Original citation: Roessler, Johannes. (2014) Reason explanation and the second-person perspective. Philosophical Explorations, Volume 17 (Number 3). pp. 346-357. Permanent WRAP url: http://wrap.warwick.ac.uk/66448 Copyright and reuse: The Warwick Research Archive Portal (WRAP) makes this work of researchers of the University of Warwick available open access under the following conditions. Copyright © and all moral rights to the version of the paper presented here belong to the individual author(s) and/or other copyright owners. To the extent reasonable and practicable the material made available in WRAP has been checked for eligibility before being made available. Copies of full items can be used for personal research or study, educational, or not-forprofit purposes without prior permission or charge. Provided that the authors, title and full bibliographic details are credited, a hyperlink and/or URL is given for the original metadata page and the content is not changed in any way. Publisher statement: "This is an Accepted Manuscript of an article published by Taylor & Francis Group in Philosophical Explorations on 27/08/2015, available online: http://dx.doi.org/10.1080/13869795.2014.946954 .A note on versions: The version presented here may differ from the published version or, version of record, if you wish to cite this item you are advised to consult the publisher’s version. Please see the ‘permanent WRAP url’ above for details on accessing the published version and note that access may require a subscription. For more information, please contact the WRAP Team at: publications@warwick.ac.uk http://wrap.warwick.ac.uk/ 1 Reason explanation and the second-person perspective Johannes Roessler Department of Philosophy, Warwick University, Coventry, U.K. Abstract: On a widely held view, the canonical way to make sense of intentional actions, is to invoke the agent’s ‘motivating reasons’, where the claim that X did A for some ‘motivating reason’ is taken to be neutral on whether X had a normative reason to do A. In this paper I explore a challenge to this view, drawing on Anscombe’s ‘second-personal’ approach to the nature of action explanation. Keywords: intentional action; rationalization; second person; Elizabeth Anscombe; disjunctivism. As Michael Thompson has pointed out, Anscombe’s ‘reason-seeking’ question ‘why?’ is usually formulated with the use of the second-person pronoun. In what Thompson calls Anscombe’s ‘fundamental scene’, my attention is caught by your doing A, I ask ‘Why are you doing A?’, and you proceed to tell me about your reasons for doing A. 1 (Your answer is not the end of the matter. It’s often the beginning of a more extended conversation.) What are we to make of Anscombe’s emphasis on ‘second-person interactions’ about practical reasons? Someone steeped in contemporary philosophy of action might suggest that we should think about this in terms of the familiar notion of first-person authority. When you are interested in X’s ‘motivating reasons’ (the reasons for which X is doing something), 1 Thompson 2011, 206. 2 she might say, it’s generally a good idea to ask X herself. For it is in the nature of motivating reasons that the agent’s view should be accorded a special weight, barring confusion or irrationality. It has to be said, though, that this account is hard to square with Anscombe’s illustrations. The questioner in her examples is someone who can be quite quarrelsome. She sometimes bluntly contradicts the agent (‘but your camera is in the cellar’), and sometimes appears to challenge the agent’s judgement (‘You did it because he told you to? But why do what he says?’). Nor are these wrangles peripheral to Anscombe’s concerns. They are in fact central to her account of what it means to explain an action in terms of a reason. ‘Roughly speaking,’ she writes, ‘it establishes something as a reason if one argues against it.’ (1957, 24) At this point, our contemporary philosopher of action might switch tack. ‘It does not appear to be motivating reasons Anscombe’s protagonists are discussing’, she will say, ‘but normative reasons — facts that count in favour of doing something. Firstperson authority, needless to say, does not extend to one’s normative reasons. Agents are often mistaken about, or ignorant of, the facts that determine what they have most reason to do, and of course disagreements about our normative reasons are part and parcel of most social relations. Anscombe’s examples are to be understood as disputes about the justification of the actions under consideration, not as disputes about their explanation.’ An initial criticism of this analysis is that the notion of justification is not quite the one we need to understand Anscombe’s (and her protagonists’) concerns. The notion encourages the expectation that the debates are about justifying reasons, considerations to the effect that acting in the relevant way is the right thing, and 3 failing to act in that way would be wrong. Furthermore, the notion of justification might lead one to think that the agent is asked to justify her activity to the questioner, where this would mean the latter must have the requisite authority to challenge the agent and hold her responsible. This is a familiar theme in recent work in moral philosophy on the ‘second-person perspective’ and its role in our thinking about certain kinds of moral reasons. (Darwall 2006) Neither of the two associations of ‘justification’ seems to me to be germane to Anscombe’s discussion. The kind of practical reason you have (or think you have) when you go upstairs in order to fetch your camera is not typically a consideration that would make it wrong not to go upstairs, as opposed to making it lazy or foolish (or possibly laying you open to no criticism at all). A reason can recommend a certain course of action without making it wrong, absent other justifying reasons, not to take it. 2 Again, in asking ‘Why are you going upstairs?’ I’m not challenging you to justify your behaviour to me. Perhaps in special circumstances I might (say, if you promised to stay downstairs); but the question is a perfectly intelligible and familiar question to ask even without that special background. It is simply ‘What is the point of your going upstairs?’. What’s requested is indeed a normative reason, a set of considerations that counts in favour of your going upstairs, but not necessarily a ‘justifying’ reason in either of the narrower senses. But this is not the main problem with the proposed analysis. The main problem is that the contemporary theorist appears to miss the point that in Anscombe’s examples normative reasons are invoked to explain why the protagonist is doing what she is Put in the terms used in recent debates about the nature of the ‘favouring’ relation, Anscombe’s reasons are arguably not exclusively ‘deontic’ or ‘peremptory’. For helpful discussion of non-deontic practical reasons, see Little 2013. 2 4 doing. By the lights of the contemporary theorist, this amounts to a category mistake. According to a widespread assumption in contemporary work on action explanation, we have two concepts of practical reasons, catering to different kinds of interest: we reflect on normative reasons when interested in deliberative or justificatory questions (What am I to do? What she right to do A?), and we invoke ‘motivating reasons’ in response to explanatory questions (Why did she to A?). Answering these questions, on what I’ll call the ‘two-concept view’ of practical reasons, are distinct kinds of enterprises. Thus the enterprise of making intentional actions rationally intelligible is seen as insulated from the rough-and-tumble of debates about normative reasons. My aim in what follows is to bring out the force of Anscombe’s challenge to this picture. We can think of the challenge as falling into two parts. First, reflection on the second-person question ‘Why?’ suggest that the two-concept view distorts the way we ordinarily think about intentional actions: it represents us as occupying a more detached perspective on action explanation than is compatible with our role as deliberators and interlocutors. The second part of the challenge maintains that the two-concept view distorts what it means to act intentionally. Understanding intentional action requires understanding the kind of the explanation requested by Anscombe’s second-person question ‘Why?’ — construed in the commonsensical, ‘argumentative’ way made vivid by her examples. I begin with a more detailed exposition of the two-concept view. In section 2, I discuss the first part of the challenge. In section 3, I turn to the second part, enlisting the support of recent work on a ‘disjunctive’ conception of reason-giving explanation. 5 1. The two-concept view The basic observation from which the two-concept view develops is that we seem to be able to make actions rationally intelligible even if the agent is mistaken about her normative reasons. The way the point tends to be put is that ‘the agent’s reason’ for (say) drinking the content of the bottle (in Williams’s example) was that he believed the bottle contained gin.3 Identifying ‘the agent’s reason’ enables us to make his drinking petrol and tonic rationally intelligible (though of course not under that description). Evidently the reason we invoke here is not a normative reason. There is nothing that counts in favour of the agent’s drinking the content of that bottle, yet he does so intelligibly. Again consider actions that reflect poor judgement or pursue bad objectives. Even though the agent has no normative reason for doing what she does, at least no undefeated normative reason, her action may be rationally intelligible. Generalizing from such cases, two-concept theorists make two claims: (a) Finding an action rationally intelligible requires identifying the operative ‘motivating reason.’ (b) ‘Motivating reason’ statements are neutral on whether the agent has a normative reason for acting in the way she does. (a) does not mean that explanations in terms of ‘motivating reasons’ are exhaustive, or should count as ‘the explanation’ of an action. Presumably almost anything can be a factor that in one way or another may influence human conduct. The claim is merely that there is a distinctive type of explanation that renders an action intelligible as an According to Jonathan Dancy, the canonical way to put the agent’s reason is: his reason was that, as he believed, the bottle contained gin. See Dancy 2000, 129. 3 6 exercise of rationality (hence ‘as an intentional action’4); and that explanations of that type must invoke ‘motivating’ reasons. There is more than one way to conceive of ‘motivating’ reasons. Some two-concept theorists think of them as psychological states of the agents, such as her beliefs and pro-attitudes. (Davidson 1963) Others favour a ‘non-psychologistic’ account, identifying ‘motivating’ reasons with the content of beliefs. (Dancy 2000) There is agreement, however, that insofar as we are interested in finding an action intelligible as intentional, we must be interested in something that is neutral on whether the agent has a normative reason. To be engaged in the business of rational explanation is to occupy a ‘detached’ point of view in this sense: the circumstances that justify or give a point to the agent’s activities are relevant at best indirectly, insofar as they might help us to identify the explanatorily critical non-factive attitudes. Note that the two-concept view may not be incompatible with regarding propositional knowledge, and specifically knowledge of considerations that are normative reasons for one’s actions, as explanatorily relevant. If almost anything can be a factor influencing human conduct, why shouldn’t the agent’s knowledge? Arguments to the effect that explanations in terms of knowledge can be correct and illuminating, insofar as they sustain distinctive sets of counterfactuals (Williamson 2000), are unlikely to cut any ice with a two-concept theorist. For such arguments leave open whether explanations in terms of knowledge form part of the explanatory schema in terms of which we find intentional actions intelligible as such. Still, work on the explanatory role of knowledge highlights a substantive assumption on the part of the two-concept 4 More precisely: as an action performed with a certain intention. Trivial actions like doodling, for example, may be intentional without being open to rational explanations. Trading convenience for precision, in what follows I’ll revert to the simpler formulation. 7 view. Why should it be assumed that rational intelligibility is to be analyzed in terms of a single explanatory schema? Pluralists might argue that the following is one — though not the only — way to find X’s doing A intelligible as an intentional action: X is doing A because, as X recognizes, the facts that p and q give her an (undefeated) normative reason to do A.5 Now there are, in the literature on ‘motivating’ reasons, some attempts to support (a) and (b) not just by generalizing from various kinds of ‘bad’ or ill-informed cases, but through reflection on the distinctive projects associated with the pursuit of normative and explanatory questions. According to Jonathan Dancy, ‘there is, it would seem, no one sort of request for explanation that both sorts of question address’ (2000, 8) Dancy is assuming here that we may understand the normative question (Did X have a good reason to do A?) as a request for an explanation of what, if anything, made the action worth doing. His claim is that the latter kind of explanation is never relevant (or ‘requested’) when we are concerned to make X’s doing A rationally intelligible (i.e. when we pursue an ‘explanatory question’ in the more straightforward sense: Why did X do A?). Note, though, that Dancy offers no support for this claim, other than describing his distinction as ‘perfectly ordinary and unproblematic’, and ‘not particularly obscure or complicated’. (2000, 1-2) In a critical discussion of Dancy’s work, Jay Wallace elaborates on the distinction. He claims that ‘the two questions’ we may answer by invoking reasons for action ‘are See Josef Raz’s discussion of what he calls the ‘normative/explanatory nexus’, in his 2011, ch. 2. Disjunctivists about acting for a reason (Hornsby 2009, McDowell 2013) make the further claim that the intelligibility of intentional actions in terms of nonfactive attitudes depends on the intelligibility provided by normative reasons, recognized as such by the agent. I’ll come back to disjunctivism in section 3. 5 8 characteristically posed from very different points of view.’ The perspective ‘within which normative reasons have their place is characteristically prospective, firstpersonal and deliberative.’ (2006, 66) In contrast, ‘when we address the question of motivation, we typically focus on an action that has already been performed, and we think about the action from a distinctively third-person perspective.’ An immediate reaction to this passage is that there are plenty of counterexamples. When we try to make X’s doing A intelligible we do assume, as Wallace rightly emphasizes, that there is an explanandum. But that’s not to say that we take it X has already done A. We may, and frequently do, ask for an explanation of facts stated by the use of the present progressive. ‘Why is X doing A?’ is a perfectly sensible question to ask and answer quite independently of whether it will be true at any point that X has done A. (See Anscombe 1957, Thompson 2008). Again, Wallace is surely wrong that consideration of normative reasons must be a matter of first-personal deliberation. In advising X, for example, you may tell her that she has most reason to do A, and if you are any good at giving advice, this will not necessarily be a claim to the effect that if you were in X’s situation, you would have most reason to do A.6 Perhaps Wallace’s point is merely that reflection on his two standpoints illuminates the nature of normative and ‘motivating’ reasons, respectively. This might be so even if our thinking about such reasons is not invariably but only ‘characteristically’ conducted from the relevant perspective. It is not implausible to think that that there is some such connection between normative reasons and practical deliberation. It is much less clear why ‘historical’ understanding should be regarded as paradigmatic of 6 As Barry Stroud has stressed. See his 2011, 111-112. 9 the project of rational explanation. Wallace offers no reason for this claim. In the next section I’ll set out an argument against it. According to Dancy, ‘there is, it would seem, no one sort of request for explanation that both sorts of question address’ (2000, 8) This, I shall argue, is easily refuted: Anscombe’s question ‘why?’ amounts to just that sort of request. 2. Sharing reasons It sometimes happens that the reason given by an agent for her current activity is accepted by the audience. You ask someone ‘Why are you setting up a camera on this pavement?’ In response you are informed ‘Because Marilyn Monroe is going to pass by.’ (Anscombe 1957: 35) Accepting the answer, you now see the action as having a recognizable point. Of course instead of answering your question, the photographer might have told to mind your own business, and instead of accepting his reason you might have thought he was deranged. But sometimes a reason offered by one speaker is accepted by another, and this kind of correspondence is plausibly seen as in some sense part of the aim of interactions involving the giving and asking for reasons.7 You might say that it’s hard to see how any of this can provide materials for a challenge to the two-concept view, given that similar examples regularly feature in the writings of two-concept theorists. To locate the disagreement, it’s useful to distinguish three elements of Anscombe’s view of the reason-seeking question ‘why?’: 7 See Moran 2005 for illuminating discussion. 10 (a) We ordinarily take it that a good way to answer the question ‘Why are you doing A?’ is to answer the question ‘What’s the point (or: What’s the good) of doing A?’8 (b) While there is more than one style of answering the latter question, the most basic way to do so is this: we set out the considerations in the light of which our doing A can be seen to be an effective way to promote some (in some way) desirable outcome.9 (c) Such explanations are often meant to be taken at face value. For example, accepting the reason given by the photographer for positioning his camera on that pavement would involve accepting (i) that Marilyn Monroe is going to pass by, and (ii) that this (plus the desirability of his taking a picture of her) gives him a good reason for acting, and (iii) that he is acting because these considerations give him good reason. It’s (c) that implies a challenge to the two-concept view. It suggests that we’re ordinarily happy to assign explanatory relevance to normative reasons, as opposed to deploying the ‘detached’ explanatory schema featuring ‘motivating reasons’ (conceived as neutral on whether the agent has a normative reason). Accordingly, while two-concept theorists may be more or less sympathetic to (a) and (b), they tend to reject (c). One might wonder whether they are committed to doing so — whether they might not simply acknowledge that our ordinary thinking about intentional Compare Vogler’s suggestion that on Anscombe’s (and Aquinas’s) view, these are ‘basically the same question’. (2002, 31) 9 Alternative styles include: ‘I’m doing A in order to do B.’ and ‘I’m doing B.’ What I mean by calling these types of answers less basic is that they are in a significant sense incomplete (which is not to say that they are not for many purposes perfectly adequate): they leave open why the fact that doing A facilitates doing B should be thought of as a consideration counting in favour of doing A. 8 11 actions does not always tally with their austere philosophical theory. I will return to this issue in the next section. In any case, as a matter of fact, two-concept theorists tend not to think of their view as in any way revisionist. They grant that the question ‘why?’ is often answered by stating non-psychological reason-giving facts rather than citing the relevant ‘motivating reason’. But they think that such talk is intended to be taken with a pinch of salt. Putatively non-psychological reason-explanations, according to Wallace, are ‘naturally taken as abbreviated ways of characterizing the agent’s state of mind’ (2006, 67, n.3) — more precisely the agent’s non-factive propositional attitudes.10 According to Davidson, such claims perform the ‘familiar dual role’ of explaining (by ‘indicating’ the non-factive mental states that constitute the ‘motivating reason’) and justifying (by making a claim that would vindicate the relevant attitudes). (1980, 8) These suggestions promise to enable us to make sense of (a) and (b) without accepting (c). The idea is that even from the participants’ perspectives, it is apparent that the normative question ‘What is the point of doing A?’ is just an indirect method for getting at the answer to the explanatory question ‘Why are you doing A?’. Answering the former will involve expressing or ‘indicating’ the non-factive attitudes that in turn provide the explanation demanded by the latter. This approach is presumably fuelled by the thought that there is no coherent alternative. You might argue, for example, that there is no conceivable mechanism by Dancy’s account of reason explanation encourages the following variant: ostensibly factive non-psychological explanations (‘X is doing A because p’) are ‘naturally taken’ as non-factive non-psychological explanations (‘X’s reason for doing A is that p’, where this is held to be consistent with ‘but not p’). Compare Dancy’s claim that ‘there are factive ways of giving an explanation which, in its own nature, is nonfactive’. (2011, 7) 10 12 which the imminent arrival of Marilyn Monroe could affect the photographer’s current behaviour. And you might say that in any case we have every reason to think that his behaviour is in fact explained by his beliefs and pro-attitudes. As Davidson writes, ‘(y)our stepping on my toes neither explains nor justifies my stepping on your toes unless I believe you stepped on my toes, but the belief alone, true or false, explains my action.’ (1980, p. 8) But we have already seen that this is inconclusive. It’s not incoherent to suggest that my stepping on your toes is explained by my knowing that you stepped on my toes. This would mean that an explanation in terms of the fact that you stepped on my toes is literally correct. It’s not of course the whole story, even the whole reason-giving story, as it does not mention my knowledge of the reason-giving fact — without which the fact could not be the reason for which I’m acting. (Hyman 1999, Hornsby 2008) But importantly, the more detailed account entails the sparser version. If your factive attitude that p makes a difference, so does the fact that p. Again, insofar as the photographer knows Marilyn Monroe is going to pass by there is no particular mystery about how that fact can make a difference to his behaviour (no more so, at least, than there is about his capacity to know that she is going to pass by). Not only is there no incoherence in taking non-psychological reason-giving explanations at face value, it’s in fact profoundly natural for us to do so when we are talking about things we are in the process of doing. In such cases, one’s relation to the activity is not theoretical but practical. Frequently one still needs to engage in practical reasoning to settle the finer details of how to go on, and quite generally it can still be a practical question whether to carry on at all. The case for (c) turns on the thought that practical reasoning has two aims that are intelligibly and, for the 13 deliberator, essentially connected. One aim is to get right what one has most reason to do. The other aim is to determine or control what one will be (or is) doing. Suppose the starting point for practical reasoning is provided, at least typically or in the basic case, by some more or less determinate end one has in view, such as taking a picture of Marilyn Monroe. The purpose of the reasoning is to help achieve that end by getting right what (else) one needs to do to achieve it. This in turn renders both of the characteristic aims of practical reasoning readily intelligible. If you are to achieve your end you certainly need to get right what course of action will enable you to achieve it; equally the outcome of that reasoning must make a difference to what you will be doing. The two things are evidently connected. What’s essential for success in practical reasoning is not just that it should get things right and that it should somehow make a difference to what one will eventually be doing, but that the facts (which one needs to get right) will make a difference to what one will be doing. These general points suggest it must be mistake to draw a contrast between the standpoints of practical reasoning and explanation. In practical reasoning we assume that our reasoning will make a difference to the action towards which we are reasoning, and this bears on how best to explain what we are or will be doing. Practical reasoning, in central cases, is concerned with a certain How?-question (How can I best do B?), but its conclusion has an immediate bearing on a certain Why?question (Why am I doing A?). Insofar as the answer to the How?-question is ‘by doing A’, it supplies a matching answer to the Why?-question: ‘because doing A is a good way to do B.’11 11 I borrow the distinction between the two questions from Hornsby 2011, 119. As Hornsby puts it, the two questions ‘go hand in hand.’ 14 To insist that the correct explanation of the agent’s doing A lies in her ‘motivating reasons’ (her non-factive attitudes) would be to sever or at least complicate the link between these questions. One would have to think about one’s action from a standpoint that’s neutral on whether one is getting things right in one’s practical reasoning, for example, whether one is discovering instrumental relations that will facilitate achievement of one’s end, or is deluded about them. It may not be impossible for an agent to adopt that kind of perspective even in the midst of deliberating and acting, but doing so would certainly go against the grain. It’s hard to stop caring about the distinction between discovery and delusion while being engaged in the project of practical reasoning, the success of which turns on that distinction. Correlatively, it is natural that when asked why she is doing A, the agent will invite others to share the view that reflects her answer to the question of how best to do B. I finish this section by trying to allay an interesting worry, to do with the connection between normative reasons and action explanation. The worry, in a word, is that Anscombe’s view makes the business of action explanation, even as conducted through second-person interactions, excessively and perhaps insanely demanding. Accepting the reason given by the agent would, in basic cases, involve accepting that there is some desirable outcome that the action helps to promote. This, you might say, makes explaining what one is doing look too much like trying to get the audience to collaborate. Without pretending to do justice to this large and interesting topic, I would make two immediate points in response. One is that it’s far from implausible that accepting the agent’s reason involves some evaluative commitments; the other is that these are fairly minimal. If the photographer’s account of his action turns on an objective we don’t consider to be in any way desirable (assassinating rather than 15 photographing Marilyn Monroe, say), we wont’ be disposed fully to share his reason. But it is a further question whether the evaluative commitment we incur in accepting someone’s reason will have any practical implications for us. We may lack the requisite skills, opportunities, temperament, and so forth, and even if the commitment generates practical reasons for us, these may be defeated by other considerations. Perhaps we will at least acquire a good reason not to interfere with the agent’s activities. But this does not look like an implausibly onerous commitment.12 3. Disjunctivism If the discussion of the last section is on the right lines, the two-concept view would seem to misrepresent the way we ordinarily find our own and our interlocutors’ intentional actions intelligible. How damaging is this criticism? You might say that the damage would be strictly limited. For we need to distinguish three questions, only the last of which, you might insist, is of any real importance to the two-concept view: (i) Do we find it natural to accept explanations of intentional actions that invoke normative reasons, recognized as such by the agent? (ii) Are such explanations sometimes correct? (iii) Should this kind of understanding count as a case of finding an intentional action intelligible ‘as such’ (as an explanation of the kind to which acting with an intention is as such open)? Two-concept theorists are committed to a negative answer to (iii). But it might be said that they can keep an open mind on (i) and (ii); and that the discussion of the previous section really only speaks to (i). Yet, if the answer to (i) is affirmative, we’d clearly need a powerful reason to accept a negative answer to (iii). It’s agreed on all hands (in the current debate) that explaining 12 For helpful discussion of reasons for non-interference see Wallace 2009. 16 an intentional action ‘as such’ is to understand it as an exercise of practical rationality. If we find it natural to explain intentional action in terms of normative reasons, recognized as such by the agent, and if, moreover, such explanation are sometimes correct, what’s the rationale for denying that they count as seeing an intentional action as a manifestation of practical rationality, hence as finding it intelligible as such? It might be said that the rationale is not far to seek. As Jonathan Dancy succinctly puts it: ‘it is not required for the purpose of explanation that the agent be right about things.’ (2011, 6) It seems right that whenever someone is intentionally doing something it’s possible to make her action (more or less) rationally intelligible without taking a view on whether there are adequate normative reasons for which she might be said to act. This is an important point, the significance of which is not always appreciated. George W. Bush, for example, was reported as urging that we should condemn, rather than try to explain, acts of terrorism. Dancy’s dictum could be used to remind Bush that the two things are perfectly compatible. But there is an alternative, stronger construal of Dancy’s point. He might be interpreted as saying that whether the agent is right about things is a question that is never immediately relevant for the purpose of rational explanation. There may of course be good reasons to take an interest in it. But when we do so in the context of trying to make sense of someone’s actions, we are, strictly speaking, conflating two distinct enterprises. Clearly the weak reading of Dancy’s point does not entail the stronger reading. That it’s always possible to make rational sense of intentional actions from a detached point of view does not mean this is the only, or always the best, way to do so. But might the plausibility of the weak reading nevertheless provide support for the 17 stronger reading? The weak reading, it might be said, encourages a certain ideal of detachment. A concern with the explanatory role of the normative reasons that give a point to an intentional action is entirely natural in cases where we relate to the action as deliberators (either directly, in the case of our own actions, or indirectly, insofar as the agent shares her reasons with us). But given that it’s always possible to find na action intelligible from a detached point of view, we should regard that kind of concern as something parochial or perspectival, to be left behind when we engage in the pure project of making rational sense of intentional actions. It’s a good question how much there is to be said, along these lines, for the ideal of detachment.13 But there is a background question, whether the explanatory link between non-factive ‘rationalizing’ attitudes and intentional actions is intelligible on its own. Disjunctivists deny this. On their view, it is not just that there is more than one way to explain intentional actions ‘as such’, but we must acknowledge what is sometimes called the ‘primacy of the good case’. The ability to make sense of actions in terms of non-factive rationalizing attitudes depends on the more basic ability to find actions intelligible as something the agent has a normative reason to do (and is doing for that reason). If disjunctivists are right about this, then realizing the ideal of detachment would deprive us even of the detached mode of understanding that the ideal promotes. I want to conclude by suggesting a way to understand and defend the disjunctivist insistence on the primacy of the good case. Start from the observation that the explanatory connection between attitudes and actions that we invoke in giving 13 Raz 2011 presents grounds for a negative answer, independently of any commitment to a disjunctivist view of rational explanation. 18 rationalizing explanations is in some ways distinctive. Compare and contrast two cases. Maria is yawning because she is tired. This may be a perfectly good explanation. There does seem to be a causal relationship between fatigue and yawning, or so at least commonsense psychology assumes. Saying this is compatible, though, with acknowledging that commonsense psychology has no insight into the reason why fatigue should give rise to yawning. The matter is opaque to most of us, though of course physiologists have produced theories that may get to the bottom of it. Things are different in the case of rationalizing explanations. Max is proceeding to the kitchen cupboard because he feels it’s time for a bite of chocolate and he believes that his chocolate is there. In this case, the explanatory connection between attitudes and action is perspicuous. The reason why the attitudes can be expected to result in the action is transparent to any mature commonsense psychologist: they make it rational or sensible for Max to go to the cupboard (other things being equal). It seems natural to describe the difference in this sort of way. But how does invoking ‘rationality’ make the explanatory link between Max’s attitudes and his action perspicuous or intelligible? One might suggest that understanding the explanatory link requires appreciating the normative import of Max’s attitudes. They give him a reason to go to the cupboard, and perhaps they suggest he ought to go to the cupboard. The trouble is that this cannot be right as it stands. Suppose the chocolate is not in the cupboard (having been transferred to the kitchen table drawer by Max’s officious mother). Then there will be no normative reason that counts in favour of Max’s going to the cupboard, and it’s not the case that he ought to go there. If, knowing this, we reflect that it is nevertheless rational for Max to go to the cupboard we must be concerned with a different question — not the question whether 19 he has reason to go, but perhaps whether his going there should attract a certain kind of criticism. We may be saying that even though he had no normative reason for doing what he did and so his doing it was in fact pointless, still his action was beyond reproach, or he may be excused for performing it. Mere beliefs (beliefs that fall short of knowledge) do not give us normative reasons for action, but they can be mitigating circumstances. I want to suggest that there are two ways in which this analysis would help to corroborate the thesis of the ‘primacy of the good case’. First, to say that certain circumstances excuse a defective performance is to say that the agent could not have been expected to exercise the relevant capacity properly. The capacity whose defective performance mere beliefs mitigate or excuse is the capacity to recognize and act for normative reasons. A proper exercise of that capacity is intelligible as such in terms of having a point or being favoured by a normative reason, recognized by the agent. ‘Bad’ cases are defective (though typically excusable) exercises of the capacities operative in a ‘good’ case, and it’s characteristic of the latter that the exercise of the capacity is explained in terms of a normative reason. Insofar as one sees an action as intentional at all, even if one explains the action in terms of merely ‘rationalizing’ attitudes, one thinks of the agent as having the capacity to act in ways that are intelligible in terms of normative reasons. Second, it is often remarked that understanding the explanatory role of mere beliefs requires certain counterfactual considerations. We appreciate what Max’s belief makes it rational for him to do, for example, by considering what, if the belief were true, there would be a good reason for him to do.14 But why should counterfactual ‘The link between explanation and justification is therefore a link between explaining the action and showing how, had things been as the agent supposed, there would have been most reason to do what he did.’ (Dancy 2000, 10) 14 20 reasons be relevant to understanding the actual explanatory connection between beliefs and actions? Disjunctivists have a ready solution to this puzzle. We understand the explanatory role of mere beliefs on the model of that of knowledge. What merely believing the chocolate to be in the cupboard makes excusable is to act in ways that may in fact be pointless but that, if one’s belief were knowledge, would be intelligible in terms of a normative reason provided, in part, by the fact that the chocolate is in the cupboard.15 Note that this way to understand the primacy of the good case does not invoke the dubious idea that mere belief is ‘subjectively indistinguishable’ from knowledge.16 Nor does it commit one to the idea that concept of knowledge is available independently of the concept of belief. The case for disjunctivism, I suggest, turns on the claim that there is no positive normative relationship in terms of which we might be said to understand the explanatory link between non-factive ‘rationalizing’ attitudes and actions. Correlatively, the challenge facing two-concept theorists is to produce an account of some such relationship. The substantive issue here must be distinguished from terminological matters. One might suggest that doing X blamelessly entails being in a certain sense justified in doing A. This may be a natural 15 This conception of the relation between the two disjuncts echoes some themes in Jonathan Sutton’s account of the nature of epistemic justification. (Sutton 2007) Hornsby’s exposition of disjunctivism is couched in a terms of a distinction between two kinds of reasons (‘fact-type’ vs ‘belief-type’), where the latter are said to ‘inherit the normativity’ of the former (2008, 250) It’s not obvious what Hornsby takes the ‘normativity’ of belief-type reasons to amount to. Perhaps she merely has in mind the weak sense of justification in which (mere) beliefs excusing someone’s doing something might be said to ‘justify’ his doing it, without any implication that there was a normative reason, of any kind, for doing it, or that there was any sense in which agent ought to do what she did. 16 Dancy assumes that to ‘enshrine the primacy of the successful case’ disjunctivists need to offer an account of the nature of ‘mere believing’ in terms of its relation to knowledge. (2008, 273) For dissent, see McDowell (2013). 21 way to use the term, but it leaves open how to understand the operative notion of justification. In particular, the question is whether being justified in doing A entails that one ought to do A, or has reason (in some sense, normative reason) to do A.17 Disjunctivists will deny this. In contrast, two-concept theorists would be well advised to argue that there are two kinds of practical ‘ought’: in Kolodny’s terminology, the ‘ought of reasons’ and the ‘ought of rationality’. (Kolodny 2005) Given his actual circumstances, Max ought to go to the kitchen table, in the ‘ought of reasons’ sense. At the same time, given his false belief about the whereabouts of his chocolate, he ought to go to the kitchen cupboard, in the ‘ought of rationality’ sense. In other words, two-concept theorists would need to argue that there is a distinctive and perhaps weak sense in which ‘motivating reasons’ are normative! This would provide us with a way to find the explanatory role of rationalizing attitudes independently intelligible, with no reliance on what disjunctivists take to be the ‘good case’. I called this task a ‘challenge’ for two reasons. One is that is has proved to be extremely difficult to provide a philosophical account of the putative normativity of ‘rational requirements’. One of the most careful recent attempts to do so, for example, turns out, in the end, to deliver an explanation, not of the ‘ought of rationality’, but of the (illusory) appearance of an ‘ought of rationality’. (Kolodny 2005) The second reason is this. Suppose some philosopher actually succeeds in constructing a credible theory of the normativity of ‘rational requirements’. There would then be an intelligible sense in which Max ought to go to the empty kitchen cupboard. But it This is left open by Davidsons’ famous remark that ‘there is a certain irreducible — thought somewhat anaemic — sense in which every rationalization justifies: from the agent’s point of view there was, when he acted, something to be said for the action.’ (1980, 9) The question is whether, and why, the agent’s view that there was something to be said for the action should be taken to imply that there was, in some sense, something to be said for the action. 17 22 would be a further question whether that sense can plausibly be seen to play any role in our ordinary understanding of the explanatory link between Max’s belief and his intentional action. The truth of the philosophical theory would not be sufficient for that. Rather its content would have to be in some way available to commonsense psychology. Otherwise it could hardly illuminate what renders the connection between the belief and action intelligible to us. A theory that might otherwise look promising may still fall at this hurdle. 4. Conclusion Let us retrace the dialectic. Anscombe’s ‘fundamental scene’, I argued in section 2, lends support to a pluralist view of the rational intelligibility of intentional actions. Sometimes your action makes sense simply in the light of the considerations that provide good reason for it. In other cases we can only make sense of an action from a more detached point of view, in terms of what you thought was to be said in favour of it. Now a two-concept theorist might concede that pluralism is an appealing picture of folk psychology, but insist that there are philosophical reasons to insist on a uniform analysis of the intelligibility of intentional actions in terms of ‘motivating reasons’. Disjunctivism then entered the scene as a way of blocking that line of argument, by denying that ‘motivating reasons’, as conceived by the two-concept theorist, are independently intelligible. But disjunctivism also bears on the question raised at the very beginning, regarding the significance of Anscombe’s ‘fundamental scene’. It suggests a way to understand the rationale of seeking the agent’s view of what makes her action intelligible. It is not that the agent enjoys first-person authority, as this is standardly conceived in the 23 literature. After all, far from being the last word, her account will often prompt challenges and arguments. Rather, the agent’s view matters because a basic question we face in trying to understand intentional actions is whether we can share her view. Acknowledgements Drafts of this paper were presented at conferences in Copenhagen, Granada, and Oxford. I’m grateful to the audiences for helpful comments, in particular to Josef Perner, Glenda Satne, Dan Zahavi, Johannes Brandl, Guy Longworth, Doug Lavin, Christopher Peacocke, Eleonore Stump, Stephen Darwall and Andrew Pinsent. Special thanks to Naomi Eilan for extremely useful suggestions on the penultimate draft. Work on this paper was supported by a grant from the British Academy for a project on The Second Person. Notes on contributor Johannes Roessler is Associate Professor in the Department of Philosophy at Warwick University. He has published articles on issues in epistemology and the philosophy of mind and action, and has co-edited three interdisciplinary volumes, including most recently Perception, Causation, and Objectivity (Oxford University Press, 2011). References Anscombe, G. E. M. 1957. Intention. Oxford: Blackwell. Dancy, J. 2000. Practical Reality. Oxford: Oxford University Press. — 2008: On How to Act — Disjunctively. In A. Haddock & F. Macpherson (eds.), Disjunctivism. Oxford: Oxford University Press. — 2011. Acting in Ignorance. Frontiers of Philosophy in China 6, 345-57. Darwall, S. 2006. The Second-Person Standpoint. Cambridge, Mass.: Harvard University Press. Davidson, D. 1980. Actions, Reasons, and Causes. In his Essays on Actions and Events. Oxford: Oxford University Press. Hornsby, J. 2008. A Disjunctive Conception of Acting for Reasons. In A. Haddock & F. Macpherson (eds.), Disjunctivism. Oxford: Oxford University Press. 24 — 2011. Actions in their Circumstances. In A. Ford, J. Hornsby & F. Stoutland (eds.), Essays on Anscombe’s Intention. Cambridge, Mass.: Harvard University Press. Little, M. O. 2013. In Defence of Non-Deontic Reasons. In D. Bakhurst, B. Hooker, and M. O. Little (eds.), Thinking about Reasons. Oxford: Oxford University Press. Kolodny, N. 2005. Why be Rational? Mind 114, 509-563. McDowell, J. 2012. Acting in the Light of a Fact. In D. Bakhurst, B. Hooker & M. O. Little (eds.), Thinking about Reasons. Oxford: Oxford University Press. Moran, R. 2005. Getting Told and Being Believed. Philosopher’s Imprint 5, 1-29. Raz, J. 2011. From Normativity to Responsibility. Oxford: Oxford University Press. Stroud, Barry 2011. Engagement and Metaphysical Dissatisfaction. Oxford: Oxford University Press. Sutton, J. 2007. Without Justification. Cambridge, Mass.: MIT Press. Thompson, M. 2011. Anscombe’s Intention and Practical Knowledge. In A. Ford, J. Hornsby & F. Stoutland (eds.), Essays on Anscombe’s Intention. Cambridge, Mass.: Harvard University Press. Vogler, C. 2002. Reasonably Vicious. Cambridge, Mass.: Harvard University Press. Wallace, J. 2006. Normativity and the Will. Oxford: Oxford University Press. — 2009. The Publicity of Reasons. Philosophical Perspectives 23, 471-497. Williamson, T. 2000. Knowledge and its Limits. Oxford: Oxford University Press.