209 121 743KB
English Pages 183 Year 2008
201 27 3MB Read more
What is an epistemic virtue? Are epistemic virtues reliable? Are they motivated by a love of truth? Do epistemic virtues
204 21 3MB Read more
"This volume features new findings by nine interdisciplinary teams of researchers on the topics of self, motivation
211 72 980KB Read more
This book examines all verses of the Quran involving knowledge related concepts. It begins with the argument that an ana
180 112 2MB Read more
180 4 2MB Read more
Charles S. Peirce, the founder of pragmatism, was also the architect of a remarkable theory of signs that continues to p
177 43 159KB Read more
'Knowledge First' constitutes what is widely regarded as one of the most significant innovations in contempora
315 5 2MB Read more
What Is Scientific Knowledge? is a much-needed collection of introductory-level chapters on the epistemology of science.
239 97 1MB Read more
Table of contents :
Table of Contents......Page 8
1. Introduction......Page 10
2. Competing Conceptions of a Cognitive Virtue......Page 26
3. A Virtue Account of Perception......Page 50
4. A Virtue Account of Memory......Page 68
5. Testimony and Cognitive Virtue......Page 86
6. Moral Expertise: The Role of Cognitive Virtues in Generating Knowledge......Page 115
7. Cognitive Virtue, Divine Hiddenness, and Reasonable Belief or Nonbelief......Page 138
8. Conclusion......Page 150
Continuum Studies in Philosophy Series Editor: James Fieser, University of Tennessee at Martin, USA Continuum Studies in Philosophy is a major monograph series from Continuum. The series features first-class scholarly research monographs across the whole field of philosophy. Each work makes a major contribution to the field of philosophical research. Aesthetic in Kant, James Kirwan Analytic Philosophy: The History of an Illusion, Aaron Preston Aquinas and the Ship of Theseus, Christopher Brown The Challenge of Relativism, Patrick Phillips Demands of Taste in Kant’s Aesthetics, Brent Kalar Descartes and the Metaphysics of Human Nature, Justin Skirry Descartes’ Theory of Ideas, David Clemenson Dialectic of Romanticism, Peter Murphy and David Roberts Hegel’s Philosophy of Language, Jim Vernon Hegel’s Philosophy of Right, David James The History of Intentionality, Ryan Hickerson Kierkegaard’s Analysis of Radical Evil, David A. Roberts Leibniz Re-interpreted, Lloyd Strickland Metaphysics and the End of Philosophy, HO Mounce Nietzsche and the Greeks, Dale Wilkerson Origins of Analytic Philosophy, Delbert Reed Philosophy of Miracles, David Corner Platonism, Music and the Listener’s Share, Christopher Norris Popper’s Theory of Science, Carlos Garcia Role of God in Spinoza’s Metaphysics, Sherry Deveaux Rousseau and the Ethics of Virtue, James Delaney Rousseau’s Theory of Freedom, Matthew Simpson Spinoza and the Stoics, Firmin DeBrabander Spinoza’s Radical Cartesian Mind, Tammy Nyden-Bullock St. Augustine and the Theory of Just War, John Mark Mattox St. Augustine of Hippo, R.W. Dyson Thomas Aquinas & John Duns Scotus, Alex Hall Tolerance and the Ethical Life, Andrew Fiala Virtue Epistemology: Motivation and Knowledge, Stephen Napier
Virtue Epistemology Motivation and Knowledge
Continuum International Publishing Group The Tower Building 11 York Road London SE1 7NX
80 Maiden Lane Suite 704 New York NY 10038
www.continuumbooks.com © Stephen Napier 2008 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN-10: HB: 0-8264-9794-2 ISBN-13: HB: 978-0-8264-9794-9 Library of Congress Cataloging-in-Publication Data Napier, Stephen. Virtue epistemology: motivation and knowledge/Stephen Napier. p. cm. Includes bibliographical references. ISBN 978-0-8264-9794-9 1. Virtue epistemology. 2. Knowledge, Theory of. I. Title. BD176.N37 2008 121—dc22
Typeset by Newgen Imaging Systems Pvt Ltd, Chennai, India Printed and bound in Great Britain by Biddles Ltd, King’s Lynn, Norfolk
This page intentionally left blank
Table of Contents Acknowledgments 1. Introduction
2. Competing Conceptions of a Cognitive Virtue
3. A Virtue Account of Perception
4. A Virtue Account of Memory
5. Testimony and Cognitive Virtue
6. Moral Expertise: The Role of Cognitive Virtues in Generating Knowledge
7. Cognitive Virtue, Divine Hiddenness, and Reasonable Belief or Nonbelief
I would like to thank the National Catholic Bioethics Center for giving me time to prepare this manuscript. I would like to thank James Bohman, Kent Staley, and Linda Zagzebski who comment on portions of the manuscript. In this regard, I thank the editors of Continuum Publishers as well. Most of all, I would like to thank my wife Carol.
I. Introduction I begin this book with a lively quotation from Brand Blanshard’s On Philosophical Style. He notes, “The fact that there is only one feeling that is always safe in thinking and writing about philosophy, and that is the one A. E. Housman has described as ‘the faintest of human passions,’ the love of truth.” 1 Blanshard goes on to note that the love of truth takes two forms, philosophy and style. The love of truth in the first of these forms, the wanting to see the facts as they are, to follow the argument where it leads, even if it leads to the painful flouting of one’s other wants, the readiness to consider all evidence, to give full weight to objections, to believe and admit that one has been wrong, this transparent honesty and objectivity of mind is rarer, I fear, even among philosophers. . . .2 I begin this essay with this quotation because they capture most succinctly the spirit and insight of virtue epistemology. Virtue epistemology attempts to give an account of knowledge and other epistemic states, with reference to the agent’s character, and the character can be both intellectual and moral. Giving full weight to someone’s objections, or evincing a readiness to consider all evidence, not only improves one’s inquiry, but is also morally good. The version of virtue epistemology with which I am aligned is fundamentally concerned with answering two questions: (a) what is knowledge? and more importantly, (b) how can one get it? There is an important connection between these two questions. Asking the latter question presupposes that knowledge is somehow in one’s control, or that an agent can perform certain actions in order to get it. Thus, if the answer to the ‘what is knowledge?’ question makes reference to an agent’s cognitive actions, then the next question to ask is exactly (b). I see this as a bonus for virtue epistemology in that it aims to address wider concerns than mere conceptual clarity. It aims to give practical guidance to one’s cognitive life. Of course, defining knowledge with reference to an agent’s actions is highly questionable in contemporary epistemology. I should then
orient the reader to the question ‘what is knowledge?’ and subsequent answers in order to motivate virtue epistemology.3
What is knowledge? The question has to be distinguished from other related questions. Other related questions which lie on the fringes of this essay include, what are the objects of knowledge? Typical answers include propositions, but states of affairs are also viable answers. Questions about other epistemic states, such as “what is a belief ?” or “what is understanding?” will also lie on the fringe of this essay. I am concerned primarily with what knowledge is. Ordinary usage suggests that knowledge is a cognitive state in an agent, and that this state is valuable. When we consider the claim ‘Cason knows that (four-year-old) Peter is outside’ we are ascribing a cognitive state to Cason, and its value is a function of the cognitive state being true or accurate. To avoid cumbersome phraseology and to stick with philosophical orthodoxy, let’s refer to these cognitive states as beliefs.4 Not all true beliefs count as knowledge. Suppose Cason guessed that Peter was outside, or relied on a hunch. We would not say that her belief amounts to knowledge, even if it were true. This suggests that true beliefs need something more in order to count as knowledge, and this more has been the locus of controversy. To see why, consider some typical proposals. One proposal has it that true belief amounts to knowledge if and only if the belief is justified. Specifics aside, the basic intuition is that Cason has to have reasons for believing that Peter is outside. But is this account sufficient for defining knowledge? Consider some examples originally put forth by Edmund Gettier.5 Suppose the president of a company assures Smith that Jones will get the new job opening. Suppose further that Smith knows that Jones has ten coins in his pocket—Smith had just counted them. Smith infers from this that ‘the person who will get the job has ten coins in his pocket.’ The belief is justified because Smith just counted the coins in Jones’ pocket and received seemingly reliable testimony from the president. Let’s suppose that Jones is not the one who gets the job, but is Smith instead and Smith happens to have ten coins in his pocket—unbeknownst to Smith himself. So, Smith’s belief that the person who will get the job has tens coins in his pocket is true, justified, and yet we don’t want to say that Smith knows this. Smith’s belief turns out to be true, but by luck or accident. Smith bases his belief on reasons, reasons that normally do “track the truth” but on this rare occasion do not. The truth of the belief just happens. Guy Axtell states, “In Gettier cases, the agent’s justification for her beliefs has come ‘unglued’ from the truth of her belief.” 6 And this ungluing allows for the truth of the belief to be arrived at by accident and not by anything stable or consistent. Getting a true belief via luck is typically taken to be knowledge destroying. Duncan Pritchard points out that there are two different kinds of luck, veritic
luck and reflective luck.7 I will comment on reflective luck below. Veritic luck refers to true beliefs that are false in relevant near-by possible worlds. We can imagine in a relevant near-by possible world in which Smith does not have ten coins in his pocket and thus, he forms a false belief. Numerous examples abound. Consider the following from Ted Warfield, I see what looks like a dog in the yard. I reason (on this occasion) as follows: there’s a dog in the yard, so there’s at least one animal in the yard. As it happens, my premise is false (there’s a realistic toy in the yard and that is what I see in the dog-less yard). But my conclusion is true because concealed behind some brush is a squirrel.8 Consider also the following example from Alvin Plantinga.9 Suppose there is an aging Austrian forest ranger who lives in the mountains. He has wind chimes hanging outside of his cottage. When the wind blows the chimes chime and the ranger forms the true belief that the wind is blowing. As he ages he becomes tone deaf to the frequency of the chimes—since he becomes deaf to only these frequencies, we can safely say that he doesn’t know that he is losing his hearing in this way. But suppose that he often hallucinates hearing these chimes ring and forms the belief that the wind is blowing, and suppose that on this particular occasion the wind is in fact blowing. On this occasion he has true belief—and again we can stipulate that he is justified—but clearly he does not know that the wind is blowing. The essential structure of Gettier-type cases is that they involve two strokes of luck, one bad, and one good. The first stroke of bad luck is that the belief I form turns out per improbable to be false. My perceptual capabilities did not pick up on the subtle difference between real dog and realistic toy dog, and one might not expect such acumen given the minute subtlety. The second stroke of luck occurs when the belief I form on the basis of my perception (and a deductive inference), turns out to be true after all—though it could have easily been false in relevant near-by possible worlds. Consistent with our intuitions, true belief via luck does not count as knowledge. Pritchard explains, One natural explanation for this fact is that knowledge is a kind of achievement, and genuine achievements are not due to luck. Instead, the achievement has to be properly attributable to the agent herself. Hitting the target with one’s arrow by luck is of no credit to you; whereas hitting the target through skill is.10 With these thoughts in mind, we can formulate basic criteria for a satisfying account of knowledge. There are three criteria we can piece together from the discussion so far:11 the account of knowledge should (i) capture the intuitive
idea that knowledge is intellectually good, or rational, or justified, (ii) it should define knowledge such that the truth is entailed by the definiens, so as to block cases of luck, and (iii) it should explain why knowledge is better than mere true belief. Virtue epistemology (VE) aims to satisfy these three criteria, and in particular obviate double-luck cases by defining knowledge with reference to properties of the agent. In what follows I shall present a version of virtue epistemology referred to as virtue responsibilism (hereafter referred to simply as responsibilism). I aim to defend this view throughout this book by applying the account to specific avenues of knowledge. In the next chapter I delineate a cluster of competitor theories i.e., virtue reliabilist theories. The overall argument is that responsibilism is a fuller account of knowledge than virtue reliabilism, and since the most promising way to dodge double-luck cases is by defining knowledge with reference to properties of the agent (as is the case with all virtue accounts of knowledge) responsibilism is the most promising account of knowledge.
II. Virtue responsibilism A general account of a virtue The idea of a virtue is indigenous to ethical theory, the general characterization of the theory being that it focuses on the ethical qualities of the agent. “In virtue ethics, the focus is on the virtuous individual and those inner traits, dispositions, and motives that qualify her as being virtuous.” 12 The moral life, according to a virtue account broadly construed, is not reducible to an agent following out a set of moral rules. Rather, the moral life for the virtue theorist concerns itself mainly with what kind of person one should be. Of course, the virtue ethicists will be quick to point out that being agent-focused does not entail that a virtue theory can give little guidance for action. Right actions are the concern of a virtue approach, but such actions are defined with reference to actions that a virtuous agent would perform.13 More can be said, but the general approach of a virtue theory is to focus the ethical analysis on the agent, and actions are analyzed secondarily. Another very general feature of a virtue ethic is the kind of terms that constitutes its analysis. Again, Slote remarks, An ethics of rules will typically characterize acts as morally right or wrong, morally permissible or obligatory, depending on how they accord with appropriate rules. Such moral epithets are called “deontic” . . ., and they contrast with another class of ethical epithets where there is less immediate or ultimate connection with rules, namely, “aretaic” (from the Greek word for excellence or virtue) ethical terms like “morally good,” “admirable,” “virtuous.” 14
Actions on the virtue account would be cashed out not with reference to terms such as obligatory, right, or wrong, but rather with reference to terms such as ignoble, courageous, cowardly, etc. Getting more specific, the enduring tradition behind virtue theory agrees that a virtue is an excellence, where an excellence includes both an element of attaining the good, and doing so in the right way. A person with the virtue of justice, not only metes out shares proportionate in the specific circumstances, but does so from the motivation to mete out proportionate shares and in the right way (e.g., serving the needs of the poor in a way that does not further disparage them). Traditionally understood, virtues are acquired habits which perfect the underlying “natural” powers. Again, traditionally understood the powers are divided generally into the will and intellect. Moral virtues perfect the emotional and motivational states of the agent so as to will the good, intellectual virtues perfect the power of the intellect so as to apprehend the truth and to avoid falsehood. Virtues are strongly associated with having the right emotions and motivations. In fact, some could say that for a virtue theory, motivations of the actor are primary, and acts are secondary. Some may think this as odd since morality concerns itself with doing the right thing whether one feels good about it or not. Given the primacy of these notions and their apparent oddity, some clarification is in order. Following Zagzebski, I take an emotion to be a state which combines both a cognitive and affective component. The affective component of an emotion is, when conscious, experienced as a feeling. Additionally, these affective states have an intentional object. “That is, an emotion is a state of feeling a certain way about something or at something or toward something of a certain description.” 15 The cognitive aspect of an emotion stems from the fact that the intentional object of an emotion is typically represented in consciousness. Since representational states are quasi-categorical—I represent something as an x—emotion has a cognitive aspect as well. “In an emotional state, the agent feels a characteristic way about an object represented as falling under a concept corresponding to the particular emotion.” Examples include the following, “pity is an emotion consisting of feeling pity for something seen as pitiful; love is an emotion consisting in loving something seen as lovable; fear is an emotion consisting in feeling afraid of something seen as dangerous. . . .” 16 Since emotion is partly a cognitive state about objects bearing some value or disvalue, “emotion is a type of value perception that feels a characteristic way.” 17 As such, there can be an adequation between emotion and object. Pity in the presence of the pitiful is apt or fitting, contempt in response to the contemptible is apt or fitting. Emotion, then, is a kind of perception of morally salient features in the environment. Motives are emotional states that initiate action. “Motives are emotions that cause acts . . .” 18 Motives, then, are subsets of emotions. They are affective states that move the agent to act, and as such serve as the causal explanation for
actions undertaken. They can exert such influence either consciously or unconsciously. They serve as components of the virtues insofar as “virtuous persons tend to have certain emotions that then lead them to want to change the world or themselves in certain ways. Virtuous persons have motives associated with the particular virtue.” 19 Examples include courage, whereby the courageous person “is motivated out of emotions characteristic of the virtue of courage to face danger when something of importance is at stake.” 20 For intellectual virtues, the motivational component is present as well. An open-minded person, Is motivated out of delight in discovering new truths, a delight that is strong enough to outweigh the attachment to old beliefs and to lead to the investigation of neglected possibilities. In doing so, she is drawn by the desire to form more true beliefs or, at least, to get closer to the truth than she was previously.21 One more feature of the virtues is important to note, and that is that virtues are context sensitive. Aristotle, for instance, remarks, Virtue, then, is a state of character concerned with choice, lying in a mean, i.e., the mean relative to us, this being determined by a rational principle, and by that principle by which the man of practical wisdom would determine it (1107a1). Christine Swanton defines a virtue as follows, “A virtue is a good quality of character, more specifically a disposition to respond to, or acknowledge, items within its field or fields in an excellent or good enough way.” 22 Similarly, Robert Audi remarks, I want now to propose an account of acting from virtue built around six notions, . . . These dimensions are, first, the field of a virtue, roughly the kind of situation in which it characteristically operates; second the characteristic targets it aims at . . .23 These quotations indicate that a virtuous person acts appropriately given the circumstances. Swanton notes, “As I shall put it, a virtue is a disposition to respond well to the ‘demands of the world.’ ” Circumstances present us with opportunities to respond well to them, and similar circumstances may require subtly different responses from the virtuous agent. Of course, there may be disagreement among virtuous persons as how best to respond to the demands of the world. I do not think this is a problem, for this may be strong evidence for an axiologically deep and pluralistic moral order. If there can be only one correct answer to our most trenchant moral problems, then one has an argument that at least one of the demurring persons is not virtuous. On the contrary, I see the
argument going the other direction, namely, if virtuous persons disagree, this indicates an axiologically complex moral order. So, a virtue includes the components of being successful in attaining the good in the specific “field” or circumstances in question, and doing so from the right motivational dispositions. But what concerns us is how the concept of a virtue can answer the “what is knowledge?” question. Clearly, knowledge that p does not require one having a settled disposition to apprehend p. Nor does it seem to be the case that knowledge is as context sensitive as the moral life is. To address the question of knowledge we need to advert our attention to the concept of a virtuous act.
Intellectually virtuous acts and the definition of knowledge Focusing moral evaluation on persons is certainly a bonus for a moral theory, but not at the exclusion of analyzing actions. We are concerned with moral evaluation of actions as well. Likewise in epistemology, we are not so much concerned about intellectually virtuous persons, but what beliefs count as knowledge. The notion of a virtuous act needs to be employed, and certainly the components of a virtue are imported into the notion of a virtuous act. Typically, a person who performs a virtuous act is successful in reaching the end of the virtue in question. A person who is motivated well has the aim in mind but if she fails to succeed in attaining the end of the virtue in question the act is not a virtuous act. “We praise success in reaching the ends of virtue and withhold praise . . . from the cases of failure. . . .” 24 Conversely, if a person is not motivated well, but succeeds in performing, say, a compassionate act, is not by that fact worthy of full praise. Additionally, being motivated correctly, and being successful is not enough to count as a virtuous act, for the success may be due to accident or luck. To appreciate the intuitive sense of these components, consider the following example, Suppose that Dr. Jones, a physician, has very good inductive evidence that her patient, Smith, is suffering from virus X. Smith exhibits all of the symptoms of this virus, and laboratory tests are consistent with the presence of virus X and no other known virus. Let us also suppose that all evidence upon which Jones bases her diagnosis is true, and there is no evidence accessible to her that counts against the diagnosis. The conclusion that Smith is suffering from virus X really is extremely probable on the evidence. But even the strongest inductive evidence does not entail the conclusion and so it is possible to make a mistake. Let us suppose that this is one of those cases. Smith is suffering from a distinct and unknown virus Y. Dr. Jones’s belief that Smith is presently suffering from virus X is false, but it is justified and undefeated by any evidence accessible to her.25
At this point in the story, we can turn this into a double-luck case as described above, according to which, we stipulate that besides suffering from virus Y, “Smith has very recently contracted virus X, but so recently that he does not yet exhibit the symptoms caused by X, nor is the laboratory evidence upon which Jones bases her diagnosis produced by X.” 26 Intuitively, we would say that Dr. Jones does not know, and yet she exhibits the components of a virtuous act, namely, she is motivated properly in that she aims at a thorough investigation and conducts her inquiry with care. Furthermore, she is successful in getting a true belief. But one component is not satisfied, namely, she is not successful because of her good inquiry. Clearly, Dr. Jones acquires her belief because of her intellectually virtuous behavior, but she does not get the truth because of her intellectually virtuous behavior. Condensing these elements of a virtuous act, we can define it as follows, An act is an act of virtue A if and only if it arises from the motivational component of A, is an act that persons with virtue A characteristically do in the circumstances, and is successful in bringing about the end of virtue A because of these features of the act.27 Extending this analysis to an act of an intellectual virtue, we get the following, An act of intellectual virtue A is an act that arises from the motivational component of A, is something a person with virtue A would (probably) do in the circumstances, is successful in achieving the end of the A motivation, and is such that the agent acquires a true belief (cognitive contact with reality) through these features of the act.28 The purpose of the ‘because of’ clause is to block cases of double luck. Once such cases are blocked, the responsibilist has the resources to define knowledge so that it is immune to Gettier-type counterexamples. Accordingly, the definition of knowledge on the responsibilist account is as follows, “Knowledge is a state of belief arising out of acts of intellectual virtue.” 29 (Notice, there is no separate reference to true belief, since performing an intellectually virtuous act entails getting the truth.) For those who think, as I do, that not all knowledge states involve a belief, the definition of knowledge may also be stated as follows, “Knowledge is a state of cognitive contact with reality arising out of acts of intellectual virtue.” 30
Clarifications All of the intellectual virtues share, in so far as they are intellectual virtues, the end of getting truth and avoiding falsehood. Open-mindedness, intellectual care, and thoroughness, all contribute in some way to the goal of truth or
avoidance of error. But the motivational component of each virtue will contain not only the specific end of the virtue in question but also a motivation for the truth. So, for example, the open-minded person is motivated to consider neglected possibilities and new viewpoints and arguments, in order to obtain new truths. Which intellectual virtue(s) is most apt for the epistemic circumstances is what a person with phronesis would exercise in those circumstances. The intellectual virtue of phronesis plays an executive function in managing what kind of intellectual behavior will bring about the truth. This clarification is important in that not just any intellectual virtue will deliver the truth or avoid falsehood. Consider intellectual courage which may be the virtue of holding fast to a belief in the face of opposition. Knowing how long to hold fast and which other virtues may be required (such as intellectual care) is determined by the person with phronesis. In concert with other virtues, the virtues the person with phronesis would exude, the end of one’s inquiry becomes alethic. To clarify then, the reference to virtue A in the definition refers to the virtue a person with phronesis would exude in the epistemic circumstance in question. Regarding the notion of an intellectually virtuous act, such an act is reliable in attaining the end of the virtue in question, but it is not a virtuous act because it is reliable (hereafter I will refer to a virtuous act to mean an intellectually virtuous act). A virtuous act puts “us in touch with the truth in a qualitatively valuable way, a way that involves self-reflectiveness and other internal properties of the agent, including a motivation to obtain knowledge and the motivations distinctive of the individual virtues.” 31 A virtuous act, then, should be thought of as including both an externalist element, namely reliability, and an internalist element, according to which the agent has access to reasons for her beliefs if the epistemic circumstances require having such reasons in order to achieve knowledge. Virtuous acts, then, are identified not only with reference to teleological ends (reliability) but to nonteleological features such as the intrinsic value of the act, or its “admirability.” I cannot give identity criteria for a virtuous act without appealing, at some point, to intuition. I suspect that avoiding such an appeal, however, is not a project worth embarking upon. We cannot avoid appealing to our intuitions, but we can inform and refine them.32 I just mentioned that the internalist element in a virtuous act is circumstance dependent. What is being said here? Consider the following two cases: suppose Dr. Naïve, a medical physician, believes that Plan B (a contraceptive drug) does not have an abortifacient effect because his priest said so. Consider Dr. Thorough who peruses the literature on the mechanisms of action for Plan B (and there is a lot of conflicting literature on the topic) and forms the belief out of a careful and thorough investigation that Plan B does not have an abortifacient effect. Assume for the moment that having an abortifacient effect is a bad thing. Suppose also that the belief is true in both cases. It seems to me that Dr. Naïve does not know that Plan B is not an abortifacient, but
Dr. Thorough does. Part of the explanation is that the roles they play as physicians place heavier epistemic demands on knowing whether Plan B functions as an abortifacient or not. The epistemic circumstances require that the physician has internally accessible evidence as obtained by Dr. Thorough. His belief is both reliable and admirable. So, responsibilism captures intuitions endorsed by both the evidentialists/internalists, and satisfies the concerns of the reliabilists by including a success component.33
Affinities within responsibilism There are several theories which bear a family resemblance to one another and the resemblance consists of defining knowledge with reference to what the agent does. Knowledge is true belief whereby the agent gets the truth not by luck but by her own cognitive action. The reason for the cognomen ‘responsibilism’ is because the agent is responsible for her true beliefs. Several theories share this feature, of which Zagzebski’s theory is one. Although I focus on a Zagzebski-type responsibilism throughout this book, others deserve elucidation. John Greco’s theory is the most notable companion to Zagzebski’s approach within the responsibilist camp. Greco defines knowledge as creditable true belief, More exactly: in cases of knowledge, S believes the truth because S believes out of intellectual virtue. On this same approach, in cases of knowledge, S deserves credit for believing the truth. This is because, necessarily, a special sort of credit accrues to success grounded in virtue. And on the present account, knowledge is success grounded in virtue.34 What counts as a virtue is the agent’s own reliable cognitive character.35 There are certainly notable differences between a Zagzebski-type responsibilism and a Greco-style one. For instance, insofar as Greco’s notion of an intellectual virtue does not necessarily include reference to a motivational component, he avoids the unmotivated belief objection—outlined in the next chapter. On the other hand, since the Zagzebski-type responsibilist defines knowledge with reference to acts, children and certain intellectually impaired adults can have knowledge in some cases. Whereas Greco’s account, which requires that the agent possess stable and reliable dispositions, has trouble explaining how such persons achieve knowledge. Of course, I have my response to the unmotivated belief objection outlined in Chapters 3 and 4, and Greco will have his explanation for why children may have knowledge on his account. I suspect that delineating further distinctions and making requisite clarifications between the two versions will not affect the conclusions arrived at in this book—i.e., that responsibilism is a superior theory to other distinct competitors. Having said that I focus on a Zagzebski-type version throughout this work.
Problems and refinements One of the main motivations behind responsibilism is that it avoids double-luck cases. Epistemic luck is typically taken to destroy knowledge if the person gets a true belief. The ‘because of’ clause in the responsibilist’s definition of knowledge ensures that the true belief the virtuous person gets is due to her epistemic behavior and not some weird or fantastic feature in the environment. But this is just where problems resurface. Duncan Pritchard has pointed out that the responsibilist is unable to deal with cases of luck after all.36 Consider the following Pritchard-inspired objection. Recall the example above of Dr. Jones. Dr. Jones conducts an impeccable inquiry in determining whether Smith has virus X. It turns out that Smith has virus Y which manifests similar symptoms as virus X. Smith did contract virus X, but the very morning Dr. Jones rendered her diagnosis—too soon for virus X to manifest any discernible symptoms of its own. Dr. Jones acquires her belief via her intellectually virtuous behavior, but she does not get the truth because of that behavior. Call this case 1. Case 2 runs just like case 1 except that Dr. Jones is correct in her judgment that Smith has virus X, and Smith contracted virus X in time for the virus to manifest discernible evidence, of which Dr. Jones descried. Here, Dr. Jones knows that Smith has virus X, but the only change between cases 1 and 2 is the incidental circumstance that Smith contracted X in time for it to manifest discernible symptoms. No difference in Dr. Jones’ cognitive character is stipulated between the two cases. If not, then how can one’s cognitive activity make the difference between knowing and not knowing? There seems to be need for an environmental condition, something akin to Alvin Plantinga’s environmental condition for warrant.37 The objection is certainly moving and important. I suspect that to some extent, the epistemic environment has to allow for virtuous activity. One cannot imagine a virtuous person performing a morally virtuous act in cases of tragic dilemmas—dilemmas for which any one action involves a violation of a wellgrounded moral principle.38 In such cases, the actor cannot perform the right act, since either choice involves breaching a moral obligation. I suspect that there are epistemic analogues to tragic dilemmas. But stipulating that the environment allow for an intellectually virtuous act is not a further condition for knowledge. Rather it is a logical condition for the realization of a virtuous act. It is not a condition specific to the concept of a virtuous act. A responsibilist can accept this weak form of an environmental condition without revision to her central claims. I suspect that this Pritchard-inspired objection aims to show more, i.e., that getting knowledge requires favorable environmental conditions, or stipulating that one’s true belief tracks the truth. The idea behind this latter stipulation is that if one’s true belief in the actual world ends up false in nearby possible worlds, then the true belief does not track the truth, and does not count as
knowledge. In case 1, Dr. Jones’ diagnosis is true in the case described but turns up false in nearby possible scenarios and therefore does not know in the actual case.39 Notice, adding the condition that one’s belief tracks the truth is adding a condition that is specifically not a virtue-theoretic condition. My response is that the objection confuses the responsibilist’s definition of knowledge with an explanation for why the conditions outlined in the definition are or are not met. The difference in the environmental conditions explains why the conditions of knowledge are met in case 2, but not met in case 1. In case 1, the explanation for why Dr. Jones fails to get knowledge is that she does not get the truth because of her virtuous behavior. In case 2, the explanation for why Dr. Jones gets knowledge is that she gets the truth because of her virtuous behavior. Pritchard has several other objections aimed at responsibilism alone. According to Pritchard, a main motivation for responsibilism is that it aims to block what he calls reflective luck. An agent S is reflectively lucky with respect to a true belief p (in the actual world) if, and only if, the grounds on which S believes p in the actual would be the same grounds in nearby worlds, and p ends up being false in these other worlds. The idea is that S’s justification is indiscriminate or more technically, it does not track the truth across relevantly similar epistemic situations. Zagzebski uses chicken sexers as an example to show the importance of reflective luck and how her theory blocks it. Briefly, chicken sexers are reliable at determining the sex of chickens. They cannot report how they are reliable. Apparently, they turn the chick over, sniff, or some other combination of perceptual inquiry; and the end result is a true belief. In nearby possible worlds, the grounds on which they form their beliefs (which are none since sexers provide little to no justification for their beliefs), could be the same and the beliefs turn our false. Zagzebski thinks that the notion of a virtuous act requires that the agent, in some epistemic circumstances, have good reflective grounds on which her belief is based. Thus, responsibilism blocks cases of reflective luck.40 Pritchard responds, The problem with understanding this type of virtue epistemology as being motivated by a desire to eliminate reflective epistemic luck is that no epistemological theory can meet this challenge. This is highlighted for us by the radical skeptic via her use of radical skeptical error scenarios (such as that one might be a brain in a vat being “fed” one’s experiences).41 The pertinence of the brain in a vat reference is to show us that one can have the exact same justification in the actual world, as in the brain-in-a-vat world, but in the latter world, one’s beliefs turn out all false. So, radical skeptical error scenarios impugn the motivation for responsibilism, namely, that responsibilism is able to block reflective luck.
I find myself unmoved by this argument, mainly because the reference to radical skeptical (or global) error scenarios do not count as nearby possible worlds—as is indicated in the definition of reflective luck. As such, it is unclear to me how global error scenarios should be considered by someone concerned about blocking reflective luck scenarios. Pritchard’s argument may be a weaker claim, namely that blocking reflective luck cases can be accomplished simply by importing an internalist-type justification requirement. If so, then responsibilism is unmotivated in that it does not gain anymore epistemic ground than what an internal-justification view would endorse. Concerning this latter point, responsibilism warrants some refinement and clarification. First, responsibilism combines into one concept, a virtuous act, both an externalist-reliability component and an internalist-justification component. The gain made is purely conceptual, but a gain nonetheless over other theories that bifurcate their conditions for knowledge. Second, for the responsibilist, the internalist-justification requirement is ensconced in the concept of a virtuous act. As such, whether internally accessible justification is required is determined by the virtuous agent. And, since a virtuous act is context sensitive, the knower would believe or conduct inquiry in the way a virtuous person would in the circumstances. Some circumstances would require internally accessible grounds, to some degree or other, and some other circumstances would not. Consider the example given above of the two physicians and how they each formed their true belief that Plan B is not an abortifacient. Clearly, a virtuous act was committed by the one who perused the literature, but not by the one who believed on nonexpert testimony. Examples abound where we would not require justification for a belief in order to count as knowledge (perception is probably such a case) and where we require different degrees or kinds of justification (the physician should have first-hand knowledge of the literature, and not base important beliefs on nonexpert testimony). A simple internalist-justification requirement on knowledge does not capture this contextual sensitivity, but the notion of a virtuous act does. I take both of these clarifications to be significant motivations for responsibilism over other nonvirtue theories. Several other objections related to the issue of luck have been generated against a Zagzebski-type responsibilism. The objection I am most concerned with here is the objection from fleeting reliability. Originally formulated by John Greco, the objection exploits the notion of a virtuous act in which an agent gets the truth on one single occasion but has no disposition to act reliably. Consider the following, Suppose that a poor math student adopts a correct algorithm for solving an equation he is working on. By hypothesis, using the algorithm is a highly reliable process . . .. But suppose the student almost never uses a correct
Virtue Epistemology algorithm; usually he adopts an incorrect one, or merely hazards a wild guess. Then it seems wrong to say that he has knowledge in the case where he happens to use a reliable process.42
Zagzebski responds43 by pointing out that the person performing a virtuous act has to be motivated to get the truth, and do what the virtuous person would do in the circumstances. By reiterating these features, Zagzebski argues that the example is underdescribed. The student does not satisfy all the criteria the responsibilist lays out. As such, the example cannot function as a counterexample. Jason Baehr takes up the Greco-inspired objection and tries again.44 Baehr stipulates that the student in Greco’s example is genuinely motivated to solve the equation (due to a momentary enthusiasm generated from a previous good grade). But the motives are fleetingly adopted and yet, they are motives the virtuous agent would evince in the circumstances. Baehr surmises that fleeting alethic motives is knowledge destroying. I do not share Baehr’s intuition on this modified case. Consider Jane, an upper division college student who is interested in sexual activity with her fiancé. She has failed her biology class as a freshman and is taking it again grudgingly. But given her interests in sexual activity and in not getting pregnant so young, she forms an intense motivation to know about the human reproductive system. She inquires about the subject with her professor, being careful and attentive to the explanations given. She studiously takes notes in class and rehearses them afterward. She takes the next biology exam and fails it, but performs impeccably on the reproductive section of the exam. Does she know the human reproductive system? I suspect that she does, even though her alethic motives were fleeting.
Further problems, motivations, and chapter summaries There are other problems for the responsibilist view which are not as easily dispatched as the ones addressed in the previous section. One in particular is what I call the problem of unmotivated belief. Since the responsibilist defines knowledge with reference to a virtuous act, which includes a motivational component, then all knowledge would be a function of one’s virtuous motivations. But clearly there are cases of easy knowledge, such as perception, memorial beliefs of the not-too-distant past and, relatedly, receiving simple testimony. Some of these problems are referred to as cases of easy knowledge, knowledge without credit, or cases of passive or simple knowledge. Whichever cognomen one uses, the central problem threatens to impugn responsibilism. I address this problem in my chapters on perception, memory, and testimony. In cases of perception and memory, I appeal to empirical evidence which indicates that one’s emotion-motivational states are in fact implicated in the genesis of
knowledge. This sounds surprising and counterintuitive, but if an objection rests on what can be empirically determined, then the evidence should be followed, not our intuitions. The basic outline of my response to the unmotivated belief objection is that perception and memory require attention and attention is mediated by one’s motives. These motives are epistemic in that they may serve as components of a virtuous act. It follows that cases of easy knowledge are not incompatible with them arising out of a virtuous act. Notice, in order to obviate the objection, I need only show that cases of easy knowledge no longer function as counterexamples. Although I do argue that easy knowledge is a function of a virtuous act, I need only argue for the weaker claim in order to render the objection inert. Given the importance of the objection, I give a detailed presentation of it in the next chapter after I present what is the main competitor to responsibilism, i.e., virtue reliabilism. I reserve a presentation of the objection until after presentation of virtue reliabilism so that the reader may see the relative merits and demerits of each theory. The overall aim in this book is to show the superiority of responsibilism over other theories. I grant that responsibilism is not without its problems and problems will remain, but what is important to see, however, is the relative superiority of responsibilism over other theories. The unmotivated belief objection has certainly served as a pertinent and even definitive objection, thus obviating it will allow us to see the distinctive contribution responsibilism makes in other ways, i.e., giving an account of testimonial, moral, and religious knowledge. The account of testimonial knowledge focuses not only on the properties of the receiver of the testimony, but also on the testifier. I define testimonial knowledge with reference to a virtuous act performed by each.45 Since a virtuous act preempts double-luck cases, the definition bears the merit of being immune to counterexample. Additionally, testimony is a variegated affair. In some cases, testimonial chains are rather complex, such as expert testimony in the court of law, and some are not so complex, e.g., giving directions to a stranger. Since a virtuous act is context sensitive—it is what a virtuous person would do in the circumstances—a virtue account of testimonial knowledge countenances our basic intuitions on successful transmission of knowledge across various contexts. A virtue epistemology modeled on virtue theory in ethics is probably most apt at giving an account of moral knowledge. Saying I will give an account of moral knowledge, however, begs important questions about the nature of moral discourse. Weakening my promise to giving an account of good moral judgment, may be more apt. I will not address the debate between moral realists and antirealists. I give an account of moral judgment that averts commitment to contentious metaethical claims. I shall use the phrase expert moral judgment to refer to good or superior moral judgments and shall use the term expertise to refer to the persons who generate such judgments. In brief, I argue that
S forms an expert moral judgment that p if, and only if, it is formed from an intellectually virtuous act(s). Spelling out what this means consists of motivating the view that moral expertise is more than having justification for p. I borrow on Michael Depaul’s study Balance and Refinement according to which he motivates the view that experience is necessary. There is a sub-category of moral experience (what he calls formative experience) that can either corrupt or convert the knower. It can cloud someone’s sensitivity to a moral issue or it can correct naiveté. How formative experience exerts its force is partly determined by the motivational state of the agent. The change of heart that takes place in response to a formative experience is a movement both cognitively and morally toward being a phronimos. Additionally, I present empirical research supporting the view that not only are emotions involved in moral judgment, they are necessary and sufficient. As such, having the right emotions, namely, the ones the phronimos would have, grounds claims to moral expertise. Only responsibilism is able to offer an account of moral expertise that countenances the role of experience that Depaul takes to be necessary, and emotion that the empirical research takes to be necessary. Religious knowledge has become the focus for discussion, not only for theists, but for atheologians as well. One atheologian, John Schellenberg, has put forth an impressive argument against God’s existence, by appealing to the hiddenness of God—which means the lack of evidence for God’s presence in the world which is something we would not expect from a good and loving God. A good and loving God would desire meaningful relationship with us, but a logically necessary condition for this is belief. Schellenberg contends, however, that reasonable nonbelief exists, and this is incompatible with God’s existence. The argument raises interesting questions in religious epistemology, some of which can be addressed by a virtue account of knowledge. Crucial to Schellenberg’s argument is the notion of adequate investigation, and reasonable nonbelief. I apply the resources of a virtue account and borrow conclusions from previous chapters, in particular the chapter on testimony, to argue that adequate investigation on a virtue account only allows the person for whom God is hidden to take the position of agnosticism, not atheism as Schellenberg contends. I end the chapter with a comment on how a virtue theory can quite naturally demarcate the limits of human inquiry. The person with phronesis will recognize the approaching limit of human knowledge and as such, can justify certain impasses in religious and moral discourse.
Competing Conceptions of a Cognitive Virtue
I. Introduction In this chapter, I elucidate the main competitor to responsibilism: virtue reliabilism. I do this to illustrate the merits and demerits of reliabilism and responsibilism. I end the chapter with a detailed presentation of the unmotivated belief objection to responsibilism.
II. Virtue reliabilism Alvin Goldman’s presentation of reliabilism serves both as an example of virtue reliabilism and as an example of clarity and breadth. My aim in this section is to capture the essential components of his theory, and defend him from Alvin Plantinga’s counterexamples. By obviating Plantinga’s counterexamples, I aim to undercut a key motivation for accepting his proper function account. I do this to pinpoint the few weaknesses reliabilism suffers from and to argue that it is the chief competitor theory to responsibilism. Reliabilism in its simplest form is the claim that a belief counts as knowledge only if it is produced by reliable processes. And reliable processes are processes which, for the most part, produce true beliefs. Reliabilism is usually classed under the general rubric of externalism. The idea behind externalist theories is that an agent need not know the conditions under which she forms a particular belief in order for that belief to count as knowledge. What is essential for knowledge is that the belief is caused or produced by processes that usually get the truth. This is an important move in that it seems to hush skeptical inquiries. John Greco adumbrates the skeptical challenge as follows, “All knowledge, says the skeptic, must be grounded in good reasons. But not any reason is a good reason—one must have reasons for believing that one’s reasons are true.” 1 But this generates a need to supply reasons for the good reasons, and we have either an infinite regress, a viciously circular set of beliefs, or the line of justification ends arbitrarily. Reliabilism stipulates that not all knowledge requires good reasons. A well-functioning or reliable process confers knowledge on the true beliefs the process produces. The simple reliabilist is quick to point out that
perception, and introspective beliefs intuitively count as knowledge, but one need not produce reasons for them. Reliabilism seems like a natural way to explain why people have or do not have knowledge in most every case. It seems clear that true beliefs formed from, say, wishful thinking, spurious conjecture, or testimony from one who desires to lie, yield beliefs which do not count as knowledge even if they turn out true. And the reason beliefs formed from such processes do not count as knowledge is because such processes are unreliable—i.e., they do not yield true beliefs often. With respect to an explanation of why one does have knowledge, reliabilism offers a fairly straightforward answer here as well. We think of knowledge as being something more than mere true belief, i.e., true belief arrived at nonaccidentally. And reliabilism explains why a true belief is formed nonaccidentally, i.e., because it is formed by a reliable process. “A belief must not only be true, but it must also be objectively likely that the belief is true in order for knowledge to obtain.” 2 For instance, most cases of perception count as knowledge. The reliabilist says that this is because perceptual beliefs are formed by reliable faculties. The same can be said about inductive beliefs and belief in other minds. More sophisticated versions of reliabilism are able to obviate Gettier-like counterexamples. The added sophistication is simply that in order for a reliable process to produce beliefs which count as knowledge, it must produce true beliefs in counterfactual circumstances as well. For instance, consider the example of Henry who is driving through a Wisconsin town in which there are numerous barns.3 Henry turns and points at a barn in this town and remarks, “That is a barn.” In such a case, we would certainly ascribe knowledge to Henry that what he saw was in fact a barn. Suppose, however, that this particular Wisconsin town has citizens who desire not to look impecunious and thus, erect fake barn facades. In fact, they erect close to ten facades for every one real barn. Suppose again that Henry is driving through this same town and looks at a barn and says, “that is a barn.” Luckily, it is a real barn, so his belief is true but certainly we would not say that he has knowledge. ‘Traditional’ justified true belief accounts cannot explain why we would say that Henry knows in the first case but not in the latter. For Henry has the same evidence in both cases, and the beliefs in both cases are true. Since any explanation for the difference in the cases here lay not in the kind of belief Henry has, the explanation must lay in the kind of processes or mechanisms used to form such a belief. And the kind of processes relevant to explain the Henry case is that the process must be reliable in relevant alternative situations. Goldman states, “. . . a cognitive mechanism or process is reliable if it not only produces true beliefs in actual situations, but would produce true beliefs, or at least inhibit false beliefs, in relevant counterfactual situations.” 4 So, in the case of Henry, he does not have knowledge in the second case because the process by which he acquired the belief could have easily produced a false belief, i.e., if the barn he looked at was a facade instead.
Competing Conceptions of a Cognitive Virtue
Whereas in the first case, the relevant alternative situation is one in which— most likely—there are numerous real barns. So, reliabilism is able to explain quite naturally many cases in which we have knowledge, (e.g., other minds, induction, perception) and many cases in which we do not (e.g., Gettier-like cases). It captures the intuitive idea that in order for true beliefs to count as knowledge they must be true in a nonaccidental way. But trouble looms in that we may ask, whether any reliable process confers knowledge? Consider Lawrence Bonjour’s famous example of Norman. Norman, under certain conditions that usually obtain, is a completely reliable clairvoyant with respect to certain kinds of subject matter. He possesses no evidence or reasons of any kind for or against the general possibility of such a cognitive power, or for or against the thesis that he possesses it. One day Norman comes to believe that the President is in New York City, though he has no evidence either for or against this belief. In fact the belief is true and results from his clairvoyant power, under circumstances in which it is completely reliable.5 This example and others like it push us to delineate reliabilism further. Typically we form beliefs from cognitive states. A cognitive state can consist of a perceptual experience, or it could be a belief. Cognitive state transitions which produce justified beliefs will be those that follow what Goldman calls J-rules. Goldman states, “On the present proposal a J-rule permits transitions from prior cognitive states to beliefs. . . . Prior cognitive states may consist of perceptual appearances, ostensible memories, [other] doxastic attitudes, or any combination of such states.” 6 So, we can understand J-rules as propositions which describe (or prescribe) cognitive state transitions that are permitted, and if the J-rule system is good then the transitions that are permitted will deliver justified beliefs. An example of a good rule may be the following;7 let M be the permission operator and let ‘/’ denote a transition state from prior beliefs to the present belief—prior beliefs will flank the left of / and the target belief will flank the right of /. Let SB be the belief operator for an agent S and p, q, r, s . . . be variables for propositions. Now, a good J-rule may read, If (p & q & r) logically entails s, then M (SBp & SBq & SBr / SBs) This rule states that if propositions p, q, r entail s, then if S believes each of p, q, and r, then S is permitted to believe s. It is not enough, however, that a belief is justified if and only if it is permitted by the right system of J-rules (to be defined further below). Suppose S forms a true belief B on the basis of a good J-rule (e.g., reliable perceptual faculties), but S also believes that her perceptual faculties are unreliable, and therefore
untrustworthy. Intuitively, B is not justified for S. Capturing this intuition, Goldman stipulates the following criteria for a justified belief, “S’s believing p at t is justified if and only if (a) S’s believing p at t is permitted by a right system of J-rules and (b) this permission is not undermined by S’s cognitive state at t.” 8 S’s cognitive state can undermine one of her beliefs if she has a defeater for that belief. Continuing, Goldman says that a J-rule system is right “if and only if R permits certain (basic) psychological processes, and the instantiation of these processes would result in a truth ratio of beliefs that meets some specified high threshold (greater than .50).” 9 Basic psychological processes include perception, memory, introspection, and reasoning. Thus, the idea is that a belief bears justification for an agent, if and only if the belief is formed as a result of a reliable cognitive process with no explicit defeaters. A belief counts as knowledge if and only if it is justified and the process that produces the belief is what Goldman refers to as being locally reliable. “Local reliability concerns only the reliability of the process in the context of the belief under assessment.”10 And he says that this context includes counterfactual situations as well. For instance, “. . . a true belief fails to be knowledge if there are any relevant alternative situations in which the proposition p would be false, but the process used would cause S to believe p anyway.”11 Notice, by adding a counter factual condition, Goldman is attempting to dodge cases in which the agent gets true beliefs via luck. The case of Henry is an example of true belief via luck. Cognitive processes that are locally reliable are the ones necessary for knowledge. We still need an idea of how to identify the relevant processes. We still do not have the conceptual tools to respond to the case of Norman or other cases, for even if Norman is counterfactually reliable, we still would be reticent to ascribe to him knowledge. Addressing these concerns requires making several distinctions regarding processes. First, there are process types, versus tokens. Additionally there are first- and second-order processes, and finally methods. I shall take each in turn. A process type is a “. . . functional operation or procedure that generates a mapping from inputs to outputs.” 12 A process token would be a single output instance of a process type. This way of distinguishing types from tokens makes appeal to some quantitative measure—i.e., how many things of one kind (e.g., beliefs) does each type produce? One motivation for distinguishing process types from tokens is that Goldman can skirt putative counterexamples to his theory. Take a process token such as perceptual processes that occur on Wednesday. Clearly such processes are not relevant for whether a belief counts as knowledge since it specifies a time at which the process or “. . . functional operation . . .” occurs. What matters in assessing knowledge claims is whether the process itself is reliable independently of a specific time at which it occurs. So any counterexample which makes reference to token processes will not work.
Competing Conceptions of a Cognitive Virtue
The distinction between type and token is also qualitative as well. For instance, a process type such as perception is content neutral, meaning that the process generates an output given inputs of varying contents. I form perceptual beliefs of varying sorts corresponding to the many different shapes, colors, and sizes of the objects in my visual environment. A process token, on the other hand, may be content specific, meaning that the process only generates one output given an input of specific content. Suppose I have a process according to which I immediately form the belief that there are 250 matches on the floor only when the matches are configured in one particular way. Such a process would be considered a process token by Goldman. And it seems reasonable not to count the products of process tokens as knowledge. First-order psychological processes are processes such as perception, memory, and reasoning. They aim to produce beliefs of various sorts. Further, they are cognitive processes that each human person has, simply by virtue of being a human person. Second-order processes and methods may vary from person to person and thus are not possessed by any particular human being simply by virtue of being human. Second-order processes are those processes that govern the acquisition of methods. It is important to note here that their aim is not the production of belief but rather pertaining to methods. And methods are heuristics, algorithms, or learnable methodologies that one can use to transition from one cognitive state to another—thus methods are like basic processes in that they aim at producing beliefs, but they are acquired whereas basic psychological process are not. An example of a method would be using a mathematical algorithm (M1) to solve various arithmetical equations. A second-order process used to acquire such a method may be one of the following, SP1: accept a mathematical algorithm on blind faith.13 or, SP2: accept a mathematical algorithm only from mathematical experts.14 Though M1 may be a reliable heuristic, it matters to Goldman how the methods are acquired. A person who forms beliefs in accordance with M1 is not by that very fact justified. If someone acquired M1 via SP1, we would say that her beliefs about mathematics are unjustified. But equally clear is the intuition that acquiring M1 via SP2 does yield justified mathematical beliefs which if true reliably, then they count as knowledge. So, a belief B is justified on Goldman’s account if and only if it is produced by a reliable process, or method. And B counts as knowledge if, additionally, B is true and the process which produces B satisfies local reliability. Before addressing Alvin Plantinga’s counterexamples we may consider some which Goldman explicitly addresses in his Epistemology and Cognition. Consider
Hilary Putnam’s case of the Dalai Lama.15 Suppose the Dalai Lama is infallible on matters of faith and morals. As such he is 100 percent correct when he renders a moral judgment or a judgment about faith. Suppose Jenny believes the Dalai Lama’s testimony on such subject matters and her only reason for taking up such beliefs is “the Dalai Lama said so.” Intuitively, Jenny fails to know, even though her beliefs are formed from a reliable process. In response, Goldman makes use of his distinction between process and method. Goldman says, The method of believing whatever the Dalai Lama says is just that, a method, in the slightly technical sense I have given that term. But my theory does not say that being produced by a reliable method is sufficient for justifiedness . . . however reliable or right a method may be, it does not confer justifiedness unless it is acquired by a suitable process, for example, metareliable processes.16 And it seems clear that Jenny does not acquire the process of trusting the Dalai Lama via a reliable process; “he said so” is not a metareliable process. In response to the Norman example, and others like it, Goldman says that Norman violates the nonundermining clause in the criteria for justified belief given above. Norman should possess a cognitive state that undermines his trust in his clairvoyant power. Goldman stipulates that Norman fails to trust his clairvoyant faculty on metareliable grounds. He should have reasoned the following way: “If I had a clairvoyant power I would have evidence of this confirmed by other reliable processes. I do not have such evidence, therefore I do not have such a power.” The point is, Bonjour has Norman acquiring a belief from a reliable process, but Goldman says he fails to acquire the trust on metareliable grounds. If he had trusted on metareliable grounds, Norman would have failed the nonundermining clause.
III. Plantinga’s objections to reliabilism Constructing a counterexample to Goldman’s theory will satisfy the following recipe: Show that (i) an agent forms a belief in accordance with a reliable process (satisfying the complications outlined in the previous two paragraphs), and (ii) this process is not knowledge conferring. Plantinga suspects that Goldman would consider the following J-rule knowledge conferring, R1: Any process is permitted whose output is a tautology. Plantinga anticipates that Goldman may not like this formulation of the rule because it stipulates that the process yields tautologies. Rather, J-rules need to be formulated with reference to the process in psychological or physiological
Competing Conceptions of a Cognitive Virtue
terms, the reliability of the processes being determined via empirical research. Conceding, Plantinga formulates the relevant rule as follows: So, choose some large class C of processes whose last items are tautological beliefs . . . there will be a (possibly complex) psychological or physiological property P had by all and only the members of C. Thus understood, the rule is as follows: R1*: Any process having P is permitted. Adding rule R1* to system S to form S* will satisfy local reliability and thus beliefs formed in accordance with R1* will count as knowledge. Now consider the following example of Paul: On Goldman’s showing, any belief permitted by S* will have a high degree of warrant, at least if it meets the condition of local reliability (as all the beliefs permitted by R1* do). But of course many such beliefs will in fact have no appreciable degree of warrant at all. Paul is in awe of logicians: if he runs across a complicated logical sounding proposition replete with p’s and q’s, he invariably believes it. He sees such a proposition in a logic book (in a “Prove or give a counterexample” exercise); it is in fact a tautology, but he can’t see that it is and has no other reason for believing it; this belief, then, has little by way of justification for him, despite its being certified by the above rule.17 The goal in this example is to specify a process which is at once intuitively vicious, and yet satisfies Goldman’s conditions for local reliability. The point Plantinga wants to draw from the example of Paul is that it matters what kinds of faculties or processes are used in order to form the belief. We may have a reliable process (any process permitted by R1* for instance) and yet the beliefs it allows clearly do not count as knowledge. According to Plantinga, if all we knew about a process was that it is reliable, we cannot infer that it is also properly functioning. Additionally, Plantinga is concerned to motivate the view that it matters how the processes or faculties were produced. To see why this matters consider the case of the Epistemically Serendipitous Lesion (hereafter I shall refer to this case simply as the lesion case). The lesion case runs as follows, As a result of this lesion (and we can stipulate, if we like, that it develops prenatally), most of my beliefs are absurdly false. It also causes me to believe, however, that I am suffering from a brain lesion. This belief of mine, pathologically caused as it is, is clearly one that has little or no warrant. But there will be a cognitive process, in Goldman’s sense, whose output is this belief;
and we may suppose that this process—call it P1—occurs only in conjunction with a lesion of the sort in question. So the result of adding R4: P1 is permitted to a right system of J-rules will itself be a right system.18 Several things are worth noting here. First, as Plantinga describes the lesion he states that it produces mostly false beliefs except there is a process P1 which, presumably, is caused by the lesion and this process produces the true belief that ‘I have a lesion’ (label this belief (L)). So, we have a putative process-type P1 which produces only the true belief (L). We can see how this is potent against Goldman in two ways. First, one may have the following intuitions; since the victim acquires (L) with no evidence and everyone else she knows does not have such a belief acquired in the way she thinks she acquired it, she clearly has a defeater for her belief that (L). The victim finds herself just believing that she is suffering from a brain lesion. The second way this example is potent against Goldman is that it focuses our attention on the pathological nature of the process in question. And, in spite of the fact that P1 is reliable, the beliefs it produces do not count as knowledge. So, we can appreciate the intuition that the victim is not justified in holding (L). I shall argue that either (1) these examples do not work against Goldman or (2) these examples do work against Goldman but they also are examples against Plantinga as well. There are two mutually exclusive routes by which one may respond to Plantinga’s example here. The first is to show that it really is not the case that Paul’s method is permitted by R1*. Goldman makes several distinctions between process types and tokens, and between second-order processes and methods. Properly understood, Paul-being-enamored-by-logicians looks more like a method than a process. The second route is to show that R1* does in fact permit only justified beliefs. I shall not argue the latter, since I think the former route is more in hand. Recall that methods are heuristics, algorithms, or learnable methodologies that one can use to transition from one cognitive state to another. Whereas processes are basic psychological processes such as perception, memory, and reasoning. But it seems clear that Paul is employing a method, i.e., logicallooking propositions are true. To be more formal, Paul seems to form his beliefs by using the following heuristic, MS: believe any logical sounding/looking proposition. This heuristic may itself be reliable in that most logic sounding propositions are true, but it seems very doubtful that MS is reliable in the case Plantinga gives since he so uses MS in a “Prove or give a counter-example” section of a logic exercise. Even worse, it seems given what Plantinga tells us about Paul, the
Competing Conceptions of a Cognitive Virtue
second-order process he used to get MS is faulty. Plantinga tells us that Paul is in awe of logicians. So, the second-order process which led to the adoption of MS may be the following, SP1: Paul is enamored with logicians and logical sounding propositions. And the force of ‘enamored’ here is important. It seems that Plantinga wants us to see Paul as someone who adopts a method of forming beliefs that is not aimed at truth. His example loses its intuitive force if we do not stipulate this. But adopting the method MS through the second-order process SP1 renders Paul’s belief-forming practices unreliable on Goldman’s showing. Goldman can say that MS is adopted using an unreliable second-order process. Perhaps Plantinga’s argument is as follows, (i) For any rule R, if R permits processes that produce true beliefs in w and worlds close to w, then R is (locally) reliable in w. Then Plantinga’s argument that R1* is reliable will look like this, (ii) R1* permits processes that produce all and only tautologous beliefs in w and worlds close to w. (iii) (The processes that R1* permits are (locally) reliable in w. (from (i)–(ii)). And (iii) is what Plantinga needs for his counterexample to work. Consider Paul again. He employs method MS. We can speak loosely here and say that Paul adopts a process of some sort—but we are to understand the term ‘process’ in a very broad, nontechnical sense to include methods as well.19 Let’s label Paul’s tautological belief that, say, ‘p → (p → p)’ as (p). Plantinga needs to say that (p) is the result of processes permitted by R1*. Recall, however, that such processes must possess the property P, where P is a physiological/psychological property had by all and only those process tokens that produce tautologous beliefs. But now, the process Paul uses to get his tautologous belief is not one that is locally reliable. Insofar as Paul forms his belief that (p) in a “Prove or give a counterexample” section of a logic book, his process of believing propositions that are logic sounding is locally very unreliable and therefore is not permitted by R1*. To see this, the proposition ‘p → (p → p)’ is a tautological proposition, which is true in all possible worlds. The belief in such a proposition, however, may not be formed in all possible worlds, or even in all the worlds within which the agent in question exists. It is unclear, given Paul’s process in world w that he would form the belief that (p) in worlds close to w. That is, a locally reliable process in w would bequeath the belief that (p) in w and worlds
close to w. The fact that the proposition is a tautology is not at all relevant to whether the process is reliable. (Processes produce beliefs, not propositions.) The only point to make here, though, is that despite the content of (p) being true in all worlds, Paul’s process in w most likely would not produce the belief that (p) in worlds close to w and again, this is especially so given the fact that Paul forms his belief in a “prove or give a counterexample” section of a logic text. Consequently, it is unclear why we should think that Paul’s process in w satisfies Goldman’s local reliability requirement. Plantinga’s example does not show that Paul’s process (or method) is locally reliable.20 The other example Plantinga gives is the case of the Epistemically Serendipitous Lesion (see above). To recapitulate briefly, the example aims to show that there is a reliable process-type P1 that reliably produces only the true belief ‘I have a lesion’ (label this belief (L)), and yet since this belief is pathologically caused, we would not consider it knowledge. So, (L) is the only belief resulting from P1, and the reason (L) is not justified is because P1 itself is pathologically caused. There are two routes Goldman can take in response. The first route aims to show that P1 is not a process type insofar as its input is content specific. The second route grants that P1 is a process type, but aims to show that it is not locally reliable. With respect to the first route, a process type is a “. . . functional operation or procedure that generates a mapping from inputs to outputs.” 21 And further, this mapping operation must have statistical properties. “Only types have statistical properties.” 22 Now P1 may possess statistical properties, even though P1 produces only one belief with the same content. The reason is that the belief may be produced on many different occasions. And insofar as there are many different occasions, there can be a truth ratio, though the truth ratio would be very easy to determine. What casts doubt on whether P1 is a process type is that it is clearly not a content-neutral operation generating outputs from inputs. Legitimate process types such as perception and reason are able to generate outputs of varying content from widely disparate inputs. For example, the process of perception covers a range of inputs. And P1 generates one output of a specific content from one kind of input. That is, though P1 performs a “functional operation” it is not a content-neutral operation, and thus the lesion is not a process type but rather a token on Goldman’s theory. One may respond and claim that Goldman gives us no reason for thinking that process tokens cannot be knowledge conferring. This is true in that Goldman leaves it an open question whether process tokens (as distinct from types qualitatively) can be knowledge conferring. My own view is that beliefs produced through process tokens need analysis on a case-by-case basis. It is not clear that tokens simpliciter are knowledge conferring. Suppose, then, that P1 is a legitimate process type. In considering the reliability of a process type, Goldman considers how the process fares in counterfactual cases. So, either (i) the lesion produces (L) in counterfactual cases, or (ii) it
Competing Conceptions of a Cognitive Virtue
does not. If (i), then the lesion is a locally reliable process. If (ii), then the lesion is not locally reliable. Of course, option (i) is the important option for us since (L) is intuitively unjustified and so if the lesion which produces (L) is locally reliable, then we have a counterexample to Goldman. And such an example is a counter example because it pulls on one of two intuitions. The first intuition seems to be either (I), (I): empirical beliefs such as (L) which simply ‘pop’ into one’s head with either no evidence or evidence against such beliefs, do not count as knowledge.23 or (PC), (PC): The process P1 is pathologically caused. Before continuing, a word or two is in order concerning (I). Most of our empirical beliefs are based on sensory data of which we are conscious. Examples include perceiving a bird, believing that I am hungry, hearing a gun fire. Empirical beliefs that just pop into one’s mind without any discernable base for such a belief—as is the case with (L)—would not count as knowledge. If Plantinga is adverting our attention to (I) in the Lesion case, then Plantinga has inadvertently produced a counterexample to his own theory. Suppose it is part of God’s epistemic plan for us (or evolutionary luck) that there is a ‘damage-detector’ faculty in the human brain according to which if there is a lesion to the cortex the agent forms the belief ‘I have a lesion’ (again, label this belief (L) as above). We can see the plausibility of this faculty being a part of our design if we think about the actual faculty of pain receptors, say, in our toe. That is, think of the damage-detector faculty in the same way pain operates when I stub my toe; upon stubbing my toe I form the belief ‘I am in pain’ or ‘I stubbed my toe’. In the same way, upon incurring a lesion on my brain I form the belief ‘there is a lesion on my brain’. So, let’s assume, 1. A damage-detector faculty is part of agent P’s epistemic design plan Ddd. And let’s define (functionally) the damage-detector faculty (DDF) as such, 2. (DDF) =df The DDF causes an agent P who possesses the faculty to form the belief ‘I have a brain lesion’ iff P incurs a lesion to the cerebrum. Now suppose, 3. P’s epistemic design plan is Ddd and P incurs a lesion to her cerebrum.
Let’s also assume for the sake of argument that DDF is functioning as it ought to in the present case. So, it follows that, 4. P forms the belief that (L): ‘I have a brain lesion’. Now, we can grant that DDF is functioning as it ought to in this case, and we may also assume, quite innocently, that DDF is functioning in the environment for which it was designed (a condition of warrant for Plantinga’s theory). And further, DDF is a reliable faculty—i.e., it is part of a good design. So, 5. (L) is the result of a faculty functioning as it ought to be in the appropriate environment, the purpose of which is truth, and this faculty is reliable. It follows that Plantinga is committed to, 6. (L) has warrant for P. (And Plantinga claims that warrant is that property which turns true belief into knowledge.) As I see it, (L) in (4) has the same properties as (L) in the lesion case used against Goldman. Both beliefs are true, and are the result of faculties or processes which are reliable, and both are cases in which the belief (from P’s perspective) pops into one’s head—thus conflicting with intuition (I) above. The only difference seems to be that (L) in (4) arises out of faculties functioning according to a design plan aimed at truth, whereas (L) in the original lesion case results from a faculty/process which gets the truth in the actual world and in relevant counterfactual cases. It is important to see that for Plantinga, a cognitive design plan aimed at truth must produce a high ratio of true beliefs in the appropriate environment. And this is very close to Goldman’s restriction that knowledge-conferring processes must be reliable in counterfactual cases, i.e., these cases being environments in which the processes get the truth in them as well. To see the present point another way, let’s compare (L) as it is produced via DDF and (L) as it is produced via P1. P1, I stipulated, is a faculty that produced truth in the actual world and in relevant counterfactual cases. DDF produces truth in the environments for which it was designed. But it is here that Plantinga does not gain anything over Goldman, for the design must be a reliable design—i.e., the faculties specified within the design plan must produce a high ratio of true beliefs in the environments for which they were designed. So DDF, insofar as it must be reliable within a specified range of environments, is on the same epistemic footing as P1 insofar as P1 must be reliable within a specified range of counterfactual cases. Therefore, if (L) is the result of either process/ faculty, then (L) is on the same epistemic footing in both cases.
Competing Conceptions of a Cognitive Virtue
One objection to the foregoing argument is to point out that there is a key difference between (L) produced by DDF and (L) produced by P1. The difference is that (L) produced by P1 is the result of a pathology—i.e., the lesion, and (L) produced by DDF is, by definition, a faculty which is part of a design plan aimed at truth. In response, I do not think that this difference is epistemically significant. Suppose, by ‘pathological process’ one means to say that the process is not part of the design plan for that possible world. But if the process is reliable, and counterfactually so, it is unclear why not being part of the design plan is so epistemically baneful. Plantinga may retort by saying that for a faculty to produce knowledge, we look not only to how it performs in this world and nearby worlds, but also to how it was caused. Etiology matters. It matters because we may deduce, then, whether the faculty is going to be locally reliable. The chief concern for Plantinga and other epistemologists is that true beliefs acquired via luck do not count as knowledge. Plantinga’s aim is to show that epistemic luck applies one stage back, namely, to the acquisition/etiology of the process. If the process is pathologically caused then if it turns out reliable, it is only by luck. The process could have easily turned out to be unreliable given its labile cause. This picture comports with Plantinga’s end project aimed to show that if one thinks that her faculties are produced in accordance with the story told by the evolutionary naturalists, then she has a defeater for her beliefs. She has no reason for thinking that her faculties would be truth conducive since evolution, the story goes, would only select for faculties that facilitate survival, not truth. I will not comment on Plantinga’s argument that naturalism is self-defeating, but I have my doubts regarding the intuition that knowledge-delivering faculties require not only reliable faculties, but also reliable faculties produced in the right way. The reason is that in conversation, Plantinga has claimed that the faculties of autistic savants deliver knowledge. But certainly a savant does not have statistically normal faculties. There is one savant who, if a radio were on, is able to discern the place in a room at which point the sound meets. It is well known that such cognitive abilities are caused by abnormal neurological functioning. But if Plantinga grants that a savant’s beliefs count as knowledge, then their faculties are not pathological—and this strikes me as odd. The question ‘what counts as a pathological faculty?’ is not so clear anymore. Suppose a pathological process is any process not caused by God—I suspect, this definition comes close to what Plantinga would have in mind. On this rough characterization, savants, he may say, are given their abilities by God. But suppose God causes Smith to be struck by lightning. The strike produces a neuro-chemical change resulting in Smith acquiring savant-like abilities. According to Plantinga, Smith’s beliefs are warranted. Suppose, however, that Smith gets struck by lightning but God merely allows it to occur, as part of God’s overall plan to ensure some irregularity and chance in the world. Smith forms savant-like abilities again. Do his beliefs count as knowledge in this second case,
even though they were not brought about by the action of God? I suspect that they do, and thus being designed by God does not matter. The point of this section is to illustrate the strengths of virtue reliabilism, and to show that a proper function theory makes little epistemic gains over virtue reliabilism. Doing so allows me to focus on virtue reliabilism as the sole competitor to virtue responsibilism (within the virtue camp). Throughout this work, I shall pause to highlight the gains of responsibilism over reliabilism especially as these gains pertain to testimonial knowledge and moral knowledge.24
IV. Virtue responsibilism In this section I lay out some deficiencies in virtue reliabilism. I then return the favor to virtue responsibilism. I aim to show both how responsibilism addresses the deficiencies in some other virtue accounts, and to pinpoint the potential problems with responsibilism which are addressed in later chapters.
Deficiencies for virtue reliabilists There are two key problems with Goldman’s reliabilism. First, reliabilism is not Gettier-proof and second, it cannot explain why knowledge is better than mere true belief. I deal with the value problem below, for now I treat the claim that reliabilism is not Gettier-proof.
I. Reliabilism is not Gettier-proof The feature that makes reliabilism open to Gettier-like cases is that it is committed to the proposition, (T): Not all beliefs produced from a reliable process are true. Goldman includes a truth condition in addition to reliably formed belief as requisite conditions for knowledge. Insofar as Goldman is committed to (T), Goldmanian reliabilism is open to what double-luck cases introduced in the previous chapter. The typical double-luck scenario would stipulate that an agent forms a reliably formed belief that perse improbable, turns out false. Then, the scenario stipulates that by a freak accident, the belief turns out true after all. I agree with Zagzebski that double-luck scenarios capture the essential structure of Gettier-type cases. Consider the following scenario developed by Zagzebski against Plantinga’s proper function account (which applies equally to Goldman’s theory): Suppose that Mary has very good eyesight, but it is not perfect. It is good enough to allow her to identify her husband sitting in his usual chair in the
Competing Conceptions of a Cognitive Virtue
living room from a distance of fifteen feet in somewhat dim light . . . She has made such an identification in these circumstances many times. Each time her faculties have been working properly and the environment has been appropriate for the faculties.25 Against this background, we can specify that in such an environment Mary acquires a false belief, i.e., the belief My husband is sitting in the living room turns out false. To quote at length, Suppose Mary simply misidentifies the chair sitter, who is, we’ll suppose, her husband’s brother, who looks very much like him. Her faculties may be working as well [or locally reliable] as they normally do when the belief is true and when we do not hesitate to say it is warranted in a degree sufficient for knowledge. It is not a question of their suddenly becoming defective or, at any rate, more defective than usual, nor is there a mismatch between her faculties and the environment. No one is trying to surprise or fool her or anything like that. Her husband and his brother may not even know she is in the house, so the normal environment has not been doctored as it is in the fake barn case . . . So her degree of warrant is as high as it usually is when she correctly identifies her husband because even in those cases it is true that she might have misidentified the chair sitter if it had been her husband’s brother instead . . . So in the case we are considering, when Mary forms the false belief, her belief is as warranted as her beliefs normally are in these circumstances. In spite of well-functioning faculties and a benign environment, she just makes a mistake . . . We can now easily amend the case as a Gettier example. Mary’s husband could be sitting on the other side of the room, unseen by her. In that case her belief My husband is sitting in the living room is true and has sufficient warrant for knowledge on Plantinga’s account, but she does not have knowledge.26 The problem, as Zagzebski sees it, is that any theory of knowledge which defines knowledge as true belief plus X where X does not entail getting the truth, is subject to a double-luck case. For Goldman’s theory, he defines knowledge as true belief plus local reliability, where being locally reliable does not entail that the belief is true. In the case of Mary, that she misidentifies the chair sitter, in spite of forming her belief with a locally reliable faculty (i.e., perception) is bad luck and that her husband happens to be in the living room is a stroke of good luck. So, when she forms the belief that ‘My husband is sitting in the living room’ she forms a reliably formed true belief that does not count as knowledge. The point here is purely a conceptual one, if the property X in a definition of knowledge does not entail truth, then it is possible to construct a case of double luck.
The logical remedy to this problem is to fill in for X conditions that entail getting the truth. For the responsibilist, the solution is as follows, In the case of Mary’s belief that her husband is in the living room, she may exhibit all the relevant intellectual virtues and no intellectual vices in the process of forming the belief, but she is not led to the truth through those virtuous processes or motives. So even though Mary has the belief she has because of her virtues and the belief is true, she does not have the truth because of her virtues.27
II. The value problem This second problem with reliabilist theories in general is the value problem. The basic intuition behind the value problem is, in Zagzebski’s words, A reliable process is good only because of the good of the product of the process. A reliable espresso-maker is good because espresso is good. A reliable water-dripping faucet is not good because dripping water is not good. Reliability per se has no value or disvalue. Its value or disvalue derives solely from the value or disvalue of that which it reliably produces. So the value of the product of the process is transferred to the process that produces it. But the value of the process is not transferred back again to the product . . . Similarly, a reliable truth-producing process is good because truth is good. But if I acquire a true belief from such a process, that does not make my belief better than it would be otherwise.28 A reliable process does not transfer value onto its product. A reliable cognitive process does not make true belief better. And if knowledge is a cognitive state better than mere true belief, then any form of reliabilism cannot account for the value of knowledge. In more syllogistic form, the argument runs, A. Knowledge is a cognitive state that is better or more valuable than mere true belief. B. The various forms of reliabilism can account only for the value of true belief. C. ∴ The various forms of reliabilism cannot account for the value of knowledge. The problem can also be formulated the following way; assume that true belief plus x (e.g., let x be justification or warrant) is sufficient for knowledge. Let’s also assume that justification is valuable only because it leads to true belief. That is, justification (or warrant) is instrumentally valuable. But then, that the
Competing Conceptions of a Cognitive Virtue
belief is justified does not add value to it over and above the value that accrues to the belief if it is true. So, any theory of knowledge which makes true belief plus something else equal to knowledge, and that something else is only of instrumental value, then that theory of knowledge cannot account for the value of knowledge (assuming the value of knowledge is more than the value of mere true belief).29 What does explain the value of knowledge? The responsibilist will say that knowledge is a state of the agent, not an output of the agent. Reliabilism and proper function theory both hold to a machine-product model of relating knower to knowledge. But as the analogy with the espresso machine shows, no machine can confer value onto its product. Zagzebski claims that a more plausible way to model the relationship between knower and knowledge is to model such a relationship on an act-agent model. “Knowing is not related to the knower as product to machine, but as act to agent. Knowing is part of the agent.” 30 But once we make this move in analysis, forming beliefs looks much like acting. And, If believing is like acting, it can be virtuous or vicious. A belief can acquire value from its motive, in addition to the value it may have in being true. The properties of true believing that make it better than mere true believing are properties that it obtains from the agent in the same way good acts obtain evaluative properties from the agent.31 The analogy with good acts is instructive. Performing acts of love are good. But consider a person who devotes several years of her life to the mission field, working tirelessly on various humanitarian and agricultural-sustainability projects. She devotes herself in this way, though, because she loves the acclaim she gets from the people for whom she works. Or suppose she devotes herself so as to patch up thoughts and feelings of an emotionally painful break-up. Though the good of her actions stand, they gain value if they are done instead out of finer motivational states, such as the motivation to promote the good of others. The analogy holds for cognitive acts as well in that acquiring true beliefs on contentious moral issues, say, are better if they are formed by virtuous inquiry as opposed to, say, close-minded conformity to a position one has long held. Though a person who is close-minded may end up with true beliefs; that is all they are. Her beliefs are not the result of good actions. To the extent, then that responsibilism defines knowledge partly with reference to virtuous acts of the agent, the value of knowledge is accounted for. To summarize, a successful account of knowledge will need to be Gettierproof and obviate the value problem. We need a definition that avoids double-luck scenarios and one that makes essential reference to cognitive acts— in order to account for the value of knowledge. The notion of an intellectual virtue modeled on the notion of moral virtue fits both criteria. If beliefs are
formed by cognitive acts and acts can be either virtuous or vicious then there is a strong motivation to analyze knowledge with reference to intellectually virtuous acts.
Deficiencies for responsibilism32 The distinctive feature of responsibilism is that a virtuous act is defined partly with reference to a motivational component. It is the motivational component that gives the responsibilist trouble. The trouble is clearly laid out by William Alston, One pressing problem concerns the question of how the account is to be applied on the “not at all voluntary” end of the spectrum. As Zagzebski indicates, ordinary perceptual and memory beliefs are obvious examples. Are we to think of perceivers as motivated by intellectual virtue to form beliefs as they do? It would seem that motivation, whether by virtue or otherwise, has nothing to do with the matter.33 The spirit of this objection is this; if knowledge requires an act of intellectual virtue and intellectual virtue has a motivational component, then in order for beliefs such as perceptual, memorial, or introspective beliefs to count as knowledge, they would require acts of intellectual virtue. But presumably, when I perceive something there is no motivational component in the description of my perceiving that X. Hence, the responsibilist version of VE cannot account for simple perceptual or memorial beliefs and yet these beliefs on suitable occasions certainly count as knowledge. I shall refer to this as the “unmotivated belief objection.” Alston and others who argue against Zagzebski along these lines are implicitly endorsing the following two intuitions: (i) The perceptual process is largely automatic or unconscious and (ii) Virtuous motivations are conscious or intentional. Can the responsibilist dodge this critical objection? I do not think that previous responses fare well. Consider Zagzebski’s own attempt to obviate it. How does a person of intellectual virtue act when it comes to forming beliefs based on sense experience or memory? Presumably, she is sometimes skeptical of her own senses, and she sometimes doubts her own memory, as in the case when it is weak and she has good contrary evidence . . . But we would assume that most of the time she does not doubt or even reflectively consider her perceptual or memory beliefs. She does not because she maintains a presumption of truth in such cases until she is given reason to think otherwise. Such an attitude is itself an intellectually virtuous one; to act otherwise is to exhibit a form of intellectual paranoia.34
Competing Conceptions of a Cognitive Virtue
To be more precise, the argument seems to look like this; (1) Holding a presumption of truth (trust) in one’s senses (or memory) is intellectually virtuous. And in order to get the conclusion that perceptual knowledge (PK) requires an act of intellectual virtue, Zagzebski needs the following premise: (2) An agent S has PK only if S trusts in the deliverances of her senses—to do otherwise is “intellectual paranoia” and clearly this does not count as knowledge. From (1) and (2) it follows that, (3) ∴ S has PK only if S is intellectually virtuous. Zagzebski’s aim is to deny that an instance of a perception can be described without any reference to virtues and thus without any reference to motives. There are several problems with the argument from (1)–(3). I see two major difficulties. First, Zagzebski’s argument for (1) is deficient in itself. But even more pertinent I shall argue that either (1) or (2) is true, but not both. And since at least one must be false, the argument is unsound. Let me begin with the first difficulty. Zagzebski’s argument for (1) seems to be that “. . . to act otherwise [than trusting in the deliverances of my senses] is to exhibit a form of intellectual paranoia” (p. 280). Presumably, if one does not trust, she exhibits intellectual paranoia—and intellectual paranoia, we may assume, is vicious. But, Zagzebski seems implicitly to conclude from this that A. If a person who distrusts her sense powers at time t is vicious, then one who trusts in them at t would be virtuous. But it is hard to see why (A) is the case from what has been said? It seems one can avoid being vicious without being virtuous. From the fact that not trusting in one’s sense powers is a vice, it does not follow that trusting in one’s sense powers is a virtue. Concerning the second difficulty, the structure of my argument is that if trust is construed strongly as having a motivational component, then (2) is false. If trust is construed weaker as being merely an act of assenting or believing in what I see, then (1) is false. Consider construing trust strongly. When we look at (2) there is no reason to think that one needs to be motivated properly in order to see that an apple is on the table—absent any reasons to doubt that it is an apple. In simple cases of perception it seems we simply accept the
deliverances of our powers without any deliberation or intentional state—we simply receive the input. Now, if we weaken the concept of trust to denote some base-level assent,35 then (2) is probably true, but (1) is suspect. It is difficult to see how a base-level acceptance of my percepts is a virtue. There is no motivational state attending it, no deliberation, no intentional state, and the capacity for such acceptance is not something I acquire. Zagzebski’s first argument, then, has not shown that a responsibilist version of VE can account for PK. In fairness, Zagzebski has responded to my argument here, attacking my claim that trust is not a virtue. She says, “. . . the kind of trust we have in the senses naturally is not a virtue, but it becomes a virtue when we learn that our senses are trustworthy . . . Some natural dispositions are good things and become virtuous when they become part of our reflective nature.” 36 And in order to become part of our reflective nature, our natural dispositions need training. “Even memory, hearing, and vision need some training . . . Even the simplest perceptual beliefs involve conceptualization of perceived ‘data.’ ” 37 Now, insofar as the use of my faculties needs to be properly trained, there is room to describe PK with reference to a virtue. Zagzebski’s response shows promise but suffers from two persistent problems. First, it is plausible to suppose that our cognitive capacities develop naturally. The neurological correlates for perceptual processing may follow a developmental path uninitiated by any of the agents’ motivations. Second, even if there were motivations to develop one’s cognitive faculties, the key issue is whether there are any motives in the description of a single belief being formed. There may be motivations required for developing one’s faculties—perfecting them in some way—but the unmotivated belief objection deals with belief acquisition (not faculty perfection) and it is still not clear whether motivations can fit into a description of belief acquisition. Abrol Fairweather attempts to obviate the unmotivated belief objection as well. Fairweather begins by affirming Sosa’s distinction between animal and reflective knowledge. Animal knowledge consists of true belief in response to the direct impact of one’s environment without reflection or understanding. Reflective knowledge is true belief in response to the direct impact of one’s environment plus an understanding of that belief—an understanding which may include knowing how the belief came about and the relationship between that belief and other beliefs one possesses. Continuing, Fairweather cites Sosa’s conclusion regarding human knowledge, Note that no human blessed with reason has merely animal knowledge of the sort attainable by beasts. For even when perceptual belief derives as directly as it ever does from sensory stimuli, it is still relevant that one has not perceived the signs of contrary testimony. A [being endowed with reason] automatically monitors his background information and his sensory input for contrary evidence and automatically opts for the most coherent hypothesis even when he responds most directly to sensory stimuli.38
Competing Conceptions of a Cognitive Virtue
The idea is that human perceptual knowledge requires both the mechanisms of perception, and a background monitoring system (BMS). And here is Fairweather’s key point, We may seek to insulate certain beliefs from falsification and hence ignore cues that suggest their falsity. This would still be a monitoring system, but a defective one. Thus, we cannot simply take it for granted that the monitoring system in reflective cognition is keyed to preserving the truth conduciveness of our doxastic system.39 Fairweather concludes from this that, “Our monitoring system must be governed by a desire for the truth, or a kindred alethic end, in order to function as Sosa describes.”40 As I see it, the best way to reconstruct this argument more clearly is the following: (4) S possesses PK only if S’s BMS is cued to alethically relevant features of the environment.41 (5) S’s BMS is cued to alethically relevant features of the environment only if S has a motivation for the truth. (6) ∴ S has PK only if S has a motivation for truth. The argument for (4) is twofold; first, perception consists of both the mechanisms of perception and the operation of the BMS. Second, our BMS must be cued to alethically relevant features of the environment, for if it were not, it would not preserve the truth conduciveness of our sensory inputs—it may make our sensory outputs unreliable. So, both (i) there is a BMS which constitutes PK and (ii) it must be aimed at alethically relevant features of the environment (in order to be reliable) suggest that S possesses PK only if BMS is cued to alethically relevant features of the environment. The argument for (5) seems to be that one cues her BMS to nonalethically relevant features of the environment she “insulate[s] certain beliefs from falsification and ignore[s] contrary evidence.” For example, beliefs produced by prudential motives or wishful thinking come about if one does not focus on the alethically relevant features of one’s epistemic environment. Presumably, since one can use her BMS in a defective way, what ensures that it is used rightly is a function of having a desire for truth. Let me begin my response with a clarification. Fairweather uses Sosa’s line about reflective knowledge to support (4). But Sosa himself does not say anything about our BMS needing to be governed by a desire for truth—at least not in the Monist article. What Sosa says about this BMS is that it is reason at work which aids our perception and memory. In concluding the very paragraph from which Fairweather cites, he (Sosa) states, “The beliefs of a rational animal hence
would seem never to issue from unaided introspection, memory, or perception. For reason is always at least a silent partner on the watch for relevant data, a silent partner whose very silence is a contributing cause of the belief outcome.”42 Sosa thinks, then, that our BMS is essentially reason on the look out for ‘fishy’ data. But if that is so, (5) should read, (5’) Reason is cued to alethically relevant features of the environment only if S has the right motivation (a motivation for the truth). Now, (5’) entails, (5”) It is not the case that both, S does not possess a motivation for truth and S’s reason is cued to alethically relevant features of S’s environment.43 But (5”) suggests that reason is not in itself aimed at truth regarding perception. That is, (5”) suggests that reason must be paired with the right motivation in order to ensure that reason (cum perception or memory) delivers truths. What I shall show presently is that if one accepts Sosa’s description of the BMS as reason (though in silent-lookout mode) and one accepts (5), we get an absurd consequence, that is, if (5) entails (5”), and (5”) entails denying knowledge to S when S clearly knows, then (5) entails an absurdity. To see my point clearly let’s take the following three cases, and let’s assume for each case that there is in fact a computer on Jones’ desk. Case 1: Jones perceives a computer on his desk and believes it is so—(no motivation). Case 2: Jones perceives a computer on his desk but out of an excessive desire to prove to his philosophical colleagues that doxastic voluntarism is true, he does not believe that he perceives a computer on his desk—(a motivation for falsehood). Case 3: Jones perceives a computer on his desk (and believes it is so), but also Jones is motivated only to believe the truth. If asked about his cognitive desires, he answers with his head cocked high, “I desire truth!”—(a motivation for the truth). Case 2 I think is not really a possibility—it amounts to the following claim: ‘P, but I do not believe P’. For my purposes then I shall stick to case 1 and case 3. In case 1 Jones has no motivation regarding his perceptions. But certainly, we would not say that Jones does not have PK in case 1. The addition of a desire for the truth in case 3, does not add anything epistemically to Jones’ perceptual belief. It is certainly nice of Jones to have such a desire, but this is irrelevant to
Competing Conceptions of a Cognitive Virtue
whether we attribute to him knowledge that his computer is sitting on his desk. But notice, (5”) entails that Jones does not know in case 1—for Jones does not have any motivation for the truth regarding his perception. But surely, if to accept (5”) forces us to deny that Jones has PK in case 1 then certainly (5”) should be rejected. In sum, my argument can be framed in the following way, either reason is aimed at truth automatically (with respect to perception) or it is not. If it is automatically keyed to truth, then there is no need to appeal to motivations in order for reason to deliver truths. If it is not keyed to truth automatically then we get the highly implausible result that Jones in case 1 does not know. The discussion thus far has left us with at least two lessons. First, the idea of a motivation is strongly associated with the concept of intentionality and/or controlled processing. And second, perceptual processing is understood to be automatic and unintentional (perception involves ‘taking in’ the environment). Therefore, the ideas of motivation and perceptual processing are disparate. If we were to weaken the concept of motivation to be compatible with automatic processing, then it is suspect whether such motivations can be considered virtuous motivations. Virtues are commonly thought of as being acquired dispositions which perfect the power in question (the will for moral virtues and the intellect for intellectual virtues). If we strengthen what we think about perceptual processing so as to include some level of control, then we run up against strong intuitions. In the next two chapters, I shall show: (I) perception/memory is not wholly automatic processes, and (II) the motivations that are essential for perceptual/memorial processing are ‘thick’ enough to fall under the rubric of a virtue. The benefit of the approach I take is that it impugns directly the intuitions supporting the unmotivated belief objection, whereas the previous attempts have not.
V. Conclusion In conclusion, I have tried to accomplish several tasks. The first was to elucidate various forms of virtue epistemology. This is important in that it provides a backdrop from which to appreciate one of the debates occurring between virtue epistemologists. One of these debates is what conception of an intellectual virtue is the right one. I have argued that responsibilism has advantages over a virtue reliabilist theory. Responsibilism is Gettier-proof and has a natural way of avoiding the value problem. I have also argued, however, that responsibilism has its own problems. In the last section I outlined the unmotivated belief objection which aims to show that responsibilism cannot account for PK. If a theory of knowledge cannot account for PK then it is a deficient theory. The important point to draw is that though responsibilism has advantages over other theories given certain criteria, it is deficient with respect to other criteria,
i.e., accounting for clear cases of knowledge. If one can show, however, that responsibilism can account for PK then we would have a theory that has the advantages of being Gettier-proof, explains the value of knowledge and is not subject to the unmotivated belief objection. Satisfying all criteria is a significant argument in favor of responsibilism over its competitors. I turn in the next two chapters to argue that responsibilism is not subject to the unmotivated belief objection. I do so by challenging the very intuitions upon which the objection rests. I illustrated above that none of the previous attempts to obviate the objection is successful because they do not address directly the intuitions which inform the objection.
A Virtue Account of Perception
I. Introduction We typically think that perception is largely an automatic process. The unmotivated belief objection rests on these very intuitions. This chapter aims to deflate the objection by showing that successful perceptual processing (P-processing) requires attention, and attention is directed by one’s epistemic motivations. The chapter references empirical research in the area of consciousness, attention, and perception. The reader should note that though I focus on visual perception, the arguments contained herein are applicable to auditory, tactile, and other modes of perception. The chapter progresses as follows: in the next section I lay out definitions of key terms and present the outline of my argument. Subsections argue for each premise in turn. In the third section I turn to some objections. This section addresses the most common objections which can be brought against the argument laid out in this chapter. The fourth and final section contains some concluding remarks.
II. Perception, attention, and consciousness There is one major fact of P-processing that suggests the presence of motivation, that is, the role of focused attention in attending to stimuli within one’s environmental field. Attention is a slippery concept but a rough characterization of it is that, attention is the directed allocation of cognitive resources to fully process a stimulus. If P-processing is a function of attention then intuitively, we have a reason for thinking that motivation is involved in P-processing. When one attends to a stimulus in her environment, this act of directing one’s attention is a function of one’s motivations. In fact, most psychologists agree upon the interrelationship between attention and motivation. Ralph Ellis states, “The idea that attention is motivated, of course, is not new in psychology, nor is the view that limbic and subcortical emotional processes play a role in directing attention.” 1 The work undertaken here is to show that attention is required in order to have perceptual knowledge and that the kinds of motivations involved in directing one’s attention are virtuous.
Attention But what notion of attention and consciousness is being used here? What notion of a motivation is being used here? In their book Inattentional Blindness, Arien Mack and Irving Rock describe attention in this way, “. . . the term attention is used to refer to the process that brings a stimulus into consciousness.” 2 Attention can be thought of as a conduit or a gate through which perception passes into consciousness. This is not to suggest, however, that perception and attention are the same process. If attention and perception were the same process, there could be no attention without perception. Clearly, however, “It is not an uncommon experience to be looking for something or keenly awaiting its appearance in the absence of perceiving it. . . .” 3 Furthermore, the objects of attention need not be perceptual objects. One can attend to one’s ideas, thoughts, or imaginations.4 So, the idea that attention and perception are the same is untenable. It is more proper to say that attention leads to conscious perception but is not identical with it.
Consciousness With respect to conscious perception, an agent consciously perceives X only if there is a subjective awareness of the content of X. The term ‘subjective’ refers to the ‘what it’s like’ property of conscious experience, and the term ‘awareness’ I will leave as primitive—though some potential misunderstandings should be dispatched.5 I do not take ‘awareness’ to be a factive state. Awareness of X does not entail the existence of X. Ordinary language usage supports this view, for example, I may become aware of Jones’ testimonial report that X, but because it is a fallacious report, X is not the case. Or, a better example for the atheist reader, it is often said by religious folk that they become aware of the presence of God in their lives. But if God does not exist, then awareness of X does not entail X. What awareness means in ordinary language use is the apprehension of some visual or conceptual content. The importance of this clarification should be made explicit. I have defined conscious perception with reference to awareness. If awareness is understood as essentially a factive state, then there would be no conscious perception in cases of delusions or hallucinations. But neither of these experiences is of a factual state, yet we would say they are cases of conscious perception. My clarification above, however, indicates that such cases are not counterexamples to my characterization of conscious perception. Another potential counterexample to my commitment that knowledge that p requires having been conscious of the contents of p comes from cases of unconscious perception. Unconscious perception (otherwise known as implicit perception) of stimulus X occurs if an agent does not verbally report perceiving X but evinces other signs of having taken in the information. A primary example of this is blindsight. Patients with blindsight suffer from cortical damage
A Virtue Account of Perception
rendering them blind to a specifiable part of their visual field. Though they report seeing nothing, they perform quite well on tasks which require them to judge the presence of an object, or even localize an object in their blind field. Another example of implicit perception are cases involving priming effects. The idea behind priming is that an unconsciously perceived stimulus can prime the subject to perceive another similar stimulus either more quickly or accurately than without the prime. This evidence indicates that some level of cognitive processing is done without attention. These cases are important since they challenge the idea above that perceptual knowledge of p requires being or having been conscious of the propositional content of p. Intuitively, without reporting any percept (as is the case with unconscious perception) there are little grounds to say that the person was conscious of the stimulus. And, intuitively, some of the cases of unconscious perception deliver perceptual knowledge. I raise this problem here, given its importance, but reserve a response for below. Related to conscious perception is an important assumption I make in my argument. The kind of knowledge with which I deal in this project is primarily propositional knowledge. In this chapter I am not concerned with epistemic states whose objects are nonpropositional objects (as is the case with, say, understanding or wisdom). Philosophers have typically understood propositional knowledge to include two other commitments, the first of which is that propositional knowledge requires belief. The second assumption is that an agent S has perceptual knowledge of p only if S is or has been conscious of the contents of p. Paradigm examples of perceptual knowledge, such as ‘there is a table in the room,’ requires the agent to have been (at some time) conscious of the table being in the room. (To avoid misunderstanding, I do not assume, nor do I think it true, that every case of conscious perception leads one to form a belief. In cases of known optical illusion, we do not believe what we consciously perceive.) Having delineated my key terms, the basic outline of my argument is as follows: 1. 2. 3. 4.
Attentionp is necessary for perceptual knowledge. Attentionp is a function of an agent’s motivation(s) M1. Motivations in class M1 are epistemic motivations. ∴ Perceptual knowledge is a function of one’s epistemic motivation(s).
The rest of this section is devoted to defend each premise in turn.
A. Perceptual knowledge requires attention There are two phenomena known to cognitive psychologists as Inattentional blindness and change blindness. Both phenomena suggest that that which is not attended to does make it to conscious perception. In their book
Inattentional Blindness Arien Mack and Irving Rock outline a series of experiments they conducted over a span of several years on over 5,000 subjects devoted to show the necessity of attention in conscious perception. The basic outline of the experimental design runs as follows: Subjects are given the perceptual task of discerning the length of two cross-bars to be presented on a computer screen in front of them. The presentation phase consisted of a fixation point followed by a blank screen for several milliseconds and then a cross-bar was presented directly over the fixation point for 200 ms. Fixation point – 200 ms
Cross-bars for 200 ms
Critical trial with critical stimulus
This was done for two trials. On the third trial, a geometrical shape was presented in one of the cross-bar’s quadrants concurrently with the presentation of the cross-bar itself, and the subjects were not informed that there would be this added stimulus. The third trial is called the critical trial and the extra stimulus (the geometrical shape) is called the critical stimulus. This experimental design is called parafoveal presentation of the critical stimulus. It is called parafoveal because the critical stimulus is presented away from the subject’s fixation point. Foveal presentation of the critical stimulus was also performed on subjects in which the cross-bar was presented parafoveally and the critical stimulus was presented at fixation. I comment on the results of foveal experimental design below. In each of these experimental designs, subjects were asked, both the following questions “what bar was longer than the other?” and, “did you see anything else?” If the subject answered ‘yes’ to the latter question, she was then asked what quadrant the ‘something else’ was located, its shape and its color. According to Mack and Rock, the results of these experiments each show a phenomenon called inattentional blindness (IB). They conclude, “Our results drew us ineluctably to the view that there is no conscious perception at all in the absence of attention, and therefore no [explicit] perceptual object can exist preattentively.” 6
Results This section outlines in detail the results of the IB trials. I do this to ground the argument that perceptual knowledge requires attention. The reader who does not need such detail may skip this section and go one to the next. The results are as follows. For the parafoveal trials (slightly modified throughout the experimental process) between 23 and 28 percent of the subjects
A Virtue Account of Perception
reported not seeing anything other than the cross-bars on the critical trial.7 Of the subjects who did report seeing something, most were adept at judging the location of the critical stimulus, but were at or below chance at judging its shape. Compared to the control group, these results were significant. The control group was simply given the task of looking for the critical stimulus. All reported seeing something, and were adept at both location judgments and shape discernment. One may draw the conclusion from the experimental-group results that perception does in fact take place; 23–28 percent is a fairly low percentage of subjects who fail to perceive the critical stimulus. This poses trouble for my defense of (1) since there were many subjects who did report seeing something, and were accurate in reporting the location of the critical stimulus. It is that we have a putative case in which perception takes place without attention, and the accuracy of the judgments certainly makes them candidates for perceptual knowledge. With respect to the judgments of location, there are at least three explanations: First, one can say, that such judgments are the result of priming effects and not conscious perception. Before filling out this response, let me say a word about this distinction. Visual priming refers to the effect that nonconsciously processed visual information can have on subsequent consciously processed visual information. More simply, information that is not processed to a conscious level can exert an effect on conscious processing of subsequent information. For example, in a typical lexical decision task, subjects are presented with a “target” word for several hundred milliseconds. They are then asked to decide whether the target word is a real or pseudo word. Performance on this task is more accurate and quicker if a semantically related “prime” word is presented for several milliseconds immediately prior to the presentation of the target word. For instance, if the target word is ‘lemon’, a prime word of ‘sour’ significantly enhances performance on the lexical task. (Whereas a prime word of, say, ‘house’, does nothing.) Now, if asked to report on any other word they may have seen, the subjects only report perceiving the target word.8 Returning to the accurate location judgments, information regarding the location of the figure may be processed to a nonconscious level, but this level of processing was sufficient to ‘guess’ accurately when asked by the experimenter if she saw anything.9 What explains the accuracy of their guesses is that full processing required for location is inchoate—all that is required is the registration of light energies. Consequently, the requisite information processing required to be primed for judgments of location would be even more incipient. So, priming effects could be very high without any attention so long as the perceptual task is not demanding (and judgments of location clearly are not). To be clear, this first response is meant to impugn our intuitions that the location judgments count as knowledge, and they do not because they may not be the result of conscious perception.
The second explanation is that some subjects did consciously perceive the location of ‘something’ but this was a function of one’s attention, a level of attention sufficient enough to merely judge location. Mack and Rock themselves give such an explanation (though for a different phenomenon). The idea is that attention is diffused throughout the boundaries of the cross-bars. Since the subjects were required to discern the respective length of the bars their attention would be diffused to include the outer perimeter of the cross. And since the critical stimulus was presented within that perimeter, the subjects would be able to discern “something” in one of the quadrants. (A reminder here is that processing sufficient to discern location is very inchoate since what is involved is not recognition or classification of the stimulus. All that is required is registering the presence of some conspicuous light energy.) So, the subjects in the parafoveal trials may consciously discern location of the critical stimulus correctly (and thus know), but this would be a function of attention. Third, the accuracy of the location judgments is a case of perceptual knowledge without attention.10 Of course, it is the third option which poses a counterexample to the first link in my argument for (1). As I shall argue, however, the next set of results from the foveally presented experiments makes the third option very implausible.11 In the foveally presented trials, subjects in the experimental group were given the same task as outlined above, i.e., to discern which of the cross-bars is longer than the other. The difference is that the center of the cross-bars was presented away from the initial fixation point and the critical stimulus was presented at fixation. The idea is that if a stimulus were presented at fixation, this would increase the chances that the image would stimulate the fovea—the back part of the eye that is responsible for translating light energies into neural codes (coded information). Images which land on the fovea, given the fovea’s understood role in vision processing, are processed better than parafoveally presented stimuli. So, the assumption going into the next set of experiments was that if the critical stimulus is presented foveally, then little IB effects would result even with the task of discerning the cross-bar lengths. In fact, the results were that between 85 and 89 percent of the subjects did not see anything at all!12 Mack and Rock offer an explanation for this result according to which the subjects actively inhibit attention to the foveal area because they know they need to perform the task of discerning the line lengths of the cross-bars. So, in spite of the fact that light energy from the critical stimulus does hit the fovea, there is no further processing of that information, but rather there is redirection of attention to the cross-bars. Not surprisingly, those in the control group were adept both at discerning whether something was presented, and what shape the critical stimulus was. The implication of these results is that they illustrate the role attention plays in fully processing external stimuli. The foveally presented experiments illustrate that even if light energies are present on the fovea, no further processing
A Virtue Account of Perception
occurs—at least no processing nearly sufficient for conscious perception. The foveal experimental design aims to maximize the level of incoming stimulus energy coming into the vision system and yet minimize the level of attention drawn to that information. The results indicate that conscious perception does not occur without attention.
B. Attention is informed by epistemic motivations Common sense tells us that ‘attention’, or ‘attending to’, or ‘focused attention’ all require some motivational state in the agent. Even if this much is granted, we are a long way from responding to the unmotivated belief objection, for the motivation involved must be a motivation that can serve as a component of a virtuous act. The task in this section is to illustrate both that attention must be motivated, and more importantly, these motivations involved are components of a virtuous act.
Enactive view of consciousness An enactive theory of consciousness comprises two theses which are that (A) consciousness is a function of subcortical brain activity (an area of the brain associated with emotional and motivational processing) and (B) consciousness can result only if actions are performed by a self-organizing system. The actions which are relevant serve the purpose of furthering the organism’s pattern of activity, and this pattern of activity is largely due to the motivational interests and plans of the agent. Ralph Ellis characterizes the enactive view the following way: On the enactive view, the element of consciousness, above and beyond mere nonconscious data processing, is contributed by brain areas associated with the organism’s motivational interest in processing certain kinds of environmental stimuli and ignoring others.13 The relevance of presenting the enactive view for my argument is that the enactivist takes it that consciousness in general, and conscious perception in particular is a function of the motivational interests of the agent. Ellis proposes the following: All phenomenal consciousness, even empirical consciousness, must be motivated; thus the “what it’s like” of phenomenal experience is inseparable from the emotions permeating it. Although we often do not pay attention to the motivational components of our perceptual and rational functions, the felt quality of these functions stems from the fact that they are continuously in a process of being motivated by the emotional purposes of the organism.
Virtue Epistemology For example, visual cortex activation is unconscious of red unless the emotional midbrain and anterior cingulate “look for red” in a motivated manner. That is, all the afferent aspects of brain function (those which receive stimulation from external sources)—such as the “sensory area” of the occipital lobe and the V4 visual area—can be completely activated, in just the way they are in a perception, but with no perceptual consciousness of the object occurring. Consciousness occurs only when the efferent system, beginning with the emotional midbrain, prompts corticothalamic loops, including frontal areas such as the anterior cingulate as well as the extended amygdala and reticular formation, to start looking for environmental items which might be of interest to the organism, in light of its emotional-motivational purposes.14
The basic point is that if one is an enactivist, then consciousness will be a function of one’s motivational interests.15 Enactivism can be defended on numerous grounds. First, Ellis points out that enactivism is confirmed by several research strands, one of which we covered above. The possible connection between inattentional blindness and enactivism is suggestive because, if attention must be engaged before contents can register in consciousness, then there may well be a subcortical and limbic [emotional/motivational] mechanism that drives attention independently of the processing of the input, and such a mechanism could well be a necessary precondition for consciousness . . .16 Ellis lays out a total of six strands of evidence in which the enactive view uniquely predicts the empirical results. Reactive theories of consciousness and informational processing, however, falter. In addition to the empirical support that the enactive theory enjoys there is also an important philosophical argument, according to which it can deflate the Knowledge Argument proposed by Frank Jackson.17 The Knowledge Argument runs as follows: The strict physicalist asserts that every conscious event E is equivalent to a set of physiological properties P such that if one were to know all of the properties of P we would then know all the properties of E. But there is at least one property of E which cannot be accounted for by the set P, and that is the phenomenal property of E namely its subjective/“what its like” aspect. That is, one can know everything there is to know about the set of physiological properties associated with E, but is bereft of knowing what it’s like to be in state E. Consequently, consciousness is not reducible to the physiological properties. This conclusion forces those with physicalist sympathies to propose a theory of consciousness that (i) does not appeal to a nonphysical cause to consciousness and (ii) “it must explain why the person executing a physical process could
A Virtue Account of Perception
have different information about it from any that is available from the standpoint of any empirical observer.” 18 Ellis argues that an enactive theory satisfies (i) and (ii). The problem for the identity theorist lay in thinking that consciousness is the result of a physiological process which happens to the agent. Ellis states: To experience a state of consciousness is not to undergo something, but to do something. . . . To ask “Why cannot you empirically perceive my consciousness?” would then be the same as asking “Why cannot you execute the actions that constitute my consciousness?” The answer would no longer be mysterious: If experiencing consciousness means executing an action, it seems obvious why one person cannot execute another’s actions.19 Insofar as an enactive theory entails that consciousness does not occur without an action performed by the agent in question and this action is the result of her motivations, then an enactive theory is able to account for consciousness without advertence to nonphysical properties, and explains the subjective nature of consciousness—i.e., they are the agent’s own actions. In the argument so far, I have made reference only to consciousness and the role motivation plays in bringing about conscious perception. On this picture, one’s motivations function as a gate or arbiter of what gets fully processed. But I have not commented on the process itself. One may rightly inquire whether we have evidence that perceptual knowledge is due to the motivational interests of the agent, and not just the routine operation of one’s perceptual processes. To argue that perceptual knowledge is a function of a virtuous act, I have to show at least that such knowledge is the result of cognitive acts. It is well established that perceptual knowledge is achieved through acts of inference and categorization. Even proponents of a view called cognitive impenetrability (the thesis that acts of cognition do not penetrate perception) circumscribe this thesis to apply only to early vision, where the informational inputs first arrive in the perceptual system. Achieving perceptual knowledge, however, requires more. The view that perceptual knowledge is achieved through inference and categorization, is called the New Look view, pioneered by Jerome Bruner.20 According to this view, the perceptual process is like science itself; it consists in finding partial clues (either from the world or from one’s knowledge and expectations), formulating a hypothesis about what the stimulus is, checking the data for verification, and then either accepting the hypothesis or reformulating it and trying again in a continual cycle of hypothesize-and-test.21 The output of this process is a categorization of the object or percept which is constitutive of knowledge.
Numerous empirical studies support the view, and it has achieved “received wisdom” in cognitive studies on perception.22 I should note that most commentators on such evidence take the terms inference and hypothesis testing rather literally. Consider the following passage from Bruner describing his view, What is interesting about the nature of inference from cue to identity in perception is that it is in no sense different from other kinds of categorial inferences based on defining attributes. “That thing is round and nubbly in texture and orange in color and of such-and-such size—therefore an orange; let me now test its other properties to be sure.” In terms of process, this course of events is no different from the more abstract task of looking at a number, determining that it is divisible only by itself and unity, and thereupon categorizing it in the class of prime numbers.23 Commenting further, Bruner states, “There is no reason to assume that the laws governing inference of this kind are discontinuous as one moves from perceptual to more conceptual activities. In no sense need the process be conscious or deliberate.” 24 The received wisdom in cognitive psychology, then, has it that a person performs cognitive acts even at the elementary perceptual level. Given this information, it becomes more intuitively plausible to claim that perceptual knowledge is the result of motivated activity. Additional indirect evidence that perceptual knowledge is due to one’s motivated activity comes from the discovery in the Mack and Rock trials that when subjects were exposed to meaningful stimuli (usually words) for 200 ms, there was no evidence of conscious perception. But these very subjects evinced implicit perception of these words, as indicated by performance on stem completion tasks. For nonmeaningful stimuli, there was no evidence for either conscious or implicit perception. The contrast here suggests that some motivational processes function farther back than attentional processing. Ellis explains that, “the anterior subcortical motivational and emotional mechanisms determine early selection; posterior attentional mechanisms then unconsciously enhance signals that have penetrated the early selection ‘gates.’ ” 25 This evidence indicates that motivational processes are acting at a very early stage in P-processing, and probably initiate the spreading activation which eventually ends in perceptual consciousness.
The motivations are epistemic If the enactive theory is true then motivational states direct attention. Additionally, if perceptual knowledge is the result of an inference, then such knowledge requires cognitive action which, in turn, is informed by unconscious motivations. Are such motivations epistemic? To answer this question I assume that one can determine the kind of motivation informing a human action by
A Virtue Account of Perception
looking at the kind of action that results. If a person goes to the kitchen to fix something to eat, we would say that she has a motivation to eat something. If an agent jumps into the water and saves a drowning person, we say she has a motivation to save this person from drowning. What then are the effects of motivation in perception? With respect to perceptual knowledge, what results from focused attention is awareness. Thus, the motivation to be aware of one’s environment is most plausibly present. A subject may be motivated immediately to perform the task of an experiment, but this requires being aware of, say, the cross-bars. Or, an agent may desire to get some food, but here too, the desire to be aware of one’s environment is required before a search for food can begin. We can see that most instances of conscious perception are rooted in a desire to be aware of the stimuli to which we aim our attention, for to be aware is a precondition for any other visual task. Furthermore, awareness is a cognitive state whose value is derivable from knowledge in that if one desired to have perceptual knowledge, one would de facto desire to be aware of the relevant stimuli. Forming perceptual beliefs that count as knowledge requires that one be aware of one’s perceptual environment. So, the motivation to be aware is present in most cases26 of perceptual knowledge.
Objections and replies
Virtues are not automatic There is a constellation of concepts such as automatic, natural disposition, and involuntary that proponents of the unmotivated belief objection may use to explain what is occurring in visual processing. They will explicitly eschew reference to such concepts as epistemic motivation and virtue. I suspect that the proponent must grant, given the evidence delineated above, that P-processing is motivated. The proponent is forced to respond to the argument by arguing that the motivations involved are not epistemic or virtuous. Such an argument may claim that the motivations informing perceptual knowledge run automatically. Now, there are different senses of automatic that the proponent may invoke here. The first sense that comes to mind is that the motivation to be aware is ‘always on’. But this cannot be pernicious to the responsibilist’s claim, since the motivation to be aware of my surroundings is not on when I am in REM sleep. But suppose such a motivation is always on in every non-REM moment. Such a supposition is still not pernicious to the concept of an epistemic motivation since an epistemic motivation could function at all non-REM times. There is nothing in the concept of an epistemic motivation that determines temporal boundaries. Virtues are not typically defined with reference to the times they are to be manifested.
Or suppose by automatic the proponent means something like this; “the motivation to be aware of one’s environment is involuntary, or relatedly, it operates unconsciously.” In response, what makes the motivation to be aware of one’s environment not wholly involuntary is because such a motivation needs to be indexed to the specific stimulus that the agent aims to fully process. So, what the responsibilist claims is that for any case of perceptual knowledge p, there is a motivation to be aware of p. But once the responsibilist indexes the awarenessmotivation to specific instances of knowledge, we can see clearly that such a motivation is part of a constellation of other desires and beliefs the agent has at that particular time. Insofar as the awareness-motivation operates as a function of one’s other motivations and beliefs, it is not a wholly involuntary motivation. In the example offered by Ellis in the quotation above, he states that one is not conscious of red unless looking for red serves some motivational interest. But becoming conscious of a red thing serves the motivational interests of the agent on that particular occasion. A related objection may run as follows; the awareness-motivation is a natural disposition. And we typically do not associate the concept of a virtue with natural dispositions. Traditionally understood, virtues perfect natural dispositions— understood as powers—but are not the dispositions themselves. In response, there are different senses of the term ‘natural’ here as well. If by ‘natural’ the critic means that the motivation is not acquired and operates automatically, then clearly the desire to be aware does not fall into this category. For the desire to be aware of p operates partly as a function of one’s specific beliefs and interests. More simply, the desire relevant for my argument is a desire to be aware of this stimulus. And certainly the desire to be so aware is not natural in this sense. It is certainly not natural that I desire to be aware of this large red object that is moving toward me. If by ‘natural’ the critic means that the motivation to be aware could not be otherwise and therefore not virtuous, then again, we do not have an accurate description of the awareness motivation. The desire for awareness of this stimulus is in fact an act on that occasion. The motivation to be aware of that stimulus could have been otherwise. Simply in virtue of the fact that such a desire is manifest on many particular occasions of perceptual knowledge does not impugn it from being an epistemic motivation.
Nonepistemic motivations What about cognition that is a function of motivation but intuitively, the motivation is not an epistemic one? Such emotions as fear and anxiety give rise to certain motivations that are not epistemic—the motivation may be to get out of the threatening situation. In response, the emotions of fear or anxiety do serve an epistemic function in that they motivate the agent to encode essential features of threatening stimuli.27 Furthermore, preserving such information in detail would have a lot of survival value in that the agent would have a blueprint of what stimuli to avoid.28 So, fear serves the function of motivating the agent to
A Virtue Account of Perception
encode features of threatening stimuli in order to avoid them in the future. (A similar argument can be run for anxiety.) In either case, the emotion does give rise to an epistemic motivation.
Moral virtues and perceptual virtue compared The responsibilist can obviate some further criticisms by comparing intuitive cases of virtuous agents with intuitive notions of ‘automaticity’ and ‘natural.’ Take a case in which an agent is born with a disposition to be compassionate toward others. From an early age, this person evinces a rather strong motivation to alleviate the suffering of others around her. In such a case of precocious moral development, we would still say that the person is virtuous. But just as someone may precociously acquire strong moral motivations, so we may acquire the awareness-motivation early on. But this in itself does not impugn the claim that the awareness-motivation is a virtuous one. Suppose an agent fully possesses the virtue of compassion. In such a case, the agent may alleviate the suffering of others quite unconsciously—at least she would not necessarily be aware of her motivation to act compassionately. It is most likely the case that she is only conscious of the suffering person. Here is a case in which the virtuous motivation functions unconsciously, and yet we would still say that it is a virtuous motivation. Similar comments can be made about the awareness-motivation. This example illustrates another interesting intuition, namely, that an agent who possesses compassion in full cannot do otherwise than she does. In this sense, it is not in her voluntary control whether to be compassionate. Her will has been so formed to do compassionate acts that she cannot do otherwise. But again, in this sense of involuntary (or automaticity) there is still no trouble for the responsibilist in calling the awareness-motivation involuntary in this latter sense. For we would still say that a person who cannot do other than be motivated to perform compassionate acts—when they are required—is virtuous. In the same way, an agent who cannot but be motivated to be aware of her environment (or of this stimulus) still exercises an epistemic motivation. I am using these clear examples of virtuous persons to illustrate that even in these cases, the virtuous person looks like she is behaving automatically or involuntarily. But we would still say that such persons are virtuous. And in the same way, the awareness-motivation shares certain properties with a fully compassionate person and a precocious person’s motivation to alleviate the suffering of others. In both cases, the agent cannot do otherwise, and her virtue functions unconsciously. Of course, as I indicated above, there is a disanalogy to note. The compassionate person cannot do otherwise, on an occasion for compassion, and the perceiver cannot do otherwise than process this stimulus of emotional salience. This disanalogy does not affect my argument. I submit that properties of an action such as automaticity and being unconscious do not impugn the claim that the awareness-motivation is a virtuous motivation.
Some anomalies and refinements Several problem cases need to be mentioned to show the limits and resources of responsibilism’s account of perceptual knowledge. First, there are three types of stimuli that have been found to capture attention in a way that seems to suggest automatic direction of attention. They are (i) emotionally salient stimuli, like a smiling face, (ii) personally meaningful stimuli like someone’s name— we experience this when, say, we are at a party and we hear our name being uttered by someone to whom we were previously not listening and (iii) sudden loud noises. Let’s address the third first. The responsibilist could say that sudden loud noises are attended to because such stimuli may signal a threat in our environment. I think this is plausible in that upon hearing a sudden loud noise, one elicits a fear response. The reason this does not offer the responsibilist a way out, however, is because the fear response is in response to what is already perceived. The emotion of fear upon hearing a sudden loud noise does not serve to direct attention, but serves other behavioral purposes. The responsibilist may try another route according to which the judgment ‘that was a bomb going off’ is plausibly a function of motivated attention in that we usually desire to judge what causes loud noises. But this requires a search in memory of things sounding like it, or an inference from other features in one’s environment. We can back up further to a more bare judgment according to which ‘that was a loud noise’ and the analysis is the same. There needs to be a search in memory, however quick, to make the judgment that X was a noise and was loud. There are other less dramatic putative counterexamples to responsibilism related to loud noises. Consider the following example: I sit at my desk near a window, trying best to concentrate on the very difficult biomedical journal in front of me. I’m very interested in it, and struggling to understand it. A motion at the window registers in my peripheral vision and I reflexively turn to see what it is. It seems that in this and like cases, I do not have the desire to be aware of what is in the window, and yet I certainly know that something (e.g., a bird) flew by the window. The force of this counterexample can be attenuated when we consider that nonconscious motivations can exert an influence on one’s behavior, both cognitively and morally. Second, these motivations can exert such an influence equivalent to that of conscious motivations. Several studies have aimed to support these claims. Tanya Chartrand and John Bargh have conducted a series of studies29 aimed to uncover the presence of and effects of nonconscious motivations. They summarize the results of these studies as follows, “These results support two tenets of the auto-motive model: first, that goal structures can be automatically and nonconsciously activated, and second, that nonconscious goals, once activated, produce the same effects as conscious goals.” 30 Chartrand and Bargh note,
A Virtue Account of Perception
“Consciously choosing a goal at some point is not necessary for that goal to become active and guide subsequent cognition.” 31 Returning to the counterexample, I grant that one does have perceptual knowledge of whatever flew by the window. But there are two implicit assumptions in the example, namely, (a) one desires to be aware only of the journal in front of him and (b) one does not desire to be aware of his surrounding environment. In response, if there are nonconscious motivations, then one need not accept assumption (b). The responsibilist will explain my attraction to the stimulus in the window via appeal to a nonconscious desire to be aware of my environment. Such a goal can be active even though I cannot report it or see it on introspection. Given the empirical research on nonconscious motivations just adverted to, this seems a very fluid way of describing one’s momentary diversion of attention. So, the critic needs to show us a case in which there is neither a conscious nor nonconscious motivation to be aware of one’s environment. As far as I can see, the case presented above is not such a case. With respect to (i) emotionally salient stimuli and (ii) personally meaningful stimuli, the explanation for why attention gets captured is because in both cases, the stimuli are emotionally relevant. Having our attention captured by hearing one’s own name is explained in terms of the motivational salience of the agent’s name. We desire to be attentive to those things that can affect us either physically or emotionally. With respect to a smiling face, we desire to be happy in our surroundings, and smiling faces facilitate this. Thus, it is understandable that there would be a preference to be aware of ‘joyful’ stimuli as opposed to emotionally neutral or stimuli with negative valence.
Cases of free scanning: knowledge without motivation We have all experienced cases of free scanning. In fact, they make up most of our perceptual experience. Our eyes are open, information is coming into the visual system, but no selective attention to particular aspects of our environment is taking place. Yet we would like to say that we acquire perceptual knowledge. Admittedly, the intuition is strong that one gets perceptual knowledge in cases of free scanning and is not motivated. There are, however, several studies that provide evidence for the claim that even in spontaneous viewing an agent is engaged in motivated looking. One piece of evidence comes from comparing a person’s perceptual preference for novel, incongruous stimuli versus more simple stimuli. Subjects consistently favor viewing the novel, incongruous scenes over the simple ones.32 A good explanation for this behavior is that in free viewing, the agent is motivated to comprehend his surroundings. This would explain why subjects fixate on more difficult stimuli than simple ones within view. Two experimental designs illustrate preferences for difficult stimuli over more simple ones. The first design has subjects view a computer screen upon
which multiple pictures are flashed on the screen. The subject is instructed to press a button corresponding to the location the picture appeared. On some trials, pictures were presented simultaneously, in which case the subjects had to choose only one picture. The reason for simultaneous presentation is to discover which picture the subject chose, i.e., which one is favored. The hypothesis being tested is that the subject will choose the picture with the highest degree of collative properties, e.g., novelty, complexity, and incongruity. This hypothesis was verified in these experiments. Daniel Kahneman notes, “[The] results show that the same collative and physical variables that determine the choice of fixation and the duration of free viewing time also control immediate and undeliberate choice in a conflict situation.” 33 The second design aims to measure the fixation and duration of fixation during free-viewing tasks. Subjects were presented with various scenes to view freely. The subject’s eye movement (saccades) is measured during this time by an eye tracker. The results indicate that across many subjects with the same stimulus array, the saccadic eye movements are remarkably similar. Collative properties of the array such as novelty, complexity, and incongruity are viewed longer than other more simple properties within the same array. Kahneman states, The best evidence has been obtained in the free-viewing paradigm, in which subjects are given control of the time they spend viewing a series of abstract pictures. The behavior of subjects who are given no specific instructions tends to be similar to the behavior of subjects instructed to linger on “interesting” stimuli, and quite different from that of subjects who follow a “pleasingness” set.34 The argument here is that in cases of free viewing, subjects are following a pattern of saccadic movement, movement which reflects their motivations to make sense of their environment. In Kahneman’s own words, he concludes, “Observations of this kind suggest that the enduring dispositions which control spontaneous attention reflect epistemic motivation, the need to perceive clearly and to reduce uncertainty.” 35 Additionally, Ralph Ellis cites several more recent studies involving free scanning and says, “Even Monkeys that are just generally looking out at a natural scene are being motivated by curiosity, and are tacitly giving that curiosity priority over weaker competing emotions . . .” 36 So, upon further investigation, even cases of free viewing do not pose a threat to responsibilism since such cases are not clear examples of unmotivated looking. Such looking is goal directed and the goal, as indicated by Kahneman and Ellis, is to understand our environment.
Blindsight and implicit perception Blindsight is the ability “of subjects with clinically blind field defects to detect, localize, and discriminate visual stimuli of which the subjects say they are
A Virtue Account of Perception
completely unaware . . . or of which they might be aware but not in the sense of experiencing a visual percept.” 37 It typically results due to lesions in the primary visual cortex, in particular the striate cortex. Blindsight is a visual phenomenon which challenges the thesis defended in this chapter. If blindsight counts as perceptual knowledge, then we have a case of perceptual knowledge without consciousness. In fact, for any case of implicit perception, if it counts as knowledge, then we have a case of perceptual knowledge without consciousness. To complete the argument of this chapter, it is important to respond to these putative counterexamples. There are several facts of blindsight that make me pause to call it a case of perceptual knowledge. To clarify, there are two experimental paradigms which test for blindsight. The first procedure involves presenting a stimulus in the blind field of the subject and asking the subject to say yes or no when a stimulus is presented. The other paradigm forces the subject to answer yes or no to when a stimulus is present. In either scenario, subjects who are not cued to when the stimulus may be presented perform very poorly. Cowey states, for example, GY [subject’s initials] never detected salient targets presented in his blind field when the moment at which they could appear was not signaled with an auditory cue, even though he did detect them easily with the temporal cue. And he states “In a further study with GY, he more often reported awareness of targets when they were preceded by a cue.” 38 The fact that cues improve performance suggests that the performance of the subjects was not attributable to veridical perception. A second important fact is that blindsight subjects learn to respond to other cognitive processing events and infer from these that a stimulus is present. Cowey explains, Blindsight necessitates that the subject is continuing to process visual events. If these events, now processed entirely by cerebral pathways that do not involve the striate cortex, have sensory-motor consequences, the subject might learn that the latter can indicate something about the visual stimulus. For example, a visual stimulus might elicit an eye movement in a particular direction as a result of the topographic representation of the retina in the superior colliculus. At first, the subject might not notice that an eye movement is caused by something unseen but could learn that when the eyes move in some particular direction or manner in blindsight tests, the movement is correlated with forced-choice guesses about “what is out there”.39 This explanation comports with the fact the blindsight patients are very poor at detecting shape, whereas detection, localization, and direction are fair. But even in these latter cases, Cowey explains that “. . . blindsight is characterized by severely impoverished discrimination of visual stimuli.” 40
So far I have provided evidence which challenges the claim that blindsight counts as a case of perceptual knowledge. I could also argue that blindsight does not challenge the crucial premise in my argument, namely, that perceptual knowledge is mediated by attention. The critic may point out, that is, that if the blindsight subjects do not consciously perceive the stimulus, then there is no attention either. If my argument fails that blindsight perception does not count as knowledge, then the critic can retreat to this argument. As with most beautiful ideas, they are challenged with ugly facts. Evidence supports the view that even in blindsight, attention modulates performance. R. W. Kentridge et al. presented cues to a blindsight subject aimed to focus his attention in a location in his blind field. The stimulus was then presented. The authors comment, “Spatially selective visual attention is demonstrated when information that stimuli are likely to appear at a specific location enhances the speed and accuracy of detection of stimuli subsequently presented at that location.” 41 The explanation for these findings varies, but one is that blindsight subjects suffer lesions to the striate cortex, but other areas governing attentional direction are spared. These areas continue to perform their visual functions just as before the lesions.42
IV. Conclusion One of the barriers to thinking that perceptual knowledge is a function of a virtuous act is that our intuitions about a virtue strongly associate it with acts of the will, or habits that perfect the underlying power, but is not the power itself. The reader may find it plausible to say that expert marksmen perform intellectually virtuous acts when they are able to shoot targets that normal people can’t even see. They have perfected their power through practice. But even this is a stretch, for not only do virtues perfect powers, but we consider them acts of the will, and perception is not an act of the will. I hope that in this chapter, I have shown that under the radar introspection is a very complicated perceptual process, which does require the agent to perform acts. The research from Ellis, especially, indicates that conscious perception requires the agent to perform acts. The strength of the argument in this chapter is, surprisingly, not that it comports with our intuitions—in fact it doesn’t—but that it is empirically informed. Philosophers may persist in their intuitions concerning what perception involves, I certainly share those intuitions. But I do not consider my intuitions a reliable datum of evidence in this case. The studies recapitulated in this chapter, and the professional commentary on them, pushes me to consider that perceptual knowledge is mediated by deep-seated motivated actions of the agent.
A Virtue Account of Memory
I. Introduction In his work Warrant and Proper Function Alvin Plantinga has the following to say about memory: “Clearly my memory beliefs are typically formed in the basic way; that is, I do not reason to them from other propositions, or accept them on the evidential basis of other propositions.” 1 And what memorial beliefs count as warranted is as follows: What counts for warrant is whether memory beliefs typically result from the proper function of our cognitive faculties in an appropriate environment, whether the function of memory is to give us true belief about the past, and whether the design plan in this area is a good one. But the fact is (as we all believe) these conditions are fulfilled. Memory beliefs, therefore, have warrant. Furthermore, they are held with great firmness. . . . I conclude that memory beliefs often have warrant and often have a great deal of warrant; many of them, if true, constitute knowledge.2 Alvin Goldman treats memory in his work Epistemology and Cognition. Memorial beliefs count as knowledge on Goldman’s showing if the memorial process is locally reliable and the beliefs are true. These accounts of memory highlight that there are different analyses of what counts as memorial knowledge. But why is it important to analyze memory in the first place? Goldman states what is at issue in the following passage: Belief is a dated affair. Though you believe a proposition now, there is no guarantee you will believe it later. That depends on memory power. . . . Similarly, though you believe a proposition now, because you seem to remember it from an earlier time, it does not follow that you did believe it earlier. That depends on the accuracy or reliability of your memory. So what you believe at any given time, whether correctly or incorrectly, often depends upon memory. Epistemology, then, is rightly concerned with properties of memory, in particular, with the power and reliability of memory.3
The gist of Goldman’s comments here is that all knowledge depends on memory. There are, to be certain, modes of acquiring beliefs, but we would not consider such beliefs knowledge if they stuck around only for several milliseconds. Acquiring beliefs is not enough, we must be able to keep them, and memory does just that. It is that faculty that enables us to keep beliefs. Conversely, if we do a poor job acquiring beliefs, beliefs that turn out false, memorial knowledge is not possible. Knowledge, then, is a communal affair. As just adumbrated, different theorists endorse different conditions for memorial knowledge. In keeping with the trajectory of my project, I aim in this chapter to defend a responsibilist analysis of memorial knowledge and obviate the unmotivated belief objection. Memory is similar to perception in that both modes of knowing seem to involve automatic processing. When I recall a belief from the not-too-distant past, there seems to be no motivation present in recalling the relevant belief, and thus memorial knowledge is a potential counterexample to responsibilism. So, memory demands treatment because first, it is that faculty upon which all knowledge depends, and second, memorial knowledge of the not-too-distant past seems to be automatic and thus bereft of any motivational factors. My aim in this chapter is to argue that memorial knowledge requires an epistemic motivation. I do not aim to show that a memorial belief requires a motivation, since a belief without an epistemic motivation does not serve as a counterexample to responsibilism. There, as in the last chapter, I assume that a condition for propositional knowledge is that one forms a belief, and that knowledge that p requires having been conscious of the contents of p. The structure of the chapter is as follows. I begin by giving a basic delineation of the memory system and then turn to delineate the actual processes of memory. A memory system refers to the kinds of memory: there are, e.g., iconic memory, short-term memory (STM), and long-term memory (LTM); and memory processes refer to the actual mechanisms used in forming a memorial judgment, i.e., encoding and retrieval. It is important for my argument that the reader understands the basic terms in this area such as encoding and retrieval, for the arguments I present in this chapter rely on an elementary understanding of how such processes work. My first argument aims to show that encoding without rehearsal requires instantiation of an epistemic motivation. But there are potential limitations to this argument, for there may be cases in which encoding is weak or incidental. Incidental encoding is encoding information with attention, but the attention is so general or dispersed that it is not likely to be informed by any motivation. A real-world case like this may be freely viewing a natural scene. If there were a short retention interval after which someone is asked to recall the scene, we would think that the recalled information counts as memorial knowledge. We have, then, a potentially strong counterexample to responsibilism. My second argument, then, aims to obviate this counterexample. I borrow my comments from the previous chapter according to which information stored during free viewing involves epistemic motivations.
A Virtue Account of Memory
A clarification is in order to acuminate the focus of this chapter. The proponent of the unmotivated belief objection will advert to intuitive cases of memorial knowledge in which either (1) a veridical belief is preserved but without attention, or (2) even if the preservation of a veridical belief requires attention, the attentional focus itself is not directed by any motivation. There is some debate on whether there are any beliefs that fall into category (1). Research on subliminal cognition would adjudicate whether category (1) is populated, but such research is a matrix of different positions and it would be specious for the objector to rest her argument on possible beliefs in category (1). Thus, in this chapter I will assume the objector has beliefs in category (2) in mind as counterexamples to responsibilism. The objector has to cite cases of memorial knowledge, which are the result of attention but the attention is not directed by an epistemic motivation.
II. Memory systems and processes It will be helpful to distinguish three memory systems: iconic memory, STM, and LTM.4 They are divided mainly in terms of duration with iconic memory being a storehouse of basically complete visual or auditory information for a very short duration—between 0.5 and 1.5s. Loss is due either to displacement from subsequent information or immediate decay.5 STM is a memory system that contains information for the duration of 1–2s to several seconds. With respect to LTM, this memory storehouse can be considered a warehouse of information that an individual has amassed throughout her lifetime. The duration of LTM is anywhere from a few minutes to many years. Its capacity is extremely large. Information in LTM is maintained through simple rehearsal, but in some cases, information gets stored quite deeply without rehearsal, for example, a traumatic event in a person’s life will get encoded quite strongly as will repeated presentations of the same stimulus. Loss is normally due to autonomous decay. The different memory systems have to be distinguished from memory processes. The memory processes are encoding, retrieval, and source monitoring. To explicate these, consider the following example. Suppose Jones decides to take an Asian history course. He attends class regularly and does his reading. Though he enrolled in the class knowing nothing about Asian history, he ends up doing quite well in the exams. He is able to recapitulate the right information in response to the relevant questions. And further he evinces some sense of connection between the events. That is, he has organized the information about Asian history fairly well. Now, at least three things are going on in Jones’ memory. First, Jones is studying and remembering information on Asian history. This process of ‘committing to memory’ is referred to as encoding. Encoding is the process of storing information into neural codes for later access. Throughout the semester as Jones studies more, the material he is learning is being
stored. Second when it is time to take the exam, the process by which he is able to recapitulate the events and interrelations is called retrieval. Encoding and retrieval are two mechanisms of the memorial process. A third thing is happening in this example as well. On test day, Jones is able to recall information pertaining to Asian history. But there has to be a mechanism which ensures that what he is recalling is information he has learned in class as opposed to information he has imagined about Asia or heard from friends who were not in the class. This mechanism is referred to as source monitoring.6 Source monitoring is a subsidiary process of retrieval. Its purpose is to ensure that the information being retrieved is from the right source. The retrieval process comes in two stages. The first stage consists of a cue or a probe that aims to extract information from a particular store. A probe could be a presentation of an object to which the cognitive system attends (e.g., I see a plane), or it could be as sophisticated as a question (e.g., where were you last night?). In either case, a probe whether in the environment or something internal to the agent (as when I think about what I learned in a previous philosophy class) aims to extract information that matches the probe. The second stage of retrieval consists of matching the probe with the stored information (or ‘traces of feature values’ as it is sometimes called). So, for example, in the first stage I am presented with the probe of a plane, in the second stage I recognize the plane as a Boeing 747-400. Robert Shiffrin presents the following model of retrieval, Events and memory traces are represented as vectors of feature values. The traces are stored separately, although some are formed through accumulation of multiple events. Traces of new events are stored in memory as incomplete and error prone vectors. Such traces are termed episodic, and contain context, physical, and meaning features. . . . Retrieval starts with the matching of a probe cue, again represented as a vector of feature values, to all memory traces in parallel, although only traces sufficiently similar are activated. . . . Matching of probe to a given trace consists of comparing corresponding feature values, and using the results to calculate a likelihood ratio: the probability that the trace is a copy of an earlier presentation of the present probe, divided by the probability that the trace is a copy of an earlier presentation of something other than the present probe.7 An example may help contextualize what is stated here. Take a simple case of object recognition. The object has certain features that distinguish it from other objects. Having been previously exposed to the object, an agent encodes such features to some specifiable level. These are feature values. Upon being exposed to the object at a later time, the information coming into the visual system contains information about the objects’ features. This information is the probe. All encoded information similar to the incoming probe is then activated and the process of matching takes place. When a match is calculated, the agent
A Virtue Account of Memory
judges, for example, “that is a Boeing 747-400.” So, in order for retrieval to take place there must be a probe or cue, and a matching process according to which the probe is matched with the stored memory trace.8
III. The argument from encoding The aim of my argument in this section is to show that an epistemic motivation is involved in forming memorial judgments that have been explicitly encoded without rehearsal. I start here because remembering with rehearsal seems to require motivation. Brown and Craik state, The key to understanding encoding in this model is attention—that is, in order for information to proceed to progressively more capacious and durable stores, the learner has to pay conscious attention to the information. The more rehearsal that the individual engages in, the greater the likelihood that the information will be transferred from short-term to permanent storage.9 The act of rehearsing is clearly a result of one’s motivations to memorize this set of information. And I will assume that rehearsal is informed by an epistemic motivation i.e., a motivation to preserve new information accurately. What is pertinent to show is that explicit encoding without rehearsal is still informed by an epistemic motivation since it is not intuitive that such encoding is the result of any motivation. (The same applies to incidentally encoded beliefs, dealt with below.) Before presenting my argument, some understanding of the encoding process is in order. There are two ways by which new information is logged into LTM, implicit and explicit encoding. Implicit encoding refers to the encoding an agent performs without consciously attending to the stimulus. Information encoded under these conditions is usually not recallable10 but it is encoded to a fairly high degree of specificity, i.e., semantic. For example, Eich11 had subjects wear stereo headphones and were asked to attend only to the information coming from the right ear-phone (or left one depending on the trial). In the unattended ear they were given homophonic words in context (e.g., taxi-FARE). Subjects were very adept at recalling the information piped into the attended ear-phone but were unable to recall at all the paired words in the unattended ear. But when asked to spell various homophonic words (words that were presented to the unattended ear), the subjects most often spelled the ‘correct’ word. For instance, if the subjects were asked to spell fare, the subject spelled ‘fare’ as opposed to ‘fair’. Of course, there may be borderline cases here. For example, I walk into an office one day to talk to a friend and notice cursorily the layout of the office. Upon returning a week later I would plausibly be able to detect that the office contains an added computer monitor or some other significant change. This may be the case even though I do not consciously try to
remember the layout of the office. If anything, I am only consciously attending to the conversation I am having with my friend. Here we would want to say that attention is necessary in order to encode the information strong enough for recall but, intuitively, the attention required does not involve any motivation since it is so general or unfocused. I will refer to such borderline cases as incidental encodings. Incidentally encoded information is information that is encoded strong enough for recall, and thus is a function of attention, but the attention does not seem to be informed by any motivation. Explicit encoding simply refers to encoding information of which one selectively attends. Most of the encoding that takes place via selective attention is recallable, at least within a specified duration. A simple example of explicit encoding is encoding what I had for breakfast this morning. (‘Selective attention’ refers to attention to particular features of a stimulus.) As mentioned above, the goal of this section is to show that an epistemic motivation is involved both in cases of explicit encoding without rehearsal, and for cases of incidental encoding. With respect to explicit encoding without rehearsal (hereafter, I will refer to this simply as encoding), the outline of my argument looks like this, a. Encoding requires a motivation M’ to accurately preserve the content of proposition p. b. M’ is an epistemic motivation. c. Memorial knowledge requires accurate encoding. d. ∴ Memorial knowledge requires an epistemic motivation. I begin by defending (a). The outline of my defense involves a sub-argument according to which selective attention is necessary in order to encode beliefs strong enough to support reliable recall, and selective attention is motivated. To illustrate the role of attention in encoding (and retrieval) without rehearsal, Craik et al.12 conducted an experiment along the following lines. All subjects were presented with 15 two-syllable words and these words were to be encoded for later recall. There were two versions of the first condition according to which the subjects had to perform a secondary task concurrently with the primary one of recalling the 15 words. The first version had subjects perform a simple visual task concurrently with being presented with the words (each word was presented for 4s each). The second version required subjects to perform a similar task in conjunction with the task of recalling the words. A second condition was simply a single task condition in which the subjects were presented with the words without having to perform a concurrent task, and again, in the retrieval period, subjects were free of any other task. In sum, there was a dualtask condition for encoding and retrieval and a single-task condition for encoding and retrieval.13 The aim of the experiment was to show the effects of attention when it is divided among tasks. The results indicate that encoding
A Virtue Account of Memory
performance in dual-task conditions was significantly less than performance in single-task/full attention conditions (4–5 words were recalled for the divided attention (DA) group and ~9 words for the full attention group). And further, the group that performed the dual task at encoding (presentation) performed significantly worse than the dual-task subjects at retrieval (~7 words for the DA group at retrieval). The response times for the DA at retrieval group, however, were significantly greater than the response times for the DA at encoding group and even greater for the full attention group.14 How are these results relevant to defend (a)? To begin, Craik et al. conclude, “These basic results thus confirm and extend the previously reported findings of Johnston and his colleagues . . . and of Baddeley. That is, division of attention at encoding reduces memory performance substantially . . .” 15 The authors took these results to indicate that encoding is subject to strategic control. “We took the latter result to mean that encoding processes are subject to strategic control, and that reduced attention is systematically related to reduced memory performance. . . .” 16 As for retrieval, since response times were significantly increased, retrieval processes could not be considered automatic either, “. . . we also found in Experiment 1 that performance on the CRT task was much more affected by [divided attention] at retrieval than by divided attention during encoding. It therefore cannot be maintained that retrieval is automatic or cost free in terms of processing resources.” 17 Thus, attention is required in order for both encoding and retrieval processes to be reliable. Less clear, though certainly very plausible, is the claim that attention is the function of one’s motivations. In the Craik experiments, the act of recalling the words required cognitive acts, i.e., paying attention to the incoming information and encoding it into somewhat durable stores for later retrieval. When the subjects were given a dual task, they were not able to perform as well as the group in the single-task condition. Even the authors conclude that this fact suggests encoding is not an automatic process, but requires the agent to focus her attention on the task at hand. And this focus required the agent to exercise “. . . strategic control.” By ‘strategic control’, I take it that the authors mean that the agents must concentrate on the presented words in order to encode them. But this concentration takes effort, effort that is a function of the agent being motivated to encode the relevant words. This conclusion comports with common experience. We have all experienced doubt as to whether we locked the car door, or turned off all of the burners in the house etc. In those cases in which we did in fact lock the car door, we still do not remember it even though we had to consciously perform locking the door. Common experiences like these suggest that at the very least, controlled attention is required to encode the relevant information. Is it the case that motivation informs attention in cases outside of the laboratory? Better put, is it always the case that motivation exerts an influence on attention? Commenting on the specific issue of whether motivation controls
attention, Derryberry and Tucker state, “First, they [motivational processes] exert a general influence that serves to broaden or narrow the breadth of attention. Second, they exert more specific effects serving to direct attention toward particular sources of information.” 18 I should mention that Derryberry and Tucker treat the role of anxious states on attentional direction and breadth, but they cite approvingly studies that argue that other motivational states control attention as well.19 So, if Derryberry and Tucker are right, then there are numerous cases in which attention is informed by one’s motivations. To argue for an even closer connection between attention and motivation, Derryberry and Reed argue that the anatomy of the brain suggests that an agent’s motivations influence attentional focus always. If there is a tight anatomical connection between two processing centers, then it is reasonable to expect the two to interact in most cases. “Many anatomists have noted the close connectivity between the limbic and frontal regions, with some viewing the frontal lobe as the cortical representative of the limbic system.” 20 (The limbic region is the neurological seat for motivation and emotion and the frontal regions are the neurological seats for attentional focus.) A tight neurological connection means that if there is activation in one area, there is a corresponding activation in the other. So, if there is an anatomical connection between motivational mechanisms and attentional mechanisms, it is more plausible than not that the two are interactive. Backing away from empirical studies, the connection between attention and motivation seems fairly intuitive. Attention is defined roughly as an allocation of cognitive resources to process information from the environment. And this allocation aims at a goal, in the case of encoding the goal is to store information. Now, a motivation is that which explains goal-directed behavior. If I ingest food, my (conscious) motivation for doing so is because I am hungry. So, insofar as attentional processing aims at a goal, there is a motivation informing that behavior. This argument aims only to establish that there is a motivation which informs attentional direction and focus. This argument does not show that such a motivation is an epistemic one. Distinguishing kinds of motivations will have to advert to the goals to which the respective motivation aims. For example, if the goal is regulating blood sugar level, the motivation is classified as a strictly physiological one carried out by the endocrine system. Given the intuitive connection between attention and motivation, the anatomical connection, and the divided attention findings, it is plausible to hold that encoding for reliable recall is a function of one’s motivations. (I say this is encoding for reliable recall because what concerns me most is encoding that counts as knowledge, and this requires encoding that is sufficiently strong to support reliable recall.) Consequently, (a) is quite plausible, in particular (a) is more probable than its denial. The denial of (a) lacks both empirical evidence on the role of attention in encoding and is antagonistic to the intuitive connection between attention and motivation. If encoding were automatic or “cost
A Virtue Account of Memory
free” we would not expect disparate performance on divided attention tasks vs. single attention tasks. We do, however, have confirmatory evidence in that if attention were a function of one’s motivations then performance differences should show up between dual task conditions and single-task conditions. The reason is that cognitive tasks can be divided very generally into (i) tasks that require attention, and (ii) tasks that do not require attention. Examples of tasks in class (ii) are subliminal perception, and implicit semantic processing. To determine whether a task belongs in (i) or (ii), one simply determines whether there are attentional “costs” in performing the task. And the encoding experiments referred to above show that in dual-task conditions performance is significantly less than singletask conditions. This would illustrate that the task involves attentional cost and thus requires attention to be directed to the specific task in order to process it fully. Now, if attention needs to be directed to specific tasks in order to process the relevant information, then attention of this sort is motivated—for there has to be something that directs the attention and the best candidate for this is an agent’s cognitive motivations. Thus, if there are differences in performance on encoding tasks (as is confirmed), then attention is motivated for these tasks.
IV. Encoding involves epistemic motivations Consistent with the procedure in the last chapter, motivations are categorized with reference to their effects. This procedure is certainly not infallible, but is the only one available for my purposes. So, suppose there is a motivation M’ that results in a cognitive state S’ (where S’ is the preservation of beliefs for later recall). Let’s stipulate further that S’ refers to a cognitive state in which information is preserved accurately. This may include preservation of false information or information the agent believes is false, as when a professor teaches her students a theory that she thinks is false, or is false. The value of cognitive state S’, however, is that it preserves this information accurately. My argument that M’ is an epistemic motivation runs as follows, 1. Motivation M’ results in cognitive state S’. 2. S’ is an epistemically valuable state. We may say that M’ is a motivation to attain S’ and thus, it follows from (1) and (2) that, 3. ∴ M’ is a motivation to attain an epistemically valuable state, i.e., a state whose value is truth. 4. An epistemic motivation is a motivation to attain a valuable cognitive state. 5. ∴ M’ is an epistemic motivation.
Premise (4) is just the definition of epistemic motivation. For premise (2), the value of S’ is determined by its accurate preservation of the information presented to the agent. And preserving information (whether it comports with reality or not) is a valuable epistemic state. Some may not find this intuitively plausible. Suppose a person has a very accurate memory but the information that she acquires is false. She therefore preserves a large set of false beliefs. How can this be valuable? I am inclined to think we should make a distinction between the set of beliefs not being valuable (since they are all false), and the accurate and detailed preservation of that information not being valuable. The latter cognitive state is clearly valuable and is one resulting from a good memory. With respect to premise (1), the sense of results in should be understood in a non technical sense. The Craik experiments show that accurate memory performance requires focused attention on the requisite stimulus in order to encode it with sufficient strength. Thus, the motivation which informs such attention aims to preserve phenomenal information accurately. I shall discuss two objections to the argument here. First, the transition from there being a motivation to there being an epistemic motivation is weak in that the motivation informing attentional focus may run automatically (like the motivation for self-preservation) or the motivation is not automatic but is a nonepistemic motivation, e.g., fear or anxiety. Neither of these motivations seems epistemic in that they seem not to issue in valuable epistemic states. But even if they did, there are problems. Sometimes attentional focus may be informed by fear and is thus, not an epistemic motivation even if the result is preserving external information accurately. So, the critic may grant that there are motivations involved in encoding, and some may be epistemic, but others are not. And this latter class of memorial beliefs poses a threat to responsibilism. The second objection is that the motivation to preserve phenomenal information is more like a natural disposition rather than a virtue in the traditional sense. It is questionable whether such a motivation is acquired, and it is often unconscious to the agent. So, if we look at the motivation itself, it does not look like a virtuous motivation. With respect to the first objection, the emotions of fear or anxiety do serve an epistemic function in that they motivate the agent to encode essential features of threatening stimuli.21 Recall that motivations are action-initiating emotions. An emotion of fear can elicit a motivation to encode detailed features of the fear stimulus so as to avoid it in the future. In this, the motivation is epistemic though the emotion of fear in itself need not be. A similar response can be made regarding anxiety. In either case, the emotion gives rise to an epistemic motivation. What about victims of violent crimes whereby she attends to the salient events at the time of the attack, but has no desire to remember those events.22 In response, the motivation to encode detailed features of the environment need not have anything to do with a motivation to retrieve those features at a later
A Virtue Account of Memory
date. A victim of a violent crime will certainly attend to salient features of the scene with no second-order desire to recall such information in the future—but the motivation to encode may just as well stop at the goal to encode. I need not establish that in order to be an epistemic motivation, the motivation must be to encode and to recall what was encoded. This extra step is not necessary for my argument. The second objection is not pernicious to responsibilism for two reasons. First, the motivation to preserve information needs to be indexed to each specific instance of encoding. So, encoding this stimulus as opposed to any other is not a function of a natural disposition per se. The motivation to encode stimulus X is a function of the organism’s current goals, beliefs, and surrounding environment—none of which is ‘natural’ in the sense indicated in the objection.
V. Incidental encoding I turn now to cases of incidental encoding (e.g., freely scanning one’s environment). I countenanced above that these cases pose the biggest threat to responsibilism in that we would consider such cases knowledge, and our intuitions are strongly opposed to the idea that such beliefs are formed via an epistemic motivation. It is the aim of this section to properly situate cases of incidental encoding within a responsibilist epistemology. There are two routes by which I do this: the first route aims to show that such judgments do not count as knowledge, and the second route grants that they do count as knowledge but argues that they are formed as a function of one’s epistemic motivations. I argue that either route has promise. With respect to the first route, we may assume, for the moment, that incidental encoding is a case of unmotivated encoding. Furthermore, let’s grant that if cases of incidental encoding are counterexamples to responsibilism, such retrieval of the encoded information must satisfy our intuitions that it counts as memorial knowledge. The nature of the first route is to argue that the encoding is so weak that retrieval is either not possible or is not veridical. One way to measure whether incidental encoding is strong enough to support reliable retrieval is to have subjects detect changes to natural scenes. The idea is that if subjects nonselectively attend to a holistic scene and then change the scene in some way, after which we have the subjects detect the change, then we could measure the retrieval accuracy of unmotivated encoding. If it turns out that such retrieval is inaccurate, we should say that incidental encoding does not yield memorial knowledge. In fact, studies on a phenomenon called changeblindness (CB) illustrate that people are in fact very poor at detecting changes to scenes—changes that are either marginal or significant. Different experimental designs are used to illustrate CB, some of which change a scene during a subject’s eye movement, or after an eye blink. Other experiments present a scene, then present a blank interval scene, and then
present the original scene changed in some way. And still others change scenes along a continuum as with changes in movie scenes.23 In all of these designs, the results are the same—subjects are poor at detecting the target changes. In the experiments employing a change during a blank interval between two pictures, subjects are exposed to a real-world picture for a short duration, and then they are presented with the picture but with a change—either marginal or significant. After the subject is exposed to the changed picture, she is presented with the original picture again and the two are cycled through until she correctly identifies the change. In a study using this very design Rensink et al. (1997) conclude, The preceding experiments show that under flicker conditions [original image (240ms)—blank screen (80ms)—changed image (240ms)—blank screen—original image . . . etc.] observers can take a surprisingly long time to perceive large changes in images of real-world scenes. This difficulty is due neither to a disruption of the information received nor to a disruption of its storage.24 And further on, the authors propose that there is no disruption of storage or of information received because if anything does get encoded, the trace is very tenuous, Visual perception of change in an object occurs only when that object is given focused attention; In the absence of focused attention, the contents of visual memory are simply overwritten (i.e., replaced) by subsequent stimuli, and so cannot be used to make comparisons.25 The point here cannot be overemphasized. In order for the contents of visual awareness to be encoded, one needs to perform a distinct act of encoding; otherwise, such contents of visual memory are overwritten by subsequent stimuli. And insofar as such information is overwritten, there is no preservation of visual information and no memorial knowledge. I should note here, however, that the central or significant changes were, as one would expect, detected more readily than peripheral changes. But this is not to suggest that unmotivated encoding takes place. Simons and Levin comment, “Given the results of the . . . flicker experiments, we know that changes will not be encoded automatically and that some effort is needed to detect changes even with a localized transient [a non-central change].” And further on, the authors explain the accuracy in detecting central changes as follows: “Apparently, the center of interest benefit derives not from automatic representation of a precise visual image but from the abstraction of a scene’s contents.” 26 The detection of central changes is a function of attending to the gist of the scene. The relevance of this result for responsibilism is that the memorial
A Virtue Account of Memory
judgments that do count as knowledge and are formed under putative incidental conditions are a function of one’s attentional focus to the abstract/general features of the scene. Thus, we do not have a clear example in which both (i) the memorial belief counts as knowledge and (ii) the belief is not the result of one’s motivation to attend to the relevant information. More startling experiments indicate that we apparently lose information between eye movements. In such experiments, subjects were given the task of viewing a set of photographs and were informed that the photograph may change. While viewing one photograph, changes were made to it during the subject’s eye movements and she was asked to report the change when she detected it. The idea is that since the subjects were given the task of recognizing changes to the photograph they would be extraattentive and would encode more durable representations of the pictures. (Of course, changes were made in a location of the picture that was disparate from the subject’s eye focus. But subjects were given time to ‘come back around’ to the changed area.) Simon and Levin comment on the results, “. . . observers failed to notice when two men in a photograph exchanged different colored hats and only 50% noticed when two people exchanged heads. In all subjects missed nearly 70% of the changes that occurred during an eye movement.” 27 Results like these push Simon and Levin to conclude, “During an eye movement, we apparently lose, or at least lose access to, many of the visual details of the previous view.” 28 And this is so even if the subjects are attentive to the scene. CB results are also evinced for tasks involving focused attention. One interesting experiment involved random people on the street—i.e., Harvard Square. An experimenter posing as a normal person asks a person (the subject) for directions. While the subject is giving directions, two people carrying a door walk in between the experimenter and the subject, during which time the experimenter switches with a different experimenter. The two experimenters were different in many ways such as height, voice tone, clothing, and build. In spite of these rather perspicuous changes, only 50 percent of the subjects noticed the change.29 The authors of this study conclude, “These studies demonstrate convincingly that paying attention to an object by no means guarantees change detection.” 30 And thus, even if encoding does take place in such incidental cases, the information so encoded is too tenuous to support reliable recall. CB effects illustrate that under certain circumstances, even attentive encoding of an object or scene does not yield encoded traces strong enough to support reliable recall. Whereas for clear cases of unmotivated encoding, CB effects are quite pervasive. Reliable encoding is not an automatic process, but requires focused attention—and even in this latter case, encoding can be very weak. Notice CB effects provide evidence for the conclusion that incidental encoding is so weak that it cannot support recognition sufficient to count as knowledge.
The second route aims to show that cases of freely viewing a scene are not cases of incidental encoding.31 I have addressed cases of free scanning in the previous chapter (see the section “Cases of Free Scanning: Knowledge without Motivation”). The evidence from the studies referenced there can be imported here in that if any feature of a scene gets encoded it is due to one’s epistemic curiosity. To strengthen my case, I should explain why such putative counterexamples to responsibilism have such an intuitive appeal. First, it would seem more adaptive that what entered into consciousness were the objects in our environment and not the inner workings of our own cognitive apparatus. So, one might suppose that if there are any motivations informing the perceptual and memorial processes, we would not expect them to be available to consciousness since consciousness should favor the objects in our environment. Being conscious of the latter would certainly be more adaptive. Second, in point of fact, “The processes that determine the locus of individual fixations are psychologically silent, and their feedback is so poor that people do not usually know precisely where they are looking.” 32 The emphasis on “feedback” is because we simply do not take note of where we are looking because what is often motivationally salient to an individual is the object of her gaze, not that she is gazing. Thus, any feedback process will often be in the “off ” position unless otherwise activated. Furthermore, the responsibilist holds that some motivations are unconscious, as may be the case in free scanning. But this is not pernicious to responsibilism since responsibilism does not entail that an epistemic motivation be conscious to the agent. Consider persons who possess the moral virtues in full. A fully compassionate person is not conscious of her motivation to alleviate the suffering of others. In the same way, we should not expect a person who possesses an intellectual virtue in full to be conscious of the respective epistemic motivation. I take it that these are three good reasons why epistemic motivations do not show up on introspection. I just made mention of there being unconscious motivations, and indicated that these motivations can exert an influence on one’s activity. I have recapitulated the empirical evidence for these claims in the last chapter, and briefly review it here. Tanya Chartrand and John Bargh argue that a person can have nonconscious motivations activated without being aware of this, and these motivations manifest effects equal to conscious motivations. They summarize their findings as follows, “These results support two tenets of the auto-motive model: first, that goal structures can be automatically and nonconsciously activated, and second, that nonconscious goals, once activated, produce the same effects as conscious goals.” 33 These results block any objection of the following form, “the subject did not consciously choose to encode X, therefore there is no motivation present to encode X.” Such an objection is inert. In summary, the combined empirical evidence on divided attention and CB suggest that epistemic motivations are essential for memorial knowledge.
A Virtue Account of Memory
Furthermore, it is plausible to hold that the requisite motivations result in cognitive states whose value is derivable from truth. Even seemingly nonalethic emotions such as fear result in epistemic states whose value is derivable from truth, i.e., encoding the central features of the fear-eliciting stimuli.
VI. Summary and conclusions In the previous two chapters I have argued for the thesis that epistemic motivations are involved in perceptual and memorial processing. The basic structure of the argument runs as follows: a. Attention is necessary for perceptual/memorial knowledge. b. One’s motivations are necessary to direct his attention. c. The kind of motivation M’ involved in directing one’s attention is discerned abductively from the results of M’. d. The results of M’ are that the agent obtains a valuable cognitive state. e. An epistemic motivation is a motivation to obtain a valuable cognitive state. f. ∴ The motivation involved in attention is an epistemic motivation. g. ∴ The motivation involved in perceptual/memorial knowledge is an epistemic one. The weakest link in this argument is (c) and (d), though I have provided some support for these claims. Premises (a) and (b) are well supported by both empirical research and by intuition. What are the implications for responsibilism? I have argued thus far that there are epistemic motivations informing the acquisition of perceptual and memorial knowledge. But the present picture of perception and memory still looks rather thin in that we are a far cry from the more robust intellectual virtues such as fair-mindedness, intellectual courage, or phronesis. In what sense, then, can we say that an agent is responsible for her perceptual or memorial knowledge? To answer this question it is important to recognize both the contrast between perceptual/memorial knowledge and other broader epistemic practices, and the contrast between the thinness of the epistemic virtues functioning in perceptual/memorial knowledge and the thickness of the epistemic virtues such as phronesis. I suggest that the significance of these contrasts lay in that perceptual/memorial knowledge is not that valuable in itself. Memorizing numbers in the phone book is not really procuring a valuable epistemic state, whereas memorizing the Torah in order to procure religious knowledge and gain holiness is very valuable. But notice in the latter case, memory is subservient to a further end, an end dictated by broader epistemic and moral goals. These further ends are what imbue memorial knowledge of the Torah with epistemic value. Notice also that these further ends of religious knowledge and moral sanctity require the
employment of more robust intellectual virtues such as diligence or intellectual perseverance. There seems to be, then, a continuum of epistemic states of which some are more praiseworthy to possess than others. So, the thickness of the virtue comports with the value of the end, the virtue in question being necessary in order to obtain the end in question. (I am not suggesting here that it does not take intellectual diligence to memorize numbers in the phone book. The direction of my conditional is simply, if the end is valuable, then the virtue requisite to attain the end is more robust.) There is a continuum, then, of valuable or praiseworthy epistemic states running from the not-so praiseworthy (e.g., perceptual knowledge in itself) to the very praiseworthy (e.g., philosophical understanding). And concurrently, there is a continuum of virtue running from the rather thin (e.g., awareness) to the rather thick (e.g., intellectual fair-mindedness). Thin virtues are intellectual virtues insofar as they possess an epistemic motivation, but such motivations function largely nonconsciously and if acquired are acquired without much conscious effort. Whereas thick intellectual virtues possess a more robust epistemic motivation and acquiring the virtue may take effort and requires consciously repeating, say, fair-minded acts. Returning to the original question, in what sense is one responsible for acquiring perceptual/memorial knowledge on the responsibilist account? The answer is to the extent that the agent’s focus of attention is a cognitive act informed by an epistemic motivation, the agent is responsible for that act. But the praise we confer on the agent is not that great. Whereas an agent who attains a high level of philosophical understanding is certainly responsible for achieving such a state insofar as he performs the requisite cognitive acts. The praise we confer on him is greater due to the value of the state and the cognitive/moral labor requisite to attain it. To summarize, there is a contrast between the value of certain epistemic states such as perceptual knowledge and philosophical understanding, and between the degrees to which one’s epistemic motivations are praiseworthy. And it is perfectly consistent with the responsibilist’s definition of knowledge that perceptual/memorial knowledge is not that praiseworthy of an epistemic state, and thus the virtuous motivations requisite to attain it are going to be thin. Insofar as responsibilism defines knowledge partly with reference to one’s epistemic motivations and low-grade knowledge requires the agent to exude an epistemic motivation, the responsibilist is able to account for low-grade knowledge. Notice, however, the key feature of responsibilism is the notion of intentionality or motivation. The reason this is a key feature is because other theories of knowledge leave it out of their analysis. In this way, responsibilism provides a needed corrective to contemporary analyses of knowledge, namely, one should incorporate the notion of intentionality into his analysis of knowledge. Empirically it seems that motivations are present in perceptual and memorial processing. But no mention of motivation or intentionality shows up
A Virtue Account of Memory
in any other contemporary treatment of knowledge. Consequently, responsibilism is better equipped both to explain at the theoretical level why some memorial/perceptual beliefs do or do not count as knowledge and it answers the practical question of how one should adjust her belief-forming practices in order to attain knowledge. To show how responsibilism can answer the practical question, let me turn to some illustrative examples of how memory and perception can go wrong, and the feature that corrects the error is one’s epistemic motivations. Schacter and Dodson mention seven sins of memory; I pick out three for my purposes. They state, The next three sins all involve distortion or inaccuracy. Misattribution involves attributing a recollection or idea to the wrong source; suggestibility refers to memories that are implanted at the time of retrieval; and bias involves retrospective distortions and unconscious influences that are related to current knowledge and beliefs.34 There are some good examples of each sin in the literature. An example of unconscious bias is illustrated by Ross et al.35 who demonstrated that if given sufficient amount of time (one month), subjects who were given a message that was pro bathing, recounted having bathed more often than the control-group subjects who were not given such a message. And the opposite effects were seen when subjects were given a negative message pertaining to bathing. The same results were obtained within a week for a group of subjects who were given proand con messages about tooth brushing. An example of suggestibility comes from Spiro’s36 study that involved presenting subjects with a videotape of a couple who were arguing. After the presentation, one group of subjects was told that the couple subsequently divorced and another group was told they stayed together. When asked to recount the story, subjects made intrusions or additions consistent with whether the subjects were told the couple stayed together or subsequently divorced. In Spiro’s study, the sin lay in that the subjects do not pay attention to the role bias plays in recall. The subjects made intrusions on the couple-story that fit their own preconceptions of what happened—preconceptions which were informed by whether they were told the couple stayed together or not. The remedy in each of these cases is for the agent to align her retrieval of information to parallel alethically orientated ways of recalling memorial judgments. But, this requires the agent to be motivated in ways to use her retrieval processes well. The remedy for the sin of suggestibility, for instance, would be for the agent to self-reflect on her own biases and any extraneous information that may distort recall of the original information. And this requires being motivated correctly, i.e., motivated to recall veridical information and to remain self-aware.
With respect to perception, there is an interesting case study in the medical field concerning a common error that surgeons make during a particular procedure.37 To oversimplify, the procedure involves transecting the common bile duct leading from the gallbladder. The error, however, usually concerned confusing the common bile duct with the cystic bile duct. Although the two are quite discernable, the error rate is quite high given current American surgical standards. A reasonable estimate is that 1 out of every 1,000 cases results in injuries due to the above-mentioned error. The results of the study show that “errors leading to laparoscopic bile duct injuries stem principally from misperception, not errors of skill, knowledge or judgment.” 38 And the misperception was due in part to what the surgeon expected before the procedure began. “The anatomic relationship between the [common bile duct] and the gallbladder in these cases mimicked the surgeon’s mental model of the relationship between the cystic duct and the gallbladder.” 39 Examples abound which illustrate how one’s expectations or beliefs can infect one’s perceptual judgments. Examples include racial profiling, induced stereotypes,40 judgments of size regarding valuable as opposed to invaluable things;41 hungry people are more adept at seeing food-related words. The needed corrective would be for one to at least be aware of her expectations and biases. But this is just to be intellectually fair-minded and self-aware. To be sure, my point here is only that responsibilism can tell us how we can modify our belief-forming practices in order to acquire knowledge. Insofar as the above-mentioned cases all involve false beliefs, none of them counts as knowledge on any theory. But that is a theoretical issue. The practical issue is how one can form beliefs that do count as knowledge in the abovementioned circumstances. And to this issue, the responsibilist has an answer whereas other theories are hardpressed to address any practical question. Thus, responsibilism is able to address the theoretical issue of analyzing knowledge in many of its modes, and it addresses the practical question of how to acquire knowledge. To both issues, the responsibilist makes reference to one’s intentions or motivations. Other theories of knowledge are unable or hard-pressed to address the practical question, and they fail to analyze knowledge with reference to one’s epistemic motivations. In this way, responsibilism alone countenances a significant feature for a robust theory of knowledge.42 In the next chapter, I broaden my treatment of modes of knowledge to include a distinctively human enterprise of knowledge acquisition, i.e., testimony. Up to date, I have treated modes of knowledge that focus on a single agent acquiring beliefs. The enterprise of testimony, however, focuses on interpersonal interaction. A robust theory of knowledge will not only treat a single knower’s epistemic practices (e.g., memory and perception), but also a community of knowers’ practices (e.g., testimony). Analyzing testimony, in addition to perception and memory, works toward completing an analysis of human knowledge.
Testimony and Cognitive Virtue
I. Introduction Discussion on the epistemology of testimony has erupted in the past decade or so. Much of the literature on this topic has been very fruitful in extending our understanding of both the importance of testimony as a mode of acquiring knowledge (many of our beliefs we consider knowledge are formed via testimony), but it has also delivered interesting proposals as to the conditions for successful testimony. The aim of this is to give a satisfactory analysis of testimonial knowledge. There are various deficiencies in previous accounts, but given the nature of the deficiencies, my goal here is not to rebut some contemporary accounts, but rather complete them. In the second section of this chapter, I lay out what a satisfactory analysis of testimony should look like. I do this by recapitulating the views of two prominent epistemologists on testimony, i.e., Tony Coady and Elizabeth Fricker. In the third section I turn to give my own analysis of testimonial knowledge which aims to further elucidate the theory of cognitive virtue outlined in this book. The argument in this chapter is that only the responsibilist can give a satisfactory account of testimonial knowledge.1
II. Some criteria for a successful account of testimony A successful account of knowledge (in general) must aim to resolve several problems. The first problem is generated when we consider that there are many modes by which we acquire and sustain knowledge such as perception, memory, testimony, inference etc. Giving a definition that is able to encompass different avenues of acquiring knowledge is one problem which an epistemologist aims to resolve. The second problem for which an epistemological theory must aim to resolve is to account for knowledge not only in its various forms, but also within each form, it must account for varying levels of complexity. For instance, some cases of successful testimony would seem to require the receiver to possess justification for the testifier’s trustworthiness. An example would be expert testimony
in the court of law.2 In other cases, we acquire testimonial knowledge simply by believing the testifier’s message. An example would be a child receiving true information from a parent, or getting directions from a stranger. I suppose one more ingenious than I could construct similar examples for perception and memory. The point is that within each form of knowledge there are varying degrees of complexity and an account of testimonial (or perceptual and memorial) knowledge must encompass these varying degrees. This second problem points out the core issue with analysis of testimony. Testimony is context sensitive, as just indicated, and thus delineating the conditions under which a receiver acquires knowledge from a testifier becomes very difficult. A basic taxonomy of positions taken to account for testimony runs basically as follows. There are two sets of conditions for which epistemologists on testimony set down. They are testifier-centered and receiver-centered conditions. Testifiercentered conditions specify the conditions a writer or speaker must satisfy in order to transmit knowledge. Receiver-centered conditions specify the conditions under which a receiver of testimony gains knowledge from the testifier. Within the receiver-centered rubric there are two positions, the first of which claims that one gets knowledge from an utterance that p simply by accepting the proposition on the testifier’s say so. This is called the non reductive or externalist position.3 It is non reductive/externalist in that testimonial knowledge is derivative of the testimony itself and not some other source of knowledge (e.g., first-hand perception or inference). There is no independent justification needed in order to know that p from a testifier’s say so that p. The second position is a reductive/internalist one in that testimonial knowledge is dependent upon the receiver possessing an independent justification for the testifier’s competency and sincerity.4 Within the testifier-centered rubric, there are two main positions as well. The first position claims that the testifier must know p himself in order to transfer knowledge that p.5 The other position simply denies that the testifier must know p in order to transfer knowledge that p.6 There are, of course, nuances within each rubric. In order to motivate a responsibilist account of testimony, we need to see the insights of other accounts. In the next section I consider C. A. J. Coady’s and Elizabeth Fricker’s accounts of receiver-centered conditions and then turn to treat testifier-centered accounts.
III. Externalism and internalism for receivers A. Coady’s externalism Tony Coady7 has set out to defend an externalist position with respect to testimony. His argument is split into two main parts, the first part being to show that one version of a reductive approach to testimony fails. The second part aims to provide a positive argument in favor of an externalist position. For the purposes of this chapter it is sufficient that I outline his positive argument.8
Testimony and Cognitive Virtue
The aim of the positive argument is to show that testimony is an independent (Coady uses the term, direct) source of knowledge. There are two different arguments to this effect. The first argument is Davidsonian in spirit and argues that we must trust the speaker in order to make sense of what the speaker is saying. The second argument is an argument from actual practice and intuitive conceptions of what counts as successful transmission of knowledge. I shall consider his intuitive argument alone since his intuitive argument is relevant for the motivation for responsibilism.
Coady’s intuitive argument Coady begins his intuitive argument for externalism by obviating an initial objection to such a position. Coady states the objection thus: After all, it may well be urged, the rational person does not believe just any and every thing he is told. His assent must be mediated by a consideration of the veracity of the witness, his reliability . . . the probability of what he says, and so on. Thus mediated, our belief in what he says must count as inferentially based and likewise for our knowledge, where it is knowledge.9 The position endorsed here is a species of reductivism in that specific testimonies hold epistemic weight only because an inference was made on the basis of other sources of knowledge or other testimonies. Coady delineates two problems with this objection the first being that what the reductivist recommends does not comport with an intuitive notion of successful transfer of knowledge. In our ordinary dealings with others we gather information without this concern for inferring the acceptability of communications from premises about the honesty, reliability, probability, etc., of our communicants. I ring up the telephone company on being unable to locate my bill and am told by an anonymous voice that it comes to $165 and is due on 15 June. No thought of determining the veracity and reliability of the witness occurs to me nor, given that the total is within tolerable limits, does the balancing of probabilities figure in my acceptance.10 Coady’s conclusion is that it would actually be irrational for one to accept what one is told—in situations like calling the phone company—only after one has made the requisite inference about the testifier’s veracity and reliability. But Coady admits that not everyone will agree on what the rational person would do in certain epistemic situations. So, Coady turns to the second problem with the above argument. The idea behind this reductivist argument is that one must possess undefeated justification concerning the testifier’s communication. The reductivist requires that one has justification that her informant is honest and competent
with respect to the communication in question. The receiver must justifiably believe that all relevant defeaters are not instantiated on the occasion in question. The problem with this is that such internalist requirements are not necessary. In the case of perception, for instance, if one believes that there is a tomato on the table, we clearly do not require that the perceiver justifiably believe that her faculties are functioning properly, the environment is conducive to veridical perception or that no one has deceptively replaced the real tomato on the table with a fake one. The analogy here is an important one. Coady notes, We do not have to establish the many propositions which, if false, would invalidate our ready assent to what we are told, unless there is already some reason to believe that their truth is in jeopardy. . . . We often take it that the testimony mechanism is functioning adequately, just as we may usually take it that the perceptual or memorial mechanism is not malfunctioning.11 Just as it is reasonable to accept without investigation that my perceptual faculties are functioning well so it is reasonable to accept without investigation that testimonial chains are truth preserving. And though testimonial chains may run the risk of not being truth preserving more so than perception, there are usually ostensible signs of someone not telling the truth—e.g., the subject matter lends itself to ulterior motivations, the speaker lacks eye contact etc. It is not clear why Coady thinks that this second response avoids the problems encountered in the first response. The first response made appeal to what the rational person would do, and it seems that that is exactly what the second response does. But here I think Coady’s move to the second response is really an explanation for why it is rational to accept the say-so of others directly. And the reason is that we clearly do not require that all relevant defeaters are defeated in forming perceptual beliefs, so therefore we should not do so for testimony either. Of course, if a defeater presents itself to the receiver, then further investigation into the testifier’s veracity would be the rational thing to do. But investigating the testifier’s veracity absent any ostensible defeaters is irrational because we clearly do not do this for other common sources of knowledge. To summarize, Coady’s point of departure for his positive defense of testimony is intuitive reflection on actual practice. When we look at how we actually accept and give testimony, in many cases it looks like we do so in a direct way.
Coady and internalist intuitions Coady does mention that the receiver should be on guard for apparent defeaters. Coady says, “. . . we do not have to establish the many propositions which, if false, would invalidate our ready assent to what we are told, unless there is already
Testimony and Cognitive Virtue
some reason to believe that their truth is in jeopardy. . . .” 12 Such comments suggest that undefeated defeaters render testimony invalid. Consider also that in the example of calling up the telephone company Coady mentions “. . . given that the total is within tolerable limits . . .” and “there is nothing hesitant or suspicious about the unknown communicant’s responses. . . .” Both comments suggest the receiver should be epistemically alert when receiving testimony even in simple cases. More pertinently, in a chapter entitled “Astonishing Reports” Coady recommends the following rather internalist recommendation for testimonial acceptance, None the less, degree of coherence with what we already know or believe about the world always has some role to play in the assessment of the credibility of a particular piece of testimony. This factor, what we might call the plausibility of the reported matter, must always be weighed against the particular circumstances of the report. To adopt the terminology of the Port-Royal Logic, we must weight the internal circumstances against the external circumstances, where the internal circumstances concern the probability of the event reported in relation to what we know and the external circumstances concern the credit of the witness, degree of confirmation by other witnesses, the internal consistency of the narration, the type of testimony, the interest, relevant beliefs, and purpose of the testifier, and the abilities of the witness in relation to what he narrates.13 Again, the appeal to actual practice leads Coady to a more internalist take on testimony when it concerns more sophisticated instances of testimony, e.g., testimony of experts, or astonishing reports. The idea to glean from Coady’s treatment of testimony is that testimony is variegated and any account of testimony must take into consideration the many different contexts within which successful testimony occurs. Elizabeth Fricker echoes these same thoughts in her treatment of testimony to which I now turn.
B. Fricker’s internalism Elizabeth Fricker’s internalism also comes in two parts. She first devotes some time to show that an externalist approach to testimony is not sound, and second, she shows that a reductive account does accord with actual practice. Fricker’s attack on externalism begins by outlining one argument in support of it. According to Fricker, the externalist claims that it is simply not practically possible to verify or investigate the trustworthiness of every testimonial chain of which one is a receiver. Additionally, we do gain knowledge from others via testimony. It follows that knowledge is gained via testimony without checking up on the alethic worth of the testimonial chain. Fricker aims to attack what she calls
the negative claim (NC) according to which it is often not practically possible for a receiver to verify the trustworthiness of her testifiers.
The possibility of reduction What is the argument against NC? Fricker starts her attack on NC by clarifying the work cut out for her. Specifically, she needs to stipulate what is required in order to have knowledge via testimony. And then she must answer how such requirements are met. If the answer to the ‘what’ question blocks skeptical worries and the answer to the ‘how’ question comports with common-sense practice, then Fricker’s theory succeeds. Let’s begin by defining what counts for a speaker to be trustworthy according to Fricker. Trus (S,U): A speaker is trustworthy with respect to an assertoric utterance U by her, which is made on an occasion O, and by which she asserts that P, if and only if (i) U is sincere and (ii) S is competent with respect to P on O, where this notion is defined as follows: If S were sincerely to assert that P on O, then it would be the case that P. [And ‘U is sincere’ means that S believes that p.]14 Now, Trus fits into Fricker’s overall argument in that if it is possible (and practical) for H to obtain independent confirmation that S is trustworthy in some cases15 then NC is false. So, the next step in Fricker’s argument is to show how one comes to know that S is trustworthy. Fricker thinks that mature language users do this quasiautomatically. To quote at length, In recognizing an utterance by a speaker as a speech act of serious assertion, with a certain content, a hearer is ipso facto engaging in a minimal piece of interpretation of the speaker—ascribing to her an intentional action of a certain kind, and hence at the very least supposing the existence of some configuration of beliefs and desires which explain that action. The theme of my account is: the epistemically responsible hearer will do a bit more of the same. She will assess the speaker for sincerity and competence, by engaging in at least a little more interpretation of her.16 And what is the nature of this interpretation? The hearer is to construct enough of a theory of the speaker, and relevant portions of her past and present environment, to explain her utterance: to render it comprehensible why she made that assertion, on that occasion. Whether the speaker’s assertion is to be trusted will fall out of this theory
Testimony and Cognitive Virtue
which explains why she made it; and it is difficult to see how sincerity and competence could be evaluated other than through the construction of such an explanation.17 Examples of how this explanation can be played out may be when an ordinarily reliable informant makes an outrageous claim. Fricker comments, “We may ask ‘Has she gone crazy? Or been elaborately tricked? Is she kidding?’—Or is the best explanation that her outrageous claim to have seen flying saucers is really true ? We feel at a loss; but it is these alternative explanatory hypotheses that we dither between.” 18 I cannot think of a less dramatic example illustrating Fricker’s point here. It is unclear to me, for instance, if Coady’s telephone caller requires such an explanation of the informant’s competence. If so, it is certainly not the same sort of explanation we offer of a testifier who makes an astonishing claim. To bracket this criticism for a moment, Fricker’s idea is that the hearer is to construct an explanation of why the speaker uttered p. If it turns out that H’s best explanation for why S uttered p on occasion O is that S is sincere and competent, then if H believes p (via S’s say so) and p is true then H knows p via S’s testimony that p.
The relevant reduction is local Fricker is sensitive to whether the reduction is practically possible and thus the kind of reduction she requires is ‘local’ and not ‘global’. Regarding global reduction Fricker states, The project of trying simultaneously to justify all our beliefs which rest in any way on testimony (or equally, to justify a single testimony-belief without appealing to any beliefs based on testimony) is not one that is properly embarked on, and we certainly do not need to seek to found these beliefs as a totality in something else.19 This is what Fricker takes to be global reduction, and she clearly does not endorse such a reduction. But she does endorse a local reduction (LR). She says, A ‘reductionist’ account of knowledge through testimony, in the context of this approach, means such a local reduction of each instance of knowledge through testimony to broader categories of knowledge, and patterns of inference.20 Fricker agrees with Coady that there cannot be a global reduction of each instance of testimony—i.e., “The greatest strength of Testimony is Coady’s
entrenched awareness of the impossibility of global reduction. He shows convincingly, through examples elaborated in many places, the diffused general dependence on past testimony in the belief system of each of us.” 21 A global reduction would require getting behind, so-to-speak, our past dependence on testimonial claims, and this is not practically possible. Fricker is in agreement with Coady on this point, but they seemingly part ways regarding the possibility of LR. Local reduction requires something much weaker than global reduction. Fricker characterizes (LR) as follows, Local Reductionist Claim: It can be the case that, on a particular occasion O when a speaker S makes an utterance U and in doing so asserts that P to a hearer H, H has, or can gain, independent evidence sufficient to warrant her in taking S to be trustworthy respect to U.22 The idea is that each instance of testimony must be assessed on its own terms. Indexing the reduction to a particular instance avoids any infinite regress or vicious circularity in her position. A regress starts only if one requires every instance of testimony to have independent justification for this speaker’s trustworthiness and not merely justification for believing the speaker on this occasion. If Fricker required justification independent of the present speaker, then either the justification is from a different speaker, or it is from a nontestimonial source. If from another speaker’s testimony, then the hearer needs a justification for her trustworthiness. But here again, since there is a requirement to obtain evidence independent of the speaker, this latter speaker’s testimony needs independent confirmation. I see no stop to this regress. If independent confirmation is from a nontestimonial source, then the hearer needs to confirm the original speaker’s trustworthiness from first-hand observation. This latter option is wholly impractical especially for domains of testimony which require expertise, or historical knowledge. So, by indexing the justification to a particular utterance, the weight of justification can be carried by testimony from a speaker (even the same one) uttered on a different occasion. The point of LR then, is that in certain circumstances one must reduce a testifier’s claim to “. . . broader categories of knowledge, and patterns of inference.” Now, if we make an inference that this testifier is competent, the content of that inference can include the testifier’s own utterances on different occasions. The justification required for LR is independent of the specific utterance, not the speaker herself.
Is local reduction practically possible? Thus far we have dealt with how an investigation of the speaker is to be carried out via LR. Is it practical though? Fricker comments,
Testimony and Cognitive Virtue
Such monitoring of speakers, and appropriate doxastic responses formed on its basis are, I suggest, usually found in ordinary hearers, at least to some extent. However, this sort of monitoring for signs of untrustworthiness in a speaker is typically conducted at a non-conscious level.23 But if it is nonconscious, how can this process provide a justification that S is trustworthy? Fricker answers, “. . . we may insist, compatibly with the subpersonal character of these perceptual or quasi-perceptual capacities, that the subject’s beliefs must not be opaque to her, in that she must be able to defend the judgment which is the upshot of this capacity. . . .” 24 So, having a justification requires having an explanation for S’s utterance and having such an explanation requires being able to articulate her belief that the testifier is competent. (By the term ‘sub-personal’ I take Fricker to mean that one is not conscious of this filter capacity being active. What we are conscious of is the informant’s utterances and the relevant nonverbal cues within which this utterance takes place.) Concluding her theory, she states, “Thus, on my account, a person may gain knowledge from others only when she has the needed conceptual framework to conceive and understand them as persons and agents. . . .” 25 I assume here that what Fricker means to say is that we can have the relevant justification only if we understand that our testifiers have many different motivations and reasons for believing what they communicate. Having the relevant justification, according to Fricker, requires being sensitive to these various psychological properties of persons. Now, there are two qualifications for this theory, corresponding to the two conditions, i.e., sincerity and competency. We are to assume on Fricker’s theory that an agent is sincere unless there are (apparent) cues telling us otherwise. She says, “. . . sincerity is the default position, in assessing a speaker . . . one is justified in taking a speaker to be sincere, unless one observes . . . symptoms of duplicity.” 26 With respect to competency, there are some areas we may assume that a speaker is competent on giving testimony. Examples include the weather, one’s name, perceptions—not too specific, and memorial judgments—not too long past. Given such qualifications, is Fricker’s theory any different from Coady’s approach? One may think that the ‘sincerity’ qualification falls within the scope of Coady’s externalism. We are to assume sincerity in all cases, unless a defeater is quite apparent—and Coady accepts the role of defeaters. As for the competency qualifier, there is a difference between Coady and Fricker. Fricker claims that the difference lies in the fact that in domains where competency is not assumed (and common sense tells us which domains these are) Fricker recommends that one search actively for cues whereas Coady does not have such a requirement. “My account requires a hearer always to take a critical stance to the speaker, to assess her for trustworthiness; while a true [externalist] thesis . . . does not.” 27
Before addressing some shortcomings of these approaches, I should highlight two important issues. First, both accounts attempt to accommodate the contextual nature of testimony. For Fricker, certain cases require the hearer to actively search for defeaters of some sort, and in other contexts, the hearer may assume the competency and sincerity of the speaker. Second, Fricker’s and Coady’s ideal hearer is proximate to what a virtuous person would do in receiving testimony. The basic idea to glean from both approaches is that the hearer should engage in different belief-forming practices depending upon the context within which a piece of testimony is given. And this is exactly what a virtuous person would do, i.e., adjust her belief-forming practices to comport with the context. We may then highlight some shortcomings with either theory.
Shortcomings of both theories One key worry with Fricker’s account is that it is very limited in its scope. It seems implausible that her account applies to writer–reader cases of testimony. In such cases, the reader does not have access to the writer’s psychological states, at least not an aspect of the writer’s psychology that may evince a defeater of some sort. Clearly, the reader is quite limited when assessing the sincerity and competency of the writer. Fricker’s account seems to entail that if a reader is not in a position to build the requisite explanation of a writer’s utterances—by being attentive to psychological cues of the relevant sort—the reader cannot have testimonial knowledge from a writer. Fricker could accept this result of her theory, but claim she never intended her theory to cover all cases of testimony. But then her account of testimony becomes very limited. A very large chunk of what we come to know comes to us via the written word. A second problem with Fricker concerns her silence on speaker-centered conditions for testimonial knowledge. Granted, she requires that the hearer have evidence independent of the speaker’s very utterance that the speaker is competent (and sincere), but it is an open question whether Fricker requires that the speaker in fact be competent. And even then, it is a further question whether, for Fricker, competency entails knowledge. According to Fricker, S is competent with respect to p iff, if S were to assert sincerely that p, then p is true. Given this definition, all we have grounds for asserting is that if S is competent, then she is a reliable witness with respect to p, but not that she knows p. Of course, Fricker may claim that she is only giving an account of how a receiver gets knowledge via testimony, and thus delineating speaker-centered conditions is irrelevant. But this response amounts to the claim that a receiver can get knowledge that p from S even if S does not know p. And to assume that a receiver can receive what the speaker does not have (i.e., knowledge) needs an argument.28 A third problem for Fricker concerns the work that the concept ‘local reduction’ (LR) plays in her account of testimony. In some cases LR is too weak, for
Testimony and Cognitive Virtue
example, cases in which a second opinion from a doctor is warranted or cases requiring multiple expert witnesses to testify in a court of law. In such cases, having justification in a single speaker’s sincerity and competency is not good enough to transfer knowledge since what is required for testimonial knowledge in those cases is corroborative testimony. Independent evidence that a single speaker is competent with respect to p is sometimes not enough to transfer knowledge that p. On the other hand, LR seems too strong in cases in which trust is the default position. In fairness, Fricker openly acknowledges that LR is not required in default situations. But this indicates that Fricker’s account of testimonial knowledge does not cover all cases. Coady runs into the same trouble as Fricker on this score. James Beilby in an illuminating critique of Coady’s Testimony sets up the contextual nature of testimony the following way, To my mind, it is important to acknowledge the differences in the way we approach the “testimonies” of (1a) those accepted as experts and (1b) those not accepted as experts. Similarly, the extent of our acquaintance with a person affects the value we ascribe to their testimony, i.e., (2a) friends and family vs. (2b) complete strangers. There are also certain classes of people that are generally deemed to be either more or less reliable, e.g., (3a) clergy . . . as compared to (3b) politicians and used car people. Lastly, there is also a difference in the way we approach testimony on subjects which we (4a) know a little about, but care a great deal, (4b) know little about, and care little about, (4c) know a great deal about, but care very little about, and (4d) know a great deal about, and care a great deal about.29 Although Coady and Fricker fail to capture a unified account of testimony, if we look at the similarities between them, we can see that they are certainly grappling with the variegated nature of testimony.30 In so doing they pave the way for a responsibilist account of testimony insofar as they dress their receivers up in virtuous clothes. For instance both Coady and Fricker, countenance alertness to the environment and to potential defeaters.31 Furthermore, Coady mentions, Contrary to what we are inclined, unreflectively, to suppose, the attitudes of critical appraisal and of trust are not diametrically opposed, though in particular cases, one cannot, in the same breath, both trust what a witness says and subject it to critical evaluation. What happens characteristically in the reception of testimony is that the audience operates a sort of learning mechanism which has certain critical capacities built into it. The mechanism may be thought of as partly innate, though modified by experience, especially in the matter of critical capacities.32
Here Coady seems to countenance the “sub-personal monitoring” Fricker recommends. Commenting on these and other passages, Fricker writes, “The position Coady is committed to in them, concerning what is required of a mature hearer, is one sketched above [Fricker’s own theory], and not a more radical [externalist] thesis.” 33 The point is that the major contemporary treatments on testimony countenance the same intuition. The intuition is, roughly, that testimony is a variegated form of knowledge acquisition. In some cases, an agent is required to have some sort of justification in her informant’s credentials and in others, there is no such requirement.
IV. Speaker-centered conditions In this section I deal with accounts which aim to outline speaker-centered conditions. This section focuses on a set of examples which highlight our intuitions on testimony. I make reference to these examples throughout the remainder of this chapter.
Plantinga on speaker-centered conditions Alvin Plantinga has the following to say about the qualities of the speaker in testimony, “. . . testimonial warrant, like water, rises no higher than its source. (Alternatively: a testimonial chain is no stronger than its weakest link.)” 34 By this, Plantinga means that a receiver of testimony t’ that p cannot possess more warrant in believing p via t’ than the testifier himself. This thesis suggests that if the receiver comes to know p from the testifier, then the testifier must also have known p at the time of communication. And here we encounter our first speaker-centered condition, i.e., A receiver R comes to know that p by accepting testifier T’s statement that p only if T knows that p.35 Now consider the following case from Jennifer Lackey, Suppose that a Catholic elementary school requires that all teachers include sections on evolutionary theory in their science classes and that the teachers conceal their own personal beliefs regarding this subject-matter. Mrs. Smith, a teacher at the school in question, goes to the library, researches this literature from reliable sources, and on this basis develops a set of reliable lecture notes from which she will teach the material to her students. Despite this, however, Mrs. Smith is herself a devout creationist and hence does not believe
Testimony and Cognitive Virtue
that evolutionary theory is true, but she none the less follows the requirement to teach the theory to her students. Now assuming that evolutionary theory is true, in this case it seems reasonable to assume that Mrs. Smith’s students can come to have knowledge via her testimony, despite the fact that she fails condition (ii) [S knows that p only if, . . . (ii) S believes that p] and hence does not have the knowledge in question herself. That is, it seems that she can give to her students what she does not have herself.36 In this case, Mrs. Smith does not even believe p, and thus does not know it, and yet her hearers can come to know p from her testimony that p. Suppose we consider the converse of Plantinga’s condition according to which, If T knows that p and T sincerely asserts that p to R, and R justifiably accepts T’s statement that p, then R comes to know that p.37 Here, that the testifier knows p is sufficient (given other intuitively obvious conditions). Unfortunately, even this formulation is specious. Consider the following example from Peter Graham, The military of a small country hopes to stage a successful coup, and pays off or threatens the reporters of the country’s newspapers to report that the President has been assassinated regardless of what in fact happens. All but one of the reporters gives in. Andy will report what really happens, and not just what the military wants him to report. As it turns out, the assassination attempt is successful and Andy is the only eyewitness. The other reporters do not even know or even inquire into what really happened. Andy writes in his bylined column that the President was assassinated. When Jenny reads Andy’s article, does she know that the President was assassinated?38 Let’s further assume that Jenny does not know Andy or have any special proclivities toward Andy’s paper. The intuition is that Jenny does not know. To see this more clearly, Peter Graham draws the analogy between this case and the Barn-facade case.39 Suppose one were driving through a small Wisconsin town and sees what looks like a barn. Our driver then forms the belief “that is a barn.” We may suppose that our driver’s perceptual faculties are functioning well and that the barn he sees really is a barn. With just this information, it seems clear that our driver knows ‘that is a barn.’ Suppose, however, we add the following condition: this particular town has erected countless barn facades, 10 facades for every one real barn. Now suppose our driver makes the same true judgment as before without the knowledge that this town has erected numerous barn facades. Even though this belief turns out to be true, we would still not say that he knows. The truth of the belief is accidental. Graham comments,
Virtue Epistemology We can then see the analogy between Jenny and the subject driving through the country-side: Andy’s report is like the barn and the other reports are like facades. Jenny, though she reads a report that is just as good as having the fact before her, does not come to know because she would have formed the same belief on the basis of the same type of cognitive state, reading a “fake” report. Andy’s report and the other reports are, for her, cognitively equivalent.40
So, Jenny gets the truth by accident (and thus does not know), and yet Andy knows that p. These are important examples, and I attempt to address them in elucidating responsibilism. Such an account outlines both receiver-centered conditions and speaker-centered conditions, both of which appeal to an act of intellectual virtue. Additionally, responsibilism countenances the variegated nature of testimony and does so with reference to a single concept i.e., a virtuous act.
V. Cognitive virtue and testimony In this section I apply the theory of cognitive virtue to testimony. Recall from the Introduction that an act of intellectual virtue is the definiendum of knowledge. Before applying this theory, some comment should be made as to why one concept is being used to define an epistemic state that is so contextually sensitive. The lesson from the previous two sections is that the conditions for successful transfer of knowledge shift from context to context. But there still must be a property that grounds saying that each case is a case of successful testimony. Fricker states that a theory of knowledge must specify “. . . a general conception K of what it takes for a belief to be knowledge, which applies to all beliefs of any individual, whatever their subject matter, and regardless of how they were acquired.” 41 Fricker calls this the Unity Constraint. I would only add what I call the plurality constraint, which is simply the claim that for testimonial knowledge there are various conditions and contexts within which testimony is successful. An account of testimony must abide by both constraints. It must be a unified account, and it must explain what counts as successful testimony in widely disparate contexts. To see how a virtuous act functions in a definition of successful testimony, it is important to delineate briefly just what testimony is. The account I defend here assumes that there are three parts to testimony. Each part is driven by relevant questions pertaining to an analysis of testimony. There is first (i) the relationship between the testifier and the proposition shared. Relevant questions include, “Must the testifier know the proposition? May she testify to what she does not even believe?” Second, there is (ii) the communicative act by the testifier. A relevant question includes, “Can a reliably false liar transfer knowledge?” The third part is (iii) the reception by the receiver. Relevant questions include,
Testimony and Cognitive Virtue
“May the receiver acquire testimonial knowledge simply by trusting the witness? Does she need to verify the reliability of the witness?” These three parts signify three distinct epistemic acts, and thus each part requires an analysis of its success. For example, testimony includes an act on behalf of the testifier to adopt the relevant proposition. If this part were not functioning well, for example, if the testifier was intellectually lazy in acquiring an important claim, then even if a true proposition gets transferred, we would not say that the receiver gets knowledge. If the communicative act is deceptive then we have another link in the chain that can break the transfer of knowledge. If the receiver acts viciously in believing the communicated proposition (e.g., the receiver does not locally reduce the testifier’s utterance when she should have) then again we have unsuccessful testimony. It is not surprising that each step is open to analysis since each step involves an epistemic act.42
A definition of testimonial knowledge Assuming there are three parts to testimony, the account of successful testimony I defend runs as follows, Responsibilist Account of Testimonial Knowledge (RTK): For any receiver of testimony R and act of testimony that p from a testifier S, R knows that p from S iff (i) S performs an epistemically virtuous act in adopting p, (ii) S performs a virtuous communicative act in communicating that p to R and (iii) R performs an epistemically virtuous act in believing that p from S. The first condition I will refer to as the acquisition condition, the second as the communication condition, and the third as the reception condition. I turn now to clarify and motivate each condition.
The acquisition condition With respect to the acquisition condition, the idea is that the testifier has to be motivated to attain an epistemic state whose value is derivable from truth and is successful in attaining the truth because of her virtuous act in adopting p and S adopts p iff S understands the propositional content of p. Three clarifications are in order for this condition. First, being motivated to attain an epistemic state whose value is derivable from truth is compatible with not believing the respective proposition(s). A journalist may be motivated to obtain coherent information about the presence of biological weapons in Iraq, but never come to believe that Iraq has biological weapons even if she obtains very good undefeated evidence for such a claim. What the journalist comes to know is what the evidence is for there being biological weapons. A teacher may be motivated to know what evolutionary theory says in order to faithfully and accurately teach
the theory to her students, irrespective of whether she believes the theory or not. Second, in getting the truth, one need not have second-order knowledge that she has the truth. The journalist may get the truth because of her expert investigative virtues, but not know that she has the truth. Third, suppose the testifier performs virtuously in adopting p, and does not believe p, but communicates p anyway. Is her communicative act vicious in that her aim is to deceive? I think this is not our only option here, since she can simply be reporting the views of others. Lackey’s example above of Mrs. Smith illustrates just this sort of compatibility. Mrs. Smith performs her role as teacher virtuously in adopting p. We can say that she is reporting the views of others, but is not intending to deceive her audience.
The communication condition Formulation of the communication condition is driven by two cases of faulty communication. The first is a case of deception whereby a true proposition gets transferred accidentally. Suppose S on a rare occasion comes to believe a false proposition p. But suppose that on this occasion S also forms the intention to deceive. Thus S communicates ~p (which is true). In such a scenario the receiver gets the truth but by accident. In this case, it is dubious whether we would attribute knowledge to the receiver. The communication condition rules in favor of this intuition in that all cases of deception do not count as transferring knowledge. The second case of faulty communication is if a falsehood gets transferred accidentally. Suppose S comes to adopt p virtuously, but she believes ~p (nonvirtuously) which is a false proposition. (The case involving Mrs. Smith illustrates how this scenario can be possible, though I argue below that it is not plausible.) Suppose that S really intends to communicate the truth and thus, she communicates what she believes and in this case, this is a falsehood (~p). Here is a case in which the testifier acts virtuously in adopting p, and intends to tell the truth, but R does not get knowledge. The communication condition requires that what gets communicated must be what gets adopted. Insofar as our testifier in this case does not communicate what gets adopted, the transfer does not count as knowledge.43
The reception condition With respect to the reception condition, we have already seen from Coady’s and Fricker’s treatments that the receiver must be alert to her environment, to the domain within which the communicated proposition falls, and to the psychological cues elicited by the testifier. Being attentive to such factors requires being motivated correctly and as Coady points out, the attitude of critical appraisal is something that is most likely acquired. Thus, the receiver is to exude virtuous behavior in assessing a testimonial claim.
Testimony and Cognitive Virtue
Reasons for RTK Turning now to reasons for RTK, the first is that it appropriately satisfies both the unity and plurality constraint. A virtuous act A is an act that persons with virtue A would characteristically do in the appropriate circumstances, and this captures the contextual nature of testimony. For example, in Coady’s telephone example, the agent is alert to the environment and to any possible defeaters for believing the anonymous employee of the telephone company. Consider cases in which one is receiving testimony on someone’s character. The testifier evinces an inapposite glee as she harangues the person’s character. The example illustrates that a virtuous receiver of testimony will be emotionally sensitive to the speaker, and this sensitivity leads to alethic belief formation. It leads to alethic belief formation (in the present example) in that it screens out potentially false beliefs. Both of these examples highlight the receiver’s virtues. There are also virtues for the testifier. The most obvious example is that the testifier intends not to deceive her receivers. This includes, however, relaying any defeaters the testifier may have if she thinks those defeaters are strong enough to share with others. A virtuous testifier will be alert to the strength of her defeaters and the purpose of the testimony itself (e.g., in a court of law the virtuous testifier would state any relevant defeater, whereas a person caught in the midst of Cartesian skeptical worries will continue to relay what she believes even though she possesses a defeater for those beliefs). Some comment is warranted on how RTK avoids Lackey’s example in which a testifier caught in Cartesian skepticism relays a piece of testimony.44 Lackey thinks, rightly, that the receiver of such testimony gets knowledge, but the testifier does not have knowledge because she holds several undefeated defeaters for her beliefs. Our intuitions would say that the receiver gets knowledge, but only because the case trades on our intuitions concerning the legitimacy of Cartesian skepticism. The responsibilist will say that the virtuous testifier will avoid disingenuousness or deceptive motivations. Doing so will involve metamonitoring the strength and relevance of the defeaters. And Cartesian-type defeaters are irrelevant to whether the testifier gives accurate directions. So, on the responsibilist account, the receiver comes to know even if the virtuous testifier does not have knowledge due to Cartesian-type defeaters.45 Let’s revise the case such that the testifier is giving witness on a contentious moral issue but withholds important defeaters. Suppose that John is a researcher on human embryonic stem cells. In giving senate testimony on federal funding for embryonic stem cell research, he discloses that such research holds out the promise of benefit for numerous human diseases due to their plasticity. John intentionally withholds, however, that no human being has benefited from the use of embryonic stem cells, and that there is no human trial expected— since such cells are notorious for forming tumors. John acts disingenuously in giving his testimony in that the domain of discourse is contentious and he
intentionally fails to communicate important defeaters. Let’s stipulate that it is true that embryonic stem cells will benefit future human beings. I suspect that the receivers do not get knowledge due to John’s disingenuous communication. The example illustrates that in some circumstances, withholding defeaters does not transmit knowledge to the receivers—it matters what kind of defeater it is. On responsibilism, a receiver can get knowledge from a testifier who does not have knowledge (due to defeaters) only if the testifier acts virtuously. In some cases, withholding defeaters is not vicious, in others it is.
Analysis of popular examples A further motivation for RTK is that it is consistent with our intuitions for various popular cases in the literature on testimony. Recall the newspaper case involving Andy (the reporter) and Jenny. The example aims to show that the following thesis is false: If S knows P and S sincerely states that P and H justifiably accepts S’s statement that P then H comes to know that P.46 For my purposes we can add that Andy performs acts of intellectual virtue in forming his belief that the president was assassinated, he performs a virtuous act of communication in sharing his beliefs to Jenny and that Jenny herself takes up Andy’s report virtuously. The intuition is that Jenny does not know. What can the responsibilist say? A crucial feature of the example is that Jenny gets the truth not because of the features of Andy’s epistemic praiseworthiness or of her own. She forms the beliefs she has because of Andy’s report, but she does not get the truth because of Andy’s praiseworthy behavior. She gets the truth by accident. So, RTK entails a result that comports with out intuitions in this case. Suppose we were to vary the newspaper case slightly in the following way. Suppose that none of the reporters accept the deal and consequently report the truth because of the virtues they exude in procuring factual information. In this case, Jenny certainly comes to know, and yet all that has changed is that the rest of the reporters (whom she does not read) report veridical information as well. Thus, if Jenny does not know in the first case, but does so in this case, then it must be something other than Andy’s or Jenny’s epistemically virtuous acts alone that makes the difference.47 Consequently, the explanation for why Jenny comes to know has nothing to do with her intellectual virtue or that of Andy’s. Fulfilling the conditions of RTK is not necessary to transfer knowledge. I have already responded to this style of objection in the Introduction. The basic line of response is that the objection confuses the definition for successful transmission of knowledge and an explanation for why the conditions are or are not met. The change in the environment shifts the explanation we give for why the conditions outlined in RTK are not met.
Testimony and Cognitive Virtue
Having said that I should point out that the contrast between these two cases highlights an interesting feature of testimony. Testimony is often an affair conducted by communities to other communities and not simply an individual to another individual. Though Andy alone communicates to Jenny that p, it is not the case that an explanation for why Jenny knows or does not know will make reference to Andy alone. But if this much is granted, the virtue theorist will advert our attention to the character of the epistemic community that acquires and shares the knowledge. So, an explanation for why Jenny gets knowledge will be a function of the virtues parasitic on the community as a whole and not any one individual. The revised newspaper case is only a counterexample if the virtue theorist explains successful transfer of knowledge strictly with reference to cognitive virtues parasitic on individuals. But there is no reason to think that the responsibilist is confined to such avenues of analysis. Inasmuch as epistemic communities can be the subjects of intellectual virtues, to that extent there is room to explain knowledge acquisition and transfer with reference to the epistemic properties of a community. The response just given runs two risks: the risk of a regress (or vicious circularity) and the risk of being ad hoc. The critic may claim that Jenny gets knowledge because of the virtues of the community, but then what explains how the community gets knowledge? In response, every testimonial chain ends in first-hand observation (‘first-hand observation’ is taken here very broadly to include both empirical knowledge and knowledge gained from reasoning). Thus, the explanation for why the relevant community has knowledge is because of the virtues of the first-hand observers. If the explanation for knowledge transfer is a function of the community, the critic may claim that this is an ad hoc way out of the counterexample. In response, a virtue theory in ethics does make essential reference to the virtues of the community. Consider two epistemic communities. In the first community, the media outlets are controlled by a cruel dictatorial regime. Information that is inconsistent with governmental power, whether factual or not, is withheld. The media, then, relays only information not censured by the government, some of it is true and some false. I suspect our intuitions would say that knowledge transfer in this community is unlikely. Consider, however, an epistemic community in which the media is not controlled by the government, and any release of information is peer reviewed with an eye toward depth, accuracy, and clarity of expression. In such an epistemic community, the prospects for successful transfer are much more likely. The difference between the two communities is attributable to the different virtues (or vices) that the communities evince.
Consideration of some hard cases How does the virtue theorist handle the case offered by Lackey involving Mrs. Smith? In this case, we have an example which shows that a testifier can transfer
knowledge even though the testifier does not even believe in the proposition transferred. Lackey’s example points out a key criterion for an account of testimony. We do not want a theory of testimony that requires the testifier to believe the transferred proposition. Requiring belief in the transferred proposition would entail, in many cases that newscasters and teachers could not relay knowledge since often enough they do not communicate what they believe. And this is gravely counterintuitive. At the same time we want an account of testimony in which such conduits of information do get the truth by how they adopt the relevant propositions. So, we want an account of testimony that is weak enough in that it does not require the conduits of knowledge to believe p, and yet we want an account to be strong enough such that the conduits get the truth in adopting that p. These two criteria are embraced by the responsibilist along the following lines. Lackey points out that successful knowledge transfer requires the testifier to be a reliable reporter, not a reliable knower.48 On the responsibilist’s showing, being a reliable testifier requires exuding the relevant virtues in relaying and acquiring the transferred information. We can imagine that if Mrs. Smith did not consult the authoritative texts, or was not intellectually careful, or intellectually thorough, we may suspect that the children come to know even if what she shares is true. Without such virtues the testifier may present the view either inaccurately or relay many false propositions nestled in with some true ones and we would not say that the children know the true propositions from Mrs. Smith in either case.49 (I assume that evolutionary theory comprises multiple propositions.) Being a reliable reporter requires exuding certain cognitive virtues. If the testifier is a reliable reporter, then her audience may come to know from her. RTK is strong enough in that it requires that the testifier is virtuous in adopting what she transfers and this blocks the charge that the testifier is accidentally reliable. For if she performs the requisite virtuous acts in the relevant circumstances, then she gets the truth because of those virtuous acts. One may wonder how it is that Mrs. Smith performs impeccably in adopting p, but fails to believe p. In response, coming to believe a proposition oneself is different than coming to understand it. And it is at least initially plausible to maintain that the virtues required to get the truth in adopting p are different from the ones required to get the truth in coming to believe p. One way to see this difference clearly is to see that Mrs. Smith is a teacher (the following comments can apply to newscasters and journalists as well). Being a teacher is a social role and with social roles come certain virtues specific to that role. The goal of being a teacher is to convey accurate information about, say, evolutionary theory. Performing well in this role requires exuding certain virtues that govern extended investigative practices (e.g., fair-mindedness, intellectual thoroughness etc.). But Mrs. Smith is also a knowing agent herself apart from her role as teacher. And her ‘role’ as a knowing agent has a different set of virtues which facilitate not just getting the truth via certain virtuous investigative
Testimony and Cognitive Virtue
practices but also coming to believe the truth. We can suppose that Mrs. Smith needs to exercise the virtue of open-mindedness in order for her to move from understanding evolutionary theory, to coming to believe it herself. An agnostic, for example, may come to understand the consistency and persuasiveness of Christian belief and its practices, but believing it for herself may require other virtues. The intellectual virtue of courage or open-mindedness may help her in moving from understanding p to believing that p. So, plausibly, there is a set of virtues V1 which are necessary and sufficient for Mrs. Smith qua teacher to be a reliable reporter of information and there is a different of virtues V2 which are necessary and sufficient in order for Mrs. Smith qua epistemic agent, to get the truth in a way sufficient for her to know.50 I should mention a clarification at this point. Newscasters and teachers and other conduits of information do not in a strict sense get the truth because of their virtues. In the Mrs. Smith case, the evolutionary theorists are the ones who get the truth. Mrs. Smith relays the truth because of her virtues in investigating evolutionary theory, but the work of getting the truth in the first place is still the work of the scientists. With this clarification, it is improper to say that the children know strictly from Mrs. Smith. Mrs. Smith is a reliable conduit, but is not herself the full explanation why the children get the truth. In this way, the epistemic virtues of the relevant community must be assessed before a judgment is made on whether the children know. Suppose the scientists were vicious in their belief-forming practices, but just happened to get the truth. Mrs. Smith investigates the views of the scientists with care, though she is bereft of the knowledge that they were intellectually vicious when they formed the relevant theory. In spite of the fact that Mrs. Smith accurately communicates the theory in question, the children fail to know insofar as the originators of the truth got the truth by accident. This clarification helps us to see that virtue theory is able to explain why the truth gets transmitted in testimony and yet does not require that the transmitters believe what gets transmitted. The scientists get the truth because of their virtues, and the truth gets transmitted because of Mrs. Smith’s virtues, and the receivers get knowledge because of the virtues evinced throughout the chain. Do my previous comments entail that the originators of the testimony know? RTK entails only that the originator adopts the relevant propositions virtuously. Suppose the evolutionary theorists in Lackey’s example exude intellectually impeccable behavior in formulating and articulating evolutionary theory. Nothing so far suggests that these very theorists believe the propositions constitutive of the theory. These theorists may lack intellectual trust, trust in the say-so of their peers doing similar research, trust in their own epistemic practices etc. In this way, one can say that they act viciously in forming their own beliefs, but not in sharing the propositions they adopt. RTK allows for scenarios like this and yet is immune to double-luck cases mentioned in the Introduction. It is immune because even though the scientists do not come to believe the propositions
themselves, if they get the truth because of their virtuous investigative practices then they are successful conduits for the truth. How does RTK handle cases of reliably false liars? Suppose that S is consistently false in what he comes to believe but he also forms the intention to lie on many occasions of testimony. Since lying involves sharing what one does not believe, S ends up reliably communicating the truth. The intuition is that the receivers come to know from S and yet S fails both the acquisition and communication condition. Let’s stipulate that R knows that S is consistently right, but does not know that S is a reliably false liar. I make this stipulation such that R satisfies the basing relation, namely, the idea that one must base her belief on the right evidence. There are only two routes by which this counterexample may be dodged: deny that S fails the acquisition and communication conditions or deny that the receivers come to know. I don’t think the first option is at all plausible, but the second is more promising. The case of a reliably false liar resembles cases of double luck introduced in the Introduction of this book. Double-luck cases, to repeat, are cases in which the agent forms a false belief via bad luck, but the agent ends up getting a true belief via a stroke of good luck. The agent ends up with a true belief, but intuitively the belief does not amount to knowledge. The reliably false liar case evinces two strokes of luck as well. The first stroke of bad luck is when the agent suffers from widespread error due to errant cognitive processing. The second stroke of luck occurs when the agent forms the intention to lie on the very occasions his cognitive processing errs. The end result is a true belief. But our intuitions in double-luck cases say the true belief does not count as knowledge, and so the receiver gets a true belief that does not amount to knowledge. That the truth gets transmitted is due to luck. There is an additional source of luck that apply to reliably false liar cases. Lackey describes such a case as the liar forming a false belief, say, that x is a dog, when really it is a cat. Intending to lie, Lackey stipulates that the liar says that x is a cat. And this stipulation introduces another source of luck. The lair could have said that x is a rabbit, or a guinea pig, or any other fury thing that is not a cat. The point is that there are many ways in which one may share what is false, but only one way in which one shares the truth. In the reliably false liar case, forming a false belief plus forming an intention to lie does not entail transmitting the truth. To construct the example in such a way such that the truth gets transferred, the case must stipulate that, per improbable, the liar happens to hit on the truth in what she shares. This is an added tier of luck that must be stipulated in order to get the case off the ground, but it is knowledge destroying luck for the receiver. Finally, consider another example according to which the acquisition condition is too strong.51 We can see why one should weaken it or even throw it out
Testimony and Cognitive Virtue
altogether if we consider the following case. Autistic savants evince very unique cognitive abilities. Some are very adept at determining very complex mathematical calculations. Intuitively, such savants do not perform virtuous acts in forming their beliefs, and yet if one of them communicates to me that the cube root of 3,796,416 is 156 in several seconds, then I will believe him and intuitively will have acquired testimonial knowledge. So, here is a case in which the receiver can acquire knowledge via testimony, and yet the testifier does not, it seems, evince a virtuous act in forming his belief. The recipe for this type of counterexample is straightforward. Simply dress up the testifier T such that, intuitively, we see that T does not perform a virtuous act in acquiring the communicated proposition. Autistic savants are real-world (human) examples of this recipe, but one can think of robots, or computers. The case highlights that there are two kinds of testifiers that pose trouble for the acquisition condition. They are either (a) testifiers that are not human, and reliable, or (b) testifiers that are reliable, but defective to the extent that they do not or cannot exercise a virtuous act. With respect to testifiers in class (a), it is not at all clear to me that computers or robots have beliefs or whether they have mental states at all. An assumption lurking behind discussions on testimony is that testimony is an act of transferring knowledge (or on the virtue account, it is an act of transferring what one understands). Given this, it is dubious to say that robots or computers have mental states like knowledge or understanding and thus do not serve as appropriate sources of testimony. The thrust of this response is to claim that RTK and most other current accounts of testimony pertain to human testimony. With respect to (b), the issue is more complex. First, the intuition behind the example is that savants do not perform virtuous acts in forming their beliefs because they lack the right motivation (though it is granted that they are successful). That is, the intuition is that savants process information automatically. There are two options the critic can take at this point. Either (1) savants process information without any virtuous motivation at all, or (2) they process information without as well-developed capacities for evincing virtuous motivations. If the critic makes the latter claim (2), then the virtue theorist can say that to the extent the savant can evince virtuous actions, to that extent the savant acquires his beliefs as a function of virtuous acts. The savant does not form his belief automatically, the responsibilist may say, but rather he forms his belief as a function of a virtuous motivation to some extent. And this is enough to be compatible with RTK. If the critic takes the former option (1), then the critic runs against empirical research on cognitive processing. My chapter on perception above argued how motivation impregnates seemingly automatic processing. Furthermore, it is known that response times grow longer for savants as the problems become more complex.52 This indicates that the processing savants perform is not automatic. The general point is that the more we dress up the testifier in
“automatic clothes” the more we drift from empirical research on how humans process information. If we leave room for nonautomatic processing, then our intuitions are not pulled enough to function as a counterexample to RTK. Finally, consider the reception condition applied to the children in the Mrs. Smith example. Do they acquire “easy knowledge” and thus serve as counterexamples to RTK? I think the response here mirrors the response to easy knowledge in the case of perception. Children, it is well known, won’t learn anything unless they are attentive. And being attentive in the classroom is clearly a function of the person’s epistemic motivations.
VI. Superiority of responsibilism To motivate this virtue theory more fully, it is not enough to illustrate how it deals with certain problems and counterexamples. One needs to show how responsibilism fairs in relation to other going theories. I have already mentioned that Coady and Fricker have difficulty providing a unified account of testimony in all its forms (e.g., reader-writer) and contexts. Responsibilism, insofar as it defines knowledge with reference to a virtuous act, and the concept of a virtue has built into it contextual sensitivity,53 is able to provide an account of testimony that is context sensitive. Furthermore, it seems clear that the virtue account applies to all forms of human testimony. I turn now to deal with other accounts of knowledge (and testimony) in order to show the superiority of virtue theory over these other accounts. Linda Zagzebski offers an example aimed to show the superiority of her account of knowledge. The example shows that any theory of knowledge that makes reference to counterfactual reliability, i.e., Nozick, Goldman, and Dretske (and Peter Graham for testimony)54 is incomplete. I have already mentioned why Plantinga’s account of testimony falls prey to Lackey’s counterexamples and how the virtue theorist is able to avoid them. I turn now to show how Goldman and others fall prey to a different kind of example, one which again, the virtue theorist avoids. The example aims to show that the properties of the agent are a necessary part of epistemic analysis. Let me briefly turn to show what a counterfactual analysis of knowledge amounts to. Alvin Goldman is a representative proponent of a counterfactual analysis of knowledge. Goldman’s commitment to such a requirement is fairly explicit. He says in one place: Another version of reliabilism was presented in my paper “Discrimination and Perceptual Knowledge.” 55 That paper also invokes counterfactual situations in accounting for knowledge attributions. . . . It says (roughly) that a true belief fails to be knowledge if there are any relevant alternative situations in which the proposition p would be false, but the process used would cause S to believe p anyway.56
Testimony and Cognitive Virtue
And applying this account to a Gettier-style counterexample, Goldman states, Consider Brian Skyrm’s case in which a pyromaniac believes that the next Sure-Fire match he strikes will light.57 It is true that it will light, but not because of the usual chemical condition of previously encountered Sure-Fire matches. This particular match is defective, but it will light because of a coincidental burst of Q-radiation. Here the pyromaniac does not know that the match will light. Why not? He does not know because there is a relevant alternative situation in which the pyromaniac’s evidence is exactly the same but the match does not light, that is, a situation in which the match is defective (as in the actual situation) but there is no burst of Q-radiation.58 The rough idea is that S’s belief that B counts as knowledge only if B is formed by a counterfactually reliable process. Now, Zagzebski’s example against such counterfactual analyses of knowledge runs as follows: Suppose that Jones is very good at identifying vintages of Bordeaux. In particular, she has no trouble distinguishing a ’94 Chateaux Margaux from very similar wines. Black knows that Jones is going to be tasting different vintages of Margaux without knowing in advance the year of the vintage she is tasting. He has installed a device in her head that can make Jones believe that the next wine she tastes is a ’94 Margaux whether it is or not. (Never mind why Black would want to do such a thing.) When Jones tastes the next wine, if she appears about to judge that it is a ’94 Margaux, the device does nothing. But if she appears about to judge that it is anything else, the device will interfere with her tasting sensations and will lead her to think it is a ’94 Margaux. Now suppose that she tastes a ’94 anyway and believes it is a ’94, and Black’s device does nothing but monitor what is going on in Jones’s nervous system. Jones’s tasting faculties and taste memory are working fine and she comes to have a true belief in the normal way.59 Zagzebski goes on to claim that Jones gets epistemic credit in this case, for the same reasons we give the victims of the standard Frankfurt cases moral credit.60 And our intuitions are to say that Jones knows in the present case. One can see that in many relevant alternative situations, ones in which a ’94 Margaux is not served, Jones gets a false belief and yet we may assume that her evidence remains the same in each of these situations. Thus, given Goldman’s analysis of knowledge Jones does not know in the actual scenario, contrary to intuition. To make this example into an example involving testimony we make the following modifications. Suppose Jones brings her friend Smith to the same wine-tasting event. And Black installs two devices, one that ensures that Jones forms the belief that the wine she tastes is a ’94 Margaux and another one to
ensure that she communicates what she believes. As before, if it appears that Jones is about to judge that the wine is a ’94 Margaux, the device does nothing. If it appears that she is about to judge that it is something else, the device activates and causes her to believe that it is a ’94 Margaux anyway (a false belief). As it turns out, the wine she is served is a ’94 Margaux, and she communicates what she believes with no hint of deception. Smith certainly comes to know that the wine Jones has tasted is a ’94 Margaux, and yet in a relevant counterfactual situation (one in which a ’94 Margaux is not served), neither Smith nor Jones comes to know. A counterfactual analysis of testimony must say that Smith (and Jones) do not come to know in the actual world. Again, such a verdict does not comport with the intuition that Smith does come to know from Jones. A virtue account, however, delivers the right result. So far I have spent some time illustrating how RTK treats some popular counterexamples. And I have raised and responded to some objections to RTK. The conclusion I draw is that RTK is immune to clear counterexamples and objections of which other theories fall prey. But what does RTK look like if we back away from borderline cases? More specifically, what are the virtues evinced by testifier and receiver in most normal cases? How does RTK satisfy the unity and plurality constraint? And more globally, what does the responsibilist say about the epistemic status of testimony in relation to other avenues of knowledge? With respect to the first question, it seems clear that the virtue attaching to the testifier in any given situation is authenticity. Authenticity is the desire to communicate truth and success at doing so. Authenticity may be enough for successful transfer in most cases, e.g., reporting one’s name or claims about the weather. In other circumstances, however, authenticity may not be enough. One may need to evince honesty. Following Zagzebski, I take honesty to not only entail a desire to communicate truth, but a concern for the truth. Zagzebski states, But it is not sufficient for honesty that a person tell whatever she happens to believe is the truth. An honest person is careful with the truth. She respects it and does her best to find it out, to preserve it, and to communicate it in a way that permits the hearer to believe the truth justifiably with understanding.61 Examples of being careful with the truth as well as being authentic are cases involving say, testimony in court, or giving directions to a place I don’t immediately remember how to get to. In the latter example, I need to take time to recall the relevant information and communicate it in a way that the hearer can memorize and understand. Getting slightly more robust, in some cases, a virtuous testifier needs to evince what I will call self-attentiveness. Self-attentiveness is an intellectual virtue according to which one is attentive to the potency of her defeaters and is able
Testimony and Cognitive Virtue
to judge whether such defeaters should be passed on. Self-attentiveness is a virtue whereby one is aware of the status of his own belief system and its coherence. In this way it is not a virtue that governs the acquisition of the belief system but rather its current status. My analysis above of Lackey’s example of a person caught in Cartesian worries illustrates this. In such a case, the self-attentive person would give the proper directions without communicating her Cartesian defeater. Jane would recognize the overall status the defeater has for her own belief system and recognize that in the past she saw the irrelevance of such worries. Contrast this case with the following case. A social psychologist is running an experiment the subjects of which are strictly undergraduate students aged from 18 to 22. The hypothesis being tested, however, makes reference to the population in general. The self-attentive experimenter would recognize the force of such a defeater and report them in his publication. My treatment of Coady and Fricker delineate kinds of receiver-centered virtues. In particular, the virtuous receiver would be intellectually sensitive; sensitive to the kind of speaker, the kind of report being uttered, and the way in which the report is being made—e.g., do they lack eye contact? Is this person experiencing too much glee in detracting from another person’s reputation? The virtuous receiver would be attentive to the epistemic context within which each utterance takes place. Backing away to a more global epistemic issue, epistemologists on testimony have attempted to assess the status of testimony. One pertinent question under this rubric is whether a belief held via testimony is as warranted as one held via perception? Alvin Plantinga says no, stating, “In many situations, while testimony does indeed provide warrant, there is a cognitively superior way.” 62 Plantinga goes on to claim that first-hand knowledge is better. On the other hand Thomas Aquinas advocates a stronger position for testimony. He says, And since among men dwelling together one man should deal with another as with himself in what he is not self-sufficient, therefore it is needful that he be able to stand with as much certainty on what another knows but of which he himself is ignorant, as upon the truths which he himself knows.63 The responsibilist placement of testimony in the fabric of knowledge comes closer to Aquinas’ assessment of it than Plantinga’s. To illustrate exactly why, take the following example from Coady (1992), Suppose p is the proposition that the lesser crested twit was in Cambridge in the early weeks of March. The bird is very rarely seen in Cambridge and even more rarely seen as early as March. It also has some resemblances to the more common, garden twit. A, Q, and W are all experienced ornithologists and although the sighting is indeed of the lesser crested twit they each, unaware
of the others’ sightings, have a certain diffidence in their judgements. They are confident enough to pass the message on and to do so without qualification. The chain of transmission goes, in each case, through expert mouths until the ornithologists down the line D, T, and Z each communicate their news to E who is editor of the ornithological magazine The Early Bird. He is somewhat doubtful when he hears from D, impressed when he hears from T, and quite convinced when he hears from Z. . . . My point here is only that where such transmission patterns exist the later hearsay testifier may be in as good if not better position than even an original witness.64 Notice the virtues exuded by the editor. He measured the individual claims against his background beliefs and the established authority of the testifiers. The editor recognized reliable authority, evinced intellectual humility in accepting the ornithologists’ corroborative word, and, relatedly, evinced trust in that authority. What makes this an example illustrating the responsibilist’s exaltation of testimony in the fabric of knowledge is the centrality of intellectual trust and recognition of reliable authority evinced by the editor. The case illustrates that without intellectual trust and recognition of reliable authority the editor would not be in as felicitous epistemic position as he is. Though this is a case in which there are multiple chains of testimony to one receiver, in cases involving a single chain, the recipe for showing how a virtuous receiver can be in a better epistemic position than the testifier is straightforward. Either make the testifier not even believe the transferred proposition (as in the Mrs. Smith example above) or have the testifier not communicate her own defeaters in a way that is still self-attentive. The testifier can refrain from transferring a defeater and remain virtuous in some cases and thus the receiver’s belief that p would hold more warrant so long as the receiver recognizes the authority in question. (Furthermore, even if the first-hand observer did transfer a defeater for p, it is certainly plausible to grant that certain defeaters hold more sway within some belief systems and not others. And it is certainly possible that the defeater for p holds sway only in the first-hand observer’s belief system and not the receiver’s. In such a case, if the receiver is self-reflective in accepting p, his belief would hold more warrant than the first-hand observer’s warrant in believing p.) To summarize, in this section I have tried to show that RTK satisfies the following criteria, (a) it is Gettier-proof in that the central concept is a virtuous act and we have seen that defining knowledge with reference to virtuous acts renders such a definition strong enough to avoid counterexample. It is so impervious because it entails that the truth is obtained (or communicated) and thus is not open to double-luck cases (ref. the Dr. Jones’ case discussed above). And yet, the definition is weak enough to avoid Lackey’s counterexamples that aim to show that the testifier need not even believe the communicated proposition. Furthermore, I have shown that RTK (b) satisfies both the unity constraint and the plurality constraint. It accounts for the contextual nature of testimony
Testimony and Cognitive Virtue
quite naturally insofar as sensitivity to context is built into the concept of a virtue. And finally, (c) RTK is able to deliver intuitive results on various queer cases of testimony. And dovetailing this last criterion, the accounts of testimony from both Goldman and Plantinga fall prey to certain counterexamples of which the responsibilist avoids. I mention this last point as a positive point in favor of responsibilism.
Moral Expertise: The Role of Cognitive Virtues in Generating Knowledge
I. Introduction The American Society for Bioethics and the Humanities published a report entitled Core Competencies for Health Care Ethics Consultation.1 In that report, they outline certain properties of ethics consultants according to which an ethics consultant must be able to critically evaluate evidence and arguments for and against alternative options. They should possess certain knowledge competencies which include knowledge of moral reasoning and various ethical theories. In the professional literature on ethics expertise, the primary approach is to cash out expertise with reference to having certain analytical skills, and that the expert can back her moral judgment up with reasons and justification. Bruce Weinstein, for instance, says, “Expert opinions are justified beliefs rather than knowledge in a domain.” 2 And further on, “In short, the concept of epistemic expertise turns on the justification, not the truth, of judgments.” 3 Jan Crosthwaite4 echoes this same commitment in that for her, providing moral reasons is an essential function of moral experts. Scott Yoder says, “The key is to see that expertise in ethics is connected with justification—a claim to ethics expertise is not based on the truth of one’s judgments but on one’s ability to provide a coherent justification for them.” 5 The common feature in all of these accounts is that they cash out expertise with reference to having justification for one’s moral judgments, and they implicitly eschew the possibility of moral knowledge. I do not wish to take issue with whether there is moral knowledge or not. I plan to steer clear of certain meta-ethical debates (for e.g., realism vs. antirealism), in formulating a moral epistemology. I use the term expertise, then, to refer to the properties of an agent who produces good moral judgments and remain silent on whether these judgments reflect the truth. In this chapter I take issue with the claim that expertise can be cashed out strictly with reference to one having justification. Such a position endorses some version of coherentism or reflective equilibrium. Reflective equilibrium is a version of coherentism according to which one balances her considered intuitions on a certain case with more general principles or rules that putatively apply to such
cases. Furthermore, one’s background philosophical worldview and theories of the person may factor in a specific moral judgment. These three elements, one’s intuitions on a case, the principles putatively governing the case, and the background philosophical theory are balanced together into a coherent whole. I think there are good reasons to reject such an approach to expertise. This basic approach has been challenged quite persuasively by Michael Depaul’s work Balance and Refinement. Depaul’s argument is that a person’s “considered intuitions” can falter, and they cannot be corrected merely through the accruement of moral beliefs, or through the further specification of principles and rules. The problem lay in a lack of what Depaul refers to as formative experience (to be explained more fully below). Depaul suggests that one’s faculty of moral insight can be blurred, and no amount of coherent justification will correct this. What the person needs is moral experience to make her moral perception more acute. Depaul’s contribution is insightful and I borrow from it heavily. But I think this is one-half of the responsibilist account of moral expertise. In so far as responsibilism is modeled on a virtue ethical theory, it is quite at home in giving an account of moral expertise for which other theories are clearly not at home. The responsibilist embraces both the requirement that the expert has justification and Depaul’s insight that moral expertise requires that one’s moral perception is acute. Both criteria are satisfied by the person with phronesis. Additionally, the phronimos possesses valuable motivational and emotional states. Endorsement of this latter criterion is certainly not popular in bioethical treatments of expertise.6 But Depaul himself admits that in order for a person to have a formative experience, she must resonate with it, and respond to it correctly. For the responsibilist, emotion is a type of cognition, and thus ‘seeing’ a moral issue properly involves having the right emotions for the circumstances in question. That is to suggest that the emotional-motivational states of a person partly constitute expertise. The structure of the chapter is as follows: in the first section I delineate what is right about characterizing expertise with reference to justification. I point out, however, that a lacuna exists in the literature on expertise regarding what kind of justification is relevant. I argue that the kind of justification required involves understanding. I explicate and defend the view that understanding is essential to expertise. In the second section I turn to Depaul’s argument that moral inquiry requires reference to the person’s sensitivity. Sensitivity is my term and I use it to refer to the moral expert’s ability to see the moral issues properly in being sensitive to the goods and evils associated with the issue. In the third section I motivate a responsibilist account of moral expertise which incorporates the insights from the previous two sections and countenances the role of the emotions in processing moral information. My overall argument is that moral expertise is a function of one’s understanding and sensitivity. The notion of a virtuous agent, i.e., a phronimos, captures both features in
one concept. The fourth section shifts gears slightly, in that I aim to argue there that one’s emotions are necessary and sufficient for making moral judgments. This section appeals largely to the empirical literature on moral deliberation, and serves the role of arguing for a key claim by the responsibilist, namely, the moral expert must be motivated correctly and have emotions which fit the circumstances. This fourth section aims to vindicate responsibilism, and argue for its superiority over other theories that do not include reference to one’s motivations.
II. From justification to understanding If we look at what characterizes expertise in other areas than morality, we get some interesting results. Chess experts, for instance, are able to ‘see’ the right move.7 They do not, for instance, reason through the implications of possible moves, and pick the best option. They see the best option based primarily on analogical relationships with other board layouts. Physics experts are able to reason forward when solving a physics problem. Novices, on the other hand, run through a series of hypotheses to see if the hypotheses fit the data presented in the problem. Medical experts are able to make accurate diagnoses based on very tenuous information whereas even sub-experts (a general practitioner as opposed to, say, a cardiologist) either will not make diagnoses at all, or are wholly inaccurate. To think about expertise in morality, we may begin by asking what it is that separates novice from expert when it concerns normative morality—morality concerned with what should be done. The first answer that comes to mind is that the novice does not have justification for her moral judgments. Novices may happen to have the right moral judgment on some matter, but are not able to back it up with reasons. The idea that expertise entails being able to justify one’s judgments in the relevant domain is not without precedent. The quotations from Weinstein, Yoder, and Crosthwaite in the introduction indicate an endorsement of this view. The problem with these accounts is that they do not fully capture an intuitive notion of expertise. In what follows, I aim to fill in what is lacking in previous accounts. In one sense, justification is a function of the evidential grounds that one can muster in support of a proposition. In this sense, justification comes in degrees. Weinstein rightly points this out when he quotes approvingly of O’Connor and Carr (1982) “ “S is Justified in believing that P” is equivalent to saying, “S had good grounds for believing that P” or, in many instances, “S has evidence to believe that P.” ” 8 But there are some options left open as to what kind of justification the expert possesses. First of all, there are three different kinds of justification, namely, propositional, doxastic, and personal. Propositional justification refers to the
justification a proposition has independently of whether anyone believes those propositions. The proposition ‘something cannot be and not be at the same time and in the same respect’ is objectively very probable independently of whether someone actually believes that proposition. Doxastic justification comes in various forms, of which I will delineate below, but the basic idea is that a belief p is doxastically justified if an agent S believes p and also believes q where q supports or grounds p. That is, a belief in p is justified insofar as the believer believes other propositions which support p. Notice, the justification here concerns a person’s belief. Personal justification refers to the justification a person may have in forming her beliefs. To see the difference with doxastic justification, suppose a person S were to form her beliefs about political candidates only by reading positive reports about the candidate who represents S’s own political commitments. Here is a case in which S may likely have lots of doxastic justification for her beliefs since they are well supported, but she lacks personal justification insofar as she is close-minded, lacks thoroughness and the like. Personal justification, then, refers to the intellectual traits or dispositions a person exercises in forming her beliefs. Now, it seems that when Weinstein speaks of strong justification, he is discussing doxastic justification exclusively. Weinstein says, “For all propositions which admit to being true (i.e., all objective propositions), experts are more likely to discover their truth than can nonexperts.” 9 But propositions are the proper object of beliefs. We believe propositions—we don’t experience them. Furthermore, the examples Weinstein uses to illustrate expertise appeals to physicians who are able to offer reasons for a particular diagnosis. But these reasons, one may likely assume, are beliefs the physician has. And thus, Weinstein’s notion of having strong justification amounts to strong doxastic justification. This much, I think, is right on. But more can be said in that, intuitively, expertise would include some reference to the intellectual traits of the expert and to the objective probability of the proposition. I will return to say more about personal justification below, but let’s turn to doxastic justification, for even here, more can be said. A rough notion of doxastic justification will say that, S is justified in believing that p if and only if both (i) S believes p, and (ii) S believes evidence e, where e is adequate evidence for p. I say this is rough because we have not spelled out the relationship between S’ belief that p and S’ belief that e. All we have said is that S believes both. But it is quite clear that in order to be justified in believing a proposition, one must believe such a proposition on the basis of adequate evidence. Suppose that S believes e, and p. But also suppose that the reasons S offers in support of p do not make reference to e, but rather to q and q does not support p at all. In such a situation, S believes e and p, but we certainly would not say that S is justified in believing p, insofar as she believes p on the wrong grounds. So, we need to include in our notion of justification the requirement that S believes p on the basis of e, where e provides adequate grounds for p.
Next, we need to specify what kind of evidence is adequate in order to be included in our notion of expertise.10 It is clear given Weinstein’s notion of strong justification that the kind of evidence relevant is doxastic. And doxastic evidence is evidence that consists of other beliefs. Take the following example of justifying one’s belief via other beliefs. Jane argues for the immorality of abortion because she believes it is always wrong to intentionally kill innocent human life. Abortion, she further believes, is the intentional killing of innocent human life. Therefore, abortion is always wrong. The evidence here consists of propositions (which represent reality in some way), but Jane believes these premises and further she believes that such premises support the conclusion that abortion is immoral. Here is a case of doxastic evidence, evidence that consists of propositions of which the agent believes. But there is also non-doxastic evidence. Suppose I form the belief that there is a computer monitor in front of me (and suppose there is one). This belief is certainly justified, though not in virtue of other beliefs I have. I do not infer from ‘my perceptual faculties are functioning properly’ and ‘I am perceiving a computer monitor in front of me’ to ‘there is a computer monitor in front of me’. My simple perceptual belief here is certainly justified—only a hard-core skeptic would deny this—but the justification is a function of an experience or a non-doxastic psychological state. The point here is that a belief can be justified, but the justifier need not be another belief. Depaul’s argument, given below, can be read as an argument that the expert must have non-doxastic justification in the form of a well-refined faculty of moral perception. So far I have argued that doxastic justification requires basing one’s belief on adequate evidence, and that adequate evidence may come in the form of other beliefs or an experience. None of the previous accounts of expertise specify criteria for adequate evidence, but on the responsibilist account, adequate evidence is determined by what the phronimos would determine it to be. There are other features of doxastic justification that must be dealt with, one of which is the exclusion of defeaters. There are at least three ways in which a defeater D may defeat one’s justification in believing p. Where e is a reason for p, a defeater D may defeat either e, the target belief p itself, or the support claim linking e and p. As an example of the first, suppose I were to believe that abortion is immoral based partially on my belief that it is a form of intentional killing. Also suppose that I happen to come across Mary Thompson’s article in which she poses her now famous analogy involving the violinist.11 The argument aims to show that in most cases of abortion, the woman does not intend to destroy the life of the fetus, but rather to live her life without the excrescent attachment of the fetus. Thompson’s argument functions as a defeater to one of my beliefs that support my other belief that abortion is immoral. It does not defeat my target belief that abortion is immoral. Specifically, Thompson’s argument, if correct, shows that my belief that abortion is intentional is false. I refer to this type of defeater as an undermining defeater insofar as it aims to show that one of my beliefs upon which are based other beliefs is false.
There are also what are called undercutting defeaters, and this kind of defeater attacks the support or connection between e and p.12 Suppose I were at a factory in which widgets were rolling off the assembly line. Suppose also that these widgets look red, and thus I form the belief that the widgets are red. Here, my reason for believing they are red is that they simply look red. Now, suppose that the factory foreman approaches me and tells me that red incandescent lights irradiate the widgets. This makes it easier to spot imperfections, but it also has the effect of making the widgets look red even if they were not. This is an example of an undercutting defeater. It undercuts my reasons or grounds for believing p but does not aim to suggest that such reasons are false— as in an undermining defeater. As an example of defeating p directly; I see a sheep in your field from a distance (100 yards, say). I see you the following day and you inform me that there are no sheep in your field—ever!—but you have a dog that looks like a sheep. Here is a case in which the information you share with me is inconsistent with my belief that there was a sheep in the field. In this case, the defeater (the information you share with me) rebuts my belief that there is a sheep in the field. This kind of defeater is called a rebutting defeater. An undercutting defeater, then, defeats the support claim linking the evidence to the target proposition, whereas a rebutting defeater is incompatible with that target proposition and an undermining defeater defeats the grounds directly—i.e., it suggests that they are false. Are there defeaters of the above sort we can tolerate in our conception of an ethical expert? It seems that we could not tolerate an expert possessing a defeater for p and continue believing that p. It sounds epistemically akratic, or worse, dogmatic and neither of these properties fit our intuitive description of an expert. I contend, however, that there are at least four conditions in which we can allow our putative expert to possess defeaters for p and go on believing that p. These four conditions are, first, S believes p (or e) more strongly than she believes the defeater D and this may be a function of having multiple justifications for p. The second condition is if S believes R where R defeats D—R is a “defeater-defeater.” The third is if S believes that there is further evidence that may reveal itself if further investigation were done—and this further evidence would function either as a defeater-defeater or raise the probability of P such that S believes p more strongly than D. The fourth is if S could intuit that p is true even in light of no evidence or even contrary evidence. The idea in this fourth case is that the expert feels right in believing p and the agent’s status as expert overrides potential defeaters. Let me further explicate each of these. Suppose S were to have a non-rebutting defeater D for p. This means that D defeats p by attacking either e, or the belief that e supports p. It seems as if S could no longer go on believing p and still be considered an expert. But this is not the end of S’s story, for we could stipulate that S has many other lines of justification for p. Suppose Jane after reading Thompson’s article, comes to believe
that abortion may not be an intentional act, but Jane further believes that the kind of action entailed in abortion is an act of killing, and this is enough to support her belief that abortion is immoral. The end result will be that S believes p either more strongly than she believes D, or in spite of her belief that D. The reader may notice why I say non-rebutting defeater. If a person possessed a rebutting defeater, there would be the presence of two incompatible beliefs (held to whatever strength within a given person’s belief nexus) and this is a deficiency—at least a deficiency we cannot tolerate in a conception of expertise.13 In the above non-rebutting scenario, there is nothing epistemically deficient about S believing that p so long as she believes p more strongly than D. Is S’s epistemic state pristine enough to be considered an expert? Here again, it seems clear that the presence of a defeater whose subjective probability is lower than the target belief may be included within a conception of expertise. In the medical field especially, physicians running a diagnosis rarely operate without the presence of any defeaters to their end diagnosis. The mark of an expert is not whether she possesses any defeaters to her beliefs, but how she weighs such defeaters. The second condition seems straightforward. Suppose S believes p and D where D is any kind of defeater. Add to S’s belief nexus the belief R where R defeats D. It seems clear that S may go on believing p and remain epistemically beyond reproach. What is involved in the third condition is that S believes p and D, S believes no other defeater-defeaters for D, and furthermore, S does not believe p with greater subjective probability than D. In such a scenario if S continues believing p, then S’s epistemic state is rather dismal. But add to this that S also believes F where F comprises a set of beliefs according to which further investigation will most likely lead either to (i) a defeaterdefeater for D, or to (ii) other beliefs which raise the subjective probability of p greater than D. Now, simply stipulating that S believes F is not enough to save S’s epistemic state from its lugubrious condition insofar as S could just find herself believing that future inquiry will yield confirmatory evidence for p. It looks like S could believe F and be dogmatic (in an intellectually vicious way). So what are the characteristics of F and in particular how can we allow such beliefs into a conception of expertise without also including dogmatism? An expert does not only possess good justification for her beliefs, but is also good at conducting inquiry. And F is a set of meta-beliefs about how inquiry should be conducted and the likely results of such. These meta-beliefs about inquiry are formed consistent with the intellectual virtues of intellectual care and humility. An agent who possesses intellectual care will not only acquire evidence for a belief, but will also meta-monitor her inquiry as to when the inquiry should stop.14 And if the domain were complicated enough to warrant a belief
that further investigation should be conducted, then S would be perfectly reasonable in continuing (or suspending) her belief in p. An agent who is intellectually humble will measure her perspective and knowledge against what she could know. And rightly positioning one’s epistemic position with reference to a field of inquiry helps such an agent gauge her inquiry along alethically orientated routes.15 So the characteristics of F are that (i) it is a set of beliefs about the inquiry, not about the target proposition p. More specifically, F comprises beliefs according to which the inquiry will likely lead to either a defeaterdefeater (i.e., R) or more evidence favoring p over D. A further characteristic of F, and one that blocks dogmatism, is that (ii) such beliefs are formed from a disposition to be intellectually careful and humble. In the jargon mentioned above, F must have personal justification behind it in order to avoid dogmatism. The fourth condition is illustrated by psychological research on moral deliberation on no-harm cases and research on nursing expertise.16 For no-harm research, subjects are presented with cases in which their moral intuitions are strongly opposed to the actions, but when asked to articulate reasons why they think the action is wrong, the subjects are unable to do so. Typically, subjects attempt to point out the harm that it would do to the players in the story, but the story is constructed in such a way that there is no harm done to either party. The subjects make such judgments from moral intuitions in the absence of evidence or even evidence to the contrary. Consider also the following description of nursing expertise: He was banging the side rails, making gurgling sounds and pointing to his endotracheal tube. He was diaphoretic, gasping, and frantic. The [nonexpert] nurse put her hand on his arm and tried to ascertain whether he had a sore throat from the tube. While she was away from the bedside retrieving an analgesic, the expert nurse strolled by, hesitated, listened, went to the man’s bedside, reinflated the endotracheal cuff, and accepted the patient’s look of gratitude because he was able to breathe again. The expert reviewed the signs of a leaky cuff with the nonexpert and pointed out that nagging the side rails and panic help differentiate acute respiratory distress from pain.17 The example indicates several interesting components of nursing expertise. First, the expert nurse evinced immediate pattern recognition. Second, she did so with little evidence, while the nonexpert was unable to discern the problem even after quite extensive search (entering the room, holding the patient’s arm etc.). The examples of judgments made on non-harming cases and the example of expert nurses serves the point that one can intuit the correct judgement in the absence of evidence or in some cases even evidence against. An expert nurse may have evidence that x is wrong with the patient, but given the overall pattern
and similarity with previous like cases, may conclude instead that y is the problem.18 In sum, we can define doxastic justification for an expert S as follows: S’s belief that p is doxastically justified if and only if, (i) S believes p and e, (ii) S believes p on the basis of e, (iii) e is the kind of evidence adequate to support p and (iv) if S possesses a defeater for p, S satisfies one of the conditions mentioned above. But having good doxastic and personal justification in one’s belief is not enough to say that the moral judgment is an expert one. An expert must be able to explain her position. Consider the following example. Suppose Jane were to argue against abortion with the argument given above; i.e., it is always wrong intentionally to kill innocent human beings. Abortion is the intentional killing of human beings. Therefore, it is always wrong to commit an abortion. What is missing in this argument? It may not be obvious, insofar as Jane provides reasons for her position on abortion. The argument itself is a valid argument. But it seems that expertise would demand more, i.e., an explanation for why those premises and not others are important. An argument as such is not enough to call someone an expert on moral matters. Suppose Jane were an introductory ethics student, and she offers this argument against abortion on an exam. Though she gets the argument right, and she believes each premise, she is not an expert unless she is able to explain the relevance of the premises, why the concepts such as human being and intentionality are important etc. Without this explanation, Jane remains simply a good intro student in ethics and not an expert.19 So expertise will include the property of having justification for any one belief—taking into account various kinds of defeaters—and the property of having an explanation for the importance and relevance of that justification to the target belief. It is clear, however, that expertise is not exhausted by one having justification for p and an explanation for why the justification cited is pertinent. The expert must have both a justification and explanation nonaccidentally, and will have such for most of the propositions within the expert domain. What would this look like? The next several paragraphs argue that ‘understanding’ is an epistemic state that blocks accidental justification and captures the intuitive idea that an expert has justification not just for one belief but also for many within her domain. Take the following example. Imagine a young Catholic schoolboy coming across the proposition (A): “St. Mark of Ephesus refused to sign on with the Roman Bishops at the Council of Florence”.
This student would likely be bereft of an explanation as to why St. Mark did such a thing, much less could he defend such an explanation to any interlocutors. Contrast this child’s state of believing (A) with a historical theologian’s belief that (A). Certainly the historical theologian is able to explain why and defend that explanation to other educated interlocutors. But she does this as a function of having a network of other beliefs, which evidentially support or logically entail others. Of course, it is not enough to simply have more beliefs than the child has. The theologian not only has more beliefs, but they are structured in a particular way. The theologian ‘sees’ the connection (whether logical or evidential) between each of the relevant propositions in her area of competence. So, as part of what constitutes understanding, we have both a comprehensive network of beliefs but also the network evinces what others in this regard call depth.20 As an example that illustrates that both comprehension and depth is important for expertise, Charles Darwin had, for many years before articulating evolutionary theory, all the information before him. In this way he had comprehension—as I am using the term—but he did not have depth of coherence until he was able to structure the information in a way that supported and made sense of his experience. Consider another example illustrating the role of structure in generating judgments. Suppose a physician approaches the in-house ethics consultant with a case involving an elderly person who is undergoing multiple organ failure. The physician wants to withhold further treatment against the family’s wishes. The bioethicist may have at her fingertips all the ethical theories on offer, she believes all of the relevant information, and yet she may not issue an answer until she structures or represents the information as a case of futility, and similarly, quality of life. If she represents the problem simply as a case of physician-family disagreement, she will likely not recommend an appropriate course of action. Framing the problem as a case of futility generates questions such as “what do you [family] plan to accomplish with further treatment” or “what do you think the patient’s wishes would be given this condition?” Framing the problem as a case of physician-family disagreement, generates questions such as “what is the transfer policy?” or “What legal risk does the physician incur in acting unilaterally?” Clearly, representation matters for ethical expertise. The information available must be structured in a way that best represents the problem at hand. And this feature of structure and representation is a function of what I am calling depth. Let’s return to (A). Suppose that our schoolboy provides an explanation and justification for why St. Mark did what he did. And just suppose that he happens to get it right. As I have described the situation, he stumbled onto the right explanation. Intuitively, moral experts, do not get their justification or explanations by accident. At least they are not reflectively lucky. There is a logos and order to their answers or conclusions. But to avoid accidental explanation, one needs to possess a network of supporting beliefs and the agent must ‘see’ the
evidential and logical relationships evinced in such a system. It is here that we encounter a common theme, i.e., the expert is able to justify, explain, and understand. The schoolboy does not have understanding, nor does the introductory ethics student. Understanding is an epistemic state characterized by comprehension and depth. Having a comprehensive belief system entails that the belief system holds many of the relevant beliefs for the respective domain. Additionally, understanding includes the feature of not only having a system of beliefs but also includes a kind of perception whereby the agent who understands sees the connections and relations between the beliefs. A belief system evinces depth if the system is structured in meaningful ways. Such meaningful ways include organization of the beliefs in some evidential, etiological, and logical order. The key idea is that having understanding preserves our intuitions that experts do not get the right explanation or justification by accident, and further they have such justification for many of the domain beliefs. Glaucon and Ademantus, for example, cannot give an encomium of the just life, but they think correctly that it is better than the unjust life. Only Socrates is able to give the logos of justice. To summarize, an expert is able to justify, explain, and do so from a comprehensive and deep network of beliefs (i.e., understanding). Comprehension is the property of having a full set of beliefs relevant to the domain in question, and depth is the property of structuring the information in meaningful ways. I will refer to the cluster of properties including justification, explanation, comprehension, and depth as understanding. Understanding is a valuable epistemic state which takes as its proper objects, relationships and structures of one’s cognitive system. Do we have a sufficient conception of moral expertise? At this point, I have moved beyond most contemporary accounts of expertise. I have given a fuller characterization of justification, fuller than Weinstein, Yoder, and Crosthwaite. And I have added to the notion of expertise the notions of explanation and more broadly, understanding. One may be tempted to end the analysis here insofar as the person who possesses these properties has seemingly sufficient features of expertise. In the next section I argue that the analysis needs to proceed. We need to treat if and to what extent non-discursive cognitive processes should be included in an analysis of expertise. I shall call the non-discursive aspect of generating expert moral judgments sensitivity.
II. Sensitivity In this section I delineate and defend the view that ethical expertise requires sensitivity, and the term sensitivity is being used here quasi-technically. The basic idea is that ethics experts ‘see’ the right answer analogous to chess experts
who ‘see’ what move to make next. But further, the ethically sensitive person will have the right emotions given the context. Much literature in moral epistemology focuses on rational deliberation and justification as contributing to moral knowledge. The role of experience and feeling has been looked at with an askance eye. The objection may run as follows: intuitions and feelings do not contribute to the justification of a moral belief insofar as justification is a function of one’s other beliefs, not her feelings or intuitive sense. And if having the right emotive response or intuitive sense does not contribute to a moral belief’s justification, why should it factor so prominently in a conception of ethical expertise? I aim to answer this very question and further clarify what is entailed by my notion of sensitivity. When one peels away the layers of justification for an ethical position (either theoretical or applied) the differences eventually trickle down to differences in intuitions. Failure to account for good vs. bad intuitions is a grave deficiency in an account of moral expertise. Michael Depaul takes on the task of distinguishing good intuitions from bad ones, and en route to doing so, shows the deficiencies of largely coherentist theories. The best summary of Depaul’s position comes from a book review of his Balance and Refinement.21 Mark Timmons lays out Depaul’s view as follows, The basic line of argument in the book can be summarized as follows, Depaul begins his study by defending the claim that a radical conception of the familiar Rawlsian wide reflective equilibrium methodology (a coherentist methodology) is preferable to the more standard conservative conception of this method, since it allows for discontinuous change in moral belief of the sort characteristic of moral conversions—a kind of change that Depaul claims may lead to an overall better outlook. Moreover, radical reflective equilibrium is a more adequate method for guiding inquirers to have rational moral beliefs since, unlike its noncoherentist competitors, it never obligates an inquirer to revise her beliefs in a manner that would run counter to the inquirer’s own epistemic standards. However, even the radical method has its shortcomings because it cannot address adequately cases where the inquirer operates with a naïve moral sensibility or with one that has been corrupted. The remedy for these failings is to move to the method of balance and refinement, a method that emphasizes the importance of an inquirer’s faculty of moral judgment and, in particular, the role of formative experiences in shaping this faculty.22 The move to focus on a person’s faculty of moral judgment is motivated by reflection on the case of Jay. Depaul gives us an example in which Jay has a romantic view of war whereby war is good in so far as it supplies a theater in which courage, honor, and bravery are realized. Jay has never been in a war but
he watches numerous movies which glorify the valor and triumph associated with it and converses frequently with the “old-timers” at the local American Legion. Depaul concludes, “Unfortunately, Jay is rather naïve. The experience . . . that grounds his view of war is very narrow; none is first hand, and the rest is exclusively of reports, essays, novels, and films that romanticize warfare.” 23 His experience is limited, and intuitively, this naiveté grounds the claim that Jay’s view of war has poor epistemic standing. Two observations are worth noting: first, Jay could have a very coherent set of beliefs. He could have well-developed responses to those who hold a nonromantic view of war. Jay’s beliefs may have strong support, but Jay’s intuitions do not serve him well. He does not “see” the issue properly and thus, does not possess sensitivity. This leads to the second observation, namely, that Depaul is implicitly presenting and responding to a familiar objection to any intuitionisttype theory of moral knowledge. Intuitionism comes in different types, but all of them are bound to say that moral intuitions, defined as moral judgments that are not based on any other judgment, bear positive epistemic status and thus, may serve as appropriate justification for one’s moral beliefs. One objection, and seemingly endorsed by many, is called by Mark Nelson as the “no-credibility objection.” Quoting R. M. Hare, The appeal to moral intuitions will never do as a basis for a moral system. It is certainly possible . . . to collect all the moral opinions of which they [various thinkers] and their contemporaries feel most sure, find some relatively simply method or apparatus which can be represented with a bit of give and take, and make plausible assumptions about the circumstances of life, as generating all these opinions; and then pronounce that that is the moral system which, having reflected, we must pronounce to be the correct one. But they have absolutely no authority for this claim beyond the original convictions, for which no ground or argument was given. The ‘equilibrium’ they have reached is one between forces which might have been generated by prejudice, and no amount of reflection can make that a solid basis for morality.24 Depaul’s strategy is to agree with the objection but only for moral intuitions that are the result of limited experience. Depaul remarks, The problem with the naïve moral judge is not that there is any information she lacks, or that the system of theories and concepts that she has built up is somehow inconsistent, or even that there is an alternative theory that she has failed to consider. The problem is that a significant number of the beliefs she is using to develop her moral theory are the products of a faculty of moral judgment that has not been used enough, or been used on a sufficient number of different types of situations, or objects, or perhaps, been used on the right kinds of situations, or objects, for it to have developed to the point
where it is competent to make the sorts of discriminations it is being used to make.25 So, Depaul thinks that one’s moral intuitions may be the result of an unrefined faculty of moral judgment and formative experience may enhance or refine this faculty. The “authority” that Hare demands would be supplied by the refined faculty itself. Wine connoisseurs have refined tastes and are authoritative for that very reason: by analogy, our faculty of moral judgment can be refined as well. Several objections and misunderstandings must be dispatched before continuing. First, reference to a moral faculty may commit one to there being mysterious entities in one’s ontology. The objection would seem to be that if there is a moral faculty it would “see” moral entities. But there are not such mysterious moral entities. Therefore, there is no moral faculty. In response, if we do generate moral judgments, then there must be some cognitive system which generates them. The system may be reducible to some amalgamation of the affective, reasoning, and visual systems or some other combination of cognitive processing. But there is a need to claim that there is some cognitive-emotive process that generates these judgments. On a related point, Timmons claims that Depaul assumes moral realism (which is a bad assumption according to Timmons). But he grants that Depaul sufficiently points out the need for formative experience. Timmons grants, that is, that Jay is naïve, and his naivete eclipses his ability to see moral issues correctly. But Timmons claims that Depaul makes the gross assumption that there is a single “objective moral order” which serves as the object of moral inquiry. There are two worries I have with Timmons’ claims. First, if Timmons grants that experience improves moral perception, then how could he find the assumption that there is an objective moral order repugnant? If experience will sharpen Jay’s moral insight, what is the nature of this sharpening if not a more perspicacious insight into the moral order? The worry here is that Timmons owes us an explanation for why he can grant the importance of formative experience and yet repudiate the assumption that there is an objective moral order. Second, even if Timmons can mark out a place for formative experience without commitment to a moral order, the idea that Depaul is necessarily committed to a realist view is not obvious. Consider the following addition to Depaul’s theory: formative experience refines the person’s faculty of moral judgment. What a refined faculty sees are solutions to moral problems similar to the way creative scientists see solutions to novel problems. The example of Archimedes is indicative. Archimedes was given a gold crown by the king and the task was to determine the volume of gold in it. Since the crown was amorphous, he could not simply measure the surface area and weigh it in order to determine the volume, thus the problem was novel given the existing background knowledge at the time. One day he stepped into his bathtub and noticed the water level rising
as he immersed his foot. At this moment he saw the analogy to the gold crown and determined that he could exact the volume of gold in it by immersing it in water. Solutions to moral problems may be characterized in similar ways, that is, seeing analogical connections, resonating with different facts of the moral problem etc., but none of this entails a commitment to an objective moral order. A person with formative experience may be presented with a moral problem the goal of which is to solve it, but the criteria for a valid solution need not be cashed out in terms of truth or falsity. The goals of the solution may be to countenance as many goods as possible. Given the latter solution parameters, a solution may be appropriate or correct, but not uniquely true or false. It seems that this addition to Depaul’s account preserves everything he says and dodges Timmons’ worries.
Augmenting Depaul’s account With these misunderstandings dispatched, there are certain additions to make to Depaul’s account. Depaul’s view does not explicitly countenance moral expertise begotten through testimony. Depaul’s view seems ego-centered in that it is the moral agent herself who must have formative experiences relevant to make refined moral judgments. But it seems plausible to say that S’s moral judgment can bear equal epistemic status having received the views of others as, say, P’s moral judgment in which P has the relevant formative experience. If S receives the views of P (one with the requisite formative experience), her judgment is not lacking epistemic status compared to P’s own judgment. If there is a lack in status, it is unclear what that lack is. Consider S’s epistemic position upon receiving the views of several persons with formative experience. The work of corroboration may bequeath greater epistemic status to the moral judgment S receives through testimony. Tony Coady relates an interesting example in which corroborative testimony enjoys greater epistemic weight than even first-hand experience.26 Coady has us imagine an editor of a bird magazine who receives several independent reports of a rare bird citing in an area not indigenous for the kind of bird in question. The expert bird watchers alone find such sightings implausible and thus their first-hand beliefs are held in doubt. The situation with the editor is different for he has access to numerous reports from expert bird-watchers, concluding that the bird in question has in fact migrated to an area formerly alien to it. Clearly the editor is in a better epistemic position and yet does not have the requisite perceptual experience to support his judgment. Likewise, S’s received moral judgment may bear greater epistemic status because of the corroborative testimony from several sources even though S herself lacks such experience. The objection from testimony has its limits. It certainly does not show that Depaul’s theory is incorrect: for the testimony to be good, the testifier must
have the requisite formative experience herself. Thus, the objection requires the correctness of at least one main aspect of Depaul’s account. The argument does point out, however, that Depaul’s theory needs to be augmented in the following way: the previous chapter on testimony argued that a successful act of testimony requires (at least) that the testifier and receiver perform a virtuous act in sharing and receiving the piece of testimony. If good moral judgments can be transmitted through testimony, then one can receive expert moral judgments without having the requisite formative experience, so long as the testifier and receiver perform virtuously in the transmission. What may this performance look like? Formative experience is meant to refine the faculty of moral judgment. It is best understood on analogy with perception, where one acquires perceptual expertise through repeated exposure to the right percepts (see Depaul’s example of apprehending different equine gaits, 200ff.). If Jay had more relevant formative experiences he would see things differently on the morality of war. For testimony, there is an analog to perception-type formative experience. In order to receive testimony from an authority already supposes that the person is ready to receive it in that she recognizes the authority of the other and trusts the person’s opinion. An agent may receive an expert moral judgment from another, but only because the agent has certain dispositions of trust and deference to reliable authorities. Suppose the V. F. W. guys share with Jay their own non-romantic views of war. Suppose also that Jay trusts their authority and “looks up to them.” Having such dispositions to receive moral judgments through testimony can be called testimony-specific formative experience. This testimony-specific formative experience would lead to (in part) deference to reliable authorities, and intellectual humility. Depaul’s theory is augmented at this very point wherein the receiver of moral testimony has to possess certain intellectual virtues in order for the received judgment to count as good. Does the role of testimony render obsolete the role of understanding? We may assume that the receiver does not have the understanding the testifier does, and yet I am endorsing the claim that the receiver can render an expert moral judgment. This is not a problem in that what makes the judgment an expert one, is not that the receiver generated it, but rather the person with understanding and sensitivity generated it. The receiver takes up an expert judgement much like Glaucon and Ademantus took up Socrates’ encomium of justice. We have seen that moral expertise is not reducible to coherence, or propositional justification alone (for Jay may have both). Depaul has shown us the need for experience which affects our faculty of moral judgment leading to sensitivity. I have augmented Depaul’s theory by including testimony as a source for expert moral judgments. Depaul’s theory initially looks like it cannot countenance testimony because his theory looks ego-centric. This is mere appearance
in that his insights on the role of formative experience can be applied analogously to moral judgments formed through testimony (though the theory needs to make reference to certain intellectual virtues).
III. Responsibilism and moral expertise The responsibilist countenances the insights garnered from the previous sections. After all, Aristotle, a paradigm virtue theorist, has the following to say about moral perception, It is thought that a young man of practical wisdom (phronesis) cannot be found. The cause is that such wisdom is concerned not only with universals but with particulars, which become familiar from experience, but a young man has no experience . . . (1142a12) And further on he defines the operation of phronesis more specifically, It is opposed, then, to intuitive reason; for intuitive reason is of the limiting premises, for which no reason can be given, while practical wisdom is concerned with the ultimate particular, which is the object not of scientific knowledge but of perception—not perception of qualities peculiar to one sense but a perception akin to that by which we perceive that the particular figure before us is a triangle . . . (1142a25–30). There are certainly interpretive issues with the passage, but two points are clear: moral expertise involves a kind of perception that is sensitive to the salient features of a moral problem. Responsibilism will also countenance understanding insofar as understanding is considered to be an intellectual virtue itself. It perfects the power of the intellect to apprehend reality in its totality. It is clearly an excellence of the intellect to say of one that she has a comprehensive and deep network of beliefs and sees the complex relations between these beliefs. Responsibilism goes further in telling us that such valuable epistemic states are generated by valuable traits in the agent.27 Responsibilism will say that a person’s motivations and overall moral character contributes to the acquisition of sensitivity and understanding. The idea that moral character infects our moral judgments may sound odd to some readers. It may look like an endorsement of doxastic voluntarism. Some further explanation is then warranted. By way of introducing this idea, I should note that Depaul alludes to the role of one’s character when he comments on the potential inertness of experience. He states,
When we ask what a minimal necessary condition is for the adequacy of a method, rather than requiring guarantees, we will simply require that it be possible for some persons to come to have warranted beliefs as a result of following the method, viz., those who come to the process in the right frame of mind, or make the right contribution to the process.28 Being in the right frame of mind and making the right contribution, though, alludes to other features of the person’s moral character. I think Depaul’s observations here are correct and as such, moral character plays a fundamental role in acquiring formative experience.29 Additionally, moral character (good or bad) constrains what moral options we consider attractive. For example, the generous person may reject a course of action that would involve ungenerosity, and do so with little deliberation. A person without the virtue of generosity may entertain such a course of action with little hesitation. There is some intuitive support, then, for the view that there is an adequation between moral character and moral judgment. Not only that, one’s moral character can direct one’s inquiry. These claims, I think, are plausible enough on their own, but will motivate them presently borrowing from Eleonore Stump’s treatment of the virtue of wisdom.30 To motivate the connection between moral character and judgment, Stump offers us some reflections on the biography of Franz Stangl who was commandant of Treblinka during World War II. Stangl was one of those tried and convicted at Nuremberg for war crimes, and Gita Sereny wrote a biography on him during his imprisonment. Initially, Stangl was stationed at a euthanasia clinic in which the very infirm were euthanized. This repulsed Stangl but he eventually talked himself into its permissibility. The reason for doing so is that he either could divest himself of his rank—something he was not willing to do—or accept the position as ordered. Stangl wanted to accept his appointment, and was not willing to give it away, and the other alternatives were not attractive to him. He was doubleminded at first, looking for a way either to see euthanasia as good, or rejecting the assignment. Since we typically do not will actions we see only as being evil, Stangl had to see euthanasia as in some way good, and focus on that good aspect. If a person wills something that is initially apprehended as evil, the will directs the mind to consider what is good about the desired object or action. Evidence in favor of the action being evil is discounted or ignored altogether. In this example, we see Stangl’s motivations (will) directing the intellect toward features of the act that would indicate its goodness. And such a direction is initiated by his behavior. The formula here is fairly simple and common. An agent desires x, but in order to get or maintain x, she must perform y. Action y is initially seen as being evil or repulsive, but the desire for x overrides this, and directs the intellect toward features of y that indicate its goodness. If someone
repeats this formula enough, then we can plausibly substitute in for x ‘being commandant of Treblinka’ and for y ‘engaging in systematic destruction of human life’. Another process is exerting an influence as well. It is a psychological fact that we desire consistency between our moral beliefs and our moral behavior. After engaging in behavior we may initially have seen as repulsive, post-behavior; we “force” ourselves to judge the action differently. The desire for consistency, then, makes it possible that behavior affects our judgments. Sereny tells us that Stangl never thought that what he did was wrong, he was never remorseful and could not understand the public ridicule he endured subsequent to the Nuremberg Trials. He saw himself as an upstanding soldier performing a necessary evil. Responsibilism, then, accepts the importance of understanding and sensitivity. But the conceptual gain for the responsibilist is that she will subsume both under the virtue of phronesis. The phronimos possesses both understanding and sensitivity. Additionally, responsibilism has the added feature of explaining how these features develop, insofar as responsibilism is modeled on a virtue ethical theory. We can now define both an expert moral judgment and a moral expert. A moral judgment J is expert, if and only if, it is a judgment the phronimos would make in the circumstances in question. And a moral expert is one who possesses phronesis. None of this should be taken to entail an individualistic approach to moral deliberation. The phronimos will not only see moral issues properly but will inquire into their nature being motivated to understand the moral problem. This may require dialogical virtues which govern deliberative discourse with other members in her epistemic community. In the next section I shift the discussion slightly aiming to vindicate the responsibilist’s claim that moral judgments involve one’s emotional and motivational states. Consistent with the project of the book, it is important to show that the responsibilist’s method of defining valuable epistemic states with reference to valuable states of the agent (i.e., motivational states) is vindicated.
IV. Motivation and moral judgment Empirical evidence suggests that motivations and emotions play an essential role in moral processing. This is not itself a remarkable claim, but what is remarkable is the claim that emotions and motivational states constitute moral judgment. Furthermore, one’s motivational states can either enhance or detract from the quality of one’s moral judgment. The argument for this latter claim is that the empirical evidence on the role of emotion/motivation in moral judgments shows that given certain emotions, the judgments are distorted. For the former claim, I recapitulate evidence which suggests that without emotive processing, no moral judgment can even be made. Let’s consider the evidence indicating the necessity of emotive processing in forming moral judgments.
Studies on the etiology of attitude formation,31 the role of bias in social judgments,32 and studies on psychopaths all indicate a strong connection between emotion33 and moral judgment. The connection is not only one of co-occurrence, but is also a function of having certain emotions. Jesse Prinz summarizes such evidence focusing on research pertaining to psychopaths. In arguing that emotion is necessary for forming moral judgments, Prinz starts with the wellestablished claim that psychopaths are deficient in experiencing negative emotions such as fear, sadness, and especially empathy. “They have remarkable difficulty even recognizing them [negative emotions] in facial expressions and speech sounds. . . . Without core negative emotions, they cannot acquire empathetic distress, remorse, or guilt.” 34 As a result of these deficiencies, psychopaths engage in immoral behavior, and yet their reasoning, perceptual, memorial, and language abilities are well intact. Regarding their ability to make moral judgments, Prinz claims that they lack sufficient understanding of moral concepts to even make them. Psychopaths may utter the proposition “killing is wrong” but Blair35 discovered that what they mean by this utterance is that “killing is prohibited by local authorities.” As such, their grasp of concepts such as ‘right’ and ‘wrong’ are too deficient to count as making moral judgments. An analogy may help here. A child may utter ‘red is a shape’ on the basis of her seeing many red things. But clearly red is not a shape, and thus, the child is not making a judgment about the color red. Similarly, in uttering “killing is wrong,” a psychopath is stating an empirical fact about the act of killing—that it is prohibited by local authorities—but this is not a statement about the moral quality of killing. In addition to the evidence presented by Blair, Prinz offers a thought experiment concerning a person who has no emotional aversion toward killing. “We might imagine a person who knows everything non-emotional about killing. She knows that killing diminishes utility and that killing would be practically irrational if we universalized the maxim, thou shalt not kill. Would we say of this person that she believes killing to be morally wrong?” 36 Prinz rightly concludes that we would not. Before continuing, an important objection to responsibilism can be obviated at this point. One may claim that simple moral judgments such as: ‘it is wrong to cut up one’s daughter into pieces until she dies a slow and painful death’ are so clearly true that it is hard to see why having the right motivations mediates this judgment. This is a species of the unmotivated belief objection encountered in the chapter on perception. The response is that the evidence outlined by Prinz suggests that psychopaths do not hold this moral judgment,37 because they do not have the right emotional resonance with it. It is part of responsibilism that moral judgments are a function of having the right emotions. In the presence of no emotional repulsion—as is the case with psychopaths—there is no moral judgment with the same content that our moral judgments have.
Further evidence for moral judgments requiring emotive-motivational states comes from the work of Antonio Damasio. He presents neuro-anatomical evidence that moral judgments are a function of emotion. Adults who suffer lesions in a very specific area of the prefrontal cortex (areas associated with decision-making and emotional processing) evince what Damasio calls “ “moral” misbehavior syndrome.” 38 By this he means that subjects with such lesions would consistently perform immoral actions, when previous to the lesion they would not. The evidence here is limited in that it only supports the thesis that moral behavior is tied to emotion, not moral judgment. To argue for the latter, Damasio presents research on lesions occurring in children, prior to the acquisition of moral norms and concepts. What is most interesting is that “whereas adultonset patients know the rules that they violate, early onset patients do not; they failed to learn those rules.” And concludes, “a sector of the prefrontal cortices is: 1) required for making adequate social and moral decisions once the relevant knowledge is acquired; and 2) probably essential initially to acquire that relevant knowledge.” 39 Such evidence suggests that emotional processing is essential for making moral judgments. If emotions are essential for making moral judgments, can they improve the quality of one’s judgment? That is, can emotions either enhance or detract from the epistemic status of one’s moral judgments? Considerable evidence suggests that judgments which are a function of certain kinds of emotion can be distorted. One such stream of evidence comes from research on “attitudinal conviction.” A person’s attitudes are intimately tied to who one thinks she is, i.e., her self-identity and moral judgments are often a function of one’s attitudes.40 Commenting on the connection between attitudes and self-identity, Kari Edwards states, Zajonc claimed that affective reactions implicate the self; Our affective responses not only reveal properties of the stimulus but also inform us of our reactions to the stimulus. Because of the self-referential nature of affect-based attitudes, counterattitudinal information may be experienced as a challenge to the self.41 Here we see that attitudes that are tied to one’s emotive state resist change because they would be tied to one’s self-identity. The point to highlight is the proximity of emotive-based judgments to the self. As such, people typically make judgments consistent with who they think they are and the values to which they are committed. Such judgments risk being subject to bias, or are obstinate to change because of the stability of one’s self-identity. Other interesting evidence of the effects of emotion and motivation on judgment comes from the phenomenon of “freezing.” Consider the following etiology of this phenomenon by Tallie Freund et al.42
A person’s tendency to “freeze” upon a given belief is partially determined by two separate motivations: the need for cognitive structure and the fear of invalidity. Freezing upon a belief refers to a state in which the individual fails to cognize plausible alternatives to a currently held hypothesis [or belief] and fails to attend to information inconsistent with that hypothesis [or belief ].43 Intuitively, motivations leading to the freezing of beliefs can infect one’s moral inquiry in deleterious ways. Finally, there is evidence suggesting that our moral judgments are tied not to the actual moral qualities of the action, but to the emotional proximity we feel with the persons involved.44 Jonathan Haidt and Jonathan Baron showed that persons determined the same action more gravely immoral if the action were done to a friend, versus a stranger. This evidence should not be surprising, but is notable for the purposes of my argument. In summary, several motivations can be highlighted that infect moral judgments: first, there are motivations for consistency in one’s belief system, there are motivations to protect one’s self-esteem (and not to be invalid), and there are motivations to form beliefs consistent with one’s self-identity. What kind of refined motivations could correct the distortions outlined just previously? It seems clear that if one had good motivations, then the corresponding judgments would be enhanced, or at least the distortions would be avoided. But what makes the motivations good? One of the arguments in favor of the theory outlined here is that it is empirically informed. The motivations one adopts are those that would avoid the pitfalls to which other motivations lead. But rather than specifying each motivation associated with good inquiry, I can give a more general characterization of good motivation in considering Ziva Kunda’s work on motivated reasoning. Ziva Kunda’s work in social cognition focuses on the effects of motivated reasoning. In this stream of research, motivations are classified under two broad categories: accuracy motivations and directional motivations. Accuracy motivations are motivations to get an accurate conclusion, and directional ones are motivations to argue for a specific conclusion. Kunda explains the distinction with a helpful metaphor: accuracy motivations are like the judge, directional motivations are like the lawyer. She notes, “The work on accuracy-driven reasoning suggests that when people are motivated to be accurate, they expend more cognitive effort on issue-related reasoning, attend to relevant information more carefully, and process it more deeply, often using more complex rules.” 45 Regarding directional motivations, the course of reasoning is not indeterminate. For directional motivations there are certainly biases that dictate the course of reasoning, but such biases are constrained. Kunda explains, People do not seem to be at liberty to conclude whatever they want to conclude merely because they want to. Rather, . . . people motivated to arrive at
a particular conclusion attempt to be rational and to construct a justification of their desired conclusion that would persuade a dispassionate observer. They draw the desired conclusion only if they can muster up the evidence necessary to support it. . . . They maintain an “illusion of objectivity.” To this end, they search memory for those beliefs and rules that could support their desired conclusion. They may also creatively combine accessed knowledge to construct new beliefs that could logically support the desired conclusion.46 The point here is that persons under the influence of directional goals access beliefs and information, but the access is selective. The conclusion is supported by reasons, yes, but the information selected is a function of the goals and thus, may only represent a subset of the relevant information. I contend that accuracy-motivations are the motivational analogue to Depaul’s formative experience. Depaul argued that in order to “see” a moral issue properly, one needed formative experience. Formative experience works by way of sharpening the moral perception of the agent and/or modifies her ability to resonate with the right features of the moral issue. Analogously, accuracy motivations do for one’s moral motivations what formative experience does for one’s moral perception. Accuracy motivations aim to sharpen or redirect or modify the moral motivations informing one’s judgments. These motivations aim to conduct inquiry that more closely approximates the truth or tracks creative solutions to moral problems.
V. Conclusion The responsibilist account of expertise is able to obviate certain objections to the notion of expertise. The strongest objection to the notion of expertise is the fact of disagreement. The idea is that if there is intractable disagreements in ethics that cannot be adjudicated through reasons, then there can be no such thing as an expert. Being an expert implies being able to generate superior moral judgments. The account offered here responds by pointing out that we should look not only to the reasons a moral agent can generate in support of her position, but what kind of person she is, how she conducted her inquiry, and how emotionally sensitive she is to the relevant moral goods. I suspect that the range of disagreement would severely truncate if the persons are equally virtuous and equally sensitive to the relevant moral goods. Disagreement may persist, which may indicate an axiologically deep and complex world, but admitting this does not imply that there are no experts who can generate superior moral judgments.47
Cognitive Virtue, Divine Hiddenness, and Reasonable Belief or Nonbelief
I. Introduction Suppose a good and loving God exists in world alpha. Now consider the evidence in alpha that could be used to ground or support belief in God. Is that evidence extant in the actual world? The question aims to highlight our intuitions about what we would expect from a good and loving God vis-à-vis manifesting His existence, or making it known to His creation that He exists. Some think that when we compare the evidence in world alpha with what we are given in the actual world, the actual world falls short. This is the basic idea behind the argument from divine hiddenness. The conclusion of the argument is that God (hereafter I shall assume a good and loving God) does not exist. The premises include at least two claims, namely, the evidence for God is too paucious to support rational belief, and we would expect a more transparent display of God’s existence if He did exist. In this chapter I outline some tempting solutions to this argument, and argue that they categorically fail. I then offer my own criticisms of the argument. I admit, however, certain limitations of my criticisms, but such limitations are a happy result in that they pinpoint certain limitations of the inquiry itself. I conclude by outlining the boundaries of good inquiry which if instantiated, dissolve the impasse regarding the hiddenness argument and religious knowledge in general. The chapter outline is as follows: I begin with a recapitulation of the hiddenness argument endorsed by J. L. Schellenberg. I then survey some of the more prominent responses to the argument. My aim is to highlight which ones fail and which ones hold promise with more detail filled in. In the last section I show the limitations of even the more promising responses and chart out reasons for dissolving the impasse.
II. The argument from divine hiddenness The argument from hiddenness begins with the observation that there are those who do not believe in God, and yet these same persons desire to know
God. Additionally, an important subset of these persons conducts their inquiry in virtuous ways. John Schellenberg refers to these persons as possessing reasonable nonbelief. There are those who have conducted their inquiry concerning God’s existence in a meritorious manner, but the evidence amassed is insufficient to support belief in God. More technically, Schellenberg defines reasonable nonbelief with reference to inculpable doubt. S is inculpably in doubt about the truth of G [God’s existence] if (1) S believes that epistemic parity obtains between G and not-G,1 and (2) S has not knowingly (self-deceptively or non-self-deceptively) neglected to submit this belief to adequate investigation.2 We may assume that (1) is satisfied since we typically take it that most people accurately report their state of mind to us. Knowing whether (2) is satisfied is more difficult. Nevertheless, Schellenberg indicates that in some circumstances we could judge that (2) is satisfied, for clearly there are some who exemplify meritorious investigative procedures, intellectual honesty, considerable time and energy, and a desire to have resolution to the issue. Schellenberg comments that such persons “will arrive at a parity belief only reluctantly and, therefore, only if careful attention to the matter seems to him to leave him with no other option.” 3 Pointing out that there are those who have inculpable nonbelief is not enough to generate an argument against God’s existence, however. Schellenberg inserts a second and key premise according to which if a personal loving God exists, then such a being would make His existence known to His creatures. Belief in His existence is a logically necessary condition for them to enter into an “explicit and positively meaningful”4 relationship. The intuition Schellenberg is capitalizing on is that in order to enter into a personal relationship with person P, I must at least believe in P’s existence. And a personal, loving God would want to enter into such a relationship; thus, belief is a logically necessary condition. The argument can be presented more succinctly as follows: (A) Reasonable nonbelief occurs. (B) If a good and loving God exists, then reasonable nonbelief would not occur. (C) Therefore, there is no good and loving God.
III. Weak responses and clarifications Some theists have been tempted to respond to the argument in ways typical of responses to the argument from evil. After all, a similar claim is being made, namely, there is some fact about the world that is inconsistent with the existence of a good and loving God. One of these responses appeals to the value of
Virtue, Hiddenness, and Belief
free will. In order to preserve the good of a personal relationship, God must arrange things such that persons enter into the relationship freely. But in order to enter into a relationship freely, the agent cannot be coerced. If God were to ostend Himself clearly enough so as to make inculpable doubt impossible, then, the claim is, persons would not enter into relationship with Him freely. Michael Murray, for instance says, My claim is that the hiddenness of God is required in order for free beings to be able to exercise their freedom in a morally significant manner. . . . If God revealed His existence in a more perspicuous fashion we would be in a situation very much like the one in the standard robbery case. . . .5 The response here is aimed to impugn premise (B). The idea is that God would not arrange things to make His existence clearly ostensible since making Himself known in such a way would compromise freedom. Others take issue with premise (A). Though this is not a popular attack, it is an intriguing one since many take (A) to be an undeniable empirical claim. Douglas Henry, however, aims to draw out the implications of the definition of inculpable doubt to show that (A) is implausible. In the definition of inculpable doubt, reference is made to the notion of adequate investigation. And adequate investigation is partially characterized as including exemplary investigative procedures. Henry observes, however, that, Reflective persons whose investigative procedures are exemplary do not usually assume that all the evidence is at hand. . . . They search for additional evidence, its quantity and quality depending upon the importance of the issue for which the evidence is being sought.6 Henry further observes that the very traits which ground our judgments of an adequate investigation (passion for truth, anxiety when in doubt, etc.) are incompatible with one who rests content with evidential parity. Henry concludes that Schellenberg is caught in a dilemma. Either, Schellenberg shows that there are persons who exude a high degree of intellectual virtue, thus making deception or self-deception unlikely. Or, Schellenberg accepts a lower standard of intellectual virtue. If the former, then the likelihood that such persons would accept evidential parity would be low for the passion for truth will lead one on past evidential impasses. If the latter, then the probability of deception or self-deception increases, or our intuitions suggest that their inquiry is not adequate. In either case, the claim that anyone satisfies inculpable doubt is unlikely. The response to both Murray’s and Henry’s arguments is to point out that theists themselves are counterexamples to their claims. Murray suggests that an ostensible display of God’s existence would curtail freedom. But theists themselves would claim that they know God exists, and this is so on basis of ample
evidence (either in the form of reflection on the natural order, or via religious experience). And yet, these very same theists would not say that their freedom is truncated because of their entrenched belief in God. Regarding Henry’s response, Schellenberg could say that not even theists exude the virtues associated with exemplary investigative procedures. God’s self-disclosure, he may say, is not apportioned to the intellectual virtues of the potential believer for if such virtues are not exuded by theists, nor should they be required for atheists. To remain consistent, Henry must say that the intellectual virtues are necessary not for everyone, but only when epistemic parity occurs for a specific agent. Schellenberg could respond, however, and say that if God exists, then epistemic parity would not occur in the first place.
IV. Stronger responses The God-human relationship and participation in goodness I have just surveyed one response to (B) and one to (A). Numerous objections to the argument abound, however. In this section I shall delineate two which highlight the key weaknesses in the hiddenness argument. The first weakness that stands out is that Schellenberg is assuming a particular conception of how God relates with created beings. In his argument for (B), Schellenberg offers a fairly detailed reflection on the nature of love and personal relationships. A representative quotation is the following: If I love you and so seek your well-being, I wish to make available to you all the resources at my disposal for the overcoming of difficulties in your life. But then I must also make it possible for you to draw on me personally—to let you benefit from my listening to your problems, from my encouragement, from my spending time together with you, and so on.7 On this view, if God is perfectly loving, He would make Himself available in a similar fashion. But making Himself available in order to make personal relationship possible, entails making belief in His existence clear enough so as to remove the possibility of inculpable doubt. If intellectually virtuous persons do not acquire belief in His existence, and belief is logically necessary for a meaningful relationship, then there is no good and loving God. I should note here that Schellenberg is only claiming that belief is a logically necessary condition to enter into a meaningful relationship. God is not, in Schellenberg’s view, forcing a relationship on someone, but rather is making His existence known so the believer may choose to enter into a more meaningful relationship. The free will of the believer is not violated in as much as the theist’s free will is not violated.
Virtue, Hiddenness, and Belief
The weakness occurs in that many theists do not believe that explicit belief in God is a necessary condition to enter into meaningful relationship with God. This should not be surprising in that Christian theology 8 indicates that the summum bonum of the Christian life is participation in the divine nature. Participating in the divine nature means that the more the “believer” grows in holiness (understood as growth in the virtues, primarily the virtue of charity) the more she participates in an aspect of God, namely goodness. Belief is not as important as the person being transformed into the likeness of God—however this end is reached. Whether this moral transformation and growth is mediated by belief matters little. Schellenberg considers such an objection endorsed by William Wainwright. To appreciate Schellenberg’s response, consider Wainwright’s own articulation of the view considered here, As Schellenberg says, “A personal relationship with God” appears to entail “belief in Divine existence. . . . For I cannot love God, be grateful to God, or contemplate God’s goodness unless I believe there is a God.” . . . But Schellenberg is mistaken. If I don’t believe that God exists, I can’t respond to God under that description. It doesn’t follow that I can’t respond to God.9 Schellenberg is unmoved by such a view of the divine-human relationship. The core of Schellenberg’s response is that, “Wainwright has substituted his own notion of relationship between God and human beings for mine here.” 10 This is an important response to take note of since it highlights the point at which theist and atheist diverge. Specifically, they diverge on intuitions of how a good and loving God would interact with His creation. I will return to this point below. For present purposes, further explication of Wainwright’s response is warranted. If I understand him correctly, there are alternative construals of the God-human relationship some of which entail that hiddenness is compatible with God’s existence. But saying there are alternative construals is not enough. Some argument is needed for why Wainwright’s alternative is more plausible than Schellenberg’s. One argument is that Schellenberg’s picture seems to entail that God is passible. Recall that for Schellenberg, love and meaningful relationship requires reciprocity, and thus requires explicit belief. Understood as such, it seems that God is passible in Schellenberg’s view. But traditional theology is unanimous in that God is impassible. Since atheism is parasitic on the conception of God being denied, the traditional theist can thank Schellenberg for showing us that a passible God does not exist. Even if Schellenberg can cash out his picture of the God-human relationship consistent with traditional theology, there is still the core point, according to which we have little intuitive support for the idea that God would desire explicit belief in order to relate to Him. The reason is that, to repeat, insofar as God is
goodness itself, anyone who exudes holiness (goodness) is participating in the nature of God. One can certainly expect that a good and loving God would desire that persons grow in holiness and charity, but to expect that God desires explicit belief is unsupported. Whether growth in charity is accomplished with or without explicit belief would seem to matter little to a good and loving God Who took participation in His nature to be the chief good. Schellenberg may respond by saying that this notion of the God-human relationship is not a personal relationship, and there is intuitive support that God being good and loving would desire a personal relationship. In response, the notion of a personal relationship for Schellenberg is tied closely to a human-tohuman model with emotive and psychological elements built into it. That is, being conscious of or having certain feelings (e.g., love) toward some person requires that there is some person there. Such a model suggests that there is someone who is the object of my affection and interaction. Traditionally understood, however, God is not an object that is located anywhere. True, Christians typically claim that they love God, but for an omnipresent being, this must be understood in radically different ways than love of a fellow human is understood.
The evidential force of testimony Another core weakness in Schellenberg’s argument is its commitment to an ego-centric notion of epistemic justification. Whereas the weaknesses highlighted in section A appealed to specific theological claims, the weaknesses outlined in this section are primarily epistemological and do not rest on assumptions buried in traditional theology. If the hiddenness argument is correct, then what is the epistemic status of theistic belief? How can the hiddenness argument avoid the conclusion that there is no reasonable belief, i.e., that theists have not based their belief on adequate investigation? I do not think Schellenberg can offer easy answers to these questions. Suppose S arrives at nonbelief after acquiring objective justification that epistemic parity obtains between G (God exists) and ~G. That is, the totality of evidence relevant to adjudicate the question of God’s existence indicates evidential parity. Since this is not what one would expect from God, God does not exist. If we assume that the notion of being reasonable involves this objective level of justification, then the argument from divine hiddenness would entail that there is no reasonable belief. If ‘reasonable’ is cashed out as objective justification, and the totality of evidence is as the proponent of the argument claims, then no theistic belief is reasonable. This seems patently false. I should say Schellenberg does not endorse delineating reasonable in terms of objective justification. I suspect he is unconcerned about the entailment I just outlined, but is sensitive to the heavy epistemic burden such a notion places
Virtue, Hiddenness, and Belief
on any proponent of the hiddenness argument. To keep the intuitive plausibility of premise (A), Schellenberg must opt for a weaker notion of justification. A more plausible approach is to argue that for a specific subject, her epistemic picture indicates that evidential parity obtains between G and ~G. But the more subjective Schellenberg makes his notion of being reasonable, the more reason we have for thinking that S’s epistemic picture does not reflect the totality of evidence, and therefore, we have little grounds for thinking that S satisfies intuitive criteria for adequate investigation. We can state the argument here in the form of a dilemma. If reasonable nonbelief is characterized subjectively, then Schellenberg owes us a reason for thinking that his notion of adequate investigation is satisfied. If reasonable nonbelief is characterized more objectively, that is, given the totality of evidence, God does not exist is more probable than not, then the theist suffers from unreasonable belief. More importantly, there are weaknesses with Schellenberg’s implicit egocentric notion of justification. Consider an intellectually virtuous theist. Being intellectually virtuous, the theist has submitted her belief to adequate investigation, and epistemic parity has not been obtained. Suppose now that our intellectually virtuous theist (call her Jane) reports to S that God exists. S knows Jane fairly well, and knows her to be very thorough in her cognitive endeavors, and very careful about the conclusions she draws from whatever evidence is presented to her. We may suppose further that Jane is psychologically well adjusted and we may even suppose that she grew up in a nonreligious home— thus circumventing any genetic or historical explanation for her theistic belief. How should S weigh or judge the importance of Jane’s testimony that God exists? There are three options for S (i) S could discount the testimony altogether, (ii) she could consider Jane’s testimony as a reason to suspend nonbelief and continue with her investigation, or finally, (iii) S could accept Jane’s report on trust. I suggest that Schellenberg is committed to holding (i) alone, since (ii) and (iii) are incompatible with the premise that reasonable nonbelief occurs, for in (ii) there is no belief and in (iii) there is no nonbelief. Additionally, I shall argue that (i) is incompatible with Schellenberg’s notion of adequate investigation. Suppose that S rejects Jane’s testimony. On what grounds could S reject it? Recall, we are assuming that Jane is an intellectually virtuous inquirer, and she evinces similar levels of intellectual care, fair-mindedness, and thoroughness on the issue of God’s existence as she does for other fields of inquiry. S has little reason to think that her epistemic standards in other areas are put aside when Jane directs her inquiry to investigate whether God exists. S is confronted with a reliable witness who has conducted her inquiry in virtuous ways. S has, as I see it, two grounds upon which to reject Jane’s testimony. First, S could argue that testimony on religious matters is inapposite. Second, S could generate
defeaters for the key grounds on which Jane bases her belief. More succinctly, S could reject the testimony head-on, or reject the character of the inquiry which generated the belief in the first place. Suppose S rejects the testimony head-on. I suspect that this strategy is most effective if the evidence upon which Jane bases her belief is not publicly available, such as religious experience. S could reject the testimony head-on even if Jane bases her belief on publicly available evidence, but the reason for rejecting it would likely generate defeaters to Jane’s key grounds, and this is the second strategy. So, suppose for purposes of clarity, that Jane bases her belief on religious experience (or a series of such) and provides testimony to S that God exists on the basis of these experiences. To comport with our intuitive grasp of adequate investigation, S would have to generate reasons to reject the veridicality of Jane’s experience. But notice, any argument or defeater to the veridicality of Jane’s experience would apply equally to S’s own if God were to manifest Himself in similarly ostensible ways. So, either S rejects the testimony with or without reason. If without reason, then S does not evince adequate investigation. If S rejects the testimony with reason, then those reasons must attack the veridicality of Jane’s religious experience. But any reason generated to discount Jane’s experience could be applied equally to one’s own. Given such epistemic commitments, God may not only be hidden to S, but must be so. The conclusion here is that preempting both (a) an entire class of testimonial reports and (b) an entire class of experiences (i.e., one’s own putative religious experiences) from carrying any evidential weight, does not comport with Schellenberg’s own notion of adequate investigation. Suppose S challenges the character of the inquiry itself. For example, assume that Jane bases her belief in God on the basis of the overall evidence of the natural world available to her. She has conducted her inquiry in virtuous ways, taking into account the seemingly intractable problem of evil. The point to make here is that S needs to ground her rejection of Jane’s testimony. But to do so, S needs to take issue with how Jane has conducted her inquiry, the inferences she drew from the evidence or some other such aspect of her inquiry. But on the assumption, we are assuming Jane has conducted her inquiry in virtuous ways. If the inquiry is genuinely virtuous, there is little room for S to challenge it. At least, the points at which a challenge can be generated would likely be points on which virtuous persons disagree. But if two virtuous agents who know each other find opposing sets of propositions likely (or unlikely), the proper epistemic course to take is (ii), not (i). The response to this line of reasoning may be that testimony from virtuous atheists should count as well. S may receive testimony from multiple sources, and we may suppose the testifiers conducted inquiry in virtuous ways. What should S do with conflicting reports from intellectually virtuous witnesses? Intuitions may differ, but I don’t see how (i) is an appropriate response to such conflicting testimony for it does not comport with Schellenberg’s own notion of
Virtue, Hiddenness, and Belief
adequate investigation. In the setting of conflicting reports from virtuous agents, the proper course is, again, to accept (ii). Another response may be that I have substituted my own view for Schellenberg’s view concerning what God would do epistemically for persons who seek Him out. Taking the same strategy as he takes in response to Wainwright above, Schellenberg can do the same here. Some reason should be given as to why God would desire some of us to believe through testimony. Laura Garcia comments approvingly of this approach saying, This method [believing through testimony] preserves the relationality of human beings better than the more direct, individualistic methods . . . and it is perhaps a better introduction into the realm of personal relationships that comprises our final destiny and that characterizes the inner life of God himself (according to the Christian doctrine of the Trinity). . . . A further benefit of this mode of transmission of the gospel message, for example, is that we must trust in the testimony of others in order to come to faith, and we are deeply indebted to those who brought us the message. . . .11 The idea here is that by ‘forcing’ us to rely on our fellow human beings and to trust their word, we are thereby linked with each other at a much deeper level than if God were to take an individualistic approach. The more trusting we are with others and the deeper ties we have to them—especially in order to trust them for such important issues—serves to cement the moral fabric God may be intending for the human community. Anecdotal evidence confirms this in that numerous persons convert due to a person or persons with whom she is familiar who displays incredible virtue and integrity. Trusting and becoming closer to such examples of holiness may likely be just what God intends. Is it unjust that God would disclose himself to a few, while letting others rely on those humans to bear virtuous testimony? Responding to this very objection Garcia notes, “There doesn’t seem to be a violation of justice involved, since as long as persons are not punished for involuntary unbelief, it’s hard to see that God wrongs anyone by failing to provide every person with the same evidential situation.” 12 I would only add that in some cases, the evidential situation favors the one who must rely on testimony. Recall my use of Tony Coady’s13 example of the ornithologists each giving an independent report of a bird sighting in a geographical region alien to the bird in question. They each make a report to the editor of a popular bird magazine. Because the sightings are so rare, the ornithologists in the field hold their perceptions in doubt. But the editor’s epistemic situation is different. He has access to numerous independent reports of the bird being in a specific geographical region previously known to be foreign for that bird. The same epistemic situation may confront the atheist who receives reports from intellectually virtuous witnesses that God exists. Such reports we may stipulate (to maintain the analogy) are based on some sort of
religious experience. The witnesses each may hold their interpretations with more reservation than the atheist who receives numerous reports with similar content and describing the object of the experience (i.e., God) in similar ways.14
V. Justification for agnosticism There is a significant limitation of the argument from testimony just given. As mentioned, there is testimony from those intellectually virtuous persons who do not believe in God. There are competing testimonies from equally virtuous persons. One may think that the testimonies cancel each other out in terms of supplying any epistemic direction. So, the epistemic situation of S would revert to what it was before considering the testimonies of others. Recall, however, that S’s original epistemic state was one of atheism. Therefore, in the setting of competing testimonies, S would remain justified in his atheism. In the previous section I argued that the appropriate response to testimony that God exists is (ii)—at least. Here we are presented with a reason for S to revert to atheism. I suggest that this is not the appropriate stance to take. Competing testimonies would justify taking a stance of further investigation while suspending belief. To see this consider the following example. Suppose Jane grows up in an academic family in which the parents are both intellectually sophisticated theists. They teach Jane the cosmological and ontological arguments for God’s existence. Jane happens to have solid reasons for believing in God. She understands the weaknesses of the arguments but also understands various replies to the objections and refinements to the arguments. Jane goes off to college and is introduced to the evidential argument from evil. Jane’s epistemic position is now radically challenged. Jane has at least two general options for her epistemic state: accept the force of the evidential argument and abandon her original beliefs, or pursue further inquiry while her religious beliefs are somewhat suspended. Let’s stipulate that the latter option involves familiarization with theistic responses and further refinements of the argument from evil, and the first option does not. It seems clear that Jane’s inquiry should take the course of the latter option. The example illustrates that in the setting of competing justification, the cognitively virtuous agent would pursue further inquiry. So it is with competing testimonies from virtuous persons.
VI. Justification for an impasse I have spent some time analyzing the hiddenness argument and motivating a position in which the virtuous agent for whom God is hidden should adopt an agnostic position, not atheism as argued by Schellenberg. Theists may be
Virtue, Hiddenness, and Belief
unmoved by the argument for many theists believe, as do I, that being good and loving does not entail being ostensibly manifest to the object of love. Love involves willing the other’s good, and the nature of the good is fulfillment of one’s nature and growth in virtue. Willing that either of these take place does not entail that the creature believes in God, or that God becomes ostensibly manifest to the potential believer. These reflections raise a problem for the virtue theorist. If both theist and agnostic are intellectually virtuous, then by the definition offered in the introduction, it seems they should both have knowledge. But they cannot both have knowledge for they believe in contrary propositions. Therefore, both cannot be intellectually virtuous contrary to the original assumption. One route by which one may stymie this reductio is to deny that they both perform intellectually virtuous acts in forming their respective beliefs. They certainly embody intellectually virtuous dispositions and traits of character. So our intuitions can be preserved and responsibilism remains unscathed. But there is another more interesting route by which to respond to the objection. What generates the reductio is the implicit assumption that certain religious claims are the proper object of human knowledge. Why is this assumption questionable? The hiddenness argument rests on the premise that a good and loving God would not allow reasonable nonbelief. Properly understood this proposition is a counterfactual of divine freedom. Assuming God is free, William Rowe’s objections aside,15 knowing what a free agent would do in a specified set of circumstances is not the proper object of knowledge. The intellectually virtuous agent would consider the domain of knowledge and measure the complexity of the domain with her epistemic position. That is, the intellectually virtuous agent would not only further her inquiry but also confine her claims to know relative to the inscrutable nature of the domain. In forming a belief that the domain is outside of one’s cognitive ken that belief itself may be a function of a virtuous act. If it is the result of one’s motivation to know (in this case, to know the limits of one’s inquiry) and the belief is true, then the agent performs a virtuous act in confining her inquiry. To clarify, by confining one’s inquiry I do not mean that the agent cut short her investigation within the domain. What is meant is that the virtuous agent will recognize the threshold point at which further inquiry exceeds the capacities of human cognition to deliver knowledge. Thus understood, the response to the reductio is that the counterfactuals of God’s freedom are not the proper objects of human knowledge. It is inapposite to say, then, that the theist and agnostic perform virtuous acts in forming the contrary beliefs they do. Properly understood, the concept of a virtuous act applies to objects that are within the domain of human knowledge. It seems clear that counterfactuals of God’s freedom are not within this domain. I am not saying that the counterfactuals of God’s freedom do not have any truth value. I am saying, rather, that there are true beliefs concerning the limitations of human cognition and these limitations themselves are the proper objects of human knowledge. If one exceeds
these limitations she is no longer performing virtuous acts, since the concept of a virtuous act defines human knowledge strictly. By definition, exceeding the limitations exceeds the proper domain of human knowledge. Is the response just given ad hoc? I think not, for in discovering the borders of human knowledge this judgment itself is arrived at by virtuous acts. Intellectual humility, care, and thoroughness contribute to one’s assessment of her cognitive position vis-à-vis the inscrutable nature of the domain. There is nothing ad hoc or arbitrary in stipulating one’s limits, but rather are themselves determined by intellectually virtuous acts. Is it intellectually virtuous to continue one’s inquiry past a point at which human knowledge is exceeded? The answer here is yes, since the boundaries of human knowledge are not fixed. They expand or are more clearly defined through further inquiry. As one furthers inquiry in a domain beyond human capacities (at the time of inquiry) one can still exude certain intellectual virtues. These virtues of inquiry would not be virtuous acts which serve as a definition for knowledge. Rather these virtues would serve as accounts for other epistemic states such as understanding or wisdom. The justification for the impasse concerning the hiddenness argument is that each side is conducting inquiry that is intuitively outside the boundaries of human insight. The claim to know, one way or the other, is inapposite. The virtue account offered here gives a reason for limiting claims to human knowledge because such an account governs one’s inquiry. Before concluding, I should clarify a tension here. My response to the hiddenness argument involves a counterfactual as well (i.e., a loving God would will the other’s good) just as Schellenberg’s defense of it does (i.e., a loving God would make his existence known to the beloved). Both propositions, I have just argued, are not the proper objects of human knowledge. But the theist may continue being a theist, for her theistic belief may not be based on a response to the hiddenness argument that would amount to knowledge. Knowing that God exists is quite different than knowing what God would do. How do other theories fare in this regard? To the extent that other epistemological theories make reference to cognitive processes (i.e., reliabilism and proper function accounts) or to justified beliefs (i.e., evidentialism) such theories are handicapped in defining the scope of one’s inquiry. Process-based theories do not have the resources to mark out the scope of human inquiry, for it is unclear what “basic psychological process” or faculty may recognize the limits of one’s inquiry in a potentially inscrutable domain. Recognition of one’s limitations is mediated by one’s intellectual virtues such as intellectual humility, thoroughness, and care. Since responsibilism includes reference to both intellectually virtuous traits, and virtuous acts, she is able to ground claims to limit inquiry.
In the concluding chapter of Linda Zagzebski’s Virtues of the Mind she comments on the role of cognitive psychology in developing a normative epistemology. She states, Psychology and cognitive science have made important contributions to epistemology in the last two decades, and we have every reason to be grateful for this and to expect it to continue. Meanwhile, accomplishments on the normative side of epistemology have lagged, perhaps overshadowed by the more spectacular results on the empirical side. My purpose in writing this book is to draw more attention to the side of epistemology that overlaps with ethics and, in particular, to show how one form of ethical theory—a pure virtue theory—can be developed in ways that are rich enough to permit the kinds of evaluations of epistemic states that are crucial to epistemology. I suggest that cognitive psychology is related to normative epistemology as moral psychology is related to ethics.1 The aim of my project has been to use empirical results in cognitive psychology to help out the normative side of epistemology. The empirical results recapitulated in this project aim to patch up developments on the normative side. The specifics of my argument concerned deflecting a strong objection to responsibilism and applying responsibilism to other areas of human knowledge staying faithful to the empirical evidence along the way. The overall purpose of this project was to show that responsibilism satisfies several important criteria for a theory of knowledge, criteria that are not all satisfied by other theories. Insofar as responsibilism satisfies many important criteria for a theory of knowledge and other theories currently on offer cannot satisfy all such criteria, we have a backhanded argument for the superiority of responsibilism. Recall from the introductory chapter that responsibilism provides us with (i) a Gettier-proof theory of knowledge, (ii) is able to account for the value of knowledge, and (iii) captures the intuitive idea that a person who knows is rational, justified, or intellectually good. Responsibilism satisfies criteria (i)–(iii) in the domains of knowledge addressed here.
I have taken some time obviating a potential defeater to responsibilism, namely, the unmotivated belief objection. I have shown that responsibilism delivers a result that comports with the empirical research on perceptual and memorial processing. Thus, we can say that responsibilism also satisfies a fourth criterion, namely (iv) it is immune to clear counterexamples (where these counterexamples are not of the Gettier-style). In addition to obviating a potent objection to responsibilism, the past three chapters illustrate that responsibilism satisfies two other important criteria for a theory of knowledge. They are (v) responsibilism is able to account for low-grade and (vi) higher-grade knowledge—higher-grade knowledge is knowledge which is the result, typically, of extended inquiry. It is certainly an argument in favor of a theory of knowledge that it can embrace the entire landscape of human knowledge. Moreover, it is certainly an argument for the superiority of a theory of knowledge that it is the only theory on offer that can do so. In fact, nonresponsibilist’s accounts of knowledge typically address themselves to perceptual, or memorial, or inferential knowledge, but are inapposite when it comes to account for philosophical knowledge, the very kind of knowledge the theories purport to give. As such, these theories are inadequate when applied self-referentially. So, there are two important implications of this project, the first of which is that it obviates a potent objection to responsibilism and second, it shows the superiority of responsibilism over other theories concerning higher-grade knowledge. If we look at the two results in toto, we have an argument for the superiority of responsibilism over other virtue theories. If it is the case that responsibilism is the best epistemology on offer, one may still ask what this means for epistemology in general. Has this project been an exercise in theoretical axe-grinding or are there important theoretical and practical results which fall out from responsibilism? If we assume that responsibilism is true, what follows from that assumption? There are several implications of responsibilism, some practical and some theoretical. One practical result is that it defines more clearly the epistemologist’s task. If acts of intellectual virtue are necessary for knowledge, the next question to ask is which virtuous acts are apropos for the epistemic circumstance in question. Research into how good scientists behave, for instance, helps us understand how we should conduct inquiry in that area. That is, there is a natural shift in the epistemologist’s task from defining what knowledge is to delineating how we can acquire it in various domains. What I am thinking of here is analogous to the natural movement in ethical discourse. If the aim of the moral life is exuding virtue, the natural question to ask is how one can develop the virtues. An answer may make reference to representative figures of the virtuous life. A Jane Austen novel or a biography on the life of a saint serves the purpose of ostending what various virtues look like and the appropriate contexts within which to exercise them. I recommend that the same study be
done on the intellectual side, namely, focus on what the intellectual virtues look like and how they are applied. A second practical result is that if knowledge is a function of one’s character, then education should include a moral component. And conversely, moral development requires the agent to form correct judgments which in turn requires performing intellectually virtuous acts. The importance of training in the moral and intellectual virtues is highlighted in the following passage from Richard Paul who is commenting on the place of virtue in education curricula: There is little to recommend schooling that does not foster what I call intellectual virtues. These virtues include intellectual empathy, intellectual perseverance, intellectual confidence in reason, and an intellectual sense of justice (fair-mindedness). Without these characteristics, intellectual development is circumscribed and distorted, a caricature of what it could and should be. These same characteristics are essential to moral judgment. The “goodhearted” person who lacks intellectual virtues will act morally only when morally grasping a situation or problem does not presuppose intellectual insight. Many, if not most, moral problems and situations in the modern world are open to multiple interpretations and, hence, do presuppose these intellectual virtues.2 The point here is twofold; first, knowledge (moral knowledge in this passage) requires instantiating certain intellectual virtues in order to be fully sensitive to the moral problem and potentially competing goods. Second, knowledge has a moral component to it. An agent must embody certain traits of character and motivations (e.g., motivations for fair-mindedness) in order to attain knowledge. If one accepts responsibilism, one must view with suspicion the idea that intellectual pursuits are not impregnated by one’s moral character. But what exactly is the connection here? In a section on the interdependence of the intellectual virtues, Paul offers us the following characterization of the intellectually virtuous person: To become aware of the limits of our knowledge we need the courage to face our own prejudices and ignorance. To discover our own prejudices in turn we must often empathize with and reason within points of view toward which we are hostile. To do this, we must typically persevere over a period of time, for learning to emphatically enter a point of view against which we are biased takes time and significant effort. That effort will not seem justified unless we have faith in reason to believe we will not be “tainted” or “taken in” by whatever is false or misleading in the opposing viewpoint. Furthermore, merely believing we can survive serious consideration of an “alien” point of view is not
enough to motivate most of us to consider them seriously. We must also be motivated by an intellectual sense of justice. We must recognize an intellectual responsibility to be fair to views we oppose. We must feel obliged to hear them in their strongest form to ensure that we do not condemn them out of our ignorance or bias.3 Given this picture of the intellectually virtuous person who is engaged in extended inquiry, it is clear that moral character cannot be separated from intellectual activity. One theoretical fall out of responsibilism is that it marks a shift away from analyzing epistemic concepts (e.g., knowledge) in terms of other epistemic concepts (e.g., justification) to analyzing epistemic concepts with reference to kinds of human activity. Let me explain. Epistemology is not concerned with knowledge simpliciter, but with human knowledge. And human knowledge is the result of activity as is clear from the preceding several chapters. Knowledge does not happen to an agent; the agent, rather, gets knowledge. A person is a candidate for knowledge only insofar as he is an agent (in the traditional sense of the term agens meaning “one who acts,” or “an actor”) and not merely a patient. Due to the centrality of agency in treating knowledge the centrality of activity likewise exists. Thus, the answer to the question ‘what cognitive state counts as knowledge?’ (a normative question) and ‘how does one acquire knowledge?’ look very similar. For the responsibilist, a cognitive state counts as knowledge iff the agent performs an act of intellectual virtue in forming that cognitive state. And, likewise, the agent procures such a praiseworthy state by performing an act of intellectual virtue. The move in focus here is subtle but crucial. Much of analytical epistemology centers on epistemic concepts, whereas, the responsibilist focuses on epistemic activity. This shift in focus allows the responsibilist to provide a satisfactory normative theory and at the same time answers crucial practical questions (i.e., how does one acquire knowledge?). One further theoretical gain for responsibilism is that it has a natural way of analyzing other higher-order epistemic states such as creativity,4 wisdom,5 and understanding. For instance, Deirdre Kramer thinks that wisdom is mediated by one’s emotional and motivational processing. Cognition and affect are interdependent and attain their meaning in the praxis, or activity, of the organism. Wisdom has often been seen as a form of intelligence, involving cognitive processes. However, such processes cannot be separated from affect which serves to motivate and sustain cognitive processing.6 One can see here that wisdom, as a cognitive state, can be analyzed quite naturally by the responsibilist insofar as the responsibilist defines knowledge (and
cognitive states like knowledge) with reference to one’s motivations and cognitive actions. Similar comments can be made for understanding and creativity. Both states involve not just intellectual incisiveness, but inquiry which is conducted over an extended period. With respect to creativity, “insight comprises selective encoding, selective combination, and selective comparison. Insightful thinking occurs when these processes are successfully applied in situations where the individual does not have a routine set of procedures for solving the problem.” 7 Here, creativity comes to the one who has persevered, and has phronesis regarding what information to select for future consideration. There are, I suppose, further conceptual problems which remain for the responsibilist. One problem, which I adumbrate here, concerns the success component of the responsibilist’s definition of a virtuous act. One may think that scientists in the past have acted with intellectual excellence and yet the truth eluded them. The intuition is that they performed intellectually virtuous acts and yet they do not get the truth i.e., they are not successful. So, here is a case in which it seems (i) an agent performs an act of intellectual virtue and (ii) is not successful. Is this a counterexample to responsibilism? I am inclined to think that it is not for two reasons. First, we must keep distinct, our intuitions concerning whether a person possesses a virtue from our intuitions concerning whether a person performs a virtuous act. And it is certainly possible that one possesses a virtue A to a high degree and yet does not perform the characteristic actions associated with A in every relevant circumstance. Thus, our scientists may intuitively possess a virtue to a high degree, but where they go wrong in interpreting the results of an experiment or formulating a theory may be a function of her not performing a virtuous act (uncharacteristically) on the occasion in question. Thus, one may certainly grant that there are many scientists of the past who possess a high degree of the relevant intellectual virtues and yet deny (i). Likewise, one may reject (ii). A historian of science with realist proclivities may contend that though our predecessors have been wrong at times, they have for the most part procured the truth.
Brand Blanshard, On Philosophical Style (Bloomington: Indiana University Press, 1954), pp. 21–22. 2 Ibid., 22. 3 I should note here that not all versions of virtue epistemology answer the “what is knowledge?” question with reference to a cognitive act. I aim to motivate responsibilism, and reserve motivating virtue reliabilism for the next chapter. 4 I should alert the reader that I do not think that all knowing states are belief states. For an excellent defense of this claim, see Kenneth Sayre, Belief and Knowledge, (Lanham, MD: Rowan & Littlefield, 1997), 117ff. 5 Edmund Gettier, “Is Justified True Belief Knowledge?” Analysis 23 (1963): 121–123. 6 Guy Axtell, “Introduction,” in Knowledge, Belief and Character: Readings in Virtue Epistemology, ed. Guy Axtell (Lanham, MD: Rowan and Littlefield Publishers, 2000), xv. 7 See Duncan Pritchard, “Virtue and Luck Revisited,” pp. 6 and 16 respectively. Retrieved from http://www.philosophy.ed.ac.uk/staff/documents/VEAndLuck Revisited3.pdf 8 Ted Warfield, “Knowledge from Falsehood,” Philosophical Perspectives vol. 19, Epistemology (Oxford: Blackwell Publishers, 2005), p. 406. 9 Alvin Plantinga, Warrant and Proper Function (New York: Oxford University Press, 1993), pp. 35–36. 10 Pritchard, “Virtue and Luck Revisited,” p. 3. 11 Linda Zagzebski has prompted formulation of these criteria—correspondence. 12 Michael Slote, Morals from Motives (New York, NY: Oxford University Press, 2001), p. 4. 13 For a defense of virtue theory regarding its ability to provide action guidance, see Roselind Hursthouse, On Virtue Ethics (New York: Oxford University Press, 1999), Chapter 1. 14 Slote, Morals from Motives, p. 4. 15 Linda Zagzebski, Divine Motivation Theory (New York: Cambridge University Press, 2004), p. 60. 16 Ibid., 61. 17 Ibid., 69. 18 Ibid., 72.
Linda Zagzebski, Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge (New York: Cambridge University Press, 1996), p. 131. 20 Ibid. 21 Ibid. 22 Christine Swanton, Virtue Ethics: A Pluralist Approach (New York: Oxford University Press, 2005), p. 19. 23 Robert Audi, Moral Knowledge and Ethical Character (New York: Oxford University Press, 1997), p. 180. 24 Zagzebski, Virtues of the Mind, p. 248. 25 Linda Zagzebski, “What is Knowledge?” in The Blackwell Companion to Epistemology, eds. John Greco and Ernest Sosa (Oxford: Blackwell Publishers, 1999), p. 102. 26 Ibid. 27 Zagzebski, “What is Knowledge?” 108. Zagzebski now calls an act of virtue, and act good in every respect. This change in terminology lends itself to less confusion given the many competing conceptions of a virtue that are on offer. 28 Zagzebski, Virtues of the Mind, p. 270. 29 Ibid., 271. 30 Ibid., 270. 31 Ibid., 311. 32 For more enlightening discussion on identifying virtuous traits, see Hursthouse, On Virtue Ethics, and Swanton’s, Virtue Ethics: A Pluralist Approach. 33 Some reliabilists will not be satisfied even with the success component, since it does not block cases of fleeting reliability. A virtuous act is just that, an act, and not a stable or reliable disposition. I address this concern in the following section. 34 John Greco, “Virtue, Luck, and the Pyrrhonian Problematic,” Philosophical Studies 130 (2006): 9–34, at 27–28. 35 See Greco, “Knowledge as Creditable True Belief,” in Intellectual Virtue: Perspectives from Ethics and Epistemology, Eds. Michael Depaul and Linda Zagzebski (New York: Oxford University Press, 2003). 36 Duncan Pritchard, “Virtue Epistemology and Epistemic Luck,” Metaphilosophy 34 (2003): 106–130. 37 See Plantinga, Warrant and Proper Function, 34ff. 38 For more on tragic dilemmas see Roselind Hursthouse, On Virtue Ethics (New York, NY: Oxford University Press, 1999), Chapter 3. 39 For an important challenge to Gettier-style counterexamples (all of them) see Stephen Hetherington, “Actually Knowing,” Philosophical Quarterly 48 (1998): 453–469. 40 For a discussion of some further complications on this basic line of thought, see Pritchard’s “Virtue Epistemology and Epistemic Luck.” 41 Pritchard, “Virtue Epistemology and Epistemic Luck,” p. 126. 42 John Greco, “Two Kinds of Intellectual Virtue.” Philosophy and Phenomenological Research 60 (2000): 181. 43 Zagzebski, “Responses,” Philosophy and Phenomenological Research 60 (2000): 209– 221, at 210. 44 Jason Baehr, “Character in Epistemology,” Philosophical Studies 128 (2006): 479– 514, at 486.
This move obviates another one of Baehr’s putative counterexamples. See, “Character in Epistemology,” 487.
Chapter 2 1
John Greco, “Virtue, Luck, and the Pyrrhoian Problematic,” Philosophical Studies 130 (2006): 9–34, at 9. 2 Jonathan Kvanvig, Intellectual Virtues and the Life of the Mind: On the Place of the Virtues in Epistemology (Lanham: Rowan & Littlefield, 1992), p. 19. 3 This version of the barn-facade case is taken from Alvin Plantinga, Warrant and Proper Function (New York: Oxford University Press, 1993), p. 33. 4 Alvin Goldman, “Discrimination and Perceptual Knowledge,” in Liaisons: Philosophy Meets the Cognitive and Social Sciences (Cambridge: MIT Press, 1992), p. 85. 5 Lawrence Bonjour, “Externalist Theories of Knowledge,” Midwest Studies in Philosophy 5 (1980): 62. 6 Alvin Goldman, Epistemology and Cognition (Cambridge: Harvard University Press, 1986), p. 77. 7 This is adapted from Goldman, EC, pp. 77–78. Goldman actually does not symbolize the rule with reference to a single cognizer as I have done, but it is clear that one must. Also, there is a distinction here between belief-dependent and beliefindependent transitions. The example given here is an example of a beliefdependent transition according to which there is a belief that operates as an input. Belief-independent transitions do not take any doxastic states as inputs, but rather take nondoxastic states as input, e.g., light energy hitting the photo-receptors on the retina. 8 Goldman EC, p. 63. Briefly, the difference between justification and knowledge for Goldman is that a piece of knowledge requires that the belief be true, whereas this is not the case for justification. Furthermore, local and global reliability are required for knowledge but not justifiedness. 9 Goldman EC, p. 106. 10 Goldman EC, p. 45. 11 Goldman EC, p. 46. 12 Alvin Goldman, “What is Justified Belief,” in Liaisons: Philosophy Meets the Cognitive and Social Sciences (Cambridge: MIT Press, 1992), p. 115. 13 See Goldman EC, p. 52. 14 It is unclear from Goldman’s treatment of these distinctions whether secondorder processes must be content-neutral like first-order processes. But I think the unclarity here can slide since what is important is that second-order processes are those processes (or cognitive states) which license the acceptance of a method. 15 Hilary Putnam, “Why Reason Can’t Be Naturalized,” in Realism and Reason (Cambridge: Cambridge University Press, 1983), p. 231. 16 Goldman, EC, p. 110. 17 Both quotations are from Alvin Plantinga, Warrant: The Current Debate (New York: Oxford University Press, 1993), p. 206. 18 Plantinga, Warrant: The Current Debate, p. 207.
Making this adjustment will not hurt the subsequent argument since the criteria for right methods are the same as the criteria for right processes except that methods have the added requirement that they are formed via reliable second-order processes. Consideration of the latter kinds of processes is not relevant to the subsequent argument. 20 Structurally the same things can be said to the rest of Plantinga’s examples on p. 206 of Warrant: The Current Debate. The strategy is to pick out a putatively reliable process (reliable because it produces mostly true beliefs) and then show that beliefs formed in accordance with such processes are intuitively unjustified. But as I have shown here, the processes used are not reliable and therefore the examples do not even get off the ground. 21 Goldman, “What is Justified Belief,” p. 115. 22 Ibid. 23 I should note that Plantinga allows for beliefs that pop into one’s head such as faith in Christ to count as warranted. See Alvin Plantinga, Warranted Christian Belief (Oxford: Oxford University Press, 2000). I should note, though, that the beliefs Plantinga counts as warranted and seem to pop into one’s head are not empirical beliefs—they are religious beliefs put there by the Holy Spirit. As (I) reads, it is only empirical beliefs that pop into one’s head that should be viewed with suspicion. 24 Zagzebski argues that Plantinga’s theory is not Gettier-proof and Goldman’s theory does not solve the value problem. Both authors, however, fall prey to both objections. (Zagzebski has, however, argued that Plantinga has trouble with the value problem in “From Reliabilism to Virtue Epistemology.” See citation note 28. below.) 25 Linda Zagzebski, Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge (New York: Cambridge University Press, 1996), 285–286. 26 Ibid., 286–287. 27 Ibid., 297 (emphasis mine). 28 Linda T. Zagzebski, “From Reliabilism to Virtue Epistemology,” in Knowledge, Belief and Character, ed. Guy Axtell (Lanham: Rowan and Littlefield Publishers, 2000), p. 113. 29 I owe this formulation of the value problem to Michael Depaul—in conversation. 30 Linda T. Zagzebski, “The Search for the Source of Epistemic Good.” Metaphilosophy 34 (2003): 12–28, at 15. One reason why one may prefer to see believing as more like acting is that agents don’t form just any belief. We form beliefs to which our attention is directed. And insofar as attention is an act of directing one’s cognitive abilities to a particular stimulus or stimulus array, it is more proper to see this as an act, and not the result of an automatic process. 31 Ibid., 16. 32 I remind the reader that I am excluding treatment of Greco’s theory. The problems outlined here are not patently problems for his version of responsibilism. I show, however, that they are not ultimately problems for the version of responsibilism I defend throughout this work. In any case, when I refer to responsibilism, I am referring to a Zagzebski-type approach, not Greco’s.
William Alston, “Virtue and Knowledge,” Philosophy and Phenomenological Research 60 (2000): 187. 34 Zagzebski, Virtues of the Mind, 279–280. Emphasis mine. 35 When trust is stripped of its motivational component, I think we are left with the fairly intuitive notion of simple assent or acceptance—with a slight twist. One should understand such acceptance functionally and not doxastically. Commenting on an issue similar to the one dealt with here, Keith Lehrer remarks, “The mental state of acceptance is a functional state, one that plays a role in thought, inference and action. We think, infer, and act in a way manifesting out trust in what we accept.” See his “A Critique of Externalism,” in A Theory of Knowledge, ed. Louis Pojman (Belmont: Wadsworth Inc., 1993), p. 319. The idea then of a baselevel acceptance is that an agent simply accepts the deliverances of her senses and this acceptance is functional, not necessarily an explicit doxastic state. 36 Correspondence. 37 Correspondence. 38 Abrol Fairweather, “Epistemic Motivation,” in Virtue Epistemology, eds. Abrol Fairweather and Linda T. Zagzebski (New York: Oxford University Press, 2000), p. 78. Fairweather quotes from Sosa’s Knowledge in Perspective: Selected Essays in Epistemology (Cambridge: Cambridge University Press, 1991), p. 240. 39 Fairweather, “Epistemic Motivation,” p. 78. 40 Ibid. 41 There may be many features of our environment which may be relevant to the formation of our beliefs. We may for instance, form beliefs out of wishful thinking or self-deception. Clearly such beliefs do not count as knowledge. Thus, alethic features of the environment are those features of the world or the agent herself which are related to the truth of a proposition. So, another way of thinking about (1) is that if the BMS were not cued to alethic features, the possibility of forming a true belief accidentally looms large. 42 Ernest Sosa, “Knowledge and Intellectual Virtue,” Monist 68 (1985): 242. 43 One should read both (5’) and (5’’) within the context of perceptual knowledge. So, for instance, (5’’) should read: It is not the case that both, S not possess a motivation for the truth (regarding perceptual beliefs) and S’s reason is cued to alethically relevant features of S’s environment.
Chapter 3 1
Ralph D. Ellis, “Implications of Inattentional Blindness for ‘Enactive’ Theories of Consciousness.” Brain and Mind 2 (2001): 299. 2 Arien Mack and Irving Rock, Inattentional Blindness (Cambridge: MIT Press, 1998), p. 25. 3 Ibid., 245. 4 Ibid., 246. 5 One could require that an agent is aware of X only if that agent can give a verbal report that she experienced X. But this won’t work since babies are certainly conscious of their environment, but cannot give verbal reports. Or suppose that there is an adult who for some reason cannot communicate with others. Certainly here,
too, we want to say that the agent can be conscious of her environment. One may incorporate these cases into the ‘verbal-report’ test by claiming that an agent is aware of X only if that agent could communicate that she saw X if her communication faculties were functioning sufficiently. But even here the boundaries can get fuzzy. For instance, an agent can be visually aware of X and yet not be able to report it simply because the information immediately decays in memory. 6 Mack and Rock, Inattentional Blindness, p. 227. 7 Ibid., 53–62. 8 See Markus Kiefer, “The N400 is Modulated by Unconscious Perceived Masked Words: Further Evidence for an Automatic Spreading Activation Account of N400 Priming Effects,” Cognitive and Brain Research 13 (2002): 27–39. 9 I should note here that there is empirical evidence for what is known as false positive responses. False positive responses in this experiment—which did occur—would be something like the following experimenter: “Did you see anything?”, subject: “ Yes.” And this response is given even though no stimulus was presented. A plausible reason for such responses is that the subject may be thinking, “Well, the experimenter would not ask me if anything were present if there were nothing present.” Other explanations in addition to this one are given by Mack and Rock, Inattentional Blindness, pp. 68–71. 10 It is not necessary for my argument to argue that attention is necessary in order for an agent to report having some kind of perception. What is required is somewhat weaker, namely that attention is required in order for an agent to have perception of a stimulus sufficient to be called knowledge. If performance is poor for some perceptual task, then we may say that even if the agent procures a true perceptual belief, it would not count as knowledge. The next step is to show that the culprit for such low performance lay in some property X (in this case attention). It follows then that X is necessary for perceptual knowledge. 11 I should mention that the three options here exhaust the relevant possible explanations for the location judgments. Logically, the first option is that the subjects’ accurate judgments were a function of attention, without conscious perception. But this is irrelevant for my argument insofar as I aim only to show that attention is a necessary condition for perceptual knowledge. The second option is that the subjects consciously perceived the location of the figure, but this was a function of attention. The third is that the subjects consciously perceived the location of the figure, but this was not a function of attention. The fourth is that the subjects neither consciously perceived the location nor were attentive to it. Suppose it is possible that one might have cases of perceptual knowledge without attention, but that when attention is directed to one location it blocks or interferes with perception at another. On the face of it, I do not see how this is a case of perceptual knowledge without attention. If one’s attention at location A blocks perception of stimuli in location B, then it is not the case that one has perceptual knowledge of stimuli in location B. 12 Mack and Rock, Inattentional Blindness, pp. 68–71. There was an experimental group that evinced a 50% IB effect, but the design in this case is somewhat different. The only point is IB effects were significantly greater even at fovea presentation. Also, what is interesting to note, though doubtful how it affects my argument, is that the subjects in the inattentive trial who did see something were correct at
both location and shape—pace the parafoveal trials in which shape was not at all accurately judged. 13 Ellis, “Implications,” p. 299. 14 Ralph D. Ellis, “Dynamic Systems as an Approach to Consciousness: Emotion, SelfOrganization, and the Mind-Body Problem.” New Ideas in Psychology 17 (1999): 244. 15 Further evidence that attention is a function of motivation can be found in Douglas Derryberry and Majorie Reed, “A Multidisciplinary Perspective on Attentional Control,” in Attraction, Distraction and Action: Multiple Perspectives on Attentional Capture, ed. Bradley Gibson (Oxford: Elsevier Science, 2001), pp. 313–348. See also, Douglas Derryberry and Don M. Tucker, “Motivating the Focus of Attention,” in The Heart’s Eye: Emotional Influences in Perception and Attention, eds. Paula Niedenthal and Shinobu Kitayama (San Diego: Academic Press, 1994), pp. 170–196. Also see, M. Bernstein et al. “The Effect of Motivation on the Stream of Consciousness: Generalizing from a Neurocomputational Model of Cingulo-Frontal Circuits Controlling Saccadic Eye Movements,” in The Caldron of Consciousness: Affect, Motivation and Self-Organization, eds. Ralph D. Ellis and Natika Newton (Amsterdam: John Benjamins, 2000), pp. 132–161. And finally, see Ellis (2005), (2001), and (1999). 16 Ellis, 2005, p. 52. 17 Frank Jackson, “What Mary Didn’t Know” Journal of Philosophy 83 (1986): 291–295. 18 Ellis, “Dynamic Systems,” p. 242. 19 Ibid., 246–247. 20 Jerome S. Bruner, “On Perceptual Readiness.” Psychological Review 64 (1957): 123–152. 21 Zenon Pylyshyn, “Is Vision Continuous with Cognition? The Case for Cognitive Impenetrability of Visual Perception.” Behavioral and Brain Sciences 22 (1999): 341–423, at 341–342. 22 Rock, Logic of perception, Palmer, vision science and Cite texts cited on p. 342 left. 23 Jerome S. Bruner, “On Perceptual Readiness,” p. 123. 24 Ibid., 124. 25 Ralph Ellis, Curious Emotions: Roots of Consciousness and Personality in Motivated Action (Philadelphia: John Benjamins Publishers, 2005), p. 70. 26 I give the exceptions below. 27 See Stephen Porter and Angela R. Birt, “Is Traumatic Memory Special? A Comparison of Traumatic Memory Characteristics with Memory for Other Emotional Life Experiences,” Applied Cognitive Psychology 15 (2001): 101–117. And also, SvenAke Christianson and Elizabeth Engelberg, “Organization of Emotional Memories,” in Handbook of Cognition and Emotion, eds. Tim Dalgleish and Mick J. Power (New York: Wiley & Sons, 1999), pp. 211–228. 28 I owe this point to Bradley Gibson. 29 Tanya L. Chartrand and John A. Bargh, “Automatic Activation of Impression Formation and Memorization Goals: Nonconscious Goal Priming Reproduces Effects of Explicit Task Instructions,” Journal of Personality and Social Psychology 71 (1996): 464–478. See also, Tanya L. Chartrand and John A. Bargh, “Nonconscious Motivations: Their Activation, Operation, and Consequences,” in Self and Motivation:
Emerging Psychological Perspectives, eds. Abraham Tesser, Diederik A. Stapel, and Joanne V. Wood (Washington, D.C.: American Psychological Association, 2002), p. 18. 30 Tanya L. Chartrand and John A. Bargh, “Nonconscious Motivations: Their Activation, Operation, and Consequences,” in Self and Motivation: Emerging Psychological Perspectives, eds. Abraham Tesser, Diederik A. Stapel, and Joanne V. Wood (Washington, D.C.: American Psychological Association, 2002), p. 18. 31 Ibid., 17. 32 Daniel Kahneman states, “Novelty and incongruity are defined by a mismatch between stimulation and a neuronal model of expectations.” Daniel Kahneman, Attention and Effort (Englewood Cliffs, NJ: Prentice-Hall, 1973), p. 53. 33 Daniel Kahneman, Attention and Effort (Englewood Cliffs, NJ: Prentice-Hall, 1973), pp. 53–54. Emphasis mine. See also, Daniel E. Berlyne, “Curiosity and Exploration,” Science 153 (1966): 25–33. Hereafter I shall cite Kahneman with author and pagination. 34 Kahneman, p. 54. 35 Ibid. Emphasis mine. 36 Ralph Ellis, Curious Emotions: Roots of Consciousness and Personality in Motivated Action (Philadelphia: John Benjamins Publishers, 2005), p. 76. See also J. Panskepp, Affective Neuroscience (New York, NY: Oxford University Press, 1998), and J. L. Gallant et al. “Neural Activity in Areas V1, V2, and V4 During Free Viewing of Natural Scenes Compared to Controlled Viewing.” NeuroReports (1998): 9. 37 Alan Cowey, “The 30th Sir Frederick Bartlett Lecture: Fact, Artefact, and Myth about Blindsight,” The Quarterly Journal of Experimental Psychology 57 A (4) (2004): 577–609, at 577. 38 Ibid., 585. 39 Ibid., 587. 40 Ibid., 588. 41 R. W. Kentridge, C. A. Heywood, and L. Weiskrantz, “Attention without Awareness in Blindsight,” Proceedings of the Royal Society London B (266) (1999): 1805–1811, 1805. 42 I relegate to a footnote one last objection because I find it completely unpersuasive though it has been mentioned by some responding to this chapter. The objection is that I am guilty of “over-intellectualizing” the mind. One anonymous reader commented that s/he does not know why one would endorse an epistemology that would preclude, say, dogs from having knowledge. Two comments are warranted. First, I am unclear what the claim of over-intellectualization amounts to. I suspect that it amounts to a rigid commitment to a purely reactive view of perceptual cognition. If so, the view conflicts with the empirical evidence recapitulated in this chapter and is thus, patently unpersuasive. Second, if the test for a good epistemology is that it accounts for why dogs, cats, or ants have knowledge, then several popular theories of knowledge are categorically excluded from consideration in addition to responsibilism, such as evidentialism. Few ants or dogs, I suspect, apportion their beliefs to the evidence given to them. In fact, any epistemology that countenances the idea that knowledge requires having a belief is suspect since it is not obvious that dogs or cats form beliefs—certainly not the sophisticated doxastic states philosophers banter about.
Chapter 4 1
Alvin Plantinga, Warrant and Proper Function (New York: Oxford University Press, 1993), p. 61. 2 Plantinga, Warrant and Proper Function, p. 64. 3 Alvin Goldman, Epistemology and Cognition (Cambridge: Harvard University Press, 1986), p. 199. 4 Stephen Palmer, Vision Science: From Photons to Phenomenology (Cambridge, MA: MIT Press, 1999), pp. 573–577. 5 Emanuel Averbach, and Abner S. Coriell, “Short-Term Memory in Vision,” Bell System Technology Journal 40 (1961): 309–328. See also Palmer, Vision Science, p. 576. 6 S. Hashtroudi, M. Johnson, and S. Lindsay, “Source Monitoring,” Psychological Bulletin 114 (1993): 3–28. Also see M. Johnson and C. Raye, “Reality Monitoring,” Psychological Review 88 (1981): 67–85. 7 Richard M. Shiffrin, “Modeling Memory and Perception,” Cognitive Science 27 (2003): 341–378. 8 Given what I have said so far in this paragraph it may look like the argument that motivations are involved in memory is quite secure. For the retrieval process looks rather intentional, and intuitively intentional acts are informed by a motivation. This conclusion cannot be drawn so fast, however, since the matching process alluded to may be part of a parallel process. That is, information expected to enter into the system is already retrieved, and the matching takes place without any intentional act. For example, when I walk into my office, I expect to see certain things, a desk, table, a computer etc. Thus, in a parallel system, such information is already retrieved from storage. (There are numerous studies illustrating how expectations can prevent one from perceiving rather significant aberrations from an expected depiction of a scene.) 9 Scott C. Brown and Fergus I. M. Craik, “Encoding and Retrieval of Information,” in The Oxford Handbook of Memory, eds. Endel Tulving and Fergus I. M. Craik (London: Oxford University Press, 2000), 93. 10 Eich mentions that “. . . as a general rule, unattended stimuli are not recallable” (p. 105). E. Eich, “Memory for Unattended Events: Remembering with and without Awareness,” Memory and Cognition 12 (1984): 105–111. 11 Ibid. 12 F. I. M. Craik et al. “The Effects of Divided Attention on Encoding and Retrieval Processes in Human Memory,” Journal of Experimental Psychology: General 125 (1996): 159–180. 13 There were two middle groups, one performed the dual task condition at encoding only, and the other performed the dual task condition at retrieval only. 14 What explains the fact that the DA subjects did recall some words—whereas the subjects in Eich’s experiment did not recall any of the nonattended information? The task for Craik’s experiment did allow for attention to the target stimuli, i.e., the words. But this attention had to oscillate between the two tasks, whereas there was likely no allowance for oscillation in Eich’s experiments. 15 Craik, F. I. M., “The Effects of Divided Attention,” p. 174. See also W. A. Johnston et al. “Divided Attention: A Vehicle for Monitoring Memory Processes,” Journal of
Experimental Psychology 83 (1970): 164–171 and A. D. Baddeley et al. “Attention and Retrieval from Long-Term Memory,” Journal of Experimental Psychology: General 13 (1984): 518–540. 16 Craik, “The Effects of Divided Attention,” p. 174. Emphasis mine. 17 Ibid., 164. 18 Douglas Derryberry and Don M. Tucker, “Motivating the Focus of Attention,” in The Hearts Eye: Emotional Influences in Perception and Attention, eds. Paula Niedenthal and Shinobu Kitayama (San Diego, CA: Academic Press, 1994), p. 168. 19 See R. A. Wise, “Sensorimotor Modulation and the Variable Action Pattern (VAP): Toward a Noncircular Definition of Drive and Motivation,” Psychobiology 15 (1987): 7–20. 20 Douglas Derryberry and Majorie Reed, “A Multidisciplinary Perspective on Attentional Control,” in Attraction, Distraction and Action: Multiple Perspectives on Attentional Capture, ed. Bradley Gibson (Oxford: Elsevier Science, 2001), p. 329. Emphasis mine. Also see W. J. H. Nauta, “The Problem of the Frontal Lobe: A Reinterpretation,” Journal of Psychiatric Research 8 (1971): 167–187, for endorsement of the claim in italics. 21 See Stephen Porter and Angela R. Birt, “Is Traumatic Memory Special? A Comparison of Traumatic Memory Characteristics with Memory for Other Emotional Life Experiences,” Applied Cognitive Psychology 15 (2001): 101–117. And also, SvenAke Christianson and Elizabeth Engelberg, “Organization of Emotional Memories,” in Handbook of Cognition and Emotion, eds. Tim Dalgleish and Mick J. Power (New York: Wiley & Sons, 1999), pp. 211–228. 22 I owe this objection to Kent Staley. 23 Daniel J. Simons and Daniel T. Levin, “Change Blindness,” Trends in Cognitive Science 1 (1997): 264. 24 R. A. Rensink, J. K. O’Reagan, and J. J. Clark, “To See or Not to See: The Need for Attention to Perceive Changes in Scenes,” Psychological Science 8 (1997): 372. 25 Ibid., 372. Emphasis mine. 26 Both quotations from Simons and Levin, “Change Blindness,” p. 263. Emphasis mine. 27 Simons and Levin, “Change Blindness,” p. 263. 28 Ibid. 29 Ibid., 266. An interesting addition to the results here is that the subjects who did notice the change were of the same age as the experimenters; the subjects who did not notice the change were older than the experimenters. 30 Ibid. 31 By free viewing I mean the subject is given no task to perform when she is viewing the scene. 32 Daniel Kahneman, Attention and Effort (Englewood Cliffs: Prentice-Hall, 1973), p. 51. Emphasis mine. 33 Tanya L. Chartrand and John A. Bargh, “Nonconscious Motivations: Their Activation, Operation, and Consequences,” in Self and Motivation: Emerging Psychological Perspectives, eds. Abraham Tesser, Diederik A. Stapel, and Joanne V. Wood (Washington, DC: American Psychological Association, 2002), p. 18. 34 Daniel L. Schacter and Chad S. Dodson, “Misattribution, False Recognition and the Sins of Memory,” in Episodic Memory: New Directions in Research (New York:
Oxford University Press, 2002), p. 72. The other four sins are transience, absentmindedness, blocking, and persistence. Transience involves decreasing accessibility of information over time. Absent-mindedness involves “. . . shallow processing that contributes to weak memories of ongoing events . . .” (p. 72) and persistence refers to intrusive memories that we wish we could forget but don’t. 35 M. Ross et al. “The Effect of Attitude on the Recall of Personal Histories,” Journal of Personality and Social Psychology 40 (1981): 627–634. 36 R. J. Spiro, “Remembering Information from Text: The ‘State of Schema’ Approach,” in Schooling and the Acquisition of Knowledge, eds. R. C. Anderson et al. (Hillsdale: Erlbaum, 1977), 137–165. 37 Lawrence W. Way et al., “Cause and Prevention of Laparoscopic Bile Duct Injuries: Analysis of 252 Cases from a Human Factors and Cognitive Psychology Perspective,” Annals of Surgery 237 (2002): 460–469. 38 Ibid., 460. 39 Ibid., 464. 40 Ziva Kunda, Social Cognition: Making Sense of People (Cambridge, MA: MIT Press, 1999), Chapters 3 and 5. 41 Ezra Stotland and Lance Canon, Social Psychology: A Cognitive Approach (Philadelphia, PA: Saunders Publishing, 1972). 42 Kent Staley has pointed out to me that responsibilism cannot claim to be superior over reliabilism insofar as reliabilism can countenance the role of motivations as well. If a process cum motivation is reliable then the beliefs it produces count as knowledge, and if not reliable then the beliefs do not count as knowledge. Wherein lies the inferiority? My objection to reliabilism is not that it cannot countenance the role of motivation in cognition, but that it cannot account for the value of knowledge. In addition, reliabilism suffers from counterexamples, examples for which the responsibilist does not fall prey. I advert the reader to the first chapter for treatment of the value problem, and to the fourth chapter for treatment of counterexamples for which the reliabilist falls prey. Even so, though the reliabilist can accept the role of motivations in reliable cognition, it may not. It is not committed either way to the epistemic status of motivation. As it stands, then, reliabilism looks rather amorphous—it can countenance the presence of many different putative criteria for knowledge. And responsibilism takes one step further and specifies just what kind of cognition is knowledge producing i.e., cognition informed by an epistemic motivation. Now, let’s assume that everything is equal between these two theories except that one specifies the conditions for knowledge whereas the other gives us merely a programmatic criterion for determining the conditions for knowledge. It seems clear that the theory which is more specific and can explain just as much should be considered the superior theory. And responsibilism does just this vis-à-vis reliabilism.
Chapter 5 1
I should mention that my thesis here may not be at odds with either Coady or Fricker insofar as they do advert us to important criteria for an adequate account of testimonial knowledge. In relations to Coady and Fricker, then, the responsibilist offers a fuller account of testimonial knowledge.
See Edward J. Imwinkelried, “The Next Step after Daubert: Developing a Similarly Epistemological Approach to Ensuring the Reliability of Nonscientific Testimony.” Cardozo Law Review 15 (1994): 2271–2294. Also, Bethany Spielman and George Agich, “The Future of Bioethics Testimony: Guidelines for Determining Qualifications, Reliability, and Helpfulness.” San Diego Law Review 36 (1999): 1043–1059. 3 See Robert Audi, “The Place of Testimony in the Fabric of Knowledge and Justification.” American Philosophical Quarterly 34 (1997): 405–422. C. A. J. Coady, Testimony: A Philosophical Study (New York: Oxford University Press, 1992). Tyler Burge, “Content Preservation.” Philosophical Review 102 (1993): 457–488. 4 See Elizabeth Fricker, “The Epistemology of Testimony.” Proceedings of the Aristotelian Society, Supp. 61 (1987): 57–83. Elizabeth Fricker, “Against Gullibility,” in Knowing From Words. Eds. Bimal Krishna Matilal and Arindam Chakrabarti (Dordrecht: Kluwer, 1994), pp. 125–161. Elizabeth Fricker, “Telling and Trusting: Reductionism and Anti-Reductionism in Epistemology and Testimony,” Mind 104 (1995): 393–411. 5 See Jonathan Adler, “Transmitting Knowledge.” Nous 30 (1996): 99–111. Robert Audi, “The Place of Testimony in the Fabric of Knowledge and Justification.” American Philosophical Quarterly 34 (1997): 405–422. Michael Welbourne, “Testimony, Knowledge and Belief.” in Knowing From Words. Eds. Bimal Krishna Matilal and Arindam Chakrabarti (Dordrecht: Kluwer, 1994), 297–313. Alvin Plantinga, Warrant and Proper Function (New York: Oxford University Press, 1993), 77–88. 6 Peter Graham, “Transferring Knowledge,” Nous 34 (2000): 131–152. Fred Dretske, “A Cognitive Cul-de-sac,” Mind 91 (1982): 109–111. Jennifer Lackey, “Testimonial Knowledge and Transmission,” The Philosophical Quarterly 49 (1999): 471–490. 7 See Coady, Testimony. See also, Tyler Burge, “Content Preservation,” and Alvin Plantinga, Warrant and Proper Function, pp. 77–88. 8 Coady offers a reductio argument against a reductivist. This is an important argument, but is not an argument for his externalist view. I therefore pretermit treatment of it. 9 Coady, Testimony, p. 143. 10 Ibid. 11 Ibid., 145. 12 Ibid. 13 Ibid., 196. Emphasis mine. 14 Fricker, “Against Gullibility,” p. 147. 15 Fricker admits that there are domains of testimony that do not require one to check up on the trustworthyness of the speaker. Such domains are determined by common sense. 16 Ibid., 148. 17 Ibid., 149. 18 Ibid. 19 Ibid., 136. 20 Ibid., 133. Emphasis mine. 21 Fricker, “Telling and Trusting,” p. 402. 22 Fricker, “Against Gullibility,” p. 133. See also “Telling and Trusting,” p. 404. 23 Fricker, “Against Gullibility,” p. 150. 24 Ibid.
Ibid., 155. Ibid., 151. 27 Ibid., 154. Emphasis mine. 28 I give such an argument below. 29 James Beilby, “Review of Coady,” C. A. J. Testimony: A Philosophical Study. Oxford: Oxford University Press, 1992. in Philosophia Christi. 1 (1999): 145. 30 A similarity not mentioned in the text but worth noting is that both Coady and Fricker are drawing on Donald Davidson’s work. Imagine being deported to a foreign island and given the task of interpreting the language of the natives. One would attempt to make sense of their utterances against the background of his own beliefs. But also, this interpreter would assume that the beliefs of these natives are true, and he is there to determine what their speech means. Coady emphasizes this latter element in Davidson, and Fricker emphasizes the former. On the one hand, we need to assume the truth of the beliefs of the natives, but on the other hand, the interpretive task requires measuring their utterances against our own beliefs. 31 See the quotation above concerning the telephone company and the quotation from the chapter on “Astonishing Reports.” 32 Coady, Testimony, p. 47. 33 Fricker, “Telling and Trusting” p. 406. 34 Plantinga, Warrant and Proper Function, p. 84. 35 This condition is adapted from Peter Graham, “Transferring Knowledge,” 131. For endorsement of this condition see Welbourne, Burge, and Plantinga cited in notes 3 and 5 above. 36 Jennifer Lackey, “Testimonial Knowledge and Transmission,” The Philosophical Quarterly 49 (1999): 477. 37 This condition is adapted from Peter Graham, “Transferring Knowledge,” p. 131. 38 Graham, “Transferring Knowledge,” p. 134. 39 Alvin Goldman, “Discrimination and Perceptual Knowledge” in Liaisons: Philosophy Meets the Cognitive and Social Sciences (Cambridge, MA: MIT Press, 1992), pp. 85–103. 40 Graham, “Transferring Knowledge,” p. 135. 41 Fricker, “Epistemology of Testimony,” p. 62. Emphasis mine. 42 I assume that the condition that a receiver understand the testifier’s utterance is subsumed under taking up the communicated proposition virtuously. It would not be a virtuous act if one were to believe p even though she did not understand it. 43 I should note an important aside. The reason why what gets adopted must be a function of one performing a virtuous act is because an account of testimonial knowledge must explain why the truth gets transferred nonaccidentally. Theories that require the testifier to know p and sincerely communicate p is one way to ensure the truth is obtained and transferred. RTK provides such an explanation, however, in a way that does not fall prey to the case involving Mrs. Smith. Performing a virtuous act entails getting the truth in the case of adopting p and sharing it in the case of communicating p. 44 Lackey, “Testimonial Knowledge and Transmission,” 484 ff. 45 Some may think that the truly virtuous knower would never succumb to Cartesian worries. Lackey’s example can stipulate that a person gets caught in the grip of 26
Cartesian worries at a particular point in time, at which the testifier is asked for directions. The virtuous testifier will eventually come out of her skepticism, but the example may focus on the short phase in which the testifier is skeptical. 46 Graham, “Transferring Knowledge,” p. 131. 47 Kent Staley has suggested this example to me. 48 Lackey, “Testimonial Knowledge and Transmission,” p. 481. 49 A parallel example is given by Plantinga, Warrant and Proper Function, p. 83. 50 It is not a requirement of the responsibilist account that an informant have secondorder knowledge that what she is sharing is true. In this way, teachers or newscasters can get the truth because of their virtues in adopting the truth, and yet never know that what they share is in fact true. 51 I owe this objection to Linda Zagzebski. 52 See Richard Cowan and Daniel P. J. Carney, “Calendrical Savants: Exceptionality and Practice,” Cognition 100 (2006): B1–B9. For endorsement of the view that savant skills are a function of long-term memory see, Ted Nettelbeck and Robyn Young, “Intelligence and Savant Syndrome: Is the Whole Greater than the Sum of the Fragments?” Intelligence 22 (1996): 49–68. 53 An intellectually virtuous act A, is in part an act “. . . a person with virtue A would (probably) do in the circumstances . . .” Zagzebski, Virtues of the Mind, p. 270. 54 See Robert Nozick, Philosophical Explanations (Cambridge, MA: Harvard University Press, 1981). See also, Goldman, Dretske and Graham cited above. 55 Alvin Goldman, “Discrimination and Perceptual Knowledge,” The Journal of Philosophy 73 (1976): 771–791. 56 Alvin Goldman, Epistemology and Cognition (Cambridge: Harvard University Press, 1986), p. 46. 57 Brian Skyrms, “The Explication of ‘X knows that p’,” The Journal of Philosophy 64 (1967): 373–389. 58 Goldman, Epistemology and Cognition, pp. 46–47. 59 Linda Zagzebski, “Must Knowers be Agents” in Virtue Epistemology. Eds. Linda Zagzebski and Abrol Fairweather (Oxford: Oxford University Press, 2002), p. 148. 60 Harry Frankfurt, “Alternative Possibilities and Moral Responsibility,” The Journal of Philosophy 66 (1969): 829–839. 61 Zagzebski, Virtues of the Mind, p. 158. 62 Plantinga, Warrant and Proper Function, p. 87. 63 Aquinas, ST. II-II q. 4 art. 1. Quoted from Coady’s Testimony, p. 17. 64 Coady, Testimony, p. 213.
Chapter 6 1
American Society for Bioethics and Humanities, Core Competencies for Health Care Ethics Consultation, 1998. 2 Bruce D. Weinstein, “What is an Expert?,” Theoretical Medicine 14 (1993): 67. 3 Ibid., 68. 4 Jan Crosthwaite, “Moral Expertise: A Problem in the Professional Ethics of Professional Ethicists,” Bioethics 9 (1995): 361–379. 5 Scott Yoder, “Experts in Ethics? The Nature of Ethical Expertise,” Hastings Center Report 28 (1998): 13.
A notable exception is Lisa Parker, “Ethical Expertise, Maternal Thinking, and the Work of Clinical Ethicists,” Ethics Expertise: History, Contemporary Perspectives and Applications. Ed. Lisa Rasmussen (Amsterdam: Springer Publishers, 2005), pp. 165– 207. 7 W. G. Chase and Herbert A. Simon, “In the Mind’s Eye,” in Visual Information Processing. Ed. W. G. Chase (New York: Academic Press, 1973), pp. 215–281. 8 Weinstein, 68, inside quotations taken from D. J. O’ Connor and B. Carr, Introduction to the Theory Of Knowledge (Minneapolis: University of Minnesota Press, 1982), p. 64. 9 Weinstein, p. 69. 10 I will abbey treatment of adequacy, in particular, I will not try to lay out standards for determining adequacy. The main reason for this is that epistemic circumstances change, and with it, so does the level of justification we think constitutes a good epistemic state. 11 Judith Jarvis Thomson, “A Defense of Abortion,” in The Problem of Abortion. Ed. Joel Feinberg (Belmont: Wadsworth, 1984), pp. 173–187. 12 This and the next example are taken from Alvin Plantinga, Warranted Christian Belief (New York: Oxford University Press, 2000), p. 359. 13 See Richard Foley, “Justified Inconsistent Beliefs,” American Philosophical Quarterly 16 (1979): 247–257. 14 One may think of intellectual care as capturing the idea that experts have good predictive ability. Experts are reliable at determining what future inquiry would look like and being intellectually careful is meant to capture this idea. 15 Some comment is in order regarding what exactly alethically orientated routes are. The reason we think dogmatism is bad is because, partly, it does not lead one to form many true beliefs, and minimizing false ones. But even if the dogmatic person ends up with true beliefs, we still do not admire her for her motivational states are not good. Since I desire to avoid commitment to realism, by ‘alethically orientated routes’ I mean to say that the inquiry makes probable the acquisition of good beliefs over bad ones which requires good motivational states. We admire, for instance, motivations that open one to more information and thorough inquiry. They are admired not only because we may think they lead to true beliefs, but because they are intrinsically worthwhile. But even here, my comments must be qualified. Not all intellectual virtues are such because they lead to an abundance of true beliefs over false ones. The virtue of creativity may lead to many false ones before the inquirer hits on the truth. What makes creativity a virtue is that it expands our knowledge base, if only at the expense of going through many false beliefs. See Linda Zagzebski’s Virtues of the Mind: An Inquiry into the Ethical Foundations of Knowledge (New York: Cambridge University Press, 1996), p. 182. 16 Jonathan Haidt, “The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgement,” Psychological Review 108 (2001): 814–834. 17 Judith A. Effken, “Informational Basis for Expert Intuition,” Journal of Advanced Nursing 34 (2) (2001): 246–255, at 247. Effken quotes from S. K. Hanneman, “Advancing Nursing Practice with a Unit-Based Clinical Expert,” IMAGE: Journal of Nursing Scholarship 28 (1996): 333. 18 For more on the reliability of intuition in the absence of evidence or evidence to the contrary, see the examples given by Gary Klein in The Power of Intuition: How to
Use Your Gut Feelings to Make Better Decisions at Work (New York: Currency Books, 2003), Chapter 2. 19 Neither Yoder nor Weinstein seems to countenance the importance of explanation in ethical discourse. 20 Richard L. Franklin, The Search for Understanding (New York: Peter Lang Publishers, 1995), Chapter 3. 21 Michael Depaul, Balance and Refinement: Beyond Coherence Methods of Moral Inquiry (London: Routledge, 1993). 22 Mark Timmons, “Review of Balance and Refinement: Beyond Coherence Methods of Moral Inquiry” in Ethics, 105 (3) (1995): 651–652. 23 Depaul, Balance and Refinement, p. 149. 24 R. M. Hare, Moral Thinking, p. 12. Quoted from Mark Nelson, “Morally Serious Critics of Moral Intuitions,” Ratio 12 (1999): 69. 25 Depaul, Balance and Refinement, 204–205. 26 See C. A. J. Coady, Testimony: A Philosophical Study (New York: Oxford University Press, 1992), p. 213. 27 See Linda Zagzebski, “Recovering Understanding,” in Knowledge, Truth, and Duty: Essays on Epistemic Justification, Responsibility, and Virtue. Ed. Matthias Steup (New York: Oxford University Press, 2001), pp. 232–256. 28 Depaul, Balance and Refinement, p. 191. Emphasis mine. To illustrate this point, see Depaul’s use of Dostoyevsky’s Crime and Punishment, pp. 142–145 of Balance and Refinement. 29 Formative experience here refers to what an experience does to the agent, not to an experience considered in the abstract. Watching We Were Soldiers may impact one person and leave another without any change. Only the first person had a formative experience. 30 Eleonore Stump, Aquinas: Arguments of the Philosophers (London: Routledge Publishers, 2002). 31 Ziva Kunda offers a review of this literature in Social Cognition: Making Sense of People (Cambridge: MIT Press, 1999), Chapters 6 and 8. 32 See Norbert L. Kerr, Geoffrey P. Kramer, and Robert J. MacCoun, “Bias in Judgment: Comparing Individuals and Groups,” Psychological Review 103 (4) (1996): 687–719, and, R. Nisbett and L. Ross, Human Inference: Strategies and Shortcomings of Social Judgment (Englewood Cliffs: Prentice Hall, 1980). 33 Since the literature focuses on the role of emotion, mention of motivation will be absent in these latter sections. 34 Jesse Prinz, “The Emotional Basis of Moral Judgments,” Philosophical Explorations 9 (1) (2006): 32. 35 R. J. R. Blair, “A Cognitive Developmental Approach to Morality: Investigating the Psychopath.” Cognition 57 (1995): 1–29. Cited by Prinz, Ibid. 36 Prinz, “The Emotional Basis of Moral Judgments,” p. 32. 37 I am using judgment and belief synonymously here. 38 Antonio Damasio, “Neuroscience and Ethics: Intersections,” American Journal of Bioethics 7 (1) (2007): 3–8, at 3. 39 Damasio, “Neuroscience and Ethics: Intersections,” p. 4. For more striking evidence see Jonathan Greene et al. “An fMRI Investigation of Emotional Engagement in Moral Judgment,” Science 293 (2001): 2105–2108.
The reader should note that statements of attitude are similar to moral judgments because they are statements of value, an evaluation of a person or commitment to more abstract values. 41 Kari Edwards, “The Interplay of Affect and Cognition in Attitude Formation and Change,” Journal of Personality and Social Psychology 59 (2) (1990): 213. Emphasis mine. 42 Freund, “The Freezing and Unfreezing of Impressional Primacy: Effects of the Need for Structure and the Fear of Invalidity,” Personality and Social Psychology Bulletin 11 (4) (1985): 479–487. 43 Ibid., 479. 44 See Jonathan Haidt and Jonathan Baron, “Social Roles and Moral Judgment of Acts and Omissions.” European Journal of Social Psychology 26 (1996): 201–218. 45 Ziva Kunda, “The Case for Motivated Reasoning,” Psychological Bulletin 108 (1990): 481. 46 Ibid., 483. 47 By superior moral judgment I do not mean one single judgment that is better than the rest. I mean to denote the idea of a moral judgment that is good in every respect.
Chapter 7 1
Schellenberg does not tell us whether epistemic parity means that G and not-G are equally probable propositions, or that the probability of G alone, or not-G alone is inscrutable. In fairness to Schellenberg, I suspect that failure to distinguish the two is inconsequential. 2 J. L. Schellenberg, Divine Hiddenness and Human Reason (Ithaca, NY: Cornell University Press, 1993), p. 64. Given the centrality of the notion of inculpable doubt in the argument from hiddenness, the argument poses particular difficulty for Alvin Plantinga’s proper function theory applied to religious knowledge. For Plantinga, religious knowledge is a function of the “sense of the divine” faculty working properly. If it works properly, it picks up on the divine presence in the world in certain circumstances. When the faculty is not working properly, it is due largely to the noetic effects of sin (i.e., Original sin). The problem is that understood as a faculty it is largely not in our control and likewise, understood as malfunctioning due to Original sin it is also not in our control. Thus, if it is not functioning properly the agent will suffer from inculpable doubt. And this grants exactly what the proponent of the hiddenness argument wants. I shall show below, however, that the responsibilist has resources to rebut this key premise. 3 Schellenberg, Divine Hiddenness, p. 66 4 Schellenberg, “The Hiddenness Argument Revisited (I),” Religious Studies 41 (2005): 202. 5 Michael Murray, “Coercion and the Hiddenness of God,” American Philosophical Quarterly 30 (1993): 34. 6 Douglas V. Henry, “Does Reasonable Nonbelief Exist?” Faith and Philosophy 18 (2001): 79. 7 Schellenberg, Divine Hiddenness, p. 18.
I confine my comments to Christian theology since it is the position with which I am most familiar. 9 William J. Wainwright, “Jonathan Edwards and the Hiddenness of God,” in Divine Hiddenness: New Essays, eds. Daniel Howard-Snyder and Paul K. Moser (New York: Cambridge University Press, 2002), 113. Wainwright quotes from Schellenberg, Divine Hiddenness, p. 30. 10 Schellenberg, “The Hiddenness Argument Revisited (I),” p. 208. 11 Laura L. Garcia, “St. John of the Cross and the Necessity of Divine Hiddenness,” in Divine Hiddenness: New Essays, eds. Daniel Howard-Snyder and Paul K. Moser (New York: Cambridge University Press, 2002), p. 87. 12 Ibid., 87. 13 C. A. J. Coady, Testimony: A Philosophical Study (New York: Oxford University Press, 1992), 213. 14 This is, I think, very close to the actual world scenario. See Caroline Franks Davis’s Evidential Force of Religious Experience (New York: Oxford University Press, 1999). 15 See William Rowe, Can God Be Free (New York: Oxford University Press, 2004).
Linda T. Zagzebski, Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge (New York: Cambridge University Press, 1996), p. 336. 2 Richard Paul, “Critical Thinking, Moral Integrity, and Citizenship: Teaching for the Intellectual Virtues,” in Knowledge, Belief, and Character: Readings in Virtue Epistemology, ed. Guy Axtell (Lanham: Rowan & Littlefield Publishers, 2000), p. 163. 3 Ibid., 170. 4 See Colleen Seifert et al., “Demystification of Cognitive Insight: Opportunistic Assimilation and the Prepared-Mind Perspective,” in The Nature of Insight, eds. Robert Sternberg and Janet Davidson (Boston: MIT Press, 1995), pp. 63–95. 5 Deirdre A. Kramer, “Conceptualizing Wisdom: The Primacy of Affect-Cognition Relations,” in Wisdom: Its Nature, Origins, and Development, ed. Robert Sternberg (Cambridge: Cambridge University Press, 1990), pp. 279–316. 6 Ibid., 280. 7 Janet Davidson, “The Suddenness of Insight,” in The Nature of Insight, eds. Robert Sternberg and Janet Davidson (Boston: MIT Press, 1995), p. 127.
Adler, Jonathan. “Transmitting Knowledge.” Nous 30 (1996): 99–111. Ahissar, Merav and Hochstein, Shaul. “The Role of Attention in Learning Simple Visual Tasks.” In Perceptual Learning, edited by Manfred Fahle and Thomaso Poggio. Cambridge: MIT Press, 2002. Ahissar, Merav and Hochstein, Shaul. “Attentional Control of Early Perceptual Learning.” Proceedings of the National Academy of Science USA 90 (1993): 5718. Alston, William P. “Epistemic Desiderata.” Philosophy and Phenomenological Research 53 (1993): 527–551. Alston, William P. Epistemic Justification: Essays in the Theory of Knowledge. Ithaca: Cornell University Press, 1989. Alston, William P. “Virtue and Knowledge.” Philosophy and Phenomenological Research 60 (2000): 185–189. American Society for Bioethics and Humanities. Core Competencies for Health Care Ethics Consultation, 1998. Audi, Robert. Moral Knowledge and Ethical Character. New York: Oxford University Press, 1997. Audi, Robert. “The Place of Testimony in the Fabric of Knowledge and Justification.” American Philosophical Quarterly 34 (1997): 405–422. Aurell, Carl G. “Man’s Triune Conscious Mind.” Perceptual and Motor Skills 68 (1989): 747–754. Averbach, Emanuel and Coriell, Abner S. “Short-Term Memory in Vision.” Bell System Technology Journal 40 (1961): 309–328. Axtell, Guy. “Introduction.” In Knowledge, Belief, and Character: Readings in Virtue Epistemology, edited by Guy Axtell. Lanham: Rowan & Littlefield Publishers, 2000. Baddeley, Alan D. et al. “Attention and Retrieval from Long-Term Memory.” Journal of Experimental Psychology: General 13 (1984): 518–540. Balota, David A. “Automatic Semantic Activation and Episodic Memory Encoding.” Journal of Verbal Learning and Verbal Behavior 22 (1983): 88–104. Bargh, John. A. and Pietromonaco, Paula. “Automatic Information Processing and Social Perception: The Influence of Trait Information Presented Outside of Conscious Awareness on Impression Formation.” Journal of Personality and Social Psychology 43 (1982): 437–449. Battaly, Heather D. “Thin Concepts to the Rescue: Thinning the Concepts of Epistemic Justification and Intellectual Virtue.” In Virtue Epistemology, edited by Abrol Fairweather and Linda T. Zagzebski. New York: Oxford University Press, 2001. Baehr, Jason. “Character in Epistemology,” Philosophical Studies 128 (2006): 479–514. Beilby, James. “Review of Coady, C. A. J.” Testimony: A Philosophical Study. Oxford: Oxford University Press, 1992. Philosophia Christi 1 (1999): 145.
Berlyne, Daniel E. “Curiosity and Exploration.” Science 153 (1966): 25–33. Bernstein, Marica. et al. “The Effect of Motivation on the Stream of Consciousness: Generalizing from a Neurocomputational Model of Cingulo-Frontal Circuits Controlling Saccadic Eye Movements.” In The Caldron of Consciousness: Affect, Motivation and Self-Organization, edited by Ralph D. Ellis and Natika Newton. Amsterdam: John Benjamins, 2000. Blair, R. James. R. “A Cognitive Developmental Approach to Morality: Investigating the Psychopath.” Cognition 57 (1995): 1–29. Blanshard, Brand. On Philosophical Style. Bloomington: Indiana University Press, 1954. BonJour, Laurence. “Externalist Theories of Empirical Knowledge.” Midwest Studies in Philosophy 5 (1980): 53–73. Brown, Scott C. and Craik, Fergus I. M. “Encoding and Retrieval of Information.” The Oxford Handbook of Memory, edited by Endel Tulving and Fergus I. M. Craik. New York: Oxford University Press, 2000. Bruner, Jerome. “On Perceptual Readiness.” Psychological Review 64 (1957): 123–152. Burge, Tyler. “Content Preservation.” The Philosophical Review 102 (1993): 457–488. Chartrand, Tanya and Bargh, John A. “Automatic Activation of Impression Formation and Memorization Goals: Nonconscious Goal Priming Reproduces Effects of Explicit Task Instructions.” Journal of Personality and Social Psychology 71 (1996): 464–478. Chartrand, Tanya L. and Bargh, John A. “Nonconscious Motivations: Their Activation, Operation, and Consequences.” In Self and Motivation: Emerging Psychological Perspectives, edited by Abraham Tesser, Diederik A. Stapel, and Joanne V. Wood. Washington, D.C.: American Psychological Association, 2002. Chase, W. G. and Simon, Herbert A. “In the Mind’s Eye,” in Visual Information Processing. edited by W. G. Chase (New York: Academic Press, 1973), pp. 215–281. Cheeseman, Jim and Merikle, Philip M. “Distinguishing Conscious from Unconscious Perceptual Processes.” Canadian Journal of Psychology 40 (1986): 343–367. Cheeseman, J. and Merikle, P. M. “Priming Without Awareness.” Perception and Psychophysics 36 (1984): 387–395. Christianson, Sven-Ake and Engelberg, Elizabeth. “Organization of Emotional Memories.” In Handbook of Cognition and Emotion, edited by Tim Dalgleish and Mick J. Power. New York: Wiley & Sons, 1999. Coady, C. Anthony J. Testimony: A Philosophical Investigation. London: Oxford University Press, 1992. Cowan, Richard and Carney, Daniel P. J. “Calendrical Savants: Exceptionality and Practice.” Cognition 100 (2006): B1–B9. Cowey, Alan. “The 30th Sir Frederick Bartlett Lecture: Fact, Artefact, and Myth about Blindsight,” The Quarterly Journal of Experimental Psychology 57A (4) (2004): 577–609. Craik, Fergus. I. M. et al. “The Effects of Divided Attention on Encoding and Retrieval Processes in Human Memory.” Journal of Experimental Psychology: General 125 (1996): 159–180. Crosthwaite, Jan. “Moral Expertise: A Problem in the Professional Ethics of Professional Ethicists.” Bioethics 9 (1995): 361–379.
Damasio, Antonio. “Neuroscience and Ethics: Intersections.” American Journal of Bioethics 7 (1) (2007): 3–8. Davidson, Janet. “The Suddenness of Insight.” In The Nature of Insight, edited by Robert Sternberg and Janet Davidson. Boston: MIT Press, 1995. Davis, Caroline Franks. Evidential Force of Religious Experience. New York: Oxford University Press, 1999. Depaul, Michael. Balance and Refinement: Beyond Coherence Methods of Moral Inquiry. London: Routledge Publishers, 1993. Derryberry, Douglas and Reed, Majorie. “A Multidisciplinary Perspective on Attentional Control.” In Attraction, Distraction and Action: Multiple Perspectives on Attentional Capture, edited by Bradley Gibson. Oxford: Elsevier Science, 2001. Derryberry, Douglas and Tucker, Don M. “Motivating the Focus of Attention.” In The Heart’s Eye: Emotional Influences in Perception and Attention, edited by Paula Niedenthal and Shinobu Kitayama. San Diego: Academic Press, 1994. Dretske, Fred. “A Cognitive Cul-de-sac.” Mind 91 (1982): 109–111. Edwards, Kari. “The Interplay of Affect and Cognition in Attitude Formation and Change.” Journal of Personality and Social Psychology 59 (2) (1990): 213. Effken, Judith A. “Informational Basis for Expert Intuition.” Journal of Advanced Nursing 34 (2) (2001): 246–255. Eich, Eric. “Memory for Unattended Events: Remembering with and without Awareness.” Memory and Cognition 12 (1984): 105–111. Elgin, Catherine Z. “Take It from Me: The Epistemological Status of Testimony.” Philosophy and Phenomenological Research 65 (2002): 291–308. Ellis, Ralph D. “Dynamic Systems as an Approach to Consciousness: Emotion, SelfOrganization, and the Mind-Body Problem.” New Ideas in Psychology 17 (1999): 237–250. Ellis, Ralph D. “Implications of Inattentional Blindness for ‘Enactive’ Theories of Consciousness.” Brain and Mind 2 (2001): 297–322. Ellis, Ralph D. Curious Emotions: Roots of Consciousness and Personality in Motivated Action. Amsterdam: John Benjamins Publishers, 2005. Fahle, Manfred. “Specificity of Learning Curvature, Orientation, and Vernier Discriminations.” Vision Research 37 (1997): 1885–1895. Fahle, Manfred. “Introduction.” In Perceptual Learning, edited by Manfred Fahle and Tomaso Poggio. Cambridge: MIT Press, 2002. Fairweather, Abrol. “Epistemic Motivation.” In Virtue Epistemology, edited by Linda T. Zagzebski and Abrol Fairweather. New York: Oxford University Press, 2000. Feldman, Richard. “Reliability and Justification.” The Monist 68 (1985): 159–174. Ferguson, Eva D. Motivation: A Biosocial and Cognitive Integration of Motivation and Emotion. New York, NY: Oxford University Press, 2000. Foley, Richard. “Justified Inconsistent Beliefs.” American Philosophical Quarterly 16 (1979): 247–257. Fowler, Carol A. Wolford, George et. al. “Lexical Access with and without Awareness.” Journal of Experimental Psychology: General 110 (1981): 341–362. Frankfurt, Harry. “Alternative Possibilities and Moral Responsibility.” The Journal of Philosophy 66 (1969): 829–839. Franklin, Richard L. The Search for Understanding. New York: Peter Lang Publishers, 1995.
Freund, Tallie. “The Freezing and Unfreezing of Impressional Primacy: Effects of the Need for Structure and the Fear of Invalidity.” Personality and Social Psychology Bulletin 11 (4) (1985): 479–487. Fricker, Elizabeth. “Against Gullibility.” In Knowing from Words, edited by Bimal Krishna Matilal and Arindam Chakrabarti. Dordrecht: Kluwer, 1994. Fricker, Elizabeth. “The Epistemology of Testimony.” Proceedings of the Aristotelian Society Supp. 61 (1987): 57–83. Fricker, Elizabeth. “Telling and Trusting: Reductionism and Anti-Reductionism in Epistemology and Testimony.” Mind 104 (1995): 393–411. Gallant, Jack L. et al. “Neural Activity in Areas V1, V2, and V4 During Free Viewing of Natural Scenes Compared to Controlled Viewing.” NeuroReports (1998): 9. Garcia, Laura L. “St. John of the Cross and the Necessity of Divine Hiddenness,” in Divine Hiddenness: New Essays, edited by Daniel Howard-Snyder and Paul K. Moser. New York: Cambridge University Press, 2002. Gettier, Edmund. “Is Justified True Belief Knowledge?” Analysis 23 (1963): 121– 123. Gettier, Edmund. “Is Justified True Belief Knowledge?” In Empirical Knowledge: Readings in Contemporary Epistemology, edited by Paul Moser. Totowa: Rowan & Littlefield, 1986. Goldie, Peter. The Emotions. New York: Oxford University Press, 2000. Goldman, Alvin. “Epistemic Folkways and Scientific Epistemology.” In Knowledge, Belief and Character, edited by Guy Axtell. Lanham: Rowan & Littlefield, 2000. Goldman, Alvin. “Discrimination and Perceptual Knowledge.” The Journal of Philosophy 73 (1976): 771–791. Goldman, Alvin. “Discrimination and Perceptual Knowledge.” In Liaisons: Philosophy Meets the Cognitive and Social Sciences, edited by Alvin Goldman. Cambridge: MIT Press, 1992. Goldman, Alvin. “What is Justified Belief.” In Liaisons: Philosophy Meets the Cognitive and Social Sciences. Cambridge: MIT Press, 1992. Goldman, Alvin. Epistemology and Cognition. Cambridge: Harvard University Press, 1986. Goldman, Alvin. Knowledge in a Social World. New York: Oxford University Press, 1999. Goldstone, Robert. “Perceptual Learning.” Annual Review of Psychology 49 (1998): 585–612. Gordon, Robert M. The Structure of Emotions. New York: Cambridge University Press, 1987. Graham, Peter. “Transferring Knowledge.” Nous 34 (2000): 131–152. Greco, John. “Agent Reliabilism.” In Philosophical Perspectives: Epistemology, edited by James Tomberlin. Atascadero: Ridgeview Press, 1999. Greco, John. “Two Kinds of Intellectual Virtue,” Philosophy and Phenomenological Research 60 (2000): 181. Greco, John. “Knowledge as Creditable True Belief,” In Intellectual Virtue: Perspectives from Ethics and Epistemology, edited by Michael Depaul and Linda Zagzebski. New York: Oxford University Press, 2003. Greco, John. “Virtue, Luck, and the Pyrrhonian Problematic.” Philosophical Studies 130 (2006): 9–34.
Greene, Jonathan et al. “An fMRI Investigation of Emotional Engagement in Moral Judgment.” Science 293 (2001): 2105–2108. Haidt, Jonathan. “The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgement.” Psychological Review 108 (2001): 814–834. Haidt, Jonathan and Baron, Jonathan. “Social Roles and Moral Judgment of Acts and Omissions.” European Journal of Social Psychology 26 (1996): 201–218. Hanneman, Sandra K. “Advancing Nursing Practice with a Unit-Based Clinical Expert.” IMAGE: Journal of Nursing Scholarship 28 (1996): 333. Henry, Douglas V. “Does Reasonable Nonbelief Exist?” Faith and Philosophy 18 (2001): 75–92. Holender, D. “Semantic Activation without Conscious Identification in Dichotic Listening, Parafoveal Vision, and Visual Masking: A Survey and Appraisal.” Behavioral and Brain Sciences 9 (1986): 1–66. Hume, David. An Enquiry Concerning Human Understanding. 2nd ed., ed. L. A. SelbyBiggs. Oxford: Oxford University Press, 1957. Hursthouse, Roselind. On Virtue Ethics. New York: Oxford University Press, 1999. Jenkins, John. Knowledge and Faith in Thomas Aquinas. New York: Cambridge University Press, 1997. Johnson, Marcia, K. and Raye, Carol L. “Reality Monitoring.” Psychological Review 88 (1981): 67–85. Johnson, Marcia, Hashtroudi, Shahin, and Lindsay, Stephen. “Source Monitoring.” Psychological Bulletin 114 (1993): 3–28. Johnston, William A. et al. “Divided Attention: A Vehicle for Monitoring Memory Processes.” Journal of Experimental Psychology 83 (1970): 164–171. Kahneman, Daniel. Attention and Effort. Englewood Cliffs: Prentice-Hall, 1973. Kerr, Norbert L., Kramer, Geoffrey P., and MacCoun, Robert J. “Bias in Judgment: Comparing Individuals and Groups.” Psychological Review 103 (4) (1996): 687 –719. Kiefer, Markus. “The N400 is Modulated by Unconscious Perceived Masked Words: Further Evidence for an Automatic Spreading Activation Account of N400 Priming Effects.” Cognitive and Brain Research 13 (2002): 27–39. Klein, Gary. The Power of Intuition: How to Use Your Gut Feelings to Make Better Decisions at Work. New York: Currency Books, 2003. Kramer, Deirdre A. “Conceptualizing Wisdom: The Primacy of Affect-Cognition Relations.” In Wisdom: Its Nature, Origins, and Development, edited by Robert Sternberg. Cambridge: Cambridge University Press, 1990. Kunda, Ziva. “The Case for Motivated Reasoning.” Psychological Bulletin 108 (1990): 481. Kunda, Ziva. Social Cognition: Making Sense of People. Cambridge: MIT Press, 1999. Kvanvig, Jonathan L. The Intellectual Virtues and the Life of the Mind: On the Place of the Virtues in Epistemology. Lanham: Rowan and Littlefield Publishers, 1992. Lackey, Jennifer. “Testimonial Knowledge and Transmission.” The Philosophical Quarterly 49 (1999): 471–490. Lang, Peter J. et. al. “Motivated Attention: Affect, Activation and Action.” In Attention and Orienting: Sensory and Motivational Processes, edited by P. J. Lang et al. Hillsdale: Erlbaum, 1997. Lehrer, Keith. “A Critique of Externalism.” In A Theory of Knowledge, edited by Louis Pojman. Belmont: Wadsworth Publishing Company, 1993.
Liu, Z. “Perceptual Learning in Motion Discrimination that Generalizes across Motion Directions.” Proceedings of the National Academy of Science USA 96 (1999): 14085–14087. Mack, Arien and Rock, Irving. Inattentional Blindness. Cambridge: MIT Press, 1998. Mack, Arien, Tang, Benyu, Tuma, Regina, Kahn, Steven, and Rock, Irving. “Perceptual Organization and Attention.” Cognitive Psychology 24 (1992): 475–502. MacLeod, Colin and Matthews, Andrew. “Anxiety and the Allocation of Attention to Threat.” Quarterly Journal of Experimental Psychology 40 (1988): 653–670. Madsen, K. B. Theories of Motivation: A Comparative Study of Modern Theories of Motivation. Kent: Kent State University Press, 1968. Merricks, Trenton. “More on Warrant Entails Truth.” Philosophy and Phenomenological Research 57 (1997): 627–631. Nauta, Walle J. H. “The Problem of the Frontal Lobe: A Reinterpretation.” Journal of Psychiatric Research 8 (1971): 167–187. Nettelbeck, Ted and Young, Robyn. “Intelligence and Savant Syndrome: Is the Whole Greater than the Sum of the Fragments?” Intelligence 22 (1996): 49–68. Niedenthal, Paula M. et al. “Emotional Organization of Perceptual Memory.” In The Heart’s Eye, edited by Paula M. Niedenthal and Shinobu Kitayama. San Diego: Academic Press, 1994. Nisbett, Richard and Ross, Lee. Human Inference: Strategies and Shortcomings of Social Judgment. Englewood Cliffs: Prentice Hall, 1980. Nozick, Robert. Philosophical Explanations. Cambridge: Harvard University Press, 1981. O’Connor, Daniel John and Carr, Brian. Introduction to the Theory of Knowledge. Minneapolis: University of Minnesota Press, 1982. Ohman, Arne. “Distinguishing Unconscious from Conscious Emotional Processes: Methodological Considerations and Theoretical Implications.” In Handbook of Cognition and Emotion, edited by Tim Dalgleish and Mick Power. Chichester: Wiley and Sons, 1999. Palmer, Stephen. Vision Science: From Photons to Phenomenology. Cambridge: MIT Press, 1999. Panskepp, Jaak. Affective Neuroscience. New York: Oxford University Press, 1998. Parker, Lisa S. “Ethical Expertise, Maternal Thinking, and the Work of Clinical Ethicists.” In Ethics Expertise: History, Contemporary Perspectives and Applications, edited by Lisa Rasmussen. Amsterdam: Springer Publishers, 2005. Paul, Richard. “Critical Thinking, Moral Integrity, and Citizenship: Teaching for the Intellectual Virtues.” In Knowledge, Belief, and Character: Readings in Virtue Epistemology, edited by Guy Axtell. Lanham: Rowan & Littlefield Publishers, 2000. Plantinga, Alvin. “Introduction.” In Naturalism Defeated? Essays on Plantinga’s Evolutionary Argument against Naturalism, edited by James K. Beilby. Ithaca: Cornell University Press, 2002. Plantinga, Alvin. Warranted Christian Belief. New York: Oxford University Press, 2000. Plantinga, Alvin. “Reliabilism, Analyses and Defeaters.” Philosophy and Phenomenological Research 55 (1995): 427–464. Plantinga, Alvin. Warrant and Proper Function. New York: Oxford University Press, 1993.
Plantinga, Alvin. Warrant: The Current Debate. New York: Oxford University Press, 1993. Pollock, John. Contemporary Theories of Knowledge. Totowa: Rowan & Littlefield, 1986. Porter, Stephen and Birt, Angela R. “Is Traumatic Memory Special? A Comparison of Traumatic Memory Characteristics with Memory for Other Emotional Life Experiences.” Applied Cognitive Psychology 15 (2001): 101–117. Posner, Michael I. and Rothbart, Mary K. “Attentional Mechanisms and Conscious Experience.” In The Neuropsychology of Consciousness, edited by A. D. Milner and M. D. Rugg. London: Academic Press, 1992, pp. 187–210. Prinz, Jesse. “The Emotional Basis of Moral Judgments,” Philosophical Explorations 9 (1) (2006): 29–43. Pritchard, Duncan. “Virtue Epistemology and Epistemic Luck.” Metaphilosophy 34 (2003): 106–130. Putnam, Hilary. “Why Reason Can’t Be Naturalized.” In Realism and Reason: Philosophical Papers. Vol. 3 edited by Hilary Putnam. Cambridge: Cambridge University Press, 1983. Pylyshyn, Zenon. “Is Vision Continuous with Cognition? The Case for Cognitive Impenetrability of Visual Perception.” Behavioral and Brain Sciences 22 (1999): 341–423. Ramachandran, Vilayanur and Braddick, Oliver. “Orientation-Specific Learning in Steropsis.” Perception 2 (1973): 371–376. Rensink, Ronald A., O’Reagan, J. Kevin, and Clark, James J. “To See or Not to See: The Need for Attention to Perceive Changes in Scenes.” Psychological Science 8 (1997): 368–373. Revelle, William. “Individual Differences in Personality and Motivation: ‘non-cognitive’ Determinants of Cognitive Performance.” In Attention: Selection, Awareness and Control, edited by Alan Baddeley and Lawrence. Weiskrantz. Oxford: Oxford University Press, 1993. Roediger, Henry L., McDermott, Kathleen B., and Goff, Lyn. “Recovery of True and False Memories: Paradoxical Effects of Repeated Testing.” In False Memories and Recovered Memories, edited by Martin Conway. Oxford: Oxford University Press, 1997. Roediger, Henry L. III. “Why Retrieval is the Key Process in Understanding Memory.” In Memory, Consciousness and the Brain: The Tallinn Conference. Philadelphia: Psychology Press, 2000. Rorty, Amelie Oksenberg. “Distinctive Measures of Epistemic Evaluation: Character as the Configuration of Traits.” Philosophy and Phenomenological Research 60 (2000): 203–206. Ross, Michael et al. “The Effect of Attitude on the Recall of Personal Histories.” Journal of Personality and Social Psychology 40 (1981): 627–634. Rowe, William. Can God Be Free. New York: Oxford University Press, 2004. Thomson, Judith Jarvis. “A Defense of Abortion,” In The Problem of Abortion, edited by Joel Feinberg. Belmont: Wadsworth Publishers, 1984. Sayre, Kenneth. Belief and Knowledge. Lanham: Rowan & Littlefield, 1997. Schacter, Daniel L. and Dodson, Chad S. “Misattribution, False Recognition and the Sins of Memory.” In Episodic Memory: New Directions in Research, edited by Alan
Baddeley, Martin A. Conway, and John P. Aggleton. New York: Oxford University Press, 2002. Schellenberg, John L. Divine Hiddenness and Human Reason. Ithaca: Cornell University Press, 1993. Schellenberg, John L. “The Hiddenness Argument Revisited (I).” Religious Studies 41 (2005): 201–215. Seifert, Colleen et al. “Demystification of Cognitive Insight: Opportunistic Assimilation and the Prepared-Mind Perspective.” In The Nature of Insight, edited by Robert Sternberg and Janet Davidson. Boston: MIT Press, 1995. Shiffrin, Richard. “Modeling Memory and Perception.” Cognitive Science 27 (2003): 341–378. Simon, Herbert. “Bottleneck of Attention.” Integrative Views of Motivation, Cognition, and Emotion: Vol. 41 of the Nebraska Symposium on Motivation, edited by William D. Spaulding. Lincoln: University of Nebraska Press, 1994. Simons, Daniel J. and Levin, Daniel T. “Change Blindness.” Trends in Cognitive Science 1 (1997): 261–267. Skyrms, Brian. “The Explication of ‘X knows that p’.” The Journal of Philosophy 64 (1967): 373–389. Slote, Michael. Morals from Motives. New York: Oxford University Press, 2001. Sosa, Ernest. “Knowledge and Intellectual Virtue,” Monist 68 (1985): 242. Sosa, Ernest. “For the Love of Truth.” In Virtue Epistemology, edited by Linda T. Zagzebski and Abrol Fairweather. New York: Oxford University Press, 2000. Sosa, Ernest. “Postscript to ‘Proper Functionalism and Virtue Epistemology.’” In Warrant in Contemporary Epistemology: Essays in Honor of Plantinga’s Theory of Knowledge, edited by Jonathan L. Kvanvig. Lanham: Rowan Littlefield Publishers, 1996. Sosa, Ernest. “Proper Functionalism and Virtue Epistemology.” Nous 27 (1993): 51–65. Sosa, Ernest. Knowledge in Perspective: Selected Essays in Epistemology. Cambridge: Cambridge University Press, 1991. Spiro, Rand J. “Remembering Information from Text: The ‘State of Schema’ Approach.” In Schooling and the Acquisition of Knowledge, edited by Richard C. Anderson et al. Hillsdale: Erlbaum, 1977. Steup, Matthias. “Proper and Improper Use of Cognitive Faculties: A Counterexample to Plantinga’s Proper Functioning Theory.” Philosophy and Phenomenological Research 55 (1995): 409–413. Stotland, Ezra and Canon, Lance. Social Psychology: A Cognitive Approach. Philadelphia: Saunders Publishing, 1972. Stump, Eleonore. “Aquinas on the Foundations of Knowledge.” In Aristotle and His Medieval Interpreters, edited by Richard Bosley and Martin Tweedale. Canadian Journal of Philosophy, Supplementary Volume 17 (1991): 125–158. Stump, Eleonore. Aquinas: Arguments of the Philosophers. London: Routledge Publishers, 2002. Swanton, Christine. Virtue Ethics: A Pluralist Approach. New York: Oxford University Press, 2005. Timmons, Mark. “Review of Balance and Refinement: Beyond Coherence Methods of Moral Inquiry.” Ethics, 105 (3) (1995): 651–652.
Wainwright, William J. “Jonathan Edwards and the Hiddenness of God,” In Divine Hiddenness: New Essays, edited by Daniel Howard-Snyder and Paul K. Moser. New York: Cambridge University Press, 2002. Warfield, Ted. “Knowledge from Falsehood,” Philosophical Perspectives: Epistemology 19, (2005): 405–416. Way, Lawrence W. et al. “Causes and Prevention of Laparoscopic Bile Duct Injuries: Analysis of 252 Cases from a Human Factors and Cognitive Psychology Perspective.” Annals of Surgery 237 (2003): 460–469. Weiner, Bernard. Theories of Motivation: From Mechanism to Cognition. Chicago: Markham Publishing, 1972. Weinstein, Bruce D. “What is an Expert?,” Theoretical Medicine 14 (1993): 57–73. Welbourne, Michael. “Testimony, Knowledge and Belief.” In Knowing from Words, edited by Bimal Krishna Matilal and Arindam Chakrabarti. Dordrecht: Kluwer, 1994. Wise, Roy A. “Sensorimotor Modulation and the Variable Action Pattern (VAP): Toward a Noncircular Definition of Drive and Motivation.” Psychobiology 15 (1987): 7–20. Scott Yoder, Scott. “Experts in Ethics? The Nature of Ethical Expertise,” Hastings Center Report 28 (1998): 11–19. Zagzebski, Linda T. “Must Knowers Be Agents.” In Virtue Epistemology, edited by Abrol Fairweather and Linda T. Zagzebski. New York: Oxford University Press, 2001. Zagzebski, Linda T. “From Reliabilism to Virtue Epistemology.” In Knowledge, Belief and Character, edited by Guy Axtell. Lanham: Rowan and Littlefield Publishers, 2000. Zagzebski, Linda T. “What is Knowledge.” In The Blackwell Guide to Epistemology, edited by John Greco and Ernest Sosa. Oxford: Blackwell Publishers, 1999. Zagzebski, Linda T. “Recovering Understanding,” In Knowledge, Truth, and Duty: Essays on Epistemic Justification, Responsibility, and Virtue, edited by Matthias Steup. New York: Oxford University Press, 2001. Zagzebski, Linda T. Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. New York: Cambridge University Press, 1996. Zagzebski, Linda T. “Responses.” Philosophy and Phenomenological Research 60 (2000): 209–221. Zagzebski, Linda T. “The Search for the Source of Epistemic Good.” Metaphilosophy 34 (2003): 12–28. Zagzebski, Linda T. Divine Motivation Theory. New York: Cambridge University Press, 2004.
Act of intellectual virtue, 8, 34–5, 90, 144–5 Adequate investigation, 16, 130–1, 134–7 Alston, William, 34, 150 Attention: and consciousness, 42 ff. as requisite for knowledge, 15, 45–7, 60–5 blindsight, 56–8 motivation, 47 ff., 65 ff. Baehr, Jason, 14, 147 Barn façade case, 89, 148 ‘Because of’ clause, 8, 11 Bonjour, Lawrence, 22, 148 C.A.J. Coady, 77, 79–81, 120 Change blindness, 69–71, 155 Cognitive methods vs. cognitive processes, 21, 24–5 Contextual nature of testimony, 86, 87, 93, 100, 104 Counterexamples to responsibilism: unmotivated belief objection, 10, 15, 17, 34 ff., 47, 51, 60–1, 125, 142 blindsight, 56–8, 153 capture of attention, 54–5 easy knowledge, 14–15, 100 free scanning, 55–6, 72 sudden loud noises, 54–5 Depaul, Michael, 117, 147, 149, 161 Double-luck cases, 4, 8, 11, 15, 30–1, 33, 97–8, 104 Dr. Jones: case of, 7–8, 11–12, 104 Dr. Naïve and Dr. Thorough: case of, 9–10 Efferent vs. afferent processing, 48 Emotion as representation, 5 Emotional processing and moral judgment, 117–19, 124–6
Enactive view of consciousness, 47–50, 150 Encoding vs. retrieval, 60–3 Epistemic motivation, 51–3, 56, 60–1, 63–9, 72–4, 150, 156 Fair-mindedness, 73–4, 96, 135, 143 Fairweather, Abrol: response to unmotivated belief objection, 36–7, 150, 159 First-order vs. second order processes, 21, 148 Frankfurt case: counterexamples to reliabilism, 101–2, 159 Fricker, Elizabeth, 77, 81–8, 90, 100, 103 Gettier, Edmund, 2, 146 Global reduction, 83–4 God: as passible, 133 as goodness, 132–4 Graham, Peter, 89, 100, 157, 158, 159 Greco, John, 10, 13–14, 17, 147, 148 Inattentional blindness, 42–4, 48, 150, 151 Incidental encoding, 60, 64, 69ff. J-rule: definition of, 19–20, 22 Justified belief, 20, 22, 148, 149 Knowledge argument, 48 Knowledge: Goldman’s definition of, 20 Responsibilism’s definition of, 8 Lackey, Jennifer, 88, 93, 95, 96, 98, 157, 158 Local reduction: definition of, 84 possibility of, 84–6 Motivation: unconscious/nonconscious, 50, 54–5, 72, 152, 153 accuracy vs. directional, 127–8
174 Motivation: unconscious/nonconscious, (Cont’d) motivation to be aware, 51–2, 55 nonepistemic motivations, 52, 68 thick vs. thin, 39, 74
Index Pritchard, Duncan, 2, 3, 11–12, 146, 147 Process type vs. process token, 20–1 Recognition vs. appearance, 46 Reflective equilibrium, 106, 117 Reflective luck, 3, 12–13
Norman: case of, 19–20, 22 Open-mindedness, 8, 97 Perception as inferential (New Look view), 49–50 Phronesis, 9, 16, 73, 107, 122, 124, 145 Plantinga, Alvin: counterexamples to reliabilism: Paul’s case, 23–6 autistic savants and proper function, 29 Case of Mary as counterexample to Plantinga, 30–2 damage-detector faculty, 27 on memory, 59 on testimony, 88ff. the epistemically serendipitous lesion, 23, 26 Port-Royal Logic, 81 p-processing, 41, 50–1 Priming effects, 43, 45, 151
Sins of memory, 75, 155 Sophistry, 1–145 Testimonial knowledge: definition of, 91 ff. adopting vs. believing that p, 91–2 authenticity vs. concern for the truth, 102 corroborative testimony (ornithologist case), 103–4, 137 epistemic status of, 103–4 The value problem, 32–4 Trust in one’s senses as a virtue, 35–6 Unconscious perception/implicit perception, 42–3, 50, 56–7 Understanding: nature of, 114–16 Unmotivated belief objection, see counterexamples Veritic luck, 3