Attention Is Cognitive Unison: An Essay in Philosophical Psychology

Attention Is Cognitive Unison

PHILOSOPHY OF MIND Series Editor David J. Chalmers, Australian National University

Thinking without Words José Luis Bermúdez The Conscious Brain Jesse Prinz Simulating Minds The Philosophy, Psychology, and Neuroscience of Mindreading Alvin I. Goldman Supersizing the Mind Embodiment, Action, and Cognitive Extension Andy Clark Perception, Hallucination, and Illusion William Fish Phenomenal Concepts and Phenomenal Knowledge New Essays on Consciousness and Physicalism Torin Alter and Sven Walter Phenomenal Intentionality George Graham, John Tienson, and Terry Horgan The Character of Consciousness David J. Chalmers The Senses Fiona Macpherson The Contents of Visual Experience Susanna Siegel Attention Is Cognitive Unison An Essay in Philosophical Psychology Christopher Mole

Attention Is Cognitive Unison An Essay in Philosophical Psychology Christopher Mole

1 2011

3 Oxford University Press, Inc., publishes works that further Oxford University’s objective of excellence in research, scholarship, and education. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2011 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 www.oup.com Oxford is a registered trademark of Oxford University Press. 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, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Mole, Christopher, 1978– Attention is cognitive unison : an essay in philosophical psychology / by Christopher Mole. p. cm. — (Philosophy of mind) Includes bibliographical references. ISBN 978-0-19-538452-9 (alk. paper) 1. Attention. 2. Philosophy of mind. I. Title. BD418.3.M64 2010 128’.2—dc22 2009053059

9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid-free paper

To my parents, with love and thanks.

Preface

There is a venerable tradition in philosophy, so venerable that even Descartes satirizes it, of presenting one’s work, not as a definitive statement, but as a contribution to an ongoing conversation. This book is not a definitive statement. Nor, however, is the conversation to which it contributes ongoing. There have been occasional rumblings, but the conversation about the metaphysics of the psychology of attention has been more or less dormant for the century since philosophy and psychology split. This is not because philosophers and psychologists have been ignoring one another. The two disciplines have been talking, but for the last century they have not been talking about attention. The silence has been detrimental to both. Philosophers, by ignoring the branch of psychology that has attention-related phenomena as its central explananda, have tended to operate with a false picture of the psychologists’ project. According to this false picture psychology gives us a theory in which cognition is thought of as a set of computations, realized somewhere in the frontal lobes, while the perceptual processes and the action-executing processes, located nearer to the back of the head, orchestrate the inputs and outputs for these frontal, cognition-constituting computations. As a picture of psychology’s theory of the mind this is misleading. Except when it borrows philosophy’s false picture for its self-conception, psychology does not take the cognitive aspects of the mind to be implemented, separately from perception and from action execution, in the poorly understood neural centers at the front of the brain. Certainly the front of the brain is doing cognitive work, but the bits of cognition for which psychology has well-worked out theories— the bits to do with attention and memory—are understood to be instinct with the perception and action processes at every stage. Psychologists, for their part, have been investigating attention without regard to philosophy, and so have spent the last century blunting ever sharper empirical tools against what is fundamentally a metaphysical problem. The psychologists mostly know that they have been doing this, but they know also that the underlying problem is somehow not theirs to address. The result is that a great deal of ingenious research succeeds in describing the ways in which various forms of attention behave, and succeeds in accounting for that behavior by reference to its basis in the brain, but nobody knows quite how all of this adds up to an explanation of attention. The literature is loaded with protestations, caveats, and complaints. We are told that ‘attention’ is a family resemblance term, or vi

Preface

vii

that it is ambiguous, or hopelessly folksy, or ill-defined, or nonreferring. But ‘attention’ remains the best name that we have for the subject matter of cognitive psychology’s central, defining research project. This book presents a single unified theory of attention, intended to apply to attention in all its forms. According to this theory, ‘attention’ is not a family resemblance term, nor is it ambiguous, or folksy, or illdefined, or nonreferring. Nor, however, is it explicable simply by identifying and describing the processes that constitute it. The theory says that the relationship of attention to the cognitive processes executed in various parts of the brain is analogous to the relationship between unison and the individual performances of the members of an orchestra. Just as there is no place in the orchestra where unison sits, so, according to this theory, there is no place in the brain where attention is located. A similar point applies to particular cognitive activities: Just as there are no particular melodies that all instances of unison involve, so there are no particular patterns of neural activity that we should expect to find in all instances of attention. The fact that a person is paying attention does not, by itself, entail anything specific about which cognitive processes are taking place or where. Nor is there an entailment running in the other direction. If attention is cognitive unison, then facts about specific cognitive processes do not, by themselves, entail facts about attention. Just as the execution of a particular activity on the part of a musician may be unison constituting on one occasion and not on another, depending on how it is related to the things that his fellow musicians are doing, so a cognitive process that plays a role in attention in one context may, in a different context, be instantiated without attention. In this way the cognitive unison theory entails that no particular cognitive activity is sufficient for attention, or even a very reliable indicator of it. The contribution that this theory is intended to make to our understanding of attention is a contribution to our understanding of its metaphysics. It is the metaphysical relationship between attention and its constituent processes that psychologists have failed to get clear on, and it is the failure to get clear on this that has led to confusion about the nature of attention and about the way in which its explanation is to be accomplished. There is more than one reason why metaphysical theories are a hard sell when the audience is composed of psychologists. The theme of metaphysics avoidance recurs throughout psychology’s history as a science, often with some justification. Even the most enthusiastic practitioners of metaphysics would admit that the period when psychology broke from philosophy was a period in which metaphysics had got itself into a mess. It is therefore understandable that psychologists have been steering clear of metaphysics ever since. But the happy fact is that over the course of the twentieth century the mess that metaphysics had got into was sorted

v iii

Preface

out. Metaphysics, as we have it now, gives us tools that can usefully be deployed on foundational issues in psychology. The development of these tools has, perhaps inevitably, upped the entry requirements for the metaphysical debates. Those debates now tend to be incomprehensible to those not familiar with the issues and with the accompanying vocabulary. And by now few psychologists are interested enough in what has been going on to think that the entry price to the metaphysical debates is worth paying. The result is that most psychologists think that, once we have all agree to be physicalists, there are no interesting metaphysical questions about mental entities to be asked. ‘Metaphysical’, they think, is just a show-off’s word for ‘spooky’. I would like to think that these misconceptions will have been abandoned by the time the reader gets to chapter 7 of this book, but I am also aware that the chances of her ever getting there can be increased by saying something now. ‘Metaphysical’ does not mean ‘spooky’. It is not a word to say in a hushed voice when the door is shut. By ‘metaphysical’ all that I mean here is ‘basic’ in the sense of ‘basic even from the point of view of physics’. The physicist, most of the time, can take it for granted that there are distinctions between existence and nonexistence, between objects and events, intrinsic and extrinsic properties, actuality and mere possibility, and so on. The job of the metaphysician, insofar as we are concerned with him here, is to give an account of what these basic categories amount to. My contention is that questions pertaining to such basic categories arise in psychology too, and that they matter for the business of building explanations. This has the consequence that the psychologist cannot avoid questions about the metaphysical category of the items that she seeks to explain. There is much in the cognitive unison theory that creates a nuisance for the project of studying attention with the methods characteristic of current cognitive psychology, but it would be quite wrong to think that the theory is intended as part of an attack on that science. The work of cognitive psychologists is indispensably valuable, and it informs the argument of this book at almost every step. The lesson for the cognitive psychologists is not that their work is no good, but that there is a need to deploy the resources of analytic metaphysics in order to understand what that work tells us. For some psychologists this will come as unwelcome news, but it cannot be avoided. The giant leaps forward that were taken by experimental psychology and by analytic metaphysics in the twentieth century were leaps in one another’s direction: Psychology stepped out into territory where metaphysical questions were unavoidable, and, at much the same time, philosophy got straight on the matter of how to address those questions. The cognitive unison theory of attention was first presented in the dissertation that I defended at Princeton University in September 2005. My fellow graduate students at Princeton, both in philosophy and out of

Preface

ix

it, were unfailingly inspiring and supportive throughout my time there. Huge sections of this book contain ideas that I could only have worked out in their company. I owe huge thanks to every one of them. The dissertation was supervised by Sean Kelly, who patiently guided me out of the blind alleys that I repeatedly charged down. Several other faculty members also read and commented on drafts of chapters and on various half-baked thoughts. They, too, made huge differences, and without them there are several confusions from which I would never have escaped. I am very grateful to them. My time at Princeton was made possible by the William Alexander Fleet Fellowship. I owe an enormous debt of gratitude to Miss Julia Fleet for her patronage. I am sad that she died without seeing this book. Between 2005 and 2006 I had the good fortune to be a postdoctoral fellow in the Philosophy-Neuroscience-Psychology (PNP) Program at Washington University in Saint Louis. There, again, I benefited from innumerable conversations with an astonishingly collegial and stimulating group of faculty and students. The most important stages of the theory’s development into a book took place between 2006 and 2009, during which time I was teaching in the School of Philosophy and in the Cognitive Science Programme at University College Dublin. Various sections of the book were tried out on a diverse group of students, many of whom took the time to provide me with valuable feedback. I am grateful to all of them. My philosophical life in Dublin would have been greatly impoverished without the friendship and philosophical stimulation provided by my colleagues in the Aporo research network. My thanks to them are especially heartfelt. In the summer of 2009 I spent a very productive month in the almost perfect research conditions afforded by the Australian National University’s Research School of Social Sciences. I am hugely grateful for that opportunity and for the many helpful conversations I enjoyed there. The final stage of the book’s preparation was completed at the University of British Columbia, where the support of my new colleagues has been faultless. I have also learned a lot from audiences at the various venues, in conferences, job talks, and colloquia, where various prototypes of this book’s arguments have been presented. Thanks to all of them too. This book is a rather impersonal, imperfect, and academic work. It seems a little inappropriate to give it a personal dedication, but I do dedicate it: to my parents—with love and thanks.

This page intentionally left blank

Contents

1 2 3 4 5 6 7

Highlights of a Difficult History The Underlying Metaphysical Issue Rejecting the Process-First View Cognitive Unison The Causal Life of Attention Consequences for Cognitive Psychology Philosophical Work for the Theory of Attention

3 24 35 47 88 117 136 171 174 183

Notes Bibliography Index

xi

This page intentionally left blank

Attention Is Cognitive Unison

This page intentionally left blank

1 Highlights of a Difficult History

1.1 THE PRELIMINARY IDENTIFICATION OF OUR TOPIC There is one remark from William James’s The Principles of Psychology that all those who study attention can quote by heart. It is the remark that ‘everyone knows what attention is’ (James 1890: 381). This remark does not, by itself, tell us very much. The quoting of it nonetheless serves two purposes, thanks to which it continues to be quoted with a frequency that nothing else in James’s masterwork can match. The first purpose is to serve as evidence of attention’s importance. Everyone knows what attention is because the role played by attention is so central to cognition that nobody could possibly overlook it. The second purpose is to stand in lieu of a definition. If everyone knows what attention is, then there is no need for psychologists interested in attention to start out by identifying which phenomenon is under investigation, and no cause for philosophical complaint when they fail to do so. James, of course, had more to say. Having told us that everyone knows what attention is, he goes on to tell us that what everyone knows is that attention is ‘the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible trains of thought’ (James 1890: 381). Philosophical controversies begin when we try to say anything more than this. The best tactic yet discovered for postponing those controversies, and so for allowing psychologists to get on with the kind of work that their methods are equipped for, is to quote James’s bold remark and to leave things at that. In this book we shall be facing up to the much postponed philosophical controversies about the nature of attention. But, as a preliminary, pretheoretical identification of our subject matter, it is still James’s remark that provides us with our best starting place. We can expand upon that remark like this: For minds like ours, in environments like ours, more than one sequence of mental states is possible. We end up with the train of thought that we actually have partly by chance, but partly because certain things catch our attention, and partly because we direct our attention onto certain things. A theory of attention is an attempt to give an account of this nonrandomness in our coming to have

3

4

Attention Is Cognitive Unison

the train of thoughts that we in fact end up with. It is an attempt to explain the selectivity of our mental engagement with the world. That is what psychologists working on attention seek to explain. It is what this book is about. James’s remark that ‘everyone knows what attention is’ has an air of unanswerability about it, and an air of finality. It suggests that a more philosophically ambitious analysis of attention, such as we shall be attempting in chapter 4, would be superfluous. But this air of unanswerability should be looked on with suspicion. It is a signature trait of the bluffer. James’s remark was, I think, a piece of bluff. When The Principles of Psychology was written it was far from clear that anybody knew what attention was. Karl Groos admitted as much, not very long after The Principles appeared: ‘To the question, “What is Attention?” ’, he wrote, ‘there is not only no generally recognized answer, but the different attempts at a solution even diverge in the most disturbing manner’ (Groos 1896, cited by Tsotsos, Itti, and Rees 2005: xxiv). James himself provides evidence that there was some confusion among his peers as to which phenomena were, and which were not, attentioninvolving. His chapter on attention includes, under the title ‘To How Many Things Can We Attend at Once?’, a discussion of experiments into the ability that we now call ‘subitizing’ (that is, the ability to see straightaway whether one, two, or three stimuli are present, without having to go through any procedure of counting them). But James goes on to say that ‘it is obvious that such observations decide nothing at all about attention, properly so called’ (384). Subitizing was by no means the only phenomenon whose relation to attention was controversial at the time when James was writing. Disagreements about which phenomena should be taken as the paradigmatically attention-involving ones were widespread at that time, as they were in the decades before and after. In this chapter I shall be arguing that, underpinning these disagreements about attention’s explanatory remit, there was a metaphysical issue. This issue is of the first importance for the attempt to provide an explanatory account of what attention really is. Various historical factors have conspired to make it an issue that has rarely been explicitly discussed. One reason for revisiting the history of this issue is to show that metaphysical matters need not be arcane ones. They are woven into our foundational assumptions about the way in which mental phenomena should be explained, and woven, also, into our assumptions about the methods by which those phenomena can be examined. The research program of cognitive psychology is so fruitful and so fast-moving that it is easy for experimental psychologists to be tempted by the idea that a priori philosophical inquiry into the metaphysical underpinnings of the phenomena that they investigate is a futile pursuit, or, at least, that it is a pursuit long made obsolete by the advent of their empirical methods. The history that we shall be revisiting here shows this attitude to be mistaken.

Highlights of a Difficult History

5

The metaphysical question that this book attempts to articulate, and to answer, is a question that has to be settled at the first stage in the construction of any explanatory theory of attention. It is a question that psychology’s founding fathers never quite squared up to. The current methods of empirical psychology do not enable us to dodge that question. On the contrary, it is our attempt to understand the explanatory importance of applying those methods that obliges us to take sides on it.

1.2 THREE APPROACHES Although it was not the first work of psychology to discuss the topic of attention at length, James’s Principles came at a time that was crucial to psychology’s development as a science, and it was unusually influential. At the time when James was writing there were at least three different ways of approaching the topic of attention, each of which had its own advocates. There were those who took the most fundamental thing about attention to be its involvement in willed action, and in the coordination of bodily movements. There were those who took the most fundamental thing about attention to be its involvement in perception, and the achievement of ‘sensory clearness’. And there were those who took the most fundamental thing about attention to be its involvement in the direction of thinking, and in the coordination of one’s internal monologue. We can produce examples that show each of these three approaches to have some intuitive appeal. On behalf of the first approach we can produce examples in which the coordination and execution of action seem to be a paradigm case of attention. One obvious example is the attention required by gymnasts. A less obvious example is the attention that one pays when trying to move a partially numb limb. Both cases involve attention, but in neither case is there a perceived object the details of which we are trying to bring into view. The attention given to such cases therefore seems to be attentive action.1 On behalf of the second approach we can produce examples in which sensory focalization seems to be a paradigm case of attention. The attention paid by the piano tuner before making some fine adjustment to the pitch of a string gives us an example along these lines. Before making his adjustment the piano tuner is not performing any action. His task is the purely perceptual one of listening to the pitch. The attention given to such a task therefore seems to be attentive perception. On behalf of the third approach we can produce examples in which pursuing a train of thought, without perceptual input or behavioral output, seems to be a paradigm case of attention. An example along these lines is the attention that we pay when performing mental arithmetic. Mental arithmetic certainly requires attention, but it may involve no

6

Attention Is Cognitive Unison

attempt to perceive anything, nor any attempt to execute an action. The attention given in such a case therefore seems to be attentive thinking. Such examples show that, among the paradigm cases that a theory of attention needs to account for, there are cases of attention in action, cases of attention in perception, and cases of attention in cogitation. The three approaches that these examples make plausible were each developed in some detail around the time when James was writing. The perception-centered approach was endorsed by several psychologists in Germany and in the United States (for a contemporary review, see Münsterberg and Kozaki 1894). It was influentially advocated by E. B. Titchener in his 1908 Lectures on the Elementary Psychology of Feeling and Attention, and again in his 1910 article ‘Attention as Sensory Clearness’. The action-centered approach was taken in the 1888 edition of Alexander Bain’s The Emotions and the Will (although in earlier editions Bain had been less sure of it). And the third approach—the approach centered on the role of attention in the coordination of reasoning—can be found in work from this period by G. F. Stout. Stout presents the view in 1891, in the introduction to an article on ‘Apperception and the Movement of Attention’, where he writes that intellectual ends are attained by an appropriate combination of movements of attention, just as practical ends are attained by an appropriate combination of movements of the body. If, therefore, we are to explain the process of thinking, we must clearly determine the nature of active Attention. (25)

It would be misleading to suggest that the plausibility of these three approaches led to a period of sustained normal scientific debate between the advocates of three well-articulated theories. What actually happened was that nobody had much of an idea about how the question that divided these three approaches might be settled. By the beginning of the twentieth century, the unsatisfactoriness of this theoretical situation had become a cause for occasional complaint. W. B. Pillsbury began his 1908 book on attention by remarking that the contemporary ‘attention theories’ were in a ‘chaotic condition’ (ix). And F. H. Bradley, in a paper from 1902, complained that some of the theories that were being offered as attempts to explain attention had lost touch with their purported subject matter entirely. He writes that ‘In the case of attention the abuse [of words] has even been carried to such a point that attention has been used to include and cover what everyone does and must call a state of inattention’ (1).

1.3 BRADLEY’S PROTEST Bradley did not tell us whose abuse of words he was objecting to when he complained that the word ‘attention’ had been used to cover ‘what everyone does and must call a state of inattention’, but it is a safe bet that

Highlights of a Difficult History

7

James Ward was among those whom he had in mind. In an article written in 1885, surveying the entire field of psychology for the ninth edition of the Encyclopedia Britannica, Ward claimed that, for scientific purposes, ‘attention’ should be construed very broadly indeed. This claim was one that Ward continued to advance in his 1918 book Psychological Principles. In chapter 2 of that book Ward recommends ‘extending the denotation of this term [‘attention’] so as to include even what we ordinarily call inattention’ (49). On the face of it Ward’s suggestion looks to be recipe for confusion, and for exactly the abuse of words that Bradley was complaining about. If we allow ‘attention’ to be used in a way that includes ‘what we ordinarily call inattention’ then it seems we allow the term to become uselessly broad. That may be a little unfair to Ward. He was not recommending that ‘attention’ be used in an entirely unconstrained way. (He was, I think, trying to advance the view that the word ‘attention’, thanks to its verbal form, ‘attend’, is better suited than the word ‘consciousness’ to play the role that the latter word has had assigned to it.) But, whether or not Ward’s view was ultimately a sensible one, we can see why Bradley thought that such liberal use of the word ‘attention’ was worth complaining about. If we use the word ‘attention’ in such a way that all conscious states involve attention to some degree, even those idle, tuned-out states that we naturally describe as instances of inattention, then we lose our grip on the selectivity and directedness that our preliminary characterization of attention took to be of its essence. Bradley’s reluctance, in 1902, to name the theorists whose unconstrained use of the word ‘attention’ he objected to may have been partly owing to the fact that Bradley himself, in 1886, had advocated a view of attention that was notable for the slightness of the constraints that it imposed. The view found in Bradley’s earlier work is that any function whatever of the body or the mind will be active attention if it is prompted by an interest and brings about the result of our engrossment with its product. (316)

That is clearly not a view that puts much constraint on the things that can be called ‘attention’. It may be no surprise, therefore, that the later Bradley, when complaining that the use of ‘attention’ had become too unconstrained, was happy for his earlier view to be neglected. A footnote to his 1902 paper mentions that the view defended there departs from his 1886 view but tells us that ‘it seems not worth while to ask in detail how much’ (2). Bradley’s earlier view was something that he never developed in any detail. The ‘chief object’ that he claimed for the paper in which that view is found was not to elucidate his positive theory. It was ‘to record a kind of protest’ (Bradley 1886: 21). The remark quoted above, about interest and engrossment, is more or less all the elucidation that the positive theory gets.

8

Attention Is Cognitive Unison

It will be one of my claims in what follows that the approach taken by Bradley’s early positive theory is, in fact, well worth pursuing, but since Bradley’s ‘chief object’ was not the development of that theory but the recording of ‘a kind of protest’, we need to know, if we are to understand him, what it was that he was protesting against. Several aspects of the contemporary treatments of attention seem to have been objectionable to him: I observe a tendency to break up the life of the soul, to divide it into active and passive factors, or to suppose a passive beginning with a supervening activity, the latter by some identified with an irreducible act of attention. I believe this tendency to be a serious obstacle to psychology, and there is another tendency not less injurious. Attention may be given such a position that the reader cannot tell if it is primary or derivative, or, if primary, whether it is an original element or something that supervenes; or, again, whether it is one of a class of activities, or one function exerted on different objects. (Bradley 1886: 305–306)

The first part of this remark—the bit about ‘a tendency to break up the life of the soul’—is certainly evocative of what might have been objectionable in the conceptual schematizing of Bradley’s contemporaries. The second part of the remark is rather too jargonized to be illuminating. What does Bradley mean by being ‘primary’, and how is the property of being primary meant to contrast with the property of being ‘an original element’? What is the notion of ‘supervenience’ that is in play here? (It clearly is not the same as the notion that metaphysicians currently denote with that word.) One suspects that, in writing this passage, Bradley was rather more concerned with illustrating the reigning confusion than with showing us a way out of it. We can bring Bradley’s positive theory more clearly into focus, together with his point of protest, by contrasting his explanatory approach not only with the approach of Ward, for whom ‘attention’ and ‘consciousness’ were virtually synonymous, but also with the approach taken by James in The Principles of Psychology. The approach that James advocated was one version of the approach that Bradley objected to, and James himself devotes a good part of his chapter on attention to what he takes to be his point of dispute with Bradley. It will be helpful, therefore, to look at James’s approach before returning to get a clearer view of Bradley’s protest, and of the theoretical position that motivated it.

1.4 JAMES’S DISJUNCTIVE THEORY At the time when he wrote The Principles of Psychology James was sympathetic to something rather like the perception-centered approach to attention. Early in chapter 11 of Principles, shortly after the remark about everyone knowing what attention is, James writes that ‘sensory

Highlights of a Difficult History

9

focalization’ is ‘of [attention’s] essence’ (381–382). Later in that chapter he attempts to provide an account of the ‘intimate nature’ of attention by identifying the particular processes that constitute attention’s instances. Two processes are identified: The accommodation or adjustment of the sensory organs; and The anticipatory preparation from within of the ideational centres concerned with the object to which attention is paid. (411)

The first of these processes is a familiar one. By ‘the accommodation or adjustment of the sensory organs’ James means such things as pointing one’s ears in the right direction, bringing one’s eyes into focus, taking a sniff, and so on. The second of the processes is described by James in somewhat less familiar terms, but his point is again quite straightforward. ‘Anticipatory preparation’ of ‘ideational centres’ is simply James’s way of characterizing imagination. His claim here is that some instances of attention consist in imagining the things one is attending to, or looking for. Taking both of these processes together, we find that James is advocating a disjunctive theory according to which attention is constituted either by the processes underpinning imagination, or else by the processes of sensory orienting. James presents this theory of attention as if it were a theory that enjoys a good deal of intuitive appeal. These ‘two physiological processes’, he writes, ‘immediately suggest themselves as possibly forming in combination a complete reply’ to the question of ‘The Intimate Nature of the Attentive Process’ (1890: 411). In suggesting that this disjunctive analysis of ‘the attentive process’ immediately suggests itself as complete James goes too far since there are instances of attention that seem to involve neither imagination nor sense organ orientation. But if James’s claim had been only that imagination and sense organ orientation can each play a central role in some instances of attention, then his claim to intuitive appeal would ring true. We can find clear examples that illustrate what James must have had in mind here, just as we could find the examples of the attentive gymnast, attentive piano tuner, and attentive mental arithmetic solver to illustrate the action-centered, perception-centered, and reasoning-centered views from the works of Bain, Titchener, and Stout. Examples illustrating the plausibility of the first disjunct of James’s claim—the claim that sense-organ adjustment can be a way of attending—are relatively commonplace: If, sitting in the audience at the symphony, I want now to attend to the strings, having previously been attending to the brass, then it seems that an appropriate reorientation of my sense organs is one way to go about doing it. This ‘overt attention’, as more recent discussions have dubbed it, is the easy case. For the second disjunct of James’s claim—the claim that preparatory imagination can be attention-constituting—the clearest examples are not quite as familiar. But once we have seen them, we can see that they are also strong in their intuitive appeal. The example mentioned by James himself is the

10

Attention Is Cognitive Unison

following: When listening to a note played on the piano one can hear a number of harmonic overtones in addition to the fundamental frequencies of the note played. Suppose you want to listen for the third overtone of a particular note. (If you are a piano tuner then this will be the sort of thing that you have to do routinely.) One way to go about attending to the note so as to hear the third overtone is by first listening to the pitch at which that overtone sounds, and then imaginatively keeping that pitch in mind while listening as the note is struck. In this case the imagining of the pitch and the attention that one pays in listening for the overtone seem to be the same thing. James took this example from Hermann von Helmholtz’s 1870 treatise on sound perception. For those of us who are not piano tuners it may not be an everyday example of attention, but such shifts in what one is listening for in a single auditory stimulus are familiar enough. They are certainly possible, and they are certainly attention-involving. We need our theory of attention to allow for such shifts. The part of James’s theory that is concerned with imagination is an attempt to do so. Some readers will be sure to have a qualm at this point, since I have just made a psychological claim on the basis of introspection and informal observation. Many psychologists disapprove of introspection and of informal observation. They have perfectly legitimate reasons for doing so. Ideally, one would like to puts James’s and Helmholtz’s ideas about attention and preparatory imaging to a more rigorous empirical test. I know of no studies that do exactly that, but Harold Pashler, in his 1997 textbook on attention, reports a study in which subjects were required to form a visual mental image of an item and then, forgetting about that image, to perform a task requiring them to detect the occurrence of a digit in a rapidly presented sequence of pictures (249). The results that Pashler observed comport nicely with what James suggested. When presented with a sequence of pictures it seems that, irrespective of the demands of their task, subjects do give attention to the pictures of recently imagined items, just as James would have predicted. The evidence that attention is paid to these items comes from the fact that they elicit an ‘attentional blink’: If the digit that the subject has been looking out for comes shortly after the imagined image, then subjects are liable to overlook it. In Pashler’s experiments, then, prior imagination of a thing does seem to bring it about that the thing is attended, much as James thought. Whatever their empirical credentials, the examples of switching attention to a different instrument in the symphony and switching attention to a different aspect of a note played on the piano give us some insight into James’s disjunctive theory, and into his reasons for finding it plausible. With that theory before us, as an example of the sort of thing that F. H. Bradley was protesting against, we can now begin to see the source of Bradley’s dissatisfaction.

Highlights of a Difficult History

11

1.5 THE SOURCE OF BRADLEY’S DISSATISFACTION We have seen that James’s claim about attention’s links to orienting and to imagining is plausible enough. What was it, then, that Bradley found in approaches such as James’s that he wanted to protest about? The problem that Bradley had with the theorists of attention who were writing at the end of the nineteenth century was not that they had failed to identify any psychological phenomena in which attention was clearly displayed. That was very far from being the problem. James, we have seen, identified phenomena in which attention seems clearly to be displayed. So did Helmholtz, Bain, Titchener, and Stout. All of their proposals seemed plausible (and they continue to seem plausible, even in the face of more recently gathered evidence). The source of Bradley’s dissatisfaction with his contemporaries was not that they had identified no clear examples of attention-involving phenomena, but that there were too many such examples, and the various examples of attention-involving phenomena that these different theorists had identified seem not to have much in common as far as their constituent processes go. Nor did any one of these examples seem to have explanatory priority over the others. It therefore seemed that the contemporary accounts of the processes that are involved in particular instances of attention did little to help us in formulating a generalizable explanatory theory of what attention is. And the project of identifying any particular process as the ‘special activity of attention’ therefore seemed wrongheaded. When Bradley titled his 1886 essay ‘Is There Any Special Activity of Attention?’ it was in order to urge that the question be answered in the negative. To readers familiar with the recent empirical literature on attention the diversity among the processes that Bradley’s various contemporaries took to be involved in paradigm cases of attention may sound like a familiar fact, and the metaphysical lesson that Bradley took from it—that there is no ‘special activity of attention’—may be thought to be something that we now find obvious. Recent discussions of attention frequently note the plurality and heterogeneity among the paradigm cases of attention, and urge us to accept that our theories of attention must be corresponding heterogeneous (for an influential example of such reasoning, see Parasuraman 1998). Those who attempt summaries of the current debates typically preface their discussions with a remark to the effect that ‘there may be numerous component processes behind our attentional abilities’ (Driver 1998: 298), or else they claim that ‘attention’ is ‘an umbrella term for a variety of psychological phenomena’ (Styles 2006: 1) and take this as showing that ‘to try to define attention as a unitary concept is not possible and to do so would be misleading’ (Styles 2006: 9). We shall return to these ideas from the current literature in chapter 6. The point to note presently is just that their familiarity should not be allowed to mislead us as to the point that Bradley

12

Attention Is Cognitive Unison

was making. The lessons that current theorists draw from the diversity among attention’s instances are not the lesson that Bradley drew. When Bradley objected to views of attention along the lines of James’s disjunctive two-process view it was not because he thought that two was not a high enough number. He was not claiming simply that there were many other varieties of attention, in addition to those that theorists such as James had taken account of in talking about sensory orienting processes and the processes of preparatory imagination. In fact, Bradley did not think that there was a great diversity among the attention constituting processes. When it comes to naming attention-constituting processes, Bradley, like James, provides a list of just two: ‘redintegration’ (that is, the following of an association between two ideas) and ‘blending’ (the forming of such an association). He writes that ‘in attention there is either no activity at all beyond the common processes of redintegration and blending, or, if the activity exists, itself is not attention’ (1886: 316). Bradley’s reason for rejecting the project of identifying attentionconstituting processes was not that he thought, as current psychologists do, that the attention constituting processes are too many and too various. What was it, then, that Bradley was objecting to? An analogy may help us to get clearer on that, and to see how Bradley’s objection differs from the current literature’s rejections of one-process views of attention. For the purposes of this analogy, imagine that we have come across a group of social scientists who are attempting to characterize the underlying nature of the process that constitutes employment. We can imagine one of these scientists, an analogue of James, observing that lots of instances of employment involve manual labor and concluding, on that basis, that the processes responsible for arm and hand movements are the underlying processes of employment. This scientist might then observe that there are lots of other instances of employment that do not involve manual labor but that do involve clerical work. He might conclude that in those cases the processes of producing and arranging documents are the employment processes. Having seen that these two theories each seem right for their own range of cases, he might become a disjunctivist. His theory of the ‘intimate nature’ of the employment process will be that it is either the process of arm and hand orientation, or the process of producing and arranging documents. Meanwhile, some other theorist of employment (analogous to Titchener) might notice that some instances of employment involve patrolling and being on the look out for certain things. He might propose a theory according to which the ‘intimate nature of the employment process’ involves perceptual processing. The analogue of James and the analogue of Titchener might take themselves to be engaged in a serious disagreement about the fundamental nature of the employment process. In fact, of course, they would both be mistaken. The point to take particular care over here is this: The scientists we are imagining would not be making a mistake about which processes

Highlights of a Difficult History

13

constitute paradigm cases of employment. Nor would they simply be mistaken about how many such processes there are. The mistake is a much more fundamental one. It is a mistake about the way in which an attempt to explain employment should proceed. In taking it that the explanation should proceed by identifying the processes that constitute employment these imagined scientists would be making a metaphysical mistake: a mistake about the sort of thing that employment is. It bears emphasis that the thing we find ourselves wanting to say about the imagined project of identifying the process of employment is not the thing that the current textbooks say about attempts to identify the process of attention. We do not give a correct diagnosis of what has gone wrong in the imagined debate about employment by saying that ‘employment’ is an umbrella term, or that it is a term for which no definition can be given, or that analogue-James and analogue-Titchener each had a part of the truth. We do not say that the theories of analogue-James and analogue-Titchener should be reconstrued as contributions to a heterogeneous, family-resemblance–based theory of employment. What we say about analogue-James and analogue-Titchener is what Bradley says of his contemporaries’ attempts to understand attention: It is all wrongheaded. For the purposes of explanation there is no point in starting a catalog of the processes that constitute employment. Any process whatever, of the body or the mind, will be a case of employment, as long as it is done in the right way and for the right reasons. A theory of employment should specify the way and the reasons. It should not catalog the constituent processes. Bradley’s complaint against a project like James’s was not that the wrong processes had been identified as the attention constitutors, or that too few processes had been identified. Bradley did not deny James’s claim that imagination and sense-organ accommodation had important roles to play in many instances of attention (Bradley 1886: 307). Nor, when discussing Bain’s claims about motor coordination, did Bradley take issue with the idea that motor coordination can be attention constituting. Nor, one imagines, would he have objected to Stout and Titchener when they claimed that there are instances of attention in which cogitation and sensory clearness are crucial. Bradley’s protest was a metaphysical one. He was objecting to the idea that the identification of these processes contributed even a part of the explanation of what attention is. His own view was that there is no explanatory value in starting a catalog of the attention-constituting processes because any process could constitute attention, as long as certain conditions were met (conditions that his remarks about ‘engrossment’ and ‘interest’ gestured at but did not adequately specify). This is what he means when he writes: We have found nothing in attention that is not derivative, nothing which could justify our placing it among the primary elements of mind. In attention there is either no activity at all beyond the common

14

Attention Is Cognitive Unison

processes of redintegration and blending, or, if the activity exists, itself is not attention. Any function whatever of the body or the mind will be active attention if it is prompted by an interest and brings about the result of our engrossment with its product. There is no primary act of attention, there is no specific act of attention, there is no one kind of act of attention at all. (Bradley 1886: 316)

James’s question about the ‘intimate nature of the attentive process’ was, in Bradley’s view, a question with a false metaphysical presupposition, not because it supposed that there was just one or two such process, but because it supposed that attention is a matter of executing particular processes at all. The metaphysical point of Bradley’s protest is worth laboring, not only because it is a point that is easily mistaken for the less radical claims that we find in the current literature, but also because it is a point that gets lost almost as soon as Bradley makes it. Bradley himself, although he was no slouch when it came to making metaphysical points elsewhere, never spelled the point out very clearly, partly because, as we have seen, he was more concerned with registering a ‘kind of protest’ than with articulating his positive view, and partly because, as we have also seen, he changed his mind before that positive view was ever brought clearly into focus. Nor did anybody else manage to give a very clear articulation of the metaphysical issue that separated Bradley from James, and from his other contemporaries. The failure to articulate this issue was partly owing to the fact that the philosopher/psychologists of this period lacked the necessary metaphysical vocabulary. Works from this time frequently display a proliferation of jargon, corresponding to the points where different authors take differing metaphysical positions, and this leads to a lack of clarity about what needed to be explained and how. An example is William Cyples’s 1880 book, An Inquiry into the Process of Human Experience, in which Cyples attempts to define ‘organization’ as ‘the interhappening of structural statics with related dynamical activities’ (10). Such definitions earned Cyples the scorn of his reviewers (Sully 1880), but they were a reflection of a widespread uncertainty about which of the concepts employed in the course of stating psychological theories were in need of clarification and about how such clarifications were to be given. Psychology did not yet have an established policy about what a psychological explanation should look like. It lacked an established vocabulary for articulating the metaphysical issues that such explanatory questions depend on. This lack of an established vocabulary in which to frame psychology’s underlying metaphysical issues prevented Bradley and James from giving a clear account of the metaphysical point about which they were disagreeing. Bradley, as we have seen, frames the issue in terms of ‘special activities’, ‘original elements’, ‘supervenient activities’, and ‘active and passive factors’ (Bradley 1886: 306). James, in contrast, frames the issue as a dispute between a ‘cause theory’ and an ‘effect theory’

Highlights of a Difficult History

15

(James 1890: 424). James’s presentation of the issue seems to be the more straightforward one, but James also tells us that settling his dispute with Bradley would require us to take a stand on ‘one of those central psychological mysteries which part the schools’: the mystery of whether the attention constituting activity was caused ‘by other brain-cells, or by some spiritual force’ (423). By the end of his discussion, James is characterizing the issue that separates him from Bradley as a dispute between ‘believers in mechanism’ and ‘believers in a spiritual force’ (429). Even without James bringing ‘spiritual forces’ into the discussion, the debate here seems fraught with confusion. Today we clearly distinguish claims about causation from claims about supervenience. By our lights, then, James and Bradley seem to have been talking past each other. James took Bradley to be denying a claim about causation. Bradley took himself to be denying a claim about supervenience and about ‘original elements’. But Bradley and James cannot just have been talking past each other. There was a genuine disagreement between them: James, as we have seen, thought that attention is the sort of thing that can be accounted for by identifying the ‘intimate nature’ of the processes that constitute it. Bradley thought that, for the purposes of explanation, the question of which processes constitute attention was a wrongheaded one. There is a genuine metaphysical question at stake here, a question about the sort of thing attention is and, correspondingly, about the sort of explanation that should be given of it, but James and Bradley never quite put their finger on the metaphysical point that lies at the heart of the matter. Because of Bradley’s abandonment of his side in the debate, and because of James’s misconstrual of that debate’s substance as being tied up with an issue about spiritual forces, the psychologist’s project of accounting for attention got under way without squaring up to the fundamental question of whether the metaphysics of attention is such that it can be explained by the identification of its constituent process, or whether, as in the case of employment, there is a fundamental mistake in the attempt to give such an explanation.

1.6 BEHAVIORISM AND AFTER In the opening decades of the twentieth century psychology’s desire to dodge metaphysical questions was matched by an enthusiasm among philosophers for approaches that attempted to show how such a dodge could be legitimated. Logical positivism’s appeal in this period was owing, in part, to its promise to do away with large swathes of metaphysics. It attempted to do so by claiming that metaphysical questions, like all questions not amenable to resolution by straightforward observation, were literally meaningless. The behaviorism that dominated psychology in the first half of the twentieth century was the result of applying this logical positivist criterion for meaningfulness to questions about the mind.

16

Attention Is Cognitive Unison

The dominance of behaviorism in this period resulted in attention being demoted from its central place in psychology. As a result, the metaphysical/explanatory issue on which Bradley and James had been divided became all but invisible. This was not because the behaviorists had nothing at all to say about attention. John Dashiell’s 1927 Fundamentals of Objective Psychology offered a behaviorist account of attention ‘as a form of posturing’ (284). And, some decades before behaviorism came to be established as a school of thought, Théodule Ribot’s 1889 book, Psychologie de l’attention, had offered a strikingly behaviorist approach to attention, which Ribot asserts with characteristic rhetorical flamboyance: Are the movements of the face, the body, and the limbs and the respiratory modifications that accompany attention, simple effects, outward marks, as is usually supposed? Or are they, on the contrary, the necessary conditions, the constituent elements, the indispensable factors of attention? Without hesitation we accept the second thesis. (25)

Despite the boldness of Ribot’s proto-behaviorist treatment of attention, the attempts, once behaviorism got under way, to identify a set of movements, postures, or respiratory modifications as essential to attention was generally recognized to be hopeless. As Gilbert Ryle notes, it is not only attention that resists simple behaviorist analysis but also the class of ‘heed concepts’ more generally: When a man is described as driving carefully, whistling with concentration or eating absent-mindedly the special character of his activity seems to elude the observer, the camera and the Dictaphone. Perhaps knitted brows, taciturnity and fixity of gaze may be evidence of intentness; but these can be simulated, or they can be purely habitual. (1949: 133)

The problem for the behaviorist is that attention may be given to all sorts of stimuli, and all sorts of responses may be performed attentively. For that reason attention is exactly the sort of thing that a behaviorist theory, given in the form of rules for stimulus/response mapping, will always struggle to accommodate. Most of the behaviorists who set the agenda for psychology in the first decades of the twentieth century were therefore happy for attention to drop out of view. Behaviorism ceased to be the dominant paradigm in psychology over the course of the 1950s. Attention returned to the psychologist’s explanatory agenda over the same period. These two tendencies were intimately linked. The Second World War had lent a new urgency to questions about such things as a person’s capacity to pay attention to multiple radar screens for prolonged periods. The research that such questions prompted revealed interesting effects. These cried out for experimental study, and for a theory to explain them. The rehabilitation of attention as a topic for scientific study was called for, but that rehabilitation required

Highlights of a Difficult History

17

the breaking of the public-observability constraints that governed the positivism-influenced varieties of behaviorism that were dominant in psychology at the time. It required the rejection of the behaviorists’ claim that any properly scientific theory should be given using only terms that refer to publicly observable entities. The most authoritative scientific rejection of that positivistic basis for behaviorism was given by Donald Broadbent in his 1958 book Perception and Communication; a book that also set the agenda for almost all thinking about attention in the decades that followed. It came at a crucial time for psychology’s development. The year before Broadbent’s book had seen the publication of B. F. Skinner’s Verbal Behaviour (1957), in which Skinner made an ambitious attempt to apply a straightforwardly behaviorist explanatory approach to distinctively human aspects of cognition. Thanks to Noam Chomsky’s vituperative review of Skinner’s book, published in 1959, the failure of Skinner’s project was conspicuous and remains famous. That failure made a major contribution to the abandonment of behaviorism and to the corresponding rise of cognitive psychology. But the shift away from behaviorism that followed was not owing to Skinner’s failure alone. It also had much to do with the increasingly detailed articulation of a scientifically respectable alternative to the behaviorist paradigm. Chomsky’s own book, Syntactic Structures, published the same year as Verbal Behaviour, and Donald Broadbent’s Perception and Communication, published the year after, together established an alternative to the behaviorist framework. In their different ways, and drawing on different conceptual resources from the then emerging field now known as ‘informatics’, they suggested a framework that would enable psychologists to study cognitive phenomena while avoiding both the methodologically unsatisfactory method of introspection and the explanatorily inadequate method of mentioning nothing but publicly observable behaviors and stimuli. Broadbent, to a greater extent than Chomsky, retained the behaviorists’ idea that a theory in cognitive psychology was answerable, in the first instance, to quantifiable data about publicly observable behavior, but both Broadbent and Chomsky took it that, contrary to behaviorism’s positivistic strictures, the postulation of internal structures and processes can be a respectable component of a scientific account of a mental phenomenon, even if these structures are not directly observable. Chomsky’s work on syntax showed how the sorts of claims that the behaviorists had rejected could be made precise in the form of claims about the sequence of operations by which a representation gets translated from one form into another. Broadbent’s distinctive contribution was to show how our understanding of a subject’s internal processing architecture could be disciplined by the then-new strategy of importing into psychology the intellectual resources used in designing information technologies. Whereas Chomsky employed the theory of computations considered in abstraction from the practicalities of their implementation.

18

Attention Is Cognitive Unison

Broadbent employed the intellectual resources that are used by communications engineers to negotiate those practicalities. Information technology was progressing dramatically at the time when Broadbent and Chomsky were writing. It was progressing rapidly, with the invention in September 1958 of the integrated circuit. And it was progressing conspicuously, with the launch of Sputnik, and with the introduction, in December of that year, of subscriber trunk dialing to telephone exchanges in the United Kingdom. Broadbent saw the theoretical resources that were used in developing such communications technologies as a tool that could be used throughout psychology. In the conclusion of Perception and Communication, he explicitly presented this as an attack on, and as an alternative to, positivism-influenced behaviorism. But at a much early stage in the discussion, before these broader methodological themes have been broached, he writes: Perhaps the point of permanent value which will remain in psychology if the fashion for communication theory wanes, will be the emphasis on problems of capacity. [. . .] The fact that any given channel has a limit is a matter of central importance to communication engineers, and it is correspondingly forced on the attention of psychologists who use their terms. (1958: 5)

This introduction of the notion of capacity limitations into discussions of perception was, as Broadbent predicted, hugely and permanently influential. It was from claims about capacity, thought of in information processing terms, that all theories of attention in the decades following Broadbent would be built. According to Broadbent’s own capacity-based theory of attention there is a single attentional bottleneck at which capacity limitations are especially pertinent, a bottleneck that arises because the information coming in from the senses is processed by two systems operating in series. The first system, Broadbent thought, has a large capacity for information processing. The second has a much smaller capacity. The bottleneck produced by the connection of the two is the locus of attention in the sense that, although all stimuli are automatically subjected to processing by the first, large-capacity system, the stimuli that make it through the bottleneck into the small-capacity system count, ipso facto, as stimuli to which attention is paid. This two-systems-and-a-bottleneck picture was intended as a communication-theoretic rendering of the everyday idea that simple features of one’s environment, such as the fact that there are people talking in the next room, come to one’s awareness involuntarily, whereas the details of these things, such as the content of the conversation that is taking place, can be detected only for one or two of the things that are going on, with the question of which things have their details detected depending on the focus of one’s attention. By introducing the information-theoretic notion of capacity to discussions of perception, Broadbent allowed various theories of attention

Highlights of a Difficult History

19

to be articulated and debated. The debate between those theories took place in a recognizably Broadbentian idiom until well into the 1980s. But the Broadbentian explanatory framework, and the theoretical skirmish that followed its eventual collapse, did nothing to bring to light the metaphysical issue that had divided James and Bradley. On the contrary, the influence of Broadbent was one of the factors that contributed to that issue’s remaining unarticulated. It did so because all of those whose research was in the Broadbentian tradition, whichever side they took in the subsequent debates about the locus of attentional selection, were, metaphysically, on the side of James rather than Bradley: They took it that the explanation of attention should proceed via the explanation and description of the processes of capacity-limitation management, and so, a fortiori, they took it that the explanation of attention should proceed via the identification and description of particular attention-constituting processes. Bradley’s claim that we need an alternative to the processidentifying explanatory approach was not considered. In the case of Broadbent’s own theory this allegiance to the Jamesian metaphysical position is especially easy to see. James, as we have said, thought that attention was, on some occasions, constituted by ‘the accommodation or adjustment of the sensory organs’. Broadbent’s view was that attention is constituted by the adjustment and accommodation of perceptual processing resources that are just two steps farther downstream: Attention for him is the adjustment, not of the sensory organs themselves, but of the resources that process their output. The late selectionists, who were Broadbent’s chief rivals, differed only in how far downstream they thought the attention-constituting processes take place. Both camps shared an allegiance to the Jamesian idea that the identification of the attention-constituting process (or of a small number of attention-constituting processes) was the correct explanatory strategy for a theory of attention to take. Nowhere in the psychological literature of the Broadbent-influenced period do we find the view that we identified in Bradley’s early work—the view according to which the attempt to explain attention by identifying its constituent processes is tied up with a false metaphysical assumption. To find that view represented in the mid-twentieth century, we need to look to the philosophical literature.

1.7 HEIRS OF BRADLEY IN THE TWENTIETH CENTURY I have said that the issue that divided James from Bradley was, although they failed to articulate it, a metaphysical issue appearing in the guise of a dispute about explanatory tactics. Being metaphysical, one might expect this issue to have been a topic of debate among philosophers, but attention was discussed remarkably little by philosophers in the twentieth century. After Broadbent’s work the philosophical neglect of attention was owing to the fact that attention had come to be regarded as a topic

20

Attention Is Cognitive Unison

for empirical inquiry rather than for a priori philosophizing. Prior to Broadbent the philosophical neglect of attention was owing to the popularity of positivistic and of Wittgensteinian qualms about the explanatory credentials of all inner acts, with attendings being just one example. In 1957, the year before Broadbent’s Perception and Communication was published, these Wittgensteinian qualms were in evidence in Peter Geach’s Mental Acts. Geach mentions attention in connection with the explanation of the way in which perception puts one in a position to form a judgment about the content perceived, but he mentions attention only in order to dismiss its explanatory credentials. ‘So far as I can see’, writes Geach, ‘it is quite useless to say the relevant sense-perceptions must be attended to, either this does not give a sufficient condition, or else “attended to” is a mere word for the very relation of judgement to sense perception that requires analysis’ (64). In the context of this Wittgenstein-inspired hostility to internal acts, the metaphysical issue that divided Bradley and James remained absent from the philosophical agenda. There were, however, occasional breaks in this silence on topics related to attention. They can be found in the literature prompted by the discussion of ‘heed concepts’ in Ryle’s 1949 The Concept of Mind. The most sustained treatment of attention in this period is Alan White’s 1964 monograph Attention. White shared Ryle’s sense that a straightforward behaviorist treatment of ‘heed concepts’ was unsatisfactory, and he thought—as Ryle himself may have done in later years—that the more sophisticated behaviourist treatment offered in The Concept of Mind failed to address these worries (see Ryle 1971, viii). White also shared Ryle’s enthusiasm for the project of extracting philosophical lessons from the norms governing discourse in ordinary language. This can sometimes disguise the depth of White’s insight. When we find him taking pains over points like the following, his philosophy can sometimes sound like a rather uninformative grammar lesson: When we say of someone that he noticed what he or another was doing, the ‘what’ is an interrogative pronoun because ‘noticing’ is here an instance of ‘noticing that’. By contrast, the ‘what’ of attending to what oneself or someone else is doing is a relative pronoun meaning ‘that which’ he is doing. (1964: 18)

Current philosophers tend to regard this fixation on natural language as a phase that philosophy did well to get out of, and understandably so, but the bad reputation of ordinary language philosophy is in some ways undeserved. White knew perfectly well that our philosophical theory of attention should tell us about attention itself, and not just about the rules governing our talk about it. His discussion of the grammar of attention talk was not intended as an end in itself. It was intended to enable the diagnosis of fallacies in the arguments about the phenomenon’s underlying nature.

Highlights of a Difficult History

21

Among the lessons that White wanted to draw from his examination of the behavior of ‘heed concepts’ was a lesson about the way in which Ryle’s treatment of those concepts needed to be modified. Ryle had treated heed concepts as belonging in the class that he dubbed ‘mongrel categoricals’ (1949: 135–138). By this he meant that statements employing heed concepts not only tell us what is taking place, but also situate that event relative to a set of dispositions and hypotheticals. Ryle provides some examples that make the point clear. ‘That bird is migrating’ is one of his paradigm cases: The description of a bird as migrating has a greater complexity than the description of it as flying in the direction of Africa, but this greater complexity does not consist in its narrating a larger number of incidents. Only one thing need be going on, namely, that the bird be at a particular moment flying south. ‘It is migrating’ tells not more stories, but a more pregnant story. (136)

Ryle’s theory is that ‘he is attending to what he is doing’, like ‘it is migrating’, tells a pregnant story in the sense that it gives us hypothetical information, about what might have been the case, along with descriptive information, about what actually is the case. The hypothetical information that it gives is information about what the attentive man does typically, or about what he would have done in similar but slightly different situations. White rejected Ryle’s idea that ‘in playing the piano, he is attending’ was a mongrel categorical of this sort. His reason, which we consider in more detail in chapter 4, was that ‘it is migrating’ can serve as an explanation of ‘it is flying south’, whereas ‘he is attending’ cannot explain the fact that he is playing the piano. At most, it explains a different fact: that his piano playing is successful. Describing the bird’s flight as an instance of migration provides us with information that zooms out and locates the flight in a broader pattern of dispositions that the bird, and birds like it, displays. Describing the pianist’s playing as an instance of attention, by contrast, zooms in and tells us more about the way in which the playing is taking place. As White puts it: ‘Whereas “He is practising on the piano” gives a more complex narration of the same incident as that narrated by “He is playing the piano”. “He is attending to what he is playing on the piano” narrates a more complex incident’ (1964:14). Ryle’s and White’s preference for talking in formal mode, about concepts and sentences rather than things in themselves, is, on this occasion, merely an expository technique. When translated into the material mode White’s claim is recognizable as a metaphysical one. It is the claim that the metaphysical relationship between attending and the activities that are performed attentively must be different in kind from the metaphysical relationship between migrating and the flights that are instances of it. The alternative picture, with which White wants to replace Ryle’s mongrel categorical account, is a picture that is close kin to the picture

22

Attention Is Cognitive Unison

found in Bradley. Bradley, as we have seen, rejected the idea that identifying the attention-constituting processes gave us what we need in order to know what attention is. He thought that attention was constituted by the processes of ‘redintegration’ and ‘blending’, but he thought that this fact was of no explanatory consequence, since a catalog of attention-constituting processes, like the catalog of employment-constituting activities, is the wrong form for an explanation of attention to take. In White’s work, the idea that the explanation of attention cannot proceed by identifying which processes take place becomes the claim that ‘attention’ ‘does not name any specific activity; it indicates the circumstances in which the activity occurs and thus signifies what, on this occasion, it amounts to or is a form of’ (1964: 6). White and Bradley are, then, in broad agreement about the negative claim—that attention is not explicable by the identification of its constituent processes. They also offer something broadly similar by way of a positive alternative. White spells it out as follows: Simply to say that someone is attending, or paying attention, gives us no more clue as to what activities he is engaged in than simply to say that he is practising. What ‘attending’ tells us is that his activities and energies, whatever they are, are directed to and focused on something which occupies him. (1964: 7)

For Bradley and White, and in a slightly different way for Ryle, claims about attention are not made true simply by the facts about which processes are taking place in the attentive thinker. In Bradley’s treatment this is because claims about attention tell us, not about which processes are taking place, but about their immediate causal context: The attentionconstituting processes are those that are ‘prompted by interest’ and that ‘bring about engrossment’ (1886: 316). In Ryle’s treatment it is because claims about attention tell us something about processes and also tell us something hypothetical: Intellectual processes constitute attention when they occur in a context such that similar processes would have taken place in a range of counterfactual situations (1949: 135–136). In White’s treatment, the idea is that claims about attention do not tell us which processes are taking place. Instead they tell us how those processes relate to the agent’s other activities and to his goals. A theory like Bradley’s or White’s (or Ryle’s) has different explanatory goals and operates at a different level from a theory like Broadbent’s, but this should not be taken as an indication that the theories are not in conflict. Although theories at different levels of explanation need not compete, this does not mean that they never exclude one another. There is a genuine disagreement between a theory like Broadbent’s and a theory like White’s. It is a disagreement over the fundamental metaphysical/explanatory question that divided Bradley and James. White, like Bradley, takes it that the identification of processes at the subpersonal level is the wrong sort of thing to provide an explanation of attention.

Highlights of a Difficult History

23

Broadbent, like James, attempts to explain attention by just such an identification. We saw above that Bradley and James lacked the vocabulary in which to articulate this disagreement. When White and Broadbent were disagreeing over this same issue the reasons why the metaphysical basis of their disagreement went unarticulated had more to do with the fragmentation of scientific and philosophical attempts to account for the mental. For these various reasons, the issue underlying these fundamental differences of explanatory approach remains one that has never been brought into focus as a topic for debate. Our analogy with employment has given us a glimpse of what that issue might be. Chapter 2 articulates it.

2 The Underlying Metaphysical Issue

2.1 EXPLANATORY TACTICS Our theory of attention should put us in a position from which we can answer both of the following questions: 1. What is attention? 2. What is it for something to be done attentively? There are two questions to be answered here, but there are not two bits of philosophical work needing to be done. Once we have an answer to either one of these questions we are in a position to derive our answer to the other: If we have an independent answer to question 1 then we may derive our answer to question 2 by saying, ‘To do something attentively is to do the thing in the manner that typifies instances of the process of attention (a process that we identified in our answer to question 1)’. If we have an independent answer to question 2 then we can derive our answer to question 1 by saying, ‘Attention is the doing of something in the attentive manner (a manner that we identified in our answer to question 2)’. We need to start with an answer to one question or the other, but it is not obvious which it should be.1 In chapter 1 we saw William James advocating a theory that took the first approach. James’s theory started by telling us what attention is. His claim was that it is either sensory orienting or preparatory imagining. From here James could, if he had wanted to, derive an account of attentiveness simply be saying that it is the manner that typifies instances of these processes. We contrasted James’s approach, as did James himself, with the approach advocated by F. H. Bradley. Bradley starts with question 2. He claims that an activity is performed attentively if and only if the agent is ‘engrossed’ and ‘prompted by interest’. As long as this manner was instantiated, Bradley suggested, we have an instance of attention. If we want an answer to the question ‘What is attention?’, then the correct way to go about getting it is by deriving that answer directly from the answer that we give to the question about performing attentively. Attention is just the doing of something in that manner.

24

The Underlying Metaphysical Issue

25

Rather than being, as James took it to be, a disagreement about causes and effects, or about spiritual and mechanical forces, the disagreement that divides these two explanatory strategies can be traced to their different commitments as to the metaphysical category into which they assign their explananda. And these different metaphysical commitments, in turn, have different methodological ramifications: Starting with the wrong view of attention’s metaphysics will lead us to look for the wrong sort of explanation, and it will lead to our using the wrong sorts of methods with which to look. In chapter 3 I present an argument that gives us a reason (although not a conclusive reason) to suspect that most current theories of attention have indeed started with the wrong metaphysics. To see that argument we shall need an account of what the underlying metaphysical issue is. This chapter is concerned with giving such an account. It will make precise the idea that if we side with James we are committed to the idea that a process, or a set of processes, can be identified as being responsible for attention, whereas if we side with Bradley we are committed to the idea that it is something other than their constitutive processes—something about manner or context—that instances of attention have in common.

2.2 THE BASIC DISTINCTION The lesson to be learned from James and Bradley, if put in maximally general terms, is that something fundamental about the metaphysical status of any phenomenon, x, is revealed by the order of explanatory priority between our answers to the questions ‘What is x?’ and ‘What is it for something to be done x-ishly?’ We can begin to make this clear by considering the order in which these questions must be taken in a pair of cases where x stands for perfectly unmysterious phenomena of intuitively contrasting metaphysical sorts. Suppose that one were faced with an intelligent but ignorant interlocutor who wanted to know what combustion is. Suppose also that our interlocutor had had some exposure to discourse about combustion and wanted answers to the two questions analogous to those with which this chapter started: (1) ‘What is combustion?’ and (2) ‘What is it for something to occur combustively?’ We would need to start our analysis with an answer to the first of these. We explain the phenomenon of combustion by identifying it with a particular process—the process of burning—and by giving a description of what that process is. Once we have identified a process of combustion (and thereby answered the first question), we can, with no further empirical or philosophical work, derive an account of what it is for something to happen combustively. We need do no more than say that, in order for something to happen combustively, it must happen in the manner of a typical instance of combustion (that is, of an instance of the process of burning). For combustion, then, our answer

26

Attention Is Cognitive Unison

to question 2 is derived from our answer to question 1. This same order of explanatory priority is required whenever we try to analyze any phenomenon that is, firstly and most fundamentally, a process. The opposite order of explanatory priority is required if we are trying to give an account of a phenomenon such as haste. To bring about an instance of haste one does not have to do anything in particular. One has only to ensure that whatever it is that one is doing is hasty. If an intelligent but ignorant interlocutor presented us with the questions ‘What is haste?’ and ‘What is it for something to be done hastily?’, then we would need to start our analysis by giving an answer to the second question first. The process-first order of explanatory priority would be unavailable since there is no process of haste by reference to which one can explain what it is for something to be done hastily. The property of hastiness—the property, that is, of being done with what might typically be rash rapidity—is explanatorily prior to haste. Our answer to question 1, ‘What is haste?’, must therefore be given via an answer to question 2, ‘What is it for something to be done hastily?’ Once we know what the property of hastiness is there is no further philosophical or empirical work to be done in saying what haste is. Instances of haste are just instantiations of that property. For haste and for combustion the orders of explanatory priority are clear. They are clearly different. And they are clearly dictated by the metaphysical character of the thing to be explained. When explaining combustion we have no choice but to answer question 1 first. When explaining haste we have no choice but to begin with question 2. This difference in the order of explanatory priority for haste and for combustion is a reflection of the fact that they belong in different metaphysical categories. James’s approach of answering question 1 first is appropriate only if attention belongs in a specific metaphysical category. Bradley’s approach of answering question 2 first is appropriate only if attention belongs in a different metaphysical category. It will be useful to have some vocabulary to refer to the metaphysical category of things that, like combustion, get explained via a question-1first mode of explanation and the category of things that, like haste, get explained via a question-2-first mode of explanation. Phenomena like combustion I will call ‘process-first’ phenomena. Phenomena like haste I will call ‘adverbial phenomena’.2 We have a good intuitive grasp on the distinction between process-first phenomena and adverbial phenomena, as reflected in the fact that there is nothing surprising in finding that the explanation of haste contrasts with the explanation of combustion. Section 2.3 articulates this distinction’s basis.

2.3 METAPHYSICAL CATEGORIES AND TAXONOMIES We can give a precise account of the distinction between process-first and adverbial phenomena by thinking about the way in which the sets

The Underlying Metaphysical Issue

27

of their instances map onto different taxonomies of events. The points that we need, in order to clarify the metaphysical distinction that we are concerned with, are points about taxonomies of events, but these points can be seen more clearly when they are applied to taxonomies of objects. It will help to consider these first. There are various different taxonomic principles that might be used to carve up the set of all objects into categories. Some of the resulting taxonomies will be useful for various purposes. Some will be completely useless. Among the useful taxonomies are those that group objects on the basis of the substance from which they are made. Since there is more than one way to individuate substances, there is more than one way to give a substance-based taxonomy of objects. One substance-based taxonomy groups all plastic things into one set, all wooden things into another, and so on. A different taxonomy might divide up the plastic things more finely, grouping all the PVC into one set, all the acrylic things into another, and so on. There are lots of ways that a taxonomy of objects might be given. Suppose now that we are given the set of objects and asked to find a way of specifying the set of all paperweights. If we are allowed to use only the classificatory resources of a substance-based taxonomy then we will not be able to do it. The problem we face is not that paperweights are ‘multiply realizable’, so that there are many substances from which a paperweight could be made. It is true that paperweights can be realized in multiple substances, but it is not this that stops us from delineating the paperweights with the resources of a substance taxonomy: The problem of multiple realizability could be solved simply by specifying a union of several sets. What stops us from being able to delineate the paperweights using the distinctions provided by a substance taxonomy is that being made of the same substance as a paperweight does not suffice for being a paperweight. That makes it impossible to rule in all of the paperweights without also ruling in some of the nonpaperweights. There is a cheating way to avoid this problem, which is to define a maximally fine-grained taxonomy of substances—a taxonomy on which, for example, ‘plastic-that-was-made-in-one-a-certain-list-of-factories-ona-Tuesday-morning’ is a different substance from ‘plastic-that-was-madein-one-of-a-certain-list-of-factories-on-a-Tuesday-afternoon’. A taxonomy like that might enable us to draw a line around all and only the paperweights. But any such taxonomy would be so fine-grained that it would cease to be useful as a taxonomy of substances. Any taxonomy that is recognizable as a taxonomy by substances is cut by the set of paperweights. When we apply it to events, this notion of a set that cuts a taxonomy enables us to clarify the metaphysical distinction that underpins the difference between those phenomena that are amenable to processfirst explanation and those that are amenable to adverb-first modes of explanation. As with the set of objects, the set of events can be carved up in different ways, and the different taxonomies that result will be useful for different purposes. One sort of taxonomy, useful for several

28

Attention Is Cognitive Unison

purposes, is the taxonomy that groups events according to the processes that they instantiate. Suppose now that we are given the task of specifying the set of instances of haste. If we are allowed to use only the classificatory resources of a process-based taxonomy then we will be unable to do it. Again the problem has nothing to do with multiple realizability. The problem is not that there are too many sorts of things that can be done hastily. The problem is that instantiating the same processes as an instance of haste is not sufficient for being an instance of haste. The set of instances of haste cuts any taxonomy by process. It would be a mistake to take these points as showing (although it might, in fact, be true) that paperweights and haste are ontologically less basic than lumps of metal or instances of combustion: No point about orders of ontological basicness can be made via a taxonomy-cutting argument since, had we chosen different taxonomies, the situation would have been the other way around. No substance taxonomy is sufficient for delineating the set of paperweights, but if we were using a taxonomy that classifies objects by function, then we could delineate the set of paperweights quite easily, and there would now be no way to delineate the set of lumps of metal. Again there is an exact parallel in the case of events. No taxonomy that classifies events by process is adequate to the task of delineating the set of instances of haste, but if we were using a taxonomy that classifies events by their manner of occurrence then we could delineate the instances of haste quite easily, and there would now be no way to delineate the set of instances of combustion. Taxonomy cutting, then, is a way of articulating the distinctions between metaphysical categories. It is not a tool for the ordering of hierarchies according to ontological basicness. The distinction between process-first phenomena and adverbial phenomena can now be analyzed as follows: The set of instances of an adverbial phenomenon cuts any taxonomy by process (as reflected by the fact that the things that can be done hastily can also be done nonhastily). The set of instances of a process-first phenomenon cuts any taxonomy by manner of performance (as reflected by the fact that imitations of the manner in which combustion occurs need not themselves be instances of combustion). The distinction that we are interested in can therefore be made precise if we can find a way to distinguish between taxonomies by process and taxonomies by manner of performance. And now our philosophical task is relatively straightforward: A distinction between these different sorts of taxonomic principles, at least in the case of taxonomies of events, is easy to draw. The distinction that we need in order to complete our account of the metaphysical difference between process-first phenomena and adverbial phenomena is a distinction between taxonomies that classify events on the basis of process and taxonomies that classify events on the basis of manner of occurrence. The natural way to characterize the contrast here is this: The question of which processes happen is settled by the facts

The Underlying Metaphysical Issue

29

about what things happen to what. The question of what manner those processes have is settled by the facts about how the things that happen happen. ‘What happens to what’ is a matter of the gaining and losing of properties by objects. ‘How the things that happen happen’ is a matter of the gaining and losing of properties by events. To put the same point more precisely: A taxonomy is a taxonomy on the basis of process if the taxonomy classifies events on the basis of the having or gaining of a property by an object. A taxonomy is a taxonomy on the basis of manner if the taxonomy classifies events on the basis of the having or gaining of a property by an event. Putting all this together gives us the result that, if a phenomenon belongs in the process-first category, then the set of instances of that phenomenon can be delineated using the resources of a taxonomy that groups events on the basis of the having or gaining of a property by an object. And if a phenomenon belongs in the adverbial category, then the set of instances of that phenomenon can be delineated using the resources of a taxonomy that groups events on the basis of the having or gaining of a property by an event. The contrapositive versions of these claims are particularly pertinent for the argument that follows and are also worth noting here: A phenomenon does not belong in the category of adverbial phenomena if the set of instances of that phenomenon cuts all taxonomies that group events on the basis of the having or gaining of a property by an event. A phenomenon does not belong in the category of process-first phenomena if the set of instances of that phenomenon cuts all taxonomies that group events on the basis of the having or gaining of a property by an object. What I have done here is to introduce and to make precise a distinction between two metaphysical categories, corresponding to a difference in two explanatory approaches. We have seen that there is a metaphysical difference between those phenomena that are explained by first answering a question like the first of those we started with and those phenomena that are explained by first answering a question like the second of those we started with. I have put this distinction on a more precise footing by showing how, via the notion of taxonomy cutting, it can be explained in terms of the metaphysical distinction between properties of objects and properties of events. All of this is intended to make precise our intuitive notion that there are some things—such as combustion, decay, digestion, and evolution—that are to be explained in one way, and other things—such as haste, stealth, wit, style, and employment— that are to be explained differently. The reader who is skeptical of the metaphysical distinction between properties of objects and properties of events (on which my analysis rests) will be skeptical about whether my discussion moves in a way that constitutes philosophical progress. I am sure some readers will be skeptical of precisely that, but the analysis has to stop somewhere, and I propose to stop it here, with the distinction

30

Attention Is Cognitive Unison

between properties of objects and properties of events as our unanalyzed analyzer. This distinction between adverbial and process-first phenomena, not a distinction between causes and effects, nor a distinction between spiritual and mechanical forces, is what James and Bradley were really disagreeing about.

2.4 ADVERBIALISM, MULTIPLE REALIZABILITY, AND NATURAL KINDS Having said what it is to be an adverbial or a process-first phenomenon, we can now clarify some important points about this distinction and about the claim, which is what we are working toward, that attention might belong on the adverbial side of it. We can, in particular, see how this claim differs from the various things that might be meant when psychologists say that attention is an ‘umbrella term’ (see chapter 1, §1.5). We have already seen that the adverbialist’s claim is not, in the first place, a claim about multiple realizability. Philosophers of mind have often focused on issues arising from the multiple realizability of mental phenomena. They have had good reasons for doing so, but multiple realizability is not the heart of the metaphysical matter that is in contention here. A phenomenon may be realized by multiple processes because it is a disjunctive kind of process; or because it is a process-genus with multiple processes as its component species; or because it is a process-family, with multiple processes as it members. Such phenomena are realized by multiple processes, but they are not ruled out of the category of processfirst phenomena. The sets of their instances can be delineated using only conjunctions or disjunctions of the distinctions given to us by a process taxonomy on the set of events. They don’t cut such taxonomies. The claim that attention is an adverbial phenomenon should not be mistaken for the claim that attention can be realized by multiple processes. Nor should it be confused with the claim that the instances of attention do not form a natural kind of process. The analogy between taxonomies of objects and taxonomies of events makes this clear: The instances of jade do not form a natural kind of substance—some instances of jade are nephrite, some are jadeite. But the set that is the union of jadeite instances and nephrite instances contains all and only the instances of jade, and so there is a taxonomy by substance that the set of instances of jade does not cut. This is exactly as it should be. The question of whether something is an instance of jade is a question that is settled by the facts about what substance the thing is made of. The metaphysics of jade is that of a substance, even though jade is not a natural kind of substance. The claim that ‘jade’ is not a substance term would be stronger than the claim that the instances of jade do not form a natural kind of substance (and it would be false, even though the second claim is true).

The Underlying Metaphysical Issue

31

This point about taxonomies of objects transfers directly to taxonomies of events. The claim that attention does not have the metaphysics of a process is stronger than the claim that the instances of attention do not form a natural kind of process. The claim that attention does not have the metaphysics of a process, unlike the claim that attention is not a natural kind of process, rules out the possibility that attention can be identified with a disjunction or family of several processes, any one of which is sufficient for the instantiation of attention. The methodological significance of this difference between the adverbialist’s claim and the not-a-natural-kind claim can be seen if we consider the standard view of the psychology of memory. The processes responsible for short-term memory are probably very different, cognitively speaking, from the processes responsible for long-term memory; and the processes responsible for episodic memory are probably very different, cognitively speaking, from those responsible for semantic memory, and so on (Baddeley 1990, chapter 3; Schacter, Wagner, and Buckner 2000). It follows from this that memory is not a natural kind of cognitive process. But that fact does not present a problem for the attempt to explain memory via the identification of cognitive processes. We can explain memory perfectly well by the piecemeal identification of various processes corresponding to its various instances. Adverbial phenomena, in contrast, cannot be explained by the piecemeal identification of their component processes. It is impossible to explain what haste is by looking closely at the component processes of various instances of haste, just as it is impossible to explain what employment is by looking at the processes that people execute in the course of being employed. However closely you study the component processes of the instance of haste that took place in my kitchen this morning, you will only find out about the processes involved in pouring water over a tea bag and adding milk. An examination of these processes will tell you nothing about haste in general and nothing about haste per se. If it turns out that attention belongs in the adverbial category then that, unlike the not-a-natural-kind claim, will rule out the possibility of explaining attention via a piecemeal process-by-process approach, of the sort that we use when explaining memory. We shall return to these methodological and explanatory consequences of adverbialism in chapter 6, once we have got a worked out version of an adverbialist theory of attention on the table. The point to note here is just that the denial of the process-first approach to attention is distinct from the denial that the instances of attention form a natural kind of process, and that the first claim, unlike the second, is incompatible with the idea that an account of what attention is can be given via the examination of attention’s constituent process. A piecemeal explanatory approach (such as some psychologists advocate when they say that attention is an ‘umbrella term’, and such as we use when explaining memory) is committed to the process-first view and incompatible with the adverbialist alternative.

32

Attention Is Cognitive Unison

2.5 ADVERBIALISM AND LEVELS OF EXPLANATION We have seen that the adverbialist’s claim should not be mistaken for a claim about multiple realizability, or for a claim about failure to correspond to a natural kind. We can also see that the adverbialist’s claim should not be mistaken for a claim about levels of explanation. The difference between adverbial phenomena and process-first phenomena is not a difference between their levels of explanation. It is a difference between types of explanation. The process-first view of attention requires that there be some taxonomy of processes not cut by the set of instances of attention. It does not need every taxonomy of processes to meet that requirement and it does not tell us at which level of description the requirement will be met. This, again, is a point that can be made clear by considering an exactly parallel case for the more familiar metaphysical category of substances. Diamond is a substance. This fact about the metaphysics of diamond entails that there is a taxonomy for substances that has the property of being a taxonomy on which being the same substance as a diamond suffices for diamondhood. There clearly is a taxonomy of substances that meets this requirement. It is a taxonomy that classifies substances by reference to such things as their elemental composition and their bonding structure, and also, perhaps, their more manifest properties (Johnston 1997). The fact that there are other taxonomies of substances that lack this property does not show that diamond is not a substance. A taxonomy that classifies substances by elemental composition alone does lack that property: Instances of graphite have the same elemental composition as instances of diamond, but they are not diamonds. What is necessary in order for diamond to have the metaphysics of a substance is for there to be some taxonomy of substances, at some level of description, on which being the same substance as a diamond suffices for diamondhood. The fact that diamond is a substance does not, by itself, tell us anything about the level at which we can describe the substance of diamonds. Similarly, the claim that attention is a process does not entail anything about the level at which we can describe the process corresponding to attention. When we deny that attention has the metaphysics of a process we are thereby committed to the claim that the question of whether or not a subject is paying attention is not settled by the facts about which processes are going on at any level of description.

2.6 TAXONOMIES AND SUPERVENIENCE RELATIONS When analyzing the process-first/adverbial distinction by reference to different taxonomic principles, we traded on the idea that the facts that make it the case that a particular process takes place are facts about what happens to what, whereas the facts that make it the case that a

The Underlying Metaphysical Issue

33

particular adverbial phenomenon occurs are facts about how the things that happen happen. The metaphysical difference between adverbial phenomena and process-first phenomena can be thought of as a difference as to the sort of facts that make it the case that the phenomenon in question occurs. The sense of ‘making it the case’ that is at work when we think about the adverbial/process-first distinction in this way is clearly not a causal sense. The facts about what happens to certain oxygen and carbon atoms when I light the gas are the facts that make it the case that there is an instance of burning, but not because these facts cause the gas to burn. They are facts about the burning itself. This noncausal metaphysical relation of making it the case is one for which we would like to have a thorough philosophical account. For present purposes it is enough to note that this variety of ‘making it the case’ is a determination relation, in the sense that claims with the form ‘p makes it the case that a is X’ correspond to supervenience claims in the following way: The fact that an item has properties p1 . . . pn makes it the case that that item is an instance of X only if it is metaphysically necessary that any two items that are the same in respect of properties p1 . . . pn are also the same in respect of whether they are instances of X.

What this says is that if any two items that are the same in respect of p1 . . . pn can differ in that one is an X and the other is not, then something other than the properties p1 . . . pn makes it the case that the one of them is an X. Or, to put it another way, it says that for some arbitrary γ that is an instance of X, the facts about p1 . . . pn make it the case that γ is an X only if the question of whether γ is an X is settled by the facts about whether γ has p1 . . . pn. The tea making that took place in my kitchen this morning was hasty, but the tea making that took place this afternoon was not hasty. The two tea makings were the same in respect of the processes that they involved (boiling water, putting a tea bag in the cup, and so on), but one was an instance of haste and the other was not. It must therefore have been something other than the facts about which processes were involved that made it the case that the tea making this morning was an instance of haste. This link between the metaphysical distinction that we have been concerned with and supervenience relations provides us with a way of framing an argument about the side of that distinction on which a particular phenomenon belongs. The fact that jade has the metaphysics of a substance entails that there is a taxonomy of substances that is not cut by the instances of jade. This entails, thanks to the link with supervenience, that for any two items that are the same substance, if one is jade then the other is jade too. The view that attention has the metaphysics of a process entails that there is a taxonomy of processes that is not cut by the instances of attention. Thanks to the link with supervenience this entails that, for any two sets of cognitive events that are the same with regard to

34

Attention Is Cognitive Unison

the processes that they instantiate, if one is attention constituting, then the other is attention constituting too. The relation to supervenience makes clear what needs to shown in order to make an argument for the claim that attention belongs on one or the other side of our distinction. In chapter 3 we shall see such an argument.

3 Rejecting the Process-First View

3.1 SUPERVENIENCE FAILURE Chapter 2 gave an account of the metaphysical distinction that underpinned F. H. Bradley and William James’s disagreement as to explanatory tactics. It concluded by showing that their two rival views of attention’s metaphysics carry different commitments concerning the sorts of cases that they take to be metaphysically impossible. The two explanatory approaches carry these different commitments because they disagree about the sort of facts that make it the case that a subject is attending, and because, in general, if it is facts about the F-ness of A that make it the case that A is G, then it is metaphysically impossible for there to be an item that fails to be G although it is indistinguishable from A in its F-ness. An adverbialist, such as Bradley, requires it to be impossible for there to be pairs of cases that are alike in manner but such that one of them is, and one of them is not, an instance of attention. He needs it to be the case that there is some way of individuating manners for which having the very same manner as an instance of attention is sufficient for being an instance of attention. Taxonomies that individuate manners on the basis superficial appearances clearly fail to satisfy this requirement, since one may attend to the task of appearing inattentive, but there may be some other taxonomy, based on the properties of less overt events, that can satisfy the adverbialist’s requirement. An adherent of the process-first view, on the other hand, requires it to be impossible for there to be a pair of cases that are alike in their constituent processes but such that one of them is, and one is not, an instance of attention. He needs it to be the case that there is some way of individuating processes for which, if two systems are the same in respect of the processes taking place with them, then either both are attention constituting or neither is. We can adjudicate between these two rival theories by looking to see which sorts of cases are indeed metaphysically impossible. In the first part of this chapter the process-first account is put to this test and found wanting. The result is not a knockdown argument against the

35

36

Attention Is Cognitive Unison

process-first view, partly because it is only half the story until we have seen an adverbial theory that can be put to the same test, and partly for reasons to do with the lack of constraint that the process-first theory places on levels of explanation, but the argument does, together with the interference argument presented in this chapter’s second half, provide a reason for looking instead at the prospects of an adverbialist treatment, such as chapter 4 will give.

3.2 THE MODAL COMMITMENTS OF THE PROCESS-FIRST VIEW The process-first view is committed to the claim that the instantiation of the same processes as are instantiated in a case of attention is sufficient for being an instance of attention, and so it is committed to claiming that no change from attention to inattention is possible without a change in the underlying cognitive processes. This commitment may look quite unproblematic. It may look like the sort of claim that any naturalistically respectable account of the mind requires. But this appearance is deceptive. The process-first view’s commitment is to a supervenience claim of broadly the sort that we use when stating the requirements of naturalistic respectability, but it is to a very strong version of such a claim, going well beyond anything that naturalistic respectability requires. The process-first view is committed to saying that nothing at all, natural or otherwise, could make a difference to whether or not a set of processes constitute attention, other than by changing the facts about which processes those are. That claim is so strong as to be implausible. We can imagine taking a set of cognitive events that constitute the paying of attention in one of attention’s simpler instances; we can imagine moving those events, without changing any of the facts about which processes are taking place, and situating them in a new, more complicated cognitive context; and we can imagine this happening in such a way that, in the new context, the events are no longer attention constituting. The fact that such cases are imaginable is a problem for the process-first view since the supervenience claim to which that view is committed says that such cases should be impossible.

3.2.1 Feature Binding The most straightforward of the problem cases for the process-first theory are the cases in which we start with an example in which the processes that constitute an instance of attention are relatively simple and relatively few in number. The attention involved in searching for a pair of co-instantiated properties in a crowded visual environment is the best studied of these cases.

Rejecting the Process-First View

37

There is an excellent and long-standing body of empirical work suggesting that attention is involved in the performance of tasks that require people to search as quickly as possible through a crowded visual display in order to find a target that is defined by certain combinations of features. If, for example, one is looking for a red triangle against a background containing other red things and other triangles, then one needs to search through the items in turn, paying attention to each of them as one goes along (Treisman and Gelade 1980). This form of attention has been studied a great deal (some examples are: Treisman and Gormican 1988; Friedman-Hill, Robertson, and Treisman 1995; Treisman 1998; Hommel 2004; Fougnie and Marois 2009; Hommel and Colzato 2009). There has been controversy about which combinations of features require attention, and controversy about how the binding together of the features is accomplished, but there is excellent evidence for the claim that in some cases the processes that constitute the attention that one pays when searching for conjunctions of properties in a speeded visual search of a crowded environment are the processes by which the brain ‘binds together’ its representations of each object’s independently registered features so as to build an ‘object file’ representing all of that object’s variously detected features (Treisman 1998). Evidence for this hypothesis comes from the predictions that it makes about the sorts of shapes that can be distinguished from one another by subjects who are not paying attention, from the predications it makes about the sorts of errors that subjects make when asked about items to which they have not been attending, and from the predictions it makes about the shapes of the functions linking search times to the size of the field searched (Treisman and Gelade 1980; Treisman 1998, 2003). I take this evidence to show that there are some contexts in which, in the initial couple of hundred milliseconds of performing a search task, where that search task involves gathering information about conjoined features from a crowded visual environment, the subject’s attention to an item is constituted by the binding of that item’s features. This much— that there are some cases in which attention is wholly constituted by feature-binding processes—is all that our argument needs. The principles about supervenience that we saw in chapter 2 entail that, if the process first view were right, then, given that the processes of feature-binding constitute attention in some contexts, it should be impossible for events constituting those same feature-binding processes to take place without constituting attention. The problem for the process-first view of attention is that this does not seem to be impossible. We can imagine those same feature-binding processes situated in a context in which visual search is irrelevant, where the thinker is devoting all of her intellectual resources to some complex and ongoing intellectual puzzle, and where perceiving the bound features of the thing before her is not at all integrated with her attempt to solve that puzzle and exerts no influence upon it. The thinker in such a case may be binding the features of

38

Attention Is Cognitive Unison

the object before her, but her bound experience of that object may be nothing more than a slight distraction from her thinking. In such a case it would seem that the processes by which the features of the object are bound do not constitute attention. The argument here does not depend entirely on our verdict about the perceptual inattentiveness of this imagined thinker. We can find realworld examples in which there is good evidence that feature-binding processes are instantiated, and good evidence that they do not constitute attention. Such cases can be found by looking at the pathological cases of inattention that are displayed by patients suffering from unilateral neglect. Patients with unilateral neglect are incapable of attending to items on the left side of space (Behrmann and Geng 2002). Some unilateral neglect patients nonetheless show clear signs that they are binding the features of stimuli presented on that side of space. These patients show priming by the semantic content of line drawings presented on the neglected side, in ways that require them to be encoding the meaning of the configured lines, not merely the collection of their unbound features (Berti and Rizzolatti 1992). There is also evidence that these patients are subject to illusions, such as the Müller Lyer illusion and the Judd illusion, even when these illusions depend on the configurations of those parts of the stimuli that fall in the neglected part of space (Ro and Rafal 1996; Olk, Harvey, Dow and Murphy 2001). We know, from the work of Rensink and Enns (1995), that when normal subjects search for the different configurations of lines that produce the Müller Lyer illusion (and the Judd Illusion), their search for one configuration against a background of others is slow and attention demanding, in exactly the way that suggests these configurations need to be bound in order for the difference between them to be experienced. The unilateral neglect patients who are susceptible to the different illusions caused by these different configurations therefore seem to be binding the neglected parts of the illusion-eliciting stimuli. This evidence suggests that neglect patients do bind the features of the stimuli that they neglect. Since those stimuli are neglected we know that the agents are not paying attention to them. In such patients, therefore, the feature-binding processes occur without constituting attention. Since those feature-binding processes do constitute attention when normal subjects perform certain sorts of visual search tasks under pressure of time, such cases suggest that it is possible for processes that constitute attention in some of their instances to occur without constituting attention in other instances. This is the possibility that the process-first account of attention’s metaphysics cannot allow for.

3.2.2 Other Cases The diversity among attention-involving phenomena makes it plausible that there will be other examples that are problematic for the

Rejecting the Process-First View

39

process-first view in just the way that feature binding is. There are, as we noted in chapter 1, paradigmatic cases of attention that are involved, not with sensory focalization and binding, but with the coordination of action or with the direction of a line of deliberative thought. It is plausible that for these varieties of attention we will again be able to find processes that constitute attention in some simple contexts but that fail to constitute attention when those same processes take place in more complex contexts, where some other matter is the focus of the thinker’s cognitive effort. The task of producing a string of entirely random numbers is one possible source of such examples. If you have a metronome at hand then set it to click at the rate of 120 beats per minute and try generating random numbers at the rate of one number per click (without cheating by producing pseudorandom sequences, such as remembered telephone numbers or PIN codes). This task proves to be remarkably attention demanding (Baddeley 1966; Jahanshahi, Saleem, Ho, Dirnberger and Fuller 2006). The numbers that pop into one’s head tend not to be random. They tend to follow obvious patterns. To quickly generate random numbers one needs to constantly switch between strategies of generation to prevent such patterns from developing. It therefore seems likely that, when one pays attention to random-number generating, one’s attention can be constituted by the processes involved in disengaging from one number-generating strategy and adopting another. If so then we can produce another case of the sort that is problematic for the process-first theory by showing that those same strategy-switching processes could occur in some other context while failing to be attention constituting. It is highly plausible that they could. In a context where the subject’s task is one that requires doggedly pursuing a single train of thought, the process of disengaging from that task and switching to another could, it seems, occur in ways that constitute failures of attention, not instances of it. If this is indeed possible then these processes present a problem for the process-first view of attention in just the way that inattentive feature binding does. They constitute attention when the agent’s task is random number generation, but the same processes constitute inattention when embedded in a context where the agent’s task is more cognitively elaborate. They are therefore counterexamples to the claim that instantiating the same processes as an instance of attention suffices for being an instance of attention. As such they suggest that the supervenience claim that the process-first view is committed to turns out to be false.

3.2.3 Caveat The fact that feature-binding processes constitute attention in cases of speeded visual search and fail to constitute attention in cases of unilateral neglect, together with the fact that processes of task switching constitute attention in cases of random number generation and fail to

40

Attention Is Cognitive Unison

constitute attention in cases of single-minded tasks, makes trouble for the process-first view of attention’s metaphysics. But these facts are not, it must be admitted, a conclusive refutation of that view. There are two reasons why not. The first bar to conclusive refutation is that the process-first view always has an available reply, although it is a reply that, if used too many times, condemns the view to triviality. The process-first view (as noted in §2.5) requires only that there be some taxonomy of processes on which being the same process as an instance of attention is sufficient for being an instance of attention. The most that the examples discussed above can show is that no taxonomy on which all the instances of feature binding are grouped together can satisfy that requirement, and that no taxonomy on which all the instances of task switching are grouped together can satisfy it. The defender of the process-first view can always respond by proposing some other taxonomy of cognitive processes that does meet the requirement. By doing so he runs the risk of finding himself with a trivial position. If one individuates processes finely enough and is prepared to tolerate a theory with the form of a long disjunction then one can always avoid having to cut across the distinctions given by a taxonomy of processes, but if the taxonomy that one needs to adopt is a gerrymandered one then the resulting theory of attention has lost its process-first character. The second reason why the argument is not conclusive is that it depends on the claim that attention can sometimes be wholly constituted by feature binding processes, or by the task-switching processes involved in random number generations. I think that the empirical data on feature integration warrant this interpretation, but it is always possible to take issue with it. One might claim that, despite the evidence mentioned above, the attention paid to serial search tasks is never constituted by feature-binding processes alone, but only by those processes together with some others, which, if they too had been included among the processes that took place in the neglect patient, would have made it the case that the neglect patient was attending. One might also claim that attention is not constituted by feature binding at all, but by quite different processes that happen to bring feature binding along with them. My claim that feature-binding processes constitute the attention of normal subjects in certain serial search tasks is based on an inference to the best explanation of diverse empirical data showing that subjects in speeded search tasks have bound representations of all and only the stimuli to which they have paid attention. But inference to the best explanation is never a conclusive matter, since one can never tell whether some further explanandum might come along that tips the balance in favor of some other view. The particular example of feature binding is not, and cannot be, a conclusive proof that the supervenience claim required by the processfirst view is false. Nor is the particular example of task switching. But

Rejecting the Process-First View

41

what these examples suggest is a form of argument that, although it may be dodged in particular cases, will be sure to find its mark somewhere. The root of the problem that these particular cases point to lies in the fact that, in sparse and undemanding contexts, where the subject’s task is simple and his time is short, the processing that constitutes his attention can be relatively simple. Once one of these simple contexts has provided us with an example of a simple attention-constituting process we can then construct a new example where we keep that simple piece of processing fixed but situate it in a more complex context, where the subject is engaged in something more sustained and cognitively elaborate. Where there is sufficient other processing to swamp the process that we started with, or where, as in the unilateral neglect case, there is some deficit that prevents that process from being properly integrated with the rest of the subject’s cognition, we have an example where it is no longer feasible that that process constitutes the subject’s attention. This is a recipe for finding a pair of cases in which the same piece of processing is in one case attention constituting and in the other case not. The problem for the process-first view of attention’s metaphysics arises from any one such pair.

3.3 THE INTERFERENCE ARGUMENT: A PUTATIVE PROBLEM FOR ADVERBIALIST ACCOUNTS Having just seen an argument against the process-first view of attention, I want now to consider an argument that seems to speak in its favor. A theory of attention must capture at least some of the features of attention that are apparent to common sense since it must be a theory of the thing that we refer to in our commonsense talk when we use the word ‘attention’. Christopher Peacocke, in his 1998 discussion of attention and self-knowledge, identifies a feature that common sense attributes to attention and that may require the truth of something like the process-first account if it is to be explained in a parsimonious way. This appears to provide a counterargument that the defender of the processfirst view could make against his adverbialist rival. Peacocke presents the argument like this: There are [. . .] some explanatory consequences of the hypothesis that there is a single general kind of attention of which perceptual and sensational attention, and conscious thought and imagination are all subspecies. It is a familiar truth about attention that any one of these kinds of attention can interrupt any one of the others. Perceptual attention can be interrupted by conscious thought; conscious thought can be interrupted by external events which capture the thinker’s attention; either of these two subspecies of the occupation of attention can be interrupted by imagination; and so on. What we have here is not merely some family resemblance between varieties of conscious states,

42

Attention Is Cognitive Unison

but apparently some form of competition for the exclusive use of a single limited faculty of attention. The familiar facts about attention are explained if there is a single, suitable high-level resource, drawn upon either by perception, conscious thought or imagination, a resource with access to some of its own recent states and to memory representations generated by its own previous states. I do not say that it is absolutely impossible to explain any one of these facts in some other way. I conjecture, though, that other explanations will be ad hoc. (66–67)

Peacocke is right that in our everyday encounters with attention we expect the occupation of our attention by one thing to explain our failure to attend to another. He does not put his conclusion in our terms, but his suggestion is that an inference to the best explanation of this will show us that common sense carries a commitment to a process-first account, of the sort we have just seen reason to be suspicious of. Peacocke’s argument is not merely an argument for a process-first view. It is an argument for the strongest version of the process-first view: the version according to which attention corresponds not merely to a disjunctive set, or genus, or family of processes, but to a single process, implemented by a single limited-capacity processing resource. The argument trades on the idea that the best explanation of interference is competition, and on the idea that it is only if the different varieties of attention compete for the same underlying cognitive apparatus that we will be able to offer a unified explanation of the ways in which different forms of attention interfere with one another. In the passage just quoted Peacocke suggests that if there is no ‘single limited faculty of attention’, then the cognitive bases for attention will be associated by no more than ‘merely some family resemblance’. This overlooks the possibility that attention could be an adverbial phenomenon. That is an oversight, but not an especially damaging one. Even when we take adverbialist approaches into account, the process-first account continues to have the best explanation of universal interference between modes of attention. This is not because the adverbialist cannot explain interference effects at all. The adverbialist about attention can explain some cases of interference between different instances of attention. Two tasks that involve the same cognitive resources will interfere whether or not those resources are specific to attention. The adverbialist therefore has no difficulty explaining why it is difficult to pay attention to two tasks both of which require manual coordination, say, or both of which require working memory. The adverbialist can also explain interference if it occurs between a pair of tasks that involve the processing of very similar forms of representation. This interference can be explained as a matter of ‘cross talk’ between separate information processing pathways that the brain fails fully to encapsulate. What the adverbial approach lacks is a good reason to expect that any two attended tasks will interfere with one another, even if they have nothing much in common apart from the fact that they are both attended. The adverbialist lacks a reason

Rejecting the Process-First View

43

to suppose that one or other of his ways of explaining interference will always apply. A strong version of the process-first account, on the other hand, can readily explain why any two instances of attention will interfere, even if the two attentively performed tasks have nothing else in common. If, as the strong version of the process-first account says, there is a process specific to attention, then just in virtue of the fact that two tasks involve attention, there is some one sort of processing that both involve, and so these tasks will interfere with one another by competing for the resources that implement that processing. We should therefore grant Peacocke the claim that interference between different kinds of attention is a phenomenon that presents itself to common sense, and we should grant that it is a phenomenon that adverbialism (or a family resemblance version of the process-first account) has difficulty explaining, but that the strong process-first account predicts. If common sense were right in identifying this phenomenon, then that would give us a good reason to favor the strong process-first account, and to reject the doubts about process-first accounts that were generated by the failure-of-supervenience argument given above. On this point, however, it has long been known that common sense turns out to be wrong. Common sense is not completely wrong. There is interference between different forms of attention. The problem is that it is not completely ubiquitous interference, occurring between any two attention-demanding tasks. When the interference to be explained is not interference of a completely ubiquitous sort it is no longer the process-first theorist who has the explanatory upper hand. The process-first theory’s explanation implies that interference should occur between any two attentiondemanding tasks. The pattern that we actually find as to when tasks can and when they cannot share attention is not the pattern that one would expect to find if there were a single general-purpose processing resource being divided between the tasks. To see this, note first that the advocate of the strong process-first approach wants to trade on the simplicity of the explanation according to which any two tasks that involve attention involve the attention process, and so compete for the resources that implement that process, and so interfere with one another. Consider the way in which this explanation applies to the interference that exists between the task of sightreading piano music and the task of counting how many times one has played a note with the index finger of one’s left hand.1 These two tasks are each attention demanding, and so, according to the strong version of the process-first view, there is some one process (the attention process) that each task must be served by if it is to be done effectively. The process-first view seeks to explain the interference between the tasks as being the result of competition for the resources that serve this common process. Or, to put the process-first account’s purported explanation in a somewhat more regimented way, the task of sight-reading and the

44

Attention Is Cognitive Unison

task of finger monitoring interfere because if one takes the attentional resources of a normal person and uses up as much as is needed by sightreading, then the resources that are left over are not sufficient to meet the demands of finger monitoring. If ‘SR’ denotes the amount of attentional resources required by sight-reading, ‘FM’ denotes the amount of attentional resources required by finger monitoring, and ‘T’ denotes the total attention resources available, then the process-first account’s explanation of the competition between these tasks requires that SR + FM > T. Some early work by Alan Allport and his collaborators shows this explanation of interference not to be the right one. Allport, Antonis, and Reynolds’s 1972 ‘disproof of the single channel hypothesis’ for attention showed, rather surprisingly, that the performance of an attentiondemanding task, such as the auditory shadowing of speech,2 was not at all detrimental to the concurrent performance of certain other attentiondemanding tasks, such as the sight-reading of piano music. The advocate of the process-first account, if he is to maintain his favored explanation of interference, must now maintain that the resources that remain available after the resources required by shadowing have been used up are sufficient for sight-reading. If we call the resources required by auditory shadowing ‘AS’, then the process-first account is now committed to the claim that AS + SR ≤ T. The two claims to which the process-first account is committed clearly imply that the resources used by finger monitoring are greater than the resources used by auditory shadowing (FM > AS). If this is true then any task that auditory shadowing interferes with must also be a task that finger monitoring interferes with (since it follows from FM > AS that, for any task, x, if we use ‘X’ to denote the resources required by that task, then X + AS > T only if X + FM > T). But this seems wrong. There are surely tasks that can be attentively performed concurrently with attentive finger monitoring but that cannot be done attentively when one is giving attention to auditory shadowing. Sight-reading involves a visual stimulus and a manual response. Shadowing involves an auditory stimulus and a verbal response. In both cases the response is strongly associated with the stimulus, and so the subject does not need to deliberately select the appropriate response. This difference between the stimuli in the two tasks, and between the responses in the two tasks, seems to play a role in explaining why it is that they do not interfere with one another. The lesson from Allport et al.’s research is that, in cases where the stimulus/response mappings involved in a pair of tasks are sufficiently familiar and sufficiently different from one another, the difficulty of either task does not affect the performance of the other. We noted above that the adverbialist could explain interference effects when the interfering tasks drew on common resources. What the adverbialist, unlike the defender of the process-first approach, could

Rejecting the Process-First View

45

not explain was why such interference effects are fully ubiquitous. To Peacocke this seemed like an explanatory advantage of the process-first approach. But what Allport et al.’s results show is that interference effects are not fully ubiquitous. This does not merely defuse an argument against adverbialism; it also makes for a point against the strong process-first view. Because the attentional interference effects are not ubiquitous they cannot be attributed to a single, general-purpose, limited capacity resource that is fully occupied when attention is employed. The kind of explanation that the adverbialist can offer looks like it must be the right one. That the sort of explanation available to the adverbialist is correct is the orthodox view among many psychologists: Botvinick, Braver, Barch, Carter and Cohen (2001) cite Cohen et al. (1990), Duncan (1996), Mozer (1991), Mozer and Sitton (1998), Navon (1985), Navon and Miller (1987), and Schneider and Detweiler (1987), as all agreeing with Allport’s 1987 claim that ‘the behavioral phenomena attributed in the past to the limited capacity of a central processor are more appropriately conceptualized [ . . . ] as the expression of crosstalk interference between parallel processes’ (Botvinick et al. 2001: 625). The argument here does not depend on a great extrapolation from a single old result, nor do we need to claim that inference effects are always cross-talk effects and never the result of competition for central resources. The argument we are interested in blocking has as a premise that ‘any one of these kinds of attention [perception, conscious thought or imagination] can interrupt any one of the others’. To block this we need only show that some kinds of attention do not interrupt one another. Although their use in other contexts may be controversial, Allport et al.’s results do at least show that.

3.4 CONCLUSION In chapter 2 we saw that the process-first view of attention is committed to a certain supervenience claim. In this chapter we have seen that that supervenience claim might very well turn out to be false. We have also undermined an argument that might have been taken to motivate the process-first view. Since the argument here depended on some of the metaphysical maneuvering of chapter 2, it may be worth taking a moment to clarify its form. The argument does not move from the fact that there are multiple attention-constituting processes to the lemma that attention is not identical to any particular process, and from there to the conclusion that attention does not belong in the metaphysical category of processes. Nor does it move from evidence that the instantiation of the featurebinding process is neither necessary nor sufficient for attention to the lemma that attention is not the feature-binding process, and from there

46

Attention Is Cognitive Unison

to the conclusion that attention is not a process at all. The first of these arguments would establish only multiple realizability, and the denial of the process-first metaphysics is, as we have seen in §2.4, a much stronger claim than that. The second move would be a patent non sequitur, inferring a universal from an existential. The argument that this chapter developed was of a different form from both of these. Its form was this: 1. If attention is a cognitive process, then, for all events x and y, if x and y instantiate the same cognitive processes, then if either one of them is an instance of attention, the other is too. 2. There are some events that are instantiations of the featurebinding process and that are instances of attention. 3. There are some events that are instantiations of those same feature-binding process and that are not instances of attention. Therefore: 4. Attention is not a cognitive process. The first premise of this argument is a consequence of the metaphysical work that we did in chapter 2. The second premise I take to be established by evidence from Anne Treisman and her colleagues, showing that there are instances of attention constituted by instantiations of the feature-binding process. The third premise I take to be established by evidence showing that some unilateral neglect patients must be binding the feature of items to which they cannot attend. I have suggested that we can also run the argument with ‘task switching’ substituted for ‘feature binding’. In this case, the grounds for believing the second premise comes from Alan Baddeley’s work on the attention demands of random number generation, and the grounds for believing the third premise comes from a thought experiment in which we imagine task-switching processes operating in a context where the subject’s task requires single-mindedness. In each case, it is the nonsufficiency of instantiating the same process as an instance of attention that enables us to infer that the matter of whether or not some event instantiates attention is not settled by the facts about the cognitive processes that the event instantiates, and it is on this basis that we conclude that attention is not a cognitive process. If these arguments are sound, then we have good reason to look away from process-first approaches and toward adverbialism. We have seen that adverbialist theories can be found in the work of Bradley and of Gilbert Ryle and Alan White, but so far I have said nothing on my own behalf about how the adverbialist alternative might be pursued and defended. In chapter 4 I give my own adverbialist theory of attention.

4 Cognitive Unison

4.1 INTRODUCTION An adverbialist theory of attention needs to identify a manner of occurrence (or a disjunctive set of such manners, or a family of them) that all and only instances of attention share. In chapter 1 we surveyed a history that included three attempts to produce such a theory, two from the twentieth century and one from the nineteenth. The arguments of chapter 3 suggest that each of these theorists was batting on the right side, metaphysically speaking. But the versions of adverbial explanation that those theorists offered face difficulties nonetheless.

4.2 THE PROBLEM WITH ATTITUDE-BASED ADVERBIALISM The version of adverbialism that F. H. Bradley proposed in 1886 was never worked out in any detail, but the details that Bradley did provide seem to be details of the wrong sort. Bradley’s proposal was that ‘any function whatever of the body or the mind will be active attention if it is prompted by interest and brings about the result of our engrossment with its product’ (316). It is easy to generate counterexamples to this proposal. If we consider tired or distracted thinkers we can find cognitive activities that are prompted by interest but that are not attentive. We can also find activities that are not prompted by interest but that are attentive. On a hot and muggy night the sleepless traveler finds that his attention keeps turning listlessly to the movements and sounds made by the fly that is circling around his ceiling. He is giving the fly his attention, but he is very far from being interested in it. If he were interested, then the way it holds his attention might be less irritating. Prompting by interest, then, is not necessary for attention. A similar example shows it not to be sufficient. Consider the devoted scholar who has spent too long at his desk. His reading of the book before him is prompted by genuine interest, but having spent the whole day trying to ignore the pangs of hunger, the children in the next room, and the radio that is playing downstairs, he is no longer able to concentrate. His reading is prompted

47

48

Attention Is Cognitive Unison

by interest, but it is not attentive. Prompting by interest, then, is not sufficient for attentiveness. The examples of tired or distracted thinkers show that a thinker’s attention and her attitudes can drift apart. They suggest not only that Bradley’s specific proposal about interest and engrossment was wrong, but that any attempt to identify attentiveness with the presence of some attitude on the part of the attentive subject would be vulnerable to similar counterexamples. The presence of a personal-level attitude seems to be the wrong sort of thing to serve as an adverbialist’s theory of the attentive manner. Bradley’s version of adverbialism about attention, in looking to the attentive thinker’s attitudes for its account of attentiveness, seems to have been looking in the wrong place.

4.3 GILBERT RYLE AND ALAN WHITE The two twentieth-century adverbialist theories outlined in chapter 1 were owing to Gilbert Ryle and to Alan White. Ryle’s theory was that the fact that someone is paying attention to something is not just a fact about what the person is doing; it is also a fact about how the person’s activities are situated relative to his various dispositions to do that thing (whatever the thing in question happens to be) and to carry on doing it, or doing something similar, in the face of potential distracters. In Ryle’s words: To describe someone as now doing something with some degree of some sort of heed is to say [. . .] that he is actually meeting a concrete call and so meeting it that he would have met, or will meet, some of whatever other calls of that range might have cropped up, or may crop up. (1949: 141)

White’s view contrasts with Ryle’s in being less concerned with what the thinker might have done had things been different and more concerned with what the thinker actually did. Instead of claiming that an agent’s attending to something is a matter of how his activities relate to what he was disposed to do, White claimed that an agent’s attending is a matter of how his actual activities relate to the various other things that he is currently doing: Simply to say that someone is attending, or paying attention, gives us no more clue as to what activities he is engaged in than simply to say that he is practising. What ‘attending’ tells us is that his activities and energies, whatever they are, are directed to and focused on something which occupies him. (1964: 7)

When we met these views at the end of chapter 1, it was mentioned that White has an explanation-based argument against Ryle’s disposition-based approach. The argument seems to me to be a sound one. And,

Cognitive Unison

49

since the view that this chapter develops sides with White, rather than with Ryle, it is worth taking a moment to spell it out.

4.4 WHITE’S ARGUMENT AGAINST DISPOSITION-BASED ADVERBIALISM The fact that an agent is attending to what she is doing can explain certain features of her performance. It can, for example, explain the fact that the performance is successful, or the fact that the performance is sustained. But it cannot explain the fact that the performance is taking place at all. One’s attention to one’s reading might explain the fact that one follows the argument or the fact that one is gripped by the plot, but it cannot explain the fact that one is reading in the first place. The point is not that reading can never be explained by attention. In cases where there is some larger task, of which reading is a part, it may be that the agent’s attention to the larger task explains the fact that she reads. A student’s attention in class, for example, may explain why she, but not her inattentive classmate, reads the footnotes on the handout. What cannot be allowed is for a piece of reading to be explained by the agent’s attention to that very reading. In general, the attention with which someone does something cannot explain his doing of that very thing. This no-self-explaining limitation is quite easily accounted for by a theory like White’s since, according to White, to be attending is to be actually doing something in a certain way. Since the way depends for its existence on the doing that instantiates it, the way cannot explain that very doing. The limitation is much less easily accounted for—is not, in fact, accommodated at all—by a theory like Ryle’s. Ryle tells us that one does something attentively if and only if one does it with a certain background of dispositions for doing that sort of thing (and for keeping on doing that sort of thing, or something appropriately similar, in the face of various possible distractions). But dispositions to ϕ are precisely the sort of thing that can explain an occurrence of ϕing. The having of this explanatory power is essential to what dispositions are. Ryle’s theory therefore entails, in contravention of the no-self-explaining limitation, that the fact that a subject is doing something attentively should be perfectly suited to providing an explanation for the fact that he is doing that very thing. White illustrates this objection using a number of examples and analogies, but the point is clearest if we take Ryle’s own. According to Ryle there is an analogy between the way in which ‘that man is attending’ might describe a man who is reading and the way in which ‘that bird is migrating’ might describe a bird that is flying south (Ryle 1949: 142). The description of the man as attending, on Ryle’s view, situates his reading relative to a set of dispositions to read in these and in similar

50

Attention Is Cognitive Unison

circumstances, just as the description of the bird as migrating situates its flying relative to a set of dispositions to fly south in these and in similar circumstances.1 If that were right then the man’s reading could be explained by his attention, just as the bird’s flight can be explained by saying that it is migrating. Since attentive reading cannot be explained in that way, Ryle must be wrong to claim that attentiveness in reading is a matter of having dispositions to read.

4.5 THE COGNITIVE UNISON THEORY So much, then, for Ryle’s view, and so much for Bradley’s. Bradley’s adverbialism fails because he attempts to identify attentiveness with the presence of a personal-level attitude. Ryle’s adverbialism fails because he attempts to identify attentiveness with the presence of a certain background of dispositions that explain the attentive agent’s action. The version of adverbialism that White proposed was neither attitude based nor disposition based. It therefore avoids the problems that Ryle and Bradley faced, and so looks altogether more promising as a starting point from which the adverbialist might build. In fact, the positive view that White articulates looks so promising, and has such an air of obviousness, that one worries that it might turn out to have no real substance. White himself states his positive claims quite concisely: Degrees of attention are not to be explained as more or less intense engagement in one specific activity, but rather as concentrating more or fewer of our activities on the one object. Starving travellers are said to talk food, think food, live food. The less attentive reader may be giving only some of his attention to the book, either because, though he is using his eyes, he is not trying very hard to follow the writer’s argument or because he is dividing his attention by looking at the book and listening to the radio. Full attention to X consists not only in the range of activities that are centred on X but in the absence of activities concerned with things other than X. Hence, the common privative force of the notion of attention; one minds one’s own business by not minding that of other people. (1964: 7)

White’s positive theory of attention is given in the first and in the next to last of these sentences. The rest of the passage serves to illustrate and elaborate it. Everything White says looks plausible. But it does not, on the face of it, look like a metaphysically revisionary theory, adequate to accounting for a centrally important cognitive phenomenon. The straightforward plausibility of White’s remarks raises the worry that, once we have spelled out what is meant by ‘activities and energies’ being ‘directed and focused’, White’s analysis will turn out to be uninformatively circular: If ‘direction’ and ‘focus’ turn out merely to be synonyms

Cognitive Unison

51

for ‘attention’, then the plausibility of White’s claims will be owing to their triviality. We can assuage this worry by finding a noncircular way of cashing out these notions of ‘direction’ and ‘focus’ and of ‘activities and energies’ being ‘centered on X’. My proposal is that we get the noncircular account that we need if we understand ‘activities and energies’ to be bits of cognitive processing (in a sense that I shall make precise) and if we understand ‘focus’ as being analogous to unison (in a sense that, again, I shall make precise). The theory of attention that emerges when White’s view is developed is the theory according to which attention is cognitive unison. This chapter presents that theory. To get ourselves into a position from which we can see how the cognitive unison theory works (and can see that it really is the metaphysically revisionary theory that the previous chapters have told us to be looking for), it will be helpful to have before us a more regimented statement of unison’s requirements, so that we may consider in turn its different elements. Here, then, in an unlovely but to-be-explained form, is the definition of cognitive unison—the notion from which the theory of attention as cognitive unison is to be built: Let α be an agent, let τ be some task that the agent is performing, and call the set of cognitive resources that α can, with understanding, bring to bear in the service of τ, τ’s ‘background set’. α’s performance of τ displays cognitive unison if and only if the resources in τ’s background set are not occupied with activity that does not serve τ.

This analysis, as it stands, is rather laden with terms of art. Definitions are needed for ‘task’, for ‘cognitive resource’, and for the notion of ‘bringing to bear with understanding’. Those definitions will be given in the course of this chapter. Once they are in place, we will have our account of ‘cognitive unison’. Turning this into a general theory of attention will require a little bit of philosophical maneuvering. We could get a simple theory of attention if we just added an equivalence claim: α performs τ attentively if and only if α’s performance of τ displays cognitive unison.

But that would be correct only as an account of what it is for the performance of a simple task to be attentive. A slightly more complex theory is needed before the theory can apply to the attention that one might give to a larger scale task, and before it gives us an account that applies to cases of partial and divided attention. The theory also needs to be supplemented before it provides us with an account of what it is for an agent to be paying attention to one object rather than another. By the end of this chapter we will have all of these additional details in place. We begin by defining the basic components.

52

Attention Is Cognitive Unison

4.6 TASKS A subject’s ‘tasks’, as these are to be understood here, are the things that the subject is in the business of doing and that she is active with. To specify the tasks in which an agent is engaged, we adopt the agent’s point of view on her own activities. Normal human tasks are such things as making a cup of tea, following a conversation, or looking for the car keys. They are activities with natural descriptions of a sort that the subject would typically accept as a description of her goal. Not all subjects will be able to describe the goals that their tasks aim at: Some subjects lack the necessary linguistic competence, and some goals are hard to describe. Nor will it generally be the case that there is just one answer that an agent can sincerely provide to the question of what she is doing. But an agent must have some understanding of what it is that she is about. What is required, in order for τ to count as a task of α’s, is for it to be the case that α has some understanding of τ on the basis of which her performance of τ is guided. This notion of guidance by understanding carries a lot of the weight at various points in the cognitive unison theory, so we shall be coming at it from various angles as this chapter progresses. For present purposes, the point to note is that it is guidance by understanding that gives the cognitive unison theory its official definition of ‘task’: A task is defined as an activity of the agent’s the execution of which is under the guidance of the agent’s understanding of that activity. An activity fails to be a task if it is not agent involving, if the agent has no understanding of that activity, or if the agent has an understanding but his understanding does not guide the way in which the activity is performed. This definition of ‘task’ is intended as a regimentation of commonsense usage rather than as an innovation. The notion is one that we have in our commonsense repertoire of psychological notions, although it is not one for which any particular term in our commonsense vocabulary has been reserved. The notion is also familiar in the psychological literature, where it figures in discussions of ‘task set’ (see, for example, Allport, Styles, and Hsieh 1994). In the empirical literature on ‘task set’ the notion of an agent’s task has not been spelled out in any detail. This is unproblematic because the tasks of agents in the lab simply correspond to whatever it is that the experimenter has told those agents they should do. Lab tasks are activities such as naming the color of the ink in which a list of words is printed, or adding three to each of the numbers in a list. Outside of the lab an agent’s tasks will tend to be more numerous, complex, nested, and overlapping. As a result the question ‘What task is this agent performing?’, when asked outside the lab, may admit of many answers, some of which are only vaguely true. Since it is part of the cognitive unison theory’s adverbialist approach that facts about manner of task performance are the grounds for facts about attention, answers to the question ‘To what is this agent attending?’ will inherit some of this vagueness.

Cognitive Unison

53

One might think that giving tasks a foundational role in the theory of attention makes for a theory that applies only to attentive action, and that has nothing to say about perceptual attention, but the definition of ‘task’ that we have just given does not have this consequence. It counts looking at a chair as a task of the agent’s (because the way in which the episode of an agent’s looking develops is guided by the agent’s understanding of looking). Lifting a chair, hammering a nail into a chair, designing a new kind of chair, remembering the details of a particular favorite chair, describing a chair—all of these are tasks in the intended sense. They are things that an agent does with an understanding of what it is that she is doing, and this understanding plays a role in determining the way in which the performance takes place. Seeing a chair typically is not a task of the agent’s, because episodes of mere chair seeing typically proceed without the agent deploying his understanding of seeing. Tripping over a chair typically is not a task either. Nor is being reminded, involuntarily, of a chair from one’s past. The subject who trips over a chair may have an understanding of what tripping is. Having such an understanding does not, by itself, mean that when she trips she has tripping as a task. The tripping becomes the subject’s task if and only if her understanding of tripping guides the activity, as it might if she is an actor feigning a trip, or if mid-trip she sees what’s happening and is able to take remedial action. Then, but only then, can we say that her tripping has taken up her attention. It is only when the tripping is guided by understanding that it is a task of the agent’s, and only when it counts as a task that it is something that can be done attentively. It would be a mistake to suppose that understanding guides an activity always or only when the activity is voluntarily performed. A subject’s understanding of tripping can guide her tripping even when the subject has not voluntarily undertaken to trip. Similarly, a subject’s understanding of looking can guide her engagement with the visual features of a thing even when the subject has not voluntarily undertaken to look at it. One’s attention can be caught involuntarily by a sudden sound at the window. If the way in which that episode of looking to the window unfolds is guided by one’s understanding of what it is that one is doing in looking to the window, then the looking is a task, although not a task that one performs voluntarily. Because it is a task the looking is something that can be done attentively. If, on the other hand, one is distracted by a sound at the window without the processing of that sound being integrated into any understanding-guided activity, then the case is one where the sound is merely distracting and is not itself attended. The distinction between tasks and nontasks should not be mistaken for the more philosophically notorious distinction between actions and mere bodily movements. The latter distinction is typically understood as being a distinction between two different sorts of bodily event—my arm going up because it is blown by the wind, say, and my arm being raised by me. The things that we want to separate into tasks and nontasks

54

Attention Is Cognitive Unison

are not all of them movements of the body, and not all are executed as acts. Some tasks involve bodily engagement, but others are more purely epistemic and involve the body rather little. Some tasks are voluntary acts, but others are things that we find ourselves engaged in.

4.6.1 The Understanding of Primitive Creatures It is important that the notion of a task be relatively undemanding, since it must be undemanding enough to allow infants and animals, in so far as we think of these as paying attention, to have tasks. The cat can have stalking its prey as a task (and so can be counted by the cognitive unison theory as stalking attentively) because the way in which it stalks is under the guidance of the cat’s understanding of stalking. The infant’s perceptual exploration of the world can be among his tasks (and so can be performed attentively) because the infant conducts himself in a way that employs an understanding of how to explore things perceptually. An objection might be raised about the putative role for understanding in the guidance of the infant’s perceptual exploration and of the cat’s staking behaviors. The objector might complain that the cat does not really understand stalking at all, that its stalking behaviors are prescribed by genetic endowment, not guided by understanding. If that were right then stalking would not be counted as a task that the cat has, from which it would follow that the stalking cat could not be counted as a case of cognitive unison, and so could not be counted by any unisonbased theory of attention as being attentive. That would be a counterintuitive result, but it would not be a result that could be used to generate a compelling objection to the cognitive unison theory. The cat’s stalking certainly appears to be attentive, but the situation where the cognitive unison theory would have to say that it is not—the situation where the stalking turns out to be a genetically given tropism—is a situation where our faith in the authority of this appearance would be undermined anyway. It is not an objection to the cognitive unison view to point out that it contradicts appearances in circumstances where those appearances are already called into question. As it happens, however, we do not need to reconcile ourselves to going against appearances in this case. This is because there is room for a notion of guidance by understanding according to which the requirements of understanding vary depending on the capacities of the creature in question. On this notion, the cat and the infant, slight though their abilities are, can count as acting under the guidance of understanding and so can count as performing tasks (and so they are not prevented from instantiating cognitive unison, and so the cognitive unison theory is not obliged to rule that they cannot pay attention). This creature-relative notion of understanding can be approached by considering the difference that understanding makes to the performance of the thinker who possesses it. Imagine that we have a pair of

Cognitive Unison

55

thinkers, each of whom is asked to prove Pythagoras’s theorem, and each of whom, in response to this request, utters just the same sentences and produces just the same diagram. Imagine, also, that these sentences and diagrams do indeed represent a proof of the theorem. From the description of the case given so far, we cannot tell whether either, or neither, or both of these thinkers understand the theorem or its proof. The two thinkers may produce matching responses, but it may be that one of the thinkers understands the theorem while the other performs parrot fashion, or just strikes lucky in the course of bluffing. Since there is no limit to the sorts of things that might be learned by rote, and no limit on the luck that a bluffer might happen to have, there are no particular responses for which we can say that the thinker with understanding will produce those responses while the thinker without understanding will fail to produce them. What the thinker who lacks understanding lacks is not a cognitive state that enables him to make an appropriate response (he may have such a state if he has been appropriately drilled). What he lacks is a capacity to redeploy the cognitive basis of his performance in a way that enables him to deal with alternatives.2 If the thinker really does not understand at all, then the cognitive basis of his performance will not equip him to deal with alternatives at all, not even with trivial variations, although some other cognitive state, imparted by some other piece of drilling, might. It may be that if we say ‘prove the Pythagorean theorem’, then the thinker without understanding knows, because he has been drilled, which moves to make, but, because these moves are not founded in any kind of understanding, he may be stumped if we say ‘show that the square of the hypotenuse is equal to the sum of the squares on the other two sides’. A thinker with a little more understanding might be able to deploy that same cognitive basis so as to respond appropriately to the alternative formulation of the question, but this thinker may still be stumped if we ask him to how to calculate the distance between any pair of points on a plane, given their Cartesian coordinates. A different thinker might have been able to answer that question purely on the basis of her understanding of the theorem. But, in a setting where the theorem is normally put to some quite different use than the calculation of distances between points, even this last thinker might count as failing to really understand the theorem. All thinkers who understand Pythagoras’s theorem will, on account of their understanding, be able to deal with some alternatives. Some of those thinkers will be able to deal with lots of alternatives. Others will be able to deal with only a few. Adding to one’s understanding always enables one to deal with more alternatives, but there is no reason to suppose that the sets of alternatives that different understanders of Pythagoras’s theorem are able to deal with on the basis of their understanding will be sets that nest inside one another. It is therefore not the case that, for any two thinkers, ϕ and ψ, if there exists an alternative situation, e, such that ϕ’s understanding

56

Attention Is Cognitive Unison

of F enables him to deal with e and ψ’s understanding of F does not enable him to deal with e, then ϕ understands F more than ψ does. Understanding does not come simply in degrees, with the bar on what counts as ‘really understanding’ being set at different heights on different occasions. Instead, what we find is that the abilities that an understanding conveys may be quite different for different creatures, depending on the repertoire of alternatives that the creatures may be expected to negotiate. Where that repertoire is small, as it may be for the cat and the infant, the bar on understanding may be easy to get over. The underlying thought here lends itself to expression in Wittgensteinian terms: Understanding is that which enables one to deal with novel situations, but the particular range of novel situations with which one is equipped to deal will depend on the sort of interaction that one typically has with world. Whether the cognitive state responsible for one’s going on is a state that enables one’s going on to be a going on with understanding depends on one’s form of life. Questions about whether some intellectual capacity on the part of an agent counts as an understanding therefore need to be relativized to a specific range of novelty appropriate to the agent’s form of life. There are various ways in which a relativized notion of understanding might be spelt out. The beginnings sketched above are already enough for us to see the points that we need to make concerning understanding and guidance by understanding, as those notions figure in the definition of ‘tasks’, and so in the analysis of attention. If the agent we are interested in is a cat, and if its form of life is one in which the catching of mice serves to feed it and its young, and not a great deal else, then having a genetic endowment that enables it to catch mice in ways and at times that are appropriate to those purposes may be enough to qualify it as having an understanding of mouse catching. If, on the other hand, the agent we are concerned with is an adult human, with a form of life in which mouse catching capacities might be called on for all sorts of reasons, then the requirements for understanding are going to be much more demanding. 3 If our theory of understanding is one in which the requirements of understanding are relativized to forms of life in something like the way just sketched then the cognitive unison theory’s consequence that only creatures capable of being guided by their understanding are able to attend will rule out many fewer creatures from the class of potential attenders than one might at first have thought.

4.6.2 The Unison-Instantiating Performance of Nontasks Is Not Attention The notion of tasks that is in play here is, as we have just seen, undemanding in the ways that it should be. But it is not totally undemanding. It does

Cognitive Unison

57

enable us to rule out some cases that might otherwise be problematic. Consider, for one such example, the case of daydreaming. Suppose we have a daydreamer whose dreaming is not done as part of any deliberate relaxation technique, or with the intention of opening himself to inspiration. He is just letting his mind wander in an idle moment. Perhaps he is experiencing the state that William James claims ‘most people probably fall [into] several times a day’: The eyes are fixed on vacancy, the sounds of the world melt into confused unity [. . .] the foreground of consciousness is filled, if by anything, by a sort of solemn sense of surrender to the empty passing of time. (1890: 382)

If the thinker is in a state like this then more or less anything that goes through his mind will serve as a contribution to his general reverie—an idea from memory, a verbalization of the content of the image in the mind’s eye, an awareness of the cooking smell coming from downstairs. All sorts of processing resources could serve the subject’s daydreaming, and all sorts of activities are such that, if they were to occupy those processing resources, then they would contribute to the subject’s daydreaming. More or less any state of the daydreamer therefore looks like it will answer to our definition of ‘unison’. This raises the threat that no theory of attention based on cognitive unison will be able to classify cases of inattentive daydreaming as being inattentive. The threat here is merely apparent because the cases in which there really are no constraints on which processes do and which do not serve the daydreaming are cases in which the daydreaming is not under the control of the subject’s understanding. In such a case the subject does not have daydreaming as a task at all. If so, then daydreaming is not something that this subject can do with cognitive unison, and there is no risk that a cognitive-unison–based theory will count the daydreamer as performing attentively.

4.7 COGNITIVE PROCESSES In the analogy between attention and unison it is ‘cognitive processes’ that play the role analogous to that played by the individual performances of the members of the orchestra. These processes are the activities whose unified occurrence constitutes an instance of attention, much as it is the individual playings of the musicians that, taken together, constitute an instance of orchestral unison. In order to play this role, ‘cognitive process’ needs to be understood in a fairly demanding sense. A cognitive process, in this sense, is a process that operates on representations that encode their contents for the agent of the task: In order for a process to count as cognitive, there must be an agent-level contentful state whose content is directly determined, at

58

Attention Is Cognitive Unison

least in part, by the content of the representations on which that process operates. More precisely, a process is a cognitive process if and only if it is an immediate consequence of the fact that this process is operating on a representation with the content . . . ξ . . . that the agent is in a position to believe that . . . ξ . . . , or to wonder whether . . . ξ . . . , or to remember . . . ξ . . . , or see that . . . ξ . . . , or to have some other attitude toward the content of the representation in question. The agent-level attitude need not be a propositional attitude, and the content of the representation in question need not be a complete proposition. In some cases the representation may be of an object, or of a property, but it may lack a form that combines these in such a way as to have a full proposition as its content. The representation may simply be of an abstract object, as it might be if there is a melody running through one’s head despite the fact that one is entertaining no propositional thoughts about that melody, or if one is casually contemplating some conjectured mathematical entity, having not yet decided whether that entity has any particular properties, or even whether it exists. Such bits of thinking are genuinely agent level, in the sense that it is the agent (in this case, a person), and not some proper part of him, who entertains the melody, or who contemplates the abstract object. The representation-handling processes that enable such bits of abstract thinking therefore count as cognitive processes by the standards of the definition that is being advocated here. It follows that unison among those processes can count as attentiveness. And this is just how it should be: Attention must, albeit in an existentially noncommittal way, be attention to something, but attentive thinking need not be the kind of thinking that has a whole proposition as its content, or that has any ‘direction of fit’ with the thinker’s environment.

4.7.1 Noncognitive Processes We have said that the definition of ‘cognitive process’ in play here is demanding, in the sense that it demands that there be a subject of experience who is put into a position to entertain the contents represented by the processes in question, and we have said that it is undemanding, in the sense that it does not place constraints on what those contents are. This definition of ‘cognitive process’ is demanding enough to rule out brain processes such as those that monitor the osmotic balance of the blood, the presence of hormones, and the rate of one’s breathing. The hypothalamus, for example, may contain representations of the blood’s osmotic balance, but those representations do not determine the content of a personal-level state in such a way as to qualify the processing that takes place there as cognitive processing. Even if we suppose that the subject is a biologically informed and skilled self-monitor who is able to make judgments about the osmotic balance of his blood, he is not put in a position to make those judgments by the representations of osmotic

Cognitive Unison

59

balance in his hypothalamus. He has to learn how to do it, and his learning is partly a matter of acquiring new representational resources. Brain processes that are not cognitive processes in the current sense may operate in some sort of unison without their unison being counted by the cognitive unison theory as a case of attention. There might be a whole battery of processing resources monitoring the heart rate, breathing, blood oxygen, and blood sugar of a jogger, for example, and another battery of resources maintaining his balance, monitoring the ground for obstacles, and controlling the movement of his legs. This set of processes is focused on the task of jogging, but, since the processes are not cognitive processes, this is not a case of cognitive unison, so it cannot be counted by a unison-based theory as attention. And that is just how it should be. One may jog while giving one’s attention, undividedly, to the question of what to have for dinner, and if one does pay attention to one’s jogging, it is not the state of one’s autonomic processes that explains why. Some readers of earlier presentations of these ideas have attempted to use an example such as this to contrive a problem case for the cognitive unison theory by designing a situation in which the subpersonal processes that do not play a role in attention would nonetheless satisfy our definition of ‘cognitive processes’. The purported problem goes like this: Suppose that the hypothalamus’s representation of the osmotic balance of the blood takes some simple and easily measurable form. Suppose also that this representation is amenable to real-time brain scanning. In that case a subject may be forming a belief about the osmotic balance of his blood on the basis of a brain scan that reads the representation of this balance in his hypothalamus. The belief about osmotic balance that the subject forms from looking at the brain scanner is a personal-level state, and the content of that belief is partly determined (via the intermediary of the brain scanner) by the content of the representations in his hypothalamus. But even in this situation we want our definition to rule out hypothalamic processing as cognitive processing. If the subject in the brain-scan case is paying attention to the osmotic balance of his blood, it is the processes by which he reads the brain scan that constitute this attention, not the processing going on in his hypothalamus. The problem that such a case is intended to pose is easily dodged. In the imagined example the hypothalamic activity determines the contents of the subject’s belief only in a very indirect way, going via a brain scan. Changes in the content of the representation in the hypothalamus may occur without producing changes in the content of the belief (if, for example, the subject is no longer looking at the brain scanner). We can rule out hypothalamic processing from the set of cognitive processes, even in this scenario, by insisting that in order for a process to qualify as cognitive the determination of the content of the personal-level state must be so direct as to make it impossible for the two contents to dissociate in this way. This can be made precise by saying that, in order for the representation operated on by a process to count as directly determining

60

Attention Is Cognitive Unison

that a personal-level state, s, has content ξ, there must be no other representation such that (1) the other representation contributes ξ to the content of s, and (2) the other representation gets its content from the first representation. We do not have this sort of unmediated determination in the case imagined. In that case there are lots of other representations operating as intermediaries between the hypothalamus and the subject’s belief—the most obvious being the representations produced by the scanner itself. To say that cognitive processes are those processes that operate on representations whose possession puts the subject in a position to entertain thoughts with the content represented is not to say that all cognitive processes are involved in the subject’s occurrent thoughts. Among the processes that constitute a subject’s attention there may be some that operate on representations whose contents are, on this occasion, not contents about which occurrent thoughts are actively entertained. Perhaps the attentive birdwatcher, unlike his inattentive mate, can give a full report of every bird he saw last week. The processes by which these facts are laid down in memory are a part of what makes it the case that he is paying attention, even if they make no difference to the train of thoughts he entertains at the time when he is being attentive. The contentful states that a cognitive process supports need to be the agent’s, but they need not be occurrently entertained at the time when the processing occurs.

4.8 POTENTIAL SERVICE OF A TASK The set of resources that must operate without processing distracters in order for the subject’s performance to be counted as an instance of cognitive unison is the set of resources that potentially serve the attended task. A task, we have said, is defined as an activity of the subject’s that is guided by his understanding. Our account of what we mean by ‘potential service’ is closely related to that definition. A resource potentially serves a task if and only if, under the guidance of the subject’s understanding, it can now be brought to bear on the performance of the task. Or, to put it another way, a resource counts as one that the agent can bring to bear in ϕing at t if and only if, given the way the agent understands ϕing, it is possible for the agent to exercise his understanding by using that resource to pursue ϕing at t. Putting the point in this latter way makes it clear that the mere metaphysical possibility of using the resource to ϕ in some possible world is not enough. A resource potentially serves ϕing at time t only if, given the agent’s understanding at t of ϕing, that resource can be brought online, at t, to further the understanding-guided pursuit of ϕing. If a resource potentially serves a task, then there are nearby possible worlds in which

Cognitive Unison

61

that resource is being deployed now as part of the agent’s attempted performance of that task. And, for these purposes, worlds count as sufficiently nearby only if they are worlds in which the agent’s understanding of his task is kept fixed. This definition does not count every resource that the agent could use to pursue his ϕing as being a resource that is potentially at the service of his ϕing. Nor does it count all or only resources that, were the agent to use them, would actually contribute to his ϕing. The agent’s current understanding is what sets the standards of relevance to the task that a resource must meet in order for that resource to have a role in determining whether the task is attended. It is because of this that the resources falling into the background set for a task can differ from those resources that would in fact further the performance of the task, and even from those resources that are involved in what the agent knows to be his best strategy for performing the task. The first of these dissociations can be seen by considering cases of agents who, being under some sort of misapprehension, undertake impossible tasks. The second can be seen by considering cases of agents who are trying to perform tasks at which they are highly practiced.

4.8.1 Impossible Tasks When an agent attempts to perform an impossible task the set of resources that would in fact serve his task comes apart from the set of resources that he can, with understanding, bring to bear in the service of that task. It is the latter set that needs to display unison in order for there to be attention. This can be seen if we consider the example of a trainee clairvoyant. The trainee clairvoyant has the task of coming to know what is in the mind of the person sitting in front of him. He attempts to perform this feat telepathically. He understands that it is to be accomplished via processes of visualization and of imagination. Part of the clairvoyant’s understanding of his task is that his mind must be kept clear and receptive. Any speech sounds that might be going on must, he thinks, be blocked out. Suppose now that the thing that is in the mind that our clairvoyant is in the business of reading is in fact an image of three wavy lines, and suppose that the person whose mind it is, believing the experiment to be over, is saying out loud: ‘The thing I had in mind was three wavy lines.’ It is clear that the resources of imagination and visualization are in fact of no use whatsoever for coming to know what is in the mind of the person sitting in front of one. Clairvoyance of that sort is impossible. But it is these resources that the clairvoyant is trying to use, and it is the unison among the activity in these resources that makes the trainee attentive. The resources that would, in fact, enable the clairvoyant to know what is in the mind of the person sitting in front of him are those that

62

Attention Is Cognitive Unison

would be involved in listening to what the subject is saying. (Remember that the subject is saying, truly, that the thing he had in mind was three wavy lines.) But, because the trainee’s understanding of his task is one on which the verbal resources must be kept clear in order to for him to be telepathically receptive, the activity that would in fact further his goal does not contribute to his attention. Instead it counts as a form of distraction.

4.8.2 Overlearned Tasks It is not merely in cases where the subject is radically mistaken that the ‘background set’ of our definition comes apart from the set of resources that would be operating if the subject were performing successfully. A less contrived example than that of the trainee clairvoyant is that of the teenage actor. Our teenage actor, let’s call him Bob, is a perfectly normal teenager. He is perfectly capable of walking from one side of the room to the other in a normal manner. Being an actor, Bob is asked to walk from one side of the room to the other, in a normal manner, in front of a large audience. Being a teenager, Bob is quite self-conscious about this and is made anxious by the people watching. The result is that Bob’s walking is not a success. He gets across the room, but the walk is a self-conscious swagger. Its manner is far from normal. Bob’s problem, of course, is that he is employing resources of selfmonitoring together with ideas about what normal walking should look like. These are resources that he can, with understanding, bring to bear in the cause of walking normally across the room. When he walks in front of the audience he pays attention. That attention involves the bringing to bear of resources involved in self-monitoring. But these are resources that, although they are in the background set, are a hindrance, rather than a help, to the task of walking normally.

4.8.3 Mastery and the Capacity for Inattentive Performance The relationship between the subject’s understanding of a task and the background set for that task is worth emphasizing because it is this feature of the cognitive unison view that gives the theory its account of why some tasks are such that, as one’s understanding of them grows, one can complete them without attention. It is also crucial to the theory’s account of why it is typically harder to give sustained attention to familiar, well-understood tasks than it is to give attention to tasks that are poorly understood. The thought in both cases is that giving one’s attention to a well-understood task involves marshaling a large set of resources, just because the task is so well understood, whereas a part of what it is for a poorly understood task to be poorly understood is that the set of resources that one can bring to bear is small and therefore brings itself into unison relatively easily.

Cognitive Unison

63

These phenomena are ones that rival accounts, especially process-first accounts, have a difficult time explaining. When attention is regarded as the activity of some central processing resource there is a tendency to portray the development of familiarity with a task, and hence of the ability to perform that task inattentively, as if it occurred by the development of an automatic neural pathway that short-circuits the central attention mechanisms (see, for example, Passingham 1998). For most cases this seems an unnatural portrayal. The tasks that one has mastered to the point where one no longer needs to pay attention in order to do them need not be tasks that one now performs like an automaton. The development of mastery is a matter of understanding more and more, not of needing less and less to understand.

4.9 SUPERORDINATE TASKS Having clarified the notion of tasks, of cognitive resources, and of what it is for a cognitive resource to be in the background set for a task, we now have a reasonably clear idea of what it is for a subject’s performance to display cognitive unison. Cognitive unison, as we said in §4.5, is what happens when the resources in the background set for an agent’s task are not occupied with activity that does not serve that task. For simple tasks this unison qualifies the agent’s performance of the task as attentive. But it was acknowledged, before we launched into the preceding analysis of our theory’s components, that more needs to be said in order to turn the notion of cognitive unison into a theory of attention that applies to more complex tasks. Complex tasks create problem cases for the theory that equates attention with simple cognitive unison. They are problematic when the agent performing them has what we might call ‘superfluous competence’.

4.9.1 Superfluous Competence An example of the problem that superfluous competence creates can be seen by considering the task of calculating the difference between twelve squared and nine squared. A subject’s understanding of mental arithmetic may enable her to approach that calculation using either one of two strategies. The first strategy is that of straight calculation. One who approaches the calculation in this way thinks: ‘Twelve times twelve is one hundred forty-four. Nine times nine is eighty-one. One hundred forty-four minus eighty-one is sixty-three, so the difference between twelve squared and nine squared is sixty-three.’ The second strategy involves some preliminary visualization. One who approaches the problem in this way thinks: ‘Twelve squared is the area of a twelve by twelve square,’ and at this point the thinker visualizes such a square. ‘Nine squared is the area of a nine by nine square.’

64

Attention Is Cognitive Unison

Again the thinker visualizes it. ‘The difference between the two values corresponds to the area that would not be covered up if we placed the smaller square in the corner of the larger one. That area would be an L-shape with a bottom portion that is 12 long and (since 12 minus 9 equals 3) 3 wide, and with a side portion that is 9 long and (for the same reason) 3 wide. The area of the bottom portion of this L-shape is 36. The area of the side portion is 27. The area of the whole shape is therefore 36 plus 27. That equals 63, so the difference between 12 squared and 9 squared is 63.’ The first of these strategies places demands on memory that the second strategy does not, since it requires that the thinker recall the multiplication tables or that she perform a calculation of twelve times twelve. The second strategy places demands on visual imagination that the first strategy does not, since it requires that the thinker visualize the two squares and, in particular, that she visualize the shape and size of the portion of the larger square that the smaller one does not overlap with. The thinker with a good understanding of what the calculation requires may, on the basis of that understanding, see that both strategies are available. If she does then, at the first stages of her performance, the resources of visual imagination and the resources of recalling the multiplication tables will both be resources that she can, with understanding, bring to bear in the service of her task. The resources of visual working memory will be members of the initial background set for this task even if the thinker adopts the strategy that does not place demands on those resources. The simple cognitive unison theory therefore has the consequence that attention to the calculation places demands on those visualization resources even when the thinker has opted for a strategy that does not use them. This is problematic because, having adopted the pure calculation strategy, it seems plausible that, contrary to the verdict of the simple unison theory, the resources of visualization need play no role in the thinker’s attention to the task. The problem here is not specific to intellectual tasks such as mental calculations. It arises whenever a thinker’s understanding of a task equips him to perform that task in two ways that differ in the cognitive resources that potentially serve them. What we need, if a unison-based theory is to accommodate cases of superfluous competence, is a way of using the agent’s adopted approach to restrict the background set so as to rule out those resources that would only have been relevant had her approach been different. It would be fatal to the theory if we tried to do this by defining the background set for a task, not as the set of resources that the agent can bring to bear in the service of the task, but instead as the set of resources that are actually brought to bear as part of the agent’s approach to the task. Since it is plausible to assume that every set of resources that is being brought to bear on a task is a set of resources that are not engaged in activity that does not serve the task, every such set displays unison trivially. If the

Cognitive Unison

65

agent’s actual approach defined the background set then the cognitive unison theory would have the absurd consequence that all task performance is attentive. The fact that an agent has chosen a certain approach to a task needs to be used to restrict the background set for that task, without restricting it to just those resources that the agent is actually using. The first step toward seeing how this can be done is to note that the tasks that create the problem of superfluous competence are superordinate tasks: They are tasks that one performs by adopting a certain strategy, and these strategies are themselves activities that the agent performs under the guidance of an understanding of what they involve, and so they, too, qualify as tasks. When τ is a superordinate task, we can allow the following modification to the simple core of the cognitive unison theory: The agent performs τ attentively just in case there is some task υ that the agent understands to be a way of performing τ, and just in case the agent is performing υ on the basis of that understanding, and performing υ in such a way that the set of cognitive resources that the agent can, with understanding, bring to bear in the service of υ does not contain resources that are occupied with activity that does not serve τ. Since every agent who understands a task, τ, understands performing τ to be a way of performing τ, the simple case is just a special case of the superordinate case. (It is the case in which τ = υ.) The formulation of the cognitive unison theory that we have just stated can therefore accommodate both cases together, without needing to go disjunctive, but it can also allow that an agent counts as performing a superordinate task attentively if he is performing a subordinate task attentively, as part of a strategy for performing the superordinate task. This enables us to deal with the problematic case of finding the difference between a pair of squared numbers. The theory now says that the agent is paying attention to finding the difference between twelve squared and nine squared if and only if there is some task that the agent understands to be a way of finding that difference, and the agent is performing that task with resources that operate in cognitive unison. The thinker who opts for the straight calculation method does indeed understand straight calculation to be a way of finding the difference, and he does perform the straight calculation with resources that operate in unison, even if his visualization resources are being deployed elsewhere. (The resources of visualization are, as they should be, absent from straight calculation’s background set.) Similarly for the thinker who opts for the visualization method: That thinker understands the visualization method to be a way of finding the difference, and applies the visualization method with resources that operate in unison. She thereby qualifies as attending to the task of calculating. This solves our initial problem, in that it enables us to count agents with superfluous competence for the performance of a task as performing it

66

Attention Is Cognitive Unison

attentively. It also solves that problem without going too far in the opposite direction: It does not allow us to arbitrarily restrict the background set, and so does not leave us with the absurd result that every performance counts as attentive. That this second problem is avoided might not be obvious. It might seem that we would run straight into the problem of overrestricted background sets by allowing the subordinate task, υ, to be arbitrarily specific. Suppose, for an example of the sort that might appear problematic, that τ is the task of getting across town, and υ is the task of driving in a slapdash and carefree manner. Then, according to the cognitive unison theory as we have just formulated it, the agent may count as performing the task of getting across town attentively if he performs the task of driving in a slapdash manner using resources that operate in cognitive unison. The problem is that slapdash driving, however unified it might be, does not look like it ought to count as a way of driving attentively. How, then, does the cognitive unison theory avoid counting slapdash driving as attentive driving? The first point to notice in replying to this challenge is that there are some cases in which driving in a slapdash fashion is attentive. Consider a would-be teenage tearaway with a James Dean complex. He takes great care to drive in a slapdash fashion, thinking that the unfortunate girl in the passenger seat will be impressed. His understanding of what slapdash driving amounts to is what guides the whole performance. In that case, he has slapdash driving as a task, and if there is unison among the resources that could serve his slap dash driving, then that driving qualifies as something that is being done attentively. This result is the correct one. To describe the driver as ‘attending to his careless driving’ sounds paradoxical only on the face of it. In context the description is clearly correct: This driver is paying attention to his driving. The cognitive unison theory says that he is paying attention to his driving because there is a task (namely, slapdash driving) that he understands to be a way of driving and the background set for that task operates in unison. This does not mean that the cognitive unison theory is stuck with the conclusion that all slapdash driving turns out to be attentive driving. Really slapdash driving, to which the driver is not paying any attention, is not guided by an understanding of what slapdash driving is. It is guided, slapdashedly, by an understanding of what driving is. Because the genuinely negligent agent is not guided by an understanding of slapdash driving, he does not perform slapdash driving as a task. For him no amount of unison among the processes that support the slapdash driving can count as a case of attention. He does perform driving as a task, but his understanding of driving defines a large background set, including resources that, due to his slapdashness, are resources that are occupied by something else. That background set does not display unison, and so the cognitive unison theory, in the form that allows for superordinate tasks, still counts genuinely slapdash driving as inattentive in the cases where it is.

Cognitive Unison

67

To show that our getting the right verdict here does not depend on the choice of example, it may be helpful to step back so as to see more abstractly how it is that the theory gets these cases right. The balance that needs to be struck is between a background set that is too inclusive and a background set that is too exclusive. We were too inclusive if the background set included resources that the agent might have used, but only if he had taken a different approach. This would have left us with the unhappy conclusion that agents with more than one available strategy are not attentive unless they somehow employ both strategies at once. We were too exclusive if the background set included only those resources that are actually engaged by the agent’s attempt at performing the task. That would have left us with the unhappy conclusion that every task is performed attentively. Striking the necessary balance depends on seeing that the different strategies that are available to the thinker with superfluous competence are themselves performed with understanding and so qualify as tasks relative to which a background set can be defined. The background set defined by the understanding that guides the performance of these subordinate tasks gets the balance right. It can be smaller than the set of all the resources that might have been brought to bear on the superordinate task, but it can also be a set that includes more resources than those that the agent is in fact using. Whenever a task is being performed we can specify a set of resources that operates in unison, simply by finding a set that includes only resources that are actually involved in that particular performance of the task. This does not lead to the conclusion that every performance is attentive because unison across the set of resources that are circumscribed on this ad hoc basis is not the unison of a background set defined by the agent’s understanding of the task by which she is engaged, and so is not the sort of cognitive unison that constitutes attention. We saw, when we considered the requirements of understanding (in §4.6.1), that if the cognitive basis for a performance is restricted so that there is just one sort of performance that that cognitive basis can enable, then it is not a cognitive basis of the sort that counts as an understanding. The maximally specific performance, using just these cognitive resources, is therefore not a task, and so we do not face the conclusion that the set of resources that the agent actually uses in performing a task forms a background set that always displays cognition unison. It is in this way that the cognitive unison theory avoids counting every agent as attentively performing the task of performing precisely as he does.

4.10 SOME FEATURES OF THE THEORY

4.10.1 Lack of Privileged Processes Some cognitive processes seem more closely linked with attention than others. Holding a representation in working memory, for example, seems

68

Attention Is Cognitive Unison

more or less to guarantee that its contents are being attended, whereas subjecting a representation to the processing in auditory cortex by which sounds are localized in azimuth seems like the sort of thing that could perfectly well happen even when the sound is not attended. From the point of view of the cognitive unison theory, however, both processes satisfy the definition of ‘cognitive process’, and both contribute to attention in the same way. This can seem like an omission. Even if one does not want to revert to a process-first theory, in which attention is identified with storage in working memory, or with the operation of any other particular processes, one might still want one’s theory of attention to pick out those processes as making some special contribution to attention. The cognitive unison view does not do any such picking out by itself, but the view can nonetheless accommodate the fact that those processes do make a special contribution to attention by showing how that special status can be derived from contingent features of our psychology. There are two ways in which a particular processing resource might, as a contingent feature of our psychology, come to have a special role in determining which tasks are performed in cognitive unison. The resources in question might have a special role because they are resources that are heavily in demand in the performance of the sorts of tasks that creatures like us are typically engaged by. If so then those resources will be ones that occur in lots of background sets, and so the facts about which tasks engage those resources will have a special role in determining which of our tasks can be performed with unison. A resource might also come to have a special role on account of being particularly good at marshaling the support of other resources to the task that it serves. If so then it will be a resource that has a special role in bringing about unison and in maintaining it. The resources of working memory seem to have a special role in human attention for at least the first of these reasons. Lots of the tasks that are performed attentively by adult humans are tasks that place demands on working memory. The fact that a task is served by working memory is therefore likely to be an important determiner of whether or not that task is served by resources that operate in cognitive unison. Working memory might also have a special role in human attention for the second reason. It might, that is, be unusually influential in determining the activities by which other cognitive resources are engaged, so that, once a task is being served by working memory, other resources are more likely to be recruited to the service of that task, with the result that a task that is served by working memory is more likely to be performed by resources that operate in unison. Nilli Lavie and her collaborators, although operating with a somewhat different picture of attention from that which I have been advocating here (Lavie, 1995), have observed a number of empirical effects that suggest working memory does indeed play such a role. Lavie finds that the attention of subjects in her experiments is

Cognitive Unison

69

much more likely to remain fixed on a task if their working memory is available to serve that task, even when the task itself makes relatively few memory-related demands. When working memory is taken up with something irrelevant, attention becomes much easier for irrelevant stimuli to capture (Lavie and De Fockert 2005). Something similar may perhaps be true of the resources that are involved in visual focalization and saccade targeting—processes whose connection to attention has been established in a series of experiments by Tirin Moore and his collaborators (see, especially, Moore and Armstrong 2003). Given the importance of vision to the cognitive lives of creatures like us, it seems likely that detailed visual processing of the stimuli that are relevant to a task might be a good recruiter of other resources to the processing of those stimuli, in a way that the processing of the locations of auditory stimuli is not. In these ways, the cognitive unison theory can account for the fact that processes such as working memory and visual focalization are special in the degree to which the tasks that occupy them are apt to become attended. Because the theory accounts for these facts by reference to contingent features of our psychology it allows that quite different processes might have this role in the attention of creatures with quite different cognitive priorities. Such latitude is a virtue. We can imagine a creature with a very large working memory capacity, sufficient to hold all the information relevant to the task in hand, and also to hold other information. For a creature such as this it seems plausible that a task may be looked after by some part of working memory while the creature gives his attention to other things. In that case working memory seems less important in determining what the creature is attending to. We allow that such creatures are genuinely capable of attention only if, as in the cognitive unison theory, we allow the importance of working memory in human attention to be treated as a contingent feature of our psychology. Something similar can be said about the visual case. For creatures like us the fact that an item is being processed by resources that represent that item’s visual details plays a strong role in bringing it about that this item is the thing to which the creature is giving his attention. For other creatures visual stimuli play a relatively slight role in providing sensory information. For these creatures the focusing of auditory and olfactory resources is much more important for the performing of sensory tasks. It is therefore natural to suppose that, in the cognitive life of the mole, for example, facts about what is attended and facts about where visual processing is focused can come apart more readily than they can in the cognitive life of creatures like us. Their importance seems like a contingent feature of our psychology, not a feature to be built into the theory of what attention is. Although these considerations depend on thought experiments about the role played by attention in the cognitive lives of creatures with a

70

Attention Is Cognitive Unison

psychology that is different from our own, and must therefore be treated with caution, they nonetheless provide some reason for thinking that the cognitive unison view is right not to prioritize any particular processes in its account of what attention is.

4.10.2 The Privative Character of Cognitive Unison The condition that the cognitive unison theorist claims must hold over the background set if the subject is to count as attentive is a negative one: It is the absence of any irrelevant processing. An alternative approach would be to offer a more positive portrayal of attention, as requiring that some quantity or proportion of the background set be actively employed on the task in hand. It is not obvious how we should choose between a privative view and these more positive, threshold-setting alternatives. The cognitive resources of a normal waking brain tend to find something to occupy them, and so one can typically expect there to be some processing going on in any given resource. Typically, then, the only way in which there can be an absence of irrelevant processing is for there to be a presence of relevant processing. It is a consequence of this that, for actual cases, a positive threshold version of the cognitive unison view will tend to coincide extensionally with the privative version in the verdicts that it makes as to whether or not a subject is attending. The cases in which the versions come apart will be rare. We can nonetheless see that there are some lights in which the privative view seems to be the right one. H. G. Alexander states one of the reasons for liking the privative approach in a 1953 paper, which, following Ryle’s Concept of Mind (1949), addresses attention together with ‘heed concepts’ more generally: If this analysis of heeding as a negative term is right, it would follow that in some respects doing something attentively is less sophisticated than doing it inattentively—although not, of course, that doing something efficiently is less sophisticated than doing it inefficiently. And that this is so can be seen by considering children, who, when young, might be said to do everything attentively. To talk of a child playing with toy soldiers inattentively seems inappropriate; and to ask him to go on playing but at the same time to ‘keep half an eye’ on something else, is to ask for the impossible. (1953: 518)

On this point I think that Alexander is right. We do not think of children as great payers of attention, but this is not because they fail to attend. It is because their attention is easily caught and so is highly vulnerable to distraction. The state of attentiveness is one that children struggle to remain in, not one that they struggle to attain. The same thing goes for animals. We said, above (in §4.6.1) that it was important to be able to say (or to provide a good excuse for not saying) that a cat can stalk attentively. But clear cases of inattentive cat behavior are harder to come by.

Cognitive Unison

71

In general, the capacity for inattentive engagement with a task is a sophisticated achievement. We see it when adults perform tasks that they have mastered. We see it hardly at all in children and animals, and only rarely when adults attempt tasks with which they are unfamiliar. This is just what the privative view would lead us to expect, since the privative definition of unison makes the inattentive performance of a task a more complex matter than its attentive performance. The positive threshold view, by contrast, has the consequence that inattention can occur on account of the fact that not very much is going on. Threshold views are therefore unable to avail of the privative view’s simple account of why attentive engagement with tasks comes earlier and easier than does inattentive engagement with them. Further reasons for favoring the privative view can be found if we look at the unusual cases in which the privative view and the threshold view do come apart extensionally. To see one such case, suppose that I am looking out of the window with the task of counting the number of birds that land on the roof opposite. Suppose also that, just at the moment, there are no birds to be seen. The resources that would be used in noting and counting the arrival of a bird are available to be brought online should they be needed, but, just now, they are not needed and are dormant. According to the threshold view I am attending to the task of vigilance only if I am actually using some resources in doing it. According to the privative view, when there is nothing for the resources to do they need not be doing anything, but they must not be doing anything else. Since the relevant resources are not being occupied in the birdcounting task, they will tend to get occupied by other processing: If the intervals between birds are longer than a minute or two I will tend to become distracted. The privative view and the threshold view disagree as to when my attention gives out. On the privative view I continue to pay attention up to the point when I become distracted. On the threshold view I am not attending during the dormant, birdless interval before distraction sets in. The verdict given by the privative view seems to be the more natural one. A different sort of case, which again seems to support a privative rather than a threshold-based version of the cognitive unison theory, is suggested by the descriptions given of the attentive states that are attained during certain sorts of meditation. There is an emphasis in some Buddhist writings on a form of meditation that involves focusing on a thought or feeling without exercising any judgment about the content thought or felt, and with a still mind, clear from any distracters (Austin 2009: 11). This meditative state of mind seems to be simple, still, and attentive. There is little more going on than the apprehension of some mantra, or image. The privative version of the cognitive unison account has a ready characterization of such a state: it is the simplest form of unison, the necessary unison of a solo performance. The positive threshold version of the view must either reject the characterization of

72

Attention Is Cognitive Unison

the meditative state as still and simple, or it must reject the characterization of such a state as attentive. Neither move seems to be a natural one, so here again, we find some reason, although not a conclusive reason, to favor the privative version of the view.

4.10.3 To Which Objects Does the Attentive Agent Attend? The cognitive unison theory, as we have stated it, gives us an account of what it is for someone to be doing something attentively, and, since it takes attention to be an adverbial phenomenon, the theory says that attention just is the doing of something in this way. This gives us our answers to the two questions with which chapter 2 began, but it does not yet specify truth conditions for claims about which objects an agent is paying attention to. Nor is it clear how we should go about specifying them. Standard philosophical tactics would tell us to start with some clear cases, and to proceed by considering the features that those cases have in common, and by coming up with thought experiments that enable us to determine which of these common features are essential. But in the present case those tactics prove to be remarkably difficult to implement, on account of a difficulty in finding any clear cases where there is a particular object that is the thing of which it is correct to say that the agent is attending to that. Even in paradigmatic cases of attention it seems to be an open question to which objects the attentive agent should be said to be attending. A skilled dancer, for example, is clearly paying attention when dancing, but what objects is he paying attention to? Perhaps it is true that the dancer is attending to the music, but his is not a clear case of attending to the music. There is a difference between attentively dancing and attending to the music while dancing: If he is attending to the music in the wrong way then his attention to the music might actually prevent him from dancing attentively. Perhaps we should say that the dancer is attending proprioceptively to his body, but, again, that does not seem quite right. Attention to one’s body is the mark of the amateur dancer, not the skilled practitioner. The only thing that it is clearly right to say is that the dancer is attending to his dancing, but that is to specify the task that is performed attentively, not the object to which attention is paid. Most natural cases of attention seem to show the same structure: Even when it is clear that attention is being paid, there is no one object that suggests itself as the one to which attention is paid. In reading now, it may be the book, the page, the sentence, or the argument that you are said to be attending to, but it seems wrong to say that you are attending to all of these, and it seems arbitrary to pick out any one of them. Particular objects of attention are equally hard to identify for the instances of attention that are studied in the lab. In many experiments there is no object to which attention is paid. The experimenter flashes

Cognitive Unison

73

up an image on a screen as a cue to the direction of attention, but that image is then removed and the thing to which the subject attends is not the image, but only a part of the visual field where that image formerly was. All of this suggests that we have been right to treat attention as a mode of engagement with a task, not as a mode of engagement with an object. In order to get an answer to the question of which objects the attentive agent attends to, our theory should start with attention to tasks, and should proceed from there to an account of attention to the objects that participate in those tasks. We can do this by defining ‘object involving’ as follows: Say that a task, τ, involves object o just in case the cognitive resources that serve τ are resources that operate on a representation of o. Most tasks will be object involving in this sense. We can now use our theory of what’s required for attention to a task to derive a theory of what’s required for attention to an object. There are more or less generous ways in which the derivation might go. A generous theory would say that, when o is an object involved in task τ, and when an agent is performing τ attentively, then and only then the agent is attending to o. A less generous theory would say that the agent is attending to the object, o, if and only if the task that is being performed attentively is one in which object o is, as we might say, accusatively involved, that is, involved as an object on which the task is being performed. The less generous theory would have the consequence that, when attentively hammering in a nail, the agent counts as attending to the nail but not to the hammer. The more generous theory would count both hammer and nail as attended. As far as I can see it is a matter of conversational pragmatics which of these standards is operative. There is no absolutely right answer about whether the hammer is among the things to which attention is paid when hammering in a nail.

4.10.4 Spectator Attention Because the facts about which objects one is paying attention to are grounded in facts about which tasks one is performing, one cannot simply comply with an instruction to pay attention to this chair without taking on some task in which the chair has a role. The task need not involve active manipulation of the chair, of course. It may be a simple perceptual task, such as getting visually acquainted with the chair, or detecting any changes that the chair might undergo. In this way the cognitive unison view collapses a distinction that Alan White, from whose suggestions the view started, was careful to draw. This is the distinction between what White called ‘agent attention’ and what he called ‘spectator attention’. White warned that ‘an assimilation of these two types of attending to what one is doing has led to mistaken views about attention’ (1964: 9), but in the White-style theory that I have been developing here, the two are indeed assimilated.

74

Attention Is Cognitive Unison

White is quite right that there is a difference between these two types of attention, and he provides an example that clearly illustrates it: A novice might watch his own driving as he would his instructor’s, to learn the right operations and correct those which are mistaken. But in attending to his own driving the novice is not necessarily, or even probably, driving attentively, and in attending to his instructor’s driving he could not be driving attentively. (1964: 9)

But, although White rightly identifies a difference here, we do not need a distinction between two kinds of attention in order to account for it. The cognitive unison theory accounts for the difference as a difference in the task attended. The novice gives his attention to the task of driving, but also to the task of learning how to drive, where the latter task is potentially served by resources involved in noticing and remembering the appropriate techniques and routines for particular situations, and by resources involved in monitoring his own driving to discover ways in which it falls short of the approved manner. When the practiced driver drives attentively, however, he just performs the task of driving. The resources in the background set of this task are only those involved in watching the road, keeping track of where he is, following the traffic laws, and so on. We do not need the distinction between agent attention and spectator attention to be a difference between two fundamentally different forms of attention. Both forms of attention can be seen to fit the single definition given by the cognitive unison theory once we realize that the spectator and the agent are performing different tasks.

4.11 DIVIDED ATTENTION Since attention is essentially selective it is not a contingent matter that dividing one’s attention between two tasks is often difficult and sometimes impossible. Our theory of attention should allow for the possibility of dividing one’s attention, but it should also account for the difficulty of doing so, and it should say something about the cases where the division of attention proves to be impossible. Not all cases in which an agent is attending to more than one task are cases in which dividing one’s attention presents its characteristic difficulties. In the mental arithmetic example that we considered in §4.9.1, the agent may be said to be paying attention to the calculation of the difference between the two squares, and she may, at the very same time, be said to be paying attention to visualizing the area of overlap between them. The task of finding the difference and the task of visualizing the overlap are distinct (since an agent may be doing either one of them without doing the other). This is therefore a case in which two distinct tasks are being given attention at one time. But it is not a case where attention is divided, since one of these tasks is being executed as a way to

Cognitive Unison

75

accomplish the other. It is only when the agent is dividing her attention between a pair of competing tasks that a distinctive difficulty arises. The nature of this difficulty differs in different cases. We can start out with the easy ones.

4.11.1 Easy Divided Attention One of the easy cases of divided attention is the case that we saw at the end of chapter 3, when we considered the surprising ease with which subjects are able to divide their attention between sight-reading piano music and performing an auditory shadowing task. What makes such a case surprising is not just the fact that subjects are able to perform both of these tasks attentively at one time. What is really surprising is that the subjects’ capacity to perform the one task seems not to be greatly affected by the attention required for the other: The piano playing of subjects who are sight-reading grade II music gets no worse when the passage of text that they are simultaneously verbally shadowing is changed from an easy piece of humorous narrative prose to a difficult extract from a textbook of early Norse history (Allport, Antonis, and Reynolds 1972: experiment II). When the opposite comparison is made some effect of difficulty is observed: The ability to shadow the text does get slightly impaired when the music that is being sight-read is made harder. But this effect is slight, and in a second session, where the subjects are more practiced at dividing their attention, the effect goes away. The cognitive unison theory has a ready account of the ease with which attention is divided between such pairs of tasks. In order for the agent to be attending to the task of auditory shadowing the resources in the background set for auditory shadowing must not be doing anything irrelevant to auditory shadowing. In order for the agent to be attending to the task of sight-reading piano music the resources in the background set for sight-reading piano music must not be doing anything irrelevant to sight-reading. The subject can simultaneously meet both sets of requirements because, since the background sets of these tasks do not intersect, there is no conflict between the requirements of unison for each of them. In this way the cognitive unison theory is able to account for the fact that the subject can, for certain pairs of tasks, pay full attention to both tasks at once. It is rare, however, to find a pair of tasks for which divided attention comes as easily as it does for sight-reading and auditory shadowing. The fact that these easy cases are rare is, for the cognitive unison theory, no surprise. The theory says that such cases occur only when the background sets do not intersect, and background sets do intersect almost always. As was mentioned in §4.10.1, some resources are understood to be useful for all sorts of tasks and therefore crop up in lots of different background sets. The result of the popularity of these resources is not only that the background sets of arbitrarily chosen pairs of tasks tend to

76

Attention Is Cognitive Unison

intersect, but also that the intersections tend to converge. The cognitive unison theory therefore predicts that, given the contingent features of our psychology that lead to processes such as working memory having a privileged role in attention, we should rarely expect to find pairs of tasks both of which can easily be given full attention simultaneously. But the theory also allows that there will be other routes by which attention can be successfully divided.

4.11.2 Easy but Unstable Divided Attention When sitting on a moving train it is relatively easy to divide one’s attention between watching the passing landscape and listening to music through one’s headphones. This state of divided attention typically comes so easily that one falls naturally into it. One settles down to listen to some music and finds oneself watching the landscape too, or decides to take in the landscape but cannot help giving some attention to the music. The state of divided attention is so much the equilibrium state here that, unless something remarkable starts taking place outside, it takes effort to pay attention undividedly to the landscape and to withdraw attention fully from the music. The division of attention in such cases is not, however, wholly stable. If something unexpected starts happening outside then it may become impossible to carry on attending to music. And, similarly, if something out of the ordinary starts coming through one’s headphones, then attention may be withdrawn from the landscape whether one wants to withdraw it or not. The cognitive unison theory’s account of such cases is only a little more complicated than its account of dividing attention between sightreading and auditory shadowing. For those two tasks the inputs were of different sorts (one visual and one auditory), the outputs were of different sorts (one manual and one verbal), and the input/output associations between the two were quite different and were well established. It was therefore plausible to claim that the background sets for these tasks are nonintersecting. The tasks of music listening and landscape watching do not have this step-by-step distinctness. The inputs are distinct (one visual and one auditory), but both tasks aim at roughly the same output, which is something like a set of occurrent thoughts in which the subject reflects on whatever is going on. Given this overlap between the sorts of things that both tasks aim at, it seems that for music listening and landscape watching, unlike for sight-reading and shadowing, we have a pair of tasks with background sets that do intersect, although each of the two sets will also have many members that fall outside the intersection. Despite the fact that their background sets intersect, both tasks may still be performed by resources that operate in unison, and so the unison theory allows that both may be attended. The possibility of simultaneous unison here is a consequence of unison’s privative character. Unison requires only that the members of the background set must not be taken

Cognitive Unison

77

up with processing that does not serve the task at hand. Members of the background set can satisfy this requirement if they are not taken up with anything at all. Tasks with intersecting background sets can therefore be simultaneously given attention if the resources in the intersection are quiescent. If the resources in the intersection are recruited to either task (as they might be if a remarkable event occurs in one or other of the stimulus streams), then, in all but a very few cases, the background set of the other task will no longer be operating in unison. When that happens the other task will cease to be attentively performed. It is for this reason that divided attention between such tasks is possible but not entirely stable. Strictly speaking, it is not only when the resources in the intersection of their background sets are quiescent that two tasks with intersecting background sets can be simultaneously attended. Both tasks can be fully attended even when the resources in the intersection of the background sets are active with processing, but in order for this to happen it must be the case that the processing that occupies the resources in the intersection is of a sort that serves the performance of both tasks. Examples along these lines are rare, but they are not impossible. Consider the trombonist in a marching band. The trombonist may be giving full and simultaneous attention to the task of marching in time and to the task of playing in time (and he may be performing both the marching and the playing as tasks in their own right, neither one of which is performed as part of a strategy for performing the other). The background sets for these two tasks clearly intersect, with the resources involved in keeping time being the most prominent members of the intersection. Since those time-keeping resources can serve both tasks by doing the same thing, the sets for both tasks can simultaneous satisfy the definition of unison. But it is only in special circumstances such as these that an activity in the intersection of two background sets is compatible with full attention to both tasks.

4.11.3 Effortful Divided Attention In the least unusual of the examples that we have considered so far—the example of looking out of the train window while listening to music—it is crucial to the possibility of divided attention that neither of the tasks is very demanding, and that both are done in a considerably less than full-on fashion. The train passenger understands what it is to look at a landscape, and that understanding is what explains the way in which he goes about it, but although he therefore has looking as a task, he is not especially committed to making sure that he remembers what he sees or to noticing all the details. This noncommittal character of his performance allows some of the resources that he could bring to bear to be left quiescent, and it is this quiescence that allows other tasks that could

78

Attention Is Cognitive Unison

draw on those resources also to be performed attentively. The noncommittal character of the performance, together with the privative character of unison, is crucial to the possibility of divided attention in this sort of case, but not all cases where attention is divided between two tasks with intersecting background sets are cases where the tasks are performed in this noncommittal way. The sorts of examples that we turn to now are examples where the background sets of the tasks between which attention is divided do intersect, and where the laid-back attitude characteristic of the train window case is absent. Maintaining divided attention in these cases requires considerably more work on the part of the agent than was required in the previous cases. Such cases are, nonetheless, fairly commonplace. Some people can pay attention to having a conversation while also paying attention to eating their dinner. Some can pay attention to the thing they are typing while also paying attention to maintaining a healthy posture at their desk. Some even claim to be able to divide their attention between listening to the radio and reading the newspaper. Dividing attention between such tasks requires effort on the part of the subject. It may also require practice. In some cases the ability to divide attention is a skill that can be acquired with training. In a classic study by Spelke, Hirst, and Neisser (1976), for example, a regime of practice enabled subjects to develop an ability to divide their attention between reading and taking dictation. Unlike in the train window example, the state of divided attention in these examples is not an equilibrium state that one naturally falls into. It is only when one makes the effort that one can divide one’s attention between typing and maintaining a healthy posture. The natural tendency is for one’s posture to get neglected as one types. Similarly, one can attend to one’s dinner and to the conversation, but effort is required. When the conversation is engrossing, one’s dinner will tend to get eaten more or less unheeded unless one makes a point noticing it. Because the tasks in these examples lack the noncommittal character of the tasks in the train window case the resources that the agent can bring to bear in performing them will tend to be ones that he does bring to bear. Among those resources are the ones that fall in the intersection of the two background sets. Those resources will therefore tend not to remain quiet, and so the quiescent-intersection route to simultaneous cognitive unison will not be available. When the intersection resources are serving one task, the other tasks into whose background set those resources fall will, ipso facto, count as being performed without cognitive unison. The cognitive unison theory is therefore committed to denying that such pairs of tasks are simultaneously given full attention. It must instead say that for these instances of divided attention, although both tasks may be performed at the same time, they are not both attended at the same time. Attention is paid now to one, now to the other.

Cognitive Unison

79

This consequence of the cognitive unison theory is quite a strong one: The theory rules that some pairs of tasks are such that simultaneous attention to the committal, full-on performance of both is impossible, however competent one becomes at performing them. The defender of the cognitive unison theory should not balk at this result. For one thing, the result seems to agree with the phenomenology of dividing attention between such tasks. The experience of typing while watching one’s posture is an experience of intermittently disengaging one’s attention from typing to make a check on how one’s posture is. Nor should the defender of the cognitive unison theory think that this consequence of the theory places it at odds with the verdicts of common sense. It is true that we find it natural to say that an agent may be dividing her attention between eating and talking, or between typing and maintaining an upright posture, but the fact that we find these natural things to say is not incompatible with the result that it is impossible to attend to both of the tasks in these pairs at one time. It is perfectly permissible to say that someone has attended to something throughout some interval even if their attention during that interval was, in fact, only intermittent, just as it is perfectly permissible to say that one played tennis throughout one’s teenage years, even if the episodes of playing were interspersed with episodes of nonplaying. In cases where attention alternates between a pair of tasks, it can still be perfectly correct to say that the subject attended to one task throughout some interval and to say that he attended to another task throughout the same interval. These claims can be understood in such a way that they do not entail that there was any one time during that interval at which both tasks were simultaneously attended. They can therefore be understood in such a way as to be compatible with the cognitive unison theory’s consequence that strictly simultaneous attention to these tasks is not possible in contexts where either task is performed with full-on commitment.

4.11.4 Mutually Exclusive Tasks The cases of divided attention that we have considered so far have been cases where the background sets of the tasks that are performed simultaneously have crucial members that fall outside of the intersection of those sets. Such task combinations present the agent with some of the difficulties that are characteristic of divided attention, but they are cases where divided attention can be achieved, and can sometimes be achieved relatively easily, if not stably, by at least some people in at least some circumstances. The cases where divided attention creates the most severe difficulties are cases where the tasks between which attention is divided are not merely tasks with intersecting background sets, but are two versions of the same thing. When the same bit of understanding is at work in two tasks the intersection of the background sets will be large, relative to the sets

80

Attention Is Cognitive Unison

themselves, and may be exhaustive of them. It may be that neither task can be performed at all except by drawing on resources in the intersection. This will be the case if, for example, both tasks are visual search tasks, with one task involving searching through the items presented on one screen for some particular kind of visual target, and the other task involving searching through the items presented on another screen for some other kind of visual target, or if both tasks are mental calculations, where one task involves adding pairs of numbers, and the other task involves subtracting them. As with the cases above, the subject who attends to both of the tasks in these pairs must, according to the cognitive unison theory, be switching between performing one attentively and performing the other attentively. There cannot be a moment when he is giving full attention to both. But in these cases something more is true. Since the tasks cannot be performed at all without drawing on the resources in the intersection of their background sets, we cannot say of these tasks, as we said of the previous ones, that the agent performs both simultaneously and counts as attending to both by switching attention between them. When we have a pair of tasks whose background sets overlap at points that include essential resources for task performance—as with a pair of arithmetic calculations, or a pair of visual search tasks, or a pair of finely coordinated actions—we have a pair of tasks that cannot be performed simultaneously at all, let alone performed with simultaneous attention. The agent who intermittently attends to both members of such a pair has to start and stop each performance when attending to the other. This consequence of the cognitive unison theory is again a strong one, but again it corresponds to something that commonsense psychology would endorse. Some pairs of tasks really do turn out to be incompatible, and it is when the tasks are demanding in similar ways that such incompatibilities arise. The cognitive unison theory’s ruling that it is impossible to divide one’s attention between tasks with substantially intersecting background sets corresponds to this intuitive point. But in ruling that divided attention in these cases is literally impossible, the cognitive unison theory may appear to go too far. It is not entirely comfortable for the advocate of the cognitive unison theory to find himself saying that dividing one’s attention between, say, a pair of visual search tasks always requires the agent to adopt a strategy of switching between them. Everyday experience may teach us that all attempts to pay attention to both tasks at a single time may be bound to fail sooner or later, but it does at least make sense to try to pay attention to both tasks, and maybe it is possible to succeed in attending to both, even if only briefly, and even if one’s performance of the tasks is far from perfect. The cognitive unison theory’s ruling may therefore seem uncomfortably strong. It says not only that it will be impossible to successfully execute pairs of tasks with substantially intersecting

Cognitive Unison

81

background sets, but also that it will be impossible ever to attend to both tasks at a single time. The strength of this consequence can be mitigated. The cognitive unison theory can allow that it might be correct to say that someone was, at a single moment, dividing her attention between a pair of visual search tasks, even when neither of the two tasks satisfies the definition of cognitive unison. It allows for this by allowing that an agent can work out how to set herself a larger task composed of the smaller ones. The background set for this larger task will be, approximately, the union of the background sets for the smaller tasks, but, since the larger task can be served by activities that contribute to either of the smaller ones, it can be performed with resources that operate in unison even while the resources in the background sets of the smaller tasks do not. For an example along these lines suppose that an agent is attempting to perform a demanding visual search of an array of stimuli presented on one screen while also attempting to perform a demanding visual search of another array of stimuli presented on another screen. We may suppose, also, that the agent in these conditions cannot perform both tasks at the same time. If so then she cannot be performing both tasks attentively at the same time. But if the tasks are not metaphysically incompatible (the two screens are not, for example, back to back), then the agent might decide to, as it were, stand back from each individual monitoring task and group the two tasks into a single, larger one. To this larger task she can then give her full attention. When the subject decides to put two smaller tasks together into a larger one, the resulting larger task is not a superordinate task of the sort that we discussed in §4.9. In the case of a superordinate task the lower level tasks are strategies for performing the higher level one. But performing just one of the visual search tasks is not a strategy for performing both. In the current case it is the performance of the larger, composite task that is a strategy for performing the individual low-level tasks. Composite tasks, like all activities, count as tasks only if they are guided by the agent’s understanding of what it is to perform them. Grouping tasks into a composite is therefore an intellectual move that requires some intelligence and insight. It is not the case that, for any two tasks that an agent may attempt under the guidance of understanding, there is a third task, composed of them, that the agent is equally capable of attempting under the guidance of understanding. An agent may understand τ1 and understand τ2, and have no understanding at all of what it would be to do both at once. If that is the case then the composite activity is not something that can be performed by that agent as a task. Failure to get a composite understanding might arise in various ways. The agent may simply fail to realize that the composite strategy for dual

82

Attention Is Cognitive Unison

task performance is available. The agent might also lack an understanding of the composite task because he realizes, or falsely believes, that there is simply no such thing as simultaneously performing both tasks at the same time. It may also be that the agent does have an understanding of the composite task but is not able to use such an understanding to guide a performance, and so is not able to adopt the composite activity as a task. This may happen if the composite task greatly outstrips the agent’s capacities, as it may if, for example, the agent is a trainee acrobat and τ1 is the task of juggling four balls while τ2 is the task of doing backward somersaults. The agent may have an understanding of juggling and an understanding of backward somersaults, sufficient for being able to perform each of these as a task. He may also know that the two tasks can be performed simultaneously, perhaps because he has seen it done. But he may nonetheless be utterly at a loss as to how to go about using his understanding to guide a simultaneous performance. In that case, the most he is able to do is to give full attention to his attempts to juggle, breaking off intermittently, and despite himself, when he attempts to add the somersaults. Using one’s understanding of the composite task to guide the task’s performance is an intellectual achievement that is not, in general, a trivial consequence of having the understandings required by the component tasks. Despite these complications, it will in many cases be possible to get to an understanding of the composite task merely by exercising some imagination and employing one’s understanding of that task’s components. The case of performing two visual monitoring tasks may be one such case. The agent who attempts to monitor one screen for visual targets and simultaneously attempts to monitor another screen for visual targets might set herself the larger task of searching both screens and might count as giving her full attention to that larger task even when her search of neither screen counts as attentive. It can be true, therefore, that during a particular interval an agent was attentively monitoring two visual streams at one time, even though the background sets for the two individual monitoring tasks massively intersect. This can be true without it being the case that the interval in question is one during which she was sometimes monitoring one and sometimes monitoring the other. It can be true because she adopted the composite task of monitoring both. As long as we are careful to avoid an equivocation of scope we can see that this is quite compatible with the restrictions on divided attention that we have seen to issue from the cognitive unison theory. It would be an equivocation of scope to think that ‘the agent is paying attention to [the monitoring of both screens]’ implies ‘the agent is paying attention to [the monitoring of screen 1] and the agent is paying attention to [the monitoring of screen 2]’. The cognitive unison theory rules out the second of these but allows for the first. In this way, and in the other ways that we have seen, the theory allows for divided attention while also accounting, in various ways, for the difficulties that it creates.

Cognitive Unison

83

4.12 DEGREES OF ATTENTION AND MERELY PARTIAL ATTENTION In the discussion above I have sometimes allowed myself to speak of full attention, rather than just speaking of attention simpliciter. This might lead one to think that there is such a thing as partial attention, and it might lead one to suspect that our theory so far gives only an inadequate account of it. Cognitive unison requires that no resources in the background set be taken up with activity that does not serve the attended task. As soon as any of those resources are doing something that does not serve that task, the thinker’s performance no longer satisfies the definition of unison, and so he is no longer counted by the cognitive unison theory as performing attentively. This may seem to make attention an all-or-nothing affair, and yet it is clear that attention comes in degrees. How, then, can the cognitive unison theory allow for this gradability of attentiveness? There are various degrees of freedom allowed by the definition of cognitive unison. These degrees of freedom, as we shall see, allow us to define dimensions of intensity along which instances of attention can vary. In this way, the theory allows us to say something about attention’s degrees. There are, nonetheless, two senses in which the cognitive unison theory does not provide a theory of partial attention: First, the theory does not assign instances of attention to a well-ordered scale of intensity and does not give us anything like a way of calculating the particular degree to which attention is paid in any given case. Second, the theory does not allow that any cognitive state that falls short of unison counts as a genuine case of attention. It is helpful to separate these points. I shall first set out the ways in which the cognitive unison theory allows for degrees of attention. I then go on to explain the sense in which true attention cannot, strictly speaking, be partial.

4.12.1 Degrees of Attention Different instances of cognitive unison can differ as to the size of the background set over which unison is displayed; they can differ as to the proportion of the background set that is actually active; they can differ as to the rate at which the active resources are performing; and they can differ as to the stability of the state of unison. Each of these dimensions of variation corresponds to a way in which a pair of performances may differ in their degree of attentiveness. The attention given to a task with a relatively large background set (such as the task of standing perfectly still) can be said to be a greater degree of attention than the attention given to a task with a relatively small background set (such as the task of blinking at regular intervals). The attention given to a task whose performance demands activity from a great many of the resources in its

84

Attention Is Cognitive Unison

background set (such as the watching of a gripping scene) can be said to be a greater degree of attention than the attention given to the task when some of the resources in the background set are quiet (as in our earlier example where one gives attention to the scenery through which one’s train is passing). Even when all the resources in a background set are working in the service of a task, they may not be working very hard. In that case the attention paid may again be said to be of a lesser degree than the attention that would be paid if all the resources were working flat out (an example here might be the difference between the degree of attention required for an easy mathematical task, such as counting down in twos, and the degree of attention required for a harder task, such as counting down in sevens). Finally, it may be that, as we see in the case of children, a subject pays close attention to one thing but is nonetheless highly susceptible to distraction from it. In such a case, we may say that the attention paid is of a lesser degree than the attention that is paid by the subject who resists distraction. There is no guarantee that these different dimensions of intensity along which instances of attention may vary will be dimensions that vary together. One instance of attention may, for example, display more stability than another, but less activity in the background set. Because of this we can say only that the attention given to a task with a large background set will tend, other dimensions of variation being equal, to be of a greater degree than the attention given to a task with a small background set. We cannot say that large-set tasks always demand, or receive, more attention than do small-set tasks. The other factors that we have identified may conspire to make the attention given to a smallset task greater than the attention given to a large-set task. The fact that these different dimensions of variation are somewhat independent of one another does not, by itself, prevent us from defining a function that would assign every instance of attention to a particular point on a well-ordered scale of degree of attentiveness. We could define such a function simply by settling on some weighting of the various dimensions of variation against one another. The cognitive unison theory does not prevent us from making a precise assignment of degrees to the various instances of attention. It does, however, refrain from doing so. We could assign instances of attention to a scale by assigning weights to the different dimensions of variation, but there is nothing in the nature of attention that would make one such assignment of weights right and another one wrong. It would be a mistake to give a theory that, for any two instances of attention, always returns a determinate answer to the question of which exhibits the higher degree of attention. Some pairs are tied, but also, and more importantly, some such comparisons make no sense. If we were asked whether the attention that a particular soccer player gives to taking a penalty kick is greater or less than the attention that a particular heart surgeon gives to a bypass operation, we would, I think, want to reject the question. One who asked such a question in

Cognitive Unison

85

the expectation of a serious answer would be making a mistake, just as he would be making a mistake if he asked whether the surgeon is more or less skilled than the soccer player. Some comparisons of attention or skillfulness make sense, but others do not. Trainee soccer players may be less skilled than are master surgeons. Trainee surgeons may be less skilled than are master soccer players. But the theorist of skills should not be looking for a theory that returns answers whenever any two performances are compared. The theorist of attention, similarly, should not be looking for a theory that assigns all cases of attention to a place on a single scale of intensity. The cognitive unison theory does allow for different instances of attention to differ in degree along various dimensions of variation. It thereby allows us to make sense of some pairwise comparisons without returning a verdict on all of them.

4.12.2 Partial Attention In giving this account of attention’s degrees the cognitive unison theory allows that a subject may be paying attention to only a low degree (and, as we saw in §4.11, it allows that in certain cases attention can be divided). But there is still a sense in which the theory does not allow for cases of attention to be partial. Instances of attention that fall at the bottom end of the dimensions of variation identified above are not instances that fall short of satisfying the definition of unison. A small background set, occupied with nothing but processing that is relevant to a simple task, is just as much a case of unison as is a large background set devoted to nothing but a complex task. A background set in which few of the resources are active, but in which none are active with non-task-serving activity, is just as much a case of unison as is a background set in which all the resources are working flat out in the service of the task. An unstable instance of unison is nonetheless a case of unison. These cases of slight attention still satisfy the definition of cognitive unison entirely. What if this were not the case? What if the background set were to fail to display full unison but were to fail in some small way, because, for example, some of the less important resources in that set were working (but not too hard) in the service of something other than the task? One might be tempted to say that this, too, is attention, although it does not entirely satisfy the definition of cognitive unison. If so, then one will also think that cases such as this ought to figure in our account of attention’s degrees, with the attention constituted by an almost unified background set being of a lower degree than that shown by a perfectly unified background set. The cognitive unison theory does not allow this. It allows for variation among instances of unison, along the dimensions of variation that were outlined above, but it says that cases where something less than unison is instantiated are not cases of attention at all. This is not a flaw in the theory. It is merely a consequence of the semantics of the adjective ‘partial’, which can figure as a privative

86

Attention Is Cognitive Unison

adjective, and which does so in this context. A candidate who is partially qualified for a job does not fully satisfy the requirements for being qualified. A statement of the criteria for being qualified need not quite apply to him. A house that is partially built does not fully satisfy the requirements for being built (nor, perhaps, for being a house). A specification of what it is for the house to be built would not cover such a case. Of course our theories should tell us something about the partial cases: A statement of the criteria for qualification does have to tell us what it is from which the partially qualified candidate falls short. A statement of what’s required for being built does have to tell us what the partially built house is on its way to being. It would be a problem for the cognitive unison view if states of partial attention were not states of partial cognitive unison. But it is no problem if states of partial attention fall somewhat short of satisfying the requirements of attention that the cognitive unison theory states. The cognitive unison theory’s insistence that, strictly speaking, the instances of partial attention are not instances of attention can have a paradoxical look about it, but we find, on closer inspection, that it does not go against our commonsense verdicts. It is true that there are some cases where common sense finds it quite acceptable to say that an agent is attending to something, even when her attention is merely partial. Common sense might allow one to say, for example, that marking a pile of essays took longer than usual because one was also paying attention to the radio. One might be allowed to say this even if listening to the radio never got any more than partial attention, having never been a task that one performed with unison. The cognitive unison theory needs to say that since there is something less than unison in the set of resources that one could bring to bear in listening to the radio, one did not, speaking strictly, pay attention to the radio at all. On reflection there is nothing ad hoc or counterintuitive about its saying this. Loose talk of this sort is a perfectly normal feature of the way in which privative adjectives behave. Forged Michelangelos are not Michelangelos, but it is perfect natural, when surveying the forger’s studio, to say things like ‘The Michelangelos are better than the Mondrians’. Stone tigers are not tigers, but it is perfectly natural to tell the stone mason that his tigers are very lifelike. We can see that this sort of thing is just loose talk because, when we change the conversational context to one in which loose talk is not permitted, the tendency to drop the privative adjectives disappears. If we were to ask the man in the forger’s studio whether there were really any Michelangelos there, he would say that there were not. When we apply this test to the case of partial attention, we find that it is indeed just loose talk that allows us to use the word ‘attention’ when speaking of instances of partial attention where the unison is incomplete. If we were to have asked the marker of essays whether she really paid attention to the radio, it would have been perfectly natural for her to say that she did not. And if we had particularly asked her to pay attention to the radio,

Cognitive Unison

87

and found that she had paid partial attention to it while marking essays, we would be entitled to conclude that our request had not be satisfied.

4.13 SUMMARY This chapter has given an adverbialist theory of attention. According to this theory there is no such thing as the attention process. Nor is there any disjunctive set, or family, or genus, of attention processes. Any process can be attention constituting, as long as it is a cognitive process, and as long as it takes place in the appropriate way, that is, as a part of a unison-displaying background set. The theory that we have given is an elaboration of White’s suggestion that the manner characteristic of attention to X ‘consists not only in the range of activities centred on X, but in absence of activities concerned with things other than X’ (1964: 7). We have seen that this theory delivers the right verdicts about when various agents are and are not attending. We have also seen that it allows for attention to come in different degrees, and that it accounts for several of the distinctive problems posed by the different varieties of divided attention. In chapter 5 we shall see that the theory is immune to a perennial form of objection to metaphysical claims about the mind.

5 The Causal Life of Attention

5.1 MENTAL CAUSATION The letters that Princess Elizabeth of Bohemia wrote to Rene Descartes in the summer of 1643 inaugurated an ongoing tradition of objecting to philosophical proposals about the metaphysical status of mental entities on the grounds that those proposals are unable to account for the causal interactions of the entities that they purport to explain. In the second half of the twentieth century philosophers realized that such objections can be raised against many more theories than just those of the Cartesian dualist. As a result of that realization the question of how to avoid causation-based objections came to be one of the main topics of discussion in the philosophy of mind (see, for example, Yablo 1992; Heil and Mele 1993; Kim 1998; Macdonald 2006). The causation-based objections that twentieth-century philosophers of mind quite rightly obsessed about can be understood as three-premise arguments in which the first premise states the contested metaphysical proposal and the second two premises reduce that proposal to absurdity. The form of the premises is this: 1. The disputed analysis identifies a mental phenomenon, m, with ξ. 2. The mental phenomenon being analyzed, m, engages in certain causal interactions. 3. ξ, the metaphysical analysans, cannot engage in those causal interactions. There is no disputing the logic of such objections. From premises with the form of 2 and 3, it follows immediately that the analysis mentioned in premise 1 cannot be correct: If m does things that ξ cannot do, then it cannot be that m and ξ are one and the same. If the cognitive unison theory of attention were vulnerable to an argument with this form, then that would be a fatal objection to the theory. More than one feature of the cognitive unison theory of attention might lead one to think that the theory is vulnerable to such an objection. This chapter examines three such apparent vulnerabilities, with a view to showing how the mental causation objections can be avoided.

88

The Causal Life of Attention

89

The goal of the chapter is to show that attention, as the cognitive unison theory understands it, can play the causal role that it needs to play, once that causal role has been understood correctly. Showing that cognitive unison can play the causal role that attention needs to play serves as a defense of the cognitive unison theory against mental causation objections. It also dispels the whiff of eliminativism that adverbial theories can sometimes seem to have. A scientifically respectable theory of the mind should make mention of phenomena only when they play a genuine explanatory role in mind-involving states of affairs. Those phenomena with no such role to play can be eliminated from our theory without loss of explanatory power, and so they should be eliminated, in the interests of parsimony. The adverbialist’s claim that attention does not correspond to any of the brain processes that explain behavior can sound as if it were an invitation to make just such an eliminativist move. It is a corollary of the results in this chapter that this eliminativist line of thought would be too quick. Although the cognitive unison theory’s adverbialism entails that attention does not correspond to any of the processes that explain our behavior, this does not commit it to casting attention in a null explanatory role, and so it does not lead to attention’s elimination from the explanatory theory of the mind.

5.2 HOW TO RESPOND TO MENTAL CAUSATION OBJECTIONS Setting out mental causation objections in the form of a three-premise reductio, as we did above, makes clear that there are, in principle, two ways in which the defender of the analysis stated in premise 1 can attempt to resist the argument’s conclusion. He can deny premise 3, by showing that the analysans is capable of the relevant causal work after all, or he can deny premise 2, by showing that the causal limitations faced by the analysans are shared by the analysandum. In most contexts the second of these tactics is horribly unappealing and so goes more or less unmentioned. This unpopularity is well deserved. Denying premise 2 typically lands us with counterintuitive consequences that ramify throughout the philosophy of mind and beyond. They ramify into epistemology, since if we were to admit that the mental phenomenon we are trying to analyze does not cause very much, then we would find ourselves having to say that rather little evidence is available for the instantiation of that phenomenon. They ramify into our etiological account of the phenomenon in question, since if a phenomenon does not cause very much, then natural selection cannot have been under very much pressure to favor creatures capable of instantiating that phenomenon. Worse still, they ramify outside of the philosophy of mind, since so much of our conception of ourselves as agents,

90

Attention Is Cognitive Unison

and of our world as a world of artifacts, depends on taking things in that world to be products of our thinking. If the metaphysical theory at issue is a theory that purports to tell us about the nature of mental states in general, or about the nature of especially cardinal mental states, such as beliefs, then when it comes to answering mental causation objections the denial of premise 2 is ruled out. Our answer must proceed via a denial of premise 3. When, however, we are considering a phenomenon like attention, it should not be taken for granted that premise 2 is nonnegotiable. There is room for dispute about which causal interactions attention enters into. Whereas assertions, for example, depend for their identity on beliefs being among their causes, it is not clear whether there are any event types that depend for their identity on standing in an etiologically demanding relation to attention. It is clear that attention must enter into some casual interactions, since, if attention caused nothing at all then it could not be the cause of our beliefs about it, and we would thereby lose our reason to think that there is anything that those beliefs rightly represent. But, before we are in a position to evaluate the prospects of using a mental-causation-based objection against the cognitive unison theory of attention, we need an accurate characterization of what attention’s causal powers are. The theory’s ability to resist mental causation objections depends on whether our characterization of attention’s causal interactions specifies a role that cognitive unison is able to play.

5.3 THE CAUSAL ROLE OF ATTENTION Although event types that are essentially caused by attention are hard to find, cases where the paying of attention matters to the production of a particular outcome are commonplace. Suppose that you find yourself needing to walk the tightrope. Suppose also that you care about the outcome and that you are not a well-practiced tightrope walker. In such a case, you will be concerned to ensure that the situation is one in which you are free from distractions and are therefore able to pay attention. Our concern with freeing ourselves from distractions in such cases is intelligible because we understand attention to play a causal role in the production of behaviors, especially those behaviors that require skill. We could make similar points about intellectual performances, or about perceptual ones. We are careful to ensure freedom from distractions when taking exams, and when keeping a lookout for hard-to-spot stimuli. This concern is intelligible because, in addition to its role in the production of skilled behavior, attention has a role to play in the occurrence of thoughts, especially clever ones, and in the occurrence of perceptual states, especially acutely accurate ones.

The Causal Life of Attention

91

Our characterization of the causal role that attention plays in the production of skilled, clever, or acute behaviors and thoughts should not proceed by specifying a set containing all and only the behaviors and thoughts that attention brings about. An enormous number of tasks may be performed attentively. The things that attention brings about are, as a result, quite various: Attended things may be seen more clearly, or acted on more precisely, or remembered more vividly, or understood more subtly. A list of the causal outputs of instances of attention would therefore need to be a broadly heterogeneous one. It would also, more problematically, need always to be hedged and always to remain openended. Attention is involved in some, but not all and not only, instances of acute vision, in some, but not all and not only, instances of precisely coordinated action, and so on. Having rejected the idea that there is any particular process of attention, it should be no surprise that there is no particular casual output that cases of attention share. Even a furrowed brow and a fixed gaze, which one might think of as the characteristic symptoms of attention, are not obviously simple effects of it (Ryle 1949: 133). They might instead be tactics that we sometimes use to help maintain attention. Just as the things that can be done attentively vary in ways that beget heterogeneity in the outputs of attentive performances, so too are the causes of attention various in ways that beget heterogeneity on the input side. We should therefore be suspicious of any attempt to identify a particular set of inputs as being the characteristic causes of attention. A light can catch one’s attention because it is bright; a man can catch one’s attention because he looks like one’s father; a sound can catch one’s attention because it is close by; an intrusive thought can catch one’s attention because it is erotic; and a possible solution to a puzzle can get one’s attention because it is mathematically elegant. These are properties of quite various items, and they are properties that are quite different from one another in their cognitive handling. None of them guarantees that its possessor will come to be attended, but each can be causally relevant to attention if the context is right. As with the list of attention’s characteristic outputs, a list of the inputs of attention would have to be heterogenous, and hedged, and open-ended. What we need, then, rather than a description of a characteristic input/output transition that attention mediates, is an account that allows attention to be relevant to the cleverness and skillfulness of a wide range of thoughts and actions without being tied to any thought or action types in particular.

5.4 ATTENTION AS AN ENABLING CONDITION The fact that attention’s causal work is not a matter of its intervening between a characteristic set of inputs and outputs is reflected in our

92

Attention Is Cognitive Unison

everyday talk about it. In most everyday contexts it is slightly unnatural to say, ‘His attention caused him to notice her coming in’. But there is nothing unnatural about saying, ‘Because he was paying attention, he noticed when she came in’. There is so little semantic difference between these sentences that their difference in naturalness must be accounted for by a pragmatic difference in focus. In the second sentence it is her coming in that is the cause of his noticing, given his attention as a background condition. In the first sentence the causal work is credited to his attention more directly. This difference in focus explains their difference in naturalness. In general, attention prefers not to occupy the causal foreground in normal conversational contexts. There is a loss of naturalness when it is made the topic of a causal sentence. It is a consequence both of the diversity of attention’s causes and effects and of attention’s preference for the causal background that when we think about causation from a mechanistic perspective we find ourselves unable to state the causal interactions that a theory of attention needs to allow for. If we think of causal relevance as requiring participation in the mechanism by which some transition is affected, or as requiring the exertion of a force on an object, then the causal role of attention remains elusive. Attention is not a push, a pull, or a trigger, nor is it a stage in a mechanism for performing any particular sort of work. If we think of causation from a mechanistic point of view then attention’s causal relations tend to disappear from view, and with them its title to be counted as a genuinely explanatory mental phenomenon.

5.5 COUNTERFACTUALS The alternative conception of causation that suggests itself once the mechanistic conception has, for these purposes, been rejected is the conception of causation as a relation of counterfactual dependency. This conception is better placed to recognize the causal relevance of variously distributed background conditions. The idea that counterfactual dependencies should provide our characterization of the causal role assigned by common sense to attention is suggested by the form of our everyday descriptions of the contexts in which attention matters. What we think is not that attention triggers the tightrope walker’s walking, but that it makes a difference, as background conditions do, to the way in which that walking turns out: We think, that is, that if the tightrope walker had not paid attention then his walk would not have been successful; that if the examination candidate had not paid attention then her thinking would have been more muddled; and that if the inspector had not been paying attention then he would not have noticed the fingerprint. These natural things to say embed counterfactual conditionals: We say that if the tightrope walker had not paid attention, then his walk

The Causal Life of Attention

93

would not have been successful. Such sentences claim that successful tightrope walking stands in a relation of counterfactual dependence to the tightrope walker’s attention. They make claims about things that did not actually happen. For that reason, they have sometimes been regarded by metaphysicians and by scientists as suspicious. I will not have anything to say here to allay these suspicions about counterfactuals per se. I will, however, defend the use of counterfactuals in characterizing the causal role of assigned by common sense to attention. At this point in the dialectic, remember, we are merely finding a way to state the role that common sense credits to attention, with a view to seeing whether that is a role that cognitive unison is able to occupy. We are, that is, filling in premise 2 of the argument schema given at the beginning of this chapter. Our project, at this point, is one of descriptive and not revisionary metaphysics. Whatever one’s ultimate view of counterfactual conditionals, our goal here is just to make precise the role that they play in our commonsense conceptual scheme.

5.6 THE CAUSAL RELEVANCE OF ATTENTION PER SE We have seen a couple of reasons for thinking that counterfactual conditionals, rather than mechanistic descriptions or sets of input/output rules, are the best form for an analysis of attention’s causal role to take, but we have not yet said which counterfactual conditionals we should use. Since causation is a relationship between events one might think that the counterfactual conditionals that we need in order to characterize attention’s causal role are simply those that link one event—the one that constitutes the subject’s attention—to another event—the one that constitutes the behavior for which attention is required. That thought would lead to the conclusion that a theory of attention is able to resist mental-causation objections if and only if it is a theory that makes the following counterfactual conditional come out true: α: If the attention-constituting events had not happened, then the attention-requiring behaviors would not have happened.

But this cannot be right. If α were the correct characterization of attention’s causal role then avoiding mental causation objections would be far too easy. According to the cognitive unison theory the events constituting a subject’s attention to a task are just the events of cognitive processing that take place in the background set for that task. These include all the bits of processing that the agent executes, with understanding, in the service of his task, and so it is of course true that if we were to remove those events then the relevant behaviors would fail to occur. This is true just because cognitive processing is the thing that brings about behavior. Counterfactual

94

Attention Is Cognitive Unison

α, then, is certainly true according to the analysis of attention offered by the cognitive unison theory, but its truth is not enough to show that the cognitive unison theory is able to account for the causal relevance of attention. The point on which the inadequacy of α hinges is analogous to a point that was emphasized in the wake of Donald Davidson’s anomalous monism (Davidson, 1980; Honderich 1982; Smith 1982). More needs to be done to secure causal efficacy for a phenomenon than just having the stuff of which that phenomenon is made be the stuff that does the causing. We do not, for example, show that voodoo is causally efficacious by showing merely that some deaths are caused by voodoo dolls. It may very well be that some deaths are caused by voodoo dolls. It may be, for example, that such dolls are left lying around and that hungry and indiscriminate dogs choke to death on them. That is not enough to demonstrate the efficacy of voodoo per se. What the demonstration of its efficacy would require is not only that the dolls cause some deaths, but that they cause those deaths in ways relative to which the property of being voodoo is relevant. What we need, as a characterization of attention’s causal role, is not just for there to be behavior that depends counterfactually on the attention constituting events, but for attention per se to be credited with a causal role. We need relevance for the events that constitute attention, and relevance also for the fact that those events have the property of being attention constituting. The counterfactual conditional α is not adequate as a characterization of the causal role that attention needs to play because its truth does not require the property of being attention constituting to be causally relevant, but we can capture the causal relevance of the property of being attention constituting, without turning our backs on counterfactual conditionals, by adding another counterfactual to our account. We get the counterfactual that we need by analyzing what it is for a property to be relevant to the holding of a causal relation as follows: Property p was relevant to the causation of e by c if and only if (1) c caused e and (2) the particular causal relation that existed between c and e would not have held if p had not been instantiated. When we apply this counterfactual theory of causal relevance to the property of being attention constituting, the characterization of attention’s casual role that results is this: β: (1) There are some causal relations between attention-constituting events and behaviors (of various skilled and clever kinds), and (2) if those events had not had the property of being attention constituting, then those events would not have stood in the same causal relations to those behaviors.

The metaphysical claim that drives this characterization of attention’s causal role—the claim that counterfactuals can be used to analyze what it is for properties to be causally relevant—is clearly in the spirit of the counterfactual analysis of causation proposed by David Lewis (1973a,

The Causal Life of Attention

95

2000). But the claim that I am making here, unlike the claim that Lewis made, is not a claim about the causal relation itself. It is only a claim about what is required for a property to be relevant to the holding of that relation. Despite the logical independence of my view from Lewis’s, many of the advantages of the Lewisian analysis are shared by the view that I am advocating.

5.7 COUNTERFACTUALS AND CAUSALLY RELEVANT PROPERTIES Metaphysical analyses, especially descriptive metaphysical analyses, need to be in step with epistemological practices. The most powerful reason for including counterfactuals in one’s analysis of the metaphysics of causation, whether in the analysis of the causal relevance of properties, as here, or in the analysis of the causal relation itself, as in the work of Lewis, is that counterfactual conditionals figure prominently in the epistemology of causal relations. When thinking about whether two events are causally related we naturally proceed by asking whether one event makes a difference to the occurrence of the other. This is so immediately natural that it does not seem to be a procedure inferred from an extraneous theory. It seems, rather, to be something that comes directly from our grasp of the concept of causation. Any theory that does not incorporate such counterfactuals therefore risks being a theory of something other than the referent of that concept. The use of counterfactuals in reasoning about the causal relevance of properties is just as natural as the use of counterfactuals in reasoning about the causal relevance of events. An example of a simple causal interaction (adapted from Honderich’s 1982 critique of Davidson) serves to illustrate the point: The placement of some pears on the tray of the grocer’s scales causes the needle to move round the gauge. Some of the properties of the pears are relevant to this causal interaction. Others are not. The pears’ weight is one of the relevant properties. This is reflected in the truth of the counterfactual conditional: If the pears had not had the weight that they actually have, then their placement on the scale would not have caused the needle to move in the way that it actually did. The pears’ color is one of the irrelevant properties. That is reflected in the fact that the corresponding counterfactual comes out false: If the pears had had a different color, then their placement on the scale would still have caused the needle to move as it did. The placement of the pears around the chameleon, on the other hand, brings about the change in the chameleon’s pigment. To this causal interaction, the color of the pears is relevant while their weight is irrelevant. The relevance of their weight to the chameleon’s change in pigment is reflected in the fact that the corresponding counterfactual now comes

96

Attention Is Cognitive Unison

out false: It is not the case that had the pears had a different weight, then their placement around the chameleon would not have stood in the same causal relation to the chameleon’s change of color. But the counterfactual conditional linking the color to the causation of the chameleon’s change comes out true. A different causal relation would have held if the color had been different. In these examples the counterfactual analysis of the causal relevance of properties returns the correct answers when the property in question is a property of an object that participates in the causing event (a property such as the color or weight of the pears). It also gets to these verdicts by a procedure that is itself recognizably commonsensical. So far so good, then, for the use of counterfactuals in assessments of causal relevance.

5.7.1 Causally Relevant Properties of Events The counterfactual analysis of what it is for properties to be causally relevant continues to accord with our intuitive judgments of causal relevance when the property whose causal relevance is in question is not a property of an object that participates in the event that does the causing, but is instead a property of that event itself. If one wants to know whether or not the hastiness of a breakfasting was relevant to its causing indigestion one asks whether it is true that, had the eating of the breakfast not been hasty, it would still have caused the indigestion. If that counterfactual conditional comes out true then the haste is relevant. If, on the other hand, the nearby worlds with nonhasty breakfasts are still worlds in which the same causal relation exists between the breakfast and the indigestion, then although hastiness might be a property that was relevant to some other effects that the breakfast brought about, such as the mess in the kitchen, it was not a property that was relevant to the breakfast’s causing of the indigestion. For all these examples the counterfactual analysis of the causal relevance of properties gives the right verdicts. All of this suggests that β makes the right use of counterfactuals in characterizing the causal role that commonsense psychology attributes to attention.

5.8 OBJECTIONS TO COUNTERFACTUAL ANALYSIS OF CAUSATION AND OF CAUSAL RELEVANCE My claim connecting counterfactual conditionals with the metaphysics of causal relevance is a proposal in the spirit of Lewis’s counterfactual analysis of the causal relation itself, but it is nonetheless distinct from it. This distinctness of my view from Lewis’s can cut both ways in the dialectic. It entails that an argument for Lewis’s view is not, ipso facto,

The Causal Life of Attention

97

an argument for the view that I am defending here. It also entails that arguments against Lewis’s view are not, at least not ipso facto, arguments against my view. The sorts of metaphysical considerations that motivate hostility to Lewis’s view need not motivate hostility to mine. If those considerations stem from a generalizable animus to facts about nonactual possibilities, then they might motivate such hostility; the point is only that they need not. If you reject Lewis’s counterfactual view of causation because you dislike counterfactual views tout court, then my counterfactual view of causal relevance is unlikely to seem appealing to you. But if you reject Lewis’s view because you think that there are specific problem cases that it cannot handle, then you might actually welcome my view. You might think, in that case, that it is in the analysis of the causal relevance of properties that counterfactual analyses find their natural home, and that by finding a role for counterfactuals here we do enough to explain why they persist in seeming relevant to causation, thereby making room for a noncounterfactual theory of causation itself, while preserving the plausible idea that counterfactuals do have an explanatory role to play somewhere in this neighborhood. In order to show that this last position is a stable one I shall show that my view remains plausible, even when we have taken account of the kinds of problem cases that make trouble for the Lewisian account of the causal relation itself.1

5.8.1 Simple Preemption Cases The sort of case that most clearly makes trouble for the Lewisian counterfactual analysis of the causal relation itself is the sort in which causal interactions occur, but occur without the corresponding counterfactual dependences. Such cases are found when the cause is backed up by a failsafe mechanism, or when it is the preempter of some other event that would have brought about the relevant effect anyway, even if the preempting cause had been absent. These cases are problematic because we have clear intuitions that the preempting cause, or the cause that is backed up with a failsafe mechanism, is still a genuine cause, even when there is some other event or latent mechanism to ensure that the effect would have happened anyway and so does not counterfactually depend on the presence of that cause. The standard example is that of two rocks that are thrown at a bottle (Lewis 2000). The first rock to hit the bottle is clearly the cause of the bottle’s breaking, even if the other rock was not far behind it and would have broken the bottle anyway. Since the second rock blocks the counterfactual dependence between the first rock and the breaking of the bottle, and since it does so without compromising the causal relation between them, it seems that that causal relation must consist in something other than that counterfactual dependence. This looks to be a problem for the Lewisian, who wants

98

Attention Is Cognitive Unison

to use the counterfactual dependence in explaining what the causal relation is. Whether or not such cases constitute an insurmountable problem for the counterfactual analysis of the causal relation itself, they do not provide the same sort of problem for the counterfactual analysis of the causal relevance of properties. This is because the intuitions that one depends on in using such cases as an argument against the counterfactual analysis of the causal relation itself are not available when a parallel case is constructed as an argument against the use of counterfactuals to assess the causal relevance of the properties of events. This point can be made clear by working through an example. My opponent aims to produce a case that challenges the counterfactual analysis of causal relevance in a way that is analogous to the way in which cases of preemption challenge the counterfactual analysis of causation itself. To do so he needs to find a case in which there is a property that is causally relevant to the production of some outcome, and he needs to augment that case by adding a ‘failsafe’ property (analogous to the second rock) that will ensure that the effect would have been brought about even if the causally relevant property were removed. Suppose that in order to do this our opponent takes the case of a hasty breakfast that causes indigestion. He starts out by specifying that this is a case in which the hastiness of the breakfast is relevant to its causing of the indigestion. He then changes the case so that the breakfasting has some additional property, playing a role analogous to that played by the second rock, which ensures that the breakfasting would have brought about indigestion anyway, even if it had not been hasty. Let this backup property in our example be the property of gluttonous indulgence. The breakfast we are now imagining was not merely hasty. It was also excessively large and greasy. The resulting case is analogous to the case that makes trouble for counterfactual analyses of the causal relation itself. It would create a problem for the counterfactual analysis of the causal relevance of properties if it were a case in which the property of hastiness remained causally relevant, despite the fact that the addition of gluttony breaks the counterfactual dependence relation between the breakfast’s causing of the indigestion and the breakfast’s having the property of hastiness. But the case that we have constructed does not seem able to do the work that our opponent wants from it. The problem does not get going because the causal relevance of the haste, unlike the causal efficacy of the first rock, is impaired when the failsafe property is added. The problem posed for the Lewisian by the case of a preemptive cause gets going only because it is clear that, by the time the second rock gets to the place where the bottle was, the causal work is already done. It is therefore clear that the presence of the second rock does nothing to rob the first rock of its status as the cause of the bottle’s breaking. But there is no such asymmetry to ground the analogous intuitions in the case of causally relevant properties. Since hastiness and gluttonousness are

The Causal Life of Attention

99

two properties of the same event, there is no analogous sense in which the haste has already done its work in bringing about the indigestion, before the gluttony gets a chance. Gluttony and haste arrive on the scene together, as properties of the same event. The haste’s claim to being the cause of the indigestion is therefore compromised by the addition of gluttony. The causally relevant property in the augmented case seems to be the conjunction of haste and gluttony, working together, or perhaps the disjunction of the two. Either way, we find that, when it is the causal relevance of the properties of events that is at issue, the adding of properties that defeat the counterfactual dependence does undermine the causal relevance of the original properties. These cases therefore do not create problems for the counterfactual analysis of causally relevant properties that are analogous to the problems that preemption cases create for the counterfactual analysis of the causal relation between events.

5.8.2 Latent Backup Mechanisms Preemption cases in which one actually occurring event is preempted by another actually occurring event are not the only sort of case that make trouble for the counterfactual analysis of the causal relation. Rather than adapting such a preemption case our opponent might instead look to the case where the truth of the counterfactual is blocked, not by another actually occurring event, but by a merely latent backup mechanism. The analogy to this case for the counterfactual analysis of causally relevant properties might seem more problematic. We no longer have the causally relevant property and its counterfactual-defeating backup arriving on the scene together as properties of the same event. Instead we have the causally relevant property arriving on the scene, but the counterfactual defeating backup is merely waiting in the wings. Again the point can be clarified by considering an example. For a case of this sort, imagine the scenario in which a huge breakfast is set before a hungry man who has just noticed that he will be late unless he hurries. The man eats a hasty breakfast, leaving most of the huge meal untouched. He suffers indigestion as a result of his haste. The haste of the breakfast is, we may stipulate, relevant to its causing of the indigestion. The counterfactual analysis that I have been advocating therefore requires there to be a counterfactual dependence here, such that if the breakfast had not been hasty it would not have stood in the same causal relation to the indigestion. But is that counterfactual true? Is it true that if the breakfast had not been hasty then there would not have been the same causal relation between the breakfast and the indigestion? The following line of thought might lead one to say that it is not true: To assess the counterfactual we need to consider the nearest possible world in which the breakfast is nonhasty. The nearest possible world in which the breakfast is not hasty is one in which the man takes his time over the

10 0

Attention Is Cognitive Unison

breakfast. But now that that extra time is allowed the man can get around to the bacon, the sausages, the fried bread, the black pudding, and the scrambled eggs, all of which his haste did not previously allow time for. In fact, our man can now get around to far too many of those items, and the result is that the world we are envisaging is one in which the breakfast causes indigestion once again. The original case is constructed to be one in which haste is relevant to the breakfast’s causing of indigestion, but it appears to be a case in which it is not true that, if the breakfast had not been hasty then there would have been no causing of the indigestion: If the breakfast had not been hasty, then it would have been gluttonous, and if the breakfast had been gluttonous, then it would have caused indigestion. A case like this could be used to show that counterfactuals cannot be used to analyze claims about the causal relevance of properties of events only if it really were a case in which the relevant counterfactual is false, but where there is causal relevance. The argument sketched above does not go far enough to show that this is such a case. It offers a line of thought that prompts us to say that the relevant counterfactual is false, but in offering this line of thought it moves us into a new context. Contexts are notoriously important in determining the way in which counterfactuals should be assessed (Lewis 1973b: 66). Some contexts lead us to assess counterfactuals in ways that are not appropriate for the testing of causal claims. The line of thought offered above does not show that the counterfactual is false on the interpretation that would be relevant to its use in the analysis of causal relevance. We can see this by following Lewis, and others, in drawing a distinction between two different ways of interpreting counterfactuals, one of which is favored by some contexts, and the other of which is favored by others. When assessing a counterfactual conditional, if P then Q, we ask whether Q is the case in the nearest possible world in which it is the case that P. Contextual factors often play a role in determining which world this is. One way to find the nearest possible world in which it is the case that P is to track back through the actual world history and to find a change that would have brought it about that P. Another way to find the nearest possible world in which P is the case is just to stipulate a world in which more or less everything is the same as it is in this world, except for the fact that P. The results of these two procedures may be quite different. Consider this sentence: ‘If more microwave ovens were being purchased in America, then the economy would be improving.’

On a backtracking interpretation this sentence might come out true, since on that interpretation one looks for a nearby world in which there is a reason why more microwave ovens are being purchased, and asks whether that world is, as it might well be, a world with an improving economy.

The Causal Life of Attention

101

On a nonbacktracking interpretation, in contrast, the sentence would almost certainly come out false. On that interpretation all we do is find a world that is just like the actual world except that more microwave oven purchases take place. Such a world would be good for the manufacturers of microwaves, but there is no reason to think that merely adding microwave purchases to a world would do very much to change the overall economy for the better. There is therefore no reason to think that the counterfactual is true on the nonbacktracking interpretation. Having seen the difference between these two ways of interpreting counterfactual conditionals it is easy to see that distinguishing between them is important when we are using counterfactuals to assess or to analyze claims pertaining to causation (Lewis 1979: 34; Bennett 2003: 482). If the sentence linking microwave purchases and economic conditions were true on the nonbacktracking interpretation then this would be causally revealing. It would be a reason to think that one could cause the economy to improve by incentivizing the purchase of microwave ovens. But the fact that the conditional is true on the backtracking interpretation provides no such clues as to what might cause an economic change. The backtracking interpretation merely tells us about a possible indicator of economic change. Backtracking interpretations are no good when using counterfactuals to assess causal claims because, when backtracking, we look for a nearby world in which something else happens to bring it about that the alleged cause is gone. But this ‘something else’ that removes the alleged causing event from the world can interfere by itself influencing the presence or absence of the alleged effect. The result is that if we permit backtracking interpretations in contexts where counterfactuals are used to assess causal claims then epiphenomenal events at the end of causal branch lines are prone to looking like real causes. Contexts that permit backtracking are therefore not appropriate contexts in which to interpret counterfactuals when those counterfactuals are being used to assess or to analyze claims of causal relevance. This is not a problem for the use of counterfactuals to assess causal claims. It means only that we need to specify that the counterfactuals that are being used to assess causal claims must be interpreted in ways that do not permit backtracking. With that stipulation in place we find that our putative problem case disappears. The line of thought that led us to say that, because the nearest nonhasty worlds are gluttonous worlds, the breakfast’s causing of the indigestion is not counterfactually dependent on the breakfast’s haste was a line of thought that encouraged us to backtrack. It suggested that a nonhasty breakfast would be a gluttonous one, but it got to that conclusion not by simply going to a world in which the haste of the breakfast has been removed—instead, it tracked back to a world in which there is something else (a greater availability of time and so more relaxed state of mind on the part of the agent) causing the breakfast not to be hasty and

102

Attention Is Cognitive Unison

also causing the breakfast to be gluttonous. This example therefore does not provide a counterexample to the analysis of the causal relevance of properties in terms of the counterfactual dependence of the casual relation on the instantiation of the property. I do not know that there is not another example out there that would do the job—the recent history of philosophical attempts to employ counterfactuals in the analysis of causal phenomena would suggest that a sufficiently crafty philosopher might be able to construct one—but it would have to be an example that does not trade on covert backtracking and in which the removal of the causally relevant property really does leave the same causal relation intact. These considerations show that some of the objections that are usually raised against counterfactual analyses of causation do not get a grip against the use of counterfactuals that I am advocating here, as providing an analysis of what it is for properties of events, or for the properties of objects that participate in those events, to be causally relevant. The upshot of all of this is that counterfactuals like those that feature in β are perfectly well suited to giving a descriptive metaphysican’s characterization of attention’s causal role. We have been looking for a way to characterize the causal contribution that, according to commonsense psychology, attention makes to the performance of the clever and skillful acts that depend upon it: acts such as tightrope walking or the taking of examinations. Natural language, together with some other commonsense considerations, suggested that the way to characterize this contribution was as a counterfactual dependence. Considerations of the sort raised in the wake of Davidson’s anomalous monism showed that there was also a need to ensure that we were characterizing a role played by attention per se, and not simply a role played by the events that happen to be attention constituting. We have seen that the counterfactuals that enable us to meet this need are counterfactuals that attribute causal relevance to the property of being attention constituting. We have now defended that use of counterfactuals in the analysis of what it is for a property to be causally relevant, and so have defended the idea that the causal role played by attention may be characterized in the claim that we labeled β: (1) There are some causal relations between attention-constituting events and behaviors of various skilled and clever kinds, and (2) if those events had not had the property of being attention constituting, then those events would not have stood in the same causal relations to those behaviors.

This account of the causal role played by attention gives us our specification of what is required for a theory of attention to be adequate to the task of explaining attention’s causal role: A theory of attention has to be one on which the causing of behaviors and thoughts can depend

The Causal Life of Attention

103

counterfactually on the fact that the events doing the causing have the property of being attention constituting. The question now is whether the cognitive unison theory satisfies this requirement. I shall address three reasons for thinking that it might not. The first is based on the fact that whether or not a set of processes is operating in unison depends on factors extrinsic to the unison-constituting processes. The second is based on the fact that cognitive unison has a privative character. The third is a more familiar sort of causal exclusion objection, turning on the fact that cognitive unison is a higher order sort of thing, whose causal significance therefore seems to be exhausted by that of the lower level processes that constitute it. We shall consider these objections in turn.

5.9 THE EXTRINSICNESS OF UNISON The first of our problems is a problem about extrinsicness. To many philosophers it has seemed natural to think that the causal relevance of a property requires that property to be intrinsic. Colin McGinn, for example, writes : What happens in the causal nexus is local, proximate and intrinsic [. . .] The causal powers of a state or property [. . .] cannot depend upon relations to what lies elsewhere. (cited in De Muijnck 2002: 125)

If McGinn is right that the ‘local, proximate and intrinsic’ nature of causation means that there is a general prohibition on the causal relevance of properties that ‘depend on relations to what lies elsewhere’, then cognitive unison would indeed fail to be causally relevant since having the property of being an instance of unison does depend upon relations to what lies elsewhere. We can make the nature of this dependence clear by considering the analogy with orchestral unison. Suppose that the strings and the woodwind are all playing the same melody, and that nobody else in the orchestra is playing anything at all. In such a case it is true of the orchestra that they are playing in unison. The activities that constitute this unison are the activities of the string players and the activity of the woodwind players. But these playings constitute unison only because of their relation to what lies elsewhere: The fact that we have a case of unison depends on the fact that the brass players are not playing some other tune. If the brass were to strike up with something different, then the orchestra would no longer be playing in unison. The same is true with cognitive unison: A set of cognitive processes that are operating in the service of a single task constitute attention on the part of the agent if and only if the other processing resources in the background set for the task do not, so to speak, strike up a different tune (that is, as long as they do not undertake processing that does not serve the task in hand). It is in this sense that unison depends on relations to what lies elsewhere.

10 4

Attention Is Cognitive Unison

There is a thought experiment that can motivate the thought that, as McGinn suggests, extrinsicness of this sort is a disqualification from causal relevance. The thought experiment starts out by taking some bit of behavior for which we think attention is causally relevant, and which is therefore such that the cognitive unison theorist must say of it that cognitive unison is relevant to its production. Having found such a behavior, the thought experiment proceeds by constructing a set of events that includes every event that could feasibly be thought to play any role in bringing about that behavior. The objection to the efficacy of unison comes from the fact that the events that are allowed into our exhaustive set of causally relevant events do not determine whether or not there was an instance of unison. Since unison requires something over and above the occurrence of these events, and since the set of events was constructed in such a way that it included every causally relevant event, it can be made to seem that the property of being in unison cannot be causally relevant. This tactic plays out as follows: Let our instance of a behavior to which attention is relevant be the catching of a ball. Now, for the purposes of this thought experiment we want to build a set of events that includes every event that could feasibly be thought to play a role in bringing about the catch. The catch itself is made on account of certain muscles contracting at the right times, so, into our set of all causally relevant events, we can put all the efferent nerve firings that triggered the muscle contractions that closed the hand around the ball. Next we move back a step in the causal chain and find the nervous activity that contributed to these efferent nerve cell firings. The events that make up this nervous activity also go into our set of causally relevant events. These neural firings are themselves triggered by some identifiable set of neural firings, for there are only so many neurons that synapse with the efferent nerves, and only a subset of these will have secreted their neurotransmitters at the relevant moment. All the neural events that contributed to the causally relevant firings get added to our set. And back we go, adding to our set of efficacious events all the firings of all the nerves that feed into the nerves that feed into the nerves that. . .feed into the efferent nerves that cause the muscles to contract. Eventually, and after an enormous number of events have been listed, we will have got back, through the motor cortex, through the ball-tracking centers of the visual cortex, to the light from the ball falling on the retina. The set that we have constructed will be a large subset of the neural events that took place in the seconds before the ball catch. It will not, however, contain anything like all of the neural events that took place in the relevant time period. The uptakes and secretions of neurotransmitters that are going on in those bits of brain that are processing the words on the side of the bus that happens to be passing near the ball catcher at the relevant time will not make it into the set. Nor, I suppose, will those bits of the brain that are responsible for the agent’s being reminded of a similar ball that he caught the week

The Causal Life of Attention

10 5

before. But the set of causally relevant events is constructed in such a way as to make it seem plausible for the objector to say that nothing outside it can plausibly be said to contribute to the psychological cause of the action. Nothing outside the set actually contributed to the firing of any of the neurons that initiate the catch. And yet, the objector will say (quite rightly) that a part of what was crucial to the incident’s being an instance of cognitive unison has not been included. The objector will be right to claim that the set he has constructed leaves out facts that play a role in determining the presence of cognitive unison. He will be right because nothing in this huge catalog of events tells us about the resources that could have, but on this occasion did not, serve the agent’s ball catching. In particular, it does not tell us whether those resources were simply doing nothing or were engaged with the processing of distracters. But this fact about resources that only potentially serve the task of ball catching can make the difference between cognitive unison and the lack of it, just as the fact that the brass players were simply doing nothing, rather than being engaged in the playing of a different tune, makes the difference between orchestral unison and the lack of it. This thought experiment makes for an objection to the cognitive unison theory because the background facts necessary for unison do not get to intervene when the firings in our imagined catalog do their causal work. These facts, like an idle cog in a machine, are peripheral to the route along which the causal impetus travels. If the difference between unison and its absence can be made wholly in the background then how, the objector asks, can unison be causally relevant? That is the extrinsicness objection to the efficacy of cognitive unison. I have tried to make this problem seem like a pressing one, but, having seen that it is causal relevance as captured in counterfactual dependence relations that we need to account for, we have already said enough for it to be clear where the objector puts his rabbit into the hat. When our objector builds the set that purports to contain all the causally relevant events, he builds it according to a principle that includes all and only the cell firings that participate in the chain that leads up to the firing of the efferent cells. But that principle is plausible only if one is thinking of causal relevance as requiring a direct contribution to the operation of a mechanism. That, as we have seen, is not the way we think of the causal role of attention. We think of the causal role of attention in terms of counterfactual dependences. Once we start thinking in these terms the problem goes away because the way in which the sequence of nerve firings proceeds can perfectly well depend counterfactually on events that were remote from the actual chain down which the energy traveled, as when the ball’s being caught depends counterfactually on the lack of intervention by potential preventers. The set of events that contribute causal oomph to the ball catching do not determine whether or not there was unison, but that fact excludes unison from causal relevance only

10 6

Attention Is Cognitive Unison

when causation is thought of mechanistically. We have seen that this is not the way we think of attention’s causal role, so it does not matter to the identification of attention with cognitive unison that cognitive unison, being somewhat extrinsic, cannot do causal work of this sort. If the objector tried to run his thought experiment again in a way that would account for this he would have to include in his set of all the feasibly causally relevant events, not only all the events that contribute energy to the firing of the efferent nerves, but also all the events that these firings counterfactually depend on. He would find himself with a very large set indeed, and now the events in the set would determine whether or not there is an instance of unison.

5.9.1 Distinguishing Relevant from Irrelevant Extrinsic Properties One might be suspicious of the ease with which the extrinsicness problem is avoided. Our standards of casual relevance for properties should exclude some extrinsic properties precisely because they are too extrinsic to do causal work. When, as in the previous section, we see that the move to a counterfactual analysis of causal relevance enables extrinsic properties to avoid the problems that threaten to bar them from causal relevance, we face the worry that the move to a counterfactual analysis of causal relevance has lowered the standards to such a degree that all sorts of properties will get to count as causally relevant, including extrinsic properties that depend on relations to remote and causally isolated times and places. I claim that this worry arises from a lack of clarity about how we should assess the relevant counterfactuals. Once we are clear about how to assess counterfactuals involving extrinsic properties for the purposes of causal evaluation it will become clear that the route by which the instantiation of cognitive unison gets to count as causally relevant is not a route that can be exploited so as to claim causal relevance for any old extrinsic property. This section is therefore devoted to showing that the use of counterfactuals that I have been advocating as a way of gauging attention’s causal role does not enable spurious causal efficacies to be claimed. What I have been suggesting is that property ε’s being relevant to C’s causing of E can be analyzed with a counterfactual conditional saying that if C (or one of the objects that participates in C) had not had ε, then C would not have stood in the same causal relation to E. When one wants to know the truth of the counterfactual conditional ‘if P had not been the case, then Q would not have been’, one goes to the nearest possible world in which P is not the case and finds out whether Q is the case in that world. The counterfactual is true only if Q is not the case in the nearest world where P is not. The point that needs to be clarified

The Causal Life of Attention

107

is a point about which of the worlds in which P is not the case counts as nearest. When the world we are looking for is the nearest world in which an extrinsic property is lacking, there are, even with backtracking interpretations ruled out, at least two sorts of world that have a claim on being the nearest. If an object, O, has an extrinsic property, ε, it will typically be the case that there are worlds in which differences with things other than O make it the case that O lacks ε at those worlds while keeping all its intrinsic properties. And there will be different worlds in which the items other than O remain the same but in which O lacks ε nonetheless, on account of a change in O’s intrinsic properties. Suppose that we are considering Ortcutt—the man with the property of being the shortest spy. There are possible worlds in which Ortcutt is not the shortest spy because there is somebody yet shorter among the spies, and there are also possible worlds in which Ortcutt is not the shortest spy because Ortcutt himself is taller in those worlds. In order to assess a counterfactual of the form ‘if Ortcutt had not been the shortest spy, then . . .’, we need to know whether the nearest world in which Ortcutt is not the shortest spy is one of those worlds in which there is somebody yet shorter among the spies, or whether it is one of those worlds in Ortcutt himself is taller.2 The question about which of these worlds is nearer is not a question about the objective geometry of modal space. It is a question about what sort of similarity metric we should adopt when evaluating the counterfactual for the purposes of assessing causal relevance. When considered abstractly this point might seem to be a moot one, but when we consider cases it becomes clear that, when it is for the purposes of assessing the causal relevance of O’s being ε that you want to find a world in which O lacks ε, you should find a world in which O lacks ε while retaining its intrinsic properties. This must be the right way of assessing such counterfactuals, on pain of ending up with far too many causally relevant extrinsic properties, and thereby trivializing the notion of causal relevance. Take ‘being the shortest spy’ as an example of an extrinsic property, and take, as a candidate for a causal interaction to which ‘being the shortest spy’ is a relevant property, the fact that Ortcutt’s infiltration of the Stunted-Gnome Gang caused the end of their reign of terror. It might seem like a perfectly natural thing to say that the success of Ortcutt’s infiltration depended on his being the shortest spy. After all, the Stunted-Gnomes are tiny, and a taller spy would have stood out a mile. But this is a case in which prima facie appearances are easily found to be deceptive. Insofar as it seems right to say that the success of Ortcutt’s infiltration depended on his being the shortest spy, it is because we are thinking of the relevant counterfactual (‘if Ortcutt had not been the shortest spy, then it would not have been the case that his

10 8

Attention Is Cognitive Unison

infiltration ended the Stunted-Gnomes reign of terror’) as being made true in the following way: We go to a world in which Ortcutt is not the shortest spy and that, we mistakenly suppose, is a world in which he is taller, and we find, in that world, that the Stunted-Gnome gang would never have taken Ortcutt—a taller man there than he was in the actual world—to be one of their own. Although it may not, on the face of it, be counterintuitive, this is the wrong way to think of the counterfactual as being made true, at least for the purposes of assessing causal relevance. If we were to assess claims about the causal relevance of extrinsic properties by assessing the relevant counterfactuals with an intrinsic-change reading, then we would get true counterfactuals to support the causal relevance of far too many of Ortcutt’s extrinsic properties. Ortcutt would have been much less successful as a spy if he had had bright orange skin, for he would then have had great difficulty hiding in the shadows. The notebook in front of me now has a very bright orange cover. But, being less brightly colored than my notebook surely was not causally relevant to Ortcutt’s success, even though, given an intrinsic-change reading, the relevant counterfactual would come out true. Intrinsic change readings would lead to all sorts of spurious properties claiming causal relevance. We want a world in which other spies are shorter, not a world in which Ortcutt is taller. Only if Ortcutt is unsuccessful in those worlds was it causally relevant that he is the shortest spy.3 If we adopt an extrinsic-change reading when assessing the counterfactual then, when we go to a world in which Ortcutt is not the shortest spy, we go to a world in which he is precisely as tall as he is now but in which there are others shorter than he who are involved in spying. In a world like that we look to see whether the effects of Ortcutt’s actions are the same. When the relevance of being the shortest spy is assessed in this way, the claim that the success of Ortcutt’s infiltration depended on his being the shortest spy is found to be false. Certainly Ortcutt’s height was relevant, but how could the absence of shorter spies have made any difference at all? The Stunted-Gnomes would have been taken in just as completely if there had been a still shorter spy; after all, the still shorter spy would probably be busy elsewhere and might never impinge on the gnomes at all. It is in the nature of spies to keep to themselves. The moral is that, when assessing the extrinsic properties of an object for their causal relevance, one should keep the intrinsic properties of that object fixed and use extrinsic changes to get to a possible world at which the extrinsic property is lacking. The counterfactual can then be assessed by looking for the presence or absence, in that world, of the causal relation to which the property in question is purported to be relevant. This is not a general argument against the causal relevance of extrinsic properties. The verdict of causal irrelevance in the case of the shortest spy

The Causal Life of Attention

10 9

depends on the fact that we can add shorter spies who are causally isolated from Ortcutt. It depends on a feature of the set of spies— specifically, on their degree of causal isolation from one another. For a contrasting example consider the relevance of Mr. Smith’s being the tallest player on the basketball court to the claim that his frequent shots caused his team to win. Although the heights of other spies in no way impinged on Ortcutt’s success, the heights of other players do impinge on Smith’s. In this case there is no problem with attributing causal relevance to an extrinsic property. Make the other players taller and they will be more effective at keeping Smith from scoring. The counterfactual—if Smith had not been the tallest player then it would not have been the case that his frequent shots caused the team to win—is true, even when we give it an extrinsic change reading. His being the tallest player on the court is thereby shown to be a causally relevant extrinsic property. Extrinsicness is a bar to the causal relevance of a property only when the extrinsic properties consist in relations to things that are remote and causally isolated. The extrinsicness of the property of being in cognitive unison does not bar cognitive unison from causal relevance because it does not reach out so far as to run into the sort of problems faced by the extrinsicality of being the shortest spy: The fact that some processing is an instance of cognitive unison requires it to take place in a context where there is no processing of distracters in the background set. It does not involve relations to anything outside the subject’s head. The facts about cognitive unison do not supervene on the facts about which cognitive processes actually contribute input to the firing of a subject’s efferent nerves, but the unison facts do supervene on all the facts about the agent’s cognitive processes (together with the fact that these are all the facts about the agent’s cognitive processes). The things that could make a difference to the subject’s state of cognitive unison are located so as to be ready to make a causal difference. Worries about extrinsicness therefore give us no reason to think that unison could not do attention’s causal work. And so they give us no reason to think that there is a problem here for the identification of unison with attention.

5.10 THE PRIVATIVE CHARACTER OF UNISON AND THE PROBLEM OF ABSENCE CAUSATION A second reason why one might think that cognitive unison cannot do the causal work that we attribute to attention is based on the privative character of unison. As was discussed in chapter 4 (see §4.10.2), the condition that processing must meet if it is to count as being in unison is a privative condition: The requirement for processing to count as attention constituting is that the activity in its background set must lack a certain property. The processing must not occur against a background

110

Attention Is Cognitive Unison

of resources that is occupied with processing that does not serve the subject’s task. In our normal thinking about causation we are frequently happy to think of the negative properties of events as being causally relevant. We are happy to think that a conversation’s not being discreet is relevant to the fact that it causes a scandal, or to think that a storm’s not being anticipated is relevant to the fact that, umbrellaless, I get wet on the way home. Some metaphysicians are nonetheless reluctant to grant causal relevance to such negative facts. Their reluctance to follow common sense in attributing casual relevance to such absences is owing to their wanting to be able to think of causal relevance as a status that can only be attained by the full-bloodedly existent (see Dowe 2000: §VI.1). Since cognitive unison has a privative character, the metaphysician who has reservations about absence causation will have reservations about causation by cognitive unison. These reservations do not, however, provide the grounds for a causation-based objection to the cognitive unison theory. If the metaphysician’s qualms could be given the status of a clearly established result then it would follow from the privative character of cognitive unison that the fact that some bit of cognitive processing exhibits unison is not relevant to that event’s causal interactions. The cognitive unison theorist would therefore be forced to accept the conclusion that a subject’s being attentive is not causally relevant. But if no absences are ever causally relevant then any theory of attention will have to accommodate a counterintuitive conclusion along these lines. If there is no causation by absences then either attention or inattention must be causally inert. It cannot be that neither of these is given a privative analysis. Any theory of attention will have to accept that one or other of these phenomena could turn out to have a metaphysically problematic causal profile. The cognitive unison theory gives attention a privative analysis and thereby secures causation by inattention at the price of making causation by attention vulnerable to the metaphysician’s qualms. A theory that did not characterize attention in a privative way would secure causation by attention, but it would do so at the price of making causation by inattention the metaphysically vulnerable one. If it turns out that there is no absence causation then one or the other bullet must be bitten. The cognitive unison theorist’s bullet is the more palatable of the two. Causation by inattention matters more to us than does causation by attention, as the following considerations show. It is important that we preserve a good psychological explanation for the effects of inattention because some of inattention’s effects are acts of the agent: When inattention leads us to start looking out of the window, or to start doodling in the margin, the train of thoughts that we inattentively entertain and the things our fingers doodlingly do are not things that simply happen to us. Our inattentive thoughts are not alien to us, and we are not automaton-like in performing our inattentive doodles:

The Causal Life of Attention

111

These are things done by us, but they nonetheless fall short of being fully deliberate actions. We do not decide to do them, or judge them to be worth doing, or have any reason that we could produce that speaks in favor of their being done. To capture the actlike character of these occurrences we need to find psychological causes for them. But we do not want to locate their causes alongside the beliefs, desires, reasons, and intentions that initiate properly deliberate actions. It is only if we acknowledge inattention as a genuine psychological cause that we can find a suitably psychological cause to answer for these actions. We would not be left wanting in the same way if metaphysical qualms about absence causation meant that attention had to be removed from the set of psychological causes. The effects of attention are things such as the noticing of a detail, the following of a convoluted argument, or the successful walking of the tightrope. The status of these as genuine actions is established by locating beliefs, desires, and intentions among their causes. We do not need the tightrope walker’s attention to serve as the cause of his tightrope walking in order to recognize the walking as being an action. His desire to get across the rope, his beliefs about how to do it, and his decision to go ahead are psychological causes enough. Given a choice between losing causation by attention and losing causation by inattention, we should prefer the loss of causation by attention since we thereby save ourselves from losing sight of the actlike character of inattentive doodlings. We do not lose sight of the actlike character of any behaviors if we have to give up attention as a cause. Further reasons for preferring to keep inattention rather than attention as a cause can be found by thinking about attention and inattention’s roles within longer causal chains. Distracting someone, in contrast with getting his attention, is a matter of acting so as to introduce a new causal impetus. If I want the students in my class to pay attention then I can only ensure that the material is interesting and take measures to prevent their attention wandering off elsewhere. If, on the other hand, I want the students to be distracted, then there are lots of things that I can actively do to introduce distraction as an influence on the proceedings. I can release a live squirrel into the room, or make the lights flicker, or tell them that there will be some gossip-worthy event that evening, or do any number of other things. I initiate a causal chain, and the inattention occupies a place in this chain. The loss of inattention as a cause leaves us with broken causal chains; the loss of attention does not. If privative causation is problematic then either causation by attention or causation by inattention is problematic. Every theory of attention would then be committed to one or other of two counterintuitive results—they must say either that attention does no causing or that inattention does no causing. The privative character of cognitive unison locates the difficulties of absence causation (if such there be) on the

112

Attention Is Cognitive Unison

side of causation by attention. That seems preferable to the alternative arrangement.

5.11 CAUSAL EXCLUSION Since instances of cognitive unison are composed from nothing over and above instances of cognitive processing, the effects of unison are composed from nothing over and above the effects of those instances. But if this is so then it seems that the effects of cognitive unison have already been accounted for, as effects of the lower level unison-constituting processes, before our theory ever gets as far as mentioning the fact that these instances constitute unison. Crediting the unison itself with causal relevance therefore seems to be redundant. This again seems to lead to the view that cognitive unison is unable to play a genuine casual role, and so to the conclusion that the cognitive unison theory of attention is vulnerable to a mental causation objection. This last argument is a special case of a general argument—the ‘causal exclusion argument’—that threatens to rob any composite entity of its causal relevance, crediting relevance only to the simple things from which such entities are composed (Block 2003; Kim 1998). My intention here is not to propose a new solution to this general problem, only to show how existing work on it can be applied to the case of unison. The existing work from which I take my lead is that by Karen Bennett (2003) and by Stephen Yablo (1992, 2003). The threat that unison will be made causally redundant by the lower level processes that constitute it can be avoided if we can acknowledge the truth of ‘the underlying cognitive processes cause the behavior’ without thereby jeopardizing the truth of ‘the cognitive unison caused the behavior’. One way in which such pairs of claims can avoid jeopardizing one another is by satisfying the following two conditions: (1) The two claims differ only in that one of them employs a more generous figure/ ground partitioning of the causal scene than the other, and (2) the events included in the more generous partitioning are not unnecessary for the production of the effect. A familiar example can help us to see what these conditions mean, and why it is that satisfying them is a way to avoid conflict between causal claims. It is uncontroversially true that the striking of a match may be what caused the match to light. The striking did not, of course, guarantee that the match would light. The match lit when struck only because there was sufficient oxygen in the room, a sufficiently high ambient temperature, no high wind at the moment of the striking, no bolt of lightning preempting the effect of the striking, et cetera ad nauseum (Mackie 1974). A whole distributed sprawl of events converged upon and contributed to the match’s lighting. It is a somewhat arbitrary and somewhat pragmatic matter how we choose to partition this sprawl of influences into figure

The Causal Life of Attention

113

(the ‘cause’) and ground—(the ‘background conditions’). A more generous partitioning, in which more of the background conditions are included in the ‘cause’, may be more appropriate for some explanatory purposes. A less generous partitioning, in which we pick out only, say, the outer surface of the match and a particular portion of the box, may be more appropriate for other explanatory purposes. But, although they may compete for explanatory appropriateness, there is no tension between the truth of these different figure/ground partionings of the causal scene. The truth of ‘the friction on the outer surface of the match head caused the match to light’ does not jeopardize the truth of ‘the striking of the match caused the match to light’, and neither of these jeopardizes the truth of ‘the striking of the match in the absence of a high wind caused the match to light’. Claims that employ more generous partitionings of the causal scene do not always avoid conflict with claims that employ less generous partitionings, but they do succeed in avoiding conflict when the more generous partitionings retain whatever naturalness is enjoyed by the less generous partitioning, and when the additional background conditions that the more generous partitioning adds into what it identifies as ‘cause’ are genuinely necessary for the production of the effect. We can move without loss of truth between ‘the striking of the match caused the match to light’ and ‘the striking of the match in the absence of a high wind caused the match to strike’ because the absence of high wind is genuinely necessary for the match’s lighting. We cannot move to ‘the striking of the match by a man in a three-piece caused the match to light’, because a partitioning that includes the man and his suit needlessly includes background conditions that are unnecessary for the production of the lighting. With this in mind, we can see that ‘the underlying processes caused the behavior’ can be shown to be compatible with ‘the cognitive unison caused the behavior’ as long as (1) the unison-citing claim differs from the process-citing claim only in the generosity of its figure/ground partitioning, and (2) the added generosity of the unison-citing claim avoids letting in background conditions that are unnecessary for the production of the attention-relevant behaviors. Finding an example with which to illustrate the way in which these conditions can be met is not wholly straightforward since, as the first part of this chapter showed, attention’s causal work is more naturally thought of as the work of a background condition than the work of a cause (§5.4). But examples of attention-caused behavior can be found. Consider the driver who, only because she is paying attention, notices an unexpected hazard on a familiar road and swerves to avoid it. The swerve is an effect of the driver’s attention, although, as §5.4 might lead us to expect, explanations that cite the swerve are not as natural as their negative counterparts: ‘Why didn’t you swerve?’, ‘Because I wasn’t paying attention!’ has a naturalness that is not shared by ‘Why did you swerve?’, ‘Because I was paying attention’.

114

Attention Is Cognitive Unison

If attention is the cause of the swerve then the cognitive unison theory requires that cognitive unison be capable of causing the swerve. The truth of ‘the unison caused the swerve’ therefore needs to be compatible with the truth of a causal history that accounts for the swerve by mentioning only the various cognitive processes that initiate the movements by which the swerve is brought about. Incompatibility between those explanations is avoided if the unison-citing explanation employs a more generous partitioning of the causal scene, but does so without letting in unnecessary parts of the background. That the unison-citing explanation does indeed employ a more generous partitioning, and so satisfies the first of our two conditions, is relatively clear. The swerve is performed under the guidance of the agent’s understanding of driving. The processes that will be mentioned by the process-citing explanation of the swerve include the cognitive process that the agent can bring to bear in the service of her driving. Those processes are therefore members of the background set for this agent’s driving and are included, together with the other members of the background set, in the cause identified by the unison-citing causal claim. The unison-citing explanation therefore cites the same events as the processciting explanation, but it partitions them more generously, packing more of the background conditions into the figure of the causal scene. The compatibility of the causal claims therefore depends on whether the second condition is satisfied: on whether the more generous partitioning needlessly includes background conditions that had nothing to do with the production of the swerve. It is slightly less clear that this second condition is met. The brain is a place in which background conditions usually do matter a great deal, but not all of them matter to everything. It may be that the processes that are cited by the process-citing explanation are processes that would have brought about the swerve whether or not the additional background processes (those that are included by the more generous partitioning of the unison-citing explanation) had been doing the things that they need to be doing in order to constitute unison. For some behaviors, the additional processes included by the unison-explanation’s partitioning will not be necessary. In these cases we should reject the unison-citing explanation in favor of the process-citing explanation. But for the behaviors to which attention is relevant the additional processes included by the unison-citing explanation will be necessary. The behaviors to which attention is relevant are the cognitively fragile ones. Responses to traffic lights and to stop signs are paradigmatically not effects of attention. Experienced drivers can make those responses whether they pay attention or not. Responses to unexpected hazards on familiar roads are reckoned to be effects of the driver’s attention only because a background of competence and wakefulness are not sufficient to ensure that they happen. They are cases where the hazard detection processes mentioned by the process-citing explanation do need a distraction-free cognitive

The Causal Life of Attention

115

background if they are to succeed in bringing it about that the swerve occurs. The distraction-free cognitive background therefore does seem to be necessary for the swerve to be brought about, and the partitioning that includes it as part of the causal foreground is therefore including genuinely necessary background. In such a case, the second criterion for avoiding causal exclusion is also satisfied. This shows that claims that attribute causal work to unison are not rendered false by the claims that attribute that work to the underlying processes. There is a further point that we can add to this, in order to show that, in addition to not being rendered false by the lower level explanation, the unison-citing explanation is also not rendered explanatorily redundant. We take this additional step by showing that there is explanatory work that unison can do, but that the lower level processes are ill-suited for. The step depends on a point about the explanatory virtue of generality—a point that Yablo takes from Hume’s Treatise: Where several different objects produce the same effect, it must be by means of some quality, which we discover to be common amongst them. For as like effects imply like causes, we must always ascribe the causation to the circumstances, wherein we discover the resemblance. (cited in Yablo 2003: 316)

The point, in the present case, is that although a particular instance of swerving may be explained by the particular set of cognitive processes that initiate it, those processes will probably not have a role in hazard avoidance in general since hazards can take a variety of forms, and can make a variety of cognitive demands. In general what one needs to do, in order to bring it about that there are such things as swerves in response to such things as unexpected hazards, is to bring it about that the cognitive resources that drivers can bring to bear in the service of their driving are not occupied elsewhere. Ensuring that there is unison is a more reliable strategy for bringing it about that hazards are avoided than is any strategy of ensuring that any particular hazard-detection or swervecoordinating resources are active. For the purposes of explaining hazard avoidance, the unison-citing explanation therefore has a generality that the process-citing explanation does not enjoy. Since this generality is an explanatory virtue, there is a loss of explanatory power if we omit to mention unison. Taking this last step is important because we wanted to secure causal relevance for unison, not only so that we could avoid the mental causation objections, but also so that we could show that the unison theory does not open the door to eliminativism about attention. The additional step shows that even the ontologically austere theorist, whose theory has room for no more psychological phenomena than those that are required to explain behavior, has a reason to include attention, as the cognitive unison theory understands it, on his list of psychological phenomena with explanatory work to do.

116

Attention Is Cognitive Unison

5.12 SUMMARY This chapter has been concerned with establishing a fit between the causal potential of cognitive unison and the causal role that we attribute to attention, with a view to showing that the cognitive unison theory of attention is not vulnerable to the mental causation objection. We saw first that the causal role attributed to attention is not that of push, pull, or trigger of any particular bits of behavior but that it should instead be thought of in counterfactual terms, as a necessary background condition for certain sorts of effect. Getting clear on the nature of the counterfactual dependence required us to present and defend an analysis of causal relevance for the properties of events. Having done this we were able to consider whether cognitive unison is capable of playing the sort of causal role that these counterfactuals pick out. Three hurdles to its doing so have been identified and surmounted: the extrinsic nature of unison, the privative nature of unison, and the aggregate nature of unison. Having surmounted these hurdles we can see that the truth of the cognitive unison theory would not require any revision to our commonsense conception of attention as playing a genuine explanatory role in the production of behavior and thought.

6 Consequences for Cognitive Psychology

6.1 PSYCHOLOGY AND METAPHYSICS Around the end of the nineteenth century, when psychology was establishing itself as a science, adverbialism dropped out of consideration as a possible way in which mental phenomena can be scientifically accounted for. This, we saw in chapter 1, was not because adverbialism was refuted. It was simply that the contemporary metaphysical theorists lacked an account of the point over which adverbialists and process-first theorists disagree. One result of this was that when theories of attention were proposed in the middle of the twentieth century—at the time when attention came to occupy the place that it currently enjoys at the center of cognitive psychology’s explanatory agenda—adverbialist theories could be found only on the philosophical side of the split that had then formed between psychological research into cognition and philosophical theorizing about it. With these historical factors conspiring to keep the possibility of adverbialism out of psychology’s field of view it has been easy for psychologists to suppose that there is something incontrovertibly right-headed about the attempt to explain all mental phenomena by identifying and describing the processes that constitute them. When we take the possibility of adverbialism seriously we see that this is a mistake. The project of process-identifying explanation, whether in psychology or out of it, carries a commitment to a substantive claim about the metaphysical category of the phenomena to be explained: It requires that those phenomena belong in the category of processes. The viability of adverbialism shows that this is a claim that goes beyond anything that right-headedness alone commits us to. It is not a claim that realism about mental phenomena requires, nor is it required in order to maintain a naturalistically respectable view of the mind. If the cognitive unison theory is right then in the case of attention the commitment to a process-first metaphysics is not merely an unargued assumption, but also a mistaken one. The cognitive unison theory’s adverbialism puts it at odds with a dominant explanatory tradition in cognitive psychology, but it does not mean that the cognitive unison theory must dismiss cognitive psychology’s

117

118

Attention Is Cognitive Unison

results: Far from it. According to the cognitive unison theory, psychology can tell us lots of important facts about attention, about the forms it comes in, about the kinds of tasks that require it, about the cases in which it breaks down, and about the ways in which it is brought about. The argument for the cognitive unison theory that was given above depended throughout on psychology providing us with such findings. The cognitive unison theory has negative consequences for psychology only because it places a limit on the sort of explanatory work that can be accomplished by psychological research that is concerned with identifying and describing the individual attention-constituting processes. Taking account of this limit requires us to make some explanatory reconstruals of certain psychological research projects. It is the job of this chapter to negotiate these reconstruals, and to note their methodological consequences. Sometimes these consequences will prove to be deflationary, and sometimes not.

6.2 THE METAPHYSICAL COMMITMENTS OF THE PROCESSIDENTIFYING PROJECT When, in chapter 2, we gave an account of the metaphysical distinction between process-first phenomena and adverbial phenomena, we did so via a distinction between two different directions of explanatory approach. It should therefore be no surprise to find that a metaphysical commitment is incurred when one or other of these directions of explanatory approach is adopted—as it is when the process-identifying research of empirical psychologists is called upon to do explanatory work in the theory of attention. The existence of this connection between explanatory approaches and metaphysical categories is, nonetheless, very far from being widely acknowledged. Instead there is a long-standing tradition of supposing that psychology’s empirical methods for process identifying mean that its explanatory project is, and ought to be, entirely free from metaphysical commitments. An early statement of this supposition is found at the beginning of Wilhelm Wundt’s 1897 Outlines of Psychology. According to Wundt: The first, or metaphysical, definition [of psychology] belongs to a period of development that lasted longer in this science than in others. But it is here too forever left behind, since psychology has developed into an empirical discipline, operating with methods of its own; and since the ‘mental sciences’ have gained recognition as a great department of scientific investigation, distinct from the sphere of the natural sciences, and requiring as a general ground-work an independent psychology, free from all metaphysical theories. (1)

Most psychologists would regard the view that Wundt here states as an accurate description of current cognitive psychology, and of its relation

Consequences for Cognitive Psychology

119

to current metaphysics. They take psychology to have left metaphysical questions behind once it established an empirical method for its inquiries. We have already seen lots of examples showing this to be a mistake. Metaphysical commitments are not a special hazard of philosophical theorizing. They are incurred whenever the results of empirical investigations are pressed into explanatory service, and so they are much less easy to leave behind that Wundt’s remarks suggest. The examples that have illustrated this throughout the preceding discussion have been of various sorts. In chapter 1, §1.5, we focused on the example of employment. There are, as we saw, metaphysical reasons why the attempt to explain employment by producing a catalog of the processes that constitute it would be bound to fail. Employment is the wrong sort of thing to be explained by the identification of its constituent processes. The sense in which it is the ‘wrong sort of thing’ is a metaphysical sense: It belongs in the wrong metaphysical category. When we imagined theorists who produce and argue over catalogs of the employment-constituting processes, we could see that such theorists would be making a mistake, not because their methods and descriptions are not refined enough, or are too homogeneous, but because employment does not belong in the category of processes at all. The process catalogs that such theorists could produce could not possibly provide an explanation of what employment is. The example that we saw in chapter 2 was somewhat simpler, and so got us closer to the metaphysical heart of the matter. There we considered the case of haste. We saw that the attempt to explain haste by identifying the processes that constitute it would be bound to fail. Its inevitable failure would again be owing to a metaphysical mistake on the part of those who undertook it. The metaphysics of haste requires that we take something other than a process-identifying explanatory approach, not because haste is a socially constructed phenomenon, as employment is, but just because it is an adverbial one. Gilbert Ryle and Alan White provided two further examples, which we discussed at the end of chapter 1 and again at the beginning of chapter 4, that illustrate different versions of the same point. Ryle’s example showed that something would inevitably be left out by any attempt to explain migration by looking only at the processes by which acts of migration are accomplished. ‘Migration’ is, in Ryle’s phrase, a mongrel categorical: It provides hypothetical information, in addition to information about the processes that actually take place. White’s example was not mongrel categorical but was, instead, what he called ‘polymorphous’. He pointed out that we cannot give an explanation of practice by describing the particular processes that take place when a man practices playing the piano. We could begin to catalog the processes that constitute instances of practice, if for some reason we wanted to, but the identification of those processes would inevitably fall short as an

12 0

Attention Is Cognitive Unison

explanation of what practice is. The reason, again, is metaphysical: It is not the facts about which processes are taking place that make it the case that the man is practicing. All these examples show that there are phenomena that, on account of their metaphysical bases, resist explanation by the identification of their constituent processes. The variety among the examples suggests that there may be lots of these phenomena and that there may be various metaphysical bases that lead to such resistance. The fact that there are any such phenomena is, by itself, enough to show that when the process-identifying explanatory approach can be applied to a phenomenon, it is thanks to the metaphysics of that phenomenon making it apt for such an explanation. The project of process-identifying explanation depends upon this metaphysical aptness on the part of its explanandum. The empirical methods of process identifying, if these are adopted in the service of an explanatory end, therefore force one to take sides on a metaphysical question. They are not, pace Wundt, a way of avoiding metaphysics altogether. The Wundtian psychologist might respond to this line of thought by conceding that psychology’s process-identifying explanatory method carries a metaphysical commitment, but he might claim that the success of the method shows the commitment to be unproblematic: If the cognitive unison theory diagnoses a mistaken metaphysical assumption in cognitive psychology, then, it might be said, so much the worse for the cognitive unison theory. The process-identifying project in cognitive psychology of attention has several explanatory successes to its name. The theory that issues a pessimistic diagnosis of that project must be mistaken. The extent to which one finds this response persuasive—either as a response to the charge of metaphysical commitment carrying or as an objection to the cognitive unison theory—will depend on one’s estimation of psychology’s success in providing a process-first explanation. There are degrees of success that, were the psychologist able to claim them, would count as vindications of any metaphysical assumptions built into his explanatory approach. If, on looking into the brain, a process or set of processes could be identified whose operation accounts for all and only the psychological effects that are attributed to attention, then we could say that the explanatory achievements of process-identifying psychology vindicated its commitment to a process-first metaphysics. But nobody thinks that this is what psychology finds when it attempts to account for attention, or that it is what psychologists should expect to find. The consensus among psychologists has long been that Alan Allport was right when, in accounting for the collapse of Donald Broadbent’s explanatory project, he wrote, ‘There is no one uniform computational function, or mental operation (in general no one causal mechanism), to which all socalled attentional phenomena can be attributed’ (1992: 203). It is this

Consequences for Cognitive Psychology

121

realization that prompts psychologists to make their frequent remarks about ‘attention’ being ambiguous, or an ‘umbrella term’, or a family resemblance concept, or ill-defined—remarks that indicate a widespread realization that there is a metaphysical issue that process-citing explanation runs into, without putting their finger on what the source of that issue is. For the more modest and less unified kinds of explanatory success that process-identifying psychology can plausibly claim to have achieved the inference from success to metaphysical vindication is much less compelling. This is because the problem that the process-identifying mode of explanation encounters, when we try to apply it to a phenomenon that is metaphysically ill-suited for such an explanation, is not that the processes constituting the phenomenon in question cannot be identified. We can discover, by empirical means if we wish, that some instances of employment are constituted by processes of laboring, that some instances of haste are constituted by processes of tea making, that some instances of migration are constituted by southward flying, and that some instances of practice are constituted by piano playing. We can, if we want to, conduct empirical research that will enable us to add further members to these lists and to describe these constituent processes at a precise level, going right the way down to the fine grain of their neural implementation. None of this does us any explanatory good. The identities and details of these constituent processes by themselves remain explanatorily unrevealing. You do not learn what haste is by discovering that the instance of haste that took place in my kitchen this morning was constituted by processes of tea making, and you do not learn anything more about haste per se by watching a bit of hasty tea making in the brain scanner. But the fact that catalogs of constituent processes do no explanatory good does not mean that those catalogs cannot be compiled. No metaphysical commitment is incurred simply by compiling them. From this it follows that no metaphysical commitment can be vindicated by one’s success in doing so. Since it is when we attempt to put process-identifying research to explanatory work that metaphysical commitments are incurred, the psychologist who remains anxious to preserve the Wundtian ideal of a metaphysics-free methodology might attempt to disavow explanatory aspirations, claiming instead that accurate description is her research’s only goal. Such a claim could not be maintained for long. Psychology depends at every stage on nondeductive modes of inference. Such inferences require that claims with explanatory import are at stake. A claim with no explanatory aspirations cannot be supported by an inference to the best explanation. Nor, plausibly, can enumerative induction be a source of reasons to believe a theory unless something is assumed about the explanation of the class of phenomena over which the induction generalizes (Harman 1965). If the desire to be free of metaphysics led

12 2

Attention Is Cognitive Unison

the psychologist to abandon explanation altogether it would leave her with a methodology that was inferentially constrained to the point of fruitlessness.

6.3 THE DIVERSE EXPLANATORY CONSTRUALS OF CURRENT PSYCHOLOGICAL RESULTS We have seen that there are various metaphysical bases on account of which a phenomenon may resist process-first explanation. And we have said that, because of this, the process-identifying project in psychology carries a commitment concerning the metaphysical bases of the phenomena that it seeks to explain. This commitment does not restrict the process-identifying psychologist to just one metaphysical position. There is variety among the metaphysical bases of the phenomena that are amenable to process-first explanation, just as there is variety among the metaphysical bases of phenomena that resist it. Process-identifying psychology brings with it a commitment to one or another version of the process-first view of attention’s metaphysics, but it does not force psychologists to settle on any one version of that view in particular. Within the range allowed by their shared allegiance to the processfirst camp psychologists are no strangers to metaphysical disagreement. Thanks to the popularity of the Wundtian idea that empirical methods make for a metaphysics-free psychology these metaphysical disagreements are often not discussed, and are often not recognized as being metaphysical. Sometimes they are not recognized at all. They are genuine disagreements, however, and are relatively easy to discern once we look in the right places. As with the more fundamental division between process-first metaphysics and adverbialist metaphysics, these smaller disagreements between different psychologists’ versions of the processfirst approach manifest themselves in their differing understandings of the way in which attention’s explanation gets put together. We can bring to light these metaphysical differences by using a particular case to illustrate the different explanatory construals that different theorists place on research that identifies an attention constituting process. Elliot Freeman’s work on the role in attention of ‘integrative horizontal connections’ in early visual cortex provides a clear example. Freeman’s research shows, using a combination of behavioral and electroencephalographic studies, that these early integrative processing systems play a role in implementing certain instances of attention that determine the way in which patterns of collinearity are used when parsing the basic contours of a visual scene (Freeman, Sagi, and Driver 2001; Freeman 2005; Khoe, Freeman, Woldorff and Mangun. 2006). How is such a finding to be taken as a contribution to the explanation of attention? One construal of Freeman’s discovery is as the identification of one particular species in the overarching genus of attention processes.

Consequences for Cognitive Psychology

12 3

Another is as the discovery of a particular effect that the central attention mechanisms can produce. A third construal takes Freeman to have identified a new member of the loosely united family of attention processes. A fourth takes the same discovery to be a contribution to the more or less heterogeneous scientific theory with which our mistaken commonsense talk about ‘attention’ should be replaced. A fifth takes the discovered process to be a new component of attention—a piece of the jigsaw that will give us a picture of attention itself only when all the pieces are viewed together (and viewed in the right configuration). A sixth construal takes Freeman’s discovery to be the discovery that the attentiondefining principles—which might be taken to be functional principles, or might be taken to be more abstract mathematical principles—are instantiated at a particular locus, early in the visual processing stream. This list of possible construals is by no means exhaustive, but it is long enough for us to see the point: Each of these different construals of the explanatory import of Freeman’s identification of an attentionconstituting process takes a slightly different view of the metaphysical relationship between attention and its underlying processes. Each therefore has a slightly different understanding of attention’s metaphysical status. Some of these construals will seem natural to some psychologists, and others will seem natural to others. A psychologist like John Duncan, who advocates a family resemblance view of the processes that constitute attention, will favor the interpretation of Freeman’s results as the identification of one member in an open-ended, potentially quite heterogeneous family (Duncan 2006). Psychologists like Claus Bundesen and Thomas Habekost, who advocate the idea that there is a distinctive set of mathematical principles governing various different instances of attention at various different levels of neural organization, will favor an interpretation of Freeman’s results as the discovery of a new instance in which these principles operate (Bundesen and Habekost 2008). A psychologist who thinks, as John K. Tsotsos, Laurent Itti, and Geraint Rees (2005: xxx) do, that there is a particular functional role that all instances of attention play will think that Freeman has described one of the processes that occupy that functional role. A psychologist who thinks, as Eric Knudsen (2007) does, that attention is a complex process built out of several simpler components will favor an interpretation of Freeman’s work as giving us an account of one of these components. And so on. Psychologists are forever negotiating these reconstruals of the explanatory importance of their work without explicitly debating the metaphysical disagreements that underpin them. They are more or less silent about these negotiations since it frequently does not matter, for the normal scientific business of designing and assessing experiments that identify the processes constituting various instances of attention, whether one takes the processes that one thereby identifies to be species in the attention genus, or determinates of the attention determinable, or implementers of the attention function, or instantiators of abstract attention principles,

124

Attention Is Cognitive Unison

or members of the attention family, or components of the attention complex, or bases for the emergence of attention, or something else entirely. Such metaphysical construals typically do not matter for the design of one’s experiments. They start to matter when, but only when, one needs to understand the explanatory import of one’s results, and the relation of those results to the results and claims of others. Since the cognitive unison theory rejects the idea that the identity of the attention constituting processes plays a role in the explanation of what attention is, the explanatory construal that it advocates for a piece of research like Freeman’s is a somewhat deflationary one. There is nothing that goes contrary to any established orthodoxy in the fact that the cognitive unison theory requires us to make such a reconstrual because, as far as the metaphysics of attention goes, there is no established orthodoxy. Explanatory reconstruals are perfectly commonplace. The particular construal that the cognitive unison theory makes of a discovery like Freeman’s is as the identification of one of the processes that can be a partial constitutor of attention under certain task conditions.

6.4 REASONS FOR DEFLATION The cognitive unison theory’s construal of a discovery such as Freeman’s breaks from established orthodoxy because it allows that this discovery might be taken in a rather deflationary way. It does so on account of two features that are found in all of the examples that we have seen of phenomena that resist process-identifying explanation. The first feature, on which philosophers have traditionally focused, is multiple realizability: Instances of employment can be constituted by manual labor, vigilant observation, or administrative organization; instances of haste can be constituted by tea makings, breakfasts, marriages, or the reaching of conclusions; instances of migration can be done by land, by air, or by sea; instances of practice can involve pianos, unicycles, or vocal recitations. These phenomena are not realized by multiple processes merely in the way that jade is realized by two distinct substances (see §2.4). Their multiple realizability is open-ended: The catalogs of possible haste constitutors, possible employment constitutors, and so on, are catalogs that could never be definitively completed. This feature can lend a deflationary tone to the adverbialist’s construal of the explanatory import of process-identifying research because it entails that, while a discovery such a Freeman’s may tell us something about the explanation of certain instances of attention, there may be any number of other instances about which it tells us nothing. The second feature of adverbial phenomena that might give our construal of Freeman’s discovery a deflationary tone is the susceptibility of adverbial phenomena to the kind of supervenience failure that drove

Consequences for Cognitive Psychology

12 5

our argument in chapter 3. Multiple realizability prevents us from making inferences that move from premises telling us which phenomenon is being instantiated to conclusions telling us about the processes that constitute the instance of the phenomenon in question. With adverbial phenomena there is also a bar on inferences going the other way. ‘He was practicing’ does not imply ‘he was playing the piano’, since he may have been practicing any number of other things, but neither does ‘he was playing the piano’ imply ‘he was practicing’, since he may have been playing the piano for any number of other reasons. Nor, more generally, does ‘he was performing the same process as was performed by a man who was practicing’ imply that he was practicing himself. This supervenience failure lends a further deflationary note to our construal of the explanatory importance of process-identifying research such as Freeman’s because (as we saw in chapter 2) it entails that the processes described by a theorist such a Freeman qualify as an instances of attention only on account of something other than the facts about which processes they are. In combination, these two features of adverbial phenomena make room for something deflationary to be going on when the cognitive unison theory construes research such as Freeman’s as telling us, not about species, or components, or determinants of attention, but about the processes that can constitute attention, or partially constitute it, in certain task conditions. The extent to which this construal turns out to be deflationary will depend on the fineness of grain with which the ‘certain task conditions’ need to be specified. If they need to be specified quite precisely then it may be that, in slightly different contexts, attention could be constituted by quite different processes, and it may be that, in slightly different contexts, the processes identified by Freeman could take place without constituting attention. If, on the other hand, the conditions in which Freeman’s processes are attention constitutors can be specified more loosely, then the range of cases that his work tells us about will be broader, and their explanatory import will be greater. In the most deflationary scenario the processes that Freeman identifies might not have any role in attention except when subjects are engaged in quite particular and peculiar high-speed visual discrimination tasks (just as the processes of tea-bag flinging might not have any role in haste, except when subjects are engaged in operations of breakfast preparing). In the least deflationary scenario, those processes might have a role in attention whenever we pay attention to a fresh scene. The cognitive unison theory need not lead us to advocate any very strong form of deflationism. The point it leads us to is that our understanding of how explanatorily important a discovery such as Freeman’s is depends on our understanding of attention’s metaphysics. Getting clear on what such results tell us is not simply a matter of getting their methodology straight.

126

Attention Is Cognitive Unison

When the cognitive unison theory raises the possibility that experiments like Freeman’s may not have much to tell us about the cases of attention that are actually out there, the problem raised needs a metaphysical answer, not simply a methodological refinement. The problem is not a complaint about ecological validity. Consider, as an analogy that illustrates this, a metaphysically misguided musicologist who is attempting to get explanatory purchase on unison by making observations of the processes that constitute it. This musicologist might find that unison can reliably be created in the lab by getting groups of undergraduates to sing ‘Happy Birthday to You’. He might then conduct perfectly accurate and well-controlled experiments that enable him to describe in detail the breathing and vocalization processes that constitute such singing. The problem with this musicologist’s research would not be that his experiments are low on ecologically validity. In fact, as far as ecological validity goes, his experiments are rather good: There are lots of wild-type cases of unison that are just like the ones he studies in the lab. The imagined musicologist’s research can lead to genuine discoveries about the processes underpinning common instances of unison. But the metaphysics of unison nonetheless requires us to make an explanatorily deflationary construal of those discoveries if they are intended to cast explanatory light on unison per se. The facts discovered by our imagined musicologist tell us about some processes that can be constitutors of unison under certain task conditions, but those facts, being inessential to an explanation of what unison is, are not be facts that tell us about unison as such. Similarly, the cognitive unison theory’s construal of process-identifying research into attention does not involve denying that such research makes genuine discoveries about the attention-constituting processes. Nor does it find fault with the methods, or the ecological validity of such research. It may be that lots of wild-type cases of attention involve the very processes that we study in the lab. But it may still be that the metaphysics of attention requires us to make a somewhat deflationary construal of the findings that we make when we examine those processes. Interpreting a piece of empirical research—even a piece of research that is as methodologically impeccable as Freeman’s—depends on our understanding of attention’s metaphysics. When we are dealing with a phenomenon that has the metaphysics of a process it is good scientific tactics to start out by describing the processes constituting the simplest cases of the phenomenon we are interested in, in the expectation that what we thereby learn about will be a common core that other instances of the phenomenon share. When we are dealing with an adverbial phenomenon we need independent reasons for thinking that this tactic will be an informative one. It is possible that features of the simplest instances of an adverbial phenomenon will be shared by its more complex instances, but we need some additional reason to suppose that this is so. It is equally possible that a description of

Consequences for Cognitive Psychology

12 7

the processes constituting the phenomenon’s simplest cases will tell us only about features of those simple cases, and explanatorily inessential features at that.

6.5 INDUCTIVELY UNRELIABLE PROPERTIES Locating the cognitive unison theory’s scientific consequences at the point of explanatory construal may make the metaphysical content of that theory sound less methodologically consequential than it really is. The possibility of an adverbial metaphysics for attention can matter to the scientific practice, not merely to the interpretation of that practice’s results. Since adverbialism about attention allows there to be many processes that can constitute attention, and since it allows that an instance of attention can be constituted by processes of the same sort as those that constitute an instance of inattention, no information about which processes are being instantiated can be anything more than an unreliable indicator of whether or not attention is being paid. In general, the multiple realizability and supervenience failure exhibited by adverbial phenomena mean that those properties that indicate which processes are taking place will not be good inductive properties of those phenomena. If attention has an adverbial metaphysics then one cannot make an inference from the fact that some instances of attention exhibit certain process properties to the conclusion that other instances of attention will exhibit them too. Nor can one infer from the fact that a particular case does not exhibit certain process properties to the conclusion that it is not a case that involves attention. Among the process properties that may fail to be good inductive properties of attention are the properties revealed by EEGs (electroencephalograms). EEGs indicate patterns of neural firing. Those patterns are indicators of which neural processes are taking place, and so, according to the adverbialist, they may only be unreliable indicators of whether attention is being paid. The fact that shifts of attention in one context reliably display a particular EEG pattern provides no reason to think that attention is not being paid in some other context where that EEG pattern is absent. It is therefore a mistake to conclude that attention is not required for the recognition of certain primed shapes on the basis of the finding that such priming effects are found in the absence of a particular EEG pattern that, in other contexts, is a sign of attention shifting. (An instance of this problematic form of reasoning can be found in Jaśkowski, van der Lubbe, Schlotterbeck and Verleger’s 2002 paper on attention’s role in unconscious priming.) Other poor inductive properties of attention include the properties of timescale. Since different processes may constitute attention on different occasions, and since these different processes will operate over

12 8

Attention Is Cognitive Unison

differing timescales, there will be no simple inductive way to establish a duration such that any difference that could be made by attention must have been made in that time. One cannot infer that all consequences of attention must become manifest within 200 milliseconds from the fact that some forms of attention have effects that are manifested within 200 milliseconds. This raises a methodological problem for research such as that conducted by Luis Pessoa and his collaborators when examining the question of whether it is possible to recognize the emotions expressed by faces to which one is not paying attention (Pessoa, McKenna, Gutierrez, and Ungerleider 2002). Pessoa and collaborators claim that the emotional expression of a face cannot be detected without attention. Their grounds for this claim are that, when emotional faces are presented for 200 milliseconds, concurrently with pairs of geometric shapes, the amygdala makes different responses to different emotional expressions only when attention is being paid to a task involving the faces, not the shapes. But the fact that attention is necessary for differential responding to faces that are presented for a fifth of a second only implies that attention is in general necessary for such expressions to be detected if it is assumed that differences between attention and inattention are always manifested within a fifth of a second. To the adverbialist this assumption looks quite unwarranted. Even if some attention-constituting processes always run to completion within a fifth of a second, there is no reason to suppose that all such processes do so.

6.6 QUESTIONS WITHOUT ANSWERS We have allowed that some matters of experimental design may be metaphysics free, even while the explanatory interpretation of the results of those experiments is metaphysically laden, but it is also true that metaphysical commitments can be incurred by the sorts of questions that our experiments are looking to answer, and so that they are sometimes incurred before the question of experimental design even arises. If attention has an adverbial metaphysics then it is not another activity in the brain in addition to the cognitive processes that constitute it, much as unison is not another activity in addition to the playing of the individual members in the orchestra, and much as, to adopt a favorite example of Ryle’s, the university is not another thing in addition to, and on the same footing as, the libraries, colleges, and laboratories (1949: 18). Since attention is not on the same metaphysical footing as the processes that constitute it, it does not compete with those processes for the occupancy of neural processing capacity or for loci in the brain, any more than the university competes with its own laboratories for real estate. This point has consequences for the ongoing debate about the function of the lateral intraparietal cortex (the LIP). The place in the architecture of the brain occupied by the LIP suggests that this area might have a role

Consequences for Cognitive Psychology

12 9

in some of the more complex cognitive functions. It receives input from several places, including higher level sensory areas, and it has outputs to areas that include those that are usually thought of as having a more or less direct role in producing planned behavior, particularly in the deliberate directing of the eyes. The conditions in which the LIP responds are not well understood. We do know that it responds to salient stimuli and that it responds just before deliberate saccadic eye movements are made. It was therefore natural for neuropsychologists to hypothesize that this part of the brain has some role in something like the planning of intentional movements, particularly in the planning of eye movements. The evidence supporting this hypothesis comes from several sources, much of it owing to research by Richard Andersen and his collaborators at Caltech (for a review, see Andersen and Buneo, 2002). Andersen’s interpretation of the LIP’s function is controversial. In the work of Michael Goldberg and colleagues we find a rival theory, according to which the LIP is involved, not in planning a movement toward a certain part of space, but in attention to that part of space. Thus Goldberg, Bisley, Powell, Gottlieb and Kusunoki (2002) tell us that ‘neuronal activity in the monkey intraparietal area has the properties appropriate for a neuronal substrate of attention: instead of all objects being represented in the parietal cortex, only salient objects are’ (205). It is easy to see how this debate arises: Andersen and colleagues observe activity in LIP just before a monkey moves its eyes toward a part of space. They conclude that this activity corresponds to the planning of the action. Goldberg and colleagues point out that there are several things that a monkey might be doing in the moments before it moves its eyes toward a part of space: It will be planning the relevant eye movements, but it will also be paying attention to the part of space in question. They interpret the activity observed in LIP as corresponding to the fact that the relevant part of space comes to be attended. It has proved extremely difficult to find a test to distinguish between the two rival interpretations of LIP activity. Goldberg et al. claim that Andersen et al.’s hypothesis (that the area is involved in the planning of eye movements) fails to account for the data: Although neurons in LIP have been implicated as being directly involved in the generation of saccadic eye movements, their activity does not predict where, when, or if a saccade will occur. The ensemble of activity in LIP, however, does accurately describe the locus of attention. (2002: 205)

Andersen et al., on the other hand, find Goldberg et al.’s interpretation of the data to be inadequate. Snyder, Batista, and Andersen’s 2000 paper claims that their experiments ‘show that there are intention and sensory related activities in the PPC [the area of which the LIP is a subregion] consistent with its proposed role in sensory-motor transformations.’ And

13 0

Attention Is Cognitive Unison

they chide their opponents for assuming carelessly that these activities are involved in attention, writing that: ‘These studies also show that care must be taken to measure intention-related signals and not assume that all task dependent modulation in the PPC reflects attention’ (1433). It is relevant that both sides in this debate are committed to the process-first theory of attention’s metaphysics. Goldberg’s team makes this commitment more or less explicit when they write that ‘the brain cannot monitor or react toward the entire world at a given time. Instead, using the process of attention, it selects objects in the world for further analysis.’ (2002: 205) Andersen’s group display the same process-first allegiance, although somewhat less overtly, when they describe the parietal cortex as ‘important for attention, target selection, coordinate transformations and other integrative functions that are closely linked to perceptual processes’ (Snyder, Batista, and Andersen 2000: 1440). This is not an explicit statement of the process-first view, but it does reveal a commitment to it. When we are told that the parietal cortex has a role in several ‘integrative functions’ and attention, it is (as Ryle would say) as if we have been told that the university has departments of art, architecture, medicine, zoology, and research and teaching, or that the violinist was playing her instrument, reading the music, and practicing. The research and teaching are not another activity of the university in addition to, and on the same footing as, the activities performed by the departments of art, architecture, and so on. The practicing is not another activity of the violinist in addition to the playing and the reading of the music. According to the cognitive unison theory it is the same with attention. Attention is not another activity in the brain in addition to, and on the same footing as, the processes of ‘target selection, coordinate transformations, and other integrative functions that are closely linked to perceptual processes’ (1440). Because they take a process-first view of attention’s metaphysics, Andersen’s team take it that if the LIP is involved in attention and in the planning of actions then it must be doing two things at once, and doing these two things with two different ‘components’ (Snyder, Batista, and Andersen 2000: 1433). The adverbialist sees this as a mistake. The planning of action can be a constituent of attention, and so for the LIP to be involved in attention and action planning it need not have two different components, one responsible for the attention and the other for the planning, any more than the violinist needs to divide herself into components in order to be both practicing and playing the violin. It is their shared commitment to the process-first view that leads Andersen, Goldberg, and their colleagues to think that their views are exclusive of one another, and so to think that when the LIP is shown to have a role in attention it is thereby shown not to have a role in motor planning, and vice versa. According to the cognitive unison theory the LIP may be involved in motor planning, and that activity may, if it is a bit

Consequences for Cognitive Psychology

131

of motor planning that operates in unison with other cognitive processes serving the same task, partially constitute the subject’s attention to a particular part of space. The attention is not another activity in addition to the planning, and so it does not give us a competing account of the work done by this part of cortex. Goldberg et al.’s account of LIP need not be in conflict with Andersen et al.’s. To say that the two theories need not be in conflict is not, of course, to say that they are supportive of one another, or that evidence that supports either one might not make the other less plausible. It is possible that only Andersen et al. or only Goldberg et al. are correct. There may be processes involved in bringing about attention that are not involved in the planning of action. There may be processes involved in the planning of actions that are not involved in the bringing about of attention. It is not that there is nothing for Goldberg et al. and Andersen et al. to disagree about. There are competing interpretations of LIP activity that differ in whether they see it as doing processing for attention or processing for intention. Goldberg et al. and Andersen et al.’s confusion is the methodological one of conducting their debate as if it were the case that, insofar as a part of the brain is found to be responsible for motor planning, it is thereby found not to have a role in attention, and insofar as a part of the brain is found to be responsible for attention, it is thereby found not to have a role in motor planning. It is this that leads them to make the mistake of thinking that the LIP must do attention, or do motor planning, or else that, if it does both, it must have distinct components dedicated to each. That understanding of the space of possibilities is a metaphysically laden one, and the adverbialist suggests that its metaphysical commitments are mistaken. The debate between Goldberg et al. and Andersen et al. illustrates adverbialism’s consequences concerning the sorts of questions that it makes sense to ask. Traditional debates about the locus of attention can be seen in the same light, as can more recent attempts to approach the questions on which those debates focused. It is common to assume that the selectivity that we have taken to be essential to attention is selectivity that results from limitations in information-processing capacity. This assumption is more or less universal, but Emanuela Bricolo and her collaborators make it unusually explicit when they write: ‘In general, set-size effects are taken to indicate that processing of the array of elements depends on limited capacity resources, that is, it involves attention’ (Bricolo, Gianesini, Fanini, Bundesen, and Chelazzi 2002: 980, emphasis added). When it is assumed that attention’s selectivity is owing to capacity limitations, it makes sense to ask if these capacity limitations occur at one or at many loci, and to ask about where these loci are. In the debates about the locus of attention that were prompted by Broadbent’s work in the middle of the twentieth century, psychologists frequently assumed that there was just one

13 2

Attention Is Cognitive Unison

such bottleneck in information capacity. We no longer think that it was reasonable to expect there to be just one locus for the attentional bottleneck, but many psychologists continue to think that questions about the location of the attentional bottleneck can still be intelligibly asked, as long as we specify the particular variety or instance of attention that we have in mind. Such a view is particularly prominent in the work of Nilli Lavie. In her 1994 work with Yehoshua Tsal she, like Bricolo et al., makes the bottleneck assumption explicit, writing, ‘Obviously, the need for selection arises because of some processing limitation’ (184). Lavie and Tsal go on to propose a theory (elaborated further in Lavie 1995, 2001) according to which there is a particular capacity limitation that explains the selectivity of attention, but according to which the locus at which this bottleneck operates moves, depending on the perceptual demands of the stimuli one is presented with. John Reynolds, Jacqueline Gottlieb, and Sabine Kastner take a similar stance, albeit on a somewhat different theoretical basis, in their 2002 review of the attention literature. They reject the idea that we should expect a single answer when they ask, ‘Where is the computational bottleneck as revealed by search tasks?’ but they retain the idea that specific loci of attention-related capacity limitations may be identified, one of them at a stage of processing when working memory is allocated, and one at an earlier stage, where feature integration is accomplished (1264–6). If attention is cognitive unison then our inability to return a single answer to Reynolds, Gottlieb and Kastner’s question has a different origin. Attempts to answer that question lead to conflicting results, not because the question has different answers on different occasions, but because it has no answer at all. According to the unison theory, the selectivity of attention need not have anything to do with a limitation in processing capacity. Unison does not owe its selectivity to capacity limitations: The orchestra can only be playing one tune in unison at any time, but this is not because there are insufficient players, or because any one section of the orchestra is less competent than the others. If attention is cognitive unison, then the attentive performance of one task will place limitations on the extent to which other tasks can be attended (as outlined in §4.11). It will do so whatever the extent of an agent’s information-processing capacities. There need not be any bottleneck to explain unison’s selectivity, and so the question of where that bottleneck occurs, even when asked of a specific case, may again be a question without an answer.

6.7 THE POSITIVE PAYOFF I said, in introducing this chapter, that the results of reconstruing cognitive psychology in the ways that the cognitive unison theory requires would not be wholly negative. One place where the cognitive unison theory’s interpretation of the empirical results is not negative at all is in

Consequences for Cognitive Psychology

133

its handling of the growing body of empirical work that takes attention to be implemented by mechanisms of ‘biased competition’. Reynolds and Duncan, together with Bob Desimone have been responsible for developing this biased competition model of attention (Desimone and Duncan 1995; Reynolds and Desimone 2000). Reynolds and Desimone summarize the central claims of the model like this: Multiple stimuli appearing together activate populations of neurons that compete with one another, possibly through mutual inhibition. When attention is directed to one of the stimuli, this causes an attentional feedback signal to be directed to the neural population activated by the attended stimulus. This feedback biases the competition in favor of the attended stimulus, enabling it to propagate its signal forward to the next cortical area. (Reynolds and Desimone 2000: 234)

This model of an attention-producing mechanism comports nicely with the cognitive unison theory’s account of what attention is. As with the selectivity of unison, the selectivity of a competition is not owing to limitations in the competing processes: There is no need to postulate any limitations in the number or in the ability of the competitors in order to explain the fact that a competition has just one winner. An architecture in which different processes compete for the control of processing resources at various levels throughout the nervous system is precisely the sort of architecture in which states of unison could emerge, and could do so without there being any one attention-explaining process to bring these states about, or any one bottleneck in capacity through which all and only the attended things have passed. The cognitive unison theory’s construal of research within the biased competition school is therefore a positive one. The theory does not give the biased competition school exclusive rights to the modeling of attention. There may be several different routes by which states of cognitive unison can be brought about and maintained. But the research from the biased competition school has, at least, given us a detailed account of the anatomy and mechanism of one of these routes, and has provided an abstract model that explains and predicts many of the attention-related effects that are revealed by neural and behavioral observations of humans and nonhuman primates. If the cognitive unison theory gives us the correct account of what attention is, then the biased competition model may give us the correct account of how many of the instances of this attention-realizing unison come about, and of how they get maintained. Notice, however, that in saying this the cognitive unison theory is once again proposing a particular explanatory construal of a body of scientific research for which no existing construal is established in the psychological literature as being the orthodoxy. Just as different psychologists, all of whom take Freeman to have made a significant discovery in his work on the forms of attention that have their basis in early visual cortex, might nonetheless differ widely in their construals of

13 4

Attention Is Cognitive Unison

that discovery’s explanatory import, so different theorists, all of whom endorse the biased competition model, might nonetheless differ widely in their construal of the way in which attention is modeled by it. Not all of the psychologists who work with the biased competition model would accept the cognitive unison theory’s construal of the model’s explanatory status. To see this last point compare Reynolds and Desimone’s presentation of the biased competition model, given above, with the following passage from the abstract of a paper by Sabine Kastner and Leslie Ungerleider: Functional brain imaging studies reveal that biasing signals due to selective attention can modulate neural activity in visual cortex [. . .]. Although the competition among stimuli for representation is ultimately resolved within visual cortex, the source of top-down biasing signals likely derives from a distributed network of areas in frontal and parietal cortex. Attention-related activity in frontal and parietal areas does not reflect attentional modulation of visually evoked responses, but rather the attentional operations themselves. (2001: 1263)

Here and elsewhere Kastner and Ungerleider take their project to be that of identifying specific ‘mechanisms of visual attention in the human cortex’ (Kastner and Ungerleider 2000), and they claim to have localized such mechanisms in a network of frontal and parietal cortical loci. The activity observed in these areas, they say, ‘reflects the attentional operations themselves.’ The competitive activity in visual cortex they take to be under the influence of a biasing signal that these attentional operations produce. According to this way of construing the biased competition model there are two different metaphysical relations in play. The competitive activity in visual cortex is influenced by attention. The activity in frontal and parietal cortex ‘reflects the attentional operations themselves’. This is not how the cognitive unison theory sees things, nor is it what Reynolds and Desimone (2000) suggest. For Reynolds and Desimone the competitions in visual cortex are not separate from yet influenced by attention. They, too, are attention constitutors. Correspondingly, Reynolds and Desimone’s construal of the biased competition model does not say that the activities in frontal and parietal cortex have any special title to the claim of being ‘the attentional operations themselves’; they are, at most, a proper part of those operations. Their role in the model is as the source of an ‘attentional feedback signal’, but that signal is caused by the fact that attention is already paid to the relevant stimulus, on account of its having won a competition within visual cortex for cortical control. Again, then, we find that the empirical literature contains different explanatory construals of the same work, although this time the work is that of model building, rather than that of process identifying. Again these different construals depend on different pictures of the metaphysical relation between attention and the processes that the empirical work

Consequences for Cognitive Psychology

135

identifies and describes. The cognitive unison theory recommends a construal of the biased competition model that is in no way deflationary. It rejects the construal according to which the competitive mechanisms that operate throughout the visual cortex are merely affected by a signal originating from an as yet poorly understood network of attention processes taking place in the frontal and parietal cortex. Instead, it endorses the view according to which the mechanisms responsible for bringing about instances of attention are, in at least some cases, the wellunderstood competitive mechanisms that are implemented throughout the visual cortex and elsewhere. Competition (especially when accompanied by a biasing feedback signal) is a way to bring about unison across a hierarchically organized network competing processes. It is these processes, when operating in unison, that constitute attention, not only the top-down processes that generate the signal by which they are biased.

7 Philosophical Work for the Theory of Attention

7.1 PUTTING ATTENTION TO PHILOSOPHICAL WORK The two deepest mysteries of the mind are the mystery of consciousness and the mystery of content. The first of these mysteries—the so-called ‘hard problem’—arises from our failure to see how to give an intellectually satisfactory account of the fact that the mental states of creatures like us are experienced by their subjects. The second mystery—‘Brentano’s problem’—arises from our failure to see how to produce an intellectually satisfactory answer to the question of how any naturally arising state could have a property that our beliefs, desires, hopes, and inquiries do have: the property of being essentially about actual or possible things, propositions, or states of affairs, rather than merely encoding information derived from those things, propositions, or states of affairs. Many psychologists and some philosophers think that the theory of attention will contribute to solving the first of these mysteries. A smaller group, this one composed mostly of philosophers, thinks that the theory of attention will contribute to solving the second. There is nothing mutually exclusive about these two positions: One might think that the theory of attention will contribute to the solution of both mysteries, either because one thinks that the two mysteries are at bottom the same or because one thinks that the theory of attention will have things to say about each. My own view is that the two mysteries are indeed intimately connected but that the theory of attention enables us to solve neither of them. Whatever light the theory casts, it is a light in which we see more clearly what the mysteries of the mind are, and see that there is no easy way by which they can be avoided.

7.2 ATTENTION AND REFERENCE

7.2.1 The Nature of the Mystery Thought’s contentfulness is so frustratingly mysterious that there is not even agreement about what exactly the mystery is. It is therefore worth 136

Philosophical Work for the Theory of Attention

137

taking a moment to give an account of what I take it to be and to say something preliminary about what I take its solution to require. The mystery originates from the fact that mental states, at least in cardinal cases, are states that have contents, and that have these contents essentially. Consider, as an example, my belief that Belfast is farther east than Dublin. That, obviously enough, is a mental state that has as its content the proposition that Belfast is farther east than Dublin. The content is essential to the belief in the sense that, if the belief had not had that content, it would have been a different belief, and if it had not had any content, it would not have been a belief at all. Examples could easily be multiplied. The desire to be in Dublin is a mental state that has as its content one’s being in Dublin. If it had not had that content, the state would have been a different desire. The same goes for hopes, fears, regrets, doubts, perceptual seemings, and all the rest. All have contents, and all have their contents essentially. There is no thought that is not a thought of things being some way or another, no state of perceptual seeming that fails to present things as being some way or another, and no desire except the desire that something or other should be the case. Whether the character of our mental states is exhausted by their contents is contentious (Block 1998; Tye 2002), just as it is contentious what sorts of entities these contents are (Bealer 1998; Speaks 2005), contentious in what format they are represented (Lycan 1996; Stalnaker 1996; Brewer 2005; Byrne 2005), and contentious also whether every mental state is a state that essentially has content (Siegel 2010). However these issues should be settled, it is at least clear that contentful mental states exist, and that they play a central psychological role, so that every thinker and, more generally, every subject of psychological states, is the subject of at least some states that have the property of being essentially contentful. This essential contentfulness creates a mystery because we lack a satisfactory account of how it can be accommodated within a naturalistically respectable theory of the mind. In the absence of such an account there are several phenomena for which we lack an explanation. We are, for example, unable to say, given mental states with particular contents, what it is that grounds the fact that it is those contents, and not any others, that the states in question have. To some ears this may not sound like a grand mystery on a par with the mystery of consciousness. Psychology, after all, is full of theories that tell us what happens when a thinker is presented with an object and comes to be in neural states that represent that object. We know a lot about the way in which the properties of a well-lit red ball, for example, can make a difference to the thinker’s sensory apparatus. We know a lot about the ways in which these sensory differences lead to various information-bearing states in the thinker’s brain. And we know that it is these states that underpin the mental states that have the property of being about the ball. All of this seems to add up to an explanation of

13 8

Attention Is Cognitive Unison

the contentful state that is the thinker’s seeing that there is a ball. But knowing all of this does not go very far toward accounting for that state’s having the property of being about the ball. It remains mysterious how it gets to be the case that, when a perceptual encounter with an object leaves information traces in the thinker’s brain, there come to be mental states that have that object as their contents. The point that remains mysterious can be brought into focus by considering psychological states other than perceptual seemings. It is true, perhaps even necessary, that in typical cases of perceptual seeming there is an overlap between the content of the psychological state and the information carried by the systems that underpin that state. An account of the information handling that underpins states of perceptual seeming can therefore look as if it were an explanation of the content of those states. But for states other than perceptual seemings the content of the psychological state and the information carried by its underpinnings do not overlap. The psychological state that you are in when you wonder whether Belfast really is to the east of Dublin (let us suppose that you are not sure) is a mental state that has as its contents Belfast being to the east of Dublin, or some subjunctive inflection thereof. But it is hugely unlikely that the neural underpinnings of that state carry any information at all about Belfast being to the east of Dublin. If they did then geographical questions could be settled with brain scanners. An explanation of how it is that that psychological state comes to have that proposition as its contents therefore requires more than is given to us by a theory of the information encoded by its neural underpinnings. The point here is not that the encoding of information is irrelevant to the explanation of content. It may be, for all anyone knows, that the explanation of contentfulness can be built from components in which information encoding is all that is talked about. The point that I am insisting upon is just that building such an explanation is a complex task, and although various suggestions have been canvassed, nobody is very sure how to go about completing it, or whether it can be completed. The contentfulness of thought remains mysterious, even if it is plausible that the relatively unmysterious phenomenon of information encoding must figure somewhere in the theory that we are looking for. If you are sympathetic to the idea that the mysteries of content are not independent from the mysteries associated with consciousness then you may also be sympathetic to the thought that, even in the case of perceptual seeming, a theory of the underlying informationbearing states cannot do all of the explanatory work that a theory of contentfulness will need to do. Consider again (although the example is a somewhat controversial one) the states that my brain goes into on those occasions when it perceptually seems to me that there is a red ball in front of me. The underlying brain states carry information about the ratios of the wavelengths of light that such balls reflect. But wavelengths per se may not figure at all in the contents of my mental

Philosophical Work for the Theory of Attention

139

states. I may be seeing a red ball, and seeing that it is red, and what I am seeing may tell me nothing about wavelength ratios per se unless I know a thing or two about the physics of light. In order to get into mental states that have wavelength ratios per se as their contents I need to do something other than just getting into a neural state that carries information about those wavelength ratios. It therefore seems that, even for the state of it perceptually seeming to me that a ball is red, an explanation of the fact that this state has the content ‘that ball is red’ (but does not have the content ‘that ball typically reflects light in such-and-such wavelength ratios’) will be an explanation that needs to mention something more than could be accounted for simply by reference to the encoding of information by that state’s neural underpinnings. The metaphysics of color, and the character of color experience, are far too controversial for these considerations to show decisively that the contentfulness, even of perceptual seemings, is inexplicable in information processing terms (for a discussion of these controversies, see Stroud 2000). But these considerations do enable us to see something of the mystery that contentfulness creates.

7.2.2 A Role for Attention? An example taken from one of Wittgenstein’s notebooks shows why it is that the mystery of contentfulness is one that our theory of attention might be expected to help with. The example is a simple one: I point with my hand and say ‘Come here’. A asks ‘Did you mean me?’ I say: ‘no, B’.—What went on when I meant B (since my pointing left it in doubt which I meant)? (1967: 21)

The case that Wittgenstein asks us to consider seems perfectly possible. The content of the thought expressed when the speaker says ‘come here’ can be a content in which B figures, rather than A, even when the speaker’s gesture leaves it in doubt which of them was meant. To make sense of that possibility we need to find something in the imagined scenario that relates the speaker to B and not to A. Wittgenstein’s case is constructed so that some of the things that elsewhere have a role to play in fixing the contents of thought cannot possibly be doing that work here. The particular words that were uttered, for example, have no established connection to B rather than A. They are just the words ‘come here’ and could have applied to either person. It is therefore of no use trying to fix the thought onto B by evoking, as Kripke (1982) and Brandom (1994) do for other sorts of cases, the communitywide norms for the use of the particular linguistic expressions that were employed. Nor does it help if we bring the speaker’s pointing gesture into the account. Pointing gestures may fix the referent in some cases of demonstrative use (Kaplan 1989). But it is part of Wittgenstein’s story

14 0

Attention Is Cognitive Unison

that the only gesture made in the case we are imagining was too vague to give us a way of determining which person was meant. It therefore seems that, in order for Wittgenstein’s scenario to be possible, there must be something else relating the speaker to B rather than to A, and it seems, although Wittgenstein himself suspects that this may be the root of our confusion, that this thing, whatever it is, must be an internal psychological event and, moreover, that it must be an event that, since the speaker knew immediately on A’s asking which person it was that he had meant, manifested itself immediately in the speaker’s experience. A natural thought is that this postulated psychological event that picks out person B is known immediately by the speaker because it is an intentional act that the speaker himself performs. With that thought in mind, we can see why one might look to the theorist of attention in the hope that at this point she will have something explanatorily useful to say. Intentional internal acts of mental selectivity—acts of the sort that Wittgenstein’s case leads us to postulate—are exactly the sort of thing that a theory of attention ought to explain. It is for this reason that one might think that the theory of attention should contribute to the explanation of how it is that thoughts come to have their contents.

7.2.3 Wittgenstein’s Doubts The remarks in which Wittgenstein introduces this example are from one of the volumes assembled by his executors after his death. Like many of those remarks, they record fragmentary ideas without much accompanying discussion. It is not entirely clear what Wittgenstein himself takes the moral of his case to be. He seems to regard it as accurate to say that one attends to B when one is referring to B (elsewhere he tells us that ‘“describing” includes “attending”’ [1980: §725]). But he also seems to regard that claim as unexplanatory. In Philosophical Investigations we find some indications of his reasons for taking such a view. By the time we get to the sections of Philosophical Investigations where the private language argument is elaborated Wittgenstein is taking it that the reasons for having a dim view of attention’s explanatory potential are already in place. Attention makes an appearance in the sections concerned with private language when, at the end of §275, we are asked to ‘consider what it means “to point to something with the attention”’. But Wittgenstein here takes it for granted that the notion of ‘pointing with the attention’ is a hopelessly obscure one. Attention appears again, in §412, when Wittgenstein is considering ‘the feeling of an unbridgeable gulf between consciousness and brain process’: When does this feeling [of ‘an unbridgeable gulf’] occur in the present case? It is when I, for example, turn my attention in a particular way

Philosophical Work for the Theory of Attention

141

onto my own consciousness, and, astonished, say to myself: THIS is supposed to be produced by a process in the brain!

But, again, attention’s appearance here is only as an indicator that Wittgenstein’s imagined interlocutor has reached a dead end in the dialectic: ‘But what can it mean to speak of “turning my attention onto my own consciousness”? This surely is the queerest thing there could be!’ (§412). The case against attention comes earlier, before the subtleties of the private language argument begin, when Wittgenstein is discussing whether pointing is a way to communicate meaning to someone who is not already a participant in the language game. The example in this part of Wittgenstein’s discussion is one where, rather than picking out one of two people, the speaker is picking out one of the properties of a piece of paper: Point to a piece of paper.—And now point to its shape—now to its colour— now to its number (that sounds queer).—How did you do it?—You will say that you ‘meant’ a different thing each time you pointed. And if I ask how that is done, you will say that you concentrated your attention on the colour, the shape, etc. But I ask again: how is that done? (§33)

Wittgenstein’s point here is not that his interlocutor will have no answer when we ask how concentrating attention on the shape or the color is done. As the discussion progresses, an answer to that question does get provided on the interlocutor’s behalf: You sometimes attend to the colour by putting your hand up to keep the outline from view; or by not looking at the outline of the thing, sometimes by staring at the object and trying to remember where you saw that colour before. (§33)

Wittgenstein does not dismiss such accounts of ‘how attention is done’ as inaccurate. His point is that, however accurate those accounts may be, the metaphysics of attention prevents them from doing explanatory work: I want to say: This is the sort of thing that happens while one ‘directs one’s attention to this or that’. But it isn’t these things by themselves that make us say someone is attending to the shape, the colour, and so on. Just as a move in chess doesn’t consist in simply moving a piece in such-and-such a way on the board. (§33)

The analogy to chess is nowadays too hackneyed for it to be clear quite how we should take it when Wittgenstein employs it here, but in what follows I shall contend that these passages do contain an important insight about attention’s limits as a bearer of explanatory weight. There is an argument to be made along broadly these Wittgensteinian lines that shows it to be a mistake to invoke acts of attention in the explanation of how thoughts get to have their contents.

142

Attention Is Cognitive Unison

If this argument is correct then it is bad news for the explanatory project of John Campbell’s 2002 book Reference and Consciousness. Campbell’s claim there is that there is an ‘intuitive link [. . .] not just between reference to a perceived object and consciousness of it, but more specifically between reference to the object and conscious attention to it’ (45). Campbell is right about the existence of a link between reference of a certain sort and attention of a certain sort. We shall get an account of exactly what he is right about in second half of this chapter. The point about which he is wrong comes when he takes this claim about an ‘intuitive link’ between reference and conscious attention as indicating that attention has a foundational role in the explanation of the most basic reference-fixing relation. The particular explanatory role that Campbell has in mind for attention is the role assigned by Bertrand Russell to acquaintance. According to Russell, ‘acquaintance’ is ‘the simplest and most pervading aspect of experience’ (1913: 5). It is more explanatorily basic than the psychological states in which reference is made to particular objects: ‘Cognitive facts [are] defined as facts involving acquaintance or some relation which presupposes acquaintance’ (45). It is also more basic, in Russell’s explanatory scheme, than subjectivity itself: ‘Subjects have been already defined [. . .] as entities which are acquainted with something, and objects as entities with which something is acquainted’ (45). Campbell modifies Russell’s scheme by assigning to attention the role that Russell here assigns to acquaintance. Campbell therefore requires attention be an absolutely basic kind of mind-to-content relation. He puts the point like this: Conscious attention to an object [. . .] is a state more primitive than thought about an object to which we can appeal in explaining how it is that we can think about the thing. [. . .] This seems to be how Russell thought of acquaintance: [. . .] a state more primitive than propositional thought about the object, which nonetheless explains how thought about the object is possible. (2002: 45)

It was the young Wittgenstein’s objections—whatever exactly they were—that led to Russell’s abandonment of the project in which he attempted to use acquaintance as an explanatory primitive from which to build a theory of contentful mental states (see Eames 1984). In what follows I shall be suggesting that the later Wittgenstein’s argument shows, pace Campbell, that attention is equally unable to deliver these explanatory goods. It is worth emphasizing, however, that there is something that Campbell is right about. If his interest were not in the mystery of contentfulness, but were instead limited to the less foundational question of how certain kinds of object-referring concepts can be used successfully in certain kinds of perceptual environments, then the problem presented by the argument below could be avoided. It may be that Campbell would

Philosophical Work for the Theory of Attention

143

be happy to settle for these more modest explanatory goals. But if—as is suggested by his invocations of Russell and, elsewhere, of Wittgenstein and Quine—it is Campbell’s intention to use the theory of attention to cast explanatory light on the mystery of contentfulness (Campbell 2002: 4), then the Wittgensteinian argument developed below shows that his project cannot be completed.

7.2.4 An Argument Against Attention-Based Explanations of Reference Our argument against any explanatory project like Campbell’s can be approached in two stages. The first stage establishes that no theory of attention could be adequate to explaining the contentfulness of mental states that directly refer to objects unless that theory were capable of distinguishing attention to objects from attention to their properties. The second stage establishes that no theory can provide an account of what it is to be attending to an object, rather than to its properties, unless that theory presupposes the existence of a subject capable of thinking about objects. From these two claims it follows that any attempt at an attention-based explanation of contentfulness inevitably lapses into a certain sort of circularity. And since we can show that this sort of circularity is a variety of vicious circularity, we can conclude that attention cannot be used to explain the contentfulness of object-referring states. Before getting to this argument’s details there are three preliminary points that it will be helpful to clarify.

7.2.4.1 On Two Varieties of Reference The first point that is worth clarifying begins with the familiar thought that there are at least two different ways in which particular objects, events, and properties come to be the things that we are thinking about. Some thoughts get to be about particular objects, properties, or events because those objects, properties, or events are referred to directly. Other thoughts get to be about particular objects (or properties or events) because those thoughts employ a description that the objects in question (or properties or events) happen to fit. There are, correspondingly, at least two ways to have a thought that refers to the shortest spy. One can come to have a thought about the shortest spy by laying eyes on him, and thinking, for example, ‘That man is unusually short’. One can also come to have a thought about the shortest spy by employing a description that the shortest spy fits. If I am thinking, ‘The shortest spy must be unusually short’, or ‘The defeater of the Stunted-Gnomes must be unusually cunning’, then it is the shortest spy that I am thinking about. My thoughts are about the shortest spy, but unlike the thoughts that are had when I lay eyes on him, they are thoughts that refer to him by description, rather than directly.

14 4

Attention Is Cognitive Unison

It is worth reminding ourselves of this familiar distinction between varieties of reference because it is worth reminding ourselves that, when the most basic explanation of thought’s contentfulness is in question, the first variety of reference is the one that we need to start with. A description only puts a thinker in a position to think about the thing described if it is a description the components of which are themselves already meaningful for that thinker.1 Descriptions expand one’s referential capacities once the possibility of contentful thought is already in place, but they cannot be the first way in which thinkers come to have contentful thoughts. Descriptive reference therefore cannot provide the point of entry for a philosophical explanation of contentfulness. It is direct reference that we need to get an explanation of first.

7.2.4.2 On Explaining Phenomena and Accounting for Their Instances The second point to be clarified, as a preliminary to our Wittgensteinian argument, is really just a reminder of the mystery whose explanation is at issue. What we are concerned with is the mystery of how it is that creatures with brains that carry information about the things and properties around them get to be subjects of mental states in which those things and properties feature as contents. I am not disputing that, once we have creatures that are equipped with the capacity to enter into content-bearing mental states, attention to an object can bring the attended object into the set of things to which those creatures’ mental states can refer. That much is highly plausible, and it is well supported by the sorts of consideration that Campbell discusses. If I want to know which man you were referring to when you said, ‘That man . . .’, then Campbell quite rightly points out that I may be able to settle the question by working out which of the nearby men was receiving your attention. There is even a sense in which your reference to the man may be explained by your attention to him. The point that is under dispute is whether the availability of this explanation alleviates the mystery of how mental states get to have the property of being about things in the first place. A familiar analogy, adapted from Peter Strawson’s Individuals (1959: 51), may help to clarify this. The analogy is with the explanation of life. There was once a mystery about how it is possible for a world composed entirely of physical processes to contain things that display the property of being alive. The work that needed to be done in order to solve that mystery was work that involved (1) getting clear on what the essential features of being alive are, and (2) giving accounts that explain how naturally emerging systems could display those features. It would have done nothing at all to alleviate the mystery if a theorist had come along and pointed out that some particular instances of life are explained by instances of birth. This theorists’ claim would be true, of course. There is a sense in which it would give us the explanation of

Philosophical Work for the Theory of Attention

145

some instances of life, including, perhaps, the instances that we care about the most. But it would leave untouched the mystery (now solved) of how to situate the phenomena of life within a world of physical processes. That mystery would still remain to be accounted for, even if the imagined theorist had a detailed story of birth’s underlying dynamics. The theorist’s problem is not that explaining lives by describing the births that produce them leads to an infinite regress (although it does threaten to do that). Nor is the problem simply that, outside the class of mammals, there are many living things that have no births to account for them. The problem is that giving birth is itself an instance of the kind of life-involving phenomenon that needs to be explained. A theory of birth alone, however detailed and complete, is not a substitute for biology as a whole. In an analogous way, it may be true enough for a theorist like Campbell to say that some instances of contentful thoughts get their contents fixed by instances of attention. That may, in a sense, be explanatory, and the sort of explanation it provides may apply to some of the instances of contentful thought that we care about the most. The point I shall be arguing for is that the sort of explanation it provides is not the sort of explanation that alleviates the mystery of contentfulness that Quine, Wittgenstein, and Russell were concerned with. Attention is itself an instance of the kind of content-involving phenomenon that needs to be explained. A theory of attention alone, however detailed and complete, is not a substitute for a theory of contentfulness.

7.2.4.3 On the Particulars to Which We Refer One last preliminary point needs to be clarified in order to complete the setup for our Wittgensteinian argument against attention-based explanations of contentfulness. Various semantic entities contribute to the contents of our thoughts. Objects figure in the contents of those thoughts, both as particulars and generically. So do events, states, processes, properties, and locations. So do various logical and modal operators. There are some thoughts in which no reference to any particular object is made. Our final point of clarification is that we can safely assume that particular objects do figure among the contents of those mental states whose contentfulness is the first to be explained. This is just an assumption on which the argument to follow is based. It may be that an argument could be made in support of this assumption, but I will not be offering one here. Strawson considers one such argument, in chapter 1 of Individuals: That argument trades on the idea that only particular material objects could provide the spatiotemporal framework within which the other possible contents of thought are located and individuated (1959: 25–26). Strawson himself expresses reservations about whether such an argument can be made to work. Here I prefer to rest the claim that particular objects are among the primitive varieties of content, not on any such argument, but merely on the fact that it has

14 6

Attention Is Cognitive Unison

always seemed pretheoretically plausible that they are. As Berkeley asks, in the introduction to Principles of Human Knowledge: ‘Is it not a hard thing to imagine that a couple of children cannot prate together of their sugar-plums and rattles, and the rest of their little trinkets, till they have first [. . .] formed in their minds abstract [or] general ideas?’. (1710: §14)

It is indeed a hard thing, just as Berkeley suggests. Children seem to be in mental states that refer to particular concrete objects as soon as they can be in contentful states at all. They enter into such states independently of any ability that they might also have to think about properties in abstraction from the objects that instantiate them, or to think about the category of objects generically. It therefore seems safe to assume that mental states referring to particular concrete objects must be among those contentful states that get explained in a way that does not assume that any other sorts of content-involving state are already in place. The reader who does not like this assumption, or who does not like the idea of making it on the basis of so little argument, should not conclude on that basis that the argument to be made here will be an unsound one. In fact, it does not really matter what we assume here. An argument of the sort that I shall develop can be made, mutatis mutandis, whatever kind of items figure in the contents of the most primitive contentbearing states. The argument depends only on the idea that those items must be particulars. That they are particular concrete objects is inessential (although it is in this case that the argument is easiest to make). When we combine this assumption—the assumption that mental states referring to particular concrete objects must be among the first contentful states to be explained—with the first of our points of clarification—the point about direct reference having explanatory priority over reference by description—we get the result that nondescriptive reference to particular objects must be among the first things to be explained in the account of how the contentfulness of thought is possible. From this it follows that no explanation of contentfulness is adequate if that explanation needs to make use of content-involving notions in accounting for nondescriptive reference to particular objects. With these points clarified, we can now embark on our argument showing that attention cannot feature in the explanation of the contentfulness of mental states.

7.2.5 Attention to Objects and Attention to Properties The first part of our argument depends on there being a real distinction between attention to objects and attention to their properties. The clearest cases showing that there is such a distinction are the everyday

Philosophical Work for the Theory of Attention

147

cases in which we attend to an instance of a property without paying attention to the object that has it. This can occur when the property is a salient one but when the object that has that property eludes our attention, as may happen if the object behaves in ways that attention is unable to cope with (perhaps by moving too quickly to be seen as distinct), or when the object is an abnormal one, perhaps because it lacks stable boundaries or moves in ways that violate the principles of cohesiveness and general good behavior that must be complied with if attention is to keep hold of it (Scholl, Pylyshyn, and Feldman 2001; VanMarle and Scholl 2003). Objects that are elusive to attention, for whatever reason, can give rise to examples in which we have attention to properties in the absence of attention to the objects that have those properties. One’s attention might be caught, for example, by a shimmering color (caused, perhaps, by sunlight on the water from a lawn sprinkler). The color may catch one’s attention, but the spray of water causing it may elude attention. Such a case shows that one may attend to the color without being able to attend to the object that is colored (if indeed, there is any such object). Something similar may be true of the colors that appear on ‘Benham’s top’, the black-and-white disk that, when rotating, appears to present colors (Von Campenhausen and Schramme 1995). There are also lab cases that point to the existence of a distinction between attention to an object and attention to its properties. In a variation on the multiple-object-tracking paradigm, Bahador Bahrami (2003) has found that subjects who are paying attention to the task of keeping track of an object on a computer screen can nonetheless fail to detect changes in the color of that object. These subjects seem to be attending to the object (since they are keeping track of it). They seem not to be attending to its color (since they fail to notice that it changes). The experiments therefore point to the possibility of attention to an object while at least some of its properties go unattended. An experiment by Detlef Wegener and his collaborators points to the converse possibility—the possibility of attention to (at least some of) an object’s properties, without attention to the object itself. In Wegener’s experiments subjects are cued to shift their attention to the color of a moving sine-wave grating. When they do so, it becomes harder for them to detect changes in that grating’s motion (Wegener et al. 2008). Wegener’s interpretation of this finding is that it shows, what our example of the shimmer of water suggested, that attention may be allocated to a property—the color—while not being allocated to the object that has that property. One might suspect that instances of what we take to be attention to an object are just composed from several instances of attention to that object’s various properties, so that what we have here is not really a distinction between two kinds of attention, but only a distinction between

14 8

Attention Is Cognitive Unison

more or less complex instances of a single kind. We can counter this line of thought by developing a more a priori line of reasoning, pointing to the same distinction, even in cases where there is no possibility of attending to a larger or smaller number of the object’s properties. Imagine that our only perceptual contact with an object comes from dimly hearing it, or that the object is one that occupies a remote position in a featureless landscape, so that we can only vaguely see it, as a formless, featureless speck in the distance. Even in cases like these, where only a single property of the object might be discernable, there seems to be a distinction between attention to the property and attention to the object that has it. That there is such a distinction is shown by the fact that there are contexts in which an episode of attending to the only perceived property might develop differently from an episode of attending to the object that has it. In cases where there is a dimly heard sound the object might remain within earshot while its auditory properties change, or, alternatively, the properties might remain, but a new object might appear in the auditory scene to which those properties are switched. In these scenarios episodes in which attention is paid to the object might develop differently from episodes in which attention is paid to the property. In one case the subject’s attention remains with the object. In the other case it switches to the new bearer of the property. This suggests that we need a distinction between object attention and property attention, even when object attention involves perception of just one of the object’s properties. It therefore suggests that object attention is genuinely distinct from property attention and not merely a complex case of the same thing. With the distinction between object attention and property attention in mind, we can separate the question of whether attention can explain the contentfulness of thought into the question of whether object attention can explain contentfulness and the question of whether property attention can explain contentfulness. Having seen, as we did above (see §7.2.4.3), that no explanation of contentfulness is adequate if it needs to make use of content-involving notions to account for direct reference to particular objects, we can now argue for the conclusion that attention cannot explain contentfulness by first showing that property attention cannot explain direct reference to particular objects, and then showing that there cannot be such a thing as object attention, as distinct from property attention, unless reference to objects is already a possibility.

7.2.6 Attention to Properties Cannot Ground Direct Thought about Objects The considerations showing that attention to properties cannot explain direct reference to objects are relatively straightforward. We can bring them into view by imagining a case that combines elements from the two

Philosophical Work for the Theory of Attention

149

Wittgensteinian examples that were discussed above. Suppose we have a thinker who points to a piece of paper and says, ‘That has changed.’ ‘Do you mean the color?’, we might say. ‘No, the color is the same. What I meant was that it is a different piece of paper.’

Referring to the paper rather than the color seems possible, just as, in the earlier example, it seemed possible to refer to the color rather than the shape, or to one man rather than another. We therefore need an account of what it is that makes it the object, rather than its properties, that figures in the content of the thought that the speaker in our imagined scenario expresses. The line of thought that led to our initial optimism about using attention to explain reference would suggest that, in the case where the thinker refers to the object rather than the property, it is because the thinker is attending to the object rather than the property. If we follow that line of thinking here then we accept straightaway that attention to properties cannot explain reference to objects. If it is always attention to x that explains reference to x then it must be attention to the object that explains reference to the object, not attention to the property. Our opponent might offer some resistance at this point. The defender of an attention-based theory of reference need not accept the principle that, for all x, reference to x is explained by attention to x. Her claim may be that reference to items in one category is more basic than reference to items in other categories and that it is only in the basic case that we can apply the principle saying that reference to x must be explained by attention to x. This is a reasonable thing for our opponent to insist upon. But, given our earlier points of clarification, it does nothing to block the argument for the claim that it must be attention to the object, as distinct from attention to its properties, that explains reference to the object. Let us grant that reference to items in one category is more basic than reference to items in other categories, and let us grant that it is only in the basic case that reference to x is explained by attention to x. Either it is reference to objects that is more basic than reference to properties, or else reference to properties is more basic than reference to objects. If reference to objects is the more basic of the two then the principle that reference to x is explained by attention to x can be applied in this case, and so the simple route to the lemma that it must be attention to objects that explains reference to objects is available after all. If, on the other hand, it is reference to the properties of objects that is more basic than reference to objects, then we would need to violate the principles that we saw above, about direct reference to objects needing to be an explanatorily basic form of reference. If we start out with attention to properties putting the thinker in a position to have thoughts about those properties then we have to achieve states in which reference is made to objects by having those states derive

15 0

Attention Is Cognitive Unison

their contents from states that refer to the properties of those objects. Such a derivation can be achieved by starting with a thought in which the property is picked out directly, and then moving to a thought that uses this property as a way of fixing a description onto the object that has the property: One can, in this way, move from a thought with the form ‘that is F’ (where ‘that’ directly refers to a property) to a thought with the form ‘the object instantiating that is G’ (where again, ‘that’ refers to a property). But in making such a move the thinker is framing a description in which reference to the property is embedded, and is combined with a concept that refers to objects generically. For the purposes of the explanatory project that we are concerned with here this is doubly problematic. To form such a description is already to be in a contentful state that has objects (generically) as a part of its content. That contradicts our assumption that reference to particular objects must be explained without presupposing reference to objects generically (see §7.2.4.3). The formation of such a description also contradicts our supposition that the basic form of reference to particular objects has to be direct reference and not reference by description (see §7.2.4.1). Attention to properties therefore cannot explain reference to objects, other than by violating the principles with which our argument began. Those principles entail that, if attention is to explain reference, then it must be attention to objects that does the explanatory work. I am not denying that the route by which attention to a property can play a role in securing reference to the object having that property is a route that may be be available to a creature who is already capable of thinking about objects. The point is that this route is only open once the possibility of thought about objects is up and running. Unless the thinker is capable of thinking about objects in general she cannot deploy a complex demonstrative concept of the sort that she must deploy if she is to think ‘the object to which that property belongs is. . .’. When explaining how content gets up and running in the first place we should not have thinking about objects in general featuring in the explanation of direct reference to particular objects, and so, if attention is to play a role in the explanation of contentfulness, then it must be attention to objects that puts the subject in a position to refer to objects.

7.2.7 The Object-Attention/Property-Attention Distinction Cannot Be Drawn Unless Thought about Objects Is Mentioned We have seen that, if attention is to play a role in the explanation of contentfulness, then it must be attention to objects that puts the subject in a position to refer to objects. The second part of our argument proceeds by showing that any attempt to use attention to objects in the explanation of reference to objects lapses into circularity.

Philosophical Work for the Theory of Attention

151

This accusation of circularity needs to be treated with care. When we are dealing with causal explanations it is clear that circularity creates a problem. It does so because circular causal explanations require metaphysically peculiar temporal loops. But the explanation of contentfulness that is at issue here is not a causal explanation. What we want to know are the facts in virtue of which mental states have their contents, not the facts about which events cause those content-bearing states. Since the explanation we seek is not a causal explanation there is no risk of finding ourselves committed to metaphysically peculiar temporal loops. In these noncausal cases it is less clear why and when circularity creates a problem, and it has sometimes been suggested that in philosophical explanations some sorts of circularity do not create a problem at all (Schaffer 2004). It is clear, however, that some sorts of circularity do. One problem that circularity can create in noncausal explanations is what we might call the problem of ‘explanatory question begging’. We can see this problem by contrasting the following two cases. Suppose that we were asked for an explanation of the fact that there is an uncle in the room—not a causal explanation, telling us how it was that the uncle got in here, but a constitutive explanation, specifying the facts in virtue of which ‘there is an uncle in the room’ pertains. One way to give such an explanation is by saying that I am in the room, am male, and have a sibling with a child. This explanation is free from problems of circularity: What it is to be in the room, what it is to be male, what it is to be a sibling, and what it is to have a child can all be accounted for without having to bring uncles into the account. Suppose now that we were looking for a similar explanation for the fact that there is a sibling in the room, and suppose that someone proposes to explain there being a sibling in the room by citing the fact that I am in the room and am an uncle. As a reason for believing that there is a sibling in the room these facts are conclusive, but as a constitutive explanation they are no good at all. My being in the room and being an uncle entails that there is a sibling in the room, but it cannot be used to give a constitutive explanation of that fact. The problem that this attempted explanation faces is that my having the status of being an uncle (who is in the room) depends on the fact that I am a sibling (who is in the room). What it is to be an uncle cannot be accounted for without bringing siblings into the account. It is in this sense that the citing of unclehood ‘begs the question’ when it is a fact about siblings that we are trying to explain.2 A general philosophical account of what this explanatory question begging amounts to would require a serious metaphysical digression, but even without having such an account before us we can see that the circularity of explanatory question begging will afflict any attempt to use object attention in the explanation of direct reference to objects. We can get that charge to stick by noting the following points.

15 2

Attention Is Cognitive Unison

First, it is because uncles depend on the sibling facts for their status as uncles that uncle facts fail to explain sibling facts. In general, the fact that a is F cannot be used in the explanation of phenomenon ϕ if a depends for its status as an F on an instance of ϕ (per se). The second point to note is that the problem of circularity that one would face if one tried to explain the fact that there is a sibling in the room by citing the fact that there is an uncle in the room is a problem that cannot be avoided by producing a description, given in purely physical terms, of the uncle in the room. The circularity does not originate with the identification of the uncle; it originates with the features in virtue of which that item has the property of being an uncle. The third point to note is that, although explanatory question begging is fatal to the explanatory usefulness of citing one fact in the constitutive explanation of another, it is perfectly compatible with the existence of entailment relations between those facts (including deductive nomological entailment relations). As we noted, the fact that there is an uncle in the room entails that there is a sibling. It is nonetheless unable to provide an explanation of that fact. From these features of explanatory question begging, together with the lemmas established earlier, we can extract the following three consequences for our present dialectic: (1) If it is in virtue of their relation to object-referring mental states that certain activities have the status of being an instance of attention to objects then one cannot give a good constitutive explanation of the contentfulness of mental states by citing those instances of attention to objects (per se). (2) The problem that is faced by attempts to explain the contentfulness of thought by citing the facts about attention cannot be avoided by noting that a description of the instances of attention can be given in purely physical terms. (3) The fact that the subject is paying attention to an object may entail that the subject has contentful mental states, but it may nonetheless be unable to provide a constitutive explanation of that fact. It is the first of these points that creates the problem for attempts to use attention in the explanation of contentfulness. The second and third points serve to seal off what might appear to have been a way out and to account for the mistaken view’s apparent appeal. The attempt to use attention to explain contentfulness fails because the need to distinguish object attention from property attention cannot be satisfied without violating the strictures that must be met to avoid the circularity of explanatory question begging. If attention to objects is to explain contentfulness in its most basic instances then the account of what attention to objects is (as distinct from attention to properties) must be given without presupposing the existence of states the content of which is entertained by the thinker. While avoiding mention of such states we can identify the processing that instances of attention to objects are constituted by. What we cannot do is say why it is that these processes have the status of instantiating attention to objects.

Philosophical Work for the Theory of Attention

153

To illustrate this point, suppose, as we did in chapter 3, that there are some contexts in which a thinker’s attention to an object is constituted by the processes by which the representations of that object’s features are bound together into an object file. Suppose also that we can give a complete subpersonal description of the processes of objectfile formation—a description telling us all about the ways in which the various representations of an object’s properties are handled, and about the ways in which that information comes to be bundled into object file format. That description may tell us what happens while one pays attention to the object. But information that is so formatted has the status of being an object file only because of a role that it is thereby able to play. Being an object file is not a status that the resulting bundle of property representations gets to have on its own. The bundle of appropriately formatted representations gets to count as an object file only on account of its being a bundle of the sort that is usable by a subject in thought about objects. The processes that we describe when we give our description of an instance of object attention depend for their status as instances of object attention on there being a subject capable of thought about objects. And so, when explanations of the subject’s capacity to think about objects mention object attention, they thereby lapse into the circularity of explanatory question begging. The force of this point does not depend on any details about the theory of object-file formation. The point is that the context that is required in order for the processing of information about an object’s properties to count as attention to that object cannot be specified without making mention of those personal-level content-bearing states that our opponent cannot allow himself to mention, on pain of lapsing into circularity. It is possible, while adhering to these circularity avoiding restrictions, to give a description of the processes by which instances of object attention are constituted, and it is possible to describe the ways in which these processes differ from the processes that constitute instances of property attention. But if these processes are described in terms that really do satisfy the strictures required for noncircularity—if, that is, we do not presuppose that there exists a subject capable of thinking about objects—then the description given can do no more work than is done by the physical description of the uncle in the room. The problem with uncle-citing explanations of sibling facts was not identifying the uncle but explaining why he has the status of being an uncle. The problem here is not that we need to mention content-bearing states in order to identify the attention-constituting processes. The problem is that we need to mention content-bearing states in order to account for these attentionconstituting processes having the status of constituting attention to an object. The content-free description of the attention constituting processes can tell us only about the information that those processes handle but it cannot specify the facts in virtue of which these processes have the

15 4

Attention Is Cognitive Unison

status of being instances of attention to objects, and so it does not give us a way in which explanatory question begging can be avoided. What we want to say here is what Wittgenstein wanted to say when his imagined interlocutor suggests that one attends to the color by holding up one’s hand or by staring at the color and trying to remember where one has seen it before. As long as the account of attention to objects is given only in the terms that allow circularity to be avoided, then: I want to say: This is the sort of thing that happens while one ‘directs one’s attention to this or that’. But it is not these things by themselves that make us say someone is attending to [the object rather than] the shape, the colour, and so on. Just as a move in chess does not consist in simply moving a piece in such-and-such a way on the board. (Wittgenstein 1953: §33)

The argument that we have extracted from this passage can be summarized as follows: Attention-based explanations of reference face the problem that, without being embedded in a context in which there are mental states referring to objects, there is nothing in virtue of which it is attention to the object that gets constituted by the processing of information about that object’s properties. But when the existence of such a context is presupposed, the resulting account of attention to objects is no longer usable in a noncircular explanation of contentfulness. The theory of attention therefore provides us with no such explanation.

7.2.8 The Explanatory Role of Modes The Wittgensteinian line of thought that we have just developed as an argument against thinking that attention plays a role in the explanation of contentfulness is independent of the cognitive unison theory. One can nonetheless see why the defender of that theory is invested in there being such an argument to make. When, in chapter 4, we were building the cognitive unison theory of attention, we presupposed the existence of a thinker who is capable of being in the content-bearing states that are involved in understanding tasks. The capacity to be in content-bearing mental states was taken for granted when specifying which activities count as tasks (in §4.5). It was taken for granted again when specifying which processes count as cognitive processes (in §4.8). If a theorist like Campbell were right, and attention had a role to play in explaining how contentfulness is possible, then these moves would be illegitimate. The cognitive unison theorist cannot accept Campbell’s picture of the explanatory relations here. The cognitive unison theory’s alternative picture of these explanatory relations stems from the fact that, being a version of adverbialism, the theory takes attention to be a mode; specifically, it takes attention to be a mode of engagement in one’s tasks. Modes of engagement, like all modes, come late in the order of explanatory priority. They come

Philosophical Work for the Theory of Attention

155

after the things or processes that they modify. Ambling, limping, and hiking, for example, are all modes of walking. To explain what these phenomena are and how they are possible one need only give a specification of the properties that modify walking in each case. For that reason, one makes the mistake of getting one’s explanatory priorities backward if one tries to use ambling (or limping or hiking) in one’s explanation of what walking is. According to the cognitive unison theory this same pattern of explanatory relations holds between attention and the kind of contentful thought that is involved in undertaking a task. The cognitive unison theorist accounts for attention as a mode of (one variety of) contentful thought. His account of what attention is and of how it is possible presupposes that contentful thought is possible and that it gets explained elsewhere. If the adverbialist explanatory approach is correct then it is quite legitimate for the cognitive unison theory to presuppose this. But if that approach is correct then any attempt to use attention in explaining the contentfulness of thought must be making the mistake of getting its explanatory priorities backward. That they are indeed doing so is what our Wittgensteinian argument suggested. So much, then, for attention-based attempts to address the mystery of contentfulness. They are not possible, and the cognitive unison theory gives us an account that traces this impossibility to the metaphysical relationship between attention and contentful states. What, then, of attention-based attempts to address the mystery of consciousness?

7.3 ATTENTION AND CONSCIOUSNESS

7.3.1 Recent Optimism The idea that there is an explanatory link between attention and consciousness has what is, by cognitive scientific standards, a relatively long history. One of William James’s early writings tells us, ‘Whoever studies consciousness, from any point of view whatever, is ultimately brought up against the mystery of interest and selective attention’ (James 1879: 8). A hundred years later—when psychologists were realizing that the whole tradition of theorizing that was inspired by Donald Broadbent’s attempts to explain attention had in some way been mistaken—this link between attention and consciousness began to seem like one of the reasons for regarding theories of attention with suspicion. In a 1980 review of research into attention and task performance, Alan Allport took the Jamesian link between attention and consciousness as an occasion for issuing the following ‘word of warning’: The study of attention began in phenomenology. [. . .] The word is still used, by otherwise hard-nosed information processing psychologists, as a code name for consciousness. (1980: 113)

15 6

Attention Is Cognitive Unison

More recently the suspicions that Allport voiced have abated. Attempts to link theories of attention to the explanation of consciousness are now regarded as scientifically promising approaches to a mystery, not as lapses into unscientific phenomenology. So it is that we find Anne Treisman suggesting that descendents of her feature integration theory of attention may provide part of the explanation for ‘the bound, unitary, interpreted, personal view of the world of subjective experience’ and going on to suggest, albeit tentatively, that the sort of explanation that such a theory will provide ‘should give us all the information there is about the conditions that create consciousness’ (2003: 111). Michael Posner, whose work in the 1970s pioneered the methods by which shifts of covert visual attention are studied, does not share Treisman’s tentativeness. He has for some time been an enthusiastic advocate of the idea that the study of attention is the way to produce an explanation of consciousness. In a 1994 article titled ‘Attention: The Mechanisms of Consciousness’, he writes that ‘an understanding of consciousness must rest on an appreciation of the brain networks that subserve attention, in much the same way as a scientific analysis of life without consideration of the structure of DNA would seem vacuous’ (7398). There is also optimism on the part of some philosophers about the idea that our theories of attention can be turned into theories of consciousness. Jesse Prinz, for example, takes the attention/consciousness relation to be very close indeed, claiming that: The neural processes underlying attention are the physical mechanisms by which the neural correlates of intermediate-level representations become conscious. How do we become conscious? We attend. (forthcoming)

To a pretheoretical eye, this optimism about using theories of attention to explain consciousness seems misplaced. There are, it seems, states of consciousness that attention cannot possibly explain since they are states that are conscious but from which attention is wholly absent.

7.3.2 Inattentive Conscious States James provided a description of one such inattentive conscious state at an early stage in the discussion of attention in The Principles of Psychology, in a passage from which we have already quoted, back in chapter 4: Attention [. . .] is a condition that has a real opposite in the confused, dazed, scatterbrained state which in French is called distraction, and Zerstreutheit in German. We all know this latter state, even in its extreme degree. [. . .] The eyes are fixed on vacancy, the sounds of the world melt into confused unity, [. . .] the foreground of consciousness is filled, if by anything, by a sort of solemn sense of surrender to the empty passing of time [. . .]

Philosophical Work for the Theory of Attention

157

Every moment we expect the spell to break, for we know no reason why it should continue. But it does continue, pulse after pulse, and we float with it, until—also without reason that we can discover—an energy is given, something—we know not what—enables us to gather ourselves together, we wink our eyes, we shake our heads, the background-ideas become effective, and the wheels of life go round again. (1890: 382)

If states of the sort that James is here describing are, as they seem to be, states in which the subject is conscious while paying no attention, then those states create a problem for the idea that attention plays a role in the explanation of consciousness since they force the theorist who wants to give an attention-based explanation of consciousness to have some other explanation of inattentive consciousness.

7.3.3 The Prima Facie Case against the Necessity of Attention for Consciousness The theorist who wants to defend the claim that attention is necessary for consciousness (and so to leave open the way for a comprehensive attention-based explanation of consciousness) has three possible responses available when presented with cases of the sort that James describes. She may deny that there are any cases of the relevant sort. She may claim that the subjects of such states are unconscious. Or she may claim that such cases do involve attention. The first of these responses—the claim that James’s description simply does not apply to any real-world states—lacks plausibility. We immediately recognize James’s description as an accurate one and can find plenty of actual states to which it applies. In some circumstances, when we are extremely tired or bored, inattentive states of the sort that James describes seem to be consciousness’s default setting. The defender of attention-based explanations of consciousness cannot plausibly claim that there are no such states. The second response—the denial that such states are states of consciousness—is, on the face of it, hardly any more plausible. James’s description of the inattentive state is a description that we immediately recognize as accurate. That recognition seems inexplicable if the state is one that we never consciously experience. The fact that one can know that one has recently been in an inattentive state, and can know it on the basis of first-person authority, seems to be best explained by the hypothesis that one is conscious when in such states. The defender of attention-based explanations of consciousness cannot plausibly claim, without additional argument, that the states James describes are not states of consciousness. The third possible response is to say that the subjects of the states that James describes are conscious but that they are also attentive. This is somewhat more plausible. It may even look as if James himself was tempted by it. The passage that gives this impression is one that I elided

15 8

Attention Is Cognitive Unison

in my original quotation. When we restore it we find that what James actually wrote is this: The eyes are fixed on vacancy, the sounds of the world melt into confused unity, the attention is dispersed so that the whole body is felt, as it were at once, and the foreground of consciousness is filled, if by anything, by a sort of solemn sense of surrender to the empty passing of time. (emphasis added)

When James says that ‘the attention is dispersed’, this could be taken as the admission that we do attend in the states he is describing, but attend in a dispersed fashion. If so then these states are not even putative counterexamples to the claim that attention is necessary for consciousness. This would, however, be the wrong way to understand what James is saying here. We can see this by considering the passage that immediately precedes the one just quoted. It is in the passage immediately preceding his remark about the attention being dispersed that James famously says: ‘Everyone knows what attention is. It is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought. Focalization, concentration, of consciousness are of its essence’ (1890: 381–382). Since James has been telling us that attention is essentially selective and focalized, just prior to his remark about the attention being dispersed, it would be strange to interpret that latter remark as the admission that this essentially selective and focalized phenomenon also has nonselective and nonfocalized forms. We should instead take seriously James’s claim that the state he is describing is ‘the real opposite’ of attention, and so should understand his remark that ‘the attention is dispersed’ in the same way that we would understand a remark such as ‘the pain is abated’ or ‘the tension is released’: For the attention to be dispersed is not for there to be attention that is no longer selective; it is for there no longer to be attention at all. This is not simply a point of Jamesian hermeneutics. Back in chapter 1 we took from James the idea that the selective directedness of our mental lives is what it is that a theory of attention needs to account for. If the defenders of attention-based explanations of consciousness try to claim that the subjects of the states that James described are paying attention, but in a uniformly distributed way, then their notion of attention has to have been stripped of its essential selectivity. Having been stripped of that, their claim is no longer recognizable as a claim about attention as we have been understanding it. With these considerations in mind, we can say that James’s state appears to be a state of consciousness, and that it appears to be a state from which attention must be absent. If these appearances are correct then attention is not necessary for consciousness. James’s discussion therefore provides the grounds for a prima facie case against the idea that attention has a role in the explanation of consciousness.

Philosophical Work for the Theory of Attention

159

7.3.4 The Refrigerator Light To maintain the currently orthodox view—that we are conscious only when we are attending, and conscious only of the things to which we attend—one needs to have an answer to the case that we have just outlined. The preferred answer in the current literature is a version of the second of the responses that we considered above: The cases of apparent inattentive consciousness are said not really to be cases of consciousness at all. The counterintuitiveness of this claim is mitigated by an attempt to explain away appearances of consciousness in inattentive states as an instance of the ‘refrigerator light illusion’. We think that we are conscious in these states, but, the story goes, we think it only because we find that we are conscious whenever we attentively check them. This gives an illusory impression that inattentive states are conscious, just as one might get an illusory impression of the refrigerator light always being on from finding that it is on whenever one opens the door to check (see O’Regan and Noë 2001: 947; Tye 2009: chap. 7). These considerations have some force. The existence of states that are found to be conscious only when attentively checked would give us no reason at all to deny that attention is necessary for consciousness. But our reasons for thinking that we are conscious when in the sorts of states that James described are not based only on a checking procedure. We think such states are conscious, not only because we find that they are conscious when we check them, but also because our consciousness while in these states has an epistemically explanatory role to play. It is because these states are conscious that we can know, in an authoritative first-person way, that we have recently come out of such a state. It is because we are conscious when in such states that we can recognize James’s description of them as a phenomenologically accurate one. And it is because we are conscious of unattended items that shifts of attention to a new item can be intentionally targeted acts, rather than shots in the dark. None of these explanations is available to the theorist who claims that there is no consciousness without attention. The refrigerator light considerations prevent direct checking from giving us a reason to believe in inattentive consciousness. They remind us that the case for inattentive consciousness has to be made on the basis of an inference to the best explanation. But they do not show that the prima facie case for inattentive consciousness is no case at all because they do not deprive us of the epistemic and phenomenal explananda that inattentive conscious states are needed to explain. Having been reminded that we are here forced to proceed by an inference to the best explanation, we should remember, also, that inference to the best explanation is only rationally compelling when it is inference to the best explanation of all the data. To settle the question of whether

16 0

Attention Is Cognitive Unison

attention really is necessary for consciousness we therefore need to determine whether, when all things are considered, we can get a simpler and more coherent explanatory picture by accepting that hypothesis or by rejecting it. I have been suggesting that reflection on our knowledge of our own inattentiveness points us toward rejecting that hypothesis, but there are various effects observed in the lab that have been thought to require that we accept it. In the following two sections I shall show that there is a simple and coherent picture of the attention/consciousness relation that is able to explain these lab effects, and the phenomenological/epistemological data that James highlighted, without needing to endorse the hypothesis that attention is necessary for consciousness. Inference to the best explanation being what it is, these considerations remain inconclusive, but, for the data that we have available at the moment, the best explanation is, I shall argue, one according to which there are conscious inattentive states. And if there are such states then optimism about attention-based explanations of consciousness is misplaced.

7.3.5 The Inattentional Blindness Paradigm The most frequently cited of the lab effects that are taken as evidence for the hypothesis that attention is necessary for consciousness are the inattentional blindness effects and the change blindness effects. The first of these were presented in Arien Mack and Irvin Rock’s influential 1998 book Inattentional Blindness. In the preface of that book Mack tells us that this research ‘brought to light some dramatic and surprising findings’ and that ‘the single most important lesson is that there seems to be no conscious perception without attention’ (ix). If Mack and Rock’s experiments really did establish such a lesson then the prima facie case from James would be defeated. The experiments by which they purport to establish it employ the following procedure: Observers were asked to report the longer arm of a cross [. . .] The cross was presented on the screen of a computer for 200 ms. [. . .] In most cases, as soon as the cross disappeared, a pattern mask appeared for 1500 ms that covered the entire area of the visible screen. [. . .] Before each presentation of the cross, a fixation mark was displayed at the centre of the screen and the subjects were asked to keep their eyes focussed on it until the mask appeared. When the mask disappeared, subjects reported which line of the cross seemed longer. [. . .] On the third or fourth trial a critical stimulus was presented in a quadrant of the cross within 2.3 degrees of fixation. [. . .] Immediately following the trial in which the critical stimulus was presented, the subjects were asked whether they had seen anything on the screen other than the cross figure. (6)

Crucial to this experimental procedure is the fact that the ‘critical stimuli’ are presented for a fifth of a second and then masked, are presented

Philosophical Work for the Theory of Attention

161

to subjects who are not expecting them, who have very little practice with the task, and who are occupied with a different task involving different stimuli. The result, which Mack and Rock describe as ‘dramatic’ and as ‘puzzling and surprising’, is that ‘on average 25% of the observers failed to detect their presence’ (13). There is nothing dramatic here for Mack and Rock to be puzzled or surprised by. It is neither puzzling nor surprising to find that there are some circumstances in which small, unexpected items escape the notice of people who are not paying attention. We all know perfectly well that there are such circumstances. It is because we know this that we are concerned to ensure that people pay attention in contexts when the items that they might otherwise fail to notice are important ones. A part of our reason for insisting that people pay attention when driving, for example, is that we know, if they fail to pay attention, that they may not notice unattended items, including such things as pedestrians. Since we know that attention is sometimes necessary for consciousness of pedestrians, it should not be surprising to learn that it is sometimes necessary for consciousness of small shapes that are presented unexpectedly, for a fifth of second, concurrently with something else that one is attending to, in an unfamiliar experimental paradigm, and followed by a pattern mask. Mack and Rock’s discovery that people sometimes fail to notice unattended stimuli in these circumstances shows, at most, that attention is sometimes necessary for consciousness. That is not news, nor is it a good reason for thinking that attention is always necessary for consciousness. Mack and Rock did make some surprising discoveries. It is surprising that people get worse, not better, when the ‘critical stimulus’ is presented in the very center of their visual field (68–69). It is surprising that subjects in these conditions fail to notice frowny-faced symbols but usually do notice smiley-faced ones (140–145). It is surprising that subjects fail to notice misspelled versions of their names but do notice if the name is spelled correctly (120–123). These findings can be interpreted as indicating that smiley faces and correctly spelled names usually attract the attention. So interpreted, the findings are consistent with Mack and Rock’s hypothesis that attention is necessary for consciousness. But this is by no means the only interpretation of them. One can just as well take the data as showing that smiley faces, since they are consciously detected even in conditions that lead to inattentional blindness, are a counterexample to the claim that attention is always necessary for consciousness. The explanation of Mack and Rock’s data does not require us to adopt the hypothesis that attention is necessary for consciousness.

7.3.6 Other Inattention Effects Some of the research inspired by Mack and Rock’s book does a better job of being surprising. Much of it succeeds in being memorable. It has

16 2

Attention Is Cognitive Unison

become well known, for example, that the appearance of a pantomime gorilla in a video of a basketball-type game is often not noticed (Simons and Chabris 1999; Most et al. 2001). Less well known are that 42% percent of people do notice the gorilla and that this number goes up markedly if the people watching are basketball experts (Memmert 2006). It has also become well known that large-scale but narratively insignificant changes to a picture can take more than ten seconds to spot when a flicker coincides with their appearance and disappearance (Rensink, O’Regan, and Clark 1997; Rensink 2002). Again, however, the interpretation of these effects as indicating that attention is necessary for consciousness is not the only interpretation available. Faces again behave in ways that suggest they might be a counterexample to the ‘attention is necessary for consciousness’ hypothesis (Ro, Russell, and Lavie 2001). More deeply problematic are the fact that narratively significant changes in the change blindness paradigm do get spotted within two to three seconds and the fact that, as Rensink and his collaborators have been careful to point out, verbal cues as to what to look for are immediately useful in detecting where the change is taking place (Rensink, O’Regan, and Clark 1997). These findings suggest that there is conscious perception outside the locus of attention, although only of the gist of the scene (Rensink, O’Regan, and Clark 1997; Koch and Tsuchiya 2007). More recent work using functional magnetic resonance imaging of distracted subjects who are presented with stimuli for as little as 130 milliseconds points to the same conclusion (Peelen, FeiFei, and Kastner 2009). As with the effects observed by Mack and Rock, there is an interpretation of the change-blindness results that is consistent with the hypothesis that attention is necessary for consciousness. But mere consistency of data with a hypothesis is never, by itself, a reason to believe that the hypothesis is true, least of all if there are other, less revisionary hypotheses that cohere with the data just as well. In this case there is such an hypothesis: the hypothesis according to which inattention leads to somewhat depleted consciousness, but leaves consciousness of gist intact. 3

7.3.7 A Jamesian Account of Inattention’s Consequences To get a more precise account of this alternative hypothesis we can start by considering unattended sounds. In the passage from Principles of Psychology that gave this discussion its starting point, James says that in the inattentive state ‘the sounds of the world melt into confused unity’ (1890: 382). This phenomenal unity of unattended sounds may be taken to consist in its being the case that, although unattended sounds are not absent from consciousness, the individual unattended sounds are nonetheless not perceived as distinct. The contents of the unattended state cannot be represented in the form: ‘This item has χ. That item has δ.

Philosophical Work for the Theory of Attention

16 3

The other item has ε’. The content is, instead, that a certain set of more or less generic properties are somewhere instantiated in the scene perceived. There is nothing ad hoc about this hypothesis. It accords, on James’s telling, with the phenomenology of inattention. It also accords with the sort of thing suggested by several strands of Treisman’s classic work on feature integration. The thought underlying much of that work is that removing attention does not remove consciousness altogether, but that it does prevent the contents of consciousness from being packaged into object files. The particulars that are experienced by the inattentive subject are therefore not perceived as particulars. This Jamesian hypothesis about the perceived unity of unattended sounds easily explains an auditory analogue of inattentional blindness. It does so by entailing that an inattentive subject’s awareness of several sounds is of the wrong sort to put her in a position to acquire knowledge about any one of these sounds in particular, and so is of the wrong sort to put her in a position to know which sound is changing. If we were to ask the subject of an auditory change-blindness task which particular sound had been changing she would be unable to tell us, not because her inattention left her unconscious of the sounds, nor even because it meant that the change in sound was not consciously registered, but just because her consciousness of the changing sounds was not structured in such a way as to be adequate to providing her with knowledge in which any one of them was picked out as a particular. The Jamesian hypothesis need not be restricted to the auditory domain. What goes for unattended sounds seems also to be true of, for example, the speckles that one consciously sees, but gives no individual attention to, when seeing that a passing hen is speckled (Chisholm 1942; Tye 2009). The hen’s speckles figure in one’s conscious experience. That experience thereby puts one in a position to know that the hen is speckled. But it does not put one in a position to know about any of the individual speckles as particulars. This idea that unattended items are experienced as a unity and not as individuals enables us to make precise the thing that was right about Campbell’s treatment of the relationship between attention and reference. Campbell was right to say that some instances of attention have an essential role to play in the explanation of how some instances of thought about particulars get to have those particulars as their contents. If unattended items are not experienced as distinct from one another, then, when there are multiple items in the thinker’s perceived environment, the inattentive thinker cannot pick out any one of those items as a particular about which he could then go on to form a thought. When multiple items are experienced, it is only the attentive thinker whose experience makes it possible for him to have a thought about any one of those items. What Campbell is right about, then, is that in crowded

16 4

Attention Is Cognitive Unison

environments (and three really is a crowd in the present context), it is attention that explains how it is that thoughts get to be about particulars, rather than about the glommed together unities of whichever items happen to be present. Campbell is wrong only in taking this to be a sign that a theory of attention should give us an account of the most basic way in which psychological states get to have the property of referring to particulars. It is, in fact, a consequence of the fact that crowded conditions mean that particularity requires selectivity, and of the fact that attention is what makes our mental lives selective. Now that we have this Jamesian hypothesis on the table, and have seen that it is able to account for our ignorance of unattended items without supposing that those items drop out of consciousness altogether, we can see that the empirical evidence that is typically adduced in support of the hypothesis according to which consciousness requires attention faces a methodological challenge. In order for that hypothesis to be established it would be necessary to rule out the Jamesian alternative. The problem is that our current ways of operationalizing attention and consciousness give us no way of settling the dispute between these hypotheses. We frequently have no way to gauge whether the subject is conscious of an item, other than by gauging the subject’s knowledge of that item. What the Jamesian hypothesis makes clear is that getting into a position from which one can gain knowledge of a particular item (in which that item is represented as a particular) may be attention demanding in ways that having conscious states in which that item figures is not. Experiments in which a subject’s knowledge is used to gauge the contents of his consciousness may therefore underestimate the number of things that figure in the contents of the subject’s conscious states. Worse, they may do so in a systematic fashion: In crowded environments those experiments will tend not to register the subject’s consciousness of unattended items. Experiments in which the subject’s beliefs are used as an indicator of the contents of the subject’s consciousness will therefore be systematically blind to precisely the sorts of case whose existence is at issue when we consider the question of whether attention is necessary for consciousness. When we use inference to the best explanation to make our choice between the hypothesis according to which attention is necessary for consciousness and the hypothesis according to which attention is necessary only for selectivity (and so necessary for consciousness of particulars as particulars when they are presented in crowded environments), we find that both hypotheses can accommodate the data from the inattentional blindness and change blindness experiments. The latter hypothesis has the advantage that it can also account for the evidence of gist perception in the absence of attention, and for the facts about our firstperson knowledge of our own inattentiveness. This gives us a reason for

Philosophical Work for the Theory of Attention

165

preferring it. If that hypothesis is true, then there are states of consciousness that do not involve attention to the thing of which we are conscious. This gives us a reason to think that the recent optimism about attentionbased explanations of consciousness is misplaced.

7.3.8 The Cognitive Unison Theory’s Account of the Attention/Consciousness Relation In §7.2.8, when considering the cognitive unison theory’s perspective on the possibility of attention playing a role in the explanation of content, we saw that it is because the cognitive unison theory characterizes attention as a mode of task performance, because task performance is content demanding, and because modes always come after the entities that they modify in the order of explanatory priority, that the cognitive unison theory rejects attention-based explanations of contentfulness. That argument did not depend on anything to do with attention being unnecessary for content. It depended on a point about the metaphysical relation of modes to things modified, and on a point about the explanatory consequences of that metaphysical relation. When considering the possibility of attention playing a role in the explanation of consciousness, the cognitive unison theory’s reasons for pessimism take a slightly more complicated form, but, again, they depend on the theory’s claims about the metaphysical basis for the relation between consciousness and attention, not simply on facts about whether either one of those phenomena necessitates the other. The considerations that we have just outlined, showing that attention is not necessary for consciousness, undermine the current optimism about attention-based explanations of consciousness, but, as far as the cognitive unison theory is concerned, they do not lie at the metaphysical heart of the matter. If the cognitive unison theory were to say that attention is a mode of consciousness, then the theory would entail that there cannot be an attention-based explanation of consciousness in just the way that it entailed that there cannot be an attention-based explanation of contentfulness. But the cognitive unison theory does not say that attention is a mode of consciousness, nor does it imply as much unless a substantive assumption is added. The additional assumption that would be needed in order to get from the cognitive unison theory’s claim that attention is a mode of contentinvolving task performance to the conclusion that attention is a mode of consciousness is this: γ: The agents of tasks are always conscious of the content represented by the cognitive processes that serve those tasks.

If γ were true then task-performance would be a phenomenon of consciousness. Cognitive unison, as a mode of task performance, would therefore be a mode of (one variety of) conscious activity, and attention,

16 6

Attention Is Cognitive Unison

as the cognitive unison theory understands it, would therefore be unable to figure in the explanation of consciousness, just because modes are unable to explain the things that they modify. One might think that the cognitive unison theory is committed to all of this by the definitions of ‘cognitive process’ and of ‘task’ that were given in chapter 4. Both of those notions were defined so as to be agent involving. An activity counts as a task of the agent if and only if it is performed under the guidance of the agent’s understanding (as we saw in §4.6). A representation-handling process counts as cognitive if and only if the representations that it handles are such that it is in virtue of those representations having the content ϕ that the agent of the task is put in a position to be in states the content of which is ϕ (as we saw in §4.7). The agents of most tasks are conscious organisms. These agents deploy their understanding consciously. The agent-level states that are mentioned in the definitions of ‘task’ and ‘cognitive process’ would, for these agents and these tasks, be conscious states. For these agents, and these tasks, γ follows. But as an argument for γ in general this line of argument would be too quick. Not all tasks have normal conscious individuals as their agents. The clearest of the exceptional cases are those in which the agent of an attentively performed task is not an individual at all but is, instead, a collective.

7.3.9 Collective Agents Can Perform Attentively The official definition of ‘task’, as given in chapter 4, says that a task is an activity performed under the guidance of the agent’s understanding. That definition allows that a collective of individuals can perform a task if that collective has an understanding of some activity and is guided by that understanding when performing it. Sports teams are good examples of collectives that have and use understandings in this way, partly because they perform an activity that is essentially collective, and partly because the division of labor between different members of the team may make it the case that no individual member of the team understands the collective activity in its entirety. Take the Tottenham Hotspur Football Team as an example. It is quite natural to say that the team understands how to play football, and also that it understands how to accomplish particular kinds of footballing task, such as how to defeat Wigan Athletic. It is also quite natural, given the team’s success on the field, to say that this understanding has a role in guiding the way in which the team plays. In such a case we must attribute understanding of the overall task to the team as a whole. Each individual member of the team may understand no more than his own role. In §4.5.1, I argued that the requirements of understanding may vary, depending on the repertoire of activities that the agent’s form of life requires it to negotiate. Since the soccer team’s form of life involves little more than the playing of soccer matches, and since the form of

Philosophical Work for the Theory of Attention

16 7

life of the individual players involves a rather more diverse repertoire of activities, it may be that the requirements that the team must meet in order to count as understanding how to defeat Wigan are quite different from the requirement that the players would have to meet in order to count individually as understanding that same thing. It is therefore quite possible for the team as a whole to count as understanding how to defeat Wigan, even when the players individually do not. In such a case the team qualifies as the agent for the task. The team, unlike its component players, is not conscious, and so the content-bearing states of this task-performing agent cannot be states of which the task’s agent is conscious. Such a case is therefore a counterexample to γ and so shows that attention is not treated by the cognitive unison theory as a mode of consciousness. It is not simply the explanatory limitations on modes that make the cognitive unison theory unsympathetic to attention-based explanations of consciousness.

7.3.10 Is Nonconscious Attention More Explanatorily Primitive Than Conscious Attention? The cognitive unison theory can allow for attention without consciousness in cases of collective agency, and, in a slightly different way, the theory can allow for attention without consciousness in cases of blindsight (Kentridge, Heywood, and Weiskrantz 1999; Kentridge and Heywood 2001; Kentridge, Nijboer, and Heywood 2008), and in conditions of ‘interocular suppression’, in which a stimulus presented to one eye is kept from conscious experience because another, more vivid stimulus, is simultaneously presented to the other eye (Jiang et al. 2006). Experiments into both of these phenomena show that attention can be influenced by the processing of stimuli that are not consciously experienced. It requires an additional nontrivial step to get from here to the conclusion that attention is paid to those very unseen stimuli (rather than, say, to the location in which those stimuli appear [see Mole 2008, Koch and Tsuchiya 2008]). But the cognitive unison theory can allow that additional step to be taken, and can accommodate the resulting cases of attention without consciousness, somewhat as it accommodates cases of collective agency, as cases in which a cognitive process fails to support a state of consciousness. The cognitive unison theory allows for cases in which attention is paid to stimuli of which the attentive subject is not conscious, but it nonetheless insists that there is something unsatisfactory about taking such cases as one’s reason for thinking that attention is more basic than consciousness, and so is available as a component from which an explanation of consciousness could be built. It may be true that there are cases in which the blindsighter is not conscious of the stimuli to which his attention is directed, since those stimuli fall within his scotoma, but the fact that his attention is directed to

16 8

Attention Is Cognitive Unison

those stimuli is manifested only in the guesses that he voluntarily makes about whether and where these unseen items are presented. These acts of guessing are thoroughly conscious. It is only because of the effect that we observe on these conscious acts that we say the blindsighter is attending. It is not clear, then, that the attention displayed by the blindsighter is really a phenomenon that is sufficiently independent of consciousness to be legitimately used in the explanation of it. Something similar is true of the attention that we see in cases of interocular suppression. The subjects in Jiang et al.’s (2006) interocular suppression experiment are conscious of the vivid patterns of visual noise that are presented to one eye, and they are not conscious of the erotic images that are presented, at a lower contrast level, to the other eye. The erotic content of the unseen images nonetheless has an effect on the subjects’ attention. But, again, the effect of attention here, although driven by unconsciously perceived stimuli, is not wholly independent of consciousness. It manifests itself in the accuracy of the subjects’ subsequent performance in a visual discrimination task that requires the subjects to make a conscious judgment about the orientations of two consciously experienced Gabor patches, presented shortly after the erotic image and the visual noise have been removed from the screen. Again, then, it is an effect on consciousness that gives us our reason for taking the phenomenon displayed in this experiment to be a phenomenon of attention. Even in the case of collective agency consciousness is not wholly absent. The soccer team itself is not conscious, but it nonetheless owes its capacity to enter into understanding-guided behavior to the conscious states of the individuals who constitute it. In the total absence of consciousness, attributions of attentiveness seem out of place. It therefore seems that these cases of attention in the absence of consciousness, although they show that consciousness is not necessary for attention, provide little reason to think that attention is more explanatorily basic than consciousness. Even if we could find cases in which attention operates wholly independently of consciousness, it seems that those cases would essentially be deviations from a consciousness-involving norm. It would only be on account of their relation to cases of conscious attention that we would recognize cases of wholly unconscious attention as cases of attention at all. In this respect the explanatory relation of unconscious attention to conscious attention is somewhat analogous to the explanatory relation between the formation of delusions and the formation of normal beliefs. Cases of delusion show that rationality is not necessary for the formation of beliefs, but they are no indication that the phenomenon of belief is more explanatorily primitive than is the phenomenon of rationality. Delusional beliefs are not explanatorily more basic than beliefs that track reality. They are, necessarily, deviations from a reality-tracking norm.

Philosophical Work for the Theory of Attention

16 9

If instances of unconscious attention are deviations from a consciousness-involving norm, and so depend on conscious states for their status as instances of attention, then it is a mistake to think that the neural correlate of consciousness can be identified by locating the neural difference between instances of unconscious attention and instances of conscious attention. Studies that locate that difference (such as those by Schurger, Cowey, Cohen, Treisman and Tallon-Baudry 2008) may tell us only about the neural correlate of a potential consciousness blocker, not about processes with an explanatory role to play in the normal bringing about of consciousness. The processes that make the difference between unconscious attention and conscious attention may be no more revealing of the explanation of consciousness than is the process that makes the difference between priests and defrocked priests revealing of the explanation of how priesthood is achieved.

7.4 PROSPECTS FOR OPTIMISM The mind does, of course, have other philosophically mysterious features, in addition to its contentfulness and its consciousness. Many of these have, at one time or another, been allocated to attention’s explanatory remit. In chapter 26 of The Principles of Psychology, William James makes one such allocation when he suggests that our theory of attention should tell us all there is to know about the controversies surrounding the phenomenon of free will: ‘The question of fact in the free-will controversy is thus extremely simple. It relates solely to the amount of effort of attention or consent which we can at any time put forth’ (1890: 1175). At an earlier time James thought that the explanatory project that gives us our theory of attention would also serve as an explanation of belief and affirmation. One of the conclusions of his early work on ‘the feeling of effort’ was that ‘attention, belief, affirmation, and motor volition, are thus four names for an identical process, incidental to the conflict of ideas alone, the survival of one in spite of the opposition of the others’ (1880: 31). At around the time that James was allocating to attention a role in the explanation of free will and of belief formation, G. F. Stout was allocating to attention a role in the explanation of thought’s relationship to language. Stout’s claim, in work that we cited in chapter 1 as an exemplar of the reasoning-based approach to attention, was that a theory of attention would ‘prepare the way’ for a treatment of ‘the function of language regarded as an instrument of thinking’ (1891: 23). A little later W. B. Pillsbury had quite other explanatory ambitions for the theory of attention. He can be found claiming in his 1908 book, Attention, that ‘taken together, the self and attention are so closely related as to be scarcely distinguishable. Conditions of attention and what we know as the self are for practical purposes identical’ (217).

170

Attention Is Cognitive Unison

More recent examples of the tendency to allocate philosophically difficult explananda to the explanatory remit of attention are no less varied. In developmental psychology, and in some philosophical works prompted by it, a number of writers have converged on the idea that a theory of joint attention should contribute to removing the mysteries that surround our capacity for knowledge of other minds (Moore and Dunham 1995; Eilan, Hoerl, McCormack and Roessler 2005). Elsewhere in philosophy we find speculation that a theory of attention should give us a more precise account of perception’s links to action (see Eilan 1998; Wu, forthcoming). We also find the suggestion that a theory accounting for attention’s role in perception will enable us to resolve the long-standing controversy about the extent to which perception is ‘cognitively penetrable’ (see Macpherson, forthcoming). In a rather different branch of philosophy the suggestion has been made (in some work of my own) that some problems with Iris Murdoch’s version of the virtue-based approach to ethics can be solved if we take seriously the idea, which Murdoch took from the works of Simone Weil, that it is forms of attention that provide the basis for the virtues (see Murdoch 1970:89; Mole 2007). Each of these suggested explanatory roles for the theory of attention deserves to be considered on its own merits. The sheer diversity of the explananda with which these various theorists have burdened attention would traditionally be taken to be an indication that, in the conception of at least some of these explanatory proposals, something might be amiss, but that strand of pessimism is owing to the process-first metaphysics that we have here been rejecting. No particular psychological process could possibly illuminate so diverse a range of philosophically puzzling phenomena. Many processes must be involved in them. But if attention is cognitive unison, then many processes can constitute it, and attention can indeed figure in all of these phenomena. Psychologists and philosophers looking to explain these phenomena, and to understand their relation to attention, should therefore look favorably on the idea that cognitive unison is what attention is.

Notes

Chapter 1 1. I suppose one might contest this, claiming that, even when the limb one is trying to move is numb, the body itself is a perceived object. The claim here is only that the example makes nonperceptual attention in action plausible.

Chapter 2 1. Couldn’t there be ‘no priority’ views, according to which the two questions get independent answers, or mutually dependent ones? I think not. Grammatical mastery alone seems to be enough to get one from an understanding of ‘attention’ to an understanding of ‘attentively’. It is therefore implausible that the two questions could receive independent answers. The idea that the two questions are mutual dependent is harder to make sense of. It seems to imply that no noncircular analysis of attentional phenomena could be given, and so we may treat it as a position of last resort. 2. Hasn’t the term ‘adverbial’ already been taken, as the name for a position in the philosophy of mind? Yes it has, but the view to which I am giving that name differs only in emphasis and application from adverbialism as traditionally conceived. Adverbialism is traditionally a tactic for avoiding commitment to problematic objects, such as sense data. The adverbialist tactic that is in play here enables us to avoid commitment to the existence of a problematic process. But the tactic operates in just the same way.

Chapter 3 1. It would take an experiment to prove that these two tasks do in fact interfere. I suppose that they do. If it were to turn out that they don’t, then an alternative example would need to be found. 2. The auditory shadowing task was widely used in earlier attention research, before a sudden and significant shift to research that focused on visual attention. The shadowing task requires subjects to repeat the words in a speech while they are hearing it (usually through headphones). Subjects are able to do this with few mistakes at surprisingly high speeds. The subjects in the Allport, Antonis, and Reynolds (1972) experiments were shadowing at 150 words per minute. In the easy shadowing task the texts were selected from ‘an anthology of humorous narrative prose containing predominantly high-frequency words’. In the difficult shadowing task, the texts were taken from ‘a textbook of early Norse history, containing a high proportion of low-frequency words’ (230).

171

17 2

Notes to Pages 50–144

Chapter 4 1. The analogy is not perfect. The dispositions of the migrating bird are typically standing dispositions. The sorts of dispositions that Ryle has in mind as characteristic of the attentive man are more transient. This does not affect the argument given. Both sorts of dispositions to ϕ can explain incidents of ϕing. 2. It is here, and not, as Ryle thought, in the analysis of attention itself that the hypothetical range of one’s performance is crucial. 3. The idea here is a development of an idea that Gareth Evans touched upon in his first treatments of the considerations that led him to propose the generality constraint on concept possession. In ‘Semantic Theory and Tacit Knowledge’, Evans considers the question of whether a rat who has become ‘bait-shy’ after being sickened by a poisoned food source deserves to be counted as believing that food from that source is poisoned. Evans makes the case against counting such a state as belief in the following passage: The rat manifests the ‘belief’ in only one way—by not eating—whereas there is no limit to the ways in which the ordinary belief that something is poisonous might be manifested. The subject might manifest it by, for example, preventing someone else from eating the food, or by giving some to a hated enemy, or by committing suicide with it. [. . .] It is of the essence of a belief state that it be at the service of many distinct projects, and that its influence on any project be mediated by other beliefs. (1985: 336–7)

Chapter 5 1. To be clear, I am officially agnostic about whether or not Lewis is right about the metaphysics of the causal relation itself. The point is only that my view, which is in the spirit of Lewis’ and is a natural accompaniment to his view, retains its appeal even if his is not a view that one wants to sign up to. 2. The distinction between supposing an extrinsic property gone by supposing an intrinsic change and supposing an extrinsic property gone by supposing an extrinsic change will break down for some cases where the extrinsic property consists in a spatial relation to just one other object. If we are to suppose that the property of being to the left of my right shoe is gone from my left shoe, we can simply imagine the shoes switched, without stipulating anything about which moved. 3. Care is needed here, for although ‘being less brightly colored than my notebook’ is not a relevant property to Ortcutt’s success as a spy, it may nonetheless be true to say that ‘Ortcutt was successful because he was less brightly colored than my notebook.’ One can use an extrinsic description as a referring expression to pick out an intrinsic property, in which context the extrinsic description has a merely reference fixing role and the sentence that appears to credit the extrinsic property with causal relevance can come out true. But in these cases the property referred to is not, despite appearances, the extrinsic property.

Chapter 7 1. In order for a description to expand a thinker’s referential repertoire it need not be that all of the components in that description are already meaningful for a thinker. Some descriptions serve as implicit definitions for terms contained

Notes to Pages 151–164

17 3

within them. What must be the case, however, is that some of the components in the description are already meaningful for the thinker. 2. One sign of the fact that what it is to be in the room, what it is to be male, what it is to be a sibling, and what it is to have a child are all explanatorily independent from unclehood is that all could have instances in a world with no uncles. Similarly, one sign of the fact that what is it to be an uncle cannot be accounted for without bringing siblings into the account is the fact that there could be no uncles in a world without siblings. These are merely signs, however; the relationship of explanatory dependence is not, I think, analyzable as equivalent to the relationship of metaphysical necessitation. 3. Alternative hypotheses have been proposed before as potential explanations of the inattentional blindness effects. Some of these have been tested empirically. One such hypothesis, discussed by Mack and Rock in their original presentation of the inattentional blindness effects, is that those effects are to be explained as consequences of fast-acting ‘inattentional amnesia’ (Wolfe 1999). A second hypothesis is that those effects are the result of ‘inattentional agnosia’ (see Simons 2000:153). A third says that inattention leads, not to a lack of consciousness, but to a lack of accessibility by the systems responsible for reasoning and the guidance of action (Block 2007). There are empirical reasons to be suspicious of the first of these hypotheses (Rees, Russell, Frith and Driver. 1999). The hypothesis that I am proposing here has more in common with the last two, and it is less easy empirically to rule it out.

Bibliography

Alexander, Henry G. (1953), ‘Paying heed’, Mind, 62 (248), 518–20. Allport, D. Alan (1980), ‘Attention and performance’, in Guy Claxton (ed.), Cognitive Psychology: New Directions (London: Routledge and Kegan Paul) 112–53. ——— (1987), ‘Selection for action: Some behavioral and neurophysiological considerations of attention and action’, in H. Heuer and A. F. Sanders (eds.), Perspectives on Perception and Action (Hillsdale, NJ: Erlbaum), 395–419. ——— (1992), ‘Attention and control: Have we been asking the wrong questions? A critical review of twenty five years’, in Sylvan Kornblum and David Mayer (eds.), Attention and Performance 14: Synergies in Experimental Psychology, Artificial Intelligence and Cognitive Neuroscience (Cambridge, MA: MIT Press), 183–218. Allport, D. Alan, Antonis, Barbara, and Reynolds, Patricia (1972), ‘On the division of attention: A disproof of the single channel hypothesis’, Quarterly Journal of Experimental Psychology, 24 (2), 225–35. Allport, D. Alan, Styles, Elizabeth A., and Hsieh, Shulan (1994), ‘Shifting intentional set: Exploring the dynamic control of tasks’, in Carlo Umiltà and Morris Moscovitch (eds.), Attention and Performance 15: Conscious and Nonconscious Information Processing (Cambridge, MA: MIT Press), 421–52. Andersen, Richard A., and Buneo, Christopher A. (2002) ‘Intentional Maps in Posterior Parietal Cortex’ Annual Review of Neuroscience, 25 (1), 189–220. Austin, James, H. (2009), Selfless Insight: Zen and the Meditative Transformations of Consciousness (Cambridge, MA: MIT Press). Baddeley, Alan D. (1966), ‘The capacity for generating information by randomization’, Quarterly Journal of Experimental Psychology, 18 (2), 119–29. ——— (1990), Human Memory: Theory and Practice Revised Edition (Hove: Erlbaum). Bahrami, Bahador (2003), ‘Object property encoding and change blindness in multiple object tracking’, Visual Cognition, 10 (8), 949–63. Bain, Alexander (1888), The Emotions and the Will, 3rd ed. (London: Longmans, Green and Co.). Bealer, George (1998), ‘Propositions’, Mind, 107 (425), 1–32. Behrmann, Marlene, and Geng, Joy J. (2002), ‘What is ‘left’ when all is said and done? Spatial coding and hemispatial neglect’, in Hans-Otto Karnath, David Milner, and Giuseppe Vallar (eds.), The Cognitive and Neural Bases of Spatial Neglect (New York: Oxford University Press), 85–100. Bennett, Karen (2003), ‘Why the exclusion problem seems intractable, and how, just maybe, to tract it’, Nous, 37 (3), 471–97. Berkeley, George (1710), Philosophical Works; Including the Works on Vision. Reprint, ed. Michael Ayers (Everyman Library; London: Dent, 1975).

174

Bibliography

175

Berti, Anna, and Rizzolatti, Giacomo (1992), ‘Visual processing without awareness: Evidence from unilateral neglect’, Journal of Cognitive Neuroscience, 4 (4), 345–51. Block, Ned (1998), ‘Is experiencing just representing?’, Philosophy and Phenomenological Research, 58 (3), 663–70. ——— (2003), ‘Do causal powers drain away?’, Philosophy and Phenomenological Research, 67 (1), 133–50. ——— (2007), ‘Consciousness, accessibility and the mesh between psychology and neuroscience’, Behavioral and Brain Sciences, 30 (5–6), 481–499. Botvinick, Matthew M., Braver, Todd, Barch, Deanna, Carter, Cameron and Cohen, Jonathan. (2001), ‘Conflict monitoring and cognitive control’, Psychological Review, 108 (3), 624–52. Bradley, F. H. (1886), ‘Is there any special activity of attention?’, Mind, 11 (43), 305–23. ——— (1902), ‘On active attention’, Mind, 11 (41), 1–30. Brandom, Robert B. (1994), Making It Explicit: Reasoning, Representing, and Discursive Commitment (Cambridge, MA: Harvard University Press). Brewer, Bill (2005), ‘Perceptual experience has conceptual content’, in Ernest Sosa and Matthias Steup (eds.), Contemporary Debates in Epistemology (Oxford: Blackwell), 217–30. Bricolo, Emanuela, Gianesini, Tiziana, Fanini, Alessandra, Bundesen, Claus, and Chelazzi, Leornardo (2002), ‘Serial attention mechanisms in visual search: A direct behavioural demonstration’, Journal of Cognitive Neuroscience, 14 (7), 980–93. Broadbent, D. E. (1958), Perception and Communication (Elmsford, NY: Pergamon Press). Bundesen, Claus, and Habekost, Thomas (2008) Principles of Visual Attention: Linking Mind and Brain (Oxford: Oxford University Press). Byrne, Alex (2005), ‘Perception and conceptual content’, in Ernest Sosa and Matthias Steup (eds.), Contemporary Debates in Epistemology (Oxford: Blackwell), 231–50. Campbell, John (2002), Reference and Consciousness (Oxford: Oxford University Press). Chisholm, Roderick (1942), ‘The problem of the speckled hen’, Mind, 51 (204), 368–73. Chomsky, Noam (1957), Syntactic Structures (Oxford: Mouton). ——— (1959), ‘Review of B.F. Skinner Verbal Behavior’, Language, 35 (1), 26–58. Cohen, Jonathan D., Dunbar, Kevin, and McClelland, James L. (1990), ‘On the control of automatic processes: A parallel distributed processing account of the Stroop effect’, Psychological Review, 97 (3), 332–361. Cyples, William (1880), An Inquiry into the Process of Human Experience (London: Strahan and Co.). Dashiell, John (1926), Fundamentals of Objective Psychology (London: George Allen and Unwin). Davidson, Donald (1980), ‘Mental events’, reprinted in his Essays on Actions and Events (Oxford: Oxford University Press), 207–25. De Muijnck, Wim (2002), ‘Causation by relational properties’, Grazer Philosophische Studien, 65, 123–37.

176

Bibliography

Desimone, Robert, and Duncan, John (1995), ‘Neural mechanisms of selective visual attention’, Annual Review of Neuroscience, 18 (1), 193–222. Dowe, Phil (2000), Physical Causation (New York: Cambridge University Press). Driver, John (1998), ‘The neuropsychology of spatial attention’, in Harold Pashler (ed.), Attention (Hove: Psychology Press), 297–340. Duncan, John (1996), ‘Cooperating brain systems in selective perception and action’, In T. Inui and J. L. McClelland (eds.), Attention and performance 16: Information integration in perception and communication (Cambridge, MA: MIT Press), 549–78. Duncan, John (2006), ‘Brain mechanisms of attention’, Quarterly Journal of Experimental Psychology, 59 (1), 2–27. Eames, Elizabeth (1984), ‘Introduction’, in Bertrand Russell, Theory of Knowledge: The 1913 Manuscript (London: Routledge), vii–xxxvii. Eilan, Naomi M. (1998), ‘Perceptual intentionality, attention and consciousness’, in Anthony O’Hear (ed.), Current Issues in Philosophy of Mind (Cambridge: Cambridge University Press), 181–202. Eilan, Naomi, Hoerl, Christoph, McCormack, Teresa, and Roessler, Johannes., eds. (2005), Joint Attention: Communication and Other Minds (Oxford: Oxford University Press). Evans, Gareth, (1985), ‘Semantic theory and tacit knowledge’, in Collected Papers (Oxford: Oxford University Press), 322–42. Fougnie, Daryl, and Marois, Rene (2009), ‘Attentive tracking disrupts feature binding in visual working memory’, Visual Cognition, 17 (1–2), 48–66. Freeman, Elliot (2005), ‘Lateral interactions between targets and flankers require attention’, in Laurent Itti, Geraint Rees, and John K. Tsotsos (eds.), Neurobiology of Attention (San Diego, CA: Elsevier Academic Press), 477–84. Freeman, Elliot, Sagi, Dov, and Driver, Jon (2001), ‘Lateral interactions between targets and flankers in low-level vision depend on attention to the flankers’, Nature Neuroscience, 4 (10), 1032–36. Friedman-Hill, Stacia R., Robertson, Lynn C., and Treisman, Anne M. (1995), ‘Parietal contributions to visual feature binding: Evidence from a patient with bilateral lesions’, Science, 269 (5225), 853–55. Geach, Peter (1957), Mental Acts: Their Content and Their Objects (New York: Humanities Press). Goldberg, Michael E., Bisley, James, Powell, Keith D., Gottlieb, Jacqueline, and Kusunoki, Makoto (2002), ‘The role of the lateral intraparietal area of the monkey in the generation of saccades and visuospatial attention’, Annals of the New York Academy of Sciences, 956, 205–15. Groos, Karl (1896), The Play of Animals, trans. Elizabeth L. Baldwin (London: Chapman and Hall). Harman, Gilbert H. (1965), ‘The inference to the best explanation’, Philosophical Review, 74 (1), 88–95. Heil, John, and Mele, Alfred, eds. (1993), Mental Causation (Oxford: Clarendon Press). Helmholtz, Hermann von (1870), Die Lehre von den Tonempfindungen, als Physiologishe Grundlage fur dir Theorie der Musik, 3rd ed. (Braunschweig: Druck und Verlag von Friedrich Vieweg und Sohn). Hommel, Bernhard (2004), ‘Event files: Feature binding in and across perception and action’, Trends in Cognitive Sciences, 8 (11), 494–500.

Bibliography

17 7

Hommel, Bernhard, and Colzato, Lorenza (2009), ‘When an object is more than a binding of its features: Evidence of two mechanisms of visual feature integration’, Visual Cognition, 17 (1/2), 120–40. Honderich, Ted (1982), ‘The argument for anomalous monism’, Analysis, 42 (January), 59–64. Jahanshahi, Marjan, Saleem, T., Ho, Aileen K., Dirnberger, G., and Fuller, R. (2006), ‘Random number generation as an index of controlled processing’, Neuropsychology, 20 (4), 391–99. James, William (1879), ‘Are we automata?’, Mind, 4 (13), 1–22. ——— (1880), ‘The feeling of effort’. Reprint, Anniversary Memoirs of the Boston Society of Natural History (Boston: Boston Society of Natural History). ——— (1890), The Principles of Psychology. Reprint (Cambridge, MA: Harvard University Press, 1981). Jaśkowski, Piotr, van der Lubbe, Rob H. J., Schlotterbeck, Erik, and Verleger, Rolf (2002), ‘Traces left on visual selective attention by stimuli that are not consciously identified’, Psychological Science, 13 (1), 48–54. Jiang, Yi, Costello, Patricia, Fang, Fang, Huang, Miner, and He, Sheng (2006), ‘A gender- and sexual orientation-dependent spatial attentional effect of invisible images’, Proceedings of the National Academy of Sciences of the United States of America, 103 (45), 17048–52. Johnston, Mark (1997), ‘Manifest kinds’, Journal of Philosophy, 94 (11), 564–83. Kaplan, David (1989), ‘Demonstratives’, in Joseph Almog, John Perry, and Howard Wettstein (eds.), Themes from Kaplan (New York: Oxford University Press), 481–563. Kastner, Sabine, and Ungerleider, Leslie G. (2000), ‘Mechanisms of visual attention in the human cortex’, Annual Review of Neuroscience, 23, 315–41. Kastner, Sabine, and Ungerleider, Leslie G. (2001), ‘The neural basis of biased competition in human visual cortex’, Neuropsychologia 38 (12), 1263–76. Kentridge, Robert W., and Heywood, Charles A. (2001), ‘Attention and alerting: Cognitive processes spared in blindsight’, in Beatrice De Gelder, Edward H. F. De Haan, and Charles A. Heywood (eds.), Out of Mind: Varieties of Unconscious Processes (New York: Oxford University Press), 163–81. Kentridge, Robert W., Heywood, Charles A., and Weiskrantz, Lawrence (1999), ‘Attention without awareness in blindsight’, Proceedings of the Royal Society of London Series B, 266, 1805–11. Kentridge, Robert W., Nijboer, Tanja C. W., and Heywood, Charles A. (2008), ‘Attended but unseen: Visual attention is not sufficient for visual awareness’, Neuropsychologia, 46 (3), 864–69. Khoe, W., Freeman, E., Woldorff, M. G., and Mangun, G. R. (2006), ‘Interactions between attention and perceptual grouping in human visual cortex’, Brain Research, 1078 (1), 101–11. Kim, Jaegwon (1998), Mind in a Physical World (Cambridge, MA: MIT Press). Knudsen, Eric I. (2007), ‘Fundamental components of attention’, Annual Review of Neuroscience, 30, 57–78. Koch, Christof, and Tsuchiya, Naotsugu (2007), ‘Attention and consciousness: Two distinct brain processes’, Trends in Cognitive Sciences, 11 (1), 16–22. ——— (2008), ‘Response to Mole: Subjects can attend to completely invisible objects’, Trends in Cognitive Sciences, 12 (2), 44–45. Kripke, Saul A. (1982), Wittgenstein on Rules and Private Language: An Elementary Exposition (Cambridge, MA: Harvard University Press).

17 8

Bibliography

Lavie, Nilli (1995), ‘Perceptual load as a necessary condition for selective attention’, Journal of Experimental Psychology: Human Perception and Performance, 21 (3), 451–68. ——— (2001), ‘Capacity limits in selective attention: Behavioural evidence and implications for neural activity’, in Jochen Braun, Christof Koch, and Joel Davis (eds.), Visual Attention and Cortical Circuits (Cambridge, MA: MIT Press), 49–68. Lavie, Nilli, and De Fockert, Jan W. (2005), ‘The role of working memory in attentional capture’, Psychonomic Bulletin and Review, 12 (4). Lavie, Nilli, and Tsal, Yehoshua (1994), ‘Perceptual load as a major determinant of the locus of selection in visual attention’, Perception and Psychophysics, 56 (2), 183–97. Lewis, David (1973a), ‘Causation’, Journal of Philosophy, 70, 556–67. ——— (1973b), Counterfactuals (Oxford: Blackwell). ——— (1979), ‘Counterfactual Dependence and Time’s Arrow’, reprinted (1986) in his Philosophical Papers, Volume II. (Oxford: Oxford University Press), 32–66. ——— (2000), ‘Causation as influence’, Journal of Philosophy, 97 (4), 182–97. Lycan, William G. (1996), ‘Layered perceptual representation’, Philosophical Issues, 7, 81–100. Macdonald, Cynthia (2006), ‘The metaphysics of mental causation’, Journal of Philosophy, 103, 539–76. Mack, Arien, and Rock, Irwin (1998), Inattentional Blindness (Cambridge, MA: MIT Press). Mackie, J. L. (1974), The Cement of the Universe: A Study of Causation (Oxford: Clarendon Press). Macpherson, Fiona (forthcoming), ‘Cognitive penetration of colour experience: Rethinking the issue in light of an indirect mechanism’, Philosophy and Phenomenological Research. Memmert, Daniel (2006), ‘The effects of eye movements, age, and expertise on inattentional blindness’, Consciousness and Cogntion, 15 (3), 620–27. Mole, Christopher (2007), ‘Attention, self and The Sovereignty of Good’, in Anne Rowe (ed.), Iris Murdoch: A Reassessment (Basingstoke: Palgrave Macmillan) 72–84. ——— (2008), ‘Attention in the Absence of Consciousness?’, Trends in Cognitive Sciences, 12 (2), 44. Moore, Chris, and Dunham, Philip J., eds. (1995), Joint Attention: Its Origins and Role in Development (Hove: Erlbaum). Moore, Tirin, and Armstrong, Katherine M. (2003), ‘Selective visual gating of visual signals by microstimulation of frontal cortex’, Nature, 421 (6921), 370–73. Most, Steven B., et al. (2001), ‘How not to be seen: The contribution of similarity and selective ignoring to sustained inattentional blindness’, Psychological Science, 12 (1), 9–17. Mozer, Michael C. (1991), The Perception of Multiple Objects: A Connectionist Approach. (Cambridge, MA: MIT Press). Mozer, Michael C., and Sitton, Mark. (1998), ‘Computational modeling of spatial attention’, in Harold. Pashler (ed.), Attention (Hove: Psychology Press), 341–393.

Bibliography

17 9

Murdoch, Iris (1970), The Sovereignty of The Good. Reprint (London: Routledge, 2001). Münsterberg, Hugo, and Kozaki, Nariaki (1894), ‘Studies from the Harvard Psychological Laboratory: The intensifying effect of attention’, Psychological Review, 1 (1), 39–44. Navon, D. (1985), ‘Attention division or attention sharing?’, in M. I. Posner and S. M. Marin (eds.), Attention and performance XI (London: Erlbaum.) 133–146. Navon, D., and Miller, J. (1987). Role of outcome conflict in dual-task interference. Journal of Experimental Psychology: Human Perception and Performance, 13, 435–448. Olk, B., Harvey, M., Dow, L., and Murphy, P. J. S. (2001), ‘Illusion processing in hemispatial neglect’, Neuropsychologia, 39 (6), 611–25. O’Regan, J. Kevin, and Noë, Alva (2001), ‘A sensorimotor account of vision and visual consciousness’, Behavioral and Brain Sciences, 24, 939–1031. Parasuraman, Raja (1998), ‘The Attentive Brain: Issues and Prospects’ in Raja Parasuraman (ed.) The Attentive Brain (Cambridge, MA: MIT Press) 3–15. Pashler, Harold (1997), The Psychology of Attention (Cambridge, MA: MIT Press). Passingham, Richard E. (1998), ‘Attention to action’, in Angela C. Roberts, Trevor W. Robbins, and Lawrence Weiskrantz (eds.), The Prefrontal Cortex: Executive and Cognitive Functions (New York: Oxford University Press), 131–43. Peacocke, Christopher (1998), ‘Conscious attitudes, attention and self knowledge’, in C. McDonald, B. Smith, and C. Wright (eds.), Knowing Our Own Minds: Essays on Self Knowledge (Oxford: Oxford University Press) 63–99. Peelen, Marius V., Fei-Fei, Li, and Kastner, Sabine (2009), ‘Neural mechanisms of rapid natural scene categorization in human visual cortex’, Nature, 460 (7251), 94–97. Pessoa, L., McKenna, M., Gutierrz, E. and Ungerleider, L. G. (2002), ‘Neuroal processing of emotional faces requires attention’ Proceedings of the National Academy of Sciences of the United States of America, 99 (17), 11458–63. Pillsbury, W. B. (1908), Attention (London: George Allen and Unwin). Posner, Michael (1994), ‘Attention: The mechanisms of consciousness’, Proceedings of the National Academy of Sciences of the United States of America, 91, 7398–403. Prinz, Jesse J. (forthcoming), ‘Is attention necessary and sufficient for consciousness?’, in Christopher Mole, Declan Smithies, and Wayne Wu (eds.), Attention: Philosophical and Psychological Essays (Oxford: Oxford University Press). Rees, Geraint, Russell, Charlotte, Frith, Christopher D. and Driver, Jon (1999), ‘Inattentional blindness versus inattentional amnesia for fixated but ignored words’, Science, 286 (5449), 2504–7. Rensink, Ronald (2002), ‘Change detection’, Annual Review of Psychology, 53, 4245–77. Rensink, Ronald A., and Enns, James (1995), ‘Preemption effects in visual search: Evidence for low-level grouping’, Psychological Review, 102 (1), 101–30. Rensink, Ronald A., O’Regan, J. Kevin, and Clark, James J. (1997), ‘To see or not to see: The need for attention to perceive changes in scenes’, Psychological Science, 8 (5), 368–73.

18 0

Bibliography

Reynolds, John H., and Desimone, Robert (2000), ‘Competitive mechanisms subserve selective visual attention’, in Alec Marantz, Yasushi Miyashita, and Wayne O’Neil (eds.), Image, Language, Brain: Papers from the First Mind Articulation Project Symposium (Cambridge, MA: MIT Press), 233–47. Reynolds, John H., Gottlieb, Jacqueline P., and Kastner, Sabine (2002), ‘Attention’, in Larry R. Squire, et al. (eds.), Fundamental Neuroscience, 2nd ed. (London: Elsevier Science), 1249–73. Ribot, Théodule (1889), Psychologie de l’attention (Paris: F. Alcan). Ro, Tony, and Rafal, Robert D. (1996), ‘Perception of geometric illusions in hemispatial neglect’, Neuropsychologia, 34 (10), 973–78. Ro, Tony, Russell, Charlotte, and Lavie, Nilli (2001), ‘Changing faces: A detection advantage in the flicker paradigm’, Psychological Science, 12 (1), 94–99. Russell, Bertrand (1913), Theory of Knowledge: The 1913 Manuscript. Reprint, ed. Elizabeth Eames (Oxford: Routledge, 1984). Ryle, Gilbert (1949), The Concept of Mind (Harmondsworth: Penguin Books). ——— (1971), Collected Papers, Volume II: Collected Essays 1929–1968 (London: Hutchinson). Schacter, Daniel L., Wagner, Anthony D., and Buckner, Randy L. (2000), Memory Systems of 1999 (New York: Oxford University Press). Schaffer, Jonathan (2004), ‘Counterfactuals, causal independence and conceptual circularity’, Analysis, 64 (4), 299–308. Schneider, W., and Detwiler, M. (1988), ‘The role of practice in dual-task performance: Toward workload modeling in a connectionist/control architecture’, Human Factors, 30, (5), 539–66. Scholl, Brian J., Pylyshyn, Zenon W., and Feldman, Jacob. (2001), ‘What is a visual object? Evidence from target merging in multiple object tracking’, Cognition, 80, 159–77. Schurger, Aaron, Cowey, Alan, Cohen, Jonathan D., Treisman, Anne and TallonBaudry, Catherine (2008), ‘Distinct and independent correlates of attention and awareness in a hemianopic patient’, Neuropsychologia, 46, 2189–97. Siegel, Susanna (2010), ‘Do visual experiences have contents?’, in Bence Nanay (ed.), Perceiving the World (Oxford: Oxford University Press). Simons, Daniel J. (2000), ‘Attentional capture and inattentional blindness’, Trends in Cognitive Sciences, 4 (4), 147–155. Simons, Daniel J., and Chabris, Christopher F. (1999), ‘Gorillas in our midst: Sustained inattentional blindness for dynamic events’, Perception, 28 (9), 1059–74. Skinner, B. F. (1957), Verbal Behavior (Century Psychology Series; East Norwalk, CT: Appleton Century Crofts). Smith, Peter K. (1982), ‘Bad news for anomalous monism?’, Analysis, 42 (October), 220–24. Snyder, Larry H., Batista, Aaron P., and Andersen, Richard A. (2000), ‘Intentionrelated activity in the posterior parietal cortex: A review’, Vision Research, 40 (10–12), 1433–41. Speaks, Jeff (2005), ‘Is there a problem about nonconceptual content?’, Philosophical Review, 114 (3), 359–98. Spelke, Elizabeth S., Hirst, William, and Neisser, Ulric (1976), ‘Skills of divided attention’, Cognition, 4, 215–30. Stalnaker, Robert (1996), ‘On a defense of the hegemony of representation’, Philosophical Issues, 7, 101–8.

Bibliography

18 1

Stout, G. F. (1891), ‘Apperception and the movement of attention’, Mind, 16 (61), 25–53. Strawson, Peter F. (1959), Individuals: An Essay in Descriptive Metaphysics (London: Methuen). Stroud, Barry (2000), The Quest for Reality: Subjectivism and the Metaphysics of Colour (Oxford: Oxford University Press). Styles, Elizabeth A. (2006), The Psychology of Attention, 2nd ed. (Hove: Taylor and Francis). Sully (1880), ‘Review of William Cyples An Inquiry into the Process of Human Experience’, Mind, 5 (18), 273–80. Titchener, E. B. (1908), Lectures on the Elementary Psychology of Feeling and Attention (New York: Macmillan). ——— (1910), ‘Attention as sensory clearness’, Journal of Philosophy, Psychology and Scientific Methods, 7 (7), 180–82. Treisman, Anne M. (1998), ‘The perception of features and objects’, in Alan Baddeley and Lawrence Weiskrantz (eds.), Attention: Selection, Awareness and Control (New York: Oxford University Press), 5–35. ——— (2003), ‘Consciousness and perceptual binding’, in Axel Cleermans (ed.), The Unity of Consciousness: Binding, Integration and Dissociation (Oxford: Oxford University Press), 95–113. Treisman, Anne M., and Gelade, Garry (1980), ‘A feature-integration theory of attention’, Cognitive Psychology, 12 (1), 97–136. Treisman, Anne M., and Gormican, Stephen (1988), ‘Feature analysis in early vision: evidence from search asymmetries’, Psychological Review, 95 (1), 15–48. Tsotsos, John K., Itti, Laurent, and Rees, Geraint (2005), ‘A brief and selective history of attention’, in Laurent Itti, Geraint Rees, and John K. Tsotsos (eds.), Neurobiology of Attention (San Diego: Elsevier Academic Press), xxiii–xxxii. Tye, Michael (2002), ‘Representationalism and the transparency of experience’, Noûs, 36 (1), 137–51. ——— (2009), Consciousness Revisited: Materialism without Phenomenal Concepts (Cambridge, MA: MIT Press). vanMarle, Kristy, and Scholl, Brian J. (2003), ‘Attentive tracking of objects versus substances’, Psychological science, 14 (5), 498–504. Von Campenhausen, Christoph, and Schramme, Jürgen (1995), ‘100 years of Benham’s top in colour science’, Perception, 24 (6), 695–717. Ward, James (1885), ‘Psychology’, Encyclopædia Britannica 9th ed. (Edinburgh: Adam and Charles Black). ——— (1918), Psychological Principles (Cambridge: Cambridge University Press). Wegener, Detlef, Ehn, Friederike, Aurich, Maike K., Galashan, F. Orlando, Kreiter, Andreas (2008), ‘Feature-based attention and the suppression of non-relevant object features’, Vision Research, 48, 2696–707. White, Alan R. (1964), Attention (Oxford: Basil Blackwell). Wittgenstein, Ludwig (1953), Philosophical Investigations (New York: Macmillan). ——— (1967), Zettel, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe (Oxford: Basil Blackwell). ——— (1980), Remarks on the Philosophy of Psychology, vol. 2, ed. G. H. von Wright and Heikki Nyman, trans. C. G. Luckhardt and M. A. E. Aue (Oxford: Blackwell).

18 2

Bibliography

Wolfe, Jeremy M. (1999), ‘Inattentional amnesia’, in Veronika Coltheart (ed.), Fleeting Memories: Cognition of Brief Visual Stimuli (MIT Press/Bradford Books Series in Cognitive Psychology; Cambridge, MA: MIT Press), 71–94. Wu, Wayne (forthcoming), ‘What is conscious attention?’, Philosophy and Phenomenological Research. Wundt, Wilhelm (1897), Outlines of Psychology (Leipzig: Wilhelm Engelmann). Yablo, Stephen (1992), ‘Mental causation’, Philosophical Review, 102 (2), 245–80. ——— (2003), ‘Causal relevance’, Philosophical Issues, 13 (1), 316–28.

Index

abduction, see ‘inference to the best explanation’ acquaintance, 142 adverbialism, 26–30, 47–50, 117, 154, 171 Alexander, H. G., 70 Allport, D. Alan, 44–45, 120, 155 amygdala, 128 Andersen, Richard, 129–131 arithmetic, 5, 63–64 attention, agent vs. spectator, 53, 72–74 causes and effects of, 5–6, 91 and consciousness, 155–169 degrees of, 83–85 divided, 42–45, 74–82 and free will, 169 involuntary, 53 motor theory of, 69 in non-humans, 54–56, 69, 167 object-directed, 72–73, 143, 146–154 partial, 85–87 privative character of, 70, 76–77, 109–112 and reference, 136–155 attentional blink, 10 auditory shadowing, 44, 73, 76, 171 background conditions, 91–92, 112–113 ‘background set’, 51, 66–67, 61–62 Bahrami, Bahador, 147 behaviourism, 15–16, 18 Benham’s Top, 147 Bennett, Karen, 112 Berkeley, George, 147 biased competition, 133–135 blindsight, 167–168 Bundesen, Claus, 123, 131 Bradley, F. H., 6–8, 11–15, 22, 24, 47

Bricolo, Emanuela, 131 Brandom, Robert, 139 ‘Brentano’s Problem’, 136, 138 Broadbent, Donald, 17–19, 120, 131, 155 Campbell, John, 142–143 capacity limitations, 131–132 causal exclusion, 112–115 causal relevance counterfactual theory of, 92–96 of extrinsic properties, 106–109 of privative properties, 109–110 change blindness, 162, 164 children, 54, 71 circular explanations, see ‘explanations, circular’ cognitive unison, definition of, 51 theory of attention, 65 cognitive processes, 57–60, 166 collective agency, 166 combustion, 25, 28 composite tasks, see ‘tasks, composite’ consciousness, 136, 155, 169 content, see ‘Brentano’s Problem’ counterfactuals, backtracking interpretations of, 100–101 and causally relevant properties, 94–96 intrinsic and extrinsic change readings of, 107–109 color, 138–139, 147, 149 Cyple, William, 14 Dashiell, John, 16 Davidson, Donald, 94 daydreams, 57

183

18 4

degrees of attention, see attention, degrees of Descartes, Rene, vi, 88 Desimone, Robert, 134 diamond, metaphysics of, 32 divided attention, see attention, divided driving, 66, 74, 114–115 Duncan, John, 123, 133–135 ecological validity, 126–127 electroencephalography (EEG), 122, 127 eliminativism, 89, 115 emotion, facial expressions of, 128 employment, 12–13, 119 enabling conditions, 92 Evans, Gareth, 172 Events, properties of, 29, 96 explanation, 24–26, adverbialist, 26 causal vs. non causal, 151 circular, 151, 171 levels of, 32 limitations on, 49, 121, modes and, 155 process-first, 26–30, 117 extrinsicness, causal relevance of, 106–109 of unison, 103–104 eye movements, see ‘saccades’ faces, 128, 161, 162 failsafe mechanisms, (see ‘preemption’) feature-binding, 36–38, 153, 156, 163 forms of life, 56, 167 Freeman, Elliot, 122–127, 135

Index

imagination, 10 impossible tasks, (see ‘tasks, impossible’) inattention (see also partial attention) causal relevance of, 110–112 in behaviour, 114 as a state of consciousness, 156–158 inattentional blindness, 160–162, 164, 173 induction, 126, 127–128 inference to the best explanation, 40, 121, 159–160, 164 informatics, 17 interocular suppression, 167–168 involuntary attention, see ‘attention, involuntary’ Itti, Laurent, 123 jade, metaphysics of, 30, 124 James, William, analysis of attention, 3–5 on consciousness and attention, 155 disjunctive theory of attention, 8–10 explanatory approach of, 24 on free-will and attention, 169 on inattentiveness, 57, 156–158, 162 similarity to Broadbent, 19, 23 Judd illusion, 38 Kastner, Sabine, 132, 134 kinds, natural, 30–31 Knudsen, Eric, 123 Kripke, Saul, 139

Geach, Peter, 20 gist perception, 162, 164 Goldberg, Michael, 129–131 Gottlieb, Jacqueline, 129, 132

Lateral Intraparietal Cortex (LIP), 128–130 Lavie, Nilli, 68–69, 132 Lewis, David, 94–102, 172 life, explanation of, 144–145, 156 logical positivism, 15, 17–18 looking, 53, 76, 77

Habekost, Thomas, 123 haste, explanation of, 26, 28, 31, 119, 121 causal relevance of, 98–101

Mach, Arien, 160–162 mastery, 62, 72, 74, 114 meditation, 71 memory, 31, 60 working, 67–69, 76

Index

metaphysics preliminary definition of, vii-viii of counterfactuals, 93 in psychology, 119, 122–124, 133–134 explanatory consequences of, 125–126 migration, 49–50, 119, 172 modes, 165 ‘mongrel categorials’, 21, 49–50, 119 motor theory of attention, see ‘attention, motor theory of’ multiple realizability, 27, 30–31, 124, 127 Müller-Lyer illusion, 38 natural kinds, see ‘kinds, natural’ natural language, 20–21 neglect, 38 object files, 37, 153–154, 163 object tracking, 147 partial attention, see ‘attention, partial’ Pashler, Harold, 10 Peacocke, Christopher, 41–45 Pessoa, Luis, 128 Pillsbury, W. B., 6, 169 Posner, Michael, 156 ‘potential service’, 60–61, 105 practice, 78, 119–121, 124 preemption, 97–99 Prinz, Jesse, 156 privative adjectives, 86 private language argument, 140–141 process-first explanation, see ‘explanation, process-first’ random number generation, 39 ‘redintegration’, 12 Rees, Geraint, 123 refrigerator light illusion, 159 Reynolds, John, 132, 133–135 Ribot, Théodule, 16 Rock, Irwin, 160–162 Russell, Bertrand, 142, 145

185

Ryle, Gilbert, opposition to simple behaviourism of, 16, mongrel categorial theory of ‘heed concepts’, 20–21, 48, 70, 119, 172 explanation and category errors, 128–130 saccades, 69, 129 selectivity, 132, 133–134, 164 sight reading, 44, 75, 76 skill, 78, 85, 90, 102 (see also, ‘mastery’) sound, 10, 53, 57, 68, 162–3 Stout, G. F., 6, 169 Skinner, B. F., 17 Strawson, Sir Peter, 144, 145 subitizing, 4 substance, 27, 30, 32–33 superfluous competence, 64 superordinate tasks, 63–67, 81 supervenience, 32–33, 124–125 tasks, 51, 52–54, 73, 166 composite, 81–82 impossible, 61–62 overlearned, 62 superordinate, 63–67, 81 ‘task set’, 52 task switching, 39 taxonomies, 27–29 timescale, 127–128 trainee clairvoyant (see ‘tasks, impossible’) Treisman, Anne, 36–38, 156, 163 Tsotsos, John, 123 understanding, guidance by, 52 of simple creatures, 54–56 and cognitive versatility, 55–56 of composite tasks, 81–82 by collective agents, 167 Ungerleider, Leslie, 128, 134 visual search, 36–37, 80–82 Ward, James, 7 Wegener, Detlef, 147

18 6

White, Alan, on agent vs. spectator attention, 73 critique of Gilbert Ryle, 20–22, 49–50 theory of attention, 22, 48, 50, 87, 119 Wittgenstein, Ludwig, on forms of life and understanding, 56

Index

doubts about explanatory power of attention, 20, 142, 145 on reference and attention, 139–141, 154 working memory (see, ‘memory, working’) Wundt, Wilhelm, 118–120 Yablo, Stephen, 88, 112–115