in the case of television (or a computer screen in an observatory), as discussed in Section 1 above, perception does not require us to locate a particular object or event relative to ourselves—in egocentric space and egocentric time. It is enough that we can locate it relative to its own surroundings—in allocentric space and allocentric time. 29 This distinction is discussed further in Chapter 6, and in Church (2000).

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

157

synthesizing, informative, and conducive to the experiencing of the objectivity of classroom dynamics.30 These sorts of cases escape the dilemma posed above, since they are cases in which the imagined spatial properties enable one to experience the objectivity of an abstract state of affairs (thus making the knowledge perceptual) without, however, presenting abstract facts as possessing those spatial properties (i.e. without threatening the veridicality of the experience). By transforming non-perceptual knowledge into perceptual knowledge, the objectivity of its objects becomes self-evident—without losing sight of the abstract nature of those objects. But can such perceptions be a source of knowledge rather than a mere expression of knowledge? Is it possible to use spatial renderings to produce new knowledge rather than merely reflect the knowledge that one already has? Although this isn’t a requirement for knowledge being perceptual, it is certainly one of the reasons philosophers have been interested in the possibility of perceiving abstract states of affairs. If we can expand our perceptual capacities in a given domain, we should be able to expand our knowledge as well.31 Wittgenstein has been an influential critic of the notion that images can give us knowledge of anything, especially of such abstract things as the meaning of a word.32 He allows that our ability to conjure up appropriate images may express our knowledge of a word’s meaning, or a mathematical function, for example, and that recognizing or producing such expressions may even be criterial of (though neither necessary nor sufficient for) having knowledge of such things.33 He argues against the temptation to think of images as supplying us with knowledge of abstract things, however. Thus, it will be useful to review his arguments before making a case for the knowledge-generating capacity of our spatial renderings of abstract things. 30 For examples of how degrees of similarity between sensory qualities generate a multidimensional “phenomenal space” in which greater similarity is expressed by greater proximity, see Clark (1998). 31 Different historical figures have had different understandings of just how the ability to perceive abstract entities might expand our knowledge of them. I explore some of the central claims of Plato and Descartes in the next section. 32 There is an interesting contrast between Wittgenstein’s early and late work in this regard— possibly correlated with a shift in focus from the meaning of words to the mental states of people, and to the objects depicted by drawings. 33 There are different ideas about what Wittgenstein meant by a criterion; the basic idea, though, seems to be that abstract entities (meanings, mental states, games, etc.) are constituted by an array of perceivable properties—each of which necessarily counts as evidence for the entity in question, but none of which counts as either necessary or sufficient for its existence.

158

THE FURTHER REACHES OF PERCEPTION

Many of Wittgenstein’s examples are designed to show that the images that accompany our knowledge are optional and variable. An important, and representative, passage is the following: ä151. . . . A writes a series of numbers down; B watches him and tries to find a law for the sequence of numbers. If he succeeds he exclaims: “Now I can go on!”—So this capacity, this understanding, is something that makes its appearance in a moment. So let us try and see what it is that makes its appearance here.—A has written down the numbers 1, 5, 11, 19, 29; at this point B says he knows how to go on. What happened here? Various things may have happened; for example, while A was slowly putting one number after another, B was occupied with trying various algebraic formulae on the numbers that had been written down. After A had written the number 10 B tried the formula an = n2 + n 1; and the next number confirmed his hypothesis. Or again, B does not think of formulae. He watches A writing his numbers down with a certain feeling of tension and all sorts of vague thoughts go through his head. Finally he asks himself: “What is the series of differences?” He finds the series 4, 6, 10 and says: Now I can go on. Or he watches and says “Yes, I know that series”—and continues it, just as he would have done if A had written down the series 1, 3, 5, 7, 9.—Or he says nothing at all and simply continues the series. Perhaps he had what may be called the sensation “that’s easy!” (Such a sensation is, for example, that of a light quick intake of breath, as when one is simply startled.) ä152 . . . . [I]t is perfectly imaginable that the formula should occur to him and that he should nevertheless not understand. “He understands” must have more in it than: the formula occurs to him. And equally, more than any of those more or less characteristic accompaniments or manifestations of understanding.34

The fact that the phenomenology of perception is neither necessary nor sufficient for knowledge of a particular fact does not mean, however, that a given image makes no epistemic contribution to our knowledge of that fact. In the first place, Wittgenstein is keenly aware of the possibility that we may be able to establish criteria for the correct application of certain terms (such as “intelligence” or “knowledge”) without being able to provide necessary and sufficient conditions. The phenomenology of perception may not be necessary or sufficient for knowledge, but if it is a criterion of knowledge

34

Wittgenstein (1953).

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

159

then its presence, together with the presence of assorted other criteria, could help to constitute a given state of knowledge. Furthermore, the fact that no particular phenomenological state is necessary for understanding a mathematical function, for example, does not show that understanding can occur in the absence of any phenomenological state whatsoever, or that the character of the phenomenology is arbitrary. The phenomenology of seeing a cube may be different from the phenomenology of feeling a cube, and neither one of these particular modes of perceiving may be necessary for knowing what a cube is, but it would be wrong to think that we could have knowledge of shapes in space without experiencing any phenomenological states whatsoever, or that the phenomenology of seeing a square is arbitrary. Finally, though, and perhaps most importantly, phenomenological states may be necessary (and perhaps even sufficient) for certain ways of knowing—more versus less direct ways of knowing, or more versus less reliable ways of knowing, for example, and these ways of knowing may constitute the necessary foundations for all other less direct or less instantaneous ways of knowing. Nothing in Wittgenstein’s example, then, shows that the phenomenology of perception is not essential to an epistemically privileged way of knowing. Wittgenstein’s deeper worry emerges out of a particular conception of what knowledge is—for Wittgenstein equates the possession of knowledge with the possession of certain abilities, and abilities are not the sort of thing that can appear “in a flash.”35 ä138 . . . . [W]e understand the meaning of a word when we hear or say it; we grasp it in a flash, and what we grasp in this way is surely something different from the “use” which is extended in time! ä139. When someone says the word “cube” to me, for example, I know what it means. But can the whole use of the word come before my mind, when I understand it in this way? Well, but on the other hand isn’t the meaning of the word also determined by this use? And can these ways of determining meaning conflict? Can what we grasp in a flash [mit einem Schlage erfassen] accord with a use, fit or fail to fit it? And how can 35

The following remarks are focused on knowledge of word meanings, but the same points can be made about knowledge of other things. Indeed, if knowledge depends on judgment, judgment depends on understanding concepts, and understanding concepts is equivalent to understanding the meaning of words, then requirements for understanding the meaning of words will be requirements for knowledge more generally.

160

THE FURTHER REACHES OF PERCEPTION

what is present to us in an instant [in einem Augenblick vorschwebt], what comes before our mind in an instant, fit a use?

If we equate the meaning of a word with its use, which is extended in time, we must either reject the claim that the meaning of a word is something we can know in an instant or we must show how the whole use of the word can come before my mind in an instant. Pursuing the first option, we could conclude that knowledge of correct use itself extends across time—specifically, across the time it takes to make appropriate use of a term. I may think that I understand the word “plus” or the word “cube” in a flash, but really such understanding can only unfold over time—the time it takes for me to solve various addition problems or to recognize and compare various cubes, for example. The problem with this conclusion (apart from its apparent discrediting of familiar “aha” experiences) is that it obliterates any distinction between right and wrong use— between uses that accord with a word’s meaning and uses that do not. (Different uses can only reflect different meanings, not misuses.) And without a distinction between right and wrong use, there is no meaning at all.36 There is another way, however, to reject the claim that meanings can be known in an instant. Wittgenstein repeatedly denies that knowing the meaning of a word (or, indeed, knowing anything at all) is a happening (either instantaneous or prolonged)—viewing it instead as the ability to do certain things with words, where abilities are dispositions rather than events. The things that we do with words take more or less time, of course, but the ability to do such things does not; abilities or dispositions may be acquired “in a flash” and they may manifest themselves across time, but they do not themselves occur in time. The mistake, then, is not the mistake of treating knowledge as requiring too little time but, rather, the mistake of treating it as a temporal event at all.37 ä154 . . . . Try not to think of understanding as a ‘mental process’ at all. —For that is the expression which confuses you. But ask yourself: in what sort of case, in what 36 This point is made in a number of different ways by Wittgenstein—as a point about the difference between following a rule and acting in accordance with a rule, and as a point about knowledge versus behavior, for example. To know how to use a word appropriately requires something more than actually using it appropriately; it requires us to experience some uses as appropriate and others as inappropriate. 37 Equating abilities with events is, in effect, a “category mistake.” Note that the same could be said of perceptual knowledge of objects—that perceptual knowledge is not itself an event but an ability to respond in appropriate ways.

3.

PERCEIVING ABSTRACT STATES OF AFFAIRS

161

kind of circumstances, do we say, “Now I know how to go on,” when, that is, the formula has occurred to me?—In the sense in which there are processes (including mental processes) that are characteristic of understanding, understanding is not a mental process. ä157. . . . “reading” meant reacting to written signs in such-and-such ways. This concept was therefore quite independent of that of a mental or other mechanism. — Nor can the teacher here say of the pupil: “Perhaps he was already reading when he said the word.” For there is no doubt about what he did. —The change when the pupil began to read was a change in his behavior; and it makes no sense here to speak of “a first word in his new state.”

Wittgenstein thus makes room for “in a flash” acquisitions of knowledge, but he does not allow for “in a flash” achievements of knowledge. For knowledge consists of abilities or dispositions, not momentary experiences. Note, however, that this view is entirely consistent with the claim that certain “in a flash” experiences are good indications that the relevant dispositions have been acquired; they can signal possession of the dispositions that constitute knowledge. As with the case of the synesthete, discussed above, some “in a flash” experiences may be indicative of something else because they happen to be reliably correlated with that other thing. A certain feeling of confidence, or the image of a watery expanse, for example, could be reliable correlated with newly acquired knowledge, such that the occurrence of this feeling or this image would justify an attribution of knowledge. But this is not usually the case. (The synesthete is unlike the person with a normal array of associations precisely because of the reliability of her associations.) There are other images, though (such as the image of lines climbing and dropping) that are reliably correlated with certain knowledge (of GNP trends, for example) and whose appearance is expressive of, not merely correlative of, the relevant knowledge. The difference, I have argued, is due to the fact that these later sorts of images have structures that replicate the structure of the facts known. So the appearance of such an image is also an appearance of the very abilities that constitute the knowledge in question. If this is right, though, then we have shown how the whole of a word’s use can come before one’s mind in an instant—through simultaneous imagining of its many dimensions.38 38

This does not challenge the important Wittgensteinian point about the infinity of appropriate uses versus the finiteness of actual uses. The imagined uses of a word, like the actual uses of a word, are finite.

162

THE FURTHER REACHES OF PERCEPTION

We are now in a position to understand how “in a flash” experiences can be a source of knowledge as well as an expression of knowledge. Insofar as the spatial structure of the image that is conjured up reflects the structure of abilities that constitute knowledge of the fact in question, then we can rely on the image not merely to illustrate our actions but to guide them. Consider the use of Venn diagrams in solving logical problems. After one learns how to make the relevant translations from sentences in ordinary language—or from sentences in a more formal, symbolic language—into images of overlapping circles, one can use those images to discover further truths. The images are certainly not necessary for such discoveries (it is always possible to reach the same conclusions “algebraically” rather than “geometically”), but neither are the diagrams mere byproducts of knowledge that has already been acquired through other means. The Venn diagram for a given sentence is nothing more than an expression or translation, but the implications of a set of sentences often become more perspicuous (to many people, anyway) through such diagrams. Wittgenstein is right to insist that it is the use of an image, like the use of a word, that gives it its meaning (so it is practices, not images, that establish the meaning of a word, and practices, not words, that establish the meaning of an image). But that does not rule out the possibility of using images to further our understanding of a word, or using words to further our understanding of an image.39 Indeed, there is good reason to think that our dispositions concerning spatial objects are more basic, and more automatic, than our dispositions concerning linguistic words and abstract properties.40 Insofar as the very structure of abstract states of affairs can be represented by Thus, if knowing the meaning of a term is equated with using it correctly, then knowledge is always indeterminate. Note that the problem here is not a problem of how to establish, with absolute certainty, that one knows the meaning of a world; there is nothing surprising about the realization that absolute certainty can never be had. 39 There is a further question about how one knows that the structure or the use of a particular image—a particular Venn diagram, for example—really does match the structure or use of a particular sentence. We can never be certain about such mappings, any more than we can be certain about a particular interlinguistic translation, but that should not preclude us from being confident in our assumptions or from succeeding in our translations (much of the time, anyway). Given enough structural similarity, there is no reason not to be confident in using one as a reliable stand-in for the other. Eventually, translations click into place (through confirmation/re-inforcement experiences—like finding one’s way back through new routes). We can then recognize “how to go on.” 40 Certainly, activities concerning spatiotemporal objects are developmentally (and evolutionarily) prior to activities concerning linguistic utterances. There is also reason to think that early forms of a language are more imagistic and metaphorical than later forms of a language.

4.

RATIONAL PERCEPTION

163

spatial images, those images can provide us with a more effective means by which to extend our knowledge of abstract things. What I hope to have shown in this section, then, is the possibility of using spatial images not merely as reliable indicators of an abstract state of affairs (as in the case of the synesthete) but also as the means by which the objectivity of a state of affairs can be disclosed (as in the case of a graph of rising and falling GNP) and the means by which further knowledge of that state of affairs may be acquired (as in the case of Venn diagrams). It is these latter two possibilities that show how spatial imagining of an abstract state of affairs can provide us with a perceptual form knowledge and, despite Wittgenstein’s apparent objections, can be a source of knowledge as well.

4. Rational Perception—Lessons from Plato and Descartes We are now in a position to offer an interpretation of some puzzling aspects of Plato’s and Descartes’ accounts of rational perception. In Plato’s allegory of the cave, the process of coming to perceive abstract objects (the Forms) is likened to a series of shifts in attention—from shadows on a wall to statues in front of a fire to living things illuminated by the sun; but knowledge of the forms is also held to be very unlike our mere opinions concerning material things. Likewise, in Descartes’ Rules for the Direction of the Mind, the notion of rational perception, or intuitus of the mind, is explicated by reference to intuitus of the eye; but vision is the epitome of unreliability, according to Descartes, while mathematical intuition is the epitome of knowledge. Why do both of these philosophers invoke visual perception in their explanations of rational perception if the differences are so significant? One could read both as making the comparison only in order to highlight the differences and prevent us from confusing one with the other. But the very possibility of confusion invites reflection on their similarities, and the relevant texts are rich with suggestions about what these similarities might be. In what follows, I try to show how both Plato and Descartes develop an understanding of perception that extends to include rational perception. The key to this extension, I suggest, is their recognition of the importance of imagining

164

THE FURTHER REACHES OF PERCEPTION

alternatives. This section, then, uses the theory I have been advocating to help illuminate some key passages in the works of Plato and Descartes; and it uses those passages, so interpreted, to help illuminate the theory I have been advocating. Plato’s allegory of the cave portrays a series of transitions that are needed in order for us to acquire genuine knowledge (as opposed to mere opinion) of what is most real (as opposed to what is merely apparent). One aspect of his portrayal involves a progressive increase in light—from the reflected light of an underground fire to the brilliant light of the sun. He draws on the familiar fact that more light frequently allows us to see things more clearly and more accurately, while sudden increases in light tend to blind us— illustrating both the possibility of (passive) increases in knowledge and the potential for resistance to such increases. Insofar as the sun, as the ultimate source of all light, is the allegorical counterpart of the Good, it points to the centrality of normativity for all knowledge (also, our tendency to resist norms that are imposed too abruptly). I return to this aspect of the story below. The more informative, and more distinctive, part of Plato’s allegory, however, concerns the progression of objects that are encountered in the ascent from the cave, and the ways that these objects are related to one another. The first objects of perception in Plato’s story are shadows; the prisoners sit chained within a cave, watching and discussing the movements of shadows on the wall (which they take to be actual figures). The second objects of perception are stone statues or puppets;41 released from their chains, the prisoners turn around to see statues of animals and people paraded along the top of a wall, casting the previously seen shadows. After emerging from the cave, it becomes possible to see actual, live animals and people.42 And, finally, it becomes possible to see the heavenly bodies and, in 41 The Greek term eikon, meaning likeness (or icon) is paired with the term eikasia, which is literally close to imagination. Eva Braun (1991), p. 37, translates it, in this context, as “image-recognition” or “the power of recognizing images as images.” “It is a noun formed from a denominative verb eikazein, meaning to do something about a likeness, an eikon. Thus eikasia can mean both the likeness itself and a claim that something is likely, a conjecture.” 42 I am skipping over the transitions from shadows of actual animals and people, to reflections of actual animals and people, to the animals and people themselves. The relations between these objects are, as will be seen, exactly like the relations between the objects here highlighted. Plato speaks of the light of the stars and moon as a kind of reflection of sunlight, much as the light on the cave wall is a reflection of firelight. It is not clear (and it doesn’t really matter for the purposes of this tale) whether he believes that moonlight and starlight are in fact reflections of sunlight.

4.

RATIONAL PERCEPTION

165

particular, the sun.43 In each case, Plato portrays the shift from one object to another as a shift from images to origins. It is the moving statues that are the origins of the flickering shadows, it is the actual animals and people that are the models for the stone statues, and it is the cycles of the heavenly bodies and the energy of the sun that are the origin of all living things. The ascent to knowledge is not likened to a passage in which more and more light is shed on the very same object so much as a passage in which our attention shifts from one object to its origin in another, and from that further object to its origin in yet another. The questions of just what the origins of things are and just what sort of perception of them is possible are thus intertwined in the allegory of the cave. While all of the perceived objects in this story are material objects, we know that Plato introduced the story to elucidate the way in which we may also progress from the perception of material objects to the perception of non-material forms; material objects are said to be related to the forms in the same way that shadows are related to the objects that cast the shadows, and in the same way as statues are related to their models.44 But what sort of relations are these, and by what process can perception of the one lead us to perception of the other? How can something material be modeled on something non-material, and how is it possible to shift our attention (which, in the allegory, requires “turning one’s entire body”) from a shadow to its source, or from a statue to its model? If we think of the shift as a shift between radically different domains—the domain of material objects versus the domain of formal objects, say, or as a shift between two different faculties, the senses versus the soul, for example—then it will remain mysterious how one can be the origin of the other and how the perception of one can lead to the perception of the other. And if we doubt the very existence of formal objects or souls, the transition will seem merely mythical. Plato’s description of the relations between these objects and of the means by which the prisoners ascend points to a method for achieving 43 There is a further transition—namely, the descent back into the cave—which is important to the political aims of the Republic ; describes the result of knowledge rather than its acquisition. 44 It is interesting to note the close connection between the words Plato uses for sight, for knowledge, and for form. Eidenai is the perfect active infinitive of the verb for “see” (in the present) or “know” (in the perfect). The idea suggested by the shift from “see” in the present tense to “know” in the perfect tense is that of the completion or perfect fulfillment of seeing that yields knowing—to have come to see something completely, to have taken in from every perspective. (Thanks to Mitchell Miller for these observations.)

166

THE FURTHER REACHES OF PERCEPTION

the requisite transitions, however—a method that emphasizes subtle shifts in attention and imagination rather than radical shifts in the domain of interest or in the mental faculties required. By considering the details of this process more closely, we arrive at an account of what it meant, for Plato, to see forms. Consider, first, how the prisoners’ attention repeatedly shifts from what is more transient to what is more constant. Watching the flickering shadows on the wall of the cave, the prisoners attend to fleeting shapes—shapes that cohere enough to suggest figures passing by, but nothing that persists. When the prisoners are released from their chains and turned away from the shadows to face the statues behind them, they encounter figures that are far more constant than the shadows they cast. Unlike the shadows that flit in and out of view on the cave wall, quickly changing shape and disappearing into darkness, these statues are steady presences, glowing in the firelight, supporting a longer gaze. (Plato describes the challenge of seeing these objects as one of getting accustomed to more light, where he has already made it clear that coming into being is like coming into light, and that passing out of being is like disappearing into darkness.45) Ascending out of the cave, the prisoners are exposed to people and other natural things that are even more persistent than the statues, and that are illuminated by a far brighter and steadier light—the light of the sun rather than the fire in the cave. Again, the greater brightness of the light—which corresponds to the greater constancy of these things—is more than the prisoners can attend to. They need to “acclimatize” themselves to this new light gradually in order, eventually, to see the things that exist outside the cave. Finally, it may become possible to look at the sun itself—the brightest and most constant object of all. The sun, though, does not just happen to be the most persistent and thus most “real” thing of all, according to Plato. The constancy of the sun, as the ultimate source of all light, is necessary for both the visibility and the existence of everything else. Without the continued existence of the sun, nothing could be seen and nothing could exist. “The sun gives to what is seen . . . not only its ability to be seen, but also birth, growth and sustenance.”46 What one sees in seeing the sun, then, is the most fundamental and 45 “[W]hen [the soul] focuses on what is mingled with darkness, on what comes into being and is destroyed, then it resorts to opinion and is dimmed.” (Republic 508d) 46 Republic 509b.

4.

RATIONAL PERCEPTION

167

most necessary thing of all. (I return to a consideration of its generative power below.) The importance of attending to constancies is also evident in the discussion that precedes the allegory of the cave: the Divided Line passage. There, the same four stages are described as a series of transitions from opinions, which “swing first one way and then the other,” focusing on “what comes into being and is destroyed,” to knowledge, which grasps unchanging forms.47 Throughout his writings, and certainly in Books VI and VII of The Republic, the forms are extolled as those things that remain constant amid the flux of changing appearances and opposing opinions.48 This is not an indication of conservative or reactionary inclinations (though it may reinforce such inclinations) so much as it is a recognition that both the intelligibility and the perceptibility of our world depend on the recognition of constancy across change. Variation can only be seen against the background of what is invariant, and differences of opinion are only intelligible insofar as they are disagreements about the same things. (If every new appearance constituted a new thing, and every new opinion concerned a new topic, there would be no difference between appearance and reality, or between opinion and knowledge.) In coming to recognize those things that are most constant, then, we come to know the most fundamental constituents of the world, and we move beyond mere opinion to knowledge.49 The possibility of knowing the forms arises in part, then, out of an everdeepening attentiveness to the constants in our experience. Consider the following progression of attention: I look up from my desk and initially see a play of dark and light, grey and green shadows outside my window. Soon I discern certain constants in this array of colors and shapes, and individual leaves and branches become the objects of my perception. Seeking patterns in these branches and leaves, I then become aware of recurring angles

47

Republic 508d. These four stages are later labeled “conjecture,” “belief,” “thinking,” and “understanding” (511d). 48 For example, in Meno 98a: “Once [true opinions] are tied down, they become knowledge and are stable. That is why knowledge is something more valuable than right opinion. What distinguishes one from the other is the tether.” 49 Compare with Stenning (2002), p. 163: “In fields of abstract concepts like logic, physics, and psychology, learning can only take place through learning what transformations leave which of the concepts invariant.” And Learning logic involves learning invariance across abstract concepts such as validity, consequence, propositions, etc., but it also involves finding invariance across all forms of thinking and reasoning.

168

THE FURTHER REACHES OF PERCEPTION

between the branches and characteristic arrangements for the grouping of leaves; and these angles and arrangements become the new focus of my attention. Taking it a step further, I may eventually come to see the mathematical ratios between different angles and between different groupings—the fact that branches extending upwards veer off at an angle that is half as wide as those that angle downward, the fact that numbers of leaves on each branch are always multiples of three, and so on. Finally, I may even come to see a deeper unity across these various ratios—a reiterated process of division through with one becomes three becomes nine, and so on.50 Thus Plato’s story of perception in and out of the cave does not just provide us with an evocative analogy, indicating the possibility of a shift from the world of ordinary perception to a world of extraordinary insight. It also provides important guidance for the accomplishment of such a transition—indicating how ordinary shifts in attention, from what fluctuates to what remains constant, might be extended to the point where mathematical relations and other organizing principles might also be perceived. He asks us to seek out what is constant across all sorts of variation and change, observing and reflecting on many different instances of trees, toes, tables, or any other type of thing, as well as different stages of individual instances of that thing. While the search for constancy begins with the automatic processing of information from our eyes (and ears, and nose, etc.), the search for constancy eventually (and with training) extends to the activity of mathematical reasoning, whereby we discover relations that remain constant across many different types of things.51 Seeing the forms involves something more than seeing what is constant, however. If the forms were merely abstract types or properties or relations— types that remain the same despite differences in how and when they are instantiated and combined, they would be essential to perception and to judgment but they would not explain why things are the way they are. They would not provide us with reasons for anything. To appreciate the status of forms as reasons, and the perception of forms as the perception of reasons, we need to consider Plato’s understanding of the way in which ordinary objects 50

Eric Havelock (1963), p. 266, claims that “the abstract ‘objects’ do not come gliding into our consciousness suspended on clouds of illumination. Rather, we have to grapple with the many and seek their conversion into ones.” See, also, Miller (1999). 51 In Meno, 85c-12, in order to achieve knowledge, the slave must be asked about “the very same things many times and in many ways.”

4.

RATIONAL PERCEPTION

169

(inside the cave) are derived from the forms (objects outside the cave) in much the way that mere appearances of objects (the shadows in the cave) are derived from ordinary objects (the statues in the cave). What sort of derivation does Plato have in mind, and how can our perceptual abilities be developed to enable us to see past appearances to their origins? Plato repeatedly described the relevant sort of derivation as one of reflecting or imitating. The figures first seen by the prisoners are reflections of the statues that are paraded in front of the fire—a relation that can be observed by the prisoners after they turn around and are able to see how the statues are manipulated in front of the firelight, and how this results in the creation of various shadows on the cave wall. The shadows are reflections of the statues insofar as they result from light deflecting off the statues in particular ways; and the shadows are imitations of the statues insofar as they are created or interpreted in particular ways. It is not easy to say just what these particular ways are, but as long as we can see the shadows and the statues, the handlers and their audience, the fire and its light, it seems plausible to suppose that we can also see that the shadows are the reflections or the imitations of the statues. But what happens when we are only shown the shadows and not the statues, or only the statues and not the living beings on which they are modeled? How is it possible to see past the reflections and imitations to the originals from which they are derived? Although Plato describes each step of the prisoners’ ascent as a change that is forced upon them, clearly he is describing a path that we can choose to pursue, and he is providing us with guidance in that pursuit. In particular, on the assumption that we already recognize the difference between objects and their reflections, or between models and their imitations (from our experience of these things “within the cave”), we are invited to now “treat as images the things which [at the previous stage of the progression] were originals.”52 But how is possible to see the origin of what appears before one when one lacks any independent experience of that origin? Leaving aside Plato’s suggestion that we have seen the originating forms in some past life—a suggestion that may be intended to be more evocative than explanatory of how we know forms,53 several parts of Plato’s story 52

Republic 510b. For an overview of alternate readings, and a suggestion of his own, see Glenn Rawson (2006). Richard J. Ketchum (1979) also develops an interesting alternative. 53

170

THE FURTHER REACHES OF PERCEPTION

indicate that we can learn the origin of an imitation by becoming attuned to the imitation’s imperfections. On the face of it, this may seem exactly the wrong way around: we need to know what something is imitating before we can know how it falls short, for we can experience failure only insofar as we know what we are aiming at. In fact, though, we are frequently aware of things going badly—in a bodily movement, in an interpersonal exchange, or in an artistic endeavor, for example—well before we know what it would be like to get things right. At the very least, it is possible for our knowledge of what counts as failure and our knowledge of what counts as success to emerge together. Imagine the prisoners’ discussion of shadows on the cave wall, for example. Because their predictions about what will appear next will sometimes fail, they have reason to suspect that the information they are given is imperfect—either because it is incomplete or because it is contaminated, much as our view through a rain-soaked windshield is both restricted and distorted. This could lead to deliberate theorizing about the missing information and the contaminating influences, or it could lead to more or less automatic adjustments in their perceptual experience—adjustments whereby they use their imaginations to “fill in” missing information, and whereby they withdraw their attention from distracting information in order to “erase” it from consciousness. When such automatic adjustments are made, then the very contents of perception shift from an imperfectly coordinated series of appearances (shadows on the wall, shapes and colors through the windshield) to the unified, three-dimensional objects from which those appearances derive.54 Thus, even in the case of the shift from shadows to statues, it is not necessary to have previous experience with the original objects in order to see the imperfections of their imitations, and to project beyond those imperfections to the originals themselves.55 How does this work in the case of shifting our attention from statues to their models—the allegorical counterpart of shifting from perception of 54 There is an interesting connection here to a Greek notion of intelligibility as a well-rounded whole. Drew Hyland (1973), p. 187, following Alex Mourelatos (1970), reads Parmenides’ insistence that Being is a “well rounded sphere, equally balanced from its center in every direction” to mean that Being must be “perspectivally neutral, accessible, and intelligible in principle to all men at all times.” 55 Dustin and Schaeffer (2006) offer an insightful discussion of how we can see (literally, not just metaphorically) the justice of a soul or of a state insofar as we recognize its underlying commitment to the unification of otherwise disparate elements. We do not project a unity behind appearances so much as we discern the projection of unity beyond appearances.

4.

RATIONAL PERCEPTION

171

ordinary physical objects to perception of the forms? Again, something that is first viewed as an object in its own right must come to be viewed as an imitation or reflection of some other, more fundamental object. But what does it take to view three-dimensional physical objects as imitations or reflections of something more basic? In the Divided Line passage of Book VI, Plato identifies mathematics as the discipline through which this crucial shift can be achieved. For mathematics (and geometry in particular) teaches us how to treat physically present squares as mere representations of something more fundamental. So their reasoning has in view the square itself, and the diagonal itself, not the diagonal they have drawn . . . The models they construct, or figures they draw, which have their own shadows, and images in water—these they treat in their turn as images, in their attempt to see the corresponding things themselves which can be seen only through thinking.56

The mathematician shifts our attention from physically drawn squares to the very idea of a square not simply by looking for commonalities or constants across various particular squares; as above, some features that are shared by all representations of a thing (e.g. slight waviness in the sides of drawn squares) are not features of the thing represented, and some features of what is represented (lines whose ends meet) may not appear in some of its representations. What is important to the mathematician is figuring out how a variety of drawn images could all be derived from a single origin. Some of this figuring out might be done by trial and error, exploring various ways in which many different drawings might all be derived from a single idea—the idea of a figure with equal sides and angles, the idea of a figure that can be replicated through double folding or unfolding, and so on. At some point (either before or after such figuring), however, we acquire more immediate, perceptual knowledge of the square which serves as the model for the assorted drawings; we see it in and through the drawings in the same way as we see a person or an animal in and through the statues that may be modeled on it—directly seeing (rather than consciously calculating) one as the origin of the other. In the case of mathematics, of course, the original objects are not themselves physical objects, but both the phenomenology (of perception) and the methodology (of looking for an unitary original 56

Republic 510e.

172

THE FURTHER REACHES OF PERCEPTION

from which multiple instances may be derived) remains the same. Mathematics enables us to extend this method beyond the realm of physical objects, thereby granting us a new sort of perceptual experience—at which point we will have crossed the divided line and climbed out of the cave. The final step in Plato’s progression—from perception of the forms (the objects outside the cave) to perception of the Good (the sun)—is again achieved by seeking the origin of what one perceives. But now Plato shifts from the methods of mathematics to the method of dialectic—a method through which differing opinions can be viewed as imperfect expressions of a common insight—an insight that is reflected in divergent opinions, an insight upon which different claims are modeled.57 By attending to differences in what is supposed, we are able to discover something shared—not just something that remains constant across differences but, more importantly, something at which each of the divergent views can be seen to aim— i.e. the truth on which the differing opinions are modeled.58 How can we know what something is aiming at without independent experience of that aim? Again, we might rather deliberately formulate certain aims as offering the best explanation of the evidence at hand or we might come to see how things fall in place when a particular aim is simply assumed. Or, and this is the most relevant possibility, we might move from a series of deliberations about better versus worse explanations to all-at-once experiences of overall fit. The study of mathematics and dialectic, according to Plato, is valuable precisely because it helps move us from the stage of deliberate theorizing to the stage of perceptual insight—from the task of figuring out the common source of multiple phenomena (the inspiration behind many different triangles, or many different opinions) to the experience of seeing those phenomena as manifestations of that source. We move from the stage of formulating a hypothesis about why certain things are as they are to the stage of seeing through to their common aim.

57 Hideya Yamakawa (2008), p. 185, suggests that mathematics and dialectics “have in common a methodological property which may be called ‘the rule of the best explanation’.” The starting points are hypotheses, in the case of the dialectician, and visible images, in the case of the mathematician. (p. 179). 58 On the importance of attending to what seems problematic, Republic, 524d: “this was what I was just trying to convey in saying that some things are apt to summon thought while others are not, defining as apt to summon it those that strike the sense at the same time as their opposites, while all those do not, are not apt to arouse intellection.”

4.

RATIONAL PERCEPTION

173

The above reflections also help make sense of Plato’s analogy between the sun and the Good. The Good, for Plato, is the singular whole that is responsible for multiple representations, renditions, and opinions—responsible for the intelligibility of representations, renditions, and opinions as such, and responsible for their very existence. An underlying singularity is necessary for the intelligibility of representations, renditions, and opinions as such, since nothing counts as a representation unless there can be multiple representations of a single object, nothing counts as a rendition unless there can be different renditions of the same fact, and nothing counts as an opinion unless there can be different opinions about a single truth. An underlying singularity is also necessary for the very existence of representations, renditions, and opinions insofar as they must be causally responsive to something singular; the construction of representations, renditions, or opinions must be guided by a singular “star.” These two types of dependence, intellectual and causal, show singularity to be normative—hence, the Good—insofar as singularity is the measure by which diverse thoughts and experiences are evaluated (as more or less true, as closer or farther from knowledge), and singularity is the aim that we ought to pursue. At each step in our ascent from the cave, then, we are led to see a singular reason for some diverse set of phenomena, a reason that is both explanatory and normative. It is only upon arrival at the absolutely singular, however, that we arrive at the singularly Good. Let us turn next to the account of rational perception that is given by Descartes in his Rules for the Direction of the Mind. For all of the evident differences between Plato’s account and Descartes’ account, I hope to bring out some deep similarities between their accounts (and between their accounts and mine)—similarities that revolve around the central role of the imagination in revealing the objectivity of abstract states of affairs. Seeing by “the light of pure reason” is a notion that is central to Descartes’ methodology, both in philosophy and mathematics. Accordingly, there has been much discussion of what makes an idea “clear and distinct,” and thus a case of knowledge, according to Descartes. In his Meditations, he discusses the limitations of the imagination with respect to knowledge—as evident by our inability to imagine a thousand-sided chiliagon, for example, and our inability to imagine the infinite number of

174

THE FURTHER REACHES OF PERCEPTION

changes that a piece of wax could undergo.59 I make some observations concerning these passages below. More relevant to our purposes, though, is his understanding of the way in which imagination can contribute to our knowledge—specifically with his account of knowledge that grasps a truth in a single act of intuition, transforming the multiple steps of a deductive or enumerative proof into a single act of intuition which, for a variety of reasons, is a preferable form of knowledge according to Descartes.60 For it is in his consideration of this transformation (which is closely related to the transformation from algebra to geometry) that Descartes shows how complex knowledge can become perceptual through the presentation of multiple relations in a single space. It is in this context also that he makes several important observations about the advantage of intuitive knowledge over deductive knowledge. Intuition, according to Descartes, is the non-sequential and unmediated apprehension of something. Deduction, in contrast, involves a sequence of thoughts whose presentation is mediated by memory: [W]e distinguish this mental intuition from deduction by the fact that into the conception of the latter there enters a certain movement or succession, into that of the former there does not. Further, deduction does not require an immediately presented evidence such as intuition possesses; its certitude is rather conferred upon it in some way by memory.61

These two aspects of intuition—the fact that the contents of an intuition appear all at once rather than sequentially across time, and the fact that our certitude about those contents does not depend on any certitude about our 59

From Meditations VI: “[I]f I desire to think of a chiliagon, I certainly conceive truly that it is a figure composed of a thousand sides, just as easily as I conceive of a triangle that it is a figure of three sides only; but I cannot in any way imagine the thousand sides of a chiliagon [as I do the three sides of a triangle].” And from Meditations II: “Is it not that I imagine that the piece of wax, being round, is capable of becoming square, or of passing from a square into a triangular figure? Assuredly such is not the case, because I conceive that it admits of an infinity of similar changes; and I am, moreover, unable to compass this infinity by imagination, and consequently this conception which I have of the wax is not the product of the faculty of imagination . . . I must, therefore, admit that I cannot even comprehend by imagination what the piece of wax is, and that it is the mind alone (mens, Lat., entendement, F.) which perceives it . . . the perception of it is neither an act of sight, of touch, nor of imagination, and never was either of these, though it might formerly seem so, but is simply an intuition (inspectio) of the mind, which may be imperfect and confused, as it formerly was, or very clear and distinct, as it is at present, according as the attention is more or less directed to the elements which it contains, and of which it is composed.” 60 This contrast is touched upon in Descartes’ Meditations, but explored much more fully in his Rules for the Direction of the Mind. 61 Descartes (1681/1952), from Rule III.

4.

RATIONAL PERCEPTION

175

memory—are, of course, related. For if the certitude of a conclusion depends on the certitude of certain premises and those premises are no longer present to mind, then we must rely on memory to assure us that the premises were indeed known with certainty.62 (As we shall see, though, the fact that intuition cannot consist of a sequence of thoughts or representations does not mean that we cannot have an intuition of a sequence.) In Rule IX of his Rules for the Direction of the Mind, Descartes likens the intuitus63 of the mind to the intuitus of the eyes, noting how a single act of apprehension, or attention, can be directed to a simple object or to many objects simultaneously. He recommends that we increase the perspicacity of our intuition by attending to just one object at a time. [I]n what way one is to use intuitus of the mind we know by comparison with the same [intuitus] of the eyes. For whoever wants to look upon many objects simultaneously in the same intuitus will see nothing of them distinctly; and equally whoever is used to attending to many things simultaneously in a single act of cognition is of confused ingenium.64

Descartes’ advice assumes that acts of the mind, and thus intuitions, are individuated by reference to changes in attention across time rather than by reference to the number of objects within one’s attention at any given time, or the clarity and distinctness of those objects. The fewer and simpler the objects of attention, the greater the likelihood of clarity and distinctness, but even simple objects can be confusing to some. (“It is true that some men are born with a much greater aptitude for such discernment than others, but the mind can be made much more expert at such work by art and exercise.”65) With practice, though, it is possible to acquire intuitive knowledge of quite complex matters. (“Now we need to see intuitively not only that 2 and 2 make 4, and that likewise 3 and 1 make 4, but further that the third of the 62 “[W]e know that the last link in a long chain is connected with the first, even though we do not take in by means of one and the same act of vision all the intermediate links on which that connection depends, but only remember that we have taken them successively under review and that each single one is united to its neighbor, from the first even to the last.” (Rule III) 63 See Sepper’s (1996) discussion of the term “intuitus,” usually translated as “seeing.” “Intuitus is the past participle of the infinitive intueir, ‘to look at, upon, or toward,’ in a more extended sense ‘to regard, observe, contemplate, consider, attend to;’ intuitus can be used as a noun meaning ‘ “a look, view.’ ” (124) See also his discussion on p. 91. 64 Descartes (1681/1952), Rule IX. Sepper (1996), pp. 91–7, discusses the term “ingenium,” which is part of Descartes’ title. 65 Rule IV.

176

THE FURTHER REACHES OF PERCEPTION

above statements is a necessary conclusion from these two.”66) What we intuit, then, is whatever we are capable of attending to all at once, in a single moment of consciousness, where a single “moment” of consciousness is equated with a single act of apprehension rather than a single object or a specific degree of certainty. Because the object of an intuition, so understood,67 is not always perspicuous (not always clear and distinct), and because it is usually possible to apprehend the parts of something more perspicuously (more clearly and distinctly) than the whole, Descartes repeatedly recommends that we break down each problem into the smallest possible units, considered in sequence, in order to ensure maximal certainty with regard to each step in our reasoning. On the other hand, at whatever point clarity and distinctness can be achieved, Descartes urges us to reverse course, as it were, and transform sequential acts of apprehension back into a single act of apprehension in which all parts of the whole may be viewed simultaneously. For such a compression lessens our reliance on memory, which is never certain.68 [D]eduction frequently involves such a long series of transitions from ground to consequent that when we come to the conclusion we have difficulty in recalling the whole of the route by which we have arrived at it.69 [T]he memory, on which we have said depends the certainty of the conclusions which embrace more than we can grasp in a single act of intuition, [is] weak and liable to fail us.70

Although we may in fact have reasons for the conclusion we reach through a series of deductive steps, we do not experience the conclusion as having reasons if those reasons are not simultaneously present to mind; and if we do not experience a conviction as having reasons, we do not experience it as knowledge.71 66

Rule III. Note that the Rule III definition of intuition emphasizes certainty and distinctness, and Rule XI specifies that “the proposition intuited must be clear and distinct.” But Rule IX is the rule that is specifically focused on the training of intuition, which would not be necessary if intuition were already guaranteed to produce knowledge, and Rule XI focuses on the non-sequential character of intuition rather than its clarity and distinctness. 68 This reason becomes even more prominent in Descartes’ later writing, especially his Meditations, which is more concerned with achieving certainty. 69 Rule VII. 70 Rule XI. 71 Whether we need to experience a state as a state of knowledge in order for it to be a state of knowledge is open to dispute—with “externalists” generally rejecting this requirement on knowledge. Descartes, clearly, was an “internalist.” 67

4.

RATIONAL PERCEPTION

177

How is such compression achieved? Descartes recommends that we repeatedly run through the whole series of steps in a deduction, keeping the imagination moving continuously in such a way that while it is intuitively perceiving each fact it simultaneously passes on to the next; and this I would do until I have learned to pass from the first to the last so quickly, that no stage in the process was left to the care of the memory, but I seemed to have the whole in intuition before me at the same time. This method will . . . relieve the memory.72

Here Descartes is allowing for the intuition of a sequence of argumentative steps that does not, however, depend on a sequence of intuitions. In the case of visual intuitions, we may see a shooting star move from left to right without experiencing a sequence of intuitions whereby we first see the star to the left, then see it in the center, and finally see it to the right. The light may strike our retina in that sequence, but the intuition of movement happens all at once.73 Similarly, in the case of mathematical intuition, we may see that A is greater than B which is greater than C which is greater than D in a single act of consciousness rather than a series of separate acts. We can encounter a series of objects in sequence, and experience them as a sequence, while also apprehending them simultaneously, in a single experience. So even if such intuitions, whether visual or intellectual, are fallible, their failures become failures of intuition rather than failures of memory. In the case of induction or enumeration (versus deduction or derivation), the process of compressing a series of intuitions or perceptions into a single intuition or perception works a bit differently. For in the case of induction, the facts that support a conclusion are always incomplete—either because the relevant set is infinitely large and therefore unavailable in principle (as is true of the infinite number of possible changes in a piece of wax), or because as a matter of fact rather than principle we are not able to attend to the entire set of relevant facts (as is true of the thousand-sided chiliagon). Nonetheless, according to Descartes, an inductive proof also can be rendered intuitive insofar as we can find appropriate “exemplars” or “abbreviations” to stand 72

Also Rule XI. Descartes wavers between the claim that this process will eliminate our reliance on memory and that it will strengthen our memory. 73 Descartes maintains that not only movement but the causes of movement can be intuited— specifically, in the case where pushing one end of a stick causes the other end to move, and where the weight on a balance raises one arm at the same time as it depresses the other.

178

THE FURTHER REACHES OF PERCEPTION

in for the larger set. We can’t contemplate every possible triangle, for example, but we can consider an “exemplar” triangle from which we may draw some conclusions regarding all triangles. And we cannot consider every animal in all of its peculiarities but we can abstract from a host of individual peculiarities to arrive at an “abbreviated” representation of an animal from which we can draw some conclusions regarding all animals. How do we arrive at such “exemplars” or “abbreviations,” and how do we determine which of their properties are attributable to all members of the exemplified or abbreviated class? Here Descartes suggests a strategy that echoes Plato’s strategy for discerning forms: Run through a number of instances in one’s imagination to determine what remains constant across those instances, and then focus one’s attention on precisely those things that remain constant across all imagined instances. A particular triangle thus becomes an exemplar insofar as one uses it to attend to its essential (versus non-essential) features; a representation of an animal is abbreviated insofar as it only possesses essential properties of animalhood. Both of the above ways of compressing sequential deliberations into intuitive insights—through temporal compression in the case of deduction, and through exemplification in the case of induction—enable us to circumvent certain limitations on the amount of information at our disposal: limitations due to failures of memory, in the one case, and limitations due to the finite character of our lives, in the other. Thus, while Descartes’ observations about the limits of the imagination (particularly in the context of his discussion of a piece of wax in Meditations II and his discussion of a chiliagon in Meditations VI) remain valid, he also advocates a particular strategy for overcoming these limitations—a strategy that actually requires us to make more rather than less use of the imagination.74 For it is precisely through imagining many different changes in the wax that we can arrive at 74 Dennis L. Sepper (1996), pp. 207–8, notes that Descartes treats the imagination as more active in his Rules for the Regulation of the Mind than in his Meditations: “[T]he model for knowing that the Regulae presents is a thinking through images: the distinct and easy grasp of the aspects they present, and the knowing power wending its way among images—noting, comparing, forming, and transforming them—in order to arrive at the truth of any question. “In the accounts of cogitation and imagination given in the Meditations, there is a difference of no little significance. Although imagination is mentioned as part of thinking in the Second Meditation, in the Sixth it is excluded from the essence of the thinking thing; and in the Third Meditation thinking is apparently transformed from the active grasping, comparing, and moving of thoughts into the more passive perception of ideas that are in oneself or one’s mind innately. One perceives ideas, one has them. Imagination itself thereby becomes just another way of having ideas. . . . Rule 12 had defined imagination as an activity,

4.

RATIONAL PERCEPTION

179

the knowledge of what is essential about the wax—that is, its extension— and extensions can only be contemplated and studied and understood insofar as they can be given a spatial exemplification. Furthermore, despite the fact that no amount of practice is likely to improve our memory to the point where every one of a thousand sides can be contemplated clearly and distinctly together at a given moment, Descartes recommends that we engage in just such practice when it comes to eight- or ten- or twelve-sided figures. There is another epistemic advantage to the compressions that Descartes describes—an advantage that is even more important to him (at least in the Rules)—namely, the capacity to engender new discoveries. The example he gives, in Rule XI, concerns a series of magnitudes—A, B, C, D, E—in continued proportion such that the proportion of A to B is the same as B to C, and so on. The ease with which one can generate a further addition to the series in which the proportion of F to E is the same as E to D is the same whether one considers this simultaneously or in sequence; but if we are asked to determine the mean between A and E, it is immensely easier if we can consider the whole sequence at once, for then it is immediately evident that C must be the mean between A and E. Most of us will recognize the truth of this comparison, but what accounts for the greater ease of discovery in this case? And is it always so? In order to intuit a sequence of items or events all at once, we must experience them as differentiated in space; for, by stipulation, they cannot be differentiated in time. (We can simultaneously represent different events on a time line, of course, but the line is itself spatial, not temporal.) Once we imagine various items within a single space, however, we automatically imagine a variety of previously unspecified relations between those objects, their parts, and the spaces between. So, insofar as mathematical truths can be presented spatially—a discovery that is itself largely due to Descartes, who demonstrated the inter-translatability of algebra and geometry—they will be presented in a manner that is bound to display more relationships than specified in the originating description. And, more broadly, insofar as any set of relationships between things can be presented spatially—relations of inclusion, exclusion, overlap, and displacement, for example—then their as the application of the vis cognoscens to the phantasia in order to form new figures. Where has the activity of thinking and imagining gone?”

180

THE FURTHER REACHES OF PERCEPTION

simultaneous presentation will also reveal a variety of further and unforeseen implications.75 If apprehending a sequence of things simultaneously amounts to imagining them in a spatial array of some sort, then a final advantage of intuition over deduction—one that is introduced less directly by Descartes—is that it establishes a more direct connection between thought and action. For insofar as a thought can be translated into the perception of a certain spatial array, our actions can respond more or less automatically, without the intermediary of deliberation. Because of the isomorphism that exists between the space of appearances and the space of movement, spatially represented thoughts provide more direct guidance for our actions than do non-spatial representations. Descartes has a very literal understanding of these translations. First, he argues that every difference that is perceived by the senses can be directly translated into geometrical figures perceived by the mind.76 Descartes insists that the impact of the external world on our senses is literally like the impact of a seal on wax. “[I]t should not be thought that all we mean to assent to is an analogy between the two.”77 In the case of touch, it is easy to see how the shape of the perceived body, or its hardness, or its roughness, is registered through the various pressures it exerts on the skin. But Descartes thinks the same is true of all the other senses. In the case of vision, for example, it is the pressure exerted by light on the eye that creates the image of an external object; Descartes even supposes that our discrimination of color is due to differences in the patterns of pressure exerted on our eyes.78 “The same argument applies to all cases; for it is certain that the infinitude of 75 Some spatial representations will be more revealing than others, of course; and some relations that appear in spatial renditions of things may not reflect any corresponding relations in what is being represented. Knowing how to “read” a representation includes knowing what is relevant and what is not. But, then, spatial renditions can be very powerful—as with Venn diagrams. See Stenning (2002). 76 Rule XII. Descartes invokes the analogy of a blind man feeling his way with a stick to establish the possibility of instantaneous cause and veridical transmission, between external object and internal registration/representation (an example from Simplicius’s commentary on Aristotle’s De Anima.). Schultz (2009), p. 44, clarifies the differences between three different faculties as discussed in these sections. The sensus communis collects and combines the information received from the external senses and stores them as images in the physical part of the brain called phantasia (Greek) or imaginatio (Latin). Productive imagining is done by the vis cognoscens or ingenium, which produces new images. 77 Also Rule XII. 78 This is not as implausible as it initially sounds. Descartes doesn’t need to defend the particular color grids he describes in order to argue that differences in color are registered (in the brain) as different spatial arrays (different interference patterns mapping onto different neural firing patterns, for example).

4.

RATIONAL PERCEPTION

181

figures suffices to express all the differences in sensible things.” Descartes then goes on to describe the transfer from patterns of pressure on the skin to mental images as being like that between the pattern of pressure on one end of a pen and the figure traced at the other. The shape and size and direction of movement at each end may be quite different, but there will always be an isomorphism that ensures that each change in pressure or movement at one end results in a corresponding change at the other end. Such a mechanism ensures a smooth transfer of information in either direction. Thus, it is not only the pattern of the world that can be inscribed, via the bodily “pen,” directly onto the mind; it is also the patterns of the mind that can issue, also via the bodily “pen,” directly in action. [W]e must conceive that the motor force on the nerves themselves derive their origin from the brain, in which the imagination is located, and that the imagination moves them in various ways, just as the external senses act on the common sense, or the lower extremity of the pen moves the whole pen. This example also shows how the imagination can be the cause of many motions in the nerves, motions of which, however, it does not have the images stamped upon it, possessing only certain other images from which the latter flows.79

The third advantage of seeing reasons over recounting reasons, then, is the advantage of presenting information in a form—that of a spatial array—that allows for maximally direct (and instantaneous) transmission of information between the inner world of thought and the outer world of sensation and action. Much has been written about the relative advantages of spatial versus linguistic thinking, images versus words, synthetic versus analytic problem solving, modeling versus proving, and right brain versus left brain processing. These contrasts do not entirely coincide with each other and they are not always drawn in the same way, but there is a recognizable overlap in their domains. For a time it was fashionable to promote spatial thinking, synthetic problem solving, modeling, and right brain processing as a needed counterbalance to what was supposed to be the domination (at least among educated westerners) of words, analysis, proofs, and the left brain. More recently, the emphasis has shifted towards ways in which different people learn in different ways, and ways in which the opposed poles of each 79

Quoted on pp. 19–20 in Sepper (1996).

182

THE FURTHER REACHES OF PERCEPTION

contrast complement one another—the way that spelling out one’s premises can keep us from being misled by appearances, the way that good writing depends on both right and left brain, and so on.80 The research and reasoning that supports these claims becomes relevant to our topic insofar as actively imagining a multiplicity of perspectives or a multiplicity of possibilities makes our thinking more spatial and more synthetic—more akin to seeing and less akin to calculating. Below I summarize, in a somewhat schematic way, the apparent advantages and disadvantages of imagining rational relations as spatial relations, with particular attention paid to the learning and use of logic and math. Analogous trade-offs, in the case of ethics and aesthetics, will be pursued in Chapter 5 and Chapter 6. One advantage of experiencing a number of things as jointly present in a unified space—i.e. constructing visual models of one’s premises rather than listing them—is that consistencies and inconsistencies become immediately apparent. Consider the following three premises, for example: 1. All birds have wings. 2. Flying is not possible without wings. 3. Not all birds can fly. Are these three claims consistent? Most people are not immediately sure, even if (perhaps especially if ) they are formalized to read: 1. B -> W 2. F -> W 3. ~(B -> F) But if you ask just about anyone whether it is possible for there to be two different circles (B and F) that both fit inside a third (W) without entirely converging, the answer is an immediate “yes.” And if you ask whether one circle can be contained inside two other circles that don’t overlap each other, then the answer is an immediate “no.” (Again, the corresponding inconsistency of A ->B, A ->C, and B -> ~C is not immediately evident to most people).81

80 Some of this research is now being questioned, though, in favor of cultivating multiple styles and multiple epistemic pathways in everyone. See New York Times, “Forget What You Know about Good Study Habits,” 9 June 2010. 81 Note that validity = inconsistency of denial.

4.

RATIONAL PERCEPTION

183

Just how to translate sentences into diagrams is not always clear, of course, and making translations (in either direction—from sentences to diagrams or from diagrams to sentences) may also be time-consuming and prone to its own confusions.82 So, while it is usually easier to determine what can be consistently pictured than what can be consistently said, there is no guarantee that in any particular case checking for consistency or validity with diagrams is more reliable or more efficient than checking for consistency or validity with sentences. It depends on how adept one is at making the relevant translations. Nonetheless, the advantage of having such translation abilities should be clear. A second, and probably more important, advantage of spatialized reasoning concerns its capacity to generate new insights as opposed to validating (or invalidating) the old, to discover new implications and relations rather than checking the consistency of implications or relations that have already been noted. Consider the following three premises: A is half the amount of B. C is half the amount of A. D is equivalent to B minus C. When the specified amounts are represented as points on a line, as follows, O---C---A---D---B

it is immediately clear that the difference between C and D is the same as the distance between A and B, that D is halfway between A and B, and so on. These facts could be deduced through a series of calculations aimed at comparing the differences between each pair of quantities, but the diagram makes these relations instantly evident—supplying us with knowledge that we may not even be seeking. Rendering the premises in space leads to discoveries that we would not otherwise make.83 82 From Meditation VI: “[I]t may happen that in imagining a chiliagon I confusedly represent to myself some figure, yet it is very evident that this figure is not a chiliagon, since it in no way differs from that which I represent to myself when I think of a myriagon or any other many-sided figure; nor does it serve my purpose in discovering the properties which go to form the distinction between a chiliagon and other polygons. But if the question turns upon a pentagon . . . I can also imagine it by applying the attention of my mind to each of its five sides, and at the same time to the space which they enclose.” 83 Most studies that have evaluated the use of diagrams in logic have focused on ease and reliability of learning, and ease and reliability of checking the validity of proofs. Less has been done to evaluate the contribution of diagrams to the discovery of new proofs. There is evidence for the importance of spatial

184

THE FURTHER REACHES OF PERCEPTION

The main reason for our greater ease in reasoning with diagrams (for cases like those above, anyway) is the non-arbitrary way in which diagrams are related to their referents—in contrast to the merely arbitrary way in which words, or symbols, are related to their referents. There is a certain isomorphism between relative quantities and relative positions on a line, with bigger differences in quantity represented by bigger distances on the line— unlike numerical representations that lack any such isomorphism; and there is an isomorphism between overlapping circles and the overlapping spaces occupied by all flying things on the one hand, all birds on the other. With diagrams, it is also possible to represent negation isomorphically by crossing out certain areas—indicating areas in which nothing appears—or by preventing certain circles from overlapping at all.84 Isomorphisms such as these allow us to see “through” a diagram to its referent more easily, and more directly, than we are able to do with sentences and symbols.85 For both Plato and Descartes, the rational realm is a realm of abstract versus spatiotemporal objects, and it is accessed directly by the soul rather than through the senses. In Chapter 1, I argued against a sensory requirement for perception; what was required was a phenomenology of selfevident objectivity—a phenomenology that was explicated in terms of simultaneously imagined alternatives. In this chapter, I have tried to show how such a phenomenology is possible with respect to abstract objects. In Section 3, above, I argued that while abstract objects may lack spatiotemporal properties, their objectivity can be experienced through the help of spatiotemporal imagining; and although the “space” of abstract objects is

representation for discovery, however, in the testimony of both mathematicians and logicians who speak of seeing a new truth and who, if pressed, will often describe what they see with spatial imagery. The translation into a proof in accordance with established rules of inference comes later. 84

The difference between these two representations of negation is normally aligned with the difference between a possibility for which there are no occurrences and an impossibility; but the logistics of complicated Venn diagrams can sometimes use crossed out intersections to represent impossibilities as well. Keith Stenning (2002), p. 134, documents hesitancy about translation of “no” or “some” as indicative of a graphic approach. “Although we know a good deal about the processing of negations in language, we have less information about how they are processed when attached to diagrams. 85 See details of four constraints on (and possibilities for) spatial representation in Gattis (2001), pp. 244–5. Both math and reasoning are about relations, and space allows for transparent ordering of elements within a dimension, directionality, and inclusion/exclusion/overlap of different dimensions. “[T]he structure provided by a spatial schema, combined with partial knowledge of a set of elements and relations between them, allows us to infer the elements or relations that are unknown. Spatial schemas do so by marking three aspects of structure that play a significant role in logical reasoning; order within a dimension, directionality within a dimension, and relations between dimensions” (p. 5).

5.

CONCLUSION

185

merely imagined, such imagining can result in genuine perception. This final section has tried to show how Plato’s and Descartes’ understanding of rational perception both support the view I have been defending. While an affinity with historical positions cannot vindicate my account, it does help to establish its importance within a long tradition of thinking about perception, insofar as perception gives us foundational knowledge of an objective world.

5. Conclusion Relying on the account of perception developed in previous chapters—an account that emphasizes the centrality of actively imagining alternatives— this chapter has explained how it is sometimes possible to perceive states of affairs that are spatiotemporally distant and how it is possible to perceive abstract states of affairs. The discussion of perceiving spatiotemporally distant states of affairs led to a clarification of the requisite role of causality, and the discussion of perceiving abstract states of affairs led to a clarification of the requisite role of spatiality. We were then able to offer a sympathetic interpretation of “rational perception” according to Plato and according to Descartes. Overall, then, this chapter has shown how an account of perception that began by addressing ordinary sense perception carries over, quite naturally, to the sorts of perception that matter most to philosophers in the rationalist tradition.

Figure 5.1. “Sled Ride Doorways” (photo by Daniel Friedman)

Chapter 5 Moral Perception The claim that moral knowledge is (or at least can be) perceptual knowledge prompts much sympathy and much suspicion. It is certainly the case that moral judgments can be as confident and as immediate as judgments about a bird in a tree—requiring no deliberation and incurring no doubt. Furthermore, insofar as our moral judgments are about particular people doing particular acts in particular contexts, they seem to have the demonstrative character of perceptions: “that was the right thing to do” as opposed to “it is right to help others in need.” On the other hand, acts that are right don’t have a distinctive appearance in the way that objects that are round or sounds that are shrill have distinctive appearances. It is much harder to agree on an objective measure of the accuracy (and thus the reliability) of our moral judgments than it is to agree on an objective measure of the accuracy of a perceptual judgment about the location of a bird or the color of some grass. So even if we feel confident about the wrongness of a friend’s lie, for example, we are frequently at a loss when it comes to defending that judgment against the equally confident judgment of someone who thinks that the friend was right to lie. Where is the independent state of affairs (like the bird in the tree) to which we can appeal? What are the optimal abilities (like color sensitivity versus color blindness) or the optimal conditions (like sunlight without a glare) that show one person’s judgment to be better than another’s? Competing claims cannot be discredited without a satisfactory way of distinguishing favorable from unfavorable “viewing conditions”—something that itself depends on having some account of just what sort of facts moral facts are. The tendency to equate favorable viewing conditions with having a good upbringing (and to equate moral qualities with “non-natural” or “secondary” qualities) only adds to the suspicion that advocates of moral perception are

188

MORAL PERCEPTION

precisely those who cling to the notion that their moral confidence needs no explanation.1 In its strongest form, as defended by Iris Murdoch for example, the perceptualist’s claim is that moral knowledge is always perceptual knowledge; that even if the right moral conclusion can be arrived at through non-perceptual means, our acceptance of that conclusion will not constitute moral knowledge because it will not have been acquired in the right way, or is not responsive to the right things. In a somewhat weaker form, the claim is that moral knowledge is always based on moral perceptions, but the relevant perceptions need not be ours, and we may use inferences from such perceptions to acquire further moral knowledge. Weaker still is the view that moral perceptions, and inferences based on those perceptions, provide us with one route, but not the only route, to moral knowledge. This chapter is concerned with the nature and the possibility of moral perception, not the exclusivity of its claims to knowledge or the possibilities for acquiring moral knowledge by other means. If perception were the only way to acquire knowledge of moral facts, there would be an important disanalogy between moral knowledge and knowledge of physical facts (where knowledge can be acquired, derivatively, from experts, and where important facts can be inferred rather than observed); and although many may welcome the distinction between moral knowledge and physical knowledge, dis-analogies between the perception of physical facts and the perception of moral facts are bound to raise doubts about whether so-called perception of moral acts is actually, literally, perception. In any case, it is only the weakest thesis about moral perception that I am committed to here—that is, the claim that we do sometimes perceive moral facts. A defense of moral perception should explain how it is possible for moral knowledge to have the epistemic character and the phenomenological character of a perception. Many accounts address the epistemic similarities between moral and non-moral knowledge but say little or nothing about their phenomenal similarities (or differences). It will come as no surprise 1 I am thinking of G. E. Moore (1912) and John McDowell, for example. The following passage from McDowell (1994) is illustrative: “Aristotle scarcely even considers that doubts might arise about the specific ethical outlook he takes for granted. On the reading I am attacking, this shows confidence that he could validate the demands of that ethical outlook by appealing to nature. But I think it shows his immunity to our metaphysical anxieties; he is simply not vulnerable to the sort of worries such a conception would address.” (p. 80)

1.

STANDARD COMMITMENTS OF PERCEPTUAL THEORIES

189

that my account of the phenomenology of moral perception will focus on the experience of objectivity and the imagining that goes into creating such an experience in the moral realm. Showing just what needs to be imagined, and just what is thereby disclosed, is the central task of this chapter. In Section 1, I identify three different theses that are usually endorsed by advocates of moral perception. I note how some writers are more focused on some commitments than on others, and how some commitments are more intertwined than others. This section simply tries to situate my arguments within a larger tradition. Section 2 offers an account of what it takes to perceive a person as a person—as an actor in the world and as a self-interpreter and self-creator. Perceiving people as people doesn’t ensure moral action and yet, I argue, it is a moral as well as an epistemic accomplishment. Among other things, perceiving people as people requires us to recognize some important limitations on our knowledge of others—not only limits on the amount of evidence we can access but also, more fundamentally, limits on what the evidence can tell us. The fact that people are self-interpreters and self-creators makes any information that we gather not only incomplete but indeterminate. The paradoxical nature of the resultant perception is explored in Section 3 through a consideration of uncanny experiences, as theorized by Freud and Cavell among others. Section 4 then turns to the possibility of perceiving what ought (or ought not) to be done. I argue that perceiving what one ought to do only makes sense insofar as we are capable of imagining relevant counterfactuals and recognizing important invariants—a process that was introduced in Chapter 3 and is developed further here. Section 5 relates my account of moral perception to my reading of Kant’s ethics (which is one of the prime targets of perceptualist criticism). If I am right, Kantian ethics not only allows for moral perception, it requires it; and moral perception is not only compatible with moral laws, it relies on them.

1. Standard Commitments of Perceptual Theories There are many different advocates of moral perception, with interestingly different things to say.2 I want to begin by focusing on three theses that seem 2

A short list: Moore (1912), Murdoch (1970, 1993), Nussbaum (1990), Blum (1994), Audi (2004), McDowell (1994), Goldie (2007a), Jacobson (2005), DeLapp (2007), Pleasants (2008).

190

MORAL PERCEPTION

to be accepted by most believers in moral perception, noting some of the different ways that these theses can be interpreted and clarifying how they are related to one another. Thesis I: Moral perception is immediate in the same way that our perception of tables and chairs is immediate. Thesis II: The moral facts that we perceive are objective in the sense that they obtain whether or not we recognize their existence. Thesis III: The moral facts that we perceive are facts about particular situations or states of affairs.

Thesis I identifies a way in which (at least some) moral knowledge is like nonmoral perceptual knowledge—namely, in being immediate rather than mediated, direct rather than indirect. As we saw in Section 2 of Chapter 1, there are different ways to understand the relevant sort of immediacy, however. Beliefs that we acquire without engaging in any conscious inference may still depend on other beliefs for their justification (they may be in need of justificatory intermediaries); while beliefs that we arrive at through careful deliberation (involving many conscious intermediaries) may still be self-standing (requiring no justificatory intermediaries). Robert Audi elaborates on this distinction to argue that moral perceptions, like visual perceptions, require justificatory immediacy rather than conscious immediacy: Much as a mere glance at a tree may not yield justified visual belief or indeed any belief, reflection on a proposition or concept, not a mere thought of it, may be needed to see even what, once it is clearly seen, is intuitive.3

Equally, he insists, beliefs that are not arrived at through any inference may nonetheless fail to be perceptual if their justificatory support depends on other beliefs: The kinds of phenomenal elements essential in drawing an inference are not required for a belief to acquire a supporting role: the “inferentiality” in question is not episodic or even phenomenal; it is a structural relation between beliefs. It implies nothing about what occurs in consciousness. We often make connections between propositions by drawing inferences; but (by the grace of God or evolution or both) the mind is capable of responding to such connections more spontaneously, without our drawing inferences. Precisely because inferential belief need not 3

Audi (2008), p. 485.

1.

STANDARD COMMITMENTS OF PERCEPTUAL THEORIES

191

arise from inference and because inferentiality need not be phenomenal, showing that intuitions are not inferential is difficult.4

Peter Goldie, on the other hand, invokes the same distinction, but maintains that moral perception involves experiential immediacy rather than justificatory immediacy: In arguing for [the possibility of moral perception], I will deploy a notion of noninferential perceptual belief or judgment according to which the belief or judgment is arrived at non-inferentially in the phenomenological sense (in the sense of involving no conscious reasoning on the subject’s part) and yet is inferential in the epistemic sense (in the sense of being justifiable by the subject after the belief or judgment has been arrived at).5

In Chapter 1, I argued for a conception of perceptual immediacy that was both phenomenal and epistemic insofar as the phenomenology of a perceptual belief must include its own justification; perceptions must be self-evident or self-justifying—vindicated from within the very experience of perception (whether or not external justification is also needed). Like Audi, I want to emphasize immediacy of justification rather than immediacy of acquisition, but also, like Goldie, I want to emphasize the phenomenal character of the relevant immediacy. Through an analysis of the self-evident character of perception, I gave an account of how it is possible to build justification into the phenomenology of an experience. On my view, then, moral perception is immediate in the same way that our perceptions of tables and chairs is immediate (Thesis I)—because its justification is likewise built into its phenomenology. Thesis II insists that the objects of perception (here, moral facts) exist independently of that perception—that the moral facts in question exist even if we fail to recognize them (perceptually, or in any other way). This claim is exactly parallel to the claim that we can’t perceive tables and chairs unless they have an objective existence. (We can hallucinate tables and

4 Audi (2008), p. 486. Jacobson (2005) also opts for justificatory rather than experiential immediacy when he argues that moral knowledge ought to be understood as having the right skills rather than the right experiences, and that it is the largely inarticulate yet reliable nature of such skills that justifies calling it “perceptual knowledge.” 5 Goldie (2007a), p. 347. Justificatory support for a perceptual belief can be viewed as either necessary or as merely possible; Goldie’s language suggests the latter. Problems attend both options, however, as noted in Chapter 1, Section 2.

192

MORAL PERCEPTION

chairs, of course, for things can appear to be objectively present even when they are not, but these experiences are not perceptions.) There are several reasons that someone might endorse Thesis I without endorsing Thesis II. If one thought that an act or an event has a moral value in virtue of its tendency to prompt certain sorts of reactions in us (feelings of shame, or resentment, or gratitude, for example, or the appropriate impulse to action), then one might plausibly suppose that our direct experience of those reactions gives us immediate self-evident knowledge of the relevant moral fact—without supposing that those moral facts exist independently of our reactions. It would be our emotional reactions or our impulses to action that give us knowledge of a moral fact and, simultaneously, make it true.6 Likewise, if one thought that a given act or event becomes valuable in virtue of our decision to value it, or our intention to act in that way, and one assumed that we have direct knowledge of our decisions and intentions, then one could agree with Thesis I but disagree with Thesis II.7 Could such knowledge—immediate knowledge of one’s emotions, one’s impulses, one’s decisions, or one’s intentions—constitute moral perceptions? Insofar as our reactions determine what is moral, it would be possible to determine what is moral by perceiving our reactions, but that is not the same thing as perceiving our reactions (let alone the objects of our reactions) as moral. We may infer moral facts from our reactions, but that is not perception of moral facts as such. It is also possible to experience our reactions as if they were moral properties of the things we react to—to project what is inner onto what is outer, in which case we will have perception-like experiences. But the fact that such projections are seriously misleading about what objects have what properties should disqualify them as genuine perceptions. To believe in moral perception, then, one must endorse Thesis II: one must believe that moral facts are objective facts. Objectivity can be understood in many different ways, however—with regard to moral facts, as with regard to facts in general.8 A strong form of 6 Blackburn and Gibbard have defended sophisticated versions of such a view, which has predecessors in Stevenson and Strawson. Among other things, their work has helped to replace disputes about whether there is such a thing as a moral “fact,” or whether moral judgments can be “true,” with disputes about just what sort of objectivity moral facts, and moral truths, do or do not have. 7 This approach is exemplified by the existentialists, and appears in a more limited form in recent work of Moran (2001). 8 In what follows, I equate facts with actual or existent states of affairs; sentences can describe states of affairs whether or not they are true.

1.

STANDARD COMMITMENTS OF PERCEPTUAL THEORIES

193

objectivity regarding a state of affairs S would claim that the existence of S is independent of any and all responses people can have to S—thus, that S would exist (or not) even if no one was capable of recognizing S. Most would agree that the fact that dinosaurs walked the earth has this sort of objectivity: that fact would exist even if nobody were capable of discovering it. (There are also stronger and weaker readings of “capable of,” which would need to be sorted out for a fuller analysis of this notion of objectivity.) A moral fact such as the fact that Nixon’s lying was wrong would be objective in this strong sense, then, if it would be true that Nixon’s lying would be wrong even in a world in which no one was disposed to view it as wrong.9 A somewhat weaker form of objectivity regarding a state of affairs S would allow that S could exist in the absence of any actual responses on our part, but that it couldn’t exist apart from our dispositions to respond in particular ways. Many think that the fact that grass is green would obtain even if no one ever looked at grass, or even if everyone suddenly became color blind, but that there are no color facts independently of the responses of normal or ideal people under normal or ideal viewing conditions. (The notions of “normal” or “ideal” would require further elaboration in a fuller analysis of this form of objectivity.) A moral fact such as the fact that killing Ruby was wrong would be objective in this weaker sense if its truth were not dependent on our actual responses to the killing but on the responses of normal (or ideal people) under normal (or ideal) conditions. A still weaker form of objectivity regarding a state of affairs S would insist on S’s independence from any particular person’s response but not from the preponderance of people’s responses. The fact that Lady Gaga is popular is independent of any response I happen to have to her, but its existence depends on the existence of enthusiastic responses from many other people. (Determining whether and when the responses of some people are more important than the responses of others is a challenge for this analysis of objectivity.) The moral fact that Clinton shouldn’t have lied to Congress would be objective in this sense even if neither Clinton nor 9

DeLapp (2007, 2009) further distinguishes between the (strong) independence of moral values from beliefs—asserting that “a moral utterance is ‘true’ iff it bears a correspondence to features of the world which exist independently of the evidence or beliefs of anyone (or everyone, or even of epistemically idealized agents)”—and the independence of moral values from human responses of any kind. He endorses the first, which he calls “realism,” while rejecting the second, which he calls “objectivism.”

194

MORAL PERCEPTION

Congress objected to those lies, but not if no one objected. (These three versions of objectivity, even in outline form, do not exhaust the possibilities; they are only meant to give an indication of the variety of ways in which objectivity may be understood.10 ) Which of these sorts of objectivity is relevant to the perception of moral facts will depend on which moral facts are supposed to be perceivable. Some moral facts may be more objective than others, and some experiences of objectivity may be truer to their objects than others. (Insofar as perceptions must be veridical, there must be a reasonably good fit between the properties an object seems to have and the properties it actually has.) Moore claimed to have perceptual knowledge of the simple “non-natural” property of goodness, which he supposed to be independent of any and all human responses (i.e. he endorsed a strong form of the objectivity thesis). Others have maintained that the moral properties we perceive are “thicker” properties such as the property of being cruel or the property of being generous, and that these properties are objective in a weaker sense only.11 Even if the perceived properties are only weakly objective, it may still be the case that they appear to us to be strongly objective—in which case our experiences of those facts are (at least partially) mistaken, resulting in misperceptions rather than (veridical) perceptions. If we experience the wrongness of an act as intrinsic to that act when it is actually a property that depends on community disapproval, then our experience of the objectivity of that moral property is (at least partially) illusory. On the other hand, if the rightness of helping someone in pain becomes apparent through the imagining of other people’s perspectives on that situation, then we will experience this as having a weaker sort of objectivity—with correspondingly weaker requirements for the veridicality of that experience. Perception, I have argued, depends on an experience of objectivity, but there are different types and degrees of perception in accordance with different types and degrees of objectivity. Let us turn, finally, to Thesis III, which concerns the particularity of the objects of perception. We are accustomed to thinking that perception is 10

For an elaboration of dispositionalist accounts and disputes over whether and what sort of realism or objectivity it entails, see DeLapp (2007, 2009). 11 Goldie (2007a) defends the view that it is only “thick” moral properties that can be perceived. McDowell (1985), on the other hand, argues for the perceivability of thick properties but suggests that thin properties may be perceivable as well.

1.

STANDARD COMMITMENTS OF PERCEPTUAL THEORIES

195

always perception of particular states of affairs, in particular spatiotemporal locations—in contrast to thoughts, which can be about more general states of affairs.12 Philosophers who object to the validity of general moral principles, therefore, tend to be defenders of moral perception. Thesis III is especially important to those who want to reject the rule-governed ethics of Kant and Mill, of deontology and of utilitarianism, in favor of a “virtue ethics” that has roots in Greek philosophy as well as (certain branches of ) Christian theology.13 Those like Ross and Audi, who think it is possible to have direct knowledge of general moral principles, are often said to be defending “moral intuitionism” rather than “moral perception.”14 It is useful to distinguish between the following two claims: (1) It is possible to perceive the rightness of a particular action rather than deriving its rightness through general principles. (2) It is only possible to perceive the rightness of a particular action, not the rightness of a type of action. While (1) is a positive thesis that serves to distinguish moral perceptualists from those who believe that our moral knowledge of a particular situation is always derived from our knowledge of more general principles, (2) is a negative thesis that rules out the possibility of perceiving certain things— general principles, or the rightness of certain types of things. (This need not preclude knowledge of such things altogether, since it may be possible to have non-perceptual, derivative knowledge of general principles. Many perceptualists want to deny the validity of any such principles, and hence the possibility of any such knowledge, but that denial is optional.) If one is antecedently committed to the view that perception, by its very nature, must be perception of spatiotemporal particulars and their spatiotemporal properties, then the truth of (2) will seem obvious. Preceding chapters have challenged that view, showing how it may be possible to perceive more abstract objects and relations (objects such as numbers, relations such as 12 There is an important sense in which abstract facts may be perceived as facts about particulars—as described in Chapter 3, Section 4. Aristotle describes moral perception as perception of particulars, but adds that such perception is “not sensory perception, but like the perception whereby we perceive that the triangle is the ultimate particular in geometry.” (Nicomachean Ethics 1142a, 28–9) 13 Also in certain branches of Chinese philosophy. 14 Audi (2008), p. 477, defends perceptual knowledge of “principles expressing obligations of justice and non-injury, of fidelity and veracity, of beneficence and self-improvement, of reparation and gratitude . . . and of liberty and respectfulness.”

196

MORAL PERCEPTION

logical implicature). More importantly, though, insistence on the spatiotemporal character of what is perceived is bound to backfire in the case of moral perception since neither rightness nor the goodness will qualify as spatiotemporal properties. If a defense of moral perception required nothing more than a defense of theses I, II, and III, I would not have much to contribute to the extensive literature on this topic. As we saw in Chapter 1, however, perceptual knowledge requires something more than phenomenal immediacy and justificatory immediacy: perceptual knowledge requires us to experience something as objective; the validity of one’s experience must be evident from within its phenomenology. And this is precisely where most accounts of moral perception come up short: they argue for the justificatory immediacy or the experiential immediacy of moral knowledge but they fail to show how the property of being right or being good can appear within the phenomenology of moral experience as an objective property. Even if a moral judgment—that I ought to stop that child from running into that street, for example, or that you are right to remain silent during this meeting—is not arrived at through inference or deliberation, and is not (in the circumstances) in need of any further justification, the property of rightness does not appear in the phenomenal field in the way that the property of greenness does. We can point to the greenness of the grass, we can trace its outline in space, we can describe its brightness; but no such thing seems possible in the case of moral judgments. The defender of moral perception could opt for a minimal or deflationary account of perception whereby experiential immediacy or justificatory immediacy suffice. But, as we saw in the case of Sam and the spider (Chapter 1, Section 3) this leaves out something distinctive about perceptual knowledge—namely, the way its phenomenology shows us the objectivity of what is known. In the following sections I offer an account of how moral knowledge is able to acquire a perceptual phenomenology of this fuller sort. I shall begin by considering the perception of persons as persons. Given the anti-Kantian bent of most defenses of moral perception, it might seem strange to regard the perception of persons as persons as a primary case of moral perception, but Kant is not alone in placing persons at the center of his moral theory, and understanding the phenomenology of perceiving persons as persons will help set the stage for an account of perceiving the right thing to do (in Section 4, below).

2.

PERCEIVING PERSONS AS PERSONS

197

2. Perceiving Persons as Persons15 Knowing that a given person is a person certainly doesn’t guarantee moral treatment of that person. Indeed, most instances of treating others immorally are also instances in which the offenders know very well that their victims are persons.16 There is a gap between knowing that others are persons and treating them as persons, and another gap between treating others as persons and treating them morally. Employers can treat others as objects (as bodies without minds) rather than persons despite knowing that they are persons, and psychological manipulators can treat others as persons without treating them morally. Without denying the all-too-familiar reality of these possibilities, I would argue that having perceptual (versus non-perceptual) knowledge of the personhood of others makes the first gap less likely— because it is harder to tolerate a contradiction between what one perceives and what one does than it is to tolerate a contradiction between background beliefs and current endeavors. It is relatively easy for the employer to dissociate his actions from his knowledge of the workers’ personhood if that knowledge is not a part of how he sees those workers.17 Perceiving persons as persons is not only valuable as a particularly effective form of moral knowledge, though; it is a moral accomplishment (and, in some cases, a moral obligation) in its own right. It matters to us that we are perceived by others as persons, not just believed to be persons (and if one is capable of perceiving others as persons, one can have an obligation to perceive them as persons). We might find it less painful to be manipulated by someone who fails to perceive us as persons rather than by someone who is keenly attuned to our personhood (because the immediacy of perception can make the mistreatment seem more personal, and thus more evil). Failing to exercise our capacity to perceive persons as persons however (whether or not that recognition leads to appropriate actions) constitutes a moral defect in its own right. If this is right, then perceiving persons as persons is morally as 15 On the assumption that there are such things as persons, I shall use the expression “seeing a person as a person” interchangeably with “seeing that a person is a person.” 16 Indeed, an action may not be immoral (as opposed to amoral, for example) unless the actor knows that the victim is a person. Hamilton (2003), p. 40, remarks that “As Nietzsche emphasized, a cruel person needs, in order that his cruelty be effective, to have a subtle and refined insight into the subjectivity of the person on whom he exercises his cruelty.” This fact is a recognized part of torture techniques. 17 I say more about the advantages of perception for disclosing contradictions, and for prompting appropriate action, in Chapter 3, Section 6.

198

MORAL PERCEPTION

well as epistemically valuable, and understanding what it involves is practically as well as theoretically significant.18 Before we can address the question of how we can perceive persons as persons, though, we need to understand something about what a person is— what it is that we are supposed to perceive. There are many interesting disputes surrounding the nature of personhood, of course, but there are two claims that I regard as sufficiently fundamental and sufficiently uncontroversial to serve as our guides for thinking about what it takes to perceive persons as persons. First, there is a requirement that a person (as opposed to a stone or a mere machine, say) has mental states—beliefs, desires, memories, hopes, feelings, and so on. And mental states cannot be had in isolation; they must exist as part of an integrated whole that is capable of acting in the world. (It is not possible to have beliefs about rain without also having beliefs about such things as water and weather and directions and size; it is not possible to desire rain without having at least some beliefs about what rain is; it is not possible to remember rain without relating rain to other objects and events; and it is not possible for one’s beliefs and desires and memories to be about rain without having the capacity to perceive and act on rain.) So having mental states at all depends on being a reasonably coherent agent.19 Second, there is a requirement that a person (as opposed to a mouse or a cat, say) be capable of a certain amount of self-reflection and self-direction. We do not need to posit a realm of freedom outside of the realm of causes in order to make room for such self-reflective agency, but we miss part of what makes persons into persons if we overlook their ability to be self-reflective and self-directive. (It is this capacity for selfinterpretation and self-direction that underwrites our willingness to hold 18 Discussions of morality tend to focus on our relations and obligations to others—indeed, “morality” is often defined in this way—but we can also adopt more or less moral attitudes toward ourselves. In particular, we can fail to see ourselves as a person. This can happen because we fix our attention too close to the surface and fail to see what lies beyond (we identify ourselves with our accomplishments, for example, or we see ourselves a mere arrangement of bodily features). Or it can happen because we simply assume a self beyond appearances without really bothering to see just how this self is made manifest in our movements, etc., or just what this self is really like (selves, so understood tend to be extremely abstractly conceived, so that we are hard pressed to distinguish one from another). 19 Classic defenses of this claim go back at least to Spinoza, and they get fuller treatment in the work of Ryle, Davidson, Block, Stich, Burge, and many others. I do not want to deny the presence of multiple kinds of psychological incoherence (within one’s self-conception as well as between one’s unreflective beliefs and desires), but a significant degree of coherence does seem necessary for having mental states and for being a person at all. For a fuller statement of my understanding of this requirement, see Church (2005), pp. 31–41.

2.

PERCEIVING PERSONS AS PERSONS

199

others responsible for their actions, for example.) Regarding a person as a person thus depends on recognizing the possibility of agency with respect to oneself as well as agency with respect to the surrounding world.20 It is tempting to suppose that human agency, whether directed at the external world or back on ourselves, is not the sort of thing that can be perceived. For, first, it seems to involve an incredibly complex intertwining of attitudes and contents that could not all be experienced in a single moment. Second, insofar as agency has a location, it seems to exist “behind” a person’s perceivable actions, “behind” a person’s perceivable skin, “beyond” the reach of perception. (Note how different it is, in this regard, from simple properties like greenness.) If perception were merely a matter of automatic and reliable belief acquisitions (a minimal or deflationary notion of perception), it would be easy to explain our ability to perceive people as people. Evolution could have designed us to make reliable judgments about what is and what is not a person automatically—whatever the underlying mechanism might be; or it could be that our environment rewards us for quick and reliable recognition of the patterns of behavior that are indicative of personhood. Explanations such as these, however, do not explain the way that the psychological complexity of a person can appear in a moment, or the way that people appear to have “inner” lives. Indeed, if reliable and automatic belief acquisition is all that is necessary for perception, experiences of human complexity and inwardness will be optional—merely indicative of personhood at best, illusory and misleading at worst. But we have rejected a minimalist understanding of perception in favor of an account that does justice to the distinctive phenomenology of perception—the phenomenology of objectivity, which is also the phenomenology of spatiality. In previous chapters, I have argued that we perceive an object such as a tree insofar as we simultaneously imagine how it would appear from other points of view. For it is the imagined convergence of different perspectives around a single state of affairs that gives us the experience of its spatiality and its objectivity. Likewise, when we perceive one event as causing another, we simultaneously imagine how differences in one event correspond to differences in the other, and it is the recognition of invariance across those 20

Kant’s claims about personhood and autonomy constitute a classic example of these claims. Frankfurt (1971) offers a more recent and influential defense.

200

MORAL PERCEPTION

changes that gives us the experience of causal dependency. When it comes to perceiving someone as an agent—as the locus of multiply intertwined mental states, I want to make a similar suggestion: to perceive that someone is a person we must imagine that person in a variety of situations—some of which are actual and many of which are merely possible, and we must discover what is invariant across those imagined actions.21 Often it is possible to discover invariants on the basis of inferences (in particular, induction) from a series of cases that are observed over time. The perception of invariants on the other hand requires that a number of different possibilities be actively imagined at one time. As we watch a thief go about his business, for example, we recognize his desire to stay hidden insofar as we simultaneously imagine his movements across a range of possible circumstances, we see his anxiety insofar as we see a certain invariance in the way we imagine him moving in moments past and moments ahead, and we see his plan for escape because we imaginatively project from his backwards glance to a variety of paths to the door should he be interrupted. It is the active imagining of these counterfactual possibilities that reveals the complexity of agency, and it is the recognition of high-level invariants that gives one the sense of forces operating behind and through the surface of things. Even as we can watch a tree bending in the wind and, through the imagining of various alternatives, come to see the intertwined forces causing it to move as it does, so too we can watch a person moving through a room and, through the imagining of various alternatives, come to see the mental states that guide his actions.22 While a certain amount of integration or unity is necessary for the existence of mental states, and thus for our perception of mental states, everyone has tendencies that are hidden from others (leading to an inevitable partialness in what is perceived) and areas of incoherence (leading to an inevitable fragmentation or multiplicity in what is perceived). Our 21 This accords with Campbell’s (1995) view that persons are experienced as the common cause of a variety of behaviors; and to Hobson’s (2005) view that we discover both ourselves and others to be similar causes through social interaction as infants. 22 Note how this account is neutral with respect to the (now dated) debate between “theory theorists” and “simulation theorists”—a debate about whether such knowledge (and the imagining that accompanies it) arises from general theories about human behavior (which sounds too intellectual), or from putting oneself in the shoes of another (which sounds too self-centered), or from something else entirely (so-called “mirror neurons,” for example). It is the role of imagination, not its sources, that is important in explaining the phenomenology of perceptual knowledge.

2.

PERCEIVING PERSONS AS PERSONS

201

aspirations to perceiving the minds of others should be limited, accordingly; there are times where we should not even try to perceive a person’s mind. Also, given the imaginative filling in that underwrites any perception, we should recognize that there can be several different (equally accurate, equally inaccurate) ways to achieve perceptual integration. These are all empirical limitations on the possibility and the authority of our perception of persons as integrated agents. Perceiving persons as self-reflective agents—the other aspect of personhood—may have greater limitations. Perceiving a person’s mood, concern, or immediate intention (even seeing them as a unified agent) is not the same as perceiving a person’s self-conception. Yet it is having a self-conception, being self-reflective and self-directive, that distinguishes persons from other animals. (The mental states ascribed to the thief above could just as well be ascribed to a thieving dog.) However long we watch the thief at work, it seems it may be impossible to know, much less know directly, how he regards himself. But if we can’t perceive that part of personhood, we lack perceptual knowledge of a morally relevant part of persons—the part that makes people self-reflective and self-directive, the part that makes people not only ends for others but ends in themselves. There are many important contexts in which we can see (albeit only partially) the nature of a person’s self-conception. We are often able to see that a person is self-conscious because we can see that their attention is focused on themselves and that this attention is self-monitoring or selfcorrective, for example. We are also, at least sometimes, able to see how people monitor and correct their states if we attend to the different sorts of attitudes they direct towards themselves. Some people adopt a dismissive or deprecating attitude when their attention is focused on themselves while others become proud, and some styles of self-deprecation seem genuine while others seem like attempts to disguise one’s pride. A sufficiently perceptive observer does not infer another’s pride from the lift of a head or the slant of a smile; rather, the pride is seen in the position of the head and the shape of the smile insofar as these belong to a larger pattern of self-reflective behavior—the memory of which, or the imagining of which, infuses one’s current experience of that person’s gestures and expressions. Furthermore, we can listen to the way that people talk about themselves (and about others), noticing what they say and how they say it, to discover things

202

MORAL PERCEPTION

about how they regard themselves.23 Hearing my neighbor talk about himself gives me knowledge, direct rather than inferential, about how he thinks about himself—knowledge conveyed through choice of words and intonations as much as the propositional content of what he says. In such cases, it is not an individual word or an individual intonation that reveals a person’s self-conception; rather, it is a pattern of word choice and inflection (whether actually observed or accurately imagined) that gives us such knowledge. Odd as it may sound, a good observer of other people can often perceive their self-conceptions better than they can themselves. Direct reports on what sort of person you take yourself to be are frequently misleading—both because of your wish to present yourself as different than you are, and because you are usually not the best observer of your own tendencies. Our perception of a person’s self-directed mental state is exactly like the perception of a person’s outwardly directed mental state insofar as it requires our present experiences of the person to be infused or overlaid with other, merely imagined (or remembered) alternatives. It is different only insofar as the relevant alternatives are concerned with second-order attitudes; they are alternatives in which people reveal the attitudes they have towards their own attitudes. This account of perceiving persons might seem to overlook a central fact about persons—their essential unknowability. Agreeing that it is possible to imagine a person’s mental states by imagining what that person would do in a variety of circumstances, and that it is possible to imagine how others view themselves by recognizing the images that unify their self-reports, one can still insist that such imaginings always fall short of knowledge (even when they turn out to be correct) because, as persons, we are always in the process of remaking ourselves. To ignore this fact would amount to both a metaphysical mistake and a moral failing. In countering one danger, this line of attack threatens to fuel another. While supposing that a person’s mental life is readily perceivable runs the danger of treating persons too much like machines, supposing that a person’s life is ultimately unknowable runs the danger of treating persons too much 23 I follow most contemporary writers here in assuming that we do not (except in exceptional situations) make an inference from words to meaning, or from the meaning of the words to the speaker’s meaning. We experience the one as transparent to the other; we see through one to the other.

2.

PERCEIVING PERSONS AS PERSONS

203

like ghosts. While treating persons as machines risks objectification, treating persons as ghosts risks skepticism. Much of Wittgenstein’s later work is dedicated to steering us away from both of these dangers. On the one hand, he tries to show us the absurdity of radical doubt. I can know that you are in pain and I can know that directly—by hearing your cries and seeing your grimacing. We would (and we should) question the sanity and the morality of anyone who continued to feel doubts in the face of behavior that is so evidently that of someone in pain. An inferential model of my knowledge of your pain not only fails to fit the phenomenology of such knowledge (I see your pain in your behavior, I don’t infer the pain from your behavior); it also serves to alienate us from others’ pain (I am less responsive to your pain if I have no direct knowledge of it). On the other hand, Wittgenstein insists that the pain is not the same thing as the pain behavior, and that recognizing persons as persons depends on recognizing a certain amount of uncertainty. His key to escaping this apparent impasse, I think, rests with his realization that when there is uncertainty, it is an uncertainty about the behavior as much as about the inner life—an uncertainty about which patterns are relevant to a person’s mental states and which are not. The fact that we are uncertain about the inner life of a person doesn’t require an inferentialist account of how we have knowledge (or fail to have knowledge) of that inner life—as though mistakes are only possible when there is a mistaken inference. Uncertainty about the presence of a fear is like uncertainty about the presence of a barn, where the meaning of what it is to be afraid is no more (and no less) precise than the meaning of what it is to be a barn. “So an uncertainty about the outer corresponds to an uncertainty concerning the inner”24—not because of the unreachability of the inner but because of our uncertainties about how to categorize the outer. “There is uncertainty and there is certainty; but from this it does not follow that there are criteria that are certain.”25 To insist on certain criteria would,

24

Wittgenstein (1980–92), p. 68. Wittgenstein (1980–92), p. 87. Also: “[I]t is not true that uncertainty in recognizing his irritation (for instance) is simply uncertainty about his future behaviour. Rather in the concept there is an uncertainty of criteria.” (p. 70) “I think the unforeseeability must be an essential property of the mental. Just like the endless multiplicity of expression.” (p. 65) Wittgenstein does raise the question of whether we would exchange our language game that “rests on ‘imponderable evidence’ and frequently leads to uncertainty” for one that did away with such uncertainty (p. 95). But he seems to think this is an 25

204

MORAL PERCEPTION

furthermore, amount to a denial of personhood, where personhood includes the ability (and the right) to re-conceive and thereby remake oneself.26 Our metaphysical attempts to explain our uncertainty by appeal to the inwardness of mental states can lead to a certain alienation from others, then; and our alienation from others can lead us to a view of the inner as hidden.27 This is a theme that has been developed at some length by Stanley Cavell. In an extended discussion of frogs and princes, statues and dolls, and automatons with and without blood, Cavell explores the conditions under which we may or may not be willing to recognize a body as a person.28 While the capacity for behavior of sufficient subtlety seems necessary for our attribution of personhood (subtlety that the frog lacks, even in the body of prince), no amount of subtlety can ensure that we see a given body as a person (versus a mere automaton). That, according to Cavell, requires a kind of moral leap of faith (I prefer to think of it as a moral commitment of the imagination) through which we overcome doubt about the very existence of a person while also acknowledging the fundamental indeterminacy of persons as such. To be a skeptic is to avoid knowledge of other persons (and the claims they have on us), and to overcome one’s skepticism is to embrace a kind of knowledge that includes uncertainty. For Cavell as for Wittgenstein, positing an inner life amounts to giving our uncertainty about others a place in our picture of the world—a way that effectively secures rather than weakens our knowledge of others as persons.

3. Uncanniness and Ambivalence The previous section tried to do justice to the phenomenology as well as the epistemology of perceiving persons as persons; but it ended with a recognition unanswerable question; it would require a radical change in our way of living, and there is no neutral standpoint from which to make the choice. 26 Uncertainty is due to indeterminacy of interpretation, not blindness. The indeterminacy is not due to a lack of access to something interior, though; it is due, instead, to the necessary incompleteness of any interpretation—especially when the interpretation depends on self-interpretation whose continuation is part of what it means to be a person. 27 See Chapter 6 for further discussion of “distancing” that is not alienating, of “alienation” that adds to rather than subtracts from intimacy and understanding. 28 Cavell (1979), pp. 398–414.

3.

UNCANNINESS AND AMBIVALENCE

205

of the essential (not just accidental) limits on such perception. Put somewhat paradoxically: to see a person as a person depends on a certain failure of perception, for it requires that our attempts at perception remain incomplete. In this section, I expand on the phenomenology of this desirable incompleteness, relating it to experiences of the uncanny and to its emotional counterpart in ambivalence. While an appreciation of uncanniness and a tolerance for emotional ambivalence are important parts of the phenomenology of moral experience, they are not required for moral perception. So this section is something of a digression from the central thread of this chapter and this book. Readers who want to stay with that thread may choose to skip ahead to Section 4. In The Uncanniness of the Ordinary, Cavell offers the following summary of his project: “I might describe my philosophical task as one of outlining the necessity, and the lack of necessity, in the sense of the human as inherently strange, say unstable, its quotidian as forever fantastic.”29 The necessity of regarding humans as strange and fantastic is, presumably, the necessity of recognizing their states of mind as inherently indeterminate; but this is not to be equated with the necessity of any radical doubt about the very existence of an inner life. What he calls the uncanniness of the ordinary is the sense that everything could appear just the way it does without there being, in fact, any people at all; and it is our response to the uncanny— either re-embracing or rejecting what was formerly familiar—that makes us either receptive to or unreceptive to, either open or closed to, the presence of others.30 Like Freud, Cavell equates a sense of the uncanny with the return of something familiar—in particular, the reanimation of what has become inanimate.31 Unlike Freud, however, he correlates a sense of the uncanny with getting something right rather than getting something wrong. One

29

Cavell (1988), p. 154. McDowell’s (1995) calls for a “re-enchantment of nature” have a similar flavor. 30 His talk of return to the familiar assumes a “normal” starting point—versus that of the sociopath. “The return of what we accept as the world will then present itself as a return of the familiar, which is to say, exactly under the concept of what Freud names the uncanny. That the familiar is a product of a sense of the unfamiliar and of the sense of a return means that what returns after skepticism is never (just) the same . . . A difference in which everything and nothing differs is uncanny”(p. 166). 31 Cavell complains that Freud denies the importance of animate/inanimate confusion (remarking that he finds Freud’s denial itself uncanny), but I read Freud as insisting that the animate/inanimate confusion is just part of the story—the other part being the return of a repressed desire.

206

MORAL PERCEPTION

reason for this difference is that Freud focuses on cases where there is, in fact, no person present—a case in which a doll seems to be alive, for example—while Cavell focuses on cases where there really is person present—for example, a case in which a body presented as a mere machine is in fact a person. But there is another reason that Cavell, unlike Freud, considers a sense of the uncanny to be indicative of insight rather than illusion: whereas Freud traces the unsettledness that characterizes the uncanny to the emergence of childhood confusions that the adult mind has sorted out, Cavell attributes the unsettledness that surrounds the uncanny to the double consciousness we experience when we retain (or regain) our confidence in the existence of other minds despite our realization that the evidence for other minds will never be certain. Put another way: according to Freud, experiences of the uncanny are the result of intrusions of a more primitive mentality in which inner and outer are merged, while according to Cavell experiences of the uncanny are the result of a reflective awareness of the unclosable gap between inner and outer. The difference is important insofar as it suggests contrasting notions of what is appropriate to seeing persons as persons: whereas Cavell seems to think that a sense of the uncanny is an essential part of appreciating the innerness of persons (their inescapable strangeness), Freud seems to think that a sense of the uncanny is reflective of our inability to distinguish between inner and outer, indicative of failure rather than success in perceiving persons as persons. Here are two examples of things we might experience as uncanny: (1) a flower looking at us; (2) the same car approaching us repeatedly throughout a road trip.32 In both cases, the uncanniness of the experience depends on the sense that there is a person or an agent where surface appearances suggest otherwise. Also, it requires a certain degree of discomfort around the mismatch, an awareness of something highly unusual or out of the ordinary. (Panpsychism, which attributes agency everywhere, would tend to dissipate the sense of something uncanny in the experience of an observant flower; and theism, which posits an omniscient and omnipresent god, would tend to dissipate the sense of something uncanny in the experience of recurring encounters. Whether such convictions actually preclude experiences of the uncanny will depend, in part, on whether they succeed at infusing the way 32

Freud (1914/1975) regards repetition as an indication of unresolved trauma; here it suggests the presence of mysterious powers.

3.

UNCANNINESS AND AMBIVALENCE

207

we perceive the world around us; as mere suppositions, they allow for the possibility that something feels strange even though we believe it is normal.) I think that Freud is right that uncanny experiences, whether experienced in response to flowers or humans, are indicative of a problem, not an insight. In response to flowers, uncanny experiences amount to perception-like experiences of a depth (of mental states, of agency) that is not there; and in response to humans, uncanny experiences indicate a failure to connect perceptions of a surface (which lacks mentality) with one’s beliefs about a depth (a mentality) that generates that surface. But Cavell is right that the gap that opens in the uncanny (the gap between surface and depth) is precisely the gap that makes humans appear as something more than objects. Seeing a person as a person requires seeing through the outer to the inner in such a way that the gap remains, but is not so unbridgeable or arbitrary as to make it uncanny. To merely believe in an inner life but fail (perhaps to refuse) to perceive it—to make no effort at imaginative synthesis—is to lose our sense of (versus belief in) others as persons. A return from skepticism to the ordinary is really a return to the processes of imaginative synthesis that make perception of persons possible; but if that return is made, there should be no more sense of the uncanny.33 There is a somewhat different form of double consciousness that remains appropriate, however, to our experience of others as others—an experience of ambivalence rather than uncanniness. The sort of ambivalence I have in mind has more to do with the give and take of interpersonal exchanges than with an inner struggle over conflicting attitudes. It is rooted in approach/ avoidance conflicts and incorporation/expulsion conflicts, and it is manifest in conflicts between curiosity and wariness about a new situation, trust and distrust of others, attraction and repulsion towards what is strange. These are not cases in which I am curious about something under one description but wary under another, or trustful under one description and distrustful under another; rather, I am confronted with something that I do not yet know

33 In describing the uncanniness that often pervades the novels of Vladimir Nabokov, John Banville (2010) suggests that for an “unenchanted reader” it can seem like “a willful chloroforming and pinning down of that brightly fluttering spontaneity that is the essence of reality,” which is also a source of terror; but, when enchanted, it gives rise to a sense of pantheistic bliss in which the fluidity of everything becomes “transfigured into something like second sight” (p. 46). What makes it uncanny, though, is the simultaneous activation of both these responses.

208

MORAL PERCEPTION

how to describe and therefore I am both curious and wary, both trustful and distrustful. Experiences of the uncanny could be viewed as extreme instances of this pattern. When a flower seems to be looking at us, we are simultaneously drawn towards and repelled by the flower’s gaze. We experience the flower, as we more rarely experience other people, as fully familiar and utterly strange. Most of our interactions with other people are not uncanny, of course, but the same sort of ambivalence is often present and appropriate. When my friend tells me good news, for example, I am drawn into her situation—asking for further details, imagining what is ahead, and making her happiness mine as well. There is (or should be) another side to my response also. I am loath to assimilate her situation to mine, to insist on a full accounting, or to feast on her happiness as though it were mine. I feel the need to allow her her space—not only as a self-imposed brake on the first impulse, which could be oppressive, but also as a felt aversion to the invasion or co-opting of her experience. So, even as I ask questions and express my own pleasure in her happiness, I draw back in appreciation of what is alien in her answers and what feelings are uniquely hers. Or consider an equally ordinary situation in which I am confronted by a panhandler on the subway. The spiel is familiar, and vaguely threatening, so I am repelled and somewhat fearful. But I am also intrigued by what I don’t know about this person, by the aspects of the situation that elude me, by the possibilities that I can hardly begin to imagine. And, again, the contrary desires both to know and not to know, to imagine and not to imagine, are not separable from an emotional ambivalence of repulsion and attraction, fear and desire, horror and fascination. It is interesting to note how some standard strategies for overcoming ambivalence seem either inappropriate or counterproductive in such cases. Frankfurt recommends that we simply opt (in so far as we are capable) for one attitude rather than the other—not on account of any reason that favors that attitude rather than the other, but for the sake of whole-heartedness.34 Here, however, the conflicting attitudes indicate a deep recognition of another person as another person; so abandoning either would amount to viewing that person as either too similar or too different. Rorty, in contrast, recommends that we think of more imaginative ways of describing and 34

Frankfurt (2004), Chapter 3, Section 15.

4.

PERCEIVING WHAT TO DO

209

resolving the apparent opposition.35 Here, however, the opposing attitudes or inclinations do not depend on different descriptions; they precede descriptions and conceptualizations. It is the fundamental strangeness or elusiveness of a situation or a person—its essential indeterminateness—that I am responding to in these cases, not two different ways of conceptualizing it. (Furthermore, attempts at developing descriptions that are thick enough to preserve the attitudinal conflict in some constructive way seem to be counterproductive insofar as such attempts are attempts at managing or coordinating difference when it is precisely our willingness to be dislodged by difference that is important in these cases.) Unlike cases in which our own psychic unity is threatened in some fairly determinate ways, these are cases in which our psychic completeness is undone by an indeterminate other. The analog is not that of a plurality of perspectives within a given community so much as it is that of an encounter with a potentially transforming work of art.

4. Perceiving What to Do Considerably more controversial than the claim that we can perceive persons as persons is the claim that we can perceive what is morally right, or what ought to be done. It is this later claim, though, that is especially important to advocates of moral perception. Even if seeing persons as persons is morally valuable in itself, it is seriously incomplete as a guide to moral action. The morally perceptive person not only sees that another person is in pain; she sees that it is right to help alleviate that pain.36 To take one of Martha Nussbaum’s favorite examples, the momentous moral insight of Maggie Verver and her father (in Henry James’ novel The Golden Bowl ) is not their standing as persons—something that was never “out of sight” between the two of them—but, rather, the fact that they ought to part. One apparent problem with the notion of perceiving what we should do as opposed to seeing what we are is the fact that the prescribed action is not 35

Rorty (2008). Seeing people as people is not even a necessary condition for seeing what is morally right. Some morally right actions may not concern people (the alleviation of suffering in animals, for example), and we can sometimes see the rightness of a particular act (rocking a crying child) even when we are not attending to their status as a person. 36

210

MORAL PERCEPTION

(yet) actual, and thus, it seems, not yet perceivable. In Chapter 4, however, I argued that we can perceive future events insofar as we can perceive how they are the necessary outcome of current conditions. We can see that a child is about to fall, or that a friend is about to smile because we recognize that the current situation can’t but result in such a fall or such a smile; and we recognize this because our imagining of various counterfactual possibilities always has this same outcome. Likewise, Maggie and her father can perceive their future parting insofar as their actions are directed, and are bound, by certain assumptions about who they are, and about what is important, such that no other outcome can be entertained. They come to see their parting as the inevitable outcome of following through on the various commitments and constraints of their lives. Just as we might see (directly, in all of its particularity, and as an objective fact) that a tree must break given the force and the direction of the wind, two people can see that, given their circumstances, they must part. Maggie and her father do not simply imagine their future parting and sense its rightness; rather, they imagine many different futures in which their central value and commitments remain, and they see how all of these futures include a parting of ways.37 Let me introduce another example to illustrate this understanding of how it is possible to perceive what we ought to do. Suppose that you are confronted with the prospect of reporting a thief; you want the thief to receive an appropriate punishment, but you are not sure of how to make that happen. Believing that the thief is a bad but not seriously dangerous person, and believing that your report will lead to months of incarceration in an unhappy but not horrible prison, you wonder whether reporting the thief is the right thing to do. It might not be possible to perceive the right thing to do (indeed, knowledge of what you should do may elude you entirely), but if you imagine a range of likely outcomes—face-to-face encounters in a court, months spent in a cell, tense interactions with guards and with other prisoners—you might discover that the stable elements of these scenarios (a public airing of the thief ’s wrongdoing, his temporary isolation from normal life, his enforced attention to issues of personal safety) fulfill your desire for appropriate punishment or you might discover that at 37

Walker (1987) develops a similar example concerning a son’s decision about the care of his ailing mother. She, however, emphasizes the plurality rather than the singularity of appropriate moral choices.

4.

PERCEIVING WHAT TO DO

211

least some of these elements (the likelihood for humiliation, or of grave bodily harm) are distinctly at odds with your desire for appropriate punishment. If there is a fit (and if your conception of appropriate punishment is valid—a condition I shall return to below), then you have perceived the rightness of reporting the thief; if there is a misfit, then you have perceived the wrongness of reporting the thief.38 In the first case it is the nature of Maggie’s relationship to her father, and in the second case it is the nature of appropriate punishment, that sets the relevant constraints on what is morally necessary or morally possible. (Given the constraints of Maggie’s character and situation, parting might be the only moral possibility, a moral necessity. On the other hand, given the constraints on appropriate punishment and given the existing penal system, reporting the thief will be one of several moral possibilities at best.) The restrictions on moral actions are considerably narrower than the restrictions on physical actions, of course; many things that are physically possible are morally impossible. All necessities, however—whether causal, conceptual, metaphysical, or moral—assume some framing of relevant possibilities. Causal necessities hold on the assumption that background conditions are normal; conceptual laws hold on the assumption that our concepts remain sufficiently stable; and metaphysical necessities hold on the assumption that the world remains sufficiently stable. Moral necessities, then, are necessities on the assumption that certain values and commitments are preserved, where the only relevant possibilities are those that respect those values and commitments. Maggie and her father must part, ceteris paribus—where the ceteris paribus clause is filled out in such a way as to ensure that their central values and commitments remain intact. In looking for, and finally perceiving, what they must do, Maggie and her father restrict the relevant possibilities to those that preserve the relevant values, and they come to see that parting is something that remains invariant across those possibilities. Likewise, a thief may be reported, ceteris paribus—where the ceteris paribus clause is filled out in such a way as to ensure that the punishment is appropriate (proportional to the crime, preserving of dignity, and so on). In trying to perceive what to do with respect to a thief, one must try to imagine various possible actions within the bounds of those constraints. Insofar as, given 38

Note how this account contrasts with that of Blum (1994) who emphasizes attention to certain facts as being “moral salient” . . . for him, a matter of certain facts “standing out.”

212

MORAL PERCEPTION

these constraints, a particular course of action can be imagined, it will be perceived as morally possible; and insofar as, given these constraints, all imagined courses of action share a particular feature, that aspect of the action will be perceived as morally necessary. There is an important difference between (a) imagining a possible state of affairs and, on the basis of that imagining, regarding it as unacceptable, and (b) finding oneself unable, or deeply resistant, to imagining that particular state of affairs.39 Maggie might imagine a scenario in which she remains with her father and both his marriage and her own suffer; not wanting to cause such damage, she could then conclude that they should part. Or she could try to project a future that preserved good relations with her father and with her husband, and between her father and his new wife, only to discover that parting from her father is the only possibility that remains imaginable under these constraints. Likewise, you might imagine what it is like to spend years in prison and decide that you are not willing to inflict that on anyone. Or you might imagine a string of events resulting from your desire to punish a thief only to find that when it comes to imagining his life in prison you can’t proceed any further with your imagining, for it quickly turns into something that no longer fits under the constraints set by your initial desire for appropriate punishment. The former sort of imagining combines with one’s preferences to reach a moral decision, while the latter sort of imagining, already operating within the constraints of morality (more exactly, the constraints of one’s own understanding of morality), directly presents certain options as morally possible or morally necessary. It is this latter type of imagining that is required for moral perception. If the constraints on one’s imagining are simply set by one’s desires— one’s desire to make a loved one happy, or one’s desire not to cause prolonged suffering—then the experience of these constraints as objective constraints will be a misleading projection of merely subjective constraints, and the resulting knowledge cannot be considered to be perceptual after all.40 Objectivity, as noted in Section 1 above, comes in different forms and strengths, and we can illuminate the nature of moral perception without 39

The phenomenon of “imaginative resistance,” in response to fiction but also in response to life choices, was described and theorized in interesting ways by Gendler (2000), whose work on this topic has been particularly influential. 40 If the desires that constrain one’s imagining are themselves a response to objective values, then what is perceived as a moral necessity (or a moral possibility) may indeed be objectively necessary (or

4.

PERCEIVING WHAT TO DO

213

taking a stand on just what sort (or sorts) of objectivity morality has; but in order for the notion of moral perception to be plausible, the objectivity of moral requirements can’t be utterly mysterious. If what one sees as morally required (or morally allowed) is really just a reflection of one’s own desires at the time, then there is no objectivity to one’s judgment, and the sense that one is perceiving something objective is an illusion. According to many, moral judgments are objectively true or false depending on whether or not they conform to the norms or commitments of the community (or communities) to which one belongs. The norms or commitments of a society might not be explicitly endorsed by members of the community, but they can provide standards against which the desires and the actions of particular individuals can be judged. This constitutes a weak sort of objectivity since it depends on community endorsement, and because the norms that are endorsed are likely to vary between communities. It is easy to see, however, how such norms could place constraints on one’s imagining such that some acts could appear to be necessary (or impossible) and where the necessity in question had this kind of objectivity. It could be that Maggie, for her own part, would rather stay with her father and risk her marriage; but the norms of her community—in which a woman’s marriage ties ought to trump her familiar ties—constrain her imagining, and she comes to perceive parting as the only possibility. Likewise, you may want a thief to suffer horribly, yet when it comes to deciding how to act you constrain your imagining with certain notions of fair punishment and comparable harm. Another option, favored by many defenders of moral perception, holds that the constraints on moral perception are given by a loving or caring attitude or by a properly developed sense of shame. (This option is not incompatible with the community norm option above, since the community norm may advocate a loving attitude and since shame may be the means by which adherence to community norms is enforced. On the other hand, one could regard love and/or shame as a moral standard that is independent of any community norms.) Nussbaum, for example, presents Maggie’s love for her father, and her father’s love for Maggie, as the overarching attitude

possible). In such cases, the constraints on one’s imagining are not “simply” set by one’s desires, or “just” a reflection of one’s desires.

214

MORAL PERCEPTION

that makes moral perception possible.41 Both love and shame can be selfserving and biased, however, calling into question their ability to provide an objective basis for moral judgment. Quite apart from questions that might be raised about different sorts of love and their distinctive dangers, it is worth questioning the scope of any morality that depends on love in this way. If moral knowledge is a kind of knowledge that only love can provide, we are pretty much precluded from distinguishing between moral and immoral treatment of strangers. What seems right, however, about Nussbaum’s appeal to love as a source of moral insight is her recognition that moral concern for another must be a concern for the preservation of another’s person, and love often enhances our perception of that person as a person.42 The requirement that we respect the personhood of persons is a central constraint of a Kantian morality. (I say more about this in the next section.) If Kant is right in thinking that respect for persons is an objective moral constraint, and that there are objective facts about what acts do (or do not) respect the personhood of persons, those facts can constrain our imagining of possible actions and can ensure that perceived necessities have an objective basis. This seems to be a stronger, and more plausible, sort of objectivity than that provided by community norms or loving attitudes—since it allows us to discount community norms on the grounds that they don’t respect personhood; and since it allows us act morally with respect to others without loving them.43 As I have argued above, being a person includes being a reasonably integrated center of agency and being reasonably effective at self-reflection and self-creation. So whatever is needed to support these two aspects of persons (one being more knowable or attainable in some situations, the other more knowable or attainable in other situations) is what will appear as morally necessary if our imagining is constrained by the need to respect the personhood of persons. Thus a Kantian account of Maggie’s moral perception might focus on her respect for her father’s

41

See, also, Noddings (1984), Watkins and Jolley (2002), Goldie (2007a). See Velleman (2006) for an insightful discussion of love as a form of respect. 43 I realize, of course, that the communitarian view will regard a Kantian concern with personhood as the central concern of a particular community, not as something that can transcend community norms; and that advocates of a loving attitude will argue that love enables us to perceive others more fully and more accurately. Here I am merely describing some candidate accounts of the objectivity of moral judgments, and moral perception, without trying to decide between them. 42

4.

PERCEIVING WHAT TO DO

215

personhood—especially insofar as respecting his personhood depends on respecting his autonomy. Before looking more closely at the Kantian account, I would like to note some of the limits of moral perception. Moral perception, by its very nature, is less automatic and less universal than perception of physical objects such as tables and chairs. More training is required to become a good perceiver. Even if it is the case that babies have moral sensitivities, it is also true that those sensitivities are seriously deficient;44 and even if it is true that our moral sensitivities are responsive to fundamental facts about human nature—about the requirements for human happiness, for example—learning those facts takes more time than learning the shape of a nearby object. Even with extensive training, our moral perceptions are much less reliable than perceptions of nearby medium-sized objects—largely because of the power and prevalence of competing desires and ideologies. Virtuous people may be less selfish, or less tempted by other goods, than the rest of us, but even virtuous people are subject to competing desires and are susceptible to defective ideologies.45 Without impugning the status of moral perception as knowledge (our justification for, or entitlement to, a judgment does not require that we are never mistaken about such things), these observations suggest that we should be much more cautious about trusting (putative) moral perceptions than we are about trusting our (putative) perceptions of ordinary physical objects. When we do succeed in perceiving the right thing to do, we may nonetheless fail to do what is right. We may be physically prevented from acting as we intend, or we may fail to align our intentions with our perceptions of what is right. This does not mean that perceptions are motivationally inert, however, for in most cases we are trying to perceive what is right because we already care enough to want to do what is right. Timothy Chappell highlights the fact that perception is designed to motivate: As a matter of the history of our species, the (original) point of perceptual capacities in a tough world must usually have been to mandate response rather than to get hold of information for its own sake. Think this way of echolocation: the bat gets hold of a pattern which is the shape of a wall in front of it, and that pattern mandates a response which is an alternation of flight-path. The general schema for such 44 45

See Bloom (2010). See Jeffrey Seidman (2005) on McDowell on this.

216

MORAL PERCEPTION

motivating representations will be: Pattern P in context C mandates response R from X. It is not obvious why this schema cannot apply to moral perception in humans just as much as to echolocation in bats.46

Even when our perceptions lack an evolutionary purpose, though, they may be directly motivating for, as Peter Goldie puts it, when comparing moral skills to chess skills, as long as we are “fully engaged in the activity in which the skill is brought to bear,” one’s actions will accord with one’s perception of what is right pretty much automatically (though not inevitably).47 Indeed, I have argued (in Chapter 3, Section 6) that one of the advantages of perceptual knowledge over other sorts of knowledge is its more automatic connection to action. The extra precariousness of moral knowledge and the extra motivational force of perception together suggest that there will be many cases in which we should be wary of even trying to perceive the right thing to do. Even when I know someone well, merely thinking about how best to support her (drawing inferences from past successes, recalling something I heard her say, and so on) may be more reliable than imagining different courses of action in an effort to determine the right way to act. It is one thing to perceive suffering as a reason to help, or to perceive sadness as a reason to be kind (these are cases of perceiving reasons, as discussed in Chapter 3); we are generally reliable at perceiving such reason-giving relations, and our awareness of competing reasons helps to moderate the actions that follow from such perceptions. Perceiving the right thing to do is a riskier (if also more rewarding) endeavor, with greater room for error and fewer moderating influences. As a result, while I am confident that it is sometimes possible to perceive the right thing to do, I am far from confident that such perception does or should make up the bulk of our moral knowledge.

5. Back to Kant We are now in a position to reconsider the alleged incompatibility between Kantian accounts of moral knowledge and perceptualist accounts of moral responsibility. Kant is famous for his insistence that we treat persons as 46

Chappell (2008), p. 434.

47

Goldie (2007a), p. 351.

5.

BACK TO KANT

217

persons (i.e. never merely as means to our ends, but always also as ends in themselves), and that has already been a part of our above discussion of moral perception. But he is even more famous for the first version of the Categorical Imperative, requiring that we act only on that maxim which we can simultaneously will to be a universal law. And it is this insistence on universalizable laws that is most commonly attacked by advocates of moral perception.48 Nussbaum (1990) is interestingly of two minds about this. She writes: “No narrative dealing in empirical particulars could see Kant’s conception in a fully sympathetic way.” But on the next page, she remarks: Of particular importance here will be to ask what role rules and universal principles can and should play inside the morality of perception, and in general what sort of systematic theoretical approach would be compatible with [perceptual] insights. This needs to be carefully considered, if we want to defend this conception as normative, especially for our public life.49

In this section I hope to show that moral perception depends on imagining that is guided by moral laws (Kantian or otherwise); and I hope to show that Kant’s understanding of how we come to know moral laws itself depends on a kind of perception. If this is right, then moral perception and Kantianism are not as diametrically opposed as is usually supposed. On the account of moral perception that was given above, Maggie held certain things constant (material resources, societal restrictions, mutual love) as she imagined various ways of proceeding forward from her current family situation. The fact that parting from her father was the only future she could imagine consistent with her constraining assumptions and values is what made parting appear to be a moral necessity. She did not first consult a moral law (“Act to enhance everyone’s autonomy” or “Put your husband before your father”) and then derive a conclusion about how to act in her particular situation. And yet, her imagining of alternate possibilities was clearly guided by her recognition of law-like relations between different aims and different situations (“Sustaining the father-daughter relationship is consistent with living apart,” “a marriage will not be happy in the absence of trust,” “It is easier to keep to one’s resolve if temptation is not present,” and so on.) Indeed, confidence about one’s imaginative projections is only possible 48 49

Blum (1994), for example, who is ardently anti-Kantian. Nussbaum (1990), p. 185 and p. 186, respectively.

218

MORAL PERCEPTION

insofar as one is confident about certain correlations and dependencies. Her current situation may be highly specific, but it can’t be so unique as to prevent reliable projections about the likely outcomes of different courses of action; and the need to part could not be experienced as a necessity unless the starting conditions were experienced as standing in certain law-like relations to various aspects of that outcome. The starting conditions do not determine all details of what must ensue (Maggie and her father could part in different ways, to different places, at different times), but they do determine some things (Maggie and her father must part); and it is precisely through the recognition that some things are so determined that we experience laws as laws.50 From what has been said so far, it may seem that our imagining must be framed by certain concerns or values (a concern for sustaining fatherdaughter relations, for example, or the valuing of a happy marriage), and then guided by what we take to be causal laws, or causal dependencies, without ever invoking anything that could be called a moral law—a law about what should happen (other things being equal) rather than what will happen (other things being equal). This description overlooks the way in which moral laws also guide our imagining, however. Maggie’s imagining is constrained not only by her love for both father and husband, and by her implicit knowledge of what causes what; it is also guided by her implicit knowledge of what should happen: husband and wife should grow closer to one another, a wife should deepen her understanding of her husband’s needs, an elderly father should be more focused on his own happiness than on his daughter’s, and so on. These may not be the sort of moral laws we are accustomed to hearing (and they are not the sort of laws that Kantians usually cite), but they are laws about what should happen rather than laws about what normally does happen, and they are precisely the sorts of rules that must guide us when we try to decide what to do. It is not enough to direct one’s imagining with a loving attitude or a desire to do right by others plus a good understanding of causal laws; one must also

50

The more general a law is, the more the conditions that get backgrounded into its ceteris paribus clause. Many writers have emphasized the existence of causal, or counterfactual, relations that do not rely on general laws. See Schiffer (1991), for example. Hitting glass can cause breakage without a law connecting hitting and breaking, for example. But insofar as counterfactual relations are modal relations, there will always be some laws in the background—concerning impact and crystalline structure, etc.

5.

BACK TO KANT

219

know a good deal about what doing right by others involves, and that is where moral laws enter. If one rejects the view that I have been advocating—namely, that perceiving anything depends on the imagining of relevant alternatives (since that is what gives one an experience of something as objective), then there will be no need to invoke laws to guide one’s imagining. But even if one accepts the need for such guidance, and even if one allows that in the case of moral perception the relevant laws must include moral laws, it is appropriate to ask how it is that we come to know such laws. I think Kant provides a compelling answer to this question, but a proper understanding of his answer requires some preliminary clarifications. First, it is important to distinguish between generalizations and laws. Consider the difference between generalizations and laws in the case of science. That crows are black, that trees are larger than bushes, that salt flats are formed when oceans recede, and that bees are attracted to flowers, are all generalizations rather than laws for there is nothing normative or necessary about any of these facts: the brownness of some crows does not need to be explained away as some sort of aberration (it is just unusual); and likewise with the fact that some bushes are taller than some trees, that oceans have sometimes receded without leaving salt flats, and that bees are not always attracted to flowers. That crows reproduce crows, that trees absorb carbon and release oxygen, that salt crystallizes in the absence of a liquid, and that flowers produce pollen are normative laws, however, where exceptions require some sort of special explanation—usually by indicating the absence of assumed background conditions (a failure of ceteris paribus). While generalizations serve a number of useful purposes, they have little to do with perception, scientific or otherwise. Knowing that most crows are black does not help me to see the blackness of this crow, and knowing that most trees are larger than bushes does not help me to see that this is a tree. Laws, on the other hand, refine our perception and enable us to recognize the particularity of what we see. Knowing that salt crystallizes in the absence of a liquid enables me to actually see the crystals forming, and it enables me to see why this particular deposit is forming in this particular way. So too, I suggest, knowing that most marriages are better if couples live away from their parents does not help me to see that this couple should live away from their parents. On the other hand, knowing that securing the bonds of marriage requires a certain break from one’s parents does enable me to see

220

MORAL PERCEPTION

what should be done in this particular situation. Kant’s emphasis is on moral laws, not moral generalizations; and moral laws are enabling rather than threatening when it comes to perception of moral particulars. Second, it is important to recognize certain parallels that exist between Kant’s account of what is necessary for the experience of spatiotemporal objects, in the Critique of Pure Reason, and Kant’s account of what is necessary for the experience of duty, in the Critique of Practical Reason. In both cases our experience of something as objective depends on our experience of that thing as governed by laws—causal laws in the case of spatiotemporal objects, moral laws in the case of duties. Also, in both cases it is the faculty of imagination that is responsible for bringing laws to bear on a particular state of affairs. Imagination, according to Kant, bridges the gap between understanding’s concepts (which are rules for organizing experience) and sensibility’s intuitions (which locate particulars in space and time). Just as our experience of physical objects depends on our recognition that they are parts of a world that persists independently of any particular observations we may make—a recognition that comes about through imagining and coordinating points of view that we do not actually occupy, our experience of moral obligations depends on our recognition that they belong to a system of duties that compel independently of any particular desires we may have—a recognition that comes about through imagining and coordinating points of view that we do not actually occupy.51 Merely bundling felt desires in an orderly way does not suffice to give us the experience of morality any more than merely bundling sensory impressions in an orderly way suffices to give us the experience of physical objects. What is additionally required for the experience of spatiotemporal particulars is a reliance on rules that determine the validity of our sensory experiences—that is, answerability to laws that require things to be one way rather than another. In the case of morality, likewise, there must be answerability to laws that require us to act in one way rather than another—rules that determine the validity of our intentions. In morality as in experience more generally, we discover objectivity by discovering rules that generate, and validate, not just our own experiences but other possible experiences within the same world. (This will also require us to discount some observations and 51

Kant clearly distinguishes desire (or inclination) from will. There is no way to give substance to the Categorical Imperative, however, without considering people’s happiness, and thus their desires.

5.

BACK TO KANT

221

intentions as mistaken because they diverge from the rules that serve to unify a wider range of observations or a wider range of actions within a single world.) The Categorical Imperative can then be understood as expressing the need for a single, integrated understanding of what is right even as the Ideas of Pure Reason express the need for a single, integrated understanding of what is true. By insisting on community at the broadest possible level, the Categorical Imperative regulates our moral experience so as to ensure its objectivity. The ideal of completeness and consistency is not reachable in the realm of morality any more than it is reachable in the realm of physics, but it must continue to function in its respective roles in order for us to experience the world and our duties in the world as objective. Kant’s answer to the question of how we come to know moral laws must ultimately rest on a story of how we come to know that the particular rules we are inclined to endorse—for example, help others in need, develop your talents, tell lies when beneficial to oneself, break promises when keeping them is inconvenient—would, or would not, result in a world that we could choose if those rules were to be followed by everyone. Once one accepts Kant’s claim that the very experience of duties as objectively valid requires that they be underwritten by relevant laws, and that just as the laws of a physical universe must be consistent with one another, so too the laws of a moral universe must be consistent with one another, one has effectively established the Categorical Imperative as a fundamental measure by which we come to know moral laws. So far, I have been addressing the way that particular moral judgments, whose truth is perceived rather than derived, nonetheless rely on moral laws to guide the requisite imagining of alternate possibilities, and the way that particular moral laws, in order to count as laws about objective duties, must rely on the general requirement of overall consistency—a requirement expressed by the Categorical Imperative. This is one way in which the proclaimed divide between perceptual theories and Kantian theories is bridged, for it gives moral laws an important role in the operations of moral perception. The bridge between moral perception and moral laws works in another direction as well, though, insofar as we can—perhaps even must—have perceptual knowledge of moral laws themselves. Kant is often viewed as insisting that we reason our way to moral wisdom by diligently checking our intentions against the broad rational constraint of

222

MORAL PERCEPTION

the Categorical Imperative, calculating out the implications of universalization, always on the alert for possible inconsistencies. This picture of Kantian ethics overlooks the extent to which Kant invokes the imagination in determining what is moral, and the extent to which imaginative syntheses are the very stuff of perception, according to Kant. Recall (from Chapter 2) that it is our ability to use our imaginations to synthesize a multiplicity of different perspectives and possibilities (past and future, actual and merely possible) that enables us to perceive an object or a sequence of events as objectively valid. Likewise, Kant’s insistence that we be able to simultaneously will a certain course of action not only for ourselves, and not only in this instance but across a multiplicity of relevant alternatives, points to a conception of moral knowledge that is deeply perceptual as well. The first version of Kant’s Categorical Imperative—act only on the maxim that you can simultaneously will to be a universal law—emphasizes the need for rules that can be projected beyond our own case to generate a unified and sustainable way of life. For in the absence of such projectability, we will not experience any rule as having an objective as opposed to merely subjective validity; and moral rules or reasons, as distinct from prudential rules or reasons, must be experienced as objectively valid. So, for example, a rule that directs us to break promises when convenient can only be an objectively valid moral law if we can consistently will the consequences of extending that rule to other people and other circumstances. If, as Kant supposes, willing the consequences of everyone breaking promises when convenient leads to inconsistency (since it would undermine the very institution of promising), then this is not an objective rule. Similarly, a rule that directs us to be kind to others only if we stand to gain thereby is not a rule that we can project successfully, since we cannot but will that others be kind to ourselves regardless of anything they gain thereby. The central role of the imagination in determining what is moral should come as no surprise, since we cannot really will that everyone do the same in similar circumstances unless we actively imagine what that would mean for the whole. To say that I am willing to have everyone break promises when convenient is a bit like saying that I can accept that every pair of numbers adds up to an even number: the acceptance lacks substance, it is merely verbal—a case of concepts that are empty for lack of the intuitions needed to give them content. We need to imagine non-actual applications of a

5.

BACK TO KANT

223

given rule in order to know when we can endorse such applications in a meaningful way. We may also need to imagine a number of different applications of a rule simultaneously, in concert, in order to check on the consistency of our endorsements and the viability of the world that we thereby will. (Kant’s first formulation of the Categorical Imperative specifies that the maxim that we follow in our own case must be one can we can “simultaneously” will others to follow as well.) It is easier to accept a rule that allows everyone to break promises when convenient if we imagine the consequences of each case in sequence rather than simultaneously, as a whole; the inconsistency that arises from universalizing this rule may not appear unless the consequences of all such cases are imagined all at once. (The need for simultaneous imagining of many different applications of a rule is also reflected in Kant’s third version of the categorical imperative which commands us to act in accordance with those rules that would bring about a Kingdom of Ends, for a Kingdom of Ends is, presumably, something that must be envisioned as a whole in order to give us guidance concerning what sorts of actions would or would not contribute to its actualization.) What I have said here should highlight an important affinity that exists between Kant’s account of what it takes to experience the objectivity of physical objects and what it takes to experience the objectivity of moral requirements. Both depend on our ability to see our experiences as constrained by reasons, and both are heavily dependent of the synthetic powers of the imagination. Although Kant insists on a sharp separation between the realm of physical objects and the realm of moral duties, the realm of cause and the realm of freedom, he is also insistent in refusing to equate this with a distinction between what is objectivity and what is subjective. Indeed, his central aim, in both the first and the second Critiques, is to show how objectivity is possible. Many questions remain about just how to square Kant’s “transcendental idealism” with his account of the objectivity of physical objects, and just how to square his account of the objectivity of moral laws with their origins in “self-governance.” It is certainly fair to say, though, that Kant is notable for his success in highlighting the parallels between what it takes to experience the objectivity of a house and what it takes to experience the objectivity of a moral law. Let me end with a remark about the third version of Kant’s Categorical Imperative, which requires us to act so as to bring about a “Kingdom of

224

MORAL PERCEPTION

Ends.” In this version, more than any other, Kant seems to recognize the limits on what can be synthesized and, thus, on what can be seen. Because we can’t really see all others as persons (versus know, in a non-perceptual way, that they are persons), we can’t really achieve the imaginative synthesis that morality requires. Stephen Mulhall, discussing Murdoch’s concern with perfection, seems to express Kant’s view as well when he writes: As finite creatures, we can never lose the sense that our moral perception is capable of further refinement, so the purification of our consciousness can never attain perfection; but neither can it shrug off its demands, and the progress we make is towards a deepening sense of the interrelatedness of the demands put upon us by moral values and dispositions—a sense of life and reality as hanging together, as making sense of a kind we have not yet begun to fathom except through a glass darkly, but to which we can always come closer than we presently are.52

This sense of the never-fully completed task of morality is one more reminder of the strong form of objectivity endorsed by Kant. It is not only the case that moral truths exist independently of any particular judgment that we make at a given time, they exist independently of even the fullest deliberation we and others are capable of across all time.

6. Conclusion I have tried to give an account of how moral perception is possible—an account that respects the three defining commitments of any view that makes perception central to moral knowledge, but goes further than that to explain how the phenomenology of moral perception is like the phenomenology of other sorts of perception. That has been done, first, by detailing what goes into seeing persons as persons and, second, by describing what goes into seeing what we ought to do. The guiding account of perception, in terms of imagining (itself indebted to Kant) is then shown to fit rather well with Kant’s moral views, which are generally criticized by defenders of moral perception. Nothing I have said here provides any guarantee that we will succeed in using our imaginations to discover what is morally right, nor that the moral 52

Mulhall (2000), p. 260.

6.

CONCLUSION

225

facts that we are able to perceive will provide us with knowledge that is adequate to the moral guidance we need. Like seeing how to carry an awkwardly shaped object through a twisted passageway, we may not be able to see our way forward and, even when we see the way, we may not succeed in following it.

Figure 6.1. Georgia O’Keeffe, Two Figs, 1923 (Oil on board, 7 3/4  5 3/4”, Bequest of Mrs. Arthur Schwab (Edna Bryner, class of 1907), 1967.31.12) The Frances Lehman Loeb Art Center, Vassar College, Poughkeepsie, New York. (# 2012 Georgia O’Keeffe Museum/The Artists Rights Society (ARS), New York)

Chapter 6 Aesthetic Perception The perception of artworks is different from the perception of ordinary objects in some important ways. The difference can be traced to the nature of art works as such (versus “mere” physical objects); or it can be traced to the nature of an “aesthetic” attitude (versus the more ordinary attitudes of our daily lives). Whereas, historically, it was usual to locate the difference in the nature of the art object (because that object embodied an ideal, because that object had special powers, or simply because it represented something beyond itself ), it has now become more common to locate the difference in the nature of the perceiving subject (especially with current interest in demystified and non-representational art).1 Insofar as the difference lies in the object rather than the subject, it is possible to have art that is not recognized as such. If the difference lies in the subject rather than the object, it is possible to transform ordinary objects into art objects merely by experiencing them as such. Fortunately, there is no need to insist on either extreme: genuine works of art need not be recognized as such, but the possibility of such recognition is a requirement for their standing as art. Likewise, merely experiencing something as art can transform ordinary objects into works of art, but there are real limits on the objects that can be so-experienced. For the perception of art as art typically requires us to look “through” surfaces in a way that the perception of ordinary objects does not; we must see beyond the paint to the depicted landscape or the expressed mood, we must hear the melody or the tension in the pitches, and we must perceive the character or the message projected by a particular actor. And not all canvases, sound sequences, and actors admit of such seeing through. 1

Kulvicki (2006) presents a thoughtful exception to this trend, focusing on the structural features of images rather than their reception in order to understand their status as images.

228

AESTHETIC PERCEPTION

In this chapter I develop an account of the double nature of objects of art and the double nature of aesthetic appreciation. These two sorts of doubling are worth considering separately but, as we will see, they are ultimately dependent on each other. The account of perception I have been developing—especially the account of moral perception that was offered in the previous chapter—can help to clarify both aspects of aesthetic perception. For perceiving the person depicted in a painting, or the melody expressed by the sounds, is analogous to perceiving the person in the body, or the worry expressed by the voice; and perceiving that a particular feature or action is aesthetically right relies on the same sort of constrained imagining that is needed in order to perceive what is morally right. The objects and the objectivity of aesthetic perception are no more (and no less) puzzling than the objects and the objectivity of moral perception. Furthermore, the double consciousness that is characteristic of aesthetic perception echoes the combination of empathy and distance that, I argued, ought to characterize our relations with other people. There are some important differences between moral perception and aesthetic perception, however; and though this chapter is primarily focused on the similarities, it ends with a consideration of some respects in which aesthetic perception is rightly distinguished from moral perception. I am not interested in determining the boundaries of what is (and what is not) “art,” or what does (and what does not) count as an “aesthetic attitude.” If there are valid works of art and/or valid ways of experiencing works of art that do not involve the doubleness I will be exploring, the relevance of this chapter will be restricted accordingly. If a splash of paint becomes a work of art merely by being dubbed as such, or if adopting an aesthetic attitude towards a splash of paint requires nothing more than appreciation of its formal properties, then there will be cases of aesthetic perception that bypass the sorts of doubling I will be considering. On the other hand, my understanding of aesthetic perception is not restricted to representational art or to the art that appears in museums. Holding the head of a broken doll on a dinner plate at home might produce the kind of object-doubling discussed in this chapter, and recollecting a painting of Judith holding Holofernes’ head might produce the relevant kind of attitudinal-doubling. As we shall see, there are many cases of layered identities, and many cases of seeing things “in” or “through” another that

1.

THE OBJECTS OF AESTHETIC PERCEPTION

229

do not depend on one thing representing another nor one thing appearing as another.2

1. The Objects of Aesthetic Perception In most cases, perceiving artworks as artworks requires us to perceive something more than the physically present object or event—something more than the colors and textures of a canvas, something more than the pitch and timbre of a series of sounds, something more than the movement of bodies on a stage. In the case of most paintings, the physically present object—a canvas with certain colors and textures and shapes—becomes the medium through which we perceive (or seem to perceive) some further objects—a tangle of leaves, for example, casting shadows on each other and receding into the distance. We see (or seem to see) the leaves through the painted canvas. In the case of music, a series of sounds with different pitches and volumes becomes a medium through which we perceive something else—a pair of melodies, for example, which weave between each other, soaring and sighing in turn. The physically present object and its properties are not merely a means for the perception of a further object with different properties, however. If we gazed at a depicted landscape without paying any attention to the painted surface as an object in its own right, we would not be perceiving the painting as a work of art. Likewise, while we might perceive the mood of a melody (as we usually perceive the meaning of a sentence), without attending to the rhythms and pitches of the words themselves, this would not suffice for perceiving that melody (or that sentence) as a work of art. Thus the metaphysics of these works of art— their double identity as surface object and revealed object—makes some important demands on the perception of art as art. 2 Most discussions of seeing-in are restricted to representational pictures. Wollheim (1968) and Hopkins (1998) argue that seeing-in is essential for the perception of pictures that depict something without really considering how it might extend to non-representational pictures. Lopes (1996) and Kulvicki (2006) maintain that seeing-in is appropriate to only some pictures, but their understanding of seeing-in is narrower than mine. Gombrich (1960) discounts the possibility of seeing both surface and depicted object at the same time, but surely the phenomenology of the experience allows for such simultaneity, and without it, it seems impossible to see that X represents Y as opposed to having merely discursive knowledge that X, which is apparently just paint, represents Y, or that Y, which is apparently a landscape, is merely represented.

230

AESTHETIC PERCEPTION

The view that sensual objects in space and time are merely veils or echoes through which more fundamental (and more real) objects can be perceived goes back to Plato, of course, and it underlies the view that attending to the sensual surface properties of things (something that art encourages) stands in the way of perceiving things as they really are. That is a view that philosophers such as Nietzsche and Foucault try to turn on its head by arguing, instead, for the primacy of sensual surfaces and the illusory nature of what lies “behind.”3 In the context of art, the very contrast between surface and depth has come under attack insofar as it suggests the need to uncover hidden objects and deep truths rather than attending to the objects that are immediately present and perceivable. Shusterman, for example, advocates interpretations that relate sensual properties to one another rather than using them to uncover hidden properties: “It is simply a mistake to think all interpretation is governed by the depth metaphor of uncovering hidden layers or kernels of meaning. Interpretation is also practiced and theorized in terms of formal structure, with the aim not so much of exposing hidden meanings but of connecting unconcealed features and surfaces so as to see and present the work as a well-related whole.”4 Likewise, with respect to music, Schenker’s “depth analyses” of music have been criticized for wrongly shifting our attention away from the richness and the interconnections of the sensual surface that is immediately present to our ears. Neither Shusterman nor the critics of Schenkerian analysis reject the value of interpretation, however (and if Nietzsche and Foucault are more radical in rejecting the value of interpretation, they are also more resigned in accepting its inevitability). Indeed, in his recent book entitled Surface and Depth, Shusterman articulates a notion of depth that is dependent on historical practices rather than on hidden essences. He endorses Dickie’s and Danto’s view that “[A]rt is defined in terms of factors that lie beneath the aesthetic surface: the background of artworld institutions, art history, and social practices in which objects are framed to make them artworks.”5 He also elaborates a definition of art as “dramatization” that, he thinks, “provides a way of reconciling the residual sense of conflict between these poles of aesthetic naturalism and sociohistorical contextualism . . . into “the 3 See Danto (1986) on a tradition of philosophical denigrations of art, and Shapiro (2000) for a summary of Nietzsche’s and Foucault’s inversions of this tradition. 4 Shusterman (2000), p. 233. 5 Danto (1986), p. 75.

1.

THE OBJECTS OF AESTHETIC PERCEPTION

231

two moments of experiential intensity and social frame” by putting a “scene” into the “frame” of a stage performance.6 Critiques of deep meanings in favor of surface meanings, thus, are not critiques of the double identity of art works so much as they are critiques of certain ways of finding meaning in and through the sensual presentation of a particular physical object (or in the case of music through a particular series of sounds). The distinction between surface and depth survive. It would be a mistake to discount either the surface identity or depth identity of a work of art, or to make any generalization about which is more important. The notion of depth, even the notion of hiddenness, need not imply the presence of metaphysically mysterious objects and properties. In the first place, the “hiddenness” of certain properties may simply indicate the difficulty we have in recognizing them—because they are complex relational properties (such as the Golden Section in painting, or a deferred resolution in music), or because they require special types of attention to discern (standing at just the right angle in front of a painting by El Greco, attending to two different meters in a performance of Raga Mawra). Secondly, there are many ways to allow for non-sensual objects without making them mysterious—as objects that are defined by their formal properties (numbers, for example), as psychological or social constructs (selves or races, for example), or as useful abstractions (the past, for example). The notions of depth and hiddenness can be explicated in a variety of ways, with different explications appropriate to different cases. (Whether all such notions are compatible with the possibility of perceiving such objects is another question, which I turn to below.) The “hidden” object in a work of art may be easier or harder to discern, and it may “reside” relatively close to the surface or more remote from the surface. The form of a waltz may be obvious upon hearing a series of sounds, and a red flower may be obvious upon seeing a painted canvas—both therefore being “close” to the surface in the sense of being automatically discerned. The underlying unity of a Boulez symphony, or the underlying unity of a Rothko painting, are more difficult to discern, but discovering their unity may depend on attending more carefully to the surface properties— the fluctuating rhythms and timbres of the notes, the different saturations of overlapping colors.

6

Shusterman (2002), p. 33.

232

AESTHETIC PERCEPTION

It is also possible, even usual, for there to be more than one “hidden” object since one can find unity or integration of sensual properties in more than one “place”—at different “depths” or in different “directions.” In discussing the paintings of El Greco, Updike describes the bodies that are portrayed (the rather obvious “hidden” objects of these paintings) but then alludes to a more abstract and supernatural object that is found in his paintings: “[T]hese wavery limbs and dwindled heads have less to do with the human body than an idea of bodies, whose basic reality lies behind or beyond their appearance” and “His supernatural bodies lift free of gravity, but at a cost: they seem insubstantial—too smooth, too rapt, too willowy, too elongated. They exist, but up there, in another world . . . ”7 An equally telling musical description can be found in Jeremy Eichler’s review of a Peter Serkin recital: “Each work feels X-ray scanned, its internal architecture laid bare . . . Emotion is not poured on, but coaxed out from the inner recesses of the score.”8 These are examples of experiencing objects “beyond” the objects that first appear through a work—the otherworldly bodies (or ideas of bodies) behind the human body, and the architectural bones beneath the intertwining melodies. In other cases, further objects appear alongside each other, as different connections and different associations are made. In the case of performance arts such as music, it is clear that different performances of the same notes will emphasize different sorts of connections; and, of course, different listeners, on different occasions, will be attuned to different sorts of unities. And writing about pictures, Kulvicki rightly notes that “As our world or merely our concepts change, we expose different aspects of the manifold possibilities for fleshing out pictures’ bare bones.”9 (In some cases, we are capable or seeing or hearing, multiple possibilities simultaneously; in other cases, attending to one such object precludes attending to the others.) To look more deeply into an object means finding a way to unify its various aspects more fully, whereas looking more widely into an object means finding a way to unify the same appearances differently, or effecting a unity of different aspects. What makes it possible to experience something as having a double identity, then, is the possibility of experiencing the surface properties of an object as different appearances or expressions of a still more 7 9

Updike (2003), p. 18. Kulvicki (2006), p. 175.

8

Eichler (2005).

1.

THE OBJECTS OF AESTHETIC PERCEPTION

233

Figure 6.2. Sanford Robinson Gifford, Tappan Zee, 1880 (Oil on canvas, 11 3/8  18 3/8”, Bequest of Robert Miller Walker and Alice Smith Walker, class of 1933, 1994.18.1). The Frances Lehman Loeb Art Center, Vassar College, Poughkeepsie, New York.

fundamental and more unified object. In the case of a painting, for example S. Gifford’s Tappan Zee (Figure 6.2), this might mean that the bright area here and the blur of color there, the rough texture in one area and the smooth sheen in another, all seem to come from the properties of an expanse of river bathed in afternoon light. In the case of sound, it may mean that the rising pitches, the slow crescendo, and the thickening timbre are all experienced as manifestations of the same building tension. Taking things a step deeper, the luminous calm of an expanse of water echoing and merging with the luminous calm of an expanse of sky may be perceived as the echoing and merging of the boundaries of our existence. And the rising tension of a musical passage together with its increasing weight may seem to express the apprehension of a tragedy. It is important to keep metaphysical questions about the nature of aesthetic doubling and aesthetic depths separate from evaluative questions about the worth of aesthetic doubling and depth. We shouldn’t assume that an increase in the number or depths of a work’s hidden objects increases its value. Nor should we assume that experiencing more or deeper unities has more aesthetic value. The appropriate degree of unity, and the appropriate sorts of unity, will depend on the details of the work and the details of the setting. As with persons, deeper persons are not necessarily better persons;

234

AESTHETIC PERCEPTION

looking for what is deep is not always appropriate, and finding what is deep is not always a good experience.10 I have said that abstract objects, or formal objects, or socially constructed objects need not be mysterious or metaphysically suspect. And I have argued, in Chapter 4, that some such objects can be perceived at least some of the time. But perception, properly so-called, requires not only an experience of objectivity but also the accuracy of such an experience. Thus, accounting for the possibility of aesthetic perception depends on accounting for the correctness of our experiences of the relevant hidden objects. If landscapes and melodies are simply objects that we imagine when looking at a painting or listening to a series of sounds, then however appropriate such imagining is, it would be wrong to claim that these objects are perceived.11 One way to ensure that the object of aesthetic perception is indeed objective would be to locate it in the mind of the creator or the observer (or some combination of the two). There is a long tradition, for example, of equating the meaning of an artwork with the artist’s intentions, and suggesting that it is the artist’s intentions that we perceive through his or her work. This suggestion is frequently criticized on the grounds that an artist’s intentions and an artist’s output are not always well aligned: the intention to paint a bed may result in a painting of a landscape, and the intention to compose a humorous piece of music may give rise to a wistful piece of music. The problem is not the unreliability of perceiving intentions through artworks, since reliability is never perfect and there will be many contexts in which the connection to artists’ intentions is quite reliable. Nor is the physical distance between artist and work necessarily a problem. In Chapter 3, I defended the possibility of perceiving distant states of affairs, so it is easy for me to accept the possibility of perceiving someone’s intentions, and perceiving them through their artistic work. The problem is, rather, that the criteria for determining the correctness of an aesthetic 10 I agree with Cohen (1998), p. 122, when he writes: “Explaining the coherence—the total sense— in a work of art is like explaining the coherent style of another person, and both are like explaining one’s own aesthetical self.” But I disagree with his following remark: “None of these explanations, in the end, is possible, and all must be attempted.” Rather: some explanations, some unities, and some experiences of depths must be possible in order for us to experience art as art, and to experience persons as persons; which explanations, unities, and depths to focus on depend on which works of art, or which people we are considering under what circumstances. 11 An insightful discussion of some relevant possibilities can be found in Kulvicki (2006), especially in the section entitled “Revelatory Realism,” pp. 226–30.

1.

THE OBJECTS OF AESTHETIC PERCEPTION

235

experience are not located in the artist’s intentions but, rather, in the work itself (and, perhaps, its socio-historical context). Determining whether a piece of music is humorous or wistful depends on properties of the music, not properties of the composer (or performer). If we learn that a piece we experience as wistful was meant to be humorous, we need not conclude that we are mistaken; we could instead conclude that the composer’s intentions were not fulfilled. These reflections might invite the view that the “hidden object” of aesthetic perception is located in the mental states of the viewer rather than the mental states of the creator of an artwork. On this view, it is my imagining of a landscape when viewing a particular painting, or the wistfulness that I feel when listening to a particular piece of music, that constitute the “deeper” object of aesthetic perception. Even though I am not explicitly focused on my states of mind, it is those states of mind that I perceive (through the paint, and through the sounds) when I perceive a work of art as art. The problem with this view is exactly analogous to the problem of the previous view, however. My reaction to a work is no more authoritative (indeed, is usually less authoritative) than the artist’s intentions. After all, I might imagine a bed when viewing a landscape painting, or I might feel wistful when hearing a piece of music that is actually humorous. To be plausible as an account of aesthetic perception, we would have to defer to appropriately attentive and knowledgeable viewers. But what establishes such viewers as appropriately attentive and knowledgeable is, precisely, their ability to pick up on the relevant aspects of a work, not aspects of their own mental states. Ultimately, then, support for any particular experience of a work must depend on the sensuous properties of that work, not the observer’s mental states.12 A more promising approach (which has been advanced by a number of authors) regards a work of art as a kind of person or generative force in its own right. The “hidden object” is in fact a subject in the sense that it is both the repository and the source of normative tendencies and dispositions in much the same way that a person is both a repository and a source of mental states. In order to contain hidden objects or meanings, a painting must be 12 One could counter this claim with a projectivist view, whereby the relevant features of objects are actually projections of one’s own (or one’s society’s) attitudes. Or one could opt for an analysis of aesthetic properties that involves some combination of artist’s intentions and viewer’s responses. I am assuming that a more objective standard is possible.

236

AESTHETIC PERCEPTION

somehow “alive,” as Wittgenstein says;13 its surface effects must be explicable in terms of an interlocking set of powers and dispositions (just as a person’s behavior is explicable in terms of an interlocking set of mental states), and those powers and dispositions must themselves have a selfgenerative character (just as a person has a capacity for self-creation). In the case of a landscape painting, for example, the particular arrangement of colors and lines will be explicable in terms of the powers and dispositions of the depicted landscape—the power to deflect light in particular ways, the disposition to be still in the late afternoon; or, more deeply, the surface features of a landscape painting may be explicable in terms of the powers and dispositions of a particular way of being in the world—a way that establishes more harmonious relations between things, a way that transcends the petty worries of everyday life. The explications in question are not causal explanations, since the depicted landscape and the depicted way of being may not exist (and even if they do, they play a very limited and indirect role in causing the painting). The explications are, rather, teleological, in the sense that they identify objects or states that serve to unify the disparate features of a painted canvas; a source of norms against which these features can be judged appropriate, or inappropriate. In a successful, or “live” work of art, the unifying object or the guiding norm are (at least partly) generated by the work itself. The distinctive way of being depicted by a fine landscape painting is not simply communicated by that painting, it is created by that painting—by its idiosyncratic mixes of color, by the way it conveys physical distance, and so on. Like a person who is (in part) the source of her own unity and her own norm, it is the painting itself (not the painter nor the actual landscape) that forges its own unity and its own norms. Works of art can be live and can be person-like without being human-like, however. The way of being conveyed by a landscape painting may not be a way of being that could be adopted by any human; indeed, part of the point of a painting, and part of its greatness, may be the presentation of a distinctively non-human sensibility. We can experience others as having lives of their own, with idiosyncratic styles and orientations, without the need to assimilate the lives of those others to our own; this holds 13 Wittgenstein (1953), p. 205: “I might say: a picture does not always live for me while I am seeing it. ‘Her picture smiled down on me from the wall.’ It need not always do so, whenever my glance lights on it.”

1.

THE OBJECTS OF AESTHETIC PERCEPTION

237

true for our relations with other humans, for our relations with non-human animals, but most especially for our relations with “live” works of art. I have maintained, throughout, that we cannot perceive states of affairs that do not obtain; we cannot perceive what is not true.14 Does this mean that the life or the generativity of artworks is only imagined or projected, not actually perceived?15 That there only appears to be a person-like object behind the surface of an artwork where no such thing exists in fact? This is certainly the easiest position to adopt, and it immediately establishes the separation between aesthetic perception and moral perception. (We don’t owe artworks the same regard that we owe living beings because artworks only appear to be living.) Alternatively, one could analyze an artwork’s “liveness” in terms of the features that it shares with genuinely living things—complex properties following a logical sequence, a formal unity, the occasional oddity, and so on—and maintain that it is these existing features that one perceives when one perceives a work’s life. But while such analyses may help to explain our experiences of a work’s life, they must still confront the question of whether the content of our experiences is accurate or not—whether the painting, or the music, or the poetry is sufficiently person-like for our experience to count as a perception. The live character of a work of art is often thought of as derivative from the live character of its creator and/or its audience. A composer or a singer expresses herself through a song, and the song thereby comes alive, or a listener hears her own hope and fear in the song and then experience it as person-like. The life of a creator and a listener is the original and undisputed life while the derivative life of a work of art may or may not be considered real in its own right. Once we begin to scrutinize the nature of live-ness in people, and the nature of personhood itself, the distinction between original 14 This was endorsed in Chapter 1, Section 1, with extensions into the future, for example, defended in Chapter 4. 15 There are interesting differences between so-called metaphorical objects and imaginary objects. In the case of music, a “line” will be considered a metaphorical object insofar as there are no literal lines in music; and in the case of drawing, a line will be considered imaginary insofar as we imagine it on the face rather than on the paper (where it actually is). Scruton (1999), p. 93, maintains that “the terms used to describe music refer to material sounds. But they refer to them under a description that no material sound can satisfy.” One could argue, however, that the terms used to describe music refer to the objects that one imagines when hearing the sounds. And one could argue that the depicted face is a metaphorical face since the lines that compose it are, literally, marks on a paper and faces cannot be composed of paper. I am not interested in defending one of these views rather than another; the important thing here is to realize the close relation between imaginary objects and metaphorical objects.

238

AESTHETIC PERCEPTION

and undisputed live-ness, on the one hand, and derivative and disputed liveness, on the other, tends to disappear. The self-generative aspects of people are no more and no less common or pervasive than the self-generative aspects of artworks, and they are no more nor less dependent on context— including the context of other people’s reactions. Works of art, like persons, integrate complex dispositions into functional unities, and the way that this integration occurs, for both people and artworks, relies on comparisons and interactions with both other works and other people. The unity of a work of art, like the unity of a person, is heavily dependent on its being treated as a unity; and the meaning of a melody is no more autonomous than the meaning of a mood. We distinguish the person from the body in much the same way as we distinguish the music from the sounds or the face from the drawing. We are reluctant to equate the boundaries of a person with the boundaries of any particular body, yet we are unable to individuate persons without referring to spatiotemporal positions. Works of art interest us and reflect back on our person just as much as persons interest us and reflect back on our works of art. There is no reason to think that the unity and autonomy of an artwork is any less real than the unity and autonomy of a person. Given the parallels, it is no surprise that persons have sometimes been regarded as imaginary constructs—unified in our imagination only, or (at least) unified in a way that is much less literal than the way that concrete physical objects are unified, but we need not, and should not, assume that constructed objects are not real objects.16 I have been glossing over the difference between hidden objects as what is represented (for example, a landscape) and hidden objects as what is expressed (for example, wistfulness). If we understand the hidden objects of art in the same way that we understand the hidden personhood of a body, however, we discover an interesting convergence between what a work represents and what it expresses. What is represented ultimately is not a scene but a sensibility, which is also what is expressed.17 16 According to Kant, all objects are constructed through syntheses of the imagination; but subjects/ persons are formal unities, not spatiotemporal objects. According to Nagel, we must assume the unity of persons/consciousness even though it is unintelligible. And within the social constructivist tradition, elements that are not independently unified come to function as unities for social purposes. 17 Items with descriptive or representational meaning are said to be truth-evaluable, for example, whereas items with expressive meaning are not; but this way of drawing the contrast simply shifts the

1.

THE OBJECTS OF AESTHETIC PERCEPTION

239

The account of expression offered here is both more and less restrictive than the well-known account offered by Nelson Goodman.18 Goodman equates expression with “metaphorical exemplification:” X expresses Y just in case X metaphorically exemplifies Y. As others have pointed out, this account allows any property whatsoever to be expressed—not just properties of subjects as such.19 A color, an angle, or even a specific length could be expressed, on Goodman’s account, as long as it could be exemplified metaphorically—that is, as long as the instantiation of some other property could function as a non-literal example of the property in question. If the heat of a blanket or the sound of a trumpet can stand as a metaphorical exemplification of the color red, then the blanket’s heat or the trumpet’s sound can express the color red. On the account I have outlined, it should be clear that the states or features expressed must be states or features of a particular sort of object; for it is only through its internal unity and generativity that an object becomes expressive. So the class of things that can be expressive on this account is significantly smaller than the class of things that can be expressed on Goodman’s account. On the other hand, my account of expressiveness does not require that the exemplification be metaphorical, and thus allows for more things to be expressed. The sweep of paint on canvas or the blare of a bugle can express boldness through literal and not merely metaphorical exemplification. Metaphorical exemplification may expand the number of things that an object can express, but there are many things that are expressed through literal exemplification.

problem onto debates over the nature of truth—especially truth in art. (Does the photograph represent the body as a desert? And, if so, is this a true representation?) More importantly, though, the alignment of truth with representation leaves expression without any positive characterization; expressive meaning enters to fill the void that remains after representational meaning has done its work. Approached this way, the positivists’ grouping of ethics, aesthetics, religion, and metaphysics alike as possessors of expressive rather than descriptive meaning makes the category of the expressive look like a mere grab bag for leftovers. One measure of just how elusive this distinction has become is the fact that deflationists like Davidson can suggest that all meaning is truth-evaluable while expressivists like Brandom maintain that all meaning is expressive. On my account, the difference between descriptive meaning and expressive meaning reflects the difference between imbuing an item with meaning by establishing a reliable connection between it and some observer-independent item, on the one hand, and imbuing an item with meaning by recognizing the agency behind it, on the other hand. 18

Goodman (1978). Roger Scruton (1999), for example. To sort the matter out we would need a fuller account of metaphor, however. If functioning as a metaphor itself depends on something like imagined animation, then Goodman’s account may be more restrictive than it seems. 19

240

AESTHETIC PERCEPTION

2. The Process of Aesthetic Perception We turn now from the question of what the hidden objects of art are, and what is expressed, to the question of how such things are perceived. Perceiving something as an artwork, like perceiving someone as a person, requires that the surface-depth contrast be perceived, not merely believed in. We can, of course, believe that something is a painting, that it depicts a landscape, and we can treat it accordingly without ever perceiving it as a painting of a landscape. Likewise, we can believe that someone is a person, not just a body, and we can treat her accordingly without ever really perceiving her as a person. But aesthetic appreciation, in contrast to merely intellectual appreciation,20 depends on perceiving what is depicted or expressed. It is not enough to know that a particular patch of blue represents water; the success of a picture qua picture requires that the lake be seen “in” or “through” the blue patch. Likewise the success of music qua music requires that certain notes be heard as a melody, a motif, or a soaring movement, for example.21 Knowing that something is a painting of a landscape and that it expresses a vibrant calm without seeing it as such, or knowing that a certain sequence of sounds expresses a kind of tragic apprehension without hearing it as such may be an intellectual accomplishment, but it is an aesthetic failure. Works of art, I have argued, are in certain respects like persons, and the perception of art as art is in certain respects like the perception of persons as persons. On the account of person perception developed in Chapter 5, the perception of a person’s inner state depends on our active imagining of relevant alternatives—alternative manifestations of the same underlying disposition. So too, I suggest the perception of an artwork’s “inner” state depends on our active imagining of relative alternatives—alternative manifestations of its underlying disposition. What these relevant alternatives are will depend on the artwork, of course. It will be useful to focus on a concrete example as we proceed. 20

There may be types of art for which merely intellectual appreciation is appropriate. Some instances of conceptual art may be like this (though many instances of so-called “conceptual art” actually require active imagining of multiple possibilities). By excluding merely intellectual appreciation from the definition of “aesthetic appreciation,” I am not committed to excluding the objects of such appreciation from the definition of art. 21 There is no correlation, however, between the ease with which we are able to perceive a landscape or a melody and the success of a painting or a piece of music. The struggle to discern what is depicted on a canvas, or just where a melody begins and ends, may or may not contribute to its aesthetic value. See the discussion in Section 3 below.

2.

THE PROCESS OF AESTHETIC PERCEPTION

241

Consider the Georgia O’Keeffe painting entitled Two Figs (Figure 6.1). One does not need to categorize the two central objects as figs or as fruit, as small or as large, in order to perceive their aesthetically relevant properties. But if one only sees interwoven patches of white, gray, purple, and green, then one fails to see most of what is aesthetically relevant in this painting. One must also see two bodies with certain textures and a certain density, two bodies that are resting on a further curved surface, and two bodies that are in contact with one another. Most people will see all of these things easily and automatically, both because the painting’s impact on the retina is sufficiently like the retinal impact of actual figs on a platter, and because the norms and conventions of representational paintings are sufficiently familiar to most people. Still, the application of those norms and conventions, though automatic, depend on active imagining of various sorts. To bring out the similarities with ordinary perceiving, it will be useful to distinguish two different sorts of imagining that are relevant to the perception of this painting. First, as I have argued in Chapters 1 and 2, to see the figs as three dimensional objects one must imagine their hidden sides—their backsides and undersides, perhaps even their insides—which means imagining them from perspectives other than the one presented by the painting. To see them as touching each other, the imagined perspectives must include some from which their contact can be observed. To perceive the figs’ texture and density one must imagine touching them in different ways—with different pressures, in different directions, with different parts of one’s hand. (Imagining what they look like from very close up might be another, less optimal alternative.) Here the analogy is between the depicted objects in the painting and the mental states of a person—beliefs, desires, fears, and so on, which can only be perceived as objective states insofar as we are able to imagine alternative manifestations of those same beliefs, desires, fears, and so on—ways in which those same states would appear if approached from a different angle (when one is not oneself upset, for example), if pressure were applied to them (and one met with relative sorts of resistance), if they were to encounter certain other objects (new threats or new temptations, for example).22 22 Note how it is difficult though not impossible to attend to the positions and textures and weights of the figs all at once, even as it is difficult though not impossible to attend to the beliefs and desires and fears of another person all at once.

242

AESTHETIC PERCEPTION

This analogy may seem flawed insofar as a person’s fear, unlike a painting’s fig, will manifest itself across many different contexts. The figs in O’Keeffe’s painting do not have any life outside that painting. We can come to see a person’s fear of failure by recollecting a variety of situations and by recognizing what remains constant in that person’s behavior across that range of situations, while O’Keeffe’s painting and its surroundings tend to remain the same. This contrast is not as great or as universal as it seems, however. Our experience of the figs in the painting is informed by our memories of other figs and other fig-like objects—and not only those that we have actually encountered but those that we have encountered in other paintings. Furthermore, our contemplation of the painting can occur in many different contexts—contexts in which it hangs alongside other still lifes versus contexts in which it is part of an exhibition of works by O’Keeffe, contexts in which we are thinking about color versus contexts in which we are thinking about form, contexts in which it is viewed from close up versus further back, and so on. Conversely, we can experience a person’s fear without having had any other encounters with that person, and although some background of experience with other animals seems necessary, this is quite analogous to the background of experiences of representational art that is needed to see the figs in the O’Keeffe painting. One important (if also flawed) way to acquire direct knowledge of another’s mental states is to imagine oneself as the one who produces the manifest behavior. When we see someone cringe before a blow, we tend to imagine the distress and apprehension that we suffer in order to act in that way, and we project those feelings onto the person we observe. In general, it may be harder to imagine oneself as the object in a painting (especially when that object is not a person) or the melody in a series of sounds.23 Identification with the objects of a work of art can sometimes enhance our perceptions of objects, however. Imagining myself as one of the O’Keeffe figs can enhance my perception of its shape in relation to its position, and imagining myself as the rising, thickening surround of sounds can enhance my perception of how these sounds fit together into a single musical gesture.

23 Scruton (1999) emphasizes imagined dancing as the means by which we are able to perceive (metaphorical) movement in a musical phrase; it is not always clear, however, whether he thinks we are imagining ourselves dancing to the music or imagining ourselves as dancing music.

2.

THE PROCESS OF AESTHETIC PERCEPTION

243

Not all imagining requires identification, however, even when it is a person who is being imagined; and not all identification is enlightening. To experience a picture as representational is not yet to experience it as a work of art; and to experience an animal as having mental states is not yet to experience it as a person. In Chapter 5, I defended the view that perceiving a person as a person requires one to perceive a being as having mental states and to perceive it as having a certain self-consciousness—that is, mental states directed at itself as well as mental states directed at the surrounding world. The analogous requirement in the case of art is this: the objects presented in a work of art—its inner objects, whether representational or not, and whether realities or not—must themselves be the objects of a further sensibility, and that sensibility must be evident from the way those objects are presented. In the case of O’Keeffe’s painting, then, perceiving it as a work of art not only requires us to perceive the qualities of the portrayed figs; it additionally requires us to perceive certain qualities of the portrayal itself. We must perceive what is constant across the way each fig is positioned, the curves of the fig, the curves in the background, the chalky surfaces, and blended edges (and so on for several other features of this painting); and we must perceive each of these features as expressions of the sensibility with which the figs are displayed. Here the imagining of relevant alternatives revolves around alternative ways in which the same objects could be presented. In some paintings, the coloring is more important than the contours, while in other paintings the contours are crucial; in some drawings an object’s texture is important, while in other drawings it is the orientation of objects vis-a`-vis each other that matters most. We don’t know which features are the important ones, however, unless and until we imagine variations that do, and variations that do not, express the same sensibility. If we imagine different surroundings for the same figs—a plate full of fruit, or a garden, for example—and we find that the underlying sensibility changes, then we know that the figs’ surroundings are aesthetically significant to that work of art. Looking beyond the evidence of any particular work of art, there is the larger oeuvre of a given artist and, indeed, the entire inventory of a school or movement that can serve not only as valuable contextual background but as manifestations of the same sensibility or the same self-consciousness. Additionally, especially in the case of performing arts such as music, dance, and theater, there is an evolving series of events that allow the

244

AESTHETIC PERCEPTION

work to manifest itself in a number of different ways across time, and, more importantly, across a number of different performances in a number of different contexts. This is analogous to the way in which a person’s family or community provide more than a background of similarities and differences against which to understand that person; they also provide further insight into that person’s reflective understanding of herself. Imagining oneself as a creator of a work, or the creator of a series of works, can be useful in one’s attempt to perceive the sensibility that is expressed by a work. What would it be like to highlight these features rather than those, to use a brush in this way rather than that? Wollheim has emphasized the importance of imaginatively replicating the action of a painter in order to appreciate what the painter was expressing. But, like trying to discover another person’s mental state by imagining oneself as behaving as that person behaves, some are better at this than others, and even under the best of circumstances such imagining can be seriously misleading. A final point of analogy between persons and artworks, and the perception of artworks, concerns their essential indeterminacy—not just uncertainty due to the limitations on our knowledge, but indeterminacy due to the open-ended character of the thing in question. As is the case with persons and their various aspects, there are always different but equally valid ways to synthesize various aspects of a work of art into a unified whole. Recognizing this fact is part of what it means to recognize a person as a person or a work of art as a work of art. Some works, like some people, are more indeterminate than others, and the attunement to the world’s indeterminacy can itself be what is expressed by a work of art.24 There is a difference, though, between assemblages that so lack unity as to fail at being a person or fail at being a work of art, and assemblages that sustain unities of many different sorts. Kant was keenly aware of the similarities between moral perception and aesthetic perception. In both domains he insisted on the role of imaginative syntheses in effecting the sorts of unities required for the experience of objectivity—the objectivity of persons and of duties in the case of morality, the objectivity of genius and of beauty in the case of aesthetics.25 Aesthetic 24

See Gorky’s The Horns of the Landscape, reproduced on this book’s cover. Although there are good reasons to question Kant’s emphasis on “genius” (with its focus on individual originality), and his emphasis on beauty (with its focus on formal properties and aesthetic pleasure), Kant’s understanding of these notions is not as simplistic, or as narrow-minded as usually supposed. I am not so much interested in his valorization of genius or of beauty, as I am interested in his syntheses of imagination that underwrites aesthetic perception. 25

2.

THE PROCESS OF AESTHETIC PERCEPTION

245

rules, like moral rules (and unlike the rules of physics), are not discovered so much as created, but they are experienced as objective insofar as they can effect the relevant syntheses. According to Kant, each feature of a successful work of art is bound to other features of that work in a rule-like way that reveals the work to be a manifestation or expression of an overarching conception—much as each intention of a virtuous person coheres with other intentions of that person in a rule-like way that reveals that intention to be a manifestation or expression of an overall conception as required by the Categorical Imperative. The light coming through the window of a Vermeer painting seems to require a certain look on the face of the seated woman, the look in turn seems to require a certain gesture from the attendant nearby, that gesture seems to demand a certain open-endedness in the scene portrayed, and that open-endedness seems to demand the distinctive light. Or to take a non-representational example, the urgent opening of a Haydn quartet seems to require a forceful reply, the completion of this exchange seems to dictate a shift to something more reflective and extended, the intensification of those reflections seems to force a return to the opening phrase, and this series as a whole comes to seem like the only possible introduction to the following movement. These are not imperatives that could have been stipulated in advance (though they may be rules that can be followed so as to generate another, equally successful work of art). Each artistic genius reveals a new “world”—of interlocking rules—revealing a previously unrecognized whole through the instantiation of new ways of binding different features together.26 Art, like morality, tries to find and to follow rules that ensure unity across an otherwise chaotic and conflicting set of possibilities. Kant maintains that, in order to experience the unity of a painting or a piece of music, the imagination needs to enter into a kind of “free play” in which shared rules of synthesis are no longer binding and new rules of synthesis are sought, in which different types of syntheses are entertained 26 This suggestion can be usefully contrasted with two other accounts of how Kant thinks concepts relate to intuitions in aesthetic judgments: (1) Sarah Gibbons’ (1995) claim (echoing Kant) that aesthetic judgments consist in the application of an indeterminate concept to intuitions; (2) Paul Guyer’s (1997) claim that aesthetic judgment involves “a sense of the unity of our experience of an object that goes beyond or surpasses whatever unity is entailed by the concept or concepts that we apply to this object.” In contrast to Gibbons, I claim that aesthetic judgments involve an indeterminate number of concepts, not an indeterminate concept; and in contrast to Guyer, I emphasize the free play of concepts in imagination (and thus the lack of governance by any one concept) rather than a sense of unity that has nothing to do with concepts.

246

AESTHETIC PERCEPTION

simultaneously. Gorky’s title “The Horns of the Landscape” (see cover), invites one to see horns of some sort and to see a landscape of some sort; but the syntheses that generate the experience of horns are (at least initially) different from the syntheses that generate the experience of a landscape. Are the horns attached to animals, however strange, whose parts are entangled below? Are the “horns” the outcropping of a rock formation, of trees, or an artist’s studio perhaps? Clearly, this painting is meant to prompt many different attempts at synthesis, to prevent the neat completion of any of them, yet to make the continuing attempt compelling. The very character of the painting—its indeterminate shape, its surface scribbles, its smearing of colors—tells us that there is no one right way to parse its “objects”—indeed, that there will always be something that escapes whatever unities we can find. The status of this painting as a work of art depends, though, on its ability to continually draw us into the project of synthesis, moving past the plurality of specific objects syntheses towards their source in an underlying sensibility, however indeterminate this sensibility may be. The task of the aesthetic imagination—to discover new forms of synthesis and to keep multiple forms of synthesis in play, recognizing the necessary incompleteness of any overarching synthesis—is closely aligned with the role of the imagination in morality. For morality not only requires us to imagine perspectives that are different from our own and free-standing in the sense that they are self-constituting (they are the “worlds” of autonomous subjects, self-ruled rather than subservient to the rules of others); morality also requires us to recognize the essential independence of the whole that is a person. The necessary incompleteness of the synthesis that we attempt when perceiving a person as a person is matched by the necessary incompleteness of the synthesis that we attempt when perceiving art as art. In all of these respects, treating works of art as works of art is a lot like treating persons as persons: it requires us to recognize the unique unity of each individual, and it impels us to seek a unity that will always exceed our understanding.

3. The Double Consciousness of Aesthetic Perception We have been considering how it is possible to perceive the inner objects of an artwork. But as we saw in Section 1, artworks have double identities, as

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

247

surface objects and as disclosed objects—the physical object that confronts us in space-time (the painted canvas on the wall, the sound waves in the air) and the revealed object (the tranquil landscape, the tragic melody. So when we perceive a work of art as a work of art we must perceive both of these identities together. In this section I offer further explication of the doubleperception or double-consciousness that is required—highlighting some similarities but also some differences between the phenomenon of “aesthetic distance” and the phenomenon of “moral distance.” The following and final section then reflects more broadly on the differences between our relations to artwork and our relations to people. It is common to speak interchangeably of seeing a landscape in a painting and seeing a painting as a landscape, of hearing a series of sounds as a melody and hearing the melody in a series of sounds. Seeing-in and seeing-as are not quite the same thing, however, and neither suffices for aesthetic perception. For we can see a bush as a bear, or an ax as a doorstop without seeing the bush as an expression of or a revelation of bearhood, without experiencing the ax as surface object and the doorstop as hidden object. Rather, in these cases, we are confronted by an object whose features are organized or utilized in one way, resulting in an object of a particular type, yet we (wittingly or unwittingly) experience them as organized some other way, resulting in some other object. We may be aware that we are replacing one organization with another, but not in such a way that we experience the bear as the object that is disclosed through the bush’s surface; and the doorstop is not the deeper meaning of the ax. Equally, not all cases of seeing-in are cases of aesthetic perception, for we can see the body in the clothes or the fish beneath in mud in virtue of the way the one causes the other to move without seeing the clothes or the mud as expressing what lies within, and without thereby engaging in aesthetic perception. While the perception of a bush and the perception of a bear are too independent (typically requiring us to flick back and forth between the two experiences), the perception of mud and the perception of a fish are too close (typically excluding the possibility of the mud revealing the fish in any way other than its shape).27

27 Wollheim (1987), p. 46, writes that “ ‘seeing-in’ is a ‘distinct kind of perception’, a biologically grounded . . . dual-aspect yet unitary experience, including seeing of the marked surface and the depicted subject as two aspects of a single experience, not two simultaneous experiences or two alternating experiences.”

248

AESTHETIC PERCEPTION

To understand the double consciousness that is characteristic of aesthetic perception we need to understand the peculiarities of what is generally known as “aesthetic distance.” Aesthetic distance, I shall argue, is responsible for the required experience of space—the experience of an opening between the surface identity and the deeper identity of a work of art, an opening that enables the one to function as an expression or revelation of the other. M. G. Benton describes aesthetic distance as the distance that enables a spectator to explore different points of view on a surface object, an exploration that in turn enables one to experience something as existing behind the surface as well. The interplay between the spectator’s sense of the represented subject and the medium in which it is cast is best conveyed by reconsidering the idea of aesthetic distance and, in particular, by invoking the concept of a shifting viewpoint. This concept facilitates not only such “looking around” but also reflects the spectator’s awareness of the continuously changing, unstable relationship between surface and scene.28

One may need to stand at some distance from a canvas, or from an orchestra, in order to perceive what is represented or expressed. It is not our distance from the canvas or an orchestra that is at issue here (though changing one’s distance can enhance aesthetic perception, enabling one to occupy more optimal points of view).29 The relevant distance is, rather, the distance that 28 Benton (1995), p. 372. The passage continues: “It acknowledges the inevitable mobility of the process and allows for greater flexibility and variety in the ways spectators operate than does the ‘viewing template’ that Wollheim’s spectator theory would place over the experience.” Benton is contrasting his view, which is indebted to Gombrich, with what he views as the overly static view of Wollheim and Hopkins. Whereas Hopkins (1998), p. 196, claims that the difference between visualizing and seeing-in is that in the case of seeing-in the surface provides a spatial perspective on the seen object, Gombrich (1960), p. 169, maintains that “ . . . the beholder must mobilize his memory of the visible world and project it into the mosaic of strokes and dabs on the canvas before him. It is here, therefore, that the principle of guided projection reaches its climax. The image, it might be said, has no firm anchorage left on the canvas . . . It is only ‘conjured up’ in our minds. The willing beholder responds to the artist’s suggestion because he enjoys the transformation that occurs in front of his eyes . . . The artist gives the beholder increasingly ‘more to do’, he draws him into the magic circle of creation and allows him to experience something of the thrill of ‘making’ . . . ” 29 It is interesting to note how imagining physical distance can contribute to this sort of psychological distance, and vice versa—how greater psychological distance can lead to the imagining of greater physical distance. Imagining oneself at a distance from a disturbing scene makes us less inclined to respond to it; helplessness in the face of danger often causes victims to imagine themselves in a location far removed from the situation they are in. Aesthetic contemplation can also create a divide between the imagined locations of two different selves: there is the self that that we imagine as immersed in the world of the painting, the world of the

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

249

separates us from the inner object of an artwork—the depicted landscape or the expressed sadness, for example. It is the distance that enables us to see the bear in the painting without wanting to run from it (as one would want to run from a bear in the woods) and to hear the sadness in the sounds without needing to share in that sadness (as one would if it were the sadness in a friend’s sigh). A standard way to understand this sort of distance renders it as a psychological distance rather than a physical distance—in particular as a dissociation between one’s perception (or perception-like experience) of some state of affairs and one’s inclination to respond to that state of affairs.30 We perceive a pair of figs without being inclined to eat them, we perceive an assault without being inclined to run or to intervene, and so on. We are “distant” to the extent that what we observe is isolated from what we do, and we are “close” to the extent that we are inclined to act in response to what we perceive. Aesthetic distance thus enables us to explore different attitudes towards the object or event without feeling any need to act on those attitudes. There are many challenges that can be raised against this standard account: Can one perceive something without having any inclination in response to that thing? What does it mean to explore attitudes if not also to explore inclinations? And isn’t art (at least sometimes) supposed to alter our inclinations? One response to these challenges depends on establishing a middle ground—an appropriate psychological stance that is neither too distant nor too close, with incipient inclinations that are effectively blocked or mediated in some way. We don’t want to be so close as to forget the fictional status of a landscape or a story, nor do we want to be so distant as to be unmoved. Edward Bullough, in his classic paper entitled “Psychical Distance,” maintains that the more we can decrease our distance from a work of art without losing it entirely, the greater our appreciation will be. Describing a jealous husband’s response to a performance of Othello, he states: The jealous spectator of Othello will indeed appreciate and enter into the play the more keenly, the greater the resemblance with his own experience—provided that

music, or the world of the sculpture; and there is the self who “stands back” from this world, receding into the background or retreating to another world entirely. Intense aesthetic experiences are often followed by a rather disconcerting sense that two selves that have gone separate ways must now unite. 30

For an illuminating account of how an interest in the perception of actual distance was transformed into an interest in emotional and imaginative distance in the eighteenth century, see Ogden (1971).

250

AESTHETIC PERCEPTION

he succeeds in keeping the Distance between the action of the play and his personal feelings.31

And illustrating the contrasting failures of too much distance versus too little distance in the case of music, he writes: Certain kinds of music, especially “pure” music, or “classical” or “heavy” music, appear for many people over-distanced; light, “catchy” tunes, on the contrary, easily reach that degree of decreasing Distance below which they cease to be Art and become a pure amusement . . . the undoubted physiological and muscular stimulus of its melodies and harmonies, no less than its rhythmic aspects, would seem to account for the occasional disappearance of Distance.32

The picture implicit in these remarks seems to be this: there is a whole series of internal responses that can follow from perception, only the final one of which is overt action, and when it comes to aesthetic perception, the closer one can get to that final gate without crossing through it, the better. Before we evaluate the above account of aesthetic distance, it is worth noting its similarity to some standard accounts of moral distance. To appreciate another person’s point of view we should empathize as fully as possible—without, however, losing sight of the fact that it is they, not we, who must live their lives. Nancy Sherman canvasses a range of views concerning the placement of the relevant location between admirable empathizing and inappropriate over-empathizing.33 Whether such imagining is understood as a kind of “off-line” simulation, a useful sort of pretending, or a primitive ability to absorb the feelings of others, moral empathy is supposed to depend on sharing in another’s experiences only to the degree that such sharing does not obliterate the distinction between one’s own perspective and that of another. In the words of Adrian Piper, there needs to be a “balance between preserving the unity and rational integrity of the self against external violation, on the one hand, and maintaining a self-enhancing connection and receptivity to external input, on the other.”34 Without this balance, our imaginative involvement “fails to recognize and respect the 31

Bullough (1912/1963), p. 239. Bullough (1912/1963), p. 244. 33 Sherman (1998), pp. 82–119, provides a useful list of reasons for thinking that morality depends on a capacity for empathy: it motivates altruism, it increases our self-understanding, it makes common discourse possible, and so on. 34 Piper (1991), p. 734. 32

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

251

ontological boundaries either of the self or of the imaginative object,” resulting in “vicarious possession” at one extreme (cases where we take another’s perspective to be our own) and “self-absorption” at the other (cases where we project our own preoccupations onto others).35 Thinking of distance versus immersion, whether aesthetic or moral, as opposite poles of a single continuum overlooks the possibility that certain sorts of distance actually enhance certain sorts of engagement, however. Sometimes it makes sense to recommend both more distance and more engagement. Oswald Hanfling, commenting on Bullough’s article, distinguishes between two sorts of distance—one sort that “cuts out the practical side of things,” and another sort that keeps our feelings separate from the feelings of a fictional character (or from the feelings expressed by a work more generally).36 He then interprets Bullough’s recommendation as a recommendation to minimize the distance between one’s own feelings and the feelings of fictitious characters (or the feelings expressed by a work of art more generally), while preserving a certain practical distance between oneself and those characters (or the work itself )—to seek maximum emotional involvement consistent with an absence of any practical involvement. Hanfling further suggests that there is an inverse relation between these two sorts of distance, since creating distance on the practical side enables one to draw closer to the feelings of fictitious characters, and being drawn into the feelings of fictitious characters helps us to distance ourselves from our practical interests in those characters (or in the work itself ). I think Hanfling is right to note that there is—often, anyway—an inverse relation between the degree to which we are practically involved with something and the degree to which we are emotionally involved. It is frequently the case that our hearts go out to someone when our hands cannot, and that our anger lessens when we are able to translate it into action.37

35

Piper (1991), pp. 735–6. Hanfling (2000) goes on to describe three other sorts of distance as well. He treats Bullough’s discussion of “pure” versus “catchy” music, cited above, as a description of too much versus too little distance in the sense of accessibility: “pure” music may not be intelligible enough to engage us while catchy tunes may be too transparent to challenge us. (This further sort of distance is discussed at greater length in his (2003).) But that reading does not account for Bullough’s suggestion that it is one’s bodily involvement with catchy tunes that gets in the way of distance. It seems more natural to read Bullough as objecting to the fact that “pure” music may not engage our emotions at all, while “catchy” music may draw us into practical activity such as dancing(!). 37 I elaborate my view of the emotions in Church (1995). 36

252

AESTHETIC PERCEPTION

Indeed, the view that art intensifies experiences precisely because it effectively insulates those experiences from any practical consequences is a familiar one. Kant’s original account of aesthetic “disinterestedness”—understood as the setting aside of our practical interests when contemplating an object—is a natural counterpart to his account of how the imagination is then able to enter a state of free play, a play that can reveal the harmony that exists between our own capacities and the contemplated object.38 Several more recent accounts also emphasize an important trade-off between a lack of practical engagement and a richness of emotional (or emotion-like) engagement. Kendall Walton, for example, maintains that the merely “pretend” character of our involvement with fictional characters and fictional events enables them to be experienced more intensely.39 And Gregory Currie describes our engagement with a work of art as a kind of “off line” simulation, whereby a range of possible responses can be more fully explored knowing that they are disconnected from their normal practical implications and risks.40 Again, there are interesting parallels to the case of morality. By bracketing off our practical interests in others, we are usually better able to empathize with their experiences; and by imagining our own experiences as quite removed from the experiences of others, we become better able to imagine those other experiences “from the inside.”41 In ethics no less than in aesthetics, finding the right distance is not a matter of finding a proper resting place between the extremes of distance and immersion so much as it is a matter of negotiating a (more delicate and destabilizing) combination of one sort of distance and another sort of immersion. The awkward yet entrancing character of this combination is effectively illustrated in Anna Deavere Smith’s performance of “Twilight: Los Angeles, 1992,” in which 38 Kant (1790/1951), especially the Analytic of the Beautiful. There are different ways of understanding what this free play involves. (Kant wrote of the free play of the understanding, which employs concepts, but he also wrote of imagining that does not depend on concepts.) It is clear, though, that the imagination is freed by dissociation from practical concerns. 39 Walton (1990). 40 Currie and Ravenscroft (2002). 41 Much has been written about the difference between imagining the position of another and imagining oneself in another’s position: Williams (1973), Wollheim (1984), Velleman (2001). Merely imagining the position of another does not seem sufficient for an “inside,” first-person experience of that position; yet imagining oneself in another’s position seems to be more a matter of self-projection than empathy. It is not the words we use that matter here, but rather the possibility of entering into a kind of double-consciousness whereby I imagine your experience “from the inside” while retaining my “outside” identity. The difficulty of locating the self in such a state stems from the fact that the self becomes a kind of split self under these conditions.

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

253

she enacts the responses of a wide range of people to the 1992 riots that followed the Rodney King verdict. As Anne Anlin Cheng points out: Smith . . . conveys the “real” presences of the characters, and yet at the same time, the audience is also made more aware of her, of the editorialization and craft that mediate such achievement. That is, even as one gets drawn into the pressing, authentic urgencies of these individuated voices, one is increasingly conscious of her mediation: the selection of monologues, their juxtaposition the resulting echoes and contrasts, the choices of beginnings and endings. The virtuosity of her very bodily performance relies precisely on the embodiment of the other through a self-erasure that nonetheless insists on the materiality of the medium of the body. . . . Anyone who has witnessed Smith’s performances understands the discomfort of being made to watch the fine lines separating speaking for, speaking as, and speaking against. Critics have read her complicity with various opposing characters as constituting a kind of community, but her complicity with everyone also marks her distance from them.42

Smith’s performance is remarkable for the way that it merges a certain moral project with a certain aesthetic project—the project of deepening our immersion in another world while increasing our sense of estrangement from it.43 Indeed, the distance we feel from the characters she embodies and from Smith herself is a result of the immersion she effects, and the effectiveness of our immersion depends on the distance she creates. These reflections have moved us away from a purely psychological account of aesthetic distance. It is not we who determine the appropriate distance between ourselves and the landscape in a painting, or the sadness in a piece of music, any more than it is we who determine the appropriate distance between ourselves and other human beings. The distance that separates me from an observer of events in Los Angeles in 1992 derives from actual physical distance—not only between New York and Los Angeles, or between 1992 and 2012, but also between a whole host of historical events that infuses that other observer’s perception versus those that infuse my own. Likewise, the distance between the landscape depicted in Gifford’s Tappan Zee and the landscape outside my window is created in part by the 42

Cheng (2000), pp. 188–9. Byatt (2001), “Ice, Snow, Glass,” p. 156: “I think artists recognize the distancing of glass and ice as an ambivalent matter, both chilling and life-giving, saving as well as threatening.” 43

254

AESTHETIC PERCEPTION

miles that stretch between my hometown Poughkeepsie and Tappan Zee, and by the years that stretch between 1880 and 2012. It is also created in part by the historical events that shaped Gifford’s sensibility versus mine. (Depictions of imaginary landscapes, and non-representational art, will still be expressive of states of mind and sensibilities that exist at a distance from my own.) An accurate perception of the “hidden object” of a work of art thus depends on an accurate appreciation of the very real distance that separates me from it. And, I suggest, it is awareness of that real distance rather than the willed adoption of a certain sort of psychological distance that determines an appropriate practical disengagement from the objects depicted in art. H. Gene Blocker points out that our practical disengagement from a work of art would cease if we were to view it as merely surface (what he calls “totally opaque”) or merely depicted objects (what he calls “totally transparent”); in either case we would cease to see the work as a representation of reality, and it is art’s representational character that makes practical engagement with it inappropriate.44 (If the surface of the painting were totally opaque, we would not hesitate to interact with it as just one more physical object among many. And if the surface of a painting were totally transparent, it would produce the illusion that what is represented is actually present, and we would act accordingly—without any “aesthetic distance.”) I would add: if we see a painted surface merely as a representation of some non-existent thing, and not as a manifestation of something real, then we are not actually seeing-in—perceiving something through the surface, and it is not really functioning as a work of art after all. Blocker goes on to claim that it is the combination of opacity and transparency, of surface and something beyond, that ensures our imaginative immersion since it requires us to engage in an active process of interpretation—a self-conscious attempt to access a hidden object rather than a passive observation of what is immediately apparent. Following Heidegger, Blocker even suggests that the real “object” of aesthetic perception is our interpretive activity—in which case our imaginative engagement is inextricable from the object with which we are engaged. [A]rtistic truth is not what is revealed but the revealing of it. Since neither total transparency nor complete opacity can achieve the latter, both must be rejected as

44

Blocker (1977).

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

255

artistic goals. Total transparency hides the act of revealing by simply showing the object. Total opaqueness fails to reveal an object. Revealing, we may say, takes two— an object and a recognized sense of discovering that object through that revelation.45

Blocker’s explanation of the necessary connection between aesthetic opacity and transparency, on the one hand, and aesthetic distancing and immersion, on the other, is illuminating. The difficulty that we have in bridging the gap between surface and depth is what draws us into an attempt to understand their connection and, simultaneously, causes us to disengage from both surface and hidden object as such. The greater the gap, the more that remains to be discovered, and the more intent one is on discovery the less inclined one is towards action. But this, I have suggested, has less to do with the representational character of an artwork than it has to do with the elusiveness of its object. It is not a particular tree or a particular hill that is hidden in a landscape painting so much as it is a way of being that is expressed by that tree or that hill. The relevant “beyond” of a work of art is not the entity depicted—the Hudson river that Gifford observed, for example, or the two figs that O’Keeffe had before her as she painted, so much as it is the particular sensibility or style or manner of synthesis that is embodied or expressed by the way these things are depicted. To the extent that these sensibilities elude one, there will be serious limits on one’s capacity for appropriate response. In this respect, also, there is an important parallel between seeing art as art and seeing persons as persons. To recognize a person as a person includes the recognition that she is not fully knowable—that any given interpretation we may offer is under-determined by our evidence and, more importantly, that people are, by their very nature, forever incomplete. Likewise with artworks: they always admit of multiple interpretations, and their very identity is forever incomplete. It is this inherent unknowability and necessary incompleteness that warrant the ambivalence discussed in Chapter 5, and that adds to the aesthetic distance discussed above.46 45 Blocker (1977), p. 228. Some of Adorno’s writing points in a similar direction. He maintains that “art expresses something while at the same time hiding it” and elaborates by referring to particular attitudes of abandonment (“the ability to abandon oneself to a particular work—is part of its enigmatic quality”) and particular activities of interpretation (“Truth cannot be identified directly . . . [It is] contained in [a work’s] elaboration and inner consistency”). Adorno (1970/1984); pp. 176, 182, and 188, respectively. 46 Wollheim (1974) makes some nice observations about the psychological importance of art’s ambiguities: “It is not simply that our understanding need not be explicit but that in many cases there

256

AESTHETIC PERCEPTION

Earlier, we noted that greater practical disengagement makes room for greater imaginative engagement, and vice versa. And we have argued that a divide between practical disengagement and imaginative engagement is the subjective counterpart of a divide between the surface of a work or a person and what lies beyond. If this is right, then the farther a work recedes behind its surface, the greater should be both our distance and our immersion. Put another way, intensifying the dual aspect of an aesthetic object should also intensify the dual character of an aesthetic attitude, and vice versa. Why wouldn’t our imaginative connection to a person or to a work of art be greater when the gap between surface and underlying sensibility is less—when a person wears her heart on her sleeve rather than hides it, when a musical passage expresses enthusiasm directly rather than indirectly, and so on? Part of the answer, I think, lies with the fact that familiar surfaces tend to elicit rather automatic responses, preventing one from seeing through surfaces to something live and fresh within. We may be quicker at seeing what a person is up to, but slower to recognize her autonomy and to regard her morally. And we may see immediately the emotion that a work expresses without engaging in that emotion ourselves. Likewise, viewing oneself in a mirror is not nearly as captivating as viewing oneself in an old photograph, or in a store window tilted at an odd angle, or in a hall of distorting mirrors— not just because of the novelty of these other reflections but, more importantly, because they demand that we project ourselves past appearances to the less obvious intentionality that animates them. Still, one might suppose, we will feel closer to the world of a Vermeer painting than to the world of a Bosch painting. This may be so; nothing in what we have said suggests that greater distance always creates better perception. Not all works, and not all humans, have a coherent sensibility behind them, and our imaginations

are dangers in explicitness, for explicitness could give rise to resistance” (p. 214); “part of understanding how it is that a work of art affects us is recognizing the confusion or the ambiguity upon which this effect in part depends” (p. 217); “There are constant vicissitudes of feeling and impulse, constant formings and reformings of phantasy, over which it is certain very general tendencies pattern themselves: but with a flexibility in which, Freud suggests, the artist is peculiarly adept.”(pp. 218–19) Wollheim makes a connection to Freud when he writes that art has a constructive role to play in “the binding of energy or, what is theoretically a related process, the building up of the ego.”(p. 219) Imaginative immersion comes in degrees, with fixation marking an extreme. We are often interested in, and imaginatively engaged in, pieces of music, stories, poems, paintings; also, in the faces or actions or outlooks of people we know. More rarely, we are riveted to them in a manner that amounts to a kind of fixation or trance: for a while, nothing else seems to exist, interruptions are intolerable, thoughts stand still, and physical agitation ceases.

3.

THE DOUBLE CONSCIOUSNESS OF AESTHETIC PERCEPTION

257

sometimes fail us. But imaginative immersion in a work of art, or in a person, is not the same thing as identifying with the work or the person; so, just as I may identify with Vermeer’s sensibility but feel more deeply drawn into Bosch’s, I may identify with a close friend quite easily, but find myself imaginatively (and emotionally) more immersed in a stranger.47 Consider the use of masks or mask-like make-up in traditional Greek theater, in Japanese Butoh, in comedia del arte performances, in Chichimecan war dances, and in clown acts. On the one hand, the presence of a mask constitutes an extra “wall” or “screen” that stands between the viewer and the performer, effectively dissociating the world outside the performance from the world inside the performance, separating the world of practical action from the world of imaginative immersion. (Compare this to the loss of distance experienced in a play in which the actors make a point of “blending in” with the audience.) On the other hand, the eyes that look through the mask, or through the make-up, can acquire an extra power over us precisely because of the presence of the mask: our eyes become locked on those of the performer as we try to discern the intentionality and sensibility behind the surface gestures. At their best, performers who use masks make us simultaneously feel that we are peering into a world that is remote and new and that we have ourselves entered into that world, leaving our normal sensibility behind. Similarly, it is the artificiality and excessiveness of a clown’s guise that both enables us to disregard his cries as we would no others, yet also to share in his evident heartbreak more powerfully than we share in that of our neighbors. Consider also the case of quiet passages in a piece of music. The softest passages of a singer’s aria, or of Bach’s cello sonatas, heard as though from a distance, are often the passages which draw us “in” most effectively; they are 47 There are some interesting similarities between the state we have been discussing and the peculiar state discussed by Freud (1899) under the title Das Unheimlich (translated as The Uncanny). A typical instance of the uncanny, he suggests, is one in which something assumed to be inanimate is suddenly recognized to be alive (as when a puppet turns out to be a person after all), or where something merely imagined or fantasized (a house from one’s dreams, for example) appears before one as a material presence. Note how both instances involve the imaginative projection of an intentionality residing behind what is visible. Unlike the examples we have been considering, however, Freud’s examples of the uncanny depend on a certain kind of doubling—of puppet and person, of dream house and actual house—which leads to a disturbed kind of distrust. See Chapter 5, Section 3, for further discussion of uncanniness. Compare to Gell (1998) on the importance of a respected idol being unlike a human; otherwise, it will seem like an uncanny robot.

258

AESTHETIC PERCEPTION

the parts that are most deeply affecting. It is not the sounds themselves that become more distant in quiet passages: the sound waves fill the same spaces, whether loud or soft, and the singer or the cello producing the sound remains in the same place. Nor is it the case that we simply imagine the singer or the cello moving farther away (as we might do when we hear “distant” horns announcing the arrival of an army). What becomes distant is, instead, the metaphorical or imagined location of the music as something existing beyond the surface of sound. In the case of loud music, the music seems to exist nearer the surface (it wears its heart on its sleeve, as it were); in quieter passages, the music recedes farther away from the surface and we must “look” more intently to see what lies beneath. As the music moves farther “back,” it becomes more autonomous, increasing the imaginative challenge and the imaginative immersion. (Likewise, for the performer, quiet passages tend to require more self-control than loud passages, forcing the performer to access resources that lie more deeply within.) The delayed applause that follows a well-performed quiet ending (versus the burst of applause that follows a well-performed loud ending) is, at least partly, an indication of our greater immersion in the quiet passages and our sense that quiet music exists someplace “far away” (versus the “all around” sound of loud music)—a place from which we must now return.48

4. Aestheticizing Morality and Moralizing Aesthetics So far, I have argued that the hidden object or the inner meaning of an artwork is (often, anyway) like the hidden self or inner mentality of a person—equally objective, and equally perceivable. I have also argued on behalf of a certain kind of double consciousness, which is analogous to a kind of interpersonal ambivalence, in both domains. Given this emphasis on the similarities between persons and artworks, and between moral perception and aesthetic perception, it is important also to say something about the differences between aesthetic perception and moral perception. In my view 48

Likewise, our recognition of the temporal distance that separates us from a Greek urn, or an Egyptian statue, can increase our imaginative immersion in it. Plato, in The Republic, admires Egyptian art for representing many different aspects of a thing, and thus bringing us closer to its form. Wyatt Mason (2010) comments on David Foster Wallace: “If his work does impose an aesthetic distance it never sought to do less than bring particular persons as close as possible.”

4.

AESTHETICIZING MORALITY AND MORALIZING AESTHETICS

259

the difference does not concern what things are perceived in each domain (although there are of course differences between artworks and people), or how they are perceived (though there are differences here too), so much as why one seeks to perceive such things—the purposes of aesthetic perception versus the purposes of moral perception. Typically, though not always, we can have an effect on the sensibilities of persons but not the sensibilities of artworks; and typically but not always we have responsibilities towards persons but not towards artworks. It is useful to approach the difference between aesthetic perception and moral perception through a consideration of contemporary worries about aestheticizing morality on the one hand and moralizing aesthetics on the other. Accusations of aestheticizing morality or moralizing aesthetics usually begin by emphasizing the difference between fantasy and reality—placing art on the side of fantasy and morality on the side of reality. A recent, muchpublicized case of “aestheticizing morality” concerned the comments of the composer Karlheinz Stockhausen following the September 11 attack on the World Trade Center. He compared the attack to a spectacular concert for which “people practice madly for ten years . . . and then die.”49 The sculptor Richard Serra responded: “What mind-set does it take to completely lose the distinction between art and reality, leading to the preposterous and hypertrophic competition between an art performance and the annihilation of thousands of people? Mr. Stockhausen made us see the extreme of a not uncommon attitude, the aestheticization of reality; in this instance, the aestheticization of terror.”50 Fellow composer Gyorgi Ligeti recommended that Stockhausen be committed to a psychiatric clinic.51 Also recent and much publicized are accusations of “moralizing aesthetics” that have been leveled against those who would censor photographer Mapplethorpe’s depictions of sadistic acts, those who accuse Damien Hirst of abusing

49

According to a tape transcript from public broadcaster Norddeutscher Rundfunk, he went on: “Minds achieving something in an act that we couldn’t even dream of in music, people rehearsing like mad for ten years, preparing fanatically for a concert, and then dying, just imagine what happened there. You have people who are that focused on a performance, and then 5,000 people are dispatched to the afterlife, in a single moment. I couldn’t do that. By comparison, we composers are nothing. Artists, too, sometimes try to go beyond the limits of what is feasible and conceivable, so that we wake up, so that we open ourselves to another world.” Note that he made these comments when asked about the attack at a press conference that was scheduled in order to publicize an upcoming celebration of his music. 50 Serra (2001). 51 Tommasini (2001).

260

AESTHETIC PERCEPTION

animals, and audiences’ who boo Harry Kupfer’s opera productions. Here too the accusations tend to revolve around a supposed failure to preserve the distinction between fantasy and reality. Resistance to these accusations tends to be focused, accordingly, on the difficulty of drawing a clear line between art and reality. Our identity as persons and our moral values are said to be “social constructions,” with art inevitably playing a major role in those constructions. Equally, in the other direction, our art can’t help but express our values (most especially, perhaps, the values that are suppressed by the dominant culture). A. O. Scott, commenting on Ron Howard’s film A Beautiful Mind, writes that the decision to change a true story “is as much an artistic (and therefore an ethical) [my italics] choice as the casting of a certain actor or the selection of a camera angle.”52 It is not the fantasy-reality distinction, its importance or its unimportance, that is at issue in these cases, however. It is not accurate to accuse Stockhausen of losing the distinction between art and reality: he compared the long-prepared attack in which many people die to a long-prepared concert in which many people die (not just pretend to die). The real offense was his apparent admiration (“a spectacular concert”) rather than condemnation— though, of course, he was quick to counter this impression. If the aesthetic comparison had been more derogatory, surely it would not have prompted such an outcry. (There was no such criticism when another commentator likened the collapse of the World Trade Center to the destruction of a woman by an erect penis as portrayed in a “huge horrid snuff film.”53 ) Commentators who praise a work for its morally uplifting character incur no such wrath. Interestingly enough, the September 11 attack also prompted widespread demands for art that is morally relevant. Anthony Tommasini, writing for the New York Times, insisted that “the questions of relevance that were forced on classical-music organizations in the aftermath of the attack are the very questions they should be addressing anyway.” Conductor Michael Tilson Thomas said “I believe that music is the most important when the music stops. When a piece ends, that’s when I really measure what effect it 52

Scott (2002). A nicely argued counter-reaction to this trend can be found in Tanner (2003). Slater (2001). “It was lewd, the atrociously erect plane, the towers collapsing like velvet, like pudding, some huge horrid snuff film.” 53

4.

AESTHETICIZING MORALITY AND MORALIZING AESTHETICS

261

had on me or those who heard it. Did it in some way or another increase our sense of compassion or courage?”54 And, conversely, it has become increasingly popular to suppose that any sophisticated moral judgment is, at bottom, an aesthetic judgment. Martin Amis, for instance, objecting to cliche´, insists that “Style is morality.”55 (A more sophisticated defense of this position can be found in Nehamas’s Nietzsche: Life as Literature.56 ) One way to mediate these disputes would be to acknowledge that works of art, like people, can have two different kinds of value (or disvalue)— aesthetic value and moral value. They may be rightly admired or praised (criticized or blamed) for having (or lacking) either type of value, but it is important that the two sorts be kept distinct. This does not mean that there is no interaction between the two sorts of values. Carroll (2001) and Stecker (2005), for example, argue that many artworks are created for the purpose of engendering certain emotional responses, and that eliciting those responses often depends on the preservation or refinement of moral outlooks. But it is the skill with which the relevant emotions are engendered, not their moral worth, that determines the aesthetic value of an artwork. Insofar as the same thing could be said about the moral and aesthetic value of persons, the difference between our evaluation of works of art and our evaluation of persons would not be a matter of which sort of value is had by people versus works of art so much as, in a given context, which sort of value matters more. What this attempt at mediation overlooks, however, is the degree to which a skillful engendering of an emotion and the value of that emotion

54

Oestreich (2002). Wagner is a classic locus of the aesthetics versus morality dispute in music. See Herzog (2000): “In my considered judgment, a taste for Wagner shows poor judgment and should therefore be changed.” “Wagner’s revolutionary fervor sweeps me along, and I become fascinated, or at least not repulsed, by the idea that humanity should be redeemed through the destruction or exclusion of its ‘lower’ or ‘inferior’ members. The appeal of Wagner’s racism lies in an ideal of purity. Wagner’s racism speaks powerfully to a longing for simplicity and wholeness, to a yearning for finality and totality, to a craving for existence without shading, complexity, compromise, or doubt. It is this, I think, more than anything else, or at least as much as anything else, that makes the music dramas so compelling” (p. 46). As many critics have pointed out (see Kieran (2003)), this rejection not only simplifies the attitudes expressed by Wagner’s operas; it also simplifies the relation between the value of what is expressed and the value of how it is expressed, or the value of its being expressed. The effective portrayal of a megalomaniac god can be more valuable precisely because of the way it makes us feel shattered and appalled. 55 As reported by Tom Chatfield in Prospect magazine, Issue 167, 1 February 2010, “Martin Amis: the Prospect Interview.” 56 Nehamas (1985).

262

AESTHETIC PERCEPTION

may both depend on the same features or traits. Miller (1998) likens aesthetic skills to skills that are more generally important—in science as well as morality: discovery, surprise, conjecture, unification, analysis, and synthesis; and Goldie (2007a) argues that artistic activity and artistic perception include intentions, motives, skills, traits, and feelings that are chosen for their own sake, and that are partly constitutive of human well being. Examples of such traits that he cites include imagination, insight, sensibility, vision, creativity, wit, authenticity, integrity, intelligence, persistence, open-mindedness, and courage.57 The similarities between skills that enable us to perceive art as art and skills that enable us to perceive persons as persons—to see unity in appearances, to recognize indeterminacy, to sustain an appropriate double-consciousness—reflect similarities in the objects of aesthetic perception and the objects of moral perception. The divided state of mind that characterizes both perception of art as art and perception of persons as persons is a state in which we attend to surfaces closely in order to discern the sensibility beyond. All of this suggests that efforts at aestheticizing what is moral and moralizing what is aesthetic can be a good thing insofar as both aesthetic perception and moral perception depend on being attentive to the surface of things—the immediate, concrete, sensuous details of the things around us— and, simultaneously, recognizing the intentional patterns, aims, and orientations expressed through those surfaces. Morality should be no less attentive to surfaces than art is, and art should be no less attentive to what lies beyond surfaces than morality is. The crucial difference between aesthetic perception and moral perception does not lie in the object or the manner of perception, but, rather, in the purposes of perception. Aesthetic perception is not insulated from practical concerns because its objects are not real (a depicted landscape or the sensibility it expresses may be quite real), nor because art has no practical implications (the art we appreciate can and should affect the way we live our lives). Aesthetic perception, however, unlike moral perception, is not in the service of discovering the right thing to do. The perception of works of art is aimed at the discovery of new syntheses and new sensibilities regardless of any role they may play in one’s actions. Aesthetic perception is “pure perception” insofar as it is dedicated to discovering a variety of ways in 57

This list draws on Woodruff (2001).

4.

AESTHETICIZING MORALITY AND MORALIZING AESTHETICS

263

which things can hang together, and a variety of objects that correspond to these unities, irrespective of how these objects or these ways of holding things together fit with other sorts of objects and other ways of fitting things together. Whereas the perception of persons as persons seeks out the underlying unity of persons (in oneself as well as in others) in order to know what needs to be done, aesthetic perception seeks out alternative unities—many of them incapable of being sustained in a single person—for no other reason than the desire to expand the domain of things that can be perceived.58

58 If these remarks suggest an overly intellectual understanding of aesthetic appreciation—an understanding that privileges knowledge rather than pleasure, this is only because we have forgotten the special sort of pleasure that we experience when we perceive something genuinely new or when we suddenly see what we have previously known only intellectually. The pleasures of entertainment—of release of tension, titillation, gratifying fantasy—are real and important, but they are not epistemic pleasures, and they are quite different from the pleasures of perceptual discovery.

This page intentionally left blank

Conclusion The argument of this book can be thought of as a series of extensions stemming from a few basic assumptions and hypotheses. The starting assumptions were these:  

Perception has a distinctive epistemology and a distinctive phenomenology. A good account of perception ought to do justice to both of these things, explaining how each is possible and explaining how they are related.

The basic hypotheses were these:   

The experience of objectivity is what gives perception both its distinctive epistemology and its distinctive phenomenology. The experience of objectivity relies on the active and synthetic imagining of relevant alternatives. Although the objects of perception must be experienced as spatial in order to be experienced as objective, there is no need for a further “sensory” component in order for our knowledge to be perceptual.

Building on these assumptions and hypotheses (defended in Chapters 1 and 2), the book considers a number of domains in which claims to perception have been particularly controversial—the domain of explanatory reasons (Chapter 3), the domain of distant and abstract states of affairs (Chapter 4), the domain of morality (Chapter 5), and the domain of aesthetics (Chapter 6). In each case, I have tried to figure out how such things could be perceived in accordance with the guiding assumptions and insights of Chapters 1 and 2; and I have shown how many prima facie objections can be overcome.

266

CONCLUSION

Establishing the possibility of perception in each of these domains is not enough to establish its actuality, of course. Indeed, in some of the domains discussed in this book, I think that perception is rare. On the other hand, the emphasis on perception (as opposed to cognition, or thought, or inference) throughout the history of philosophy, and across many fields of interest, suggests that people have acquired knowledge of mathematics, of chemistry, of psychology, and of morality in ways that are importantly similar to the ways in which we normally acquire knowledge of the physical objects around us. So, although I do not begin by assuming that knowledge in these domains really is perceptual, the fact that my account allows for perception in these domains is a point in its favor. Indeed, as noted in the Introduction, if one begins by accepting that there are genuine cases of causal, rational, psychological, or moral perception, then this book’s understanding of perception (one that focuses on experiences of objectivity rather than sensory content) becomes compelling precisely because of its ability to make sense of such cases. In an attempt to understand both the epistemic and the phenomenological similarities between various perception-like experiences, I have provided an account of perception that is more unified than most—an account that emphasizes the crucial role of the imagination. Even those who choose to focus on the differences between what counts as perception across various domains should find this unified account helpful in clarifying the reasons why many people claim to perceive things that are not present to the senses. Furthermore, for those who are interested in the nature of perception within a particular domain—causal perception, or moral perception, or aesthetic perception, for example—this book offers to situate such perception within a broader theory of what perception involves.

References Adorno, Theodor. 1970/1984. “Enigmatic quality, truth content, metaphysics.” In Aesthetic Theory. Trans. C. Lenhardt. London: Routledge. Akins, Kathleen (ed.). 1996. Perception. Oxford: Oxford University Press. Arditi, A., Holzman, J. A., and Kosslyn, S. M. 1988. “Mental imagery and sensory experience in congenital blindness.” Neuropsychologia 26: 1–12. Audi, Robert. 2004. The Good in the Right. Princeton, NJ: Princeton University Press. —— 2008. “Intuition, Inference, and Rational Disagreement in Ethics.” Ethical Theory and Moral Practice 11.5: 475–92. Ayers, Michael. 1993. Locke: epistemology and ontology. London: Routledge and Kegan Paul. Banville, John. 2010. “The Still Mysterious Enchanter.” New York Review of Books, July 15. Beebee, Helen. 2003. “Seeing causing.” Proceedings of the Aristotelian Society 103: 257–80. —— et al. (eds). 2009. The Oxford Handbook of Causation. Oxford: Oxford University Press. Benton, M. G. 1995. “The self-conscious spectator.” British Journal of Aesthetics 35.4: 361–73. Bermu´dez, Jose´ Luis. 1995. “Non-conceptual content: From perceptual experiences to subpersonal computational states.” Mind and Language 10: 333–69. —— 2007. “What is at stake in the debate on non-conceptual content?” Philosophical Perspectives 21: 55–72. —— and Gardner, Sebastian (eds) 2003. Art and Morality. London: Routledge. Berntsen, Dorthe and Bohn, Annette. 2010. “Remembering and forecasting: The relation between autobiographical memory and episodic future thinking.” Memory and Cognition 38. 3: 265–78. Blackburn, Simon. 1990. “Filling in Space.” Analysis 50. 2: 62–5. —— 1993. “Circles, finks, smells and biconditionals.” Language and Logic 7: 259–79. —— 2001. “Normativity a` la mode.” Journal of Ethics 5.2: 139–53. Block, Ned (ed.). 1981. Imagery. Cambridge, MA: The MIT Press. Blocker, H. Gene. 1977. “A new look at aesthetic distance.” British Journal of Aesthetics 17.3: 219–29. Bloom, Paul. 2010. “The Moral Life of Babies.” The New York Times Magazine, May 9.

268

REFERENCES

Blum, Lawrence. 1994. Moral Perception and Particularity. New York: Cambridge University Press. Boghossian, Paul. 2001. “Inference and insight.” Philosophical and Phenomenological Research 63.3: 633–40. Bonjour, Laurence. 1985. The Structure of Empirical Knowledge. Cambridge MA: Harvard University Press. —— 1998. In Defense of Pure Reason. New York: Cambridge University Press. —— 2001. “Replies.” Philosophical and Phenomenological Research 63.3: 673–98. Brandom, Robert. 1994. Making it Explicit. Cambridge, MA: Harvard University Press. —— 1995. “Knowledge and the social articulation of the space of reasons.” Philosophy and Phenomenological Research 55.4: 895–908. —— 2001. “Reason, expression, and the philosophic enterprise.” In What is Philosophy?, ed. C. P. Ragland. New Haven: Yale University Press. —— 2002. “Knowledge and the internal revisited.” Philosophy and Phenomenological Research 64.1: 97–105. Brandt, Richard. 1996. Facts, Values, and Morality. Cambridge: Cambridge University Press. Braun, Eva. 1991. The World of the Imagination: sum and substance. Savage, MD: Rowman and Littlefield. Brewer, Bill. 1998. “Experience and reason in perception.” In Current Issues in Philosophy of Mind, ed. A. O’Hear. Cambridge: Cambridge University Press. —— 1999. Perception and Reason. Oxford: Clarendon Press. Briscoe, Robert Eamon. 2008. “Vision, action, and make-perceive.” Mind and Language 23.4: 457–97. Bullough, Edward. 1912/1963. “ ‘Psychical distance’ as a factor in art and an aesthetic principle.” British Journal of Psychology, 5: 85–118. Reprinted in Aesthetics and the Philosophy of Criticism, ed. Marvin Levich. Random House. Burge, Tyler. 2003. “Perceptual entitlement.” Philosophy and Phenomenological Research 67.3: 503–48. —— 2010. Origins of Objectivity. Oxford: Oxford University Press. Butterworth, G. E. and Cochran, E. 1980. “Toward a mechanism of joint visual attention.” International Journal of Behavioral Development 3: 253–72. Byatt, A. S. 2001. On Histories and Stories. Cambridge, MA: Harvard University Press. Byrne, Alex. 2005. “Perception and conceptual content.” In Contemporary Debates in Epistemology, ed. E. Sosa and M. Steup. London: Blackwell. Byrne, Alex. 2007. “Possibility and imagination.” Philosophical Perspectives. 21.1: 125–44.

REFERENCES

269

Byrne, Alex and Logue, Heather (eds). 2009. Disjunctivism: Contemporary Readings. Cambridge, MA: The MIT Press. Byrne, Ruth. 2007. “Rational imagination: How people create alternatives to reality.” Philosophical Psychology 20.2: 269–73. Camp, Elisabeth. 2004. “The generality constraint, nonsense, and categorical restrictions.” Philosophical Quarterly 54: 209–31. —— 2007. “Thinking with maps.” Philosophical Perspectives 21:1. Oxford: WileyBlackwell, 145–82. Campbell, John. 1994. Past, Space and Self. Cambridge, MA: The MIT Press. —— 1995. “The body image and self-consciousness.” In The Body and the Self, ed. Jose´ Luis Bermudez, Anthony Marcel, and Naomi Eilan. Cambridge, MA: The MIT Press. —— 1997. “Sense, reference, and selective attention.” Aristotelian Society, Supplementary 7: 55–74. —— 2002. Reference and Consciousness. Oxford: Clarendon Press; New York: Oxford University Press. —— 2006. “What is the role of the location in the sense of a visual demonstrative? Reply to Matthen.” Philosophical Studies 127: 239–54. Carroll, Noe¨l. 2001. “Moderate moralism.” In his Beyond Aesthetics. Cambridge: Cambridge University Press. Cartwright, Nancy. 1999. The Dappled World: A Study of the Boundaries of Science. Cambridge and New York: Cambridge University Press. Casey, Edward S. 1976. Imagining: A Phenomenological Study. Bloomington and London: Indiana University Press. Cassam, Quassim. 2007. The Possibility of Knowledge. Oxford: Clarendon Press; New York: Oxford University Press. Cavell, Stanley. 1979. The Claim of Reason. New York: Oxford University Press. —— 1988. In Quest of the Ordinary: Lines of Skepticism and Romanticism. Chicago: University of Chicago Press. Chappell, Timothy. 2008. “Moral perception.” Philosophy 83: 421–37. Cheng, Anne Anlin. 2000. The Melancholy of Race. Oxford: Oxford University Press. Church, Jennifer. 1995. “L’emotion et l’interiorisation des actions.” Raisons Pratiques 6: La couleur des pensees, ed. P. Paperman and R. Ogien. Paris: Editions de l’Ecole des Hautes Etudes en Sciences Sociales: 219–36. —— 1998. “Two sorts of consciousness?” Communication and Cognition 31.1: 58–72. —— 2000. “ ‘Seeing As’ and the double bind of consciousness.” Journal of Consciousness Studies 7.8/9: 99–111.

270

REFERENCES

Church, Jennifer. 2002. “Taking it to heart: What choice do we have?” Monist: An International Quarterly Journal of General Philosophical Inquiry. 85.3: 361–80. —— 2005. “Reasons of which reason knows not.” Philosophy, Psychiatry, and Psychology, 12.1: 31–41. Church, Jennifer. 2008. “The hidden image: A defense of unconscious imagining and its importance.” American Imago 65.3: 379–404. —— 2010. “Seeing reasons.” Philosophy and Phenomenological Research 80.3: 638–70. Clark, Austen. 1993. Sensory Qualities. Oxford: Clarendon Press. Close, Darly. 1976. “What is non-epistemic seeing?” Mind 85.338: 161–70. Coates, Paul. 2007. The Metaphysics of Perception: Wilfred Sellars, perceptual consciousness and critical realism. New York: Routledge. Cocude, Marguerite, et al. 1997. “Latency and duration of visual mental images in normal and depressed subjects.” Journal of Mental Imagery 21.1–2: 127–42. Cohen, L. Jonathan. 1992. An Essay on Belief and Acceptance. Oxford: Oxford University Press. Cohen, Ted. 1998. “On consistency in one’s personal aesthetics.” In Aesthetics and Ethics: Essays as the Intersection, ed. Jerrold Levinson. Cambridge: Cambridge University Press. Collins, John, Hall, Ned, and Paul, L. A. (eds). 2004. Counterfactuals and Causation. Cambridge, MA: The MIT Press. Costa, Michael. 1989. “Hume and causal realism.” Australian Journal of Philosophy 67.2: 172–90. Currie, Gregory and Ravenscroft, Ian. 2002. Recreative Minds. Oxford: Clarendon Press; New York: Oxford University Press. Danto, Arthur. 1986. The Philosophical Disenfranchisement of Art. New York: Columbia University Press. Davidson, Donald. 2001. Subjective, Intersubjective, Objective. Oxford: Clarendon Press. DeLapp, Kevin M. 2007. “Moral perception and moral realism: An ‘intuitive’ account of epistemic justification.” Review Journal of Political Philosophy 5: 43–64. DeLapp, Kevin Michael. 2009. “The merits of dispositional moral realism.” The Journal of Value Inquiry 43: 1–18. Dennett, Daniel Clement. 1991. Consciousness Explained. Illustrated by Paul Weiner. Boston: Little, Brown and Co. Descartes, Rene´. 1637/1993. Discourse on Method and Meditations on First Philosophy. Trans. Donald Cress. Indianapolis: Hackett. —— 1637/2001. Discourse on Method. Trans. with an introduction by Paul J. Olscamp. Indianapolis: Hackett.

REFERENCES

271

—— 1681/1952. Rules for the Direction of the Mind. Trans. Elizabeth Haldane and G. R. T. Ross. Chicago: Encyclopaedia Britannica. Dowe, Phil. 2000. “The conserved quantity theory defended.” Theoria 15. 37: 11–31. Dretske, Fred. 1969. Seeing and Knowing. Chicago: University of Chicago Press. —— 1981. Knowledge and the Flow of Information. Cambridge, MA: The MIT Press. Dummett, M. 1954. “Can an effect precede its cause.” Proceedings of the Aristotelian Society Supplement 28: 27–44. —— 1964. “Bringing about the past.” Philosophical Review 73: 338–59. Dustin, Christopher and Schaeffer, Denise. 2006. “Looks matter: Beholding justice in the ‘Republic’.” Review of Politics 68.3: 449–73. Dylan, Bob. 2004. Chronicles, vol.1 New York: Simon and Schuster. Eichler, Jeremy. 2005. “Exploring the Inner Voices and Musical Architecture.” The New York Times, May 3. Evans, Gareth. 1982. The Varieties of Reference, ed. John McDowell. Oxford: Oxford University Press. Fine, Kit. 2002. “The varieties of necessity.” In Conceivability and Possibility, ed. T. S. Gendler and J. Hawthorne. Oxford: Oxford University Press. Flew, Antony. 1957. “Causal disorder again.” Analysis 17: 81–6. Fodor, Jerry. 1995. “Review of Christopher Peacocke’s A Study of Concepts.” London Review of Books, 34.20: 10–11. Foucault, Michel. 1970. The Order of Things: An archaeology of the human sciences. London: Tavistock. Frankfurt, Harry. 1971. “Freedom of the will and the concept of a person.” Journal of Philosophy 68: 5–20. Frankfurt, Harry G. 2004. The Reasons of Love. Princeton, NJ, Princeton University Press. Freud, Sigmund. 1899/2003. The Uncanny. Trans. David McLintock. London: Penguin Group. —— 1914/1975. “Remembering, repeating, and working through.” The Standard Edition of the Complete Psychological Works of Sigmund Freud. Trans. under the general editorship of James Strachey, in collaboration with Anna Freud. London: Hogarth Press. Friedman, Michael. 1996. “Exorcising the philosophical tradition: Comments on John McDowell’s Mind and World.” The Philosophical Review 105.4: 427–67. Gardner, Sebastian. 1993. Irrationality and the Philosophy of Psychoanalysis. Cambridge: Cambridge University Press. Gattis, Meredith (ed.). 2001. Spatial Schemas and Abstract Thought. Cambridge, MA: The MIT Press.

272

REFERENCES

Gell, Alfred. 1998. Art and Agency: An Anthropological Theory. Oxford: Clarendon Press. Gendler, Tamar Szabo. 2000. “The puzzle of imaginative resistance.” Journal of Philosophy 97.2: 5–81. Gendler, T. S. 2006. “Imaginative contagion.” Metaphilosophy 37.2: 183–203. Gendler, T. S. and Hawthorne, J. (eds) 2006. Perceptual Experience. Oxford: Clarendon Press. Genter, Dedre, and Stevens, Albert (eds) 1983. Mental Models. Hillsale, NJ: L. Erlbaum Associates. Gettier, Edmund L. 1963. “Is justified true belief knowledge?” Analysis 23.60: 121–3. Gibbons, Sarah. 1995. Kant’s Theory of Imagination: Bridging Gaps in Judgement and Experience. Oxford: Clarendon Press, New York: Oxford University Press. Glymour, Clark. 1998. “Psychological and normative theories of causal power and probabilities of causes.” In Uncertainty in Artificial Intelligence, ed. G. F. Cooper and S. Moral. San Francisco: Morgan-Kaufmann. Goldie, Peter. 2007a. “Seeing what is the kind thing to do: perception and emotion in morality.” Dialectica 61.3: 347–61. —— 2007b. “Towards a virtue theory of art.” British Journal of Aesthetics 47.4: 372–87. Gombrich, E. H. 1960. Art and Illusion. London: Phaidon. Goodman, Nelson. 1978. Ways of Worldmaking. Indianapolis: Indiana University Press. Guyer, Paul. 1997. “Kant’s theory of imagination.” The Journal of Aesthetics and Art Criticism. 55.3: 337–40. Hacking, Ian. 2005. “A new way to see a leaf.” New York Review of Books, 7 April, 70–2. Hall, Ned. 2006. “Philosophy of causation: Blind alleys exposed; promising directions highlighted.” Philosophy Compass 1.1: 86–94. Hamilton, Christopher. 2003. “Art and moral education.” In Art and Morality, ed. Jose Luis Bermudez and Sebastian Gardner. London and New York: Routledge. Hanfling, Oswald. 2000. “Five kinds of distance.” British Journal of Aesthetics 40.1: 89–102. —— 2003. “Paradoxes of aesthetic distance.” British Journal of Aesthetics 43.2: 175–86. Harman, Gilbert. 1973. Thought. Princeton, NJ: Princeton University Press. Haugland, John. 1996. “Direct Perception.” In Perception, ed. K. Akins,. Oxford: Oxford University Press. Havelock, Eric. 1963. Preface to Plato. Cambridge, MA: Harvard University Press. Heck, Richard. 2000. “Nonconceptual content and the ‘space of reasons’.” The Philosophical Review 109.4: 483–523. Heil, John. 1982. “Seeing is believing.” American Philosophical Quarterly 19.3: 229–39.

REFERENCES

273

Herzog, Patricia. 2000. “Akrasia and aesthetic judgement.” The Journal of Aesthetics and Art Criticism 58.1: 39–49. Hobson, Peter. 2005. What puts the jointness into joint attention?” in Joint Attention: Communication and Other Minds, ed. Naomi Eilan, Christoph Hoerl, Teresa McCormack, and Johannes Toessler. Oxford: Oxford University Press. Hopkins, Robert. 1998. Picture, Image and Experience: A philosophical inquiry. Cambridge and New York: Cambridge University Press. Howes, David (ed.). 2005. Empire of the Senses: The sensual culture reader. Oxford; New York: Berg. Hume, David. 1777/1978. Enquiries Concerning Human Understanding and Concerning the Principles of Morals, ed. L. A. Selby-Bigge and P. H. Nidditch. Oxford: Clarendon Press. Hurley, S. L. 1998. Consciousness in Action. Cambridge, MA: Harvard University Press. Husserl, Edmund. 1913/2001. Logical Investigations. Trans. J. L. Findlay. London: Routledge. Hyland, Drew. 1973. The Origins of Philosophy. New York: G. P. Putnam’s Sons. Jacobson, Daniel. 2005. “Seeing by feeling: virtues, skills, and moral perception.” Ethical Theory and Moral Practice 8.4: 387–410. Johnson, Mark. 1990. The Body in the Mind. Chicago: University of Chicago Press. Johnson, Paul. 1993. Wittgenstein: Re-thinking the Inner. London: Routledge. Johnson-Laird, Philip N. 1983. Mental Models. Cambridge, MA: Harvard University Press. Kahneman, Daniel. 1995. “Varieties of counterfactual thinking.” In What Might Have Been: The Social Psychology of Counterfactual Thinking, ed. Neal J. Roese and James M. Olson. Mahwah, NJ: Lawrence Erlbaum. Kant, Immanuel. 1787/1929. Critique of Pure Reason. Trans. Norman Kemp Smith. London: Macmillan. —— 1788/2002. Critique of Practical Reason. Trans. Werner S. Pluhar; introduction by Stephen Engstrom. Indianapolis: Hackett. Kant, Immanuel. 1790/1951. Critique of Judgement. Trans., with an introduction, by J. H. Bernard. New York: Hafner. Kelly, Sean. 2003. “Husserl and phenomenology.” In Blackwell Companion to Continental Philosophy, ed. Robert C. Solomon. Oxford: Blackwell. —— 2005. “Seeing things in Merleau-Ponty.” In The Cambridge Companion to Merleau-Ponty, ed. Carman Taylor. Cambridge: Cambridge University Press. —— 2005. “Temporal awareness.” In Phenomenology and Philosophy of Mind, ed. David Woodruff Smith and Arnie L. Thompson. Oxford: Clarendon Press.

274

REFERENCES

Ketchum, Richard. 1979. “Knowledge and recollection in the Phaedo.” Journal of the History of Philosophy 17.3: 243–53. Kieran, Matthew. 2003. “Forbidden knowledge: the challenge of immoralism,” in Art and Morality, ed. Jose Luis Bermudez and Sebastian Gardner. London: Routledge. Kind, Amy. 2001. “Putting the image back in imagination.” Philosophy and Phenomenological Research 62.1: 85–109. Kitcher, Patricia. 1990. Kant’s Transcendental Psychology. New York: Oxford University Press. Kornblith, Hilary. 2002. Knowledge and its Place in Nature. Oxford: Oxford University Press. Korsgaard, Christine. 2003. “Realism and constructivism in twentieth century moral philosophy.” Philosophy Documentation Center, 28: 99–122. Kosslyn, Steve. 1994. Image and Brain: The resolution of the imagery debate. Cambridge, MA: The MIT Press. Kukla, Rebecca. 2006. “Objectivity and perspective in empirical knowledge.” Episteme: A Journal of Social Epistemology 3.1: 80–95. Kulvicki, John V. 2006. On Images: Their structure and content. Oxford: Clarendon Press. Kung, Peter. 2010. “Imagination as guide to possibility.” Philosophy and Phenomenological Research 81.3: 620–63. Kunzendorf, Robert. 2011. “Depression, unlike normal sadness, is associated with a ‘flatter’ self-perception and a ‘flatter’ phenomenal world.” Imagination, Cognition, and Personality 30.4: 441–61. Larmore, Charles. 2002. “Attending to Reasons.” In Reading McDowell, ed. Nicholas Smith. New York: Routledge. Lear, Jonathan. 1998. Open Minded: Working Out the Logic of the Soul. Cambridge, MA: Harvard University Press. Leher, Keith. 1990. Metamind. Oxford: Oxford University Press. Lewis, David. 1986. Philosophical Papers. Oxford: Oxford University Press. Llinas, Rondolfo and Ribary, Urs. 1994. “Perception as an oneiric-like state modulated by the senses.” In Large-Scale Neuronal Theories of the Brain, ed. Christof Koch and Joel Davis. Cambridge, MA: The MIT Press. Lopes, Dominic. 1996. Understanding Pictures. Oxford: Clarendon Press. Lopes, Dominic McIver. 2000. “What is it like to see with your ears? The representational theory of mind.” Philosophy and Phenomenological Research 60. 2: 439–53. Martin, M. G. F. 2002. “The transparency of experience.” Mind and Language 17: 376–425. Mason, Wyatt. 2010. “Smarter than You Think.” New York Review of Books, July 15.

REFERENCES

275

McDowell, John. 1982. “Criteria, defeasibility, and knowledge.” Proceedings of the British Academy 17: 455–79. —— 1984. “Wittgenstein on following a rule.” Synthese 58: 325–63. —— 1985. “Values and secondary qualities.” In Morality and Objectivity: A Tribute to J. L. Mackie, ed. Ted Honderich. London: Routledge and Kegan Paul. —— 1986. “Singular thought and the extent of inner space” in Subject, Thought, and Context, ed. P. Pettit and J. McDowell. Oxford: Clarendon Press. —— 1994. Mind and World. Cambridge, MA: Harvard University Press. —— 1998. Mind, Value, and Reality. Cambridge, MA: Harvard University Press. —— 2002a. “Replies.” In Reading McDowell: on Mind and World, ed. Nicholas H. Smith. London and New York: Routledge. —— 2002b. “Knowledge and the internal revisited.” Philosophy and Phenomenological Research 64.1. —— 2006. “Replies Response to Jennifer Church.” Teorema 25.1: 97–100. McGinn, Colin. 1997. Ethics, Evil, and Fiction. Oxford: Oxford University Press. —— 2004. Mindsight: image, dream, meaning. Cambridge, MA: Harvard University Press. Mengzi. 1998. Mencius. Trans. David Hinton. Washington DC: Counterpoint. Menzies, Peter. 2009. “Platitudes and Counterexamples.” In The Oxford Handbook of Causation, ed. Helen Beebee et al. Oxford: Oxford University Press. Menzies, Peter and Price, Huw. 1993. “Causation as a secondary quality.” The British Journal for the Philosophy of Science 44.2: 187–203. Michotte, A. 1963. The Perception of Causality. London: Methuen. Miller, Alan. 2007. “What the disjunctivist is right about.” Philosophy and Phenomenological Research 74.1: 176–98. Miller, Mitchell. 1999. “Figure, ratio, form: Plato’s five mathematical studies.” Apeiron 32.4: 73–88. Miller, Richard W. 1998. “Three versions of objectivity: aesthetic, moral, and scientific.” In Aesthetics and Ethics: Essays as the intersection, ed. Jerrold Levinson. Cambridge: Cambridge University Press. Millikan, Ruth Garrett. 1984. Language, Thought, and Other Biological Categories: New foundations for realism. Cambridge, MA: The MIT Press. —— 2000. On Clear and Confused Ideas. Cambridge and New York: Cambridge University Press. Milner, A. David and Goodale, Melvyn. 2006. The Visual Brain in Action. Oxford and New York: Oxford University Press. Moore, G. E. 1912. Ethics. New York: H. Holt and Co. Moran, Richard. 1994. “The expression of feeling in imagination.” Philosophical Review 103.1: 75–106.

276

REFERENCES

Moran, Richard. 2001. Authority and Estrangement: An Essay on Self-Knowledge. Princeton, NJ: Princeton University Press. Mourelatos, Alex. 1970. The Route of Parmenides. New Haven, CT: Yale University Press. Mulhall, Stephen. 1990. On Being in the World: Wittgenstein and Heidegger on seeing aspects. London and New York: Routledge. —— 2000. “Misplacing freedom, displacing the imagination: Cavell and Murdoch on the fact/value distinction.” Philosophy and Public Affairs 47 Supplement: 255–77. Murdoch, Iris. 1970. The Sovereignty of Good. New York: Schocken Books. Murdoch, Iris. 1993. Metaphysics as a Guide to Morals. London: Penguin. Mylayne, Jean-Luc. 2002. (accessed 30 July 2012). Nagel, Thomas. 1974. “What is it like to be a bat?” Philosophical Review 83: 435–50. Nehamas, Alexander. 1985. Nietzsche: Life as Literature. Cambridge, MA: Harvard University Press. —— 1996. “What should we expect from reading? (There are only aesthetic values).” Salmagundi 111: 27–58. Neiman, Susan. 1994. The Unity of Reason. New York: Oxford University Press. Neisser, Ulric. 1976. Cognition and Reality. San Francisco: W. H. Freeman. Noddings, Nell. 1984. Caring: A Feminine Approach to Ethics and Moral Education. Berkeley: University of California Press. Noe¨, Alva. 2002 “On what we see.” Pacific Philosophical Quarterly 83: 57–80. —— 2003. “Causation and perception: the puzzle unraveled.” Analysis 63.2: 93–100. —— 2004. Action in Perception. Cambridge, MA: The MIT Press. —— 2006. “Experience without the head.” In Perceptual Experience, ed. T. S. Gendler and J. Hawthorne. Oxford: Clarendon Press. —— 2008. “Reply to Campbell, Martin, and Kelly.” Philosophy and Phenomenological Research 76.3: 691–706. Nussbaum, Martha Craven. 1990. Love’s Knowledge: Essays on philosophy and literature. New York: Oxford University Press. O’Callaghan, Casey. 2007. Sounds: A Philosophical Theory. Oxford: Oxford University Press. O’Shaughnessy, Brian. 1990. “The appearance of a material object.” Philosophical Perspectives 4: 131–51. Oestreich, James R. 2002. “Michael Tilson Thomas: Maverick in a City of Same.” The New York Times, February 10. Ogden, John. 1971. “From spatial to aesthetic distance in the eighteenth century.” Journal of the History of Ideas 35: 158–84.

REFERENCES

277

Paul, L.A. 2009. “Counterfactual theories.” In The Oxford Handbook of Causation, ed. Helen Beebee et al. Oxford: Oxford University Press. Peacocke, Christopher. 1983. Sense and Content: Experience, Thought, and their Relations. Oxford: Oxford University Press. —— 1986. Thoughts: An Essay on Content. Oxford: Blackwell. Piper, Adrian M. S. 1991. “Impartiality, compassion, and modal imagination.” Ethics 101: 725–27. Plato. 1968. The Republic. Trans. Allan Bloom, with an introduction and interpretive essay. New York: Basic Books. —— 1980. Meno. Trans. G. M. A. Grube. Indianapolis: Hackett Publishing. Pleasants, Nigel. 2008. “Institutional wrongdoing and moral perception.” Journal of Social Philosophy 39.1: 95–115. Quine, W. V. and Ullian, J. S. 1970. The Web of Belief. New York: Random House. Rawson, Glenn. 2006. “Platonic recollection and mental pregnancy.” Journal of the History of Philosophy 44.2: 137–55. Reed, Lisa. 1999. “Necessary versus probable cause.” Journal of Philosophical Logic 28: 289–326. Reid, Thomas. 1792/2001. “Of power.” The Philosophical Quarterly 51.202: 3–12. Rorty, Amelie. 2008. “Ambivalence: Imagination and epistemic responsibility.” Talk at Pacific Division APA meetings. Russell, James. 1987. “Reasons for retaining the view that there is perceptual development in childhood.” In Philosophical Perspectives on Developmental Psychology, ed. James Russell. Oxford: Blackwell. Russow, Lilly-Marlene. 1978. “Some recent work on imagination.” American Philosophical Quarterly 15: 57–66. Ryle, Gilbert. 1949. The Concept of Mind. New York: Harper and Row. Sabin, Theodore R. 1998. “Believed-in imaginings.” In Believed-In Imaginings: The Narrative Construction of Reality, ed. Joseph De Rivera and Theodore R. Sabin. Washington DC: American Psychological Association. Sacco, Giuseppe and Ruggieri, Vezio. 1997. “Mental imagery and symptom patterns.” Imagination, Cognition and Personality 17.4: 313–21. Sacks, Mark. 2003. Objectivity and Insight. Oxford: Clarendon Press. Sacks, Oliver. 2003. “What the blind see.” The New Yorker, 28 July, 48–59. Schiffer, Stephen. 1991. “Ceteris paribus laws.” Mind 100.397: 1–17. Schultz, Alexander. 2009. Mind’s World. Seattle, WA: University of Washington Press. Scott, A.O. 2002. “Critic’s Notebook: A ‘Mind’ Is a Hazardous Thing to Distort.” The New York Times, March 21.

278

REFERENCES

Scruton, Roger. 1974. Art and Imagination: A Study in the Philosophy of Mind. London: Methuen and Co. —— 1999. The Aesthetics of Music. New York: Oxford University Press. Searle, John. 1992. The Rediscovery of the Mind. Cambridge, MA: The MIT Press. Seemann, Axel. 2008. “Person perception.” Philosophical Explorations 11.3: 243–62. Segal, S. 1970. “Imagery and reality: Can they be distinguished.” In Origins and Mechanisms of Hallucinations, ed. W. Keup. New York: Plenum Press. Seidman, Jeffrey. 2005. “Two sides of silencing.” The Philosophical Quarterly 55.218: 68–77. Sellars, Wilfred. 1956/1997. “Empiricism and the philosophy of mind.” In Minnesota Studies in the Philosophy of Science, ed. H. Feigl and M. Scriven. Reissued (1997), with commentary by Robert Brandom, as Empiricism and the Philosophy of Mind. Cambridge, MA: Harvard University Press. Sepper, Dennis L. 1996. Descartes’ Imagination: Proportion, Images, and the Activity of Thinking. Berkeley: University of California Press. Serra, Richard. 2001. “Stockhausen; Aestheticizing Terror.” The New York Times, October 21. Shapiro, Gary. 2000. “This is not a Christ: Nietzsche, Foucault, and the geneology of vision.” In Why Nietzsche Still?, ed. Alan D. Schrift. Berkeley: University of California Press. Sherman, Nancy. 1998. “Empathy and imagination.” Midwest Studies in Philosophy 22: 82–119. Shusterman, Richard. 1997. “The end of aesthetic experience.” Journal of Aesthetics and Art Criticism 55: 29–41. —— 2000. Performing Live: Aesthetic Alternatives for the Ends of Art. Ithaca: Cornell. —— 2002. Surface and Depth. Ithaca: Cornell. Sibley, F. N. 1971. “Analyzing seeing.” In his Perception: A Philosophical Symposium. London: Taylor and Francis. Siegel, Susanna. 2009. “The visual experience of causation.” Philosophical Quarterly 59.236: 519–40. Slater, Lauren. 2001. “Talk To Me.” The New York Times Magazine, September 23. Smith, Nicholas (ed.). 2002. Reading McDowell. New York: Routledge. Snowdon, Paul. 1979. “Perception, vision and causation.” Proceedings of the Aristotelean Society 81: 175–92. —— 1990. “The objects of perceptual experience.” Proceedings of the Aristotelean Society Supplement 64: 121–50. Sosa, Ernest. 2007. A Virtue Epistemology. Oxford: Clarendon Press; New York: Oxford University Press.

REFERENCES

279

Soteriou, Matthew. 2008. “The epistemological role of episodic recollection.” Philosophy and Phenomenological Research 27.2: 472–92. —— 2010. “Perceiving events.” Philosophical Explorations 13.3: 223–41. Stanley, Jason and Williamson, Timothy. 2001. “Knowing how.” The Journal of Philosophy 98.8: 411–44. Stecker, Robert. 2005. “The interaction of ethical and aesthetic value.” British Journal of Aesthetics 45.2: 138–50. Stenning, Keith. 2002. Seeing Reason: Image and language in learning to think. New York: Oxford University Press. Stevenson, R. J. and Case, T. I. 2005. “Olfactory imagery: a review.” Psychonomic Bulletin and Review 12: 244–64. Strawson, P. F. 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen. —— 1974. “Perception and imagination.” In his Freedom and Resentment, and other Essays. London: Methuen. Strevens, Michael. 2008. Depth: An Account of Scientific Explanation. Cambridge, MA: Harvard University Press. Suppes, Patrick. 1970. A Probabilistic Theory of Causation. Amsterdam: North-Holland. Tanner, Michael. 2003. “Ethics and aesthetics are ___?” In Art and Morality, ed. Jose Luis Bermudez and Sebastian Gardner. London: Routledge. Thomas, N. J. T. 2003. “Mental imagery, philosophical issues about.” In Encyclopedia of Cognitive Science, vol. 2, ed. L. Nadel. London: Nature Publishing. (accessed 8 November 2010). Tommasini, Anthony. 2001. “Music; The Devil Made Him Do It.” The New York Times, September 30. Updike, John. 2003. “Singular in everything.” New York Review of Books, 6 November, 14–18. Vega, Manuel de. 1996. Models of Visuospatial Cognition. New York: Oxford University Press. Velleman, J. David. 2001. “Identification and identity.” In The Contours of Agency: Essays on Themes from Harry Frankfort, ed. Sarah Buss and Lee Overton. Cambridge, MA: The MIT Press. —— 2006. “Love as a moral emotion.” In his Self to Self: Selected Essays. Cambridge: Cambridge University Press. Velleman, J. David. 2006. Self to Self: Selected Essays. Cambridge: Cambridge University Press. Vygotskii, Lev Semenovich. 1962. Thought and Language. Ed. and trans. Eugenia Hanfmann and Gertrude Vakar. Cambridge, MA: The MIT Press. Walker, Margaret Urban. 1987. “Moral particularity.” Metaphilosophy 18.3/4: 171–85. Walker, Ralph C. S. 1978. Kant. London: Routledge and Kegan Paul.

280

REFERENCES

Walton, Kendall. 1990. Mimesis as Make-Believe: On the foundations of the representational arts. Cambridge, MA: Harvard University Press. Warnock, Mary. 1976. Imagination. Berkeley: University of California Press. Watkins, Michael and Jolley, Kelly Dean. 2002. “Pollyanna realism: Moral perception and moral properties.” Australasian Journal of Philosophy 80.1: 75–85. Williams, Bernard. 1973. “Imagination and the self.” In his Problems of the Self. Cambridge: Cambridge University Press. Williamson, Timothy. 2002. Knowledge and its Limits. Oxford; New York: Oxford University Press. Wittgenstein, Ludwig. 1953. Philosophical Investigations. Trans. G. E. M. Anscombe. Oxford: Blackwell. —— 1967. Zettel. Ed. G. E. M. Anscombe and G. H. von Wright. Trans. G. E. M. Anscombe. Berkeley: University of California Press. —— 1980. Remarks on the Philosophy of Psychology, Ed. G. E. M. Anscombe and G. H. von Wright. Trans. G. E. M. Anscombe. Oxford: Blackwell. Wittgenstein, Ludwig. 1982–1992. Last Writings on the Philosophy of Psychology. Ed. G. H. von Wright and Heikki Nyman. Trans. C. G. Luckhardt and Maximilian A. E. Aue. Chicago: University of Chicago Press. Wollheim, Richard. 1968. Art and its Objects. New York: Harper and Row. —— 1974. On Art and the Mind. Cambridge, MA: Harvard University Press. —— 1984. The Thread of Life. Cambridge, MA: Harvard University Press. —— 1987. Painting as an Art. London: Thames and Hudson. Woodruff, David. 2001. “A virtue theory of aesthetics.” Journal of Aesthetic Education 35: 23–36. Wright, Edmond. 1981. “Yet more on non-epistemic seeing.” Mind 90: 586–91. Yablo, Stephen. 2003. “Causal relevance: Mental, moral, epistemic.” Philosophical Issues 13: 316–29. Reprinted in Thoughts, Stephen Yablo. Oxford University Press. —— 2004. “Advertisement of a sketch of an outline of a prototheory of causation.” In Counterfactuals and Causation, ed. J. Collins, N. Hall, and L. A. Paul. Cambridge, MA: The MIT Press. Yamakawa, Hideya. 2008. Visible and Invisible in Greek Philosophy. Lanham, MD: University Press of America. Young, Michael. 1988. “Kant’s view of imagination.” Kant-Studien: Philosophische Zeitschrift der Kant-Gesellschaft 79: 140–64. Zahavi, Dan. 2006. Subjectivity and Selfhood. Cambridge, MA: The MIT Press.

Index abstract states of affairs 151–63 action animal 38 direct result of perception 133–6, 180–1, 215–16 as evidence of imagining 61 and freedom 120–2 imagined 55–6 on Noe¨’s view of perception 31–5 and perceiving what to do 209–16 and perception of persons 197–200 and perspectives 50, 77–8 in relation to art 251–7 in relation to perceiving vs. imagining 73–5 seeing action as action 105 aesthetic distance 246–58 agency and cause 91 n. 8, 111 n. 39 and personhood 198–200 ambivalence 207–9, 255–8 Aristotle on cause 89, 105 n. 32 on moral upbringing 118 on moral perception 188 n. 1, 195 n. 12 artworks appreciation of 240–6 defining 227–40 and morality 258–64 representational 229 n. 2, 238 n. 17, 241–3, 254–5 see also aesthetic distance Audi, Robert 189 n. 2, 190–1, 195 autonomy 215, 238, 256 Ayers, Michael 28–9 Beebee, Helen 91 n. 8, 92 belief in perception 12–16 vs. perception 35–6, 52–3, 95, 107, 123–36, 167, 240 Blocker, Gene H. 254–5 Bonjour, Laurence 26, 148 n. 15, 152–3 Brandom, Robert

on expression 238–9 n. 17 as internalist 23 n. 22 on justificatory immediacy 22 n. 19, 39, 77 n. 56, 146 n. 10 vs. McDowell 117 n. 53, 118 n. 58 on space of reasons 112–14 Brewer, Bill on capacity to imagine 48–50 inferential perceptual knowledge 146 n. 10, 155 on non-demonstrative perceptual knowledge 27 n. 26 richness of perception 69 n. 38 Bullough, Edward 249–51 Burge, Tyler externalism 23 n. 22, 141 n. 3 and multiple perspectives 32 n. 35, 51–2, 83 n. 61 on objectivity 29 n. 30, 39, 77 n. 54 on perceptual immediacy 20, 145 n. 9 Byrne, Alex non-conceptual content 14 n. 6 and images 64 n. 30 sensory content 39 Campbell, John on egocentric space 50 on object as cause 29 n. 29, 200 n. 21 on expectations 54 n. 11 Carroll, Noe¨l 261 Cassam, Quassim on, perception and space 33 n. 36 on object independence 67 n. 34 categorical imperative 217, 220–4, 245 causation backwards 148–51 counterfactual accounts 29 n. 29, 91–6 perception of 90–101 as requirement for perception 29 n. 29, 40, 147–51 Cavell, Stanley on perceiving persons 204 on uncanniness 204–7

282

INDEX

ceteris paribus conditions 96–7, 108–10, 211, 218 n. 50, 219 Chappell, Timothy 215–16 Cheng, Anne Anlin 253 Cohen, Ted 37 n. 41 Cohen, Jonathan 134–5 color 32–4, 40–1, 90 n. 5 Descartes on 180 objectivity of 193 optimal viewing of 187 consistency 30, 110 n. 37, 125–8, 182–3, 221–3, 255 n. 45 content conceptual vs. non-conceptual 13–15, 132 n. 81 demonstrative 27 n. 26, 36 n. 39, 116, 141–4, 187 representational 15–16, 20, 30 n. 31, 39, 60 n. 23, 63 sensory 38–44 Currie, Gregory propositional imagining 68 n. 36 on simulation 68 n. 37, 70 n. 41–2, 75–6, 252 deflationary accounts 24–5, 196 Descartes on deduction 174, 176–9 on epistemic security 123–4 on intuitus 163, 175 Meditations 173, 174 n. 59, 176 n. 68, 178 on rational perception 163–4, 173–85 Rules for the Direction of the Mind 163, 173–84 see also images Dretske, Fred on causal requirement on perception 111 n. 38, 148 on non-epistemic seeing 13 on secondary seeing 90 n. 7, 155 emotion and art 251–3, 261 and imagining 74 n. 48 and morality 192 and perception 134–5 see also ambivalence entitlement 17, 23 Evans, Gareth 36 n. 39, 141–5 expression in art 232–9, 243–8

Foucault 133 n. 82, 230 Frankfurt, Harry G. 199 n. 20, 208 Freud, Sigmund 205–7, 256 n. 46, 257 n. 47 Friedman, Michael 119, 121 Gifford, Sanford Robinson, Tappan Zee Illus. 233, 253–5 Goldie, Peter on definition of perception 39 on imagination and perception 110 n. 37 on moral perception 189 n. 2, 191, 194 n. 11, 216, 262 Goodman, Nelson 239 Gorky, Arshile, The Horns of the Landscape Cover Illus. 244 n. 24, 246 graphs 155–6, 184 n. 84 Hanfling, Oswald 251 Harman, Gilbert 20 hearing 32–5, 37 n. 41, 50, 57, 59–60, 66 n. 33 Husserl, Edmund 43–4, 56–7 images 63–77, 80, 95, 103, 112, 125 n. 68, 132, 135 n. 85, 155–6 Descartes 178 n. 74, 180–1, 227 n. 1, 248 n. 28 camera 142–5, 149 Plato 164 n. 41, 165, 169, 171, 172 n. 57 Wittgenstein 157–61 imagining actual vs. possible 48–58 Husserl 56–7 introspective awareness of 58–60, 87 n. 1 productive 129–30, 180 n. 76 reproductive 83, 129 n. 73 stipulative 68 unconscious 58–61, 98 n. 24 immediacy causal 19–20 experiential 20–2, 190–1 justificatory 22–5, 190–1 see also justification indeterminacy re art 244–6 re imagination 69–71 re perception 38–9 re persons 202–5, 209 re rules 55–6

INDEX

inference 22 n. 20, 23 n. 21, 27 n. 26, 51–2, 141 n. 2, 155 n. 27 and space of reasons 113, 116 intuition 43–4, 163–77, 195 justification self-evidence 25–9, 35–8, 56–7, 124, 152 externalist 17 n. 13, 23–4, 26–7, 51, 141 n. 3. 176 n. 71 see also immediacy Kahneman, Daniel 136 Kant, Immanuel on aesthetics 244–5, 252 on cause 98–100 on concepts and intuitions 115–17 on morality 214, 216–4 on objectivity 58 n. 18, 129 n. 73 on spontaneity 118–20 Kelly, Sean 34 n. 38, 53–4, 67 n. 34 knowledge "in a flash" 130–1, 135, 159–62 and perception 12–14, 188 mathematical 39, 40, 42, 158–60, 163, 171–2, 173–80 Kulvicki, John V. 227 n. 1, 229 n. 2, 232, 234 laws causal 91–3, 96–100 moral 217–23 realm of 117–22 Lehrer, Keith 26 McDowell, John on perceptual knowledge 12 n. 3, 16 n. 11, 26 n. 24 re deflationism 24 n. 23 re moral perception 110 n. 37, 188 n. 1, 189 n. 2, 194 n. 11 on space of reasons 112–22 re "re-enchantment" of nature 205 n. 29 McGinn, Colin 68 n. 36, 69–71, 93–4 n. 15 Martin, M. G. F. 32 n. 34 Merleau Ponty 53–4 Miller, Mitchell 168 n. 50 Miller, Richard 262 Moore, G. E. 188–9 n. 1–2, 194 motivation 133–6, 215–16 Murdoch, Iris 188–9, 224

283

Neisser, Ulric 65 Nietzsche 197 n. 16, 230, 261 Noe¨, Alva 30–2, 36 n. 40, 37 n. 41, 59 n. 22, 77 n. 55 Nussbaum, Martha Craven 189 n. 2, 211, 213–14, 217 O’Keeffe, Georgia, Two Figs, Illus. 226, 241–3, 255 objectivity in morality 213–14, 220–4 in perception 25–38 weaker vs stronger forms 192–4 see also justification (self-evidence) Peacocke, Christopher 92 n. 12, 132, 145 n. 9 perception, see knowledge, immediacy, justification, sensory content, objectivity physiology of perception 62–3 Piper, Adrian M. S. 250–1 Plato 164–73, 258 n. 48; see also images Ravenscroft, Ian 68 n. 36–7, 70 n. 41–2, 75–6, 252 n. 40 reasons as explanations 88–90 causal 90–100 constitutive 101–6 justificatory 106–12 space of 112–22 Sanford, Robinson Gifford Tappan Zee Illus. 233 schema 65–6, 184 n. 85 Scott, A. O. 260 seeing-as and seeing-in aesthetic perception 230 n. 2, 247–9, 255 vs. believing 103–4 vs. normal perception 76 n. 52 vs. thinking-as 48 n. 1 see also Wittgenstein self-evidence, see justification self-reflection 198, 201–2 Sellars, Wilfred 16 n. 11, 48 n. 1, 112–13, 117 sense modalities 32–6, 42, 66; see also hearing Serra, Richard 259 Shusterman, Richard 230–1 Siegel, Susanna 16 n. 12, 87 n. 1, 93 n. 15, 100 n. 29 simulation 68–9, 74–5, 136, 200 n. 22, 251–2 Smith, Anna Deavere 252–3

284

INDEX

Snowdon, Paul 16 n. 11, 29 n. 29, 36 n. 39, 40 n. 44, 148 n. 15 space egocentric 49–50, 142–55, 156 n. 28 and objectivity 11 n. 1, 33–8, 49–57 and sensory content 41–2 see also action (perspectives), aesthetic distance, graphs, sensory modalities Stecker, Robert 261 Stockhausen, Karlheinz 259–60 Strevens, Michael 87 syntheses of the imagination 56–8, 146, 207, 222–4, 238 n. 16, 244–6 Thomas, N.J. T. 63 n. 28, 65 transcendental arguments 98–100 truth and art 239 n. 17, 254, 255 n. 45 requirement for perception 15–16, 31, 54

uncanny, the 205–9, 257 n. 47 Walton, Kendall 252 Wittgenstein, Ludwig on following a rule 55, 71 on images 76, 93 n. 14, 117, 157–63 on “live” pictures 236 on perceiving persons 203–4 on seeing-as 130–1 on seeing and doing 135–6 see also images Wollheim, Richard on art and ambiguity 255 n. 46 on iconic memory 101 n. 30 on imagining another’s position 244, 252 n. 41 on seeing-in 229 n. 2, 247 n. 27 Young, Michael 75