Anne Meylan Foundations of an Ethics of Belief

PRACTICAL PHILOSOPHY

Herausgegeben von / Edited by

Herlinde Pauer-Studer • Neil Roughley Peter Schaber • Ralf Stoecker

Band 15 / Volume 15

The aim of the series is to publish high-quality work that deals with questions in practical philosophy from

a broadly analytic perspective. These include questions in meta-ethics, normative ethics and "applied" ethics, as well as in political philosophy, philosophy of law

and the philosophy of action. Through the publication of work in both German and English the series aims

to facilitate discussion between English- and Germanspeaking

practical philosophers.

Anne Meylan

Foundations of an Ethics of Belief

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliographie;

detailed bibliographic data is available in the Internet at http://dnb.ddb.de

North and South America by Transaction Books Rutgers University

Piscataway, NJ 08854-8042 trans@transactionpub.com

United Kingdom, Ire, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited

White Cross Mills Hightown

LANCASTER, LA1 4XS sales@gazellebooks.co.uk

Livraison pour la France et la Belgique: Librairie Philosophique J.Vrin

6, place de la Sorbonne ; F-75005 PARIS Tel. +33 (0)1 43 54 03 47 ; Fax +33 (0)1 43 54 48 18

www.vrin.fr

2013 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm nr. Frankfurt

www.ontosverlag.com

ISBN: 978-3-86838-189-4

2013

No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise

without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work

Printed on acid-free paper

ISO-Norm 970-6 This hardcover binding meets the International Library standard

Printed in Germany

by CPI buch bücher.de

To the members and the friends of Episteme. This research has been supported by the Swiss National Science Foundation.

TABLE OF CONTENTS Introduction .................................................................................................................................. 1

The initial intuition ............................................................................................................................. 1 Main objective ....................................................................................................................................... 3 Preliminary clarificatory remarks .......................................................................................................... 3 Two central problems ........................................................................................................................ 4 The problem of control and responsibility ......................................................................................... 5 The normative problem .............................................................................................................................. 6 Abstracts of the chapters ................................................................................................................... 6 Chapter 1: What the philosophy of action teaches us ..................................................................... 6 Chapter 2: The impossibility of acquiring beliefs directly for reasons .................................... 9 Chapter 3: Pascalian and theoretical control ..................................................................................... 9 Chapter 4: Doxastic responsibility as responsibility for consequences ................................ 11 Chapter 5: Epistemic praiseworthiness and epistemic blameworthiness ........................... 12 Chapter 6: Beyond epistemic justifiedness ....................................................................................... 13 Chapter 7: Epistemic justifiedness and non-epistemic justifiedness ..................................... 14

Chapter 1: What the philosophy of action teaches us .................................................. 17 Actions and happenings ...................................................................................................................17 Non-reductionist conception of action ............................................................................................... 18 Reductionist conception of action ........................................................................................................ 22 Actions, happenings and activities ..............................................................................................27 Acting for reasons ........................................................................................................................................ 28 Three distinctions about reasons ................................................................................................32 Motivating reasons vs. normative reasons ........................................................................................ 32 Internalism vs. externalism about reasons ....................................................................................... 33 Humean vs. anti-Humean conception of motivation ..................................................................... 34 Back to the doxastic realm ..............................................................................................................35 Epistemic reasons, non-epistemic reasons and evidence ........................................................... 35 Delineating the interesting issue ........................................................................................................... 38

Chapter 2: The Impossibility of directly acquiring beliefs for reasons ................. 41 Direct and indirect belief acquisitions .......................................................................................41 Direct/indirect acquisitions of belief and epistemic/non–epistemic reasons ................... 42 Williams’ argument ...........................................................................................................................45 “To believe that p is to believe that p is true” ..........................................................................52 Believing vs. imagining .............................................................................................................................. 53 Transparency ................................................................................................................................................. 55 The teleological account ........................................................................................................................... 57 Conclusions ..........................................................................................................................................60

Chapter 3: Theoretical and Pascalian control ................................................................ 64 Two forms of indirect doxastic control ......................................................................................68 Theoretical control ...................................................................................................................................... 69 Pascalian control .......................................................................................................................................... 79 Indirect doxastic influence on belief acquisitions .......................................................................... 82 Unlimited doxastic control considered ......................................................................................87 Ryan’s unlimited doxastic control ........................................................................................................ 87 Pieces of evidence vs. motivating reasons ......................................................................................... 89 Steup’s unlimited doxastic control ....................................................................................................... 92

Chapter 4: Doxastic Responsibility as Responsibility for Consequences ............. 97

Responsibility for consequences ..................................................................................................97 Responsibility for basic actions ............................................................................................................. 99 Responsibility for the consequences of actions ........................................................................... 103 Responsibility for resultant belief acquisitions, theoretical and Pascalian control ...... 119 Responsibility for resultant belief acquisitions and indirect doxastic influence ........... 124 Responsibility for believing .................................................................................................................. 125 Chapter 5: Epistemic praiseworthiness and blameworthiness ............................ 129

Epistemic and non-epistemic desirability ............................................................................. 129 The fundamental epistemic end ................................................................................................ 130 Other epistemically desirable states ................................................................................................. 131 The fundamental epistemic end: some specifications ............................................................... 134 Epistemic and non-epistemic ends: summary .............................................................................. 135 Varieties of epistemic goodness* .............................................................................................. 136 Final and instrumental epistemic goodness .................................................................................. 136 Epistemic rationality and epistemic commendability ............................................................... 137 Varieties of epistemic praiseworthiness and blameworthiness ................................... 141 Final and instrumental epistemic praiseworthiness and blameworthiness .................... 141 Epistemic praiseworthiness/blameworthiness for rational belief acquisitions ............ 144 Epistemic praiseworthiness for epistemically commendable belief acquisitions and epistemic blameworthiness for epistemically non-commendable belief acquisitions 146

Chapter 6: Beyond epistemic justifiedness .................................................................. 151 Accessibilism, mentalism, and externalism .......................................................................... 154 Accessibilism and perceptual disjunctivism .................................................................................. 157 Normative properties .............................................................................................................................. 160 Valuable, rational, commendable belief acquisitions and the threefold classification of justifiedness ................................................................................................................................. 163 Externalism: the goodness* of instrumental goodness ............................................................. 164 Mentalism: the goodness* of rationality ......................................................................................... 164 Accessibilism: the goodness* of commendability ....................................................................... 167 The reliabilist and the accessibilist explanation of the goodness* of justifiedness 172 The reliabilist explanation of the goodness* of justifiedness ................................................. 173 The credit explanation of the goodness* of justifiedness ........................................................ 175 Accessibilist explanation of the goodness* of justifiedness .................................................... 179

Chapter 7: Epistemic and non-epistemic justifiedness ............................................ 185 The divergence thesis .................................................................................................................... 185 The “pragmatic” refutation of the divergence thesis: Clifford and James .................. 191 Clifford’s ethics of belief ......................................................................................................................... 192 James’ ethics of belief .............................................................................................................................. 196 The point of agreement .......................................................................................................................... 202 The divergence of rationality ..................................................................................................... 203 The objection against the divergence of rationality ................................................................... 204

Conclusion ................................................................................................................................ 207 Bibliography ............................................................................................................................ 211

1

INTRODUCTION

Our daily intellectual life involves lots of evaluations of actions. We consider that driving above the speed limit is dangerous, that giving one’s bus seat to older people is polite, that stirring eggs with a wooden spoon does not make any difference for cooking tasty scrambled eggs, that caring for the environment is smart. On the ground of these evaluations, we get blamed or praised by others for our actions. The present study has been sparked by the intuition that something analogous is correct regarding acquisitions of belief. Acquisitions of belief seem to be proper objects of negative or positive evaluations, and it is at least occasionally appropriate to blame or praise people for the beliefs they have acquired.

THE INITIAL INTUITION Suppose that Henry is an experienced mountain guide, organizing ski excursions around Mont Blanc. This morning Henry is supposed to guide a group of four Italian tourists. While getting ready to leave, Henry is deliberating about the itinerary he wants to take. He has to make a choice between two routes. Henry would prefer to take the first route because it offers fantastic scenery. By contrast, the second route is really less impressive and will be much more boring for the Italian tourists. The only concern which holds him back from opting for the first route, is that it contains two difficult sections that only good skiers can manage. While he is trying to find out whether it is not too risky to take the first itinerary with the group of the four Italians, he remembers that the Italian tourists he met previously were, most of the time, really good skiers. On the ground of memory and without checking further, Henry acquires the belief that the four Italian are sufficiently good skiers and makes the decision to take the first route. When the group reaches the first difficult section of the circuit, a very steep slope, one of the tourists gets panicky. When Henry realizes the danger of the situation, he immediately makes his best efforts to reassure and help him but to no avail. While trying to go down the slope, the Italian tourist falls, tumbles all the way down, and finally breaks his left arm and two of his teeth. Most of us are, I suppose, inclined to blame Henry in these circumstances. We are tempted to say: “Henry is guilty for believing that they are sufficiently experienced, without having checked whether they really are as good at skiing as the other Italian tourists he has met”,

2

“He should not have believed that they are without verification”. What we reproach Henry for is neither his evaluation of the first route nor his choice to take the first route. As such and given what he sincerely believes, opting for the first, most interesting circuit seems to be an irreproachable decision. Our feeling is that Henry is blameworthy for having acquired the false belief that the four Italians are sufficiently good skiers. The sort of assessment at work in Henry’s case is very frequent. When someone’s action unintentionally brings undesirable and objectionable consequences, the pressing question, in order to establish his guilt, bears upon his right to ignore the possibility that such an action is undesirable or that such an action will have such displeasing consequences.1 Ignorance is very often offered as an excuse for a misdeed and, every time this is the case, the next question will concern the entitlement to ignore. To say it differently, the next question will bear upon the blameworthiness or innocence of the subject’s ignorance. It happens also that we evaluate someone’s acquisition of a belief negatively, without bearing in mind the potential or actual consequences of this belief acquisition. If my neighbour Ralph is convinced that eating raspberries on Tuesdays is unhealthy, I will probably consider his belief stupid or ridiculous. The negative assessment of his belief acquisition is not inevitably grounded on its potential or actual consequences. Even if my neighbour (unbeknownst to him) is allergic to raspberries and will never enjoy the pleasure of eating raspberries anyway, I will keep evaluating his belief acquisition negatively. Note that there are also cases in which we praise people for their belief acquisitions. We attribute prizes and rewards to scientists for their discoveries, we congratulate children for what they truly believe about snails’ reproduction, etc. These examples manifest our tendency to evaluate belief acquisitions and our inclination to assess people for what they believe, just as we

1 As explained in Schulthess (1991), the determination of the blameworthiness of an agent requires that we consider various forms of ignorance. We need to determine whether the agent is entitled to ignore the rules that he violates by performing his action, but also whether he is entitled to ignore the circumstances in which he acts just as the nature of the action he performs. For the sake of simplicity, I gather the last two notions under the label “ignorance of the consequences of the action” in this work.

3

undeniably evaluate actions and people for their performances of these actions.

MAIN OBJECTIVE The goal of this work is not to offer a theory determining the conditions under which a subject deserves to be praised rather than blamed for the belief he acquires.2 It is not to establish which specific virtues, norms or values should guide belief acquisitions. Answering this sort of question amounts, I believe, to elaborating an ethics of belief. 3 The objective of this work is rather to provide the theoretical framework in which it is possible to assess people for what they believe. It is to cast light on and, hopefully, answer the problems which need to be solved before that we can begin to elaborate an ethics of belief. As Henry’s example is supposed to show, we frequently feel the need to assess a subject’s right to ignore what he ignores or to believe what he believes. Preliminary clarificatory remarks Before presenting the two main problems that philosophers need to solve if they want to be able to assess people for their belief acquisitions, a few preliminary remarks are useful. First, it is important to note that the attitude of believing a proposition is a doxastic attitude and that there are other sorts of doxastic attitude: the attitude of disbelieving a proposition and the attitude of suspending one’s judgement regarding a proposition. The hypotheses, theses, conclusions, etc. of this work are mainly formulated in terms of the acquisition of beliefs without mentioning the other two kinds of doxastic attitude. This choice is dictated by the interests of brevity and simplicity. It does not reflect any priority given to beliefs, by contrast to one or the other kind of doxastic attitude. Everything which is asserted about belief acquisitions, is, mutatis mutandis, correct regarding the other kinds of doxastic attitude.

2 For the sake of simplicity, I do not always mention the third option, the possibility of being neither praiseworthy nor blameworthy for a belief acquisition. 3 The ethics of belief is very often considered a deontological theory, briefly said, as a theory presupposing the existence of norms of believing and implying the ascription of duties to believe. But an ethics of belief can also be grounded on a system of values or on a system of virtues and vices. The notions which help to differentiate various kinds of moral philosophy can be transferred to the doxastic realm.

4

There is a second point to make regarding the scope of this inquiry. As I have just said, believing a proposition is a doxastic attitude. Now attitudes are mental states that we can, among other things, acquire, maintain and revise. In this work, I assume that the acquisition, the maintenance and the revision of a belief are temporally related in the following way:

1. The acquisition of the belief that p, by a subject S, occurs at an instant t. This is the instant at which S precisely begins to be in the state of believing that p; 2. From then on, S might maintain himself in the state of believing that p. This is something S would do during a period of time that begins at t; 3. From then on, S might revise his doxastic attitude toward proposition p. He might stop believing p. This is something S would do at some instant t+n which is the time at which S would either acquire the disbelief that p or would start suspending his judgment toward p.

In brief, the acquisition of the belief that p is an event constituting the boundary from which we are in the state of believing that p. Once we are in the state of believing that p, it is possible to maintain ourselves in this doxastic state or to revise our doxastic situation. I content myself with speaking of acquisitions of belief, but most of what I will say is, mutatis mutandis, valid for maintenances and revisions of belief as well. Finally, I consider that the acquisition of the belief that p, like the maintenance or the revision of the belief that p, are states of affairs. More accurately, they are exemplifications, by a subject S, of the property of acquiring/maintaining/revising the belief that p at some instant/during a certain period of time. Such states of affairs are what the present study takes as its object.

TWO CENTRAL PROBLEMS The ascription of blameworthiness or praiseworthiness for belief acquisitions, i.e. the ascription of doxastic blameworthiness or praiseworthiness, raises two main problems, the problem of control and responsibility and the normative problem, with which I deal separately

5

in the two main parts of this work.4 The first part consists of chapters 1 to 4 while the second part includes chapters 5 to 7. The problem of control and responsibility The first problem bears upon the possibility of our being responsible for our belief acquisitions. It begins with the observation that it is incorrect to blame a person for what is not under his control because this is not something for which he is possibly responsible. I cannot be blamed for my cousin getting a cavity, for my aunt falling in love with her dentist, for the plane being late. At least, this is true as long as I do not exercise any control over these states of affairs; as long as they are not state of affairs for which I am possibly responsible. Now, it does not seem that we control our belief acquisitions as we control the performances of our actions. I cannot acquire the belief that I am 10 years younger just because I want to, as I can raise my arm if I want to ask a question at a lecture. When I acquire the belief that the weather is rainy today as a result of my drawing the curtains, the acquisition of this belief is something which merely happens to me. It does not seem to be something over which I exercise a form of control. Hence, we could be tempted to conclude that any attempt to elaborate an ethics of belief is flawed from the beginning in virtue of the impossibility of being responsible for our belief acquisitions. The reasoning would be the following:

1. The ethics of belief is essentially a theory specifying when we are praiseworthy rather than blameworthy for our belief acquisitions or conversely; 2. It is incorrect to praise or to blame a subject for the occurrence of a state of affairs if he is not responsible for the occurrence of this state of affairs; 3. Our being responsible for the occurrence of a state of affairs requires that we control the occurrence of this state of affairs; 4. We do not control the occurrence of the states of affairs consisting in our belief acquisitions; Then 5. The elaboration of an ethics of belief is impossible.

Obviously we shall have to discuss step 4. in full details to avoid the conclusion 5. above. 4 Engel already emphasizes the importance of distinguishing between these two problems in Engel (2001a).

6

The normative problem Besides the possibility of our belief acquisitions’ being under our control, the second main problem to tackle, in order to account for the possibility of blaming or praising people for their belief acquisitions, is the normative problem. The normative problem bears upon the identification of the properties which turn beliefs into, roughly speaking, “good” or “bad” things to acquire.5 If there is no property such that a belief possessing this property is either good or bad, there is no ground for evaluating belief acquisitions and, in turn, there will, obviously, be no possibility of blaming or praising someone for his belief acquisition. Fortunately, among the properties characterizing beliefs some of them clearly seem to influence their general value or desirability. The most obvious candidate is the property of being true. Acquiring true beliefs seems better than acquiring false beliefs. The acquisition of rational beliefs and the acquisition of well grounded beliefs also seem desirable. The main purpose of the second part of this work is to put some order in this diversity. As we will see, the end of acquiring true beliefs plays a central role when it is a matter of epistemic evaluation. But belief acquisitions can be assessed along non-epistemic lines as well. The acquisition of a belief might be good in virtue of its intrinsic pleasantness, its happy consequences, its being the result of a very elegant method, etc. In these cases, the desirability of the belief acquisition has not necessarily something to do with the desirability of acquiring true beliefs rather than false ones. The last chapter of this work is devoted to clarify the connection between the epistemic and the non-epistemic realm of evaluation.

ABSTRACTS OF THE CHAPTERS Chapter 1: What the philosophy of action teaches us The first chapter of this work constitutes an attempt to clarify the notion of “control” by appealing to the philosophy of action. The starting point is that the form of control of belief that epistemologists are tracking in

5 In this work, I presume that such things as normative properties of beliefs exist independently of our minds, that the goodness or the badness of a belief does not exclusively consist in its being the object of a positive or a negative attitude. I discuss the notion of “normative property” in more detail in the chapter 5 of this work.

7

the doxastic area is the one exercised when we perform movements for reasons, i.e. when we perform full-blooded actions, by contrast to mere activities. To control one’s belief acquisition in the relevant way is to acquire this belief for a reason. Therefore, the first chapter is mainly devoted to improving our understanding of the necessary conditions under which one of my movements qualifies as a movement performed for a reason, with the underlying idea that the same conditions apply to belief acquisitions. As I have just said, the exercise of a form of control of belief acquisitions is a necessary condition for the attribution of responsibility, which is itself required by the ascription of doxastic praise or doxastic blame. Hence, doxastic praiseworthiness and doxastic blameworthiness necessitates our being able to acquire beliefs for reasons. To say it differently, the specification of the notion of control in terms of reasons drives me to the following formulation of the classical problem of doxastic voluntarism:

1. It is incorrect to praise or to blame a subject for one of his belief acquisitions if he cannot be responsible for his belief acquisitions; 2. Our being responsible for our belief acquisitions requires that we control our belief acquisitions; 3. What we control is our actions, i.e. the movements we perform for reasons, by contrast to what happens to us and by contrast to our mere activities; 4. Beliefs cannot be acquired for reasons;

Then

8

5. It is incorrect to praise or to blame anyone for one of his belief acquisitions.6

From there, what is at stake in the next chapters of this work is the correctness of the fourth premise. Is it really impossible to acquire beliefs for reasons while it is possible to perform movements for reasons? Two important distinctions are made at the end of the first chapter. First, I note that our motivation to acquire beliefs is twofold. We frequently desire to acquire true beliefs about a topic but we also sometimes desire to acquire pleasant, reassuring, or morally good beliefs. That is, unlike our reasons to perform movements, our reasons to acquire beliefs can be epistemic and/or non-epistemic. Consequently, we should reformulate the question under scrutiny in the two following chapters:

Is it really impossible to acquire beliefs for epistemic or non-epistemic reasons, as it is possible to perform movements for reasons?

Second, I distinguish between the evidence supporting the truth of a proposition and a motivating epistemic reason to acquire a true belief

6 Compare to the following more traditional formulation of the problem of doxastic voluntarism: 1. If we have any epistemic obligations, then doxastic attitudes must sometimes be under our voluntary control; 2. Doxastic attitudes are never under our voluntary control; 3. We do not have any epistemic obligations. See Feldman (2000), Sharon (2003), Weatherson (2008). First, traditional formulations of the problem of doxastic voluntarism generally include deontological concepts. My formulation is more general. It raises the question of our being possibly assessed for our belief acquisitions. It does not specify whether we deserve to be praised/blamed because we satisfy/violate a duty, because we acquired a valuable/disvaluable belief or because we have been virtuous/vicious. Second, traditional formulations of the problem of doxastic voluntarism include more explicit reference to the “ought implies can” principle. In the traditional formulation just mentioned, the “ought implies can” is underlying premise 1. Facing such a traditional formulation, epistemologists who want to find a solution to the problem of doxastic voluntarism have the choice to argue against the “ought implies can” principle (see Feldman 2000, Ryan 2003) and/or to argue against the impossibility of controlling belief acquisitions (Ryan 2003, Steup 2000, etc.) This alternative does not arise with my own formulation since it does not clearly rely on the “ought implies can” principle.

9

about a topic. To have evidence is not to have a motivating reason. Indeed, to have evidence is to be in a state characterized by the world to mind direction of fit and being in such a state is not sufficient to motivate to act. Chapter 2: The impossibility of acquiring beliefs directly for reasons The main goal of the second chapter is to defend the idea that beliefs cannot be acquired directly for reasons. As we will see, a helpful way of formulating the same idea is to say that belief acquisitions cannot be assimilated to basic actions, i.e. bodily movements performed for reasons. First, I try to show that Williams’ argument against the possibility of believing at will misses its target but, also, that his argument can be amended in a way such that it finally succeeds in proving that we cannot acquire beliefs directly for non-epistemic reasons. The second objective of this chapter is to stress the importance of the following premise of Williams’ argument: it is impossible for a subject S to consider both that the mental state that he acquired is a belief and that he acquired it for a non-epistemic reason. Not only does this premise capture what is distinctive of beliefs compared to other sorts of mental states but it also provides an appropriate account of the feature of doxastic deliberation called “transparency”. Chapter 3: Pascalian and theoretical control One of the conclusions of the second chapter of this work is that belief acquisitions cannot be categorised with basic actions for the reason that we cannot acquire beliefs directly for reasons while basic actions, i.e. bodily movements can be performed directly for reasons. But this clearly leaves the door open to the possibility that belief acquisitions are non-basic actions, that is to say, the possibility that we can acquire beliefs indirectly, i.e. by performing a more basic action. The suggestion that we are able to exercise such an indirect form of control over our belief acquisitions is never really questioned. It is indeed very plausible that we possess such an ability. My goal in the chapter 3 is to develop this proposal. As I said, our motivations for belief acquisitions might be epistemic and/or non-epistemic. Accordingly, I believe that our indirect control over belief acquisitions can take two forms:

10

1. We exercise theoretical control over the acquisition of a belief if and only if we acquire this belief by performing one or several more basic action(s) consisting in a truth-oriented modification of our current set of evidence and our performance of this (these) more basic action(s) is motivated by an epistemic reason, i.e. a desire to acquire a true belief about a topic; 2. We exercise Pascalian control over the acquisition of a belief if and only if we acquire this belief by performing one or several more basic action(s) consisting in a manipulative modification of our current set of evidence and our performance of this (these) more basic action(s) is motivated by a non-epistemic reason, i.e. a desire to acquire a pleasant, useful, reassuring belief, etc.

The rest of this chapter is principally devoted to answering Alston’s argument7 against the relevance of these two kinds of control for the ascription of doxastic responsibility. This mainly leads me to elucidate the nature of theoretical control. When I exercise theoretical control —for instance, when I exercise theoretical control over the acquisition of the belief that there is no milk left in the fridge, by opening the door of the fridge— I exercise control over the acquisition of a determinable belief, the content of which regards my stock of milk. But I do not exercise control over the acquisition of the determinate belief that there is no milk left. By appealing to an example in which a subject is responsible for his action even if he only controls a determinable form if it, I conclude that theoretical control is a fully agentive sort of control. It is perfectly able to ground the ascription of responsibility for scientific discoveries or for the sort of doxastic negligence displayed by Henry in the example above. Finally, I try to reply to both Ryan’s and Steup’s suggestions to the effect that doxastic control is not restricted to the indirect form I favour. The common difficulty of their two arguments, I claim, is that they rely on a confusion surrounding the notion of “a subject’s reason to believe”; more accurately, on the confusion between a subject’s evidence

7 See Alston (1988), Alston (2006).

11

supporting the truth of proposition p and his motivating reason to acquire a true belief about a topic. Chapter 4: Doxastic responsibility as responsibility for consequences The lesson of chapters 2 and 3 is that we control our belief acquisitions only indirectly. We control our belief acquisitions as we control our non-basic actions of cooking a boeuf bourguignon, or losing weight, that is to say, actions that we perform by performing one or several more basic actions. The main result of the last two chapters is that we should conceive doxastic responsibility on the model of responsibility for non-basic actions. Now to be responsible for a non-basic action is to be responsible for the consequence of one or several basic actions. For instance, to be responsible for the non-basic action of killing my neighbour’s cat by driving over it is to be responsible for the occurrence of the state of affairs consisting in the death of my neighbour’s cat, which is the consequence of several of my basic actions. Analogously, to be responsible for, let me say, the acquisition of the belief that I am a reliable friend is to be responsible for the resultant acquisition8 of the belief that I am a reliable friend, which is the consequence of my performing one or several basic actions consisting in the modification of my current set of evidence. The goal of chapter 4 is to provide a compatibilist account of the conditions under which a subject S is responsible for his resultant belief acquisition by trying to apply a reasons-responsiveness model of responsibility for consequences to resultant belief acquisitions.

8 The expression “the acquisition of the belief that p” might refer to the complex state of affairs which consists in my actions of manipulating my current set of evidence causing the instantaneous acquisition of the belief that p. But it might also precisely refer to the instantaneous acquisition of the belief that p, which is entailed in this complex state of affairs and which delineates its occurrence. From chapter 4 on, I distinguish, when necessary, between the two meanings of the expression “the acquisition of the belief that p” by speaking of “the resultant acquisition of the belief that p” to refer to the simple state of affairs. From chapter 4 on, the expression “the acquisition of the belief that p” refers to the more complex and causal state of affairs.

12

Chapter 5: Epistemic praiseworthiness and epistemic blameworthiness If I am right to think that we are sometimes responsible for our belief acquisitions, then it is certainly possible to be either praiseworthy or blameworthy for our belief acquisitions. In chapter 5, I consider different sorts of epistemic praiseworthiness and blameworthiness. They correspond to the various ways in which belief acquisitions can be epistemically good things, that it to say, good things in regard to the achievement of the final epistemic goal of acquiring true beliefs. In more detail, I distinguish between four sorts of epistemic goodness characterizing belief acquisitions:

1. Final goodness: the goodness characterizing the acquisition of a true belief; 2. Instrumental goodness: the goodness characterizing the acquisition of a belief which causes the acquisition of a true belief;

3. Goodness of rationality: the goodness characterizing the acquisition of a belief which is finally good and/or instrumentally good according to what the subject who acquires this belief thinks at the time he acquires it;

4. Goodness of commendability: the goodness characterizing the acquisition of a belief which results from the believer trying hard enough to acquire a belief which is finally good and/or instrumentally good.

The first conclusion of this chapter is that it is impossible to be responsible for our belief acquisitions being either rational or irrational. Hence, it is inappropriate either to praise or to blame someone for his rational or irrational belief acquisition. The second conclusion is that, I cannot, let me say, be praised for my finally or instrumentally valuable acquisition of a belief without deserving to be praised for my commendable acquisition of this belief. But the contrary is not true. If I am praised for having acquired a belief by trying hard enough, I am not necessarily to be praised for having acquired a finally or instrumentally valuable belief.

13

Chapter 6: Beyond epistemic justifiedness Chapter 6 relates the previous considerations with three epistemological accounts of the justification of belief: the externalist account of justification, the mentalist account of justification and the accessibilist account of justification. These are traditionally considered as antagonist and incompatible views. The first goal of this chapter is to provide some further support for the claim, defended mainly by Alston9, according to which epistemologists who disagree about the notion of justifiedness of beliefs do not have a common pre-theoretical conception of what they are debating. My project is to bring support to this view by showing that the various conceptions of justifiedness are actually reflecting the distinct sorts of goodness presented in chapter 5. The starting point of this project is that externalists, mentalists and accessibilists do, at least, agree on the fact that a justified belief acquisition is a belief acquisition which is good in a way. From there, an important part of chapter 6 is devoted to showing that externalists pre-theoretically conceive justifiedness as instrumental goodness, that mentalists pre-theoretically conceive justified belief acquisitions as “good because rational” belief acquisitions and that accessibilists pre-theoretically conceive justifiedness as the goodness of commendability. In the last part of chapter 6, I discuss what I call the problem of the explanation of the goodness of justifiedness and consider two influential explanations: the reliabilist explanation and the so-called credit explanation. This leads me to suggest an accessibilist explanation of the goodness of justifiedness, i.e. an explanation grounded on what I previously described as the accessibilist conception of the goodness of justifiedness. Finally I try to show that an accessibilist explanation of the goodness of justifiedness does not encounter the objection which is traditionally addressed against the credit explanation. The accessibilist explanation is able to account for our being sometimes justified when we believe the testimony of others.

9 See Alston (1985), (1993) and (2006).

14

Chapter 7: Epistemic justifiedness and non-epistemic justifiedness In chapters 5 and 6, my attention is focussed on the various kinds of epistemic goodness which can characterize belief acquisitions, on the various ways in which belief acquisitions can be epistemically justified or unjustified, while leaving aside the non-epistemic form of justifiedness or unjustifiedness that belief acquisitions seem to display as well. The last chapter of this work is devoted to the consideration of the divergence thesis, i.e. the claim according to which it is possible for a belief acquisition to be simultaneously epistemically justified and non-epistemically unjustified and possible for a (another) belief acquisition to be simultaneously epistemically unjustified and non-epistemically justified. The most immediate way of refuting the divergence thesis is to claim that the acquisition of true beliefs is not finally desirable, that it is only instrumentally and non-epistemically desirable, i.e. desirable in virtue of the non-epistemic goal it helps to achieve. I do not hope to settle the question of the final desirability of the goal of acquiring true beliefs. I simply try to emphasize that, beyond their discrepancies, Clifford’s and James’ ethics of belief both consider the goal of acquiring true beliefs as being merely instrumentally and non-epistemically desirable. I conclude this chapter by emphasizing that the divergence thesis is, at least, clearly wrong if we suppose that a justified belief acquisition is nothing other than a rational belief acquisition.

THE PROBLEM OF CONTROL AND RESPONSIBILITY

17

CHAPTER 1: WHAT THE PHILOSOPHY OF ACTION TEACHES US

We saw in the introduction that the project of elaborating an ethics of belief is tied to the possibility of ascribing responsibility for the acquisition of belief, which I call doxastic responsibility. The ascription of responsibility I have in mind does not merely consist in a straightforward ascription of causality.10 When I say that a tempest is responsible for the tree’s being uprooted, I merely say that some causal relationship holds between the occurrences of two states of affairs. Causality is ascribable to various kinds of metaphysical entities: processes like tempests, objects, etc. The responsibility that I am tracking in the epistemic domain is only ascribable to persons. It is the sort of responsibility that we, sometimes, have for our actions; the sort of responsibility that presupposes the ascription of causal agency. Hence, the fundamental condition that the acquisition of belief has to satisfy, in order to be something for which we can be held responsible, is being something that we do. To be clear, what is required is not that every single acquisition of belief could be considered as an action but that, at least, some of them could.

ACTIONS AND HAPPENINGS There is a fundamental and much-debated difference between my actions and my movements when things happen to my body, e.g. between my action of getting up in the morning and opening a door, and a mere bodily happening like my stumbling over a root while walking in the forest. This difference has direct influence on the attribution of responsibility. I am never responsible for my movements when it is something that happens to my body and not something that I do. I am not responsible for being dazzled when I go out of the cinema, for getting stuck in the lift, etc. Analogously, acquisitions of belief won’t be something for which we are sometimes responsible if they always are things that happen to us and not things that we do. Addressing the question of the conditions under which someone is responsible for the acquisition of a belief requires first showing that

10 I take up here Feinberg’s distinction between “straightforward ascriptions of causality” and “ascriptions of causal agency”. See Feinberg (1970), pp. 130-133.

18

belief acquisitions are able to fulfil the condition which differentiates actions from mere happenings. Before focussing on this issue, my main goal in this chapter is to offer an overview of some theories of action with the purpose of finally drawing out a joint opinion regarding the features that a movement needs to display in order to qualify as an action.11 Many philosophers of action agree about, at least, some necessary conditions. These are the common assumptions on which I would like to rely in order to investigate responsibility for belief acquisition. Understanding the very nature of action is a notoriously difficult task, which requires much more than the formulation of a couple of necessary conditions. I would obviously have to take account of many refinements if my ambition were to take this issue forward. However, this would be beyond the scope of this chapter, the aim of which is merely to clear the ground for the understanding of responsibility for belief. I also think that such refinements would have no crucial influence on my overall results. Non-reductionist conception of action As I said, the division between the two categories of actions and bodily happenings has been much discussed and has given birth to several different conceptions. In the following pages, I would like to present the outlines of the two most influential conceptions. The first kind is the non-reductionist conception of action.12 Its main objective is to defend the idea that actions do not constitute a specific sort of bodily happenings but constitute a sui generis category of entities. The difficulty of the task appears once we notice that the truth-makers of the propositions referring to happenings and the propositions referring to actions both involve the occurrence of a state of affairs, which is a happening. Not only the truth of the proposition: “a gun’s being fired takes place” but also the truth of the proposition: “I fire the gun” require

11 To my knowledge, few epistemologists questioning the possible existence of doxastic responsibility really pay attention to the important amount of philosophical work devoted to the distinction between happenings and actions. I believe the writings of philosophers of action can help delineate accurately the ins and the outs of this question. 12 The following paragraphs are largely inspired by the very enlightening account of the non-reductionist conception of action offered by Moya. See Moya (1990), pp. 18-29.

19

the occurrence of the state of affairs consisting in the gun’s being fired. It is not true that I fire the gun if the gun’s being fired does not take place. Consequently, if something radically distinguishes the category of actions from the category of happenings, it must lie in the way the gun’s being fired is brought about. The action of firing the gun occurs when the gun’s being fired is brought about in an agentive way while the happening of the gun’s being fired takes place when the gun’s being fired is not brought about in this specific way. Hence, the genuine challenge for the supporters of the non-reductionist conception is to provide a non-reducible account of the agentive way of bringing about the gun’s being fired. It is not possible simply to claim that the bringing about of the gun’s being fired is an action when the gun’s being fired is brought about by an agent. Indeed, the gun’s being fired can be brought about by an agent without being an action. If, for instance, I cause the gun’s being fired by stumbling on a root, I did not perform the action of firing the gun. Alternatively, we could be tempted to say that the bringing about of the gun’s being fired amounts to the action of firing a gun only if I bring about the gun’s firing by performing an action, for instance, the action of pulling the trigger. The problem is that the action of pulling the trigger can be further analysed as the bringing about of a happening as well, e.g. as the bringing about of the trigger’s being pulled. What is distinctively agentive in my action of pulling the trigger has to be found in the way I bring about the trigger’s being pulled. Therefore, it is not sufficient to appeal to the action of pulling the trigger in the analysis of “I bring about the gun’s being fired” in order to account for what is specifically agentive in the bringing about of the gun’s being fired. Obviously the same difficulty will recur if we try to analyse “I bring about the trigger’s being pulled” by appealing to another more basic action. “I bring about the trigger’s being pulled” can probably be analysed in terms of “I bring about the trigger’s being pulled by moving my finger”. But, once again, an action of moving my finger can be identified to a bringing about of the movement of my finger in a specific way. Then, the appeal to the action of moving my finger is not sufficient to point to what is distinctively agentive in the bringing about of the trigger’s being pulled.

20

This is where basic actions enter into the picture. Here is the following general definition by Danto:

Basic actions are actions which are not caused to happen by the man who performs them.13

As this definition makes clear, basic actions are the most fundamental actions by which I perform other less basic actions, like bodily movements. They are not themselves induced by my performance of something else. Basic actions are conations, that is, volitions, willings or tryings (the defenders of the non-reductionist conception of action do not agree on the exact nature of these conations14). Trying or willing to move my finger is a basic action, for the most fundamental action that I perform, when I move my finger, consists in trying or in willing to move my finger. The appeal to basic actions is supposed, then, to bring to an end the regress just presented. Even if it is true that my action of trying to move my finger also implies the occurrence of a happening —a trying to move my finger—, such a happening cannot be treated in the same way as the gun’s being fired, the trigger’s being pulled, the movement of my finger. The difference is the following: when the movement of my finger is brought about, I am not necessarily performing the action of moving my finger. In contrast, when a trying to move my finger is brought about, I am inevitably also trying to move my finger. A trying to move my finger is not something that can be brought about in a non-agentive way.15 Finally, supporters of the non-reductionist conception of action claim that what distinguishes the bringing about of the gun’s being fired when it is an action from the bringing about of the gun’s being fired when it is a happening lies in the fact that I perform the former but not the latter by performing a basic action. “Basic actions are, so to speak, the source of agency; they transmit agency to other things we do.”16 13 See Danto (1965), pp. 141-142. 14 See Moya (1990), pp. 19-29. 15 Moya addresses the following objection to the non-reductionist solution: basic actions, tryings, willings etc., are either mental or physical states of affairs of our minds. If they are mental states of affairs, we face the problem of interactionism: how can purely mental acts cause physical happenings? If they are physical events, then, it is not true that they cannot by reductively analysed in terms of the bringing about of a mere happening. If they are physical states of affairs, the appeal to basic actions does not help stop the regress. See Moya (1990). 16 Moya (1990), p. 14.

21

Moya’s non-reductionism Carlos Moya’s account of the distinction between actions and mere happenings is a non-reductionist one. However, Moya’s criterion of agency does not resort to basic actions but to meaningful actions. Examples of meaningful actions are signalling for a turn, making an offer, holding a lecture. The crucial thing to note about meaningful actions is that they cannot be reductively analysed as non-agentive bringing about of happenings. For the proposition “I make an offer” to be true, an offer’s being made has to take place. But the bringing about of an offer’s being made is essentially agentive in the sense that, inevitably, when an offer’s being made takes place, someone makes an offer. Similarly, it is not possible for the state of affairs consisting in a turn’s being signalled to occur without someone performing the action of signalling for a turn. According to Moya, meaningful actions “seem to be pure actions”17 and, therefore, constitute “good starting points to analyse agency. The conditions that allow the performance of such actions will provide a good insight into the nature of agency”18. Here is the condition Moya has in mind:

“Pure actions are such in virtue of there not being the possibility that they be unintentional, or, in other words, in virtue of being necessarily intentional. There is no such thing as greeting, signalling for a turn unintentionally. To do it intentionally is a necessary condition of greeting, signalling for a turn or marrying but not of shooting a gun, killing someone or scoring a goal.”19

When he claims that pure actions are necessarily intentional, what Moya means, more accurately, following the Davidsonian tradition, is the following: when an action is described as a pure action, it is necessarily

17 Moya (1990), p. 39. 18 Moya (1990), p. 40. Despite the crucial discrepancy between Davidson and Moya’s approach, Davidson has a similar idea in mind when he claims: “Perhaps, then, being intentional is the relevant distinguishing mark. If it were, it would explain why some verbs imply agency, for some verbs describe actions that cannot be anything but intentional; asserting, cheating, taking a square root, and lying are examples”, Davidson (1980), p. 44-5. 19 Ibid. p. 52.

22

intentional under this description. That is, the action of firing a gun is not a pure action since “I fired the gun” can describe a piece of behaviour which is intentional or non-intentional. This finally leads Moya to formulate the following criterion of agency:

“A certain behaviour is an action, if and only if it is intentional (or intentionally performed) under some description”.20

Reductionist conception of action Moya’s criterion is clearly very similar to Davidson’s:

“A person is the agent of an event if and only if there is a description of what he did that makes true a sentence that says he did it intentionally”.21

The crucial difference between Moya and Davidson’s criteria lies in their respective conception of intentionality. Davidson’s view of

20 Moya (1990), p. 53. Many of what we consider to be actions cannot be described as intentional in Moya’s sense. While listening to a philosophical talk, I scratch my head or play with a pen. These are clearly actions but they are not the result of a commitment to perform them. In normal circumstances, when I play with my pen, I did not previously intend to play with my pen. Therefore, Moya distinguishes between two kinds of intentionality. When I spontaneously scratch my head while listening to a talk, my action only reflects an immediate intention in the sense that my scratching my head does not serve future goals. By contrast, a full-blooded intentional action —the action of signalling for a turn, the intentional action of going to the marketplace— serves future goals: so my intention to go back home, my intention to have fresh fruits for dinner, etc. To summarize, Moya distinguishes between two kinds of actions: what he calls “purposive behaviour” that only reflects an immediate intention and the full-blooded action that results from a more complex intention that also serves future goals. See Moya (1990), pp. 57-60. 21 Davidson (1980), p. 46.

23

intentionality supports a reductionist conception of action while Moya’s conception of intentionality is an attempt to defend non-reductionism.22 According to a reductionist conception23, an action is just a happening with a specific sort of cause. Now there are clearly various ways of spelling out this thesis depending which sort of cause you identify as the relevant one. According to the Davidsonian account of action, for instance, “we can properly say that they [beliefs and desires, my comment] are causes of intentional actions, and when we say this we draw upon the concept of ordinary event causality.”24. That is to say, the bringing about of the gun’s being fired is an action of firing a gun only if the cause of this happening is the agent’s desire and belief. Let me have a brief look at two famous difficulties encountered by the Davidsonian conception of action. First, consider the following case:

George is about to give a talk about the possible extinction of platypuses. George desires to win the sympathy of his audience and believes that crying will suffice to satisfy his desire. Unfortunately for him, he is unable to cry at will. Out of frustration, he bursts into tears.

22 Moya’s views regarding intentionality are not crystal-clear. I will try to offer a brief account below. Moya’s main claim is that intentions are commitments. Intentions are commitments to do things in the same way that signalling for a turn is a commitment to turn. Intentions are not similar to plans, desires or aims, because the latter do not involve a commitment. Planning to go to the market place is not like intending to go to the marketplace in this respect. Intending to go to the marketplace commits myself to performing this action, as signalling for a turn commits myself to turning. As a result of being conceived as a commitment, intention is also a normative notion. Its normativity appears once you note that there is something wrong if someone who commits himself to turning by signalling for a turn does not turn. By contrast, I can plan to perform an action and finally perform another without there being anything wrong. Finally, the identification of intentions with commitments protects them from being reductively conceived in terms of mere happenings. Moya claims that a reductive conception in terms of mere happenings cannot possibly account for the normative feature of intentionality. 23 Traditional reductionist conceptions of action are mainly Armstrong’s and Davidson’s. See Armstrong (1973) and Davidson (1980). 24 Davison (1980), p. 49, n. 7.

24

His bursting into tears does not seem to count as something he does. Nevertheless, it has been caused by the combination of his belief and desire.25 A conceivable reply consists in claiming that, in such a case, the agent’s desire and belief do not manifest their characteristic causal powers in the sense that his belief only accidentally produces what satisfies his desire. Indeed, it is matter of coincidence that the mechanism of frustration produces a reaction that is able to satisfy the agent’s desire according to the agent’s belief. To say it differently, a possible way out consists in narrowing the scope of the Davidsonian model to pieces of behaviour that are not accidentally caused by beliefs and desires.26 Even if this —quite ad hoc— solution is accepted as such,27 it is insufficient to save the Davidsonian conception of action. Indeed, the Davidsonian model faces a second difficulty, which is of greater interest for some of my future considerations. There are cases —cases of Freudian slips mostly— in which a piece of behaviour is non-accidentally caused by a belief and a desire but does not seem to qualify as a full-blooded action.28 Consider the following example:

While he is about to sleep, James believes that he has a professional interview the morning after. He also believes that he needs to get up at 8 am at the latest to arrive on time. Therefore, he thinks that he had better switch on his alarm clock. As a result of his suspicion that this kind of job won’t suit him and his unconscious reluctance to get the job, he unconsciously set his alarm clock one hour too late.

It sounds quite counterintuitive to claim that James setting his alarm clock one hour too late is a full-blooded action. After all James did not 25 This is, slightly modified, Velleman’s example. See Velleman (2000a), pp. 7-8. 26 See Velleman (2000a), p. 8. Note that Davidson does not believe in this solution: “beliefs and desires that would rationalize an action if they caused it in the right way —through a course of practical reasoning as we might try saying— may cause it in other ways. If so the action was not performed with the intention that we could have read off from the attitudes that caused it. What I despair of spelling out is the way in which attitudes must cause actions if they are to rationalize the action”. Davidson (1980), p. 79. 27 For a more sophisticated solution to the problem of causal deviant chains, see, for instance, Bishop (1990). I thank Pascal Engel for this reference. 28 Velleman (2000a), pp. 8-9.

25

want to get up too late. Nevertheless, his piece of behaviour is the result of his reluctance to get the job combined with his belief that if he is late, he won’t obtain it. Moreover, his piece of behaviour does not accidentally match his belief and desire. The second difficulty of the Davidsonian conception of action points to a distinction between mere activities and full-blooded actions. One of its main flaws lies in its inability to account for this distinction. The Davidsonian conception of action is “a model of activity but not of action”.29 I will say more about the distinction between activities and actions —which, I believe, plays an important role in an accurate understanding of doxastic responsibility— just below. Before this, I would like to take a step back and try to cash out the two lessons jointly taught by non-reductionism and reductionism. The first one is quite obvious. An action, as opposed to a mere happening, necessarily results from a conation: either a basic action, according to the non-reductionist view, or a conative state, i.e. a specific sort of mental state, according to the reductionist view. As I use it, the concept of conation or conative states is the broadest possible. It refers to a class of states of affairs (actions or events) —like the willing to perform action A, the desire to perform action A, an urge to perform action A— which are internal to the mind of the agent, have the world-to-mind direction of fit, and causally lead to the performance by this agent of the relevant bodily movement in normal circumstances. That is, conations or conative states are states which causally explain the performance by an agent of his bodily movement. The notion of conation/conative state needs to be distinguished from the notion of a reason to act. As we saw with the case of James, people can be caused to act by a desire to act in a certain way, without having a reason to act in this way. That is, a bodily movement, which causally results from a desire to perform this movement, is not necessarily performed for a reason.30 I leave aside the notion of reason for the moment and come back to the first lesson commonly taught by distinct theories of action:

29 Velleman (2000a), p. 10. 30 I do not want to deny that having a reason to perform action A involves having a desire, a motivation to perform action A. My only claim here is that there is, at least, a conceptual difference between my reason to perform an action and what causes me to perform this action.

26

For my movement to qualify as an action, as opposed to a mere happening, it has to be the causal result of one of my conative states, for instance a desire.31

Clearly, this is a very general lesson. And, as I said, generality is actually what I am pursuing. Indeed, I wish to ground the upcoming discussion of doxastic responsibility on some shared conception of action to avoid ending up with results which would depend on a specific theory of action. The second lesson regards the distinction between basic and non-basic actions. As we saw, the existence of basic actions, which are conations, plays a crucial role for many non-reductionist conceptions of action. Basic actions are called “basic” because they are actions that we cannot perform by performing a more basic action. The non-reductionist appeal to basic actions has its origin in the iterative way in which some actions seem to be connected.

Performing action A consists in making a state of affairs E happen by performing action A’ which itself consists in making a state of affairs E’ happens by performing action A’’, etc.

Most philosophers of action —supporters of the reductionist conceptions of action included— agree with this model and with the general idea that there are actions which are more basic than others. In the example above, A’’ is more basic than A’ which itself is more basic than A’. To make myself clear, I do not mean that every philosopher of action accepts the idea that there are basic actions. Obviously reductionists do not think so. I suggest rather, then, that philosophers of action: 1. jointly accept that a relation of “being more/less basic than” connects actions forming a chain and also that 2. they share the same metaphysical conception of action: a (non-basic) action consisting in the bringing about of a state of affairs.

31 In what follow I often use the term “desire” in its technical, philosophical sense, as synonymous to “conative state”, “pro-attitude”. According to this technical sense, every action is the causal result of a desire.

27

ACTIONS, HAPPENINGS AND ACTIVITIES Until now, I have contented myself with differentiating between actions and happenings. Examples of Freudian slips, like the one in which James is involved, seem to call for a refinement of the distinction. As I said, James’ unconsciously setting his alarm clock one hour too late does not qualify as a full-blooded action but it is not a mere bodily happening either. James’ movement seems to possess an intermediate status. Roughly, the reason why we are tempted to categorize James’ movement in such an intermediate way is that his setting his alarm clock one hour too late was only caused by his more or less conscious reluctance to get this job. It is not also something that James performs for a reason. Then, as a first approximation of the distinction between activities and actions, let me say that:

An activity is a bodily movement of an agent S, which is caused by one of his conative states without S’s performing this movement for a reason.

By contrast: A full-blooded action is a bodily movement of S that is caused by one of his conative states but which is also a bodily movement that S performs for a reason.32

The formulation of the conditions under which an action is performed for a reason is a highly delicate task. I will discuss it in more detail below. Two remarks have to be made beforehand. First, note that Freudian slips are not the only kind of case which seem to display the intermediate status of activities. When I distractedly play with my pen while listening to a philosophical talk, when I scratch my head while playing chess, when I cross my legs as soon as I sit down, these are all really movements I do, most of the time, without having a reason to perform them. They do not seem to qualify as full-blooded actions, in the way that my movement of raising my finger when I have a question, my movement of turning down the volume of the radio, etc., do. The category of activities not only includes Freudian slips, but also 32 This is Velleman’s terminology. See Velleman (2000a), p. 10. Velleman explicitly takes up Frankfurt’s distinction between two ways of being active: “To drum one’s finger on a table, altogether idly and inattentively, is surely not a case of passivity: the movements in question do not occur without one’s making them. Neither is it an instance of action, however, but only of being active.” Frankfurt (1988), p. 58.

28

movements performed out of habit, and bodily tics, as long as they are not performed for a reason. My second remark is terminological. Henceforth, I will call the category of entities including actions and activities the category of purposive behaviour. What I regarded as the distinction between mere happenings and “actions” will now be considered as the distinction between happenings and pieces of purposive behaviour. Acting for reasons Trying to provide a comprehensive and systematic account of the conditions under which an action is performed for a reason would clearly take me too far from my main objective in this study. Hence, I content myself with presenting a few generally accepted elements of the issue, which are directly relevant for my purpose. Beyond many potential disagreements, some general and common ideas seem to guide the conception of philosophers interested in the topic of our reasons for acting. My goal now is to formulate these rather uncontroversial ideas and to show, finally, that they are sufficient for the purpose of distinguishing between actions and activities. First, it is commonly accepted that performing an action for a reason involves being motivated to perform that action. I do not want to discuss—at least not for the moment— the nature of such motivation. “Is the reason constituted by the combination of desire and cognitive consideration? Do desires inevitably play a role when I perform an action for a reason?” These are questions that I consider at the end of this chapter. Another shared view is that an action performed for a reason can be explained by referring to this reason. The reason, which motivates me to perform this action, also explains why I perform this action.33 More importantly for the rest of this work, another commonly accepted idea is that the reason, which explains why I perform a bodily

33 See of course Davidson (1980), Audi (1986), Velleman (2000a), Crisp (2006). The problem faced by philosophers who claim my reason to perform action A causally explains my performance of A is the one at which I hinted with the presentation of George’s example above. If my reason to perform A causally explained my performance of A, then, my performance of A needs to be causally related to my desire and my cognitive consideration in an appropriate way. An appropriate way is a way which avoids the difficulty raised by cases in which my desire is accidentally satisfied by the bodily movement it causes.

29

movement, is also something that I am able to mention as an explanation of why I am performing this bodily movement while I am performing it. A shorter way of formulating this idea is to say that I perform an action for reason r only if the reason, which motivates me to perform this action, is avowable.34 This is the case because, when I am moving my body for a reason, I am not only a well-placed observer of my behaviour being caused by one of my motives, I am acting in the light of the reason, so to say.35 “Avowability” is the feature which grounds the distinction between activities and full-blooded actions. In the case of an activity like a Freudian slip, the agent cannot provide, at least not immediately, a meaningful explanation by appealing to what motivates him to behave in this way. What I mean by “not immediately” is that, even if the agent eventually ends up with an explanation, he will have to seek further or to introspect to explain why he is performing such a bodily movement. To be clear, the condition of avowability has the following four features:

a. it does not require that I have the ability to believe that I am performing this bodily movement for a reason. Children perform bodily movements for reasons even if they do not have the concept of reason. b. it does not require either that I am presently conscious of the reason why I am performing an action while I am performing it. I may raise my arm for a reason while being buried in thought.36 This could be considered a bodily movement that I perform for a reason even if I am not simultaneously conscious that I am raising my arm because I want to be polite to my neighbours. When I perform an action for a reason, like my action of greeting my neighbour, I am disposed to become immediately conscious of my motives. The condition of

34 See Crisp (2006), p. 38. For a similar idea, see also Hieronymi (2008). According to Hieronymi, an action performed intentionally —an action performed for a reason— is an action for which I am answerable, that is to say, is an action for which I can be asked for reasons. As Hieronymi notifies, this is originally Anscombe’s idea. According to Anscombe, whenever a subject performs action A intentionally, that is for a reason, one can rightly ask him: “Why did you perform action A?”, where this question does not look for an explanation of what causes this subject to perform action A but for what was his reason for performing action A. See Anscombe (1957). 35 See Audi (1986), pp.517-518. 36 See Audi (1986), p. 520.

30

avowability only necessitates a disposition to provide an explanation of my action. c. it does not rule out that I can forget the reason why I performed this bodily movement after having performed it. The condition of avowability only necessitates me being able to explain why I am performing this bodily movement while I am performing it. d. it does not apply to actions which are not as basic as our bodily movements. As I will explain in more detail in chapter 3, it seems that a non-basic action, like the action of baking bread, can be performed for a reason even if the agent forgets the initial reason why he performs this action while he is performing it.

Here is what is, at least, necessarily37 the case when an agent S performs a bodily movement for a reason r:

1. The condition of motivation: S’s performance of this bodily movement is motivated by r; 2. The condition of explanation: the reason r explains why S performs this bodily movement; 3. The condition of avowability: more importantly, S is disposed to provide an immediate explanation of why he performs this bodily movement by mentioning the reason r which motivates him to perform it.

Now let us see how this can account for the intuitive distinction between activities and full-blooded actions. I mentioned two kinds of cases in which it seems that my behaviour qualifies as something less than a full-blooded action: Freudian slips and 37 These conditions are probably not sufficient. But they are, at least, sufficient for the purpose of distinguishing between action and mere activities. For instance, Audi considers that it is also necessary that I could have another motivation than the one I actually have. If nothing, not even the desire to avoid the destruction of humanity, is able to defeat the motivation I actually have, then it is doubtful that I act for a reason when I act as a result of this actual motivation. As Audi noticed, this has to be distinguished from the ability to behave differently when I have a certain motivation. It could be that my motivation to smoke is invincible in the sense that I cannot bring myself to act contrary to it. It is another question whether the possession of this motivation itself is unavoidable. See Audi (1986), pp. 533-34.

31

pieces of behaviour like my drumming with my finger on a table, my playing with a pen, etc., that are not mere reflexes but do not seem to be full-blooded actions either. I believe that, in these two kinds of cases, at least, one of the conditions above is not satisfied. While getting up the day after, James will certainly be astonished to see that he is one hour late and his surprise is certainly the manifestation of his inability to explain immediately why he set his alarm clock one hour too late. Regarding pieces of behaviour like my drumming with my fingers, it is not clear that these pieces of behaviour are unconsciously motivated like Freudian slips. If they are not motivated in such a way, they differ from a bodily movement performed for a reason in virtue of the fact that they do not satisfy condition 1. Even if they are motivated in such a way, it seems that I won’t be able to provide an explanation of my drumming with my finger by appealing to a reason. Suppose that someone interrupts you while you are drumming with your fingers and asks you why you are behaving like this. You could answer that this is something you enjoy, because you want to be able to explain your action. But the pleasure you draw from drumming with your finger is not the reason why you perform the action of drumming with your fingers, as the pleasure I draw from jogging is the reason why I go jogging regularly. The pleasure you draw from drumming with your fingers is not the consideration in the light of which you drum your fingers. It seems rather to go with the action of drumming with your fingers. It possibly causes the action of drumming with your fingers, but it does not explain why you perform this action. Let me summarise the last considerations. The third lesson that the philosophy of action teaches us is that we should prefer a threefold categorisation between actions, activities and mere happenings in order to account for the intuitive difference between two levels of “agency”. My action of instinctively scratching my head when I think about a problem does not seem to be as “agentive” as my action of going to the marketplace in order to buy some fresh fruit. Now the philosophy of action also provides us with a very simple way to account for the distinction between activities and full-blooded actions. Full-blooded actions are pieces of purposeful behaviour performed for reasons while mere activities are not. That is, for the performance of a

32

bodily movement to qualify as a full-blooded action, it needs, at least, to satisfy the three conditions mentioned above. Once again, this way of accounting for the distinction has the advantage of being sufficiently general to be compatible with many theories of action.38

THREE DISTINCTIONS ABOUT REASONS Motivating reasons vs. normative reasons A crucial distinction has to be made as soon as one tackles the notion of reasons to act.39 It is crucial to differentiate between:

1. The motivating reason to perform action A; 2. The normative reason to perform action A.40

There are two ways of using the notion of reason which address two different questions. Suppose that I am suddenly interrupted by my friend Emma running up into my office while I am working. I can ask myself two distinct questions regarding Emma’s behaviour. First I can wonder why Emma performs this action, what are the considerations in the light of which she performs this action. In this case, I try to determine what is her motivation for performing this action and I expect an answer referring to her motivating reason. For instance, Emma runs up into my office because she thinks the building is on fire and she wants to warn me. The second question I can ask myself is the following: does Emma have any reason to run into my office? This time my question is normative. What I am trying is settle is whether there is something which speaks in favour of Emma’s action. I expect an answer mentioning something which makes the action of running up into my office the appropriate, the

38 This is obviously compatible with the reductionist conception but also with the non-reductionist conception of action. It is clearly possible to defend the idea that conations, like tryings, are basic actions that cannot be reduced to mere happenings while saying that a subject’s action of shooting someone by trying to shoot him is a full-blooded action only if he performs the basic action of trying to shoot this person for a reason. That is to say, only if his trying to shoot is motivated by a desire to achieve an end, his trying is explained by this desire and his desire is also something that he is immediately disposed to mention as an explanation of why he tries to shoot this person. 39 I am very grateful to Pascal Engel for having pressed me to clarify this point. 40 This is Dancy’s terminology in Dancy (2002). For a similar distinction, see Crisp (2006), Mulligan (2009a) (2009b), Parfit (1997).

33

right action to perform. For instance, Emma is right to run up into my office because the building is in fire. Because there are different interrogations, the answers to these questions might be completely divergent. If Emma runs up into my office for the same motivating reason but I perfectly know that there is no fire, I can explain why she runs up into my office but I cannot find much to say in favour of what she does, I cannot defend her behaviour. Until now, I used the term “reasons” to refer to motivating reasons exclusively. For instance, when I put forward the conditions, which need to be satisfied when a piece of purposeful behaviour is performed for a reason, what I was obviously trying to mention were the conditions under which a piece of purposeful behaviour counts as an action performed for a motivating reason. In the rest of this work, I will adopt the habit of explicitly distinguishing between normative and motivating reason every time I am speaking of normative reasons. Internalism vs. externalism about reasons My goal here is obviously not to settle this highly complicated debate. Simply, I would like to give the broad outlines of the discussion in order to avoid any confusion between the internalism vs. externalism and the aforementioned conceptual distinction between motivating and normative reasons. Internalism holds that an agent A only has a good reason to Ω if, were A to know all the relevant facts, and deliberate rationally, A would be motivated to Ω. According to internalism, the set of normative reasons in Emma’s situation is limited by what Emma would be motivated to do. For instance, suppose Emma would not have been motivated to warn me if she had rather believed that there was a bomb in the building. Hence, according to internalism, there being a bomb in the building does not constitute a normative reason to warn me in Emma’s situation. Roughly, externalism about normative reasons is the view according to which the presence of a bomb in the building might be a normative reason for Emma to run up into my office and warn me even if she would not have been motivated to perform this action if she had if she had believed that there was a bomb in the building. According to externalists, what gives us normative reasons for acting are not “facts about our own motivation, but facts about our own or other

34

people's well-being, or facts about other things that are worth achieving”.41 Humean vs. anti-Humean conception of motivation As I briefly said above, another much debated issue concerns the exact nature of our motivating reasons. According to what is traditionally called the Humean theory, when I perform an action for a motivating reason, my motivating reason is constituted by a suitable combination of a cognitive consideration and an independent desire. When an agent performs action A for a reason, he is motivated by an independent desire D to achieve an end E together with a cognitive consideration B that his performing action A will contribute to the achievement of E.42 According to the Humean conception of motivation, no belief can motivate an agent to act on its own. It needs to be combined with an independent desire, that is to say, a desire which is not itself produced by this belief.43 Briefly, the anti-Humean reaction can take two forms. First, it might be claimed that an independent desire is not always necessary to motivate us to perform an action. Some beliefs are able to motivate us to perform an action by causing us to have some new desire.44 For instance, “the belief that the action would be unjust could cause in me a desire to have nothing to do with it, and the belief–desire combination thus formed would be what motivates me to abstain”.45 The second form of anti-Humeanism is the view that belief alone is capable of motivating action. This is the view called “pure cognitivism” by Dancy who is also one of its most famous contemporary supporters.46 Like the Humean, Dancy claims that motivation requires a combination of two elements but unlike him he holds held that both could be beliefs.

41 Parfit (1997), p. 102. 42 The agent might believe either that the performance A constitutes a means to achieve E or that to perform A is to achieve E. 43 The most famous contemporary supporter of the Humean theory of motivation is certainly Donald Davidson. 44 Dancy and Parfit both attribute the first anti-Humean position to Thomas Nagel in Nagel (1970). For a discussion of the Humean vs. anti-Humean conception of motivation, see also Engel (1991). 45 Dancy (2002), pp. 12-13. On a variant of this anti-Humean view, the desire to abstain is not induced by the belief that the action is unjust. The desire to abstain is not a distinct mental state, it consists in my being moved by the belief that the action is unjust. 46 See Dancy (1993) and (2002).

35

“On my view, a desire is never a necessary part of what motivates. So we have two beliefs which together motivate. One of these is about how things are, and the other is about how they would be if the action were successfully performed”.47

Now, it is important to note that Dancy considers it would be unwise to contend that desire plays absolutely no role in motivation. This is because every time I am motivated to perform an action, I am, by definition, in a state whose success conditions consist in things coming to be in a certain way. That is to say, when I am motivated, I am inevitably in a state which has the mind to world direction of fit. And such a state is a desire not a belief. Hence, Dancy claims: “Desire never motivates, on my account. (If one is thirsty, one is motivated by the prospect of drinking.) But there can be no motivation without desire”.48 One again, I do not want to take a stance in the debate regarding the nature of our motivation when we act for a reason. I will often express myself in a quite Humean way without mentioning explicitly that there are other plausible conceptions of motivating reasons. This does not manifest a clear preference for the Humean model and has rather to be interpreted as a brevity measure. This very brief presentation of the Humean and anti-Humean positions aims at showing that even such a drastic anti-Humean conception as Dancy’s does not deny that to be motivated to perform an action is to have a desire towards the performance of this action. To say it differently, whenever I act for a reason, I can truly be said to have wanted or desired to act as I did. I am going to make use of this important lesson more than once.

BACK TO THE DOXASTIC REALM Epistemic reasons, non-epistemic reasons and evidence An important part of this work is grounded on the idea that our motivating reasons for acquiring a belief can take two main forms. Epistemic reasons to acquire beliefs

47 Dancy (2002), p. 13. 48 Dancy (2002), p. 14.

36

My reason for acquiring a belief is epistemic if and only if what I am motivated to achieve is the epistemic end of acquiring true rather than false beliefs.49

The simplest way of delineating the other set of reasons people can have for acquiring beliefs is to say that they are all conceivable reasons for acquiring beliefs that are non-epistemic. Non-epistemic reasons to acquire beliefs

My reason for acquiring a belief is non-epistemic if and only if what I am motivated to achieve is not the epistemic end of acquiring true rather than false beliefs but rather the end of acquiring pleasant, useful, reassuring, etc. beliefs.

The crucial thing to note here is the very specific use I make of the notion of epistemic reason. In this work, “epistemic reason” refers to a motivating reason. “Epistemic reason” is not synonymous with “evidence” as is commonly the case in the philosophical literature. “Evidence” is another very complex notion, which would deserve a full study for itself. In this work, I will make a quite unsophisticated and unproblematic use of the notion.50 I take it that pieces of evidence are indications —“evidence” and “truth indications” are synonymous in this work— of the truth of a proposition. To have evidence51 supporting proposition p is to have indications that p is true. For a subject to have evidence supporting proposition p is for the subject to be in a mental state which indicates that p is true. When I will say that my acquisition of the belief that p at t is grounded on — or supported by— the evidence I have at t, this is just a shorter way to say that I have evidence supporting the truth of p and that this is what causes me to acquire a belief with such a propositional content. But, strictly speaking, pieces of evidence do not support or ground the acquisition of belief, they support or indicate that a certain proposition is true. Even if it is not inappropriate to call my evidence supporting the truth of p my reason to believe that p, pieces of evidence are not motivating reasons.

49 I will say more about the epistemic end of acquiring true rather than false beliefs in chapter 5. 50 This is, I believe, the use privileged by the evidentialist conception of justification. See Feldman (2004a). 51 I say more about what it is for a subject to have evidence in chapter 6 below.

37

Let me first explain why I think it is appropriate to call “my evidence supporting the truth of p”, “my reason to believe”. Then I will try to show why pieces of evidence cannot be considered to be motivating reasons. According to its most general meaning, the term “reason” means simply “salient explanatory cause”. When I say, for instance: “the reason why Henry was late is that the road was congested”, I explain Henry’s being late by referring to what saliently caused it, but I clearly do not presume anything regarding his motivation to be late. It is appropriate to speak of the “reason” why a person did such or such a thing without simultaneously saying anything about what motivated him to do this thing. Something analogous happens when I mention my evidence in an explanation of why I believe such or such a thing. When I say: “the reason why I acquired the belief that p is that I remembered that p is true”, I explain why I acquired the belief that p by mentioning the salient explanatory cause of the acquisition of this belief. I do not say anything regarding my potential motivation to acquire a belief regarding the truth of p. Now, even according to the strictest anti-Humeanism, embodied by Dancy’s theory, if pieces of evidence are motivating reasons, my having pieces of evidence which indicate that p is true will involve my being in a state which has the mind to world direction of fit. But my having evidence supporting the truth of p does not seem to involve my desiring anything or my being in any sort of conative state. My having evidence supporting the truth of p involves my being in a state which has the world to mind direction of fit exclusively. On its own, my having evidence supporting the truth of p can only cause me to acquire the belief that p like my having evidence of a mouth-watering smell is able to cause me to be hungry. It is not something which motivates me to acquire the belief that p any more than the evidence of a mouth-watering smell constitutes a motivation to be hungry. I do not deny, of course, that my having evidence supporting the truth of a proposition p can motivate me if it involves a desire. But, on its own, my having evidence supporting the truth of p merely consists in my being in a state which indicates that p is true. As such, it is not something, which can motivate me to do one thing or another. For being

38

in a state characterized by the world to mind direction of fit is not sufficient to motivate me to act. In the rest of this work, I will take care to distinguish between:

1. My having an epistemic reason for acquiring a belief which is my being epistemically motivated to acquire a belief. When I have an epistemic reason to acquire a belief, I am in a state which involves a desire, more accurately, a desire to acquire true rather than false beliefs. 2. My evidence supporting the truth of a certain proposition p which is not a motivating reason.52

To be clear, I agree with philosophers53 when they distinguish between non-epistemic reasons for acquiring non-epistemically desirable beliefs, and evidence. I simply add to this model. Epistemic reasons are the exact analogue of the non-epistemic reasons in the epistemic realm. They are epistemic reasons to acquire epistemically desirable beliefs, i.e. true beliefs. Delineating the interesting issue Recall that the general goal of the first part of this work is to consider whether it is appropriate —at least sometimes—to hold people responsible for their belief acquisitions. As I said already, there is one obvious condition that the occurrence of a state of affairs needs to satisfy in order to be something for which someone is possibly responsible. People are not responsible for mere happenings that take place in the world independently of what they do. I am not responsible for earthquakes, or for your cat being crushed by a car. The things for which we can be held responsible are our purposeful behaviours: activities or actions.54 Then, I have a very simple test at hand to start checking whether people can possibly be held responsible for having acquired a belief: I need to

52 My having evidence indicating that p is true can be considered to be a normative reason since my having evidence indicating that p is true seems “to speak in favour” of my acquiring the belief that p. This is the normative view presupposed by the evidentialist conception of the justification of belief acquisition. 53 See Engel (1999), p. 8. 54 And the predictable consequences of our purposeful behaviour. I consider the issue of the conditions under which someone is responsible for the consequences of his behaviour in detail below.

39

investigate whether the acquisition of the belief that p can satisfy the condition under which a performance counts as an activity or an action. The reason why I previously stressed the distinction between activities and actions appears more clearly now. I was getting ready to delimit accurately the philosophical issue that I would like to address in the following two chapters. It needs to be made explicit that the interesting philosophical question does not concern the possibility of belief acquisitions qualifying as activities. It is hardly deniable that belief acquisitions are not always the faithful results of what we perceive as being true about our environment. There are definitely circumstances in which people unconsciously acquire beliefs as a result of their fears, their desires, etc.

“There is the familiar experience of wishful thinking, of belief arrived at through the influence of desire and wishes, or the formation of beliefs in self-deception and other abnormal or pathological cases. Literature, philosophy and psychiatry are full of such episodes”.55

The fascinating philosophical question rather regards the possibility of people acquiring beliefs for reasons, as it is possible for them to perform actions for reasons. Consequently, the doxastic responsibility that I try to understand in the following chapters is the analogue of the responsibility that an agent possibly has for his purposeful behaviour when it is a full-blooded action that he performs for a reason. Certainly, agents are also sometimes responsible for their activities. But this is not the kind of responsibility that I am going to track in the following chapters. The possibility of acquiring a belief for a reason is very controversial. Even if I am really highly motivated to get rid of the thought that I am a very bad tennis player and also consider that a good way of achieving this end would be to acquire the belief that I won my last match against my brother, it does not seem that I could acquire this belief, directly, for this reason.56 It seems that I cannot acquire a belief directly for a reason, as I can raise my finger directly for the reason that I desire to ask a question. But why is it so? This is the question tackled in the following chapter of this work. 55 Engel (1999), p. 5. 56 For the sake of brevity, I will mention the cognitive component of a motivating reason only occasionally when understanding requires it.