Reasons for Actions and Desires

ULRIKE HEUER REASONS FOR ACTIONS AND DESIRES (Accepted 22 October 2003) ABSTRACT. It is an assumption common to many theories of rationality that all practical reasons are based on a person’s given desires. I shall call any approach to practical reasons which accepts this assumption a ‘Humean approach’. In spite of many criticisms, the Humean approach has numerous followers who take it to be the natural and inevitable view of practical reason. I will develop an argument against the Humean view aiming to explain its appeal, as well as to expose its mistake. I focus on just one argument in favour of the Humean approach, which I believe can be constructed as the background idea of many Humean accounts: the argument from motivation. I first present the argument from motivation and explain why it seems so compelling. However, I then develop an equally compelling objection to desire-based approaches to reason, showing that they cannot accommodate the justificatory role of reasons. I show that this objection suggests that at least one of the premises of the argument from motivation must be false. And, finally, I argue that we should reject the premise that claims that only desires can explain actions. This result is fatal for desire-based views of practical reason. My conclusion is that practical reasons should be based not on desires, but on values. KEY WORDS: desire, explanation, justification, motivation, practical reason, W. Quinn, B. Williams It is an assumption common to many theories of rationality that all practical reasons are based on a person’s given motives, or desires.1 I shall follow common usage and call any approach to practical reasons which accepts this assumption – in one form or another – a ‘Humean approach’. In spite of many criticisms, the Humean approach has numerous followers who take it to be the natural and inevitable view of practical reason. I will develop an argument against the Humean view aiming to explain its appeal, as well as to expose its mistake. I will focus on just one argument in favour of the Humean approach, which – even though not always stated – I believe to be an important background assumption of many Humean accounts. It is also the Philosophical Studies 121: 43–63, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands. 44 ULRIKE HEUER argument which I regard to be the most powerful one in favour of the Humean view: the argument from motivation. My strategy will be as follows. I begin by presenting the argument from motivation and explaining why it seems so compelling. However, in Section 2, I will develop an equally compelling objection to desire-based approaches to reason, showing that they cannot accommodate the justificatory role of reasons. In Section 3, I will argue that this objection suggests that at least one of the premises of the argument from motivation must be false. And, finally, in Section 4 I will argue that we should reject the premise that claims that only desires can explain actions (or alternatively that desire-based explanations of actions are of primary importance). This result is fatal for desire-based views of practical reason. My conclusion will be that practical reasons should be based not on desires, but on values. 1. THE ARGUMENT FROM MOTIVATION The crucial first premise for the argument from motivation is this: A reason for action must be capable of explaining why a person acted in a certain way, if she actually acts for this reason. Adding ‘if she actually acts for this reason’ may make the claim sound trivial. But it brings out an important restriction. It is quite conceivable that a person has a reason to act which cannot explain her action for any of the following reasons: (a) she has a reason to φ, but acts for a stronger reason to do something else, (b) she has a reason to φ, and actually φs, but not for that reason,2 (c) she has a reason to φ, but does not φ (out of weakness of will, say, or just lethargy or even forgetfulness). In all these cases, no Humean would want to deny that the agent’s reason could serve as her reason for acting. But since the agent does not act for this reason, it is not a reason which is capable of explaining her action. Justificatory and explanatory reasons can come apart in yet another sense. Even if a person does act for a reason, the reason might not be sufficient to justify her action. This can happen in at least two kinds of situations: 1. Explanatory reasons may rest on mistakes without any loss to their explanatory force. If I believe that it would be a good idea REASONS FOR ACTIONS AND DESIRES 45 to φ, and it turns out later that it wasn’t, my belief can still explain why I performed the action. It can justify it, however, only if I was right. False beliefs normally don’t justify.3 2. Even if the explanatory reasons for an action don’t contain mistakes, they may not amount to a justification. The most common example for this are some cases of weakness of the will: Why did I smoke again, even though I had decided to quit? Because it was really tempting: it was nice to join the smokers, being around them provided opportunities for pleasant conversation, etc. These are reasons, I take it, but the explanation may not amount to a justification. It is quite conceivable that I myself conclude that all things considered I shouldn’t have smoked, despite its manifest attractions. Neither of these cases should be seen as contradicting the first premise of the argument because in both of them the reasons are in principle “fit to justify”. They would justify the action, in the first case if the belief was true, and in the second if there had not been a stronger reason to do something else. However restricted the claim of the first premise might appear given these observations, it is important to make it explicit, since it is crucial for almost all accounts of practical reasons. I am going to call it the identity thesis. By this I mean that when an agent acts for a (specific) reason that very reason is also the explanation (or at least part of the explanation) of why she did what she did. Normative or justificatory reasons and explanatory reasons are the same reasons in such a case, and not different kinds of reasons altogether.4 As I already pointed out, “identity” in this sense does not imply the claim that justificatory and explanatory reasons can never come apart. On its current, quite restricted interpretation the identity thesis picks out a subset of reasons which are both explanatory and justificatory: the reasons a person acts for who is neither weak-willed, nor mistaken about her reasons. So, all that is being claimed by the identity thesis is that justificatory reasons explain if a person acts for them; it does not imply that all explanatory reasons justify. According to the Humeans we need to add only one more premise to get to her theory of practical reasons, namely the claim that actions can only be explained by attitudes which are motivation- 46 ULRIKE HEUER ally relevant. In its philosophical usage, the notion of a desire is sometimes regarded as a term of art which covers all the various kinds of motivational attitudes that there are. Therefore, we could also say that actions can only be explained by desires. This thesis – let me call it the Humean thesis – and the identity thesis are the two premises of the argument from motivation. Taken together they lead to the Humean conclusion that all practical reasons must be based on desires. To recapitulate, once again the argument in its schematical form: (1) (2) (3) A practical reason (in the sense of justificatory reason) must be capable of explaining an action that was performed for that reason (= identity thesis). Actions can only be explained by motivationally relevant attitudes, i.e. desires (= the Humean thesis). All practical reasons must be based on desires (= the Humean view of practical reasons). Thus, it seems that the Humean view is vindicated by a simple and compelling argument.5 The Premises of the Argument Next, let me examine the premises more closely. Neither one appears to be problematic on its face. The identity thesis – the claim that justificatory and explanatory reasons are the same reasons – seems crucial for our understanding of ourselves as practically rational.6 Remember that it should be taken in a weak sense, allowing there to be justificatory reasons that did not motivate (hence, that cannot explain a person’s action) and explanations which are not justificatory. It amounts simply to the claim that justificatory reasons are in principle capable of explaining what we do. What would be the consequence of denying even this weak form of identity between justification and explanation? Utilitarians are sometimes said to deny it, and so did Frankena7 who (in some sense) got the contemporary debate about reasons and motives started.8 What we are motivated to do, and what we should do, they assure us, are quite different things. But the identity thesis does not contradict this claim. It simply ascertains that the reasons that explain and the reasons that justify are not always and not necessarily different. REASONS FOR ACTIONS AND DESIRES 47 Justificatory reasons can (special cases apart) be explanatory. Therefore, I take it, that at first sight the identity thesis is very plausible. It can be interpreted as expressing nothing more than the everyday assumption that we sometimes act for reasons, that is: that we do something because it is right or justified.9 And how should we understand the “because” if not as saying that the reason which justified the action also explains why the person acted in a certain way?10 But isn’t this weakening of the first premise perhaps more severe than I admitted, and forces us to reformulate the premise, i.e. to change (1) A practical reason must be capable of explaining an action that was performed for that reason to (1′ ) Some practical reasons must be capable of doing so? I believe that this change would be misleading, since it suggests that we can distinguish between those reasons which are motivating and those which are not by focusing on the reasons. Yet, all the explanations that I offered why a justifying reason might not be motivating depend on the special circumstances of the agent (her other reasons, her psychological states like being weak-willed, lethargic or forgetful), not on any specifics of the reasons. In different circumstances, the very same reason might well be motivating. As to the second premise, the only comment I wish to make now is that it seems overstated. Of course, there can be other explanations of actions, such as physiological or neurological ones (or at least other explanations of the events that constitute the action). Yet, the Humean claims that there is an explanation of every action qua action (as opposed to a mere event) which explains it by pointing out that it satisfied or contributed to satisfying an agent’s desire, or was at least, perhaps mistakenly, believed by her to be doing so. I am going to discuss the second premise at some length later (in section 4). For now, I assume that it is prima facie acceptable, especially if “desire” is understood in the broad philosophical sense to cover all motivational attitudes. Is the Argument Valid? If the premises so far seem to be acceptable, there is still the question whether it is a valid argument or not. After all, the conclusion states rather vaguely that all practical reasons must be ‘based’ on desires. 48 ULRIKE HEUER Why not say that practical reasons are desires? The Humean’s answer to this question is (I believe) that desires are starting points of deliberations about the best way of satisfying a given desire. The motivational force of the desire needs to be transmitted from its original object to others. In the simplest case it will be transmitted to the means necessary for satisfying the desire. Since these kinds of deliberation generate reasons, it would be too narrow to state the conclusion in a way which restricts reasons to desires only. The vagueness in the conclusion may thus be defensible.11 Therefore, I take it that the argument itself is valid. Desires as Reasons What does it mean to say that all practical reasons are based on desires? What is a desire? The technical character of this term as it is used in philosophical discussions is brought out by the fact that it is sometimes replaced by expressions such as “pro-attitude”. A desire is taken to be a pro-attitude towards a possible fact or towards a proposition. The most common analysis of propositional attitudes of this kind is that they are functional states: a person has a desire that p, if she has a disposition to act (under certain conditions) in ways that (are likely to) bring it about that p.12 What would it mean to regard desires so understood as reasons? Well, they are reasons to perform actions which are suited to satisfy the desire: The desire that p is a reason to perform an action which brings it about that p. If p is the case, the desire is satisfied. Note that this is just a logical sense of desire satisfaction. It is not necessarily related to any psychological state, such as feeling content, being pleased or happy. So much for a rough sketch of the basic meaning of “practical reason” in Humean theories, and of what is, perhaps, the most important argument in their favour. I will now discuss an objection, which does not attack the argument directly, but instead contests the basic meaning of “practical reason” in Humean accounts. I will then show how the objection reflects on the argument. REASONS FOR ACTIONS AND DESIRES 49 2. AN OBJECTION AGAINST THE HUMEAN ACCOUNT According to the Humean account, there are no restrictions at all on what a person can desire. It seems doubtful that this could be right, if we assume that desires are to be reasons that are capable of justifying and explaining actions. Would it really explain and justify an action if we pointed to just any proposition p that is likely to be made true as a result of the action and that would satisfy one of the agent’s desires? The question here is not whether this amounts to (say) a moral justification, but just whether it counts in favour of an action at all – even if only from the agent’s own perspective – that the action satisfies some of the agent’s desire. Elisabeth Anscombe raises a point which might help to state the objection:13 Consider someone who says, “I want a bowl of mud”, and acts in a way which brings it about that she has a bowl of mud as a consequence of her action. But on closer investigation the person has nothing to say that explains why she wants a bowl of mud, or why it seems good to her to have one, i.e. the person can neither explain nor justify her desire. But couldn’t the desire still be a reason for the action of satisfying it? Is it a justification, and an explanation for the person’s action to get the mud, that it satisfies her desire? According to Anscombe it is neither. She even takes the objection a step further: Not only is the agent’s action unintelligible, we cannot even attribute a desire for mud to her. We can only attribute a desire to a person if she could say something that explains why the thing in question appeared desirable to her. A possible answer could be: “I thought that we would throw the mud at each other, and it would be fun!” Anscombe regards the possibility of some such ‘desirability characterisation’ as a necessary condition of desire attribution. But – one might feel inclined to object – is saying that it would be fun to throw mud at someone any different from saying that one desires it? If it weren’t, the characterisation of the desired object as desirable would in the end again come down to the simple fact that one desires the object. Statements of the kind “it would be fun”, or “I would enjoy it” are different from desires, however, in at least three respects: 1. In claiming that something would be fun, one can be mistaken. I.e. this is a judgement which can be true or false. At least, on 50 ULRIKE HEUER some accounts of desires, one can also be mistaken in claiming that one desires that p. But even if so, it is only the belief about the desire that can go wrong, not the desire itself. Desires in the Humean sense aren’t judgements at all, and a fortiori not judgements about whether an activity would (or would not) be fun. 2. “That it would be fun” is semantically not equivalent to “that it satisfies a desire”, since desire satisfaction is not necessarily connected to hedonic states. “I want to go to the dentist” is not the same as saying “I will enjoy going there”. But perhaps desire satisfaction is a condition for experiencing pleasure, enjoyment or fun? Granted, it cannot be a sufficient condition, as the example of the dentist shows. But might it be a necessary one? It might, but it isn’t. After all, there are unexpected pleasures like pleasant surprises where the experience was not preceded by a desire. But if so, it seems that desire satisfaction and experiencing pleasure are different things. 3. The following consideration is even more important: It counts directly in favour of an action that it would be fun, or pleasurable. It shows (at least an aspect of) the good that the agent saw in performing the action. It doesn’t matter whether it also satisfies a desire or not. Finding pleasure in something isn’t a good thing only if, or because, it was desired. This last point concerns the normative relevance of pleasure. The Humean claims that only desire fulfilment matters, whereas if I am right, pleasure matters quite independently of desire fulfilment. I am going to call judgements that describe an action as desirable ‘value judgements’. I assume that the following is true of them: they express what the agent saw as counting in favour of her action; they can be true or false; they are reasons for normative judgements such as “x ought to φ”. That an action will be fun, or pleasurable is only one example of a value judgement. Yet it is an example, which in many situations puts an end to further questioning. Having said this, however, we are now in a position to see what is wrong with Anscombe’s claim that a desirability-characterisation is needed to make a person’s intentional actions intelligible. In particular, the claim that we can attribute a desire to a person only if REASONS FOR ACTIONS AND DESIRES 51 she can provide some characterisation of its content as desirable, invites objection. Can I not desire things that I do not in the least take to be desirable or good? Evidently, I can desire what I know to be bad for my health, for instance. Anscombe might answer that I must consider the unhealthy pursuit as being good in some other respect. Maybe I believe that the immediate pleasure outweighs later damages. Often enough, this will be the right answer. But is it really constitutive of a desire that it matches evaluative judgements? Isn’t it possible to desire something which one takes to be bad in every possible respect? It might be an expression of a depressed or desperate mood to do so, but even so it shows that desires and evaluations are not as tightly linked as Anscombe claims. The possibility that the two can come apart, might be explained by the various differences between ‘desiring’ and ‘believing to be desirable’ that I discussed above.14 Yet there is also something right about Anscombe’s view. Let me therefore present a modified version of the objection. I shall start with an example of Warren Quinn’s: Imagine that I had a disposition to turn on every radio within my reach, but not because I want to listen to music or news; as Quinn puts it: “indeed, I do not turn them on in order to hear anything”.15 In the functionalist sense of desire that most Humeans rely on I doubtlessly should be credited with having a desire to turn on all the radios within my reach. Furthermore, if someone has a reason to perform an action, in case the action is suited to satisfy a desire of hers, then surely I have a reason to turn on radios. Quinn, however, takes this as evidence that there is something wrong with the Humean account of reasons. “I cannot see how this bizarre functional state in itself gives me even a prima facie reason to turn on radios . . .”.16 His objection is a close relative of Anscombe’s but leads to a modified view. First of all, it allows that my disposition is a sufficient ground for attributing a desire to me. If anyone wants to know what is meant by desire here, the answer is this: I have a disposition to perform actions that are suitable to turn on the radios. Thus – contrary to Anscombe – Quinn does not claim that regarding the desired object as desirable is constitutive of desires. It is quite conceivable that there are desires for which this is false. According to Anscombe “the desirable” imposes some kind of normative constraint on the 52 ULRIKE HEUER formation of desires – comparable to the role of truth with respect to the formation of beliefs. If I believe that p, I also believe that p is true. If I find out that p is false, I don’t have the belief anymore. Similarly, Anscombe’s claim regarding desires is that if I desire that p, I also believe that p is desirable or good. If I find out that p is undesirable or bad, I don’t desire p any longer. Thus, Anscombe denies the central claim of Humean accounts: that one can desire anything. She thinks that there is a formal constraint on desiring. Quinn, on the other hand, agrees with the Humean that there aren’t any constraints on desiring. He is not proposing a narrower conception of desire. He denies, however, that desires can justify actions for just this reason! The objection is this: The mere fact that an action leads to desire satisfaction cannot justify the action.17 This becomes obvious in cases where the agent herself does not have anything to say about why the action looked desirable to her. In such cases, she herself would hardly claim that she was justified. Therefore, the problem cannot be explained away by claiming that the agent doesn’t share the observers’ standards of justification. Desire satisfaction all by itself simply fails to provide any justification. The justification from the person’s own point of view refers to the desirable features of her action, and not to her desire. Quinn’s example is of a desire stripped naked in order to show that neither the desire nor the satisfaction of it is all by itself reason giving. Yet, his claim is weaker than Anscombe’s since he does not maintain that the action cannot be explained by the fact that it satisfies the desire, but only that it cannot be justified. Luckily, most desires don’t come naked: usually the agent has something to say if asked why the object of a desire seemed desirable to her. Do we have desires like the one in Quinn’s radio-example at all then; are they perhaps indicating a pathology whenever they occur? It doesn’t really matter for my purposes whether we do have such desires, or whether they are extremely rare, or compulsory. I introduced the example to bring out a problem of all accounts of reasons that endorse two theses: (1) anything can be the object of a desire, and (2) actions can be justified (at least in some rudimentary sense) by showing that they are suited to lead to the satisfaction of a desire. Humean theories typically underwrite REASONS FOR ACTIONS AND DESIRES 53 both of these assumptions. The example shows that desires in the sense of (1) have no justificatory force at all. It also shows that the logical sense of desire satisfaction is irrelevant for justification. Asked for a justification people don’t normally mention desire satisfaction, but rather evaluative judgements. These judgements both make desires intelligible, and show why a person thought that she is justified in acting in a certain way. “That the action fulfils a desire” does not add to the justification, even if true. Only in very peculiar and perhaps somewhat desperate situations would a person mention desire satisfaction, if she were to justify her actions: in situations, that is, where a desire becomes so painful, or absorbs so much attention that one would rather satisfy it, in order to get rid of it. If all this is right, there must be something wrong with the Humean account of reasons. The claim that it is never a reason for an action that it satisfies a desire directly contradicts the conclusion of the argument from motivation that a person has a reason to act precisely if the action leads to the satisfaction of one of her desires. I presented two versions of the objection, using Anscombe’s and Quinn’s examples to bring out the difference between them. They differ in two main respects: (1) Quinn does not claim that evaluative judgements are constitutive of desires, whereas Anscombe does; (2) Anscombe claims that desires in the Humean sense neither justify nor explain actions. Quinn, on the other hand, denies only the justificatory force of desires, not the explanatory one. I will assume that his view is more attractive on both counts. But how does this objection relate to the argument from motivation? As I said already: it contradicts its conclusion. But it does so not by attacking either of the premises. After all, it wasn’t put forward as an objection to the argument. Let me try to sort this out in some more detail. 3. THE OBJECTION AND THE ARGUMENT As we all know, something must be wrong with the argument if its conclusion turns out to be false. However, the objection may be completely independent of the argument. But is it? Just to remind you: the second premise claimed that nothing could explain an action but motivational states, i.e. desires.18 This premise is 54 ULRIKE HEUER compatible with Quinn’s version of the objection which only claims that desires don’t justify. Even though he does not discuss action explanation, and therefore certainly does not endorse any claim which is as general as the second premise, it seems at the very least that there is no need for him to disagree with the Humean thesis. Quinn’s objection was that desires are not sufficient to justify an action. The second premise claims that desires are necessary for action explanations. There is no direct clash here. The claim is that my desire to turn on radios does not justify my action, but it may well help to explain it. But if the objection stands, and the second premise is alright, the first premise must be false. It easily follows from this interpretation of the objection that this is indeed so. I called the first premise the identity thesis, because it claims that justificatory and explanatory reasons are the same if a person acts for those reasons. But on the current interpretation of the objection justificatory and explanatory reasons would simply never be the same reasons. Explanatory reasons may be desires (plus the belief that the action in question is suited to satisfy the desire), but only value judgements justify, and desire satisfaction does not even add to the justification. The value judgement might explain the desire, but cannot explain the action itself. Therefore, anyone who endorses both the objection and the second premise has to reject the identity thesis. But then the identity thesis seemed crucial for making sense of the idea that we actually act for reasons. Therefore, this way of reconciling the objection and the argument seems doubtful. Perhaps it is rather the second premise which ought to be abandoned even though on first view it didn’t seem to be in conflict with the objection. If we drop the claim that only desires can explain actions, and allow for the possibility that evaluative judgements can explain them just as well, we can retain the identity thesis. And why should it be the case that only desires can explain actions? If an evaluative judgement can explain a desire, why couldn’t it also explain the action itself? If I believe that there is an interesting program on the radio, the judgement can explain my desire to turn on the radio. But couldn’t it just as well explain my action – turning on the radio? It seems that the Humean trades on an equivocation: she first says that only motivationally relevant attitudes can explain actions, and then adds that ‘desire’ REASONS FOR ACTIONS AND DESIRES 55 in its philosophical usage is just another word for ‘motivationally relevant attitude’. But if so, it should not exclude any plausible candidate from being a ‘motivationally relevant attitude’. However, the Humean notion of desire clearly excludes evaluative judgements. And this is not because they could not possibly be motivationally relevant, but rather because they aren’t desires. This shows that the employed notion of desire is not equivalent to an unbiased use of the notion of ‘motivationally relevant attitude’. Let me summarise the results so far: I first presented the strongest argument for a Humean account of reasons, the argument from motivation. I then discussed two versions of an objection against Humean accounts. Even though this objection is not directed against either of the two premises of the argument directly, it turns out that if the objection stands at least one of the premises of the argument has to go. If we keep the identity thesis, and reject the Humean thesis, there is nothing left of the Humean picture of reasons. In that case, the appeal to desires would not be necessary – neither for justification nor for explanation. If, on the other hand, we keep the second premise and reject the first, it would be possible to defend a Humean theory of motivation – but it is purchased at the price of making the Humean approach irrelevant for the explanation of justificatory reasons. – I take it that this is nonetheless the move that a Humean would be inclined to make. Michael Smith, for instance, seems to pursue this possibility.19 Thus it seems that the simple objection – that a desire cannot justify an action unless the agent can explain why the object of her desire appeared desirable to her – raises very serious problems for the perhaps best established approach to practical reasons. Facing the alternative of either giving up Humeanism and retaining the identity thesis, or hanging on to a Humean account of motivation at the expense of giving up on the identity of justificatory and explanatory reasons, I now want to argue that we should prefer the first option: give up Humeanism altogether. I will show that the second premise is false.20 56 ULRIKE HEUER 4. ACTION EXPLANATIONS Confronted with the problem of having to give up on one or the other of the premises, the question is which to drop. The only move that retains at least some of the Humean spirit would be to keep the second premise in place, the premise which claims that only desires can explain actions. That move allows us to tell a Humean story about motivation, even though it requires us to tell a quite different story about justificatory reasons. In this last section I want to show that this is in fact not an option. Let me go back to Quinn’s example. Even if the desire to turn on radios cannot justify the person’s action, it might still explain it. This observation counts in favour of the Humean theory of motivation. Or does it? The desire in the example is understood as a functional state: as the disposition to turn on radios. How can the functional state explain the action? The explanation would be something like this: ‘The person has this disposition and there was a radio within her reach. So she turned it on.’ Is this really an explanation of the action? The interesting explanatory element would have to be “she has this disposition”. But the reference to the disposition doesn’t seem to go beyond saying ‘if the opportunity arises, the agent will turn on the radio’: that the agent regularly turns on radios, and that she will do so again. But this was not the question. I did not ask “Is there a pattern?” but rather “Why did the agent perform this particular action?” Referring to a bare disposition does nothing to explain this. Perhaps it can be understood as saying: there can be no specific explanation of the particular case, but the only thing we can explain is the pattern. Thus, it would limit the range of possible explanations by rejecting a certain kind of explanation. But even if this rejection is appropriate – even if we can explain the particular action only by explaining the pattern – the answer “she has this disposition” does not explain the pattern either. It just states that there is one. If a person has such a disposition, the explanation will take a form which is different from other kinds of action explanation. And therefore, being told that she has such a disposition is an important bit of information. It would be like saying: ‘There is no point in looking for an explanation by focussing on the specific features of this case. We have to look elsewhere.’ This is most likely true in the radio case: the explanation of such an action would be the expla- REASONS FOR ACTIONS AND DESIRES 57 nation of a pattern, and might proceed along psychoanalytic lines. It might tell a story of childhood experiences and anxieties which in some ways involved radios. But the desire in the functionalist sense is not that which explains. It is – quite on the contrary – the disposition that we have to explain in order to be able to explain the action. If the explanation is the psychoanalytic one that I suggested, it wouldn’t give a justificatory reason, but it also wouldn’t rely on desire. It explains the action by explaining how the disposition was acquired. My conclusion is that dispositions like the ones in the example don’t explain.21 I don’t mean to say that they are irrelevant to explanation, since they might alter the form of the explanation, as we can explain the specific action only through explaining the disposition. It follows that desires, at least in this sense – far from being the only thing which could explain an action – don’t explain at all. You may want to object – and perhaps you wanted to object much earlier – that all this comes out the way it does only because I am employing a funny notion of desire. If my understanding of the concept of a desire were more adequate, everything would look different. Let me try out a completely different account of desire. Perhaps desire should be identified with some felt quality. The desire in the case of the radio example might consist in the urge to turn on radios together with some feeling of anxiety that the person experiences when she tries to resist it. This description leads towards a different kind of explanation: the feeling might explain the particular action directly, without making it out as part of a pattern of that person’s behaviour. She turns on the radios because she experiences a certain anxiety or uneasiness if she doesn’t. A notion of desire which captures these felt qualities of the experience would be much closer to the original Humean concept of desires as a specific kind of passions. But even if this looks in some sense promising, perhaps especially if we look out for the possibility of giving explanations, it comes with too many problems. Just to mention a few:22 (1) How is desire in this sense linked to functional states? Are there specific felt qualities which accompany all dispositions? That seems very unlikely. 58 ULRIKE HEUER (2) Moreover, is the experience of such felt qualities a necessary condition for every action? That seems plainly false. Think of everyday routines, or other kinds of very unemotional actions. But if it is not necessary, it wouldn’t help at all to establish the second premise. (3) Also: is the felt quality a necessary condition of desiring? Does one only desire something if there is an experiential quality to it? That again is unlikely. Think of wanting to be a good citizen, a good mother, etc. (4) Is it always the same quality, in those cases where there is any? In the radio example, we mentioned the feeling of anxiety. But one would much rather think of the expectation of pleasure as a feeling which accompanies many cases of desire. Why should all these different qualities be grouped together under the same heading? Therefore, I take it that this way of modifying the employed concept of desire would not be of any help. But which other possibility is there that also preserves the distinction between judgement and desire? Let me try out yet another candidate for an improved notion of desire: a less simple-minded functionalism. The functionalist need not refer to bare dispositions, but might rather think of desires as psychological states which play a certain functional role that is richer than just that of a disposition to do something. Pettit and Jackson write: “Functionalism is the doctrine that, for very many kinds of psychological states, to be in a psychological state of that kind is to have in one a state playing a certain role between inputs, outputs, and other internal states. It is the nature of that role, not the nature of the occupant of that role, which matters.”23 Kim clarifies further: “. . . functional properties are second-order properties defined in terms of causal/nomic relations between first order properties.”24 If desires were such secondary properties that can be defined in terms of “causal/nomic” relations between neural states and actions (and various other psychological states) they may well figure in interesting explanations of actions. One problem with this suggestion is, however, that these explanations are at best in the offing. So far, it isn’t more than a promissory note to claim there are any “causal/nomic” relations between neural states and actions. REASONS FOR ACTIONS AND DESIRES 59 We haven’t seen a single case of such an explanation yet. But in the meantime, we are already able to explain actions: We refer to the agent’s reasons, and even in cases where the agent doesn’t seem to act for a reason – as in Quinn’s radio example – we might well be able to explain what she does in other ways. But admittedly, at some point in the future the more sophisticated functionalist might be able to offer explanations of actions by desires, conceived as functional states. Would these explanations vindicate the Humean theory of motivation? Let me explain why I doubt it. As remarked earlier, explanations are generally plural. But the Argument from Motivation draws it strength from pointing out that explanation by desire is a particularly important and interesting kind of explanation of actions, since it is the only explanation which explains action qua action, and is therefore available for every action. It is a rationalizing explanation. But a rationalizing explanation is an explanation by the agent’s reasons for acting. Therefore, having seen that desires don’t provide reasons thoroughly undermines the special standing of explanation by desire. The special kind of explanation we are very often interested in is explanation by reason, rationalizing explanation – indeed. If desires cannot provide this kind of explanation, explanations involving desires will be much less interesting and important. So even if the sophisticated functionalist might eventually be able to offer an account of explanations by desires, those explanations will not occupy the central role that is assigned to desire-explanations in Humean accounts. Only explanations by reasons can do this. Hence, even the sophisticated functionalist’s explanations are not relevant for an account of explanatory reasons. The discussion in this section has shown that there can be various different kinds of explanations of actions – even if one does not focus on the central case where the justificatory reason explains the action. But none of them seems to require the presence of a desire. Furthermore, I started this section by granting to the Humean that we are willing to abandon the identity thesis. However, the identity thesis leads right into the centre of our understanding of rational agency. It explains what “acting for a reason” means. It in fact accounts for the special importance of rationalizing explanations. 60 ULRIKE HEUER The upshot of this is that the identity thesis is indispensable if we are trying to account for rational agency whereas desires are not. Therefore, I conclude that the Humean account of reasons has to go, even in its minimalist form of trying to give an account of explanatory reasons only.25 NOTES 1 Compare for instance: Bernard Williams, ‘Internal and External Reasons’, in B.W. Moral Luck (Cambridge: Cambridge University Press, 1981); J.L. Mackie, Ethics (Penguin: Harmondsworth, 1977); Davidson, ‘Actions, Reasons, and Causes’ (1963); Smith, ‘The Humean Theory of Motivation’; and for a rather nuanced version of the Humean view, Smith and Pettit, ‘Backgrounding Desire’ (1990). 2 For examples see Louise Antony, ‘Anomalism and the Problem of Explanatory Force’ (1989), pp. 157ff. 3 This can be brought out by looking at the structure of some excuses: if a false belief could justify, there would be no need for an excuse in such cases. But in fact, we do excuse ourselves by claiming that the situation – say, our specific epistemic situation – was such that it wasn’t possible for us to get it right. Only, because this is so – because it was impossible or extraordinarily difficult for us to form the right belief – are we excused (if we are). 4 Some take reasons to be mental states, such as beliefs; others hold that they have an independent reality. The thesis needs to be modified accordingly. If reasons exist independently of beliefs the identity would hold only between the propositions involved. 5 The Argument from Motivation is certainly one of the arguments that Williams relies on in his seminal article “Internal and External Reasons”. 6 In Practical Reality, J. Dancy defends a version of the first premise, denying however that it requires the identity of justificatory and explanatory reasons. Yet it seems to me that his view could well serve to illustrate an account of reasons that is firmly committed to the truth of this premise. 7 Frankena, ‘Obligation and Motivation in Recent Moral Philosophy’ (1958). 8 A more recent denial can be found in Parfit, ‘Reasons and Motivation’ (1997). 9 Perhaps it is not necessary that there is an explicit thought that the action is right. The person may just respond to something that she took to count in favour of the action. The phenomenology of ‘responding’ is not confined to the case where one thinks that the action is right. 10 Famously, Davidson in ‘Actions, Reasons, and Causes’ (1963) argued that it is the “because” of a causal explanation. 11 But one might object that – even with all the concessions made so far – the argument remains invalid. A version of this objection is sometimes attributed to Thomas Nagel. He claims that we can trivially attribute a desire to a person REASONS FOR ACTIONS AND DESIRES 61 whenever she acts. Thus Nagel seems to underwrite the second premise. But in the case of (as he puts it) “motivated desires” we can also give reasons for the desire, which in turn are not based on desires. And this is like saying that the conclusion of the argument is false. Since, in his use of the term “internalism” he also seems to endorse the identity thesis, his point could be constructed as an attempt to show that the conclusion would not follow even if the premises of the argument were true. However, I think that Nagel’s point is better understood as an attack on the second premise, and not on the validity of the argument. Nagel’s claim that we can ascribe desires trivially is only convincing, if we substitute “intention” for “desire”, and add that every action is an instance of intentional behaviour. It follows trivially from these assumptions that every action is intentional (under some description) – otherwise it wouldn’t even be an action. Saying that an action is intentional is tantamount to saying that the agent did what she did under some description. But ‘being intentional’ in this sense does not explain the action – it, rather, determines which action it is. The second premise, however, requires that the ascription of a desire will do some explanatory work. But in this sense, it is by no means trivial. The trivial bit might be – to use one of Davidson’s examples – that I intentionally moved my finger as a way of flicking the switch. “Flicking the switch” is the description under which I performed the action. It doesn’t explain the moving of my finger. A desire that might explain my flicking the switch could be that I wanted to turn on the light. But the ascription of such a desire is not trivial. The description of my behaviour as intentional on the other hand, is trivial in the sense that if it is known that the moving of my finger was an action of mine, it follows that there must be some description under which it is intentional. Consequently, Nagel’s claim is either not relevant to establishing the second premise, or it is false. 12 Cf., for instance, Michael Smith, The Moral Problem (Oxford and Cambridge, MA: Blackwell, 1994). 13 G.E.M. Anscombe, Intention, 2nd edn. (1963) (Ithaca and New York: Cornell University Press, 1957), p. 37ff. 14 For a more comprehensive discussion of the relation of desires and judgements about desirability compare David Velleman, ‘The Guise of the Good’, in Nous, . . . 15 Warren Quinn, ‘Putting Rationality in its Place’, in W.Q. Morality and Action (Cambridge: Cambridge University Press), p. 236. 16 Op. cit., p. 237. 17 This conclusion has been reached by various philosophers in recent years on different paths. Compare for instance, Christine Korsgaard, ‘The Normativity of Instrumental Reason’, in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason (Oxford: Clarendon Press, 1997), pp. 215–254; Jean Hampton, The Authority of Reason (Cambridge: Cambridge University Press, 1998), part II; Joseph Raz, ‘Incommensurability and Agency’, in R. Chang (ed.), Incommensurability, Incomparability, and Practical Reason (Cambridge MA: Harvard University Press, 1998), pp. 110–128. The argument that I put together here, following W. Quinn, simply lays out the, to me, most convincing way of reaching the conclusion. 62 ULRIKE HEUER 18 Or – a bit more modestly – for every action there is, and needs to be an explanation of this kind. 19 Smith (1994), op. cit. 20 I will not try to argue, however, that the first premise is true, beyond the remarks that I made above. 21 This amounts to rejecting Quinn’s point that functional states may explain, after all. But, I think, we should hold on to his suggestion that value judgements are not constitutive of desires, and also to the idea that not every explanation of an action needs to be a justifying explanation. 22 Compare M. Smith, The Moral Problem, chapter 4. 23 Jackson and Pettit, ‘Functionalism and Broad Content’, Mind 97 (1988), p. 381. 24 Jaegwon Kim, Mind in a Physical World (Cambridge: MIT Press, 1998), p. 20. 25 I would like to thank John Collins, David Enoch, Victoria McGeer, Jimmy Lenman, Joseph Raz and Jay Wallace for very helpful comments, suggestions and discussions. An earlier attempt of mine to deal with these problems was published in German as “Sind Wünsche Handlungsgründe?” in Analyse und Kritik Vol. 21, 1999, pp. 1–24. Yet, the conlcusion there is quite different and less developed. REFERENCES Anscombe, G.E.M. (1957): Intention, Ithaca/New York: Cornell University Press (2nd edition 1963). Antony, L. (1989): ‘Anomalism and the Problem of Explanatory Force’, The Philosophical Review 98, 153–187. Dancy, J. (2000): Practical Reality, Oxford: Oxford University Press. Davidson, D. (1963): ‘Actions, Reasons, and Causes’. Reprinted in Donald Davidson (1980): Essays on Actions and Events, Oxford: Oxford University Press, pp. 3–21. Frankena, W. (1958): ‘Obligation and Motivation in Recent Moral Philosophy’. Reprinted in K.E. Goodpaster (ed.) (1976): Perspectives on Morality. Essays by William K. Frankena, Notre Dame and London. Hampton, J. (1998): The Authority of Reason, Cambridge: Cambridge University Press. Heuer, Ulrike (1999): ‘Sind Wünsche Handlungsgründe?’, in Analyse und Kritik 21, 1–24. Jackson, F. and Pettit, P. (1988): ‘Functionalism and Broad Content’, Mind 97, 381–400. Kim, J. (1998): Mind in a Physical World, Cambridge: MIT Press. Korsgaard, C. (1997): ‘The Normativity of Instrumental Reason’, in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason (pp. 215–254), Oxford: Clarendon Press. Mackie, J.L. (1977): Ethics. Inventing Right and Wrong, Penguin: Harmondsworth. REASONS FOR ACTIONS AND DESIRES 63 Nagel, T. (1970): The Possibility of Altruism, Princeton and Oxford: Princeton University Press. Parfit, D. (1997): ‘Reasons and Motivation’, Proceedings of the Aristotelian Society 71(Suppl.), 99–130. Quinn, W. (1993): ‘Putting Rationality in its Place’, in W. Quinn (ed.), Morality and Action (pp. 228–255), Cambridge: Cambridge University Press. Raz, J. (1998): ‘Incommensurability and Agency’ in R. Chang (ed.), Incommensurability, Incomparability, and Practical Reason (pp. 110–128), Cambridge, MA: Harvard University Press. Smith, M. (1994): The Moral Problem, Oxford and Cambridge, MA: Blackwell. Smith, M. (1987): ‘The Humean Theory of Motivation’, Mind 96, 36–61. Smith, M. and Pettit, P. (1990): ‘Backgrounding Desire’, The Philosophical Review 99(4), 565–592. Velleman, D.: ‘The Guise of the Good’, Nous 26, 3–26. Williams, B. (1981): ‘Internal and External Reasons’, in B. Williams (ed.), Moral Luck (pp. 101–113), Cambridge: Cambridge University Press. University of Pennsylvania 433 Logan Hall Philadelphia, PA 19104-6304 USA E-mail: ulrike@phil.upen.edu