Published online by Cambridge University Press: 08 January 2010
Among philosophical questions about human agency, one can distinguish in a rough and ready way between those that arise in philosophy of mind and those that arise in ethics. In philosophy of mind, one central aim has been to account for the place of agents in a world whose operations are supposedly ‘physical’. In ethics, one central aim has been to account for the connexion between ethical species of normativity and the distinctive deliberative and practical capacities of human beings. Ethics then is involved with questions of moral psychology whose answers admit a kind of richness in the life of human beings from which the philosophy of mind may ordinarily prescind. Philosophy of mind, insofar as it treats the phenomenon of agency as one facet of the phenomenon of mentality, has been more concerned with how there can be ‘mental causation’ than with any details of a story of human motivation or of the place of evaluative commitments within such a story.
1 This is Smith's question in ‘The Structure of Orthonomy’, the paper he presented to the conference on which the present volume is based. The quotation from Smith below is taken from the handout he used at the conference.
At the conference Michael Smith responded to my own paper by saying that the standard story could be retold so as to avoid my objections to it. The present, much revised, version is aimed at showing that that which I find objectionable in the standard story cannot simply be evaded. I thank Michael for his contribution to discussion there, Tom Pink and Miranda Fricker for comments they gave me on a draft of the earlier version, and the editors for comments on a draft of the present version.
3 In ‘The Possibility of Philosophy of Action’, in Human Action, Deliberation and Causation, ed. Bransen, Jan and Cuypers, Stefaan (Dordrecht: Kluwer Academic Publishers, 1998), 17–41CrossRefGoogle Scholar, Michael Smith defends the standard story, which he there calls the basic Humean story, to the hilt. He aims to show just how widely the story has application (however much one might need to embellish it in order to deal with all of the various cases). My present concern, one might say, is not with any of the particular claims in that paper, but with the general picture of agency that lies behind it.
4 For a defence of the idea that actions are described in terms of effects they have and the thesis about the individuation that underlies this claim, see Davidson's ‘Agency’, reprinted in Essays on Actions and Events, cit. n.2, 43–61.
5 Cited in previous note. That paper beings with the question ‘What events in the life of a person reveal agency?’. The question puts in place the assumption that the phenomenon of human agency will be delimited when it is said which events are actions.
6 The doubt about assimilating actions to bodily movements mentioned a few paragraphs back is different. The question there was this: Assuming that an agent moved her body on some occasion when there was an action, then is it her bodily movement with which the action is to be identified? Here the question is this: Should the class of actions be so circumscribed that it is required that an agent move her body for there to be an event which is an action?
7 There are no such things as negative particulars: cp. Mellor, D. H.The Facts of Causation (London: Routledge, 1995), pp.131–4CrossRefGoogle Scholar. Notice that the presence of the word ‘not’ in a verb-phrase that applies to some agent need not correspond to there being no action: it could correspond to the occurrence of an action that is negatively described. Equally a positive description can sometimes be given of cases where there is no action (in the standard story's sense): she spoiled the show by not turning up.
8 The simplification assumes that w is a cause of that of which d and b are a cause. It is actually unclear how the ‘and’ of ‘d and b caused m’ (a desire and a belief caused a movement) is supposed to work: cp. nn.13 and 15 infra. But I take it that those who tell the standard story will assent to ‘Desires cause actions’, just as they assent to ‘Desires and beliefs cause actions’.
9 I cannot defend the idea that action-explanation is causal in the present paper. About it, I would say what I say, in n.5 supra about the idea that actions are events
10 I don't think that I need to take issue with Davidson's claims in ‘The Logical Form of Action Sentences’ [reprinted in his Essays on Actions and Events, cit. n.2]: one can accept that some of the sentences that give the explanantia of some action-explanations implicitly contain an existential quantifier whose domain is events, without thinking that any of the explanations is focused on the occurrence of an event. The objection that may be raised here will be that someone who allows any sort of equivalence between ‘a Φ-d’ and ‘there was an event of a's Φ-ing' is compelled to think that action- explanation is the explanation of events’ occurrences. But I think the objection relies on a failure to appreciate the hypersensitivity of ‘explains why …’ contexts. Consider that there may be circumstances in which we are interested to know why Mary stole the bicycle, and other circumstances in which we are interested to know why Mary stole the bicycle; and different answers to the questions will satisfy our interests in the different circumstances. (See Bennett, Jonathan, §14 of Events and Their Names (Oxford University Press, 1988) 32–3Google Scholar (for a spelling out of this example.) My claim is that when someone seeks an action-explanation, typically what she is interested to know is why someone did something.
11 Steward's, Helen challenge, in Part II of The Ontology of Mind: Events, Processes and States (Oxford University Press, 1997)Google Scholar, is devastating. There is no ontological category into which can be lumped both the things which those who tell the standard story call ‘states’ and the things which they call ‘events’.
12 ‘Cause’ is sometimes used in such a way that any ‘because’-statement (or, perhaps, any ‘because’-statement which is genuinely causal) licenses a statement of the form ‘C causes E’. I needn't quarrel with this usage, insofar as ‘Q because P’ might be equivalent to ‘C causes E’ where ‘C’ and ‘E’abbreviate ‘the fact that P’ and ‘the fact that Q’. My quarrel here is with the move from ‘Q because P’ to ‘c caused e’, where ‘c’ and ‘e’ are taken to name something in the category of particulars.
13 A different way to undermine this habit is to show that there are no intelligible causal statements which mix together things in a category of events with things in a category of conditions (where so-called ‘token’ states would need to be reckoned in the category of conditions). This is the conclusion of an argument of Davidson's ‘Causal Relations’ (reprinted in his Essays on Actions and Events, cit. n.2, 149–62). For a spelling out and endorsement of the relevant argument, see Helen Steward, ‘On the notion of cause “philosophically speaking” ’, in Proceedings of the Aristotelian Society, XCVII, 1997, 125–40. From the perspective of Steward's article, it must seem an irony that Davidson's writings about action should so much have influenced those who tell the standard story.
14 I hope that it will be evident now that the view about the operation of causality that I put into question need not be founded in the standard story's conception of events. At the outset of §1, I noted that some philosophers draw on a different conception: an example would be Jaegwon Kim. The criticisms of the standard story in §1 have relied upon a specific conception of events (upon the only conception, I should say, which allows that they are genuinely particulars). But I believe that my claims against “events”—based accounts have application also when “events” is understood in different (but all of them philosophically familiar) ways.
16 Hume's denial of causal powers is well-known of course. But I think that its consequences for an account of human agency are insufficiently appreciated. (I take these consequences to be revealed in the particular passage, though no doubt there could be more argument about this. I don't suggest that Hume really thought that no one can move their limbs. His compatibilist arguments always take it quite for granted that human beings can take their place in the causal nexus.)
18 In Ch. VII of op. cit. n.19, Nagel discusses a question which ‘applies even to the activity of spiders’ before he introduces two different problems relating specifically to human agency. I have not been careful to distinguish here between Nagel's various problems, thinking as I do that they should all be solved together. I have said more about the problem that Nagel calls the problem of autonomy in ‘Agency and Causal Explanation’, in Mental Causation, eds. Heil, J. and Mele, A. (Oxford University Press, 1993) 129–53Google Scholar (reprinted in Philosophy of Action, ed. Mele, A., Oxford Readings in Philosophy 1997).Google Scholar
19 David Velleman introduces his problem about agency by reference to Nagel's puzzle in ‘What Happens When Someone Acts?’, Mind 101, pp. 461–81 (reprinted at 123–43 in his The Possibility of Practical Reason, Oxford University Press, 2000)Google Scholar; but he settles for the view that the standard story is adequate if someone's agency falls short of what is needed for a case of ‘agency par excellence’. I think then that Velleman fails to address what Nagel had supposed to be a quite general puzzle. I discuss Velleman's treatment in this and subsequent of his papers in my ‘Agency and Alienation’, in Naturalism in Question, ed. De Caro, Mario and Macarthur, David (Harvard University Press, 2004).Google Scholar
20 The paper by Bratman on which I mainly focus in the present section is ‘Two Problems About Human Agency,’ Proceedings of the Aristotelian Society (2000–2001) 309–26.Google Scholar (I have just quoted from p.311.) I also draw on his ‘Reflection, Planning, and Temporally Extended Agency’, The Philosophical Review, 109 (2000) 35–61Google Scholar: the phrase ‘embedded in the event causal order’ occurs there. In this, Bratman not only endorses ‘naturalistic psychology’, but also maintains that an agent's identity over time should be treated as consisting ‘primarily in overlapping strands of various kinds of psychological ties’ between ‘states and attitudes’. It is a good question whether such a reductionist conception of personal identity is enforced by ‘naturalistic psychology’, but certainly no accident that someone should subscribe to both.
21 See his Intention, Plans and Practical Reason (Harvard University Press, Cambridge Mass., 1987).Google Scholar
22 ‘Two Problems … ’, op. cit. n.20, p.312.
25 I say this in order to point towards a different direction from Bratman's in which to go in order to answer questions like Bratman's. I don't mean to deny that there are plenty of good questions in the region of Bratman's own.
26 See ‘Reflection, Planning’, op. cit. n.20, p.39
28 I introduce the initial capitals in ‘Agent Causation’ in order to suggest a distinctive doctrine—that of e.g. Richard Taylor (see further, next n.). Kane, Robert, in The Oxford Handbook of Free Will (2002) p. 23Google Scholar, talks about the common practice of introducing a hyphen in ‘agent causation’ in order to indicate that a special kind of relation is intended. Well, many philosophers have had particular theoretical intentions when they have defined notions of agent causation; save for that, I don't think that we'd be inclined to think that there's anything special about it.
Others who are called agent causationists include Timothy O'Connor and John Bishop. There is much agreement between the view I put forward here and theirs. But O'Connor and Bishop both define actions as relations (and perhaps, then, they do introduce a notion which is ‘special’ in the sense meant by Kane). Thus they abandon the idea in the background of events-based accounts, which, as it seems to me, can be perfectly acceptable.
29 And that is why it is not an ordinary notion of agent causation which is used when Agents are said to Cause actions (see preceding n.). Bringing in an agent to do some of the causal work of the states and events of an events-based theory can be a consequence of confusing actions with their effects or results. One sees this confusion in the following passage from Taylor's, RichardAction and Purpose (Prentice-Hall, Englewood Cliffs, 1966.Google Scholar The quotation is at p.lll of the 1973 edition reprinted by Humanities Press, New York; I have added the numbering):
(1) In acting, I make something happen, I cause it…. (2) It seem[s] odd that philosophers should construe this as really meaning, not that I, but rather some event process or state not identical with myself should be the cause of that which is represented as my act. (3) It is plain that … I am not identical with any such event process or state as is usually proposed as the ‘real cause’ of my act. (4) Hence, if … I sometimes cause something to happen, … it is false that an event process or state not identical with my self should be the real cause of it.
The philosophers, the oddness of whose construal Taylor points out at (2), claim that an event, state or process causes his action. But in order to arrive at their claim, one has to confuse ‘acting’ with the ‘something’ that he makes happen, or causes, in acting. In putting the cup on the table (say) that which he makes happen, or causes, is that the cup is on the table. His action, however, is his putting the cup on the table. Thus to represent that some event process or state caused the cup to be on the table is not to represent that some event state or process caused his putting it there. Nothing is said about the cause of his putting it there. Thus Taylor's assertion at (4) can be rejected. An event not identical with the agent (sc. an action) is a cause of that which an agent causes to happen. This does not conflict with Taylor's claims at (1) and (3), which are obviously true.
30 In Ch.VII of Actions (Routledge, 1980),Google ScholarPubMed I suggested that the question of the irreducibility of agent causation comes down to the question whether ‘is an action of can be analysed in terms of event-causal notions. It now seems to me that the principal thesis of that book—namely that actions are events that we always describe in terms of their effects—leads rather directly to the answer No. We don't know which events are a's actions unless we know what a—the agent—caused. And we couldn't know what it is for something to be s's action without knowing that things like a can cause things.