Johannes Roessler (Warwick): The Epistemic Role of Intentions

The Epistemic Roleof Intentions

johannes roessleruniversity of warwick

p r o c e e d i n g s o f t h e a r i s t o t e l i a n s o c i e t y

issue i | volume cxiii | 2012 - 2013

1 8 8 8 | c e l e b r a t i n g 1 2 5 y e a r s | 2 0 1 3

D r a f tP a p e r

p r o c e e d i n g s o f t h e a r i s t o t e l i a n s o c i e t y1 3 4 t h s e s s i o n

i s s u e n o . 1 v o l u m e c x 1 1 12 0 1 2 - 2 0 1 3

t h e e p i s t e m i c r o l e o f i n t e n t i o n s

j o h a n n e s r o e s s l e ru n i v e r s i t y o f w a r w i c k

m o n d a y, 5 n o v e m b e r 2 0 1 2

1 7 . 3 0 - 1 9 . 1 5

t h e w o b u r n s u i t es e n a t e h o u s eu n i v e r s i t y o f l o n d o nm a l e t s t r e e tl o n d o n w c 1 e 7 h uu n i t e d k i n g d o m

This event is catered, free of charge, &open to the general public

c o n t a c t

[email protected]

© 2012 the aristotelian society

b i o g r a p h y

Johannes Roessler is Associate Professor in the Department of Philosophy at Warwick University. He has published articles on issues in the philosophy of mind, epistemology and cognitive development, and has co-edited three interdisciplinary volumes: Agency and Self-Awareness (2003), Joint Attention: Communication and Other Minds (2004) and Perception, Causation and Objectivity (2011).

e d i t o r i a l n o t e

The following paper is a draft version that can only be cited with the author’s permission. The final paper will be published in Proceedings of the Aristotelian Society, Issue No. 1, Volume CXIII (2013). Please visit the Society’s website for subscription information: www.aristoteliansociety.org.uk.

t h e e p i s t e m i c r o l e o f i n t e n t i o n s

j o h a n n e s r o e s s l e r

1. introduction

There is a sharp divide, in the recent literature, between two attitudes to first-person knowledge of intentional actions. Some philosophers take it to be a mark of such knowledge that it requires no grounds. Others regard it as a condition of adequacy on a philosophical theory of such knowledge that it should exhibit the relevant grounds. A suggestive way to make sense of these attitudes has been proposed by David Velleman. The idea that intending to do something typically provides for ‘groundless’ knowledge of what one will intentionally do or is doing, Velleman suggests, is a central element of the ‘commonsense psychology’ of agency. We are indebted to Anscombe and Hampshire for articulating that idea. But we are philosophers as well as commonsense psychologists, and one question we need to address, as philosophers, is whether our putative knowledge in this area ‘is worthy of the name’. (1989: 105) If an affirmative answer to that question is to be defended, we need to establish that, and how, our beliefs in this area meet ‘the usual requirements of evidential support.’ (1989: 25) A central aim of Velleman’s own well-known theory of practical self-knowledge is to show that the sense in which, from a naive standpoint, we take such knowledge to be groundless is consistent with the sense in which, as philosophers, we can make it intelligible in terms of its (highly distinctive) ‘evidential support’.

In what follows I will not be concerned so much with the details of Velleman’s account of how the ‘manifest’ and the ‘philosophical image’ of practical self-knowledge fit together as with the very idea that there is a ‘manifest image’ of the way we know our own intentional actions. Velleman devotes much less attention to this idea than to the philosophical theory that develops from it, and his critics have followed him in this.1 Yet the idea plays a pivotal role in this theory: it supplies the phenomenon the theory is designed to explain, that practical self-knowledge ‘seems to be’ independent of evidence or any other ‘method of discovery.’ (See 1989: 47) What Velleman describes as the commonsense view is of course often regarded as philosophical commonsense. But his claim that practical self-knowledge ‘seems to be’ groundless says more than that philosophers tend to find the idea intuitive. The claim is that the idea is part of the pre-theoretical understanding of practical self-knowledge we have as reflective intentional agents.

That is a substantive philosophical claim in its own right, and one that will strike at least some philosophers as counterintuitive. Opposition to it might take the form of insisting that we are in fact perfectly familiar with some ‘method of discovery’ by which we gain knowledge of our intentional actions. One might maintain, for example, that we typically infer, and can be aware of inferring, what we will do from our current intentions, or that we know what we are currently doing on the basis of a distinctive kind of experience, whose epistemic role is itself manifest to us. On the other hand, one might be sceptical about the very idea of a ‘pre-theoretical’ view of

¹ For illuminating critical discussions of Velleman’s theory of intention and practical self-knowledge, see Bratman 1999 and Langton 2004.

1

practical self-knowledge. Claims to knowledge in this area, one might argue, are not usually challenged or scrutinized. We ordinarily take no particular interest in the sources and credentials of such knowledge. It is not that we are committed to regarding our knowledge as ‘groundless’. We simply have no view, either way; the matter is of no concern to commonsense.

I am interested in a line of argument that can be extracted from Velleman’s discussion, and that, if successful, would provide an effective response to these forms of opposition. The argument can agree with much of what might be said in defence of the second, ‘indifferentist’ view, but it insists that there is a reason we ordinarily show no interest in the source of self-knowledge, and that the reason lies precisely in our pre-theoretical understanding of such knowledge as ‘groundless’. The argument echoes themes familiar from the work of Anscombe and Hampshire, but it is distinctive in its explicit concern with a ‘pre-theoretical’ view of self-knowledge. I begin by sketching the line of argument (section 2). I will then offer a reason to be sceptical about Velleman’s version of it (section 3), and present an alternative, and I think more promising, version, drawing on Stuart Hampshire’s discussion in Freedom of the Individual (section 4). Finally I consider how the more promising version bears on Velleman’s distinction between a naïve and a philosophical conception of self-knowledge. My main point will be that the naïve conception is richer than Velleman allows: it involves a distinctive conception of what makes our knowledge in this area ‘worthy of the name’. In turn, this suggests that the relationship between the naïve and the philosophical perspective is more complicated, and less harmonious, than Velleman assumes.

ii. commonsense psychology

We can distinguish three conceptions of the relation between intending to Φ and knowing that one will, or is Φ-ing. ‘Inferentialists’ hold that intentions constitute part of the evidence from which knowledge of one’s intentional actions is typically derived. ‘Cognitivists’ about intentions (and, on one reading, Anscombe2) hold that in favourable circumstances, to intend is to know.3 The third conception is the one Velleman attributes to commonsense psychology. On this view, intending to f can explain how one is in a position to know one will Φ, or is Φ-ing; specifically, how one is in a position to know this without relying on prior evidence or any ‘method of discovery’. I will label the claim that ‘commonsense psychology’ is committed to this view CS. The view is evidently incompatible with the ‘inferentialist’ conception. On the other hand, it is not obvious that the view is committed to a ‘cognitivist’ conception, though it’s the burden of Velleman’s philosophical theory that in some sense the latter underpins, and helps to vindicate, the commonsense view.

Velleman’s case for CS, as I understand it, is in two parts. The first part formulates a datum that, according to the second part, reflects CS:

² Compare Velleman’s interpretation of Anscombe’s view as holding that intentions ‘embody’ knowledge (see 2007: xiii-xiv).³ Cognitivsm, in the sense made popular by Bratman, consists in the identification of intention with a kind of belief, and practical reasoning with a kind of theoretical reasoning. (See Bratman 1999)

2

3

(1) Suppose you claim that you will attend a certain meeting tomorrow. And suppose someone were to ask you ‘How can you tell?” or ‘What makes you think you will come to the meeting?’ You, or any reflective agent, would find these questions not just surprising but out of place. You would insist, and take yourself to be entitled to insist, that you know without having to ‘tell’. (See 1989: 21)

(2) Your response would be explained and rationalized by the fact that you subscribe to the following explanatory schema: when someone knows what she is, or will be, doing without prior evidence, ‘the reason (..) is that the action was her idea to begin with: it’s what she had in mind to do’. (See 1989: 24)4

How plausible is (1)? As advocates of the ‘indifferentist’ view considered earlier will be quick to point out, the contingency of anyone asking the two questions in (1) is remote. But this is no objection. What matters for Velleman’s purposes is that however unlikely the scenario, your hypothetical response tells us something about your actual understanding of self-knowledge. He might add, of course, that (2) also offers a plausible explanation of why no-one in their right mind would ask these questions. A more serious worry is that some might simply reject the putative datum. Inferentialists may concede that the inferential step from intention to prediction is usually performed automatically and perhaps non-consciously.5 So they may be happy to accept that you would be unlikely to be very articulate about your evidence. But they will question whether you, let alone reflective agents in general, would regard the request for evidence as inappropriate.

Ultimately it is hard to see how this dispute is to be settled if not by reflection on whether it would make sense for you to dismiss the request for evidence. This brings us to (2). It is worth highlighting two (connected) normative claims at work here: that CS provides a good, or the best, explanation of (1), and that the understanding attributed to you by CS makes it rational for you to insist that you know without evidence. Let me labour this latter point. The rationale for attributing a ‘pre-theoretical’ understanding of self-knowledge to reflective agents must be the thought that this enables us to make sense of things they think or do. Your hypothetical response to the request for evidence is perhaps just a melodramatic example. A more commonplace explanandum may simply be our conviction, not usually articulated let alone challenged, that (e.g.) your claim that you will attend the meeting is expressive of knowledge. We think it is because you intend to come that you are in a position to know that you will, without relying on evidence.

This element of rational explanation, which I am suggesting does crucial work in the second phase of the argument, complicates Velleman’s distinction between the naïve and the philosophical standpoint. Part of the motivation for the distinction is to make proper room for the question whether the commonsense view of practical self-knowledge is actually correct. (He charges Anscombe and Hampshire with neglecting that question and accepting the commonsense view uncritically.) That question, he insists, can only

⁴ My summary leaves out some of Velleman’s more contentious claims. In particular, I have made no mention of the putative ‘spontaneity’ of practical self-knowledge, nor of the (apparently connected) idea that we know what we are doing by ‘inventing’ a conception of what we are doing. Neither of these ideas is without its obscurities, and Velleman offers no support for attributing either of them to ‘commonsense psychology’, beyond his case for CS.⁵ See Paul 2009a for an ‘inferentialist’ account along these lines.

4

be answered in the tribunal of a philosophical theory of self-knowledge. This might suggest that whereas our analysis of the commonsense view is simply concerned with factual matters (How do we actually conceive of self-knowledge?), the philosophical theory will tackle normative issues (Are we right? Is our putative knowledge ‘worthy of the name’?). The complication is that a certain — weak — normative claim plays an indispensable role even in articulating the commonsense view, at least insofar as this involves defending CS. The explanatory schema invoked in (2) must make it rational for us to dismiss certain questions and endorse certain attributions of knowledge. The claim is weak insofar as it does not entail the correctness of the schema. It leaves room for the ‘external’ question whether the commonsense schema is philosophically defensible. Still, it suggests that there is also an ‘internal’ normative question: what makes it rational for us to regard someone’s intending to Φ as a reason to think she is in a position to know she will Φ?

I want to suggest that this complication, in turn, poses a challenge to Velleman’s account. On a natural interpretation of his view, commonsense has no insight into the true connection between intending and knowing, which is accessible only from the philosophical standpoint. In light of this, it is unclear what kind of reason we might see to credit intentional agents with knowledge of their actions, and indeed to insist that they know without any need for evidence. That result, however, would undermine the case for CS. In the next section, I consider some possible responses to this challenge. In section 4, I look at an alternative way to fill out the case for CS, which I suggest is not vulnerable to the challenge.

iii. cognitive psychology

The natural interpretation of Velleman’s conception of the two points of view on self-knowledge is as follows. To occupy the naïve point of view is to credit intentional agents with ‘groundless’ knowledge of their actions, and to regard such knowledge as unmysterious, given the explanatory connection between intending and knowing. A proper understanding of that connection requires ascending to the philosophical standpoint. From this standpoint, the claim is, we can see that the commonsensical denial that the knowledge is ‘derived from’ prior evidence is consistent with the philosophical thesis that it ‘rests’ on certain evidence. While your belief that you will come to the meeting is not formed on the basis of adequate prior evidence, there is, once you have formed the belief, a sound evidential basis to support it. The basis includes your holding that very belief. This is not a piece of evidence from which you could have derived your belief. But it can nevertheless help to make it intelligible how your belief qualifies as knowledge, given certain complex features of your psychology, including your desire, or ‘drive’, to do what you believe you will be doing. Finally we need to realize that to intend to go to the meeting just is to hold the self-fulfilling predictive belief that you will. In summary, to say that you intend to attend the meeting is to say that you hold a belief that is epistemically justified in virtue (very roughly) of being of a kind that is normally self-fulfilling. This gives us a good reason to think that, if true, the belief qualifies as knowledge. So your knowledge is ‘worthy of the name’.

The puzzle is how there can be any reason, on this picture, for those occupying the naïve point of view to think that what they get from intending is something worthy

of the name of knowledge. As far as I can see, there are three possible lines of response to this puzzle.

(a) Given only the materials available to the naïve point of view, we have no reason to think that ‘groundless’ first-person claims about one’s current or future actions are sometimes expressive of knowledge.

(b) We do have such a reason: it is provided by our grasp of a naïve theory of intentions, according to which an intention to Φ tends to go with (groundless) knowledge that one will Φ. Commonsense, though, is ignorant of why the generalization holds.

(c) We do have such a reason: it is provided by our (possibly implicit) grasp of the nature of intentions as self-fulfilling predictive beliefs, as set out in Velleman’s theory.

There are grounds for taking (a) to be Velleman’s official view. He elaborates on the commonsense view by drawing a contrast between ‘cognitive psychology’, concerned with the question of ‘how you come into possession of self-knowledge’, and ‘epistemology’, concerned with ‘how self-knowledge is to be justified’. The point of the contrast is to deny that commonsense has anything to say about the epistemology of self-knowledge. Indeed he accuses Anscombe and Hampshire of a kind of category mistake: they read into the naïve view of groundless self-knowledge — a view that belongs to (folk) ‘cognitive psychology’ — a commitment to the epistemological thesis that the notion of evidence is ‘irrelevant’ when it comes to justifying claims to knowledge in this area (1989: 25). Of course, in general an account of how someone came into possession of knowledge that p (e.g. in terms of her observing that p) has an utterly immediate bearing on the assessment of her claim to knowledge. Velleman’s thought seems to be that in the special case of explaining how someone ‘comes into possession’ of knowledge of what she will do in terms of her intending to do it, commonsense psychology is merely concerned with the causal question of how she acquired her attitude, not with whether what she has is knowledge. But if this is right, it is unclear why we should regard the request for evidence as improper. All we could say is that as a matter of fact you consulted no evidence in arriving at your claim.6 The obvious reply would be that you had better start consulting evidence now. Under (a), we would have to agree, insofar as we are rational. This would undermine the case for CS: it would make no sense for us to insist that we know without having to ‘tell’.

(b) offers a somewhat looser reading of the ‘cognitive psychology’ point. The thought would be that while we have no insight into the epistemology of practical self-knowledge, we are familiar with a rough generalization regarding the link between intention and knowledge, a generalization that makes it reasonable for us to attribute knowledge to intentional agents, and to fend off the request for evidence. The idea is reminiscent of functionalism in the philosophy of mind. Grasp of a functional characterization associated with a given mental state is often thought to provide us with the capacity to give psychological explanations, even though we are ignorant of why the relevant generalizations hold. To find out, we’d need to discover the ‘realizer

⁶ Note that on Velleman’s considered view this is not quite right. See p. 67. As Wilson points out, ‘Despite some initial formulations that suggest otherwise, Velleman does not deny that the weighing of certain kinds of pertinent evidence is fundamentally involved in the ‘spontaneous’ practical foresight he has in mind.’ (2000: 4) The evidence in question includes facts regarding one’s motives, as well as one’s practical abilities and opportunities.

5

6

state’ that occupies the functional role in question. Velleman’s account of intentions as self-fulfilling expectations may be interpreted as an attempt to identify the psychological states realizing the commonsense picture of intentional, autonomous agency. 7

This reading may look promising, but I suspect it is precisely the analogy with functionalism that proves to be its downfall. Why should anyone accept the crucial generalization linking intentions and knowledge? Presumably, as with the ‘theory theory’ in the philosophy of mind, the idea would be that our naïve theory is supported by evidence, insofar as it enables us to explain and predict observable phenomena. But then once again, cognitive psychology would have to encompass epistemology: any evidence properly supporting the generalization would need to be evidence that the attitude usually associated with intentions is indeed knowledge. It is hard to see how such evidence could be available to commonsense, if the epistemology of self-knowledge is a closed book to us. We could have no reason to accord groundless self-ascriptions of actions the status of knowledge.

In one of his retrospective glosses on Practical Reflection, Velleman characterizes his theory as holding that ‘knowledge without observation can meet the justificatory standards of internalist epistemology.’ (2004: 229, my emphasis) This remark may encourage (c). My puzzlement over ordinary agents’ reasons to think they know what they are doing may simply rest on a misunderstanding. The goal of Velleman’s theory, you might say, is best described as that of providing a theoretical articulation of ordinary reflective agents’ indistinct, pre-theoretical awareness of the evidential basis of practical self-knowledge. On this reading, the theory is ‘internalist’ in the sense that identifies evidence available to the subject. If the evidence on which practical self-knowledge rests is available to us, we will be able, at least retrospectively, to find our knowledge intelligible and defensible in terms of it, even if our knowledge could not have been derived from that evidence.

But is the putative evidence available to us? This is not a straightforward matter. True, Velleman’s theory in Practical Reflection invokes ordinary propositional attitudes, such as the desire to know what one is doing, or the belief that one will act in a certain way partly as result of one’s expectation that one will. But he cautions that it should not be assumed that the attitudes attributed to us ever surface in speech or conscious thought. (1989: 89) In later work, he develops this theme by relegating the aim of knowing what one is doing to the status of a ‘sub-agential’ aim, an aim that is ‘not represented in our practical reasoning’ and does not ‘operate as a reason for the agent’. (2000: 21) These points suggest that the evidence that is supposed to make claims to self-knowledge intelligible and defensible is not available to us in the sense that matters in the current context: it is not evidence we are supposed ever to take up in ordinary reasoning, which would surely involve speech or conscious thought. It is hard to see, therefore, how it can help to rationalize our conviction that we know without having to ‘tell’.

⁷ This reading accords with Velleman’s general description of his project as that of constructing a ‘hypothetical theory’ designed to ‘identify (..) a kind of personal constitution that would exemplify [put differently, realize] our conception of autonomy’ (1989: 7) It would also be consistent with Velleman’s later claim that his hypothetical theory finds support from empirical research in social psychology. (2007: xvii – xix)

7

iv. practical reasoning

When answering the question whether you will go to the meeting, Hampshire writes, your ‘primary concern in the normal case must be that the policy adopted should not be a mistaken one.’ (1965: 73) A strong basis for an affirmative answer would consist of considerations convincing you that it would be ‘an unqualified mistake not to go to the meeting’ (72). It is natural to think that it is here that we find the true rationale for CS. The reason you take your claim not to be intelligible in one way — in terms of adequate evidence, or some other ‘method of discovery’ — is that you do find it intelligible in another way, in terms of what you take to be good reasons for attending the meeting. Practical reasoning is something that is missing from Velleman’s austere characterization of the commonsense view. This is unsurprising, given that Velleman is interested only in those aspects of ‘commonsense psychology’ that he thinks his philosophical theory is capable of vindicating, and given that the manifest image of practical reasoning is not one of them.8 However, if Hampshire’s account of practical self-knowledge is on the right lines, the two ingredients of the manifest image are inextricably linked. To say merely that we subscribe to a functionalist-style platitude linking intentions and knowledge is to distort commonsense psychology by omitting our grasp of why the generalization holds. Hampshire’s account suggests it’s our understanding of practical reasoning that reveals to us the categorical basis of intentions’ tendency to be attended with knowledge of what one is doing.

The leading character in Hampshire’s discussion is a distinctive kind of statement or claim. A naïve ‘epistemology’ of practical self-knowledge emerges from an analysis of its import and basis. The statement is introduced as follows. Consider two postcards, both addressed to the secretary and both bearing the message ‘Hampshire will attend the meeting’. Here is something the two postcards have in common: they both aim to inform the secretary that Hampshire will attend the meeting. To succeed in this aim they must express propositional knowledge. The difference is that while the message on the first postcard, written by a friend, expresses knowledge based on inference from the premise that Hampshire thinks it would be a great mistake not to attend the meeting (no doubt in conjunction with other premises), the message on the second postcard, composed by the man himself, is a statement directly based on Hampshire’s practical judgement that it would be a mistake for him not to go to the meeting. We can distinguish two theses Hampshire makes about the second statement. One is that it has a ‘double aspect’: it is subject to two distinct ‘criteria’ or ‘tests’ for correctness. The second thesis is that practical reasoning can provide an adequate warrant for the statement. Let’s consider the two theses in turn.

According to the first thesis, there are two kinds of norms the statement has to satisfy to count as correct. First, it must be ‘a reliable guide to the future’. Second, ‘the intention announced’ must not be ‘a misguided or confused one’. (1965: 72) To satisfy the first criterion, it is natural to suppose, the statement must express knowledge. In this respect, it is no different from assertions in general: knowledge is the constitutive rule for assertion. (Williamson 2000) Admittedly, Hampshire seems to endorse a weaker, disjunctive position. He characterizes the two statements on the postcards as ‘claims to knowledge of the future, or at least to well-supported beliefs about the future’. Yet it is

⁸ Corroborating this judgement is a task I cannot attempt in this paper. A useful starting point would be Velleman’s explicit rejection of a ‘reactive’ picture of practical reasoning, on which it is matter of ‘calculation of how to cope with a given predicament’, in favour of his constructivist view that ‘our predicaments aren’t given at all: we invent them’. (1989: 9)

8

noteworthy that, as far as the second, practical reasoning-based statement is concerned, he clearly takes the first criterion to be met in virtue of the statement’s expressing knowledge. (He makes no mention at all of the second disjunct in this context.) As we will see, there is an element of a ‘knowledge first’ epistemology avant la lettre here. It may seem, then, that Hampshire must be committed to a ‘cognitivist’ conception of the relation between intention and knowledge: if correct, the statement will announce an intention that ‘embodies’ knowledge. But that reading is certainly not demanded, and I think not licensed, by the text. The ‘double aspect’ claim is more naturally read as saying that the statement simultaneously expresses two attitudes, intention and knowledge.9

But how can practical reasoning provide an adequate warrant for a claim to knowledge, as the second thesis can now be seen to imply? At times Hampshire himself appears mildly puzzled by this. ‘How can I pass (..) from an evaluation of the practical possibilities to a statement of fact about the future?’, he asks. (1965: 74) His attitude, however, is not that of a philosopher who finds himself confronted by a ‘hard problem’. He seems intrigued rather than mystified. I think he really takes it as a given, in a Moorean fashion, that his claim that he will attend the meeting can be warranted by his practical reasoning. It is sometimes said to be a structural element of practical reasoning that a judgement that one has most reason to Φ licenses the formation of an intention to Φ, with no need for an extra ‘step’ consisting (say) of an expression of the desire to do what one takes oneself to have most reason to do. (Stroud 2003) Hampshire can be seen to make a related point. It is a structural element of practical reasoning, he claims, that an intention to Φ based on successful practical reasoning licences the ‘double-aspect’ claim that one will Φ, with no need for an extra premise to the effect, say, that one is likely to act as intended. Hampshire frequently emphasizes that the license depends on one’s conviction that it will be in one’s power to Φ. Not that the content of this conviction is needed as an extra step preceding the claim that one will Φ. Rather it must inform one’s prior practical reasoning, aimed at finding a workable ‘policy’ for attaining some given end. We might add that if the claim that one will Φ is to be not just rational, but correct, by the lights of the first ‘criterion’, — i.e. if it is to express knowledge —, a conviction, even a rational conviction, would not be enough: one surely has to know that it is in one’s power to Φ.10

We can now see more clearly what is at stake in the debate over CS. If Hampshire is right about the structural element, there is an intelligible link between successful practical reasoning and being in a position to make a warranted claim to knowledge. Reflection on one’s practical reasons for Φ-ing can thus provide one with a good reason to think one knows one will Φ, and can make it intelligible to one how one is in a position to know this.11 Evidence may of course play a vital role in establishing whether it is in

⁹ See also Falvey 2000 and Broome 2009. Broome identifies the two attitudes expressed by the statement as intention and belief.¹⁰ Setiya 2008 emphasizes the role of ‘knowledge how’ in making practical self-knowledge possible, though in the framework of a reductive account of intention as a kind of belief — a framework apt to make it hard to see how that knowledge explains knowledge of what one is doing if not by providing evidence for belief. For helpful discussion of these issues, see Paul 2009 and Setiya 2009.¹¹ Does this mean, as Charles Taylor maintains, that we gain knowledge of our intentional actions by articulating the content of our intentions? (See Taylor 1985) It does not seem implausible that someone’s inability to get herself to say, or think, what she is up to can justify the diagnosis that she is not fully aware of what she is doing. One problem with Taylor’s description, though, is that it might suggest that we do after all deploy a certain ‘method of discovery’. That is not so, on Hampshire’s analysis. In any case, what is necessary for possession of knowledge here is arguably the ability to articulate one’s intention, not one’s having exercised that ability.

9

one’s to power to Φ. But if one knows that it is, and on that basis forms the intention to Φ, one is entitled to judge or assert that one will Φ, without any further mediating step. What entitles one to the judgement is not evidence but successful practical reasoning. The absence of any intermediate step would make it unsurprising that we tend to use the same statement to express both attitudes, intention and knowledge.12 Those who reject CS deny that commonsense psychology is committed to the putative structural element. On their view, practical reasoning, as ordinarily conceived, will get you to a judgement that you have most reason to Φ, or possibly to the formation of an intention to Φ, but it cannot license a ‘statement of fact about the future’.

Doubtless there are cases that seem to fit this latter conception of practical reasoning. You may without irrationality intend to move a log blocking your driveway without being convinced that you will succeed.13 So you don’t think your practical reasoning warrants the claim that you will move the log, and surely you may be right about this. Cases such as this may be thought to bring to light a general gap between intentions and warranted ‘statements of fact about the future’. In response, Hampshire might say that while the practical reasoning in such cases need not be faulty, it is not wholly successful either. Intuitively, your plan to move the log is suboptimal. It would be even better to find a method that only involves things you know you’ll be able to do. This intuition, one may argue, reflects the essential objective of ‘calculative’ practical reasoning, to find a means that will enable one to attain a given end.14 To reach that objective the reasoning must be informed by facts as to what it is in one’s power to do. In the log case, that is so only in the attenuated form of your believing you may be able to move the log. In the case of more successful practical reasoning, facts regarding the reasoner’s practical abilities will be available to her in the shape of knowledge, or at least well-supported belief. Accordingly, the reasoning will warrant a ‘statement of fact about the future’. If it did not settle the question of what she will do, but merely the question of what she is to do, the reasoning could not be an unqualified success, simply considered as practical reasoning.15

iv. epistemology

A way to summarize the argument so far is to say that Velleman’s strong commitment to an ‘evidentialist’ epistemology leads him to distort the way we ordinarily think about the epistemic role of intentions. He assumes that the only conceivable way to corroborate our claims to knowledge in this area lies in an account of the way our beliefs meet the requirements of evidential support. This leaves him bereft of an explanation of how it can seem reasonable to us to claim to be able to know what we do without evidence, undermining his case for CS. It is a virtue of Hampshire’s analysis

¹² It would be instructive, on this view, to compare the demand for supporting evidence with Lewis Carroll’s tortoise’s insistence that a rule of inference must be represented as an additional premise if one is to have any reason to accept the conclusion. See Stroud 2003 and Stroud 2011 for illuminating discussion of Carroll’s point and its bearing on practical reasoning.¹³ See Bratman 1987 and Holton 2009 for discussion. Compare also Anscombe’s remark that ‘the less normal it would be to take the achievement of the objective to be a matter of course, the more the objective gets expressed only by ‘in order to’. E.g. ‘I am going to London in order to make my uncle change his will; not ‘I am making my uncle change his will’. ‘ (1957: 40). ¹⁴ See Vogler 2002 for a rich exposition and defence of the ‘calculative’ conception of practical reasoning.¹⁵ This is a toned-down version of a point forcefully made by Moran (2000: 127).

10

i.

that it does offer such an explanation. On the other hand, that Velleman misrepresents ‘commonsense psychology’ is of course no reason to think the evidentialist approach itself is mistaken. One might acknowledge that we ordinarily find it intelligible how intentions can put us in a position to know what we are doing in a way that has nothing to do with ‘evidential support’. At the same time, one might insist that unless we can identify some kind of evidential basis underpinning the commonsense explanatory schema, we should be sceptical about the ‘folk epistemology’ embodied in the schema. I want to conclude by raising some doubts about the force of this conditional claim.

Consider first the worry that forming beliefs about one’s future actions without prior evidence would be ‘epistemically suspect’. (Setiya 2008, Paul 2009b) The obvious response, from Hampshire’s perspective, is that it can hardly be suspect to believe p when one knows that p and has a good reason to think one knows that p. If you form a firm and realistic intention, it is reasonable for you to think you know what you will do.16 Successful practical reasoning, after all, puts you in a position to make a warranted statement of fact — hence a warranted claim to knowledge — about the future. But that response will hardly satisfy those exercised by the worry. The worry will now be that it would be ‘epistemologically suspect’ to take the link between well-grounded intentions and knowledge as primitive. To earn the right to rely on the link, it might be said, we need an independent philosophical account of how beliefs based on practical reasoning acquire the ‘status of knowledge.’

There are a variety of ways one might try to construct such an account, but I think they come up against the same basic difficulty. Not only does the commonsense schema make no mention of any grounds for forming beliefs about one’s actions, but the way in which it makes self-knowledge intelligible seems to preclude an understanding of such knowledge in terms of adequate grounds for belief. Consider the suggestion that we form beliefs about our intentional actions by the use of the following procedure: we answer the practical question of what to do by expressing our intentions, and then endorse the result as a description of what we are, or will be, doing.17 If this were so, we should find our ‘statements of fact’ intelligible, not simply in terms of successful practical reasoning but in terms of our use of the procedure. So the suggestion would amount to a revision, not a vindication, of the commonsense explanatory schema. One might seek to avoid that problem by switching to a purely reliabilist version of the proposal: to say that expressions of intentions are an intelligible epistemic basis for beliefs about the future is

¹⁶ It important that success in practical reasoning is not ‘transparent’ to the subject, in the sense of transparency discussed by Williamson 2000: the subject is not always in a position to know whether her reasoning is successful. Indeed on the ‘calculative’ conception of practical reasoning, it is natural to think that success in reasoning requires success in action: only practical reasoning resulting in an action leading to the attainment of one’s end will count as a successful case of reasoning about how to attain that end. The point is important as it suggests a plausible response to what might otherwise look like a serious concern about Hampshire’s account. The concern is: how can something that ‘falls short of’ the fact that one is Φ–ing — something that is consistent with one’s not Φ–ing — put one in a position to know that one is Φ–ing? The response is that truly successful practical reasoning does not fall short of the intentional action performed on account of it.¹⁷ This would be one way to read Falvey’s claim that agents have a ‘general warrant’ to ‘present’, or ‘employ’, their expressions of intentions as descriptions of what they are doing. (2000: 23, 37) The reading is at least encouraged by what seems to be an implication of Falvey’s account, that the ‘presenting’ or ‘employing’ is to be distinguished from the act of expressing one’s intention itself. On this reading, the ‘general warrant’ is a warrant to believe that one is doing Φ–ing on the basis of expressing an intention to Φ.

11

not to say that the subject must find her beliefs intelligible in this way. Like Velleman’s analysis of ‘commonsense psychology’, this would make the subject completely ignorant of how it is that intentions generally provide for knowledge. It would therefore, like Velleman’s analysis, be at a loss about how it can be rational for the subject to regard the request for evidence as inappropriate. A further variation of the current proposal would be to argue that our beliefs in this area are justified in virtue of meeting a requirement of practical rationality. Some philosophers have argued that an intention to Φ commits one to the belief that one will. (Moran 2000, Wilson 2000) This line of thought may seem to be more in keeping with the spirit of the commonsense schema. But it faces a daunting challenge of its own, to explain how meeting the demands of practical rationality can provide for epistemic justification (as opposed to securing the status of something like a ‘postulate of practical reason’).

These brief remarks give a flavour of the kinds of challenges facing the programme of a philosophical vindication of the commonsense schema. It would be rash to conclude that the programme must fail, but the difficulties confronting it at least make it worth posing the question what would follow from its failure. Should we conclude, as uncompromising ‘evidentialists’ should argue, that the ‘manifest image’ of practical self-knowledge has been debunked or discredited? Or should we deny that the explanatory value of the commonsense schema has to be underwritten by an independent account of how the relevant beliefs acquire the status of knowledge? Some will dismiss the ‘manifest image’ as a mere prejudice, others will use that label to brand their opponents’ ‘evidentialist’ epistemology. We find ourselves confronted here, once again, by the two attitudes from which we started. But how is the disagreement to be adjudicated?

It seems to me that while Hampshire is primarily concerned to articulate the naïve view, not to defend it against any general challenge, his analysis nevertheless enables us to see why it would be difficult to formulate and sustain such a challenge. One response to the challenge might be to redeploy what Velleman describes as the natural retort to the ‘first-order’ challenge ‘how can you tell you will Φ?’, viz. ‘I don’t have to “tell”. What makes this a sensible response, as Hampshire explains, is that careful practical reasoning provides a warrant for a claim to knowledge, where this means the knowledge is not in need of evidential support. One might invoke the same warrant to show that evidentialist philosophers’ scruples about the ‘manifest image’ are misplaced. They may protest that this begs the question; their challenge is intended to question precisely whether practical reasoning can actually ever provide such a warrant. On the face of it, though, this amounts to questioning whether practical reasoning can ever be successful, or perhaps whether there is such a thing as practical reasoning as ordinarily conceived. If the evidentialist philosophers’ answer to that question is negative, the next question would be whether this leaves room for the thought that people sometimes act intentionally, i.e. for what they take to be good reasons. Each of these moves would require detailed examination. What they would show, if correct, is that scepticism about the commonsense schema would take with it much more than one might at first realize.

Department of Philosophy

Social Sciences Building

University of Warwick

Coventry CV4 7AL

12

references

Anscombe, E. 1957: Intention. Oxford: Blackwell.

Bratman, M. 1987: Intention, Plans, and Practical Reason. Cambridge, Mass.: Harvard University Press.

— 1999: ‘Cognitivism about Practical Reason’, in his Faces of Intention. Cambridge: Cambridge University Press.

Broome, J. 2009: ‘The Unity of Reasoning’, in S. Robertson et al, (eds.), Spheres of Reason. Oxford: Oxford University Press.

Falvey, K. 2000: ‘Knowledge in Intention’, Philosophical Studies 99, 21-44.

Hampshire, S. 1965: Freedom of the Individual. London: Catto and Windus.

Langton, R. 2004: ‘Intention as Faith’, in J. Hyman and H. Steward (eds.), Agency and Action. Cambridge: Cambridge University Press.

Moran, R. 2001: Authority and Estrangement. Princeton: Princeton University Press.

Paul, S. 2009a: ‘How we know what we’re doing’, Philosophers’ Imprint 9, 1-24.

— 2009b: ‘Intention, Belief and Wishful Thinking: Setiya on “Practical Knowledge”’, Ethics 119, 546-557.

Setiya, K. 2008: ‘Practical Knowledge’, Ethics 118, 388-409.

Stroud, S. 2003: ‘Weakness of Will and Practical Judgement’, in S. Stroud and C. Tappolet (eds.), Weakness of Will and Practical Irrationality. Oxford: Oxford University Press.

Stroud, B. 2011: Engagement and Metaphysical Dissatisfaction. Oxford: Oxford University Press.

Taylor. C. 1985: ‘Hegel’s philosophy of mind’, in his Human Agency and Language (Philosophical Papers 1). Cambridge: Cambridge University Press.

Velleman, D. 1989: Practical Reflection. Princeton: Princeton University Press.

— 2004: ‘Précis of the Possibility of Practical Reason’, Philosophical Studies 121, 225-238.

— 2007: ‘Introduction to the David Hume Series Edition’ [of Practical Reflection]. Stanford: CSCI Publications

Vogler, C. 2002: Reasonably Vicious. Cambridge, Mass.: Harvard University Press.

Williamson 2000: Knowledge and its Limits. Oxford: Oxford University Press.

Wilson, G. 2000: ‘Proximal Practical Foresight’, Philosophical Studies 99, 3-19

president: Sarah Broadie (St. Andrews)

president-elect: E.J. Lowe (Durham)

honorary director: Lucy O’Brien (UCL)

editor: Matthew Soteriou (Warwick)

lines of thought series editor: Scott Sturgeon (Oxford)

executive committee: Ben Colburn (Glasgow) / Alison Hills (Oxford) / Rosanna Keefe (Sheffield)Marie McGinn (UEA) / Samir Okasha (Bristol) / Ian Rumfitt (Birkbeck) / Robert Stern (Sheffield)

executive administrator: Mark Cortes Favis

assistant editor: David Harris

editorial assistant: Lea Salje

t h e a r i s t o t e l i a n s o c i e t y

w w w. a r i s t o t e l i a n s o c i e t y . o r g . u k