RECENT WORK MORAL RESPONSIBILITY. Introduction

Transcription

1 Oxford, PHIB Philosophical /j x 46 4Original Blackwell UK Article Publishing, Books Ltd. RECENT WORK MORAL RESPONSIBILITY ELINOR MASON The University of Edinburgh Introduction In this account of recent work on moral responsibility I shall try to disentangle various different sorts of question about moral responsibility. In brief, the tangle includes questions about whether we have free will, questions about whether moral responsibility is compatible with free will, and questions about what moral responsibility involves. As far as possible I will ignore the first sort of question, be as brief as possible on the second sort of question, and focus on the third question. This is partly just in the interests of space the total literature generated by the three questions together would be impossible to summarise here For survey articles see John Martin Fischer, Recent Work on Moral Responsibility, Ethics, 110 (1999), pp ; there are several excellent survey articles online: see the Stanford Encyclopedia of Philosophy entries on Moral Responsibility (by Andrew Eshleman), Compatibilism, and the supplement: Compatibilism: The State of the Art (by Michael McKenna), Incompatibilism (by Kadri Vihvelin) and Incompatibilist Non-Deterministic Theories of Free Will (by Randolph Clarke). For collections of recent articles see Gary Watson (ed.), Free Will (2nd edition, containing 11 new contributions, Oxford University Press, 2003); David Widerker and Michael McKenna (eds.), Moral Responsibility and Alternative Possibilities (Ashgate Press, 2002); Robert Kane (ed.), The Oxford Handbook of Free Will (Oxford University Press, 2002); Sarah Buss and Lee Overton (eds.), Contours of Agency: Essays on Themes from Harry Frankfurt (MIT Press, 2002); Laura Ekstrom (ed.), Agency and Responsibility (Westview Press, 2001); Philosophical Perspectives, vol. 14, Action and Freedom, (Blackwell, 2000). Recent book length discussions of free will and moral responsibility include Randolph Clarke, Libertarian Accounts of Free Will (Oxford University Press, 2003); Daniel Dennett, Freedom Evolves (Viking Press, 2003); Alfred Mele, Motivation and Agency (Oxford University Press, 2003); Derk Pereboom, Living Without Free Will (Cambridge University Press 2001); Laura Ekstrom, Free Will: A Philosophical Study (Westview Press, 2000); Timothy O Connor, Persons and Causes: The Metaphysics of Free Will (Oxford University Press, 2000); Saul Smilansky, Free Will and Illusion (Oxford University Press, 2000); Hilary Bok, Freedom and Responsibility, (Princeton University Press, 1998); John Martin Fischer and Mark Ravizza, Responsibility and Control: An Essay on Moral Responsibility (Cambridge University Press, 1998); Ishitaque Haji, Moral Appraisability (Oxford University Press, 1998); Robert Kane, The Significance of Free Will (Oxford University Press, 1996); R. Jay Wallace, Responsibility and the Moral Sentiments (Harvard University Press, 1994). Gary Watson s collected papers on the topic have just been published as a book: Agency and Answerability (Oxford University Press, 2004). 343, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

2 It is impossible to ignore the issues of free will and determinism entirely. Moral responsibility is a problem because our best theories about our physical world tell us that our actions are caused by mechanistic processes originating outside of us. It seems that if our actions are ultimately caused in the same way that avalanches are caused, then ultimately we are no more responsible for what we do than a rock is for knocking into another rock. The problem has led some philosophers to argue that determinism is false, and that either agents are special causes (and thus morally responsible for what they do) or that indeterministic events can account for moral responsibility. I will leave aside these contributions to the debate in this brief survey, and concentrate on recent compatibilist positions. Compatibilists argue that although our actions are indeed ultimately caused by events outside of us, we can still make sense of moral responsibility. 2 There are two sorts of project for the compatibilist, and it might be useful for a moment to compare them to two projects in moral philosophy. Moral philosophers often think of themselves as falling into two camps those who are concerned with meta-ethics questions about the truth of our moral judgements, and those who are concerned with normative ethics questions about how our moral judgements should be formed, what sort of shape our moral reasons have, and so on. At first glance there is a parallel distinction in discussions of moral responsibility: on the one hand we might ask, what are we doing when we make judgements about moral responsibility? The map of possible answers looks very like the map of possible answers about moral judgements. At one end of the spectrum the answer can be a non-cognitivist one: we are expressing an attitude. On a common interpretation of Strawson s famous and influential article, this is the position he is advocating. 3 Strawson s argument in a nutshell is that we should understand responsibility in terms of our practice of engaging in reactive attitudes. We emotionally react to what we perceive to be the quality of other people s wills, and according to Strawson, the truth of determinism has no bearing on this practice. Strawson s framework for thinking about responsibility remains extremely influential, though contemporary writers have modified the view in all sorts of ways. At the other end of the spectrum there is cognitivism: the view that when we assign moral responsibility to someone for something we are making a cognitive judgement that certain conditions obtain. Scanlon s view is an example of this strategy, 4 and more recently Hilary Bok has defended a Kantian cognitivist account of responsibility. There is also some logical space in the 2. Sometimes the word compatibilism is used slightly differently, to refer to the view that free will is compatible with determinism. Hence Fischer and Ravizza (Responsibility and Control ) refer to their view as semi-compatibilism because they believe that determinism is compatible with moral responsibility but not with free will. 3. P.F. Strawson, Freedom and Resentment, Proceedings of the British Academy, 48 (1962), pp Other non-cognitivist views are presented in Daniel Dennett, Elbow Room (Cambridge University Press 1984) and Ted Honderich, A Theory of Determinism, (Oxford University Press, 1988). R. Jay Wallace, Responsibility and the Moral Sentiments (Harvard University Press 1996) develops Strawson s account, though, as I say above, his view is partially cognitivist. 4. See Thomas Scanlon, The Significance of Choice, in Sterling M. McMurrin (ed.), The Tanner Lectures on Human Values, vol. 8 (University of Utah Press, 1988). 344

3 middle Jay Wallace, for example, argues that our stance of holding people responsible is a sui generis stance it has a cognitive element in that it is essentially about what we take the agent to have done, but it necessarily involves a reactive emotion too. It seems that we can ask a different sort of question about the conditions that excuse someone from responsibility for something they have done (or, the broader question: what excuses a being from having responsibility at all?). In this arena we encounter debates about whether a deprived childhood is an excusing condition, whether lack of knowledge is an excusing condition, whether threats and duress count as excusing conditions, and so on. The literature about this question is extremely varied, and not all of it is explicitly about moral responsibility, but I would include debates about action, autonomy, intention, agency, weakness of will, moral luck and the principle that ought implies can. 5 It seems that we can set aside the meta-question about what we are doing when we assign responsibility, as we do in ethics, and concentrate on the normative questions. However, there is an important difference between the cases of moral judgements and judgements about moral responsibility. In moral philosophy, debates between deontologists and consequentialists go on at the normative level without much concern for the meta-ethical level. There is an assumption that whichever way the meta-ethical questions are resolved, there will still remain questions about the shape of the right normative ethical theory. In contrast, the meta-level challenge to our judgements about responsibility appears to undermine our normative level judgements, and so the two projects (the meta-level and the normative level projects) have been combined in the literature on moral responsibility in a way that has not happened so much in moral philosophy. 6 The explanation for this is what Wallace calls the generalisation strategy : the sorts of excuse that we accept as defeating responsibility in particular cases appear to generalise to show that if determinism is true then we do not have responsibility at all. For example, Bill explains that he was under hypnosis when he killed the cat. A natural thought is that if he was under hypnosis then he could not have done otherwise, and so he is not responsible. The 5. For example, Alfred Mele has done a lot of work on autonomy and action theory: see Alfred Mele, Autonomous Agents: From Self-Control to Autonomy (Oxford University Press, 1995); Motivation and Agency; Agents Abilities, Noûs, 37 (2003), pp ; and Ultimate Responsibility and Dumb Luck, Social Philosophy and Policy, 16 (1999), pp ; see also David Velleman, The Possibility of Practical Reason (Oxford University Press, 2000). On the Ought implies Can principle see Walter Sinnott-Armstrong, Moral Dilemmas (Oxford, 1988); Roy Sorenson, Unknowable Obligations Utilitas, 7 (1995), pp ; Frances Howard-Snyder, The Rejection of Objective Consequentialism, Utilitas, 9 (1997), pp ; Elinor Mason, Consequentialism and the Ought implies Can principle, American Philosophical Quarterly, 40 (2003), pp On the role of knowledge, see Michael Zimmerman, Moral Responsibility and Ignorance, Ethics, 107 (1997), pp This is not a complete bibliography of the work on moral responsibility outside the free will debate, rather a representative sample. 6. I am just describing the history of the debates here it may be that the truth of some metaethical position undermines some or all normative views, and of course, if determinism undermines our normative judgements about excuses, then it also undermines our normative theories about rightness. Philosophers engaged in discussions of consequentialism versus deontology have thus shelved the debate about moral responsibility as well as the debate about meta-ethics. 345

4 incompatibilist seizes on this, and points out that if determinism is true, then we can never do other then what we do, so by our own lights, we are never responsible. Thus it seems that the two compatibilist projects must go hand in hand: the compatibilist has to show that our normative level judgements about excuses and exemptions do not assume the falsity of determinism. 7 This is the challenge for compatibilist theories of moral responsibility. Compatibilists must show that there is a sensible concept of moral responsibility a way of making judgements about who is responsible and when that does not incorporate an assumption that determinism is false. In which case, many of our natural every day explanations for why someone is not responsible, such as he couldn t do anything else and he had no alternative must be re-parsed in such a way that they are not vulnerable to the generalisation strategy. There is an obvious pitfall to be avoided the problem must not be shifted to another level. It is no use saying that an agent is responsible for an action when she was in control of it, or when she chose it, or when it reflected her rational capacities. Terms such as ability, capacity, choice, control and so on are ambiguous they have the stink of metaphysical freedom about them. The compatibilist has to show that the most basic terms of her explanation are non-question begging. The compatibilist must equally avoid the complaint at the other extreme, that her view does not capture what is deep or important about moral responsibility. Undergraduates are quick to point out that Hume s compatibilism seem unsatisfactory because although the agent may be able to do what she chooses to do, the truth of determinism implies that she could not have chosen otherwise. 8 On the other hand, it would be unfair if the complainer were simply assuming that moral responsibility requires the falsity of determinism. The traditional landscape divides up into two opposing views. On the one hand are real self views that locate moral responsibility in the agent s real self. On the other hand are what I will call intellectual process views, that assign moral responsibility to agents who have gone through a particular intellectual process in arriving at their decision. In what follows I shall provide a brief history of each of these positions, and go on to discuss recent work in more detail. Real Self Views Real Self views try to show that we hold (or, at least, should hold) an agent morally responsible when her action issues from her real self. 9 The hard question for this sort of view is what is the real self? what makes an action one that really is the agent s? Different theorists have tried to answer this in 7. This exposition of the problem owes a lot to Wallace s discussion (see Responsibility and the Moral Sentiments). 8. Hume, Of Liberty and Necessity in An Inquiry Concerning Human Understanding. Saul Smilansky presses the shallowness objection in Free Will and Illusion (Oxford University Press 2000) and in Compatibilism: The Argument from Shallowness, Philosophical Studies, 115 (2003), pp These views are also known as structuralist views or internalist views (see Fischer and Ravizza, Responsibility and Control ), and Self-disclosure views (see Gary Watson, Agency and Answerability). 346

5 different ways. Harry Frankfurt was one of the first to develop the view. 10 Frankfurt argues that we are responsible for our acts when we act according to our second order volitions. A second order volition is a desire that a first order desire be effective, i.e. issue in action. So, for example, I might have a first order desire to eat some tofu, and a second order desire that I follow through and eat it. On other hand I might have a first order desire to eat some candy that is not endorsed at the second order my hand reaches out and grabs the candy, but without my consent as it were. This action is not one that I am responsible for, on Frankfurt s view. Frankfurt on Alternate Possibilities Part of Frankfurt s compatibilist strategy is his argument that alternate possibilities are not necessary for moral responsibility. 11 His argument involves a thought experiment: imagine that a powerful scientist has control over your brain, and will intervene if you are about to choose A and make sure that you choose B. If you are about to choose B, the scientist will do nothing. Frankfurt contends that if you choose B on your own, you are responsible for doing so it doesn t matter that you had no alternative. The fact that you had no alternative plays no part in the story it is irrelevant. So the argument against the principle of alternate possibilities (PAP, as it is often called in the literature) supports the view that moral responsibility depends on the act being the agent s own. This aspect of Frankfurt s argument has independently spawned an enormous literature. 12 In the interests of space I shall leave aside this issue here. Back to Real Self Views Gary Watson criticises Frankfurt s version of the real self view on the grounds that the higher order desires cited by Frankfurt might themselves be arbitrary. Frankfurt responds by suggesting that an agent is morally responsible when her action is wholehearted, and changes his mind again in The Faintest Passion where he suggests that the crucial notion is that an agent be satisfied with her higher order desire. 13 Frankfurt s progression here represents his struggle with 10. Harry Frankfurt, Alternate Possibilities and Moral Responsibility, Journal of Philosophy, 66 (1969), pp , and Freedom of the Will and the Concept of a Person, Journal of Philosophy, 68 (1971), both reprinted in The Importance of What We Care About (Cambridge University Press 1988). His more recent work is collected in Necessity, Volition and Love (Cambridge University Press 1999). 11. Frankfurt, Alternate Possibilities and Moral Responsibility. 12. Recent discussions of Frankfurt cases include: Ishtiyaque Haji, Flickers of Freedom, Obligation, and Responsibility, American Philosophical Quarterly, 40 (2003), pp ; Alternative Possibilities, Luck, and Moral Responsibility, The Journal of Ethics, 7 (2003), pp ; Alfred Mele and David Robb, Rescuing Frankfurt-Style Cases, Philosophical Review, 107 (1998), pp For a survey of some of the work on the flicker of freedom view see Fischer, Recent Work. Moral Responsibility and Alternative Possibilities: Essays on the Importance of Alternative Possibilities is useful collection of recent articles on Frankfurt cases. 13. Watson, Free Agency, reprinted in Agency and Answerability. Frankfurt responds in Identification and Wholeheartedness reprinted in The Importance of What We Care About, and in The Faintest Passion (reprinted in Necessity, Volition and Love). 347

6 the issue of what is involved in an agent identifying with her motivation or, to put it another way, which actions really do come from the agent s real self? Gary Watson develops another real self view. Watson rejects Frankfurt s picture of a hierarchy of desires, and instead suggests that it is the agent s evaluations that are crucial. Thus a distinction is often made between hierarchical real self views (Frankfurt), and non-hierarchical views (Watson). In Free Agency, Watson defended the view that one is morally responsible for actions that one has evaluatively endorsed. By using the notion of evaluation as opposed to mere desiring, Watson hoped to secure the relevant sense in which an agent identifies with her action. In later work he raised a worry for this view: there appear to be cases ( perverse cases ) of responsible action that do not reflect the agent s values sometimes we knowingly do things that we do not value. We might do something we think vulgar just because it is fun, for example. 14 Michael Bratman takes up the issue of identification, and argues that the key to understanding it is not in valuing, but in the way we think of our own desires. According to Bratman, we identify with a desire when we treat it as reason-giving. 15 To treat a desire as reason-giving is to make a decision to treat a particular desire as a justifying reason for a course of action. This view can deal with Watson s perverse cases: the agent may not value her action, but she has nonetheless treated her desire to do it as reason-giving on this occasion, and that is why she is responsible for it. A rather different real self view is developed by Derk Pereboom, who builds on Ted Honderich s view that although determinism undermines one sort of responsibility, it does not undermine another sort. 16 Pereboom calls himself a hard incompatibilist : by that he means that he believes that if agents are not the ultimate causes of their actions (as determinism implies), then they are not ultimately responsible. However Pereboom argues that his view allows for all the things that matter our life hopes (Honderich s term), our personal relationships, and our chances of finding meaning in our lives. Pereboom s argument can be classified as a real self view because his point is that what we are matters to us much more than whether we are really free. We care about and evaluate people for traits that we do not take to be voluntary beauty, for example. 17 Nomi Arpaly and Timothy Schroeder have suggested a view that is probably best categorised as a real self view. Arpaly and Schroeder focus on the behaviour of the fictional character Huckleberry Finn. Huck believes that he should return an escaped slave, Jim, to his master. However, Huck cannot bring himself to do what he believes to be the right action, and helps Jim to escape. Arpaly and Schroeder call this inverse akrasia, and they argue that Huck is praiseworthy for his action. 18 However, his praiseworthiness cannot be explained 14. Free Action and Free Will reprinted in Agency and Answerability. 15. Michael Bratman, Identification, Decision and Treating as a Reason, Philosophical Topics, 24 (1996), pp. 1 18, reprinted in Faces of Intention (Cambridge University Press, 1999). 16. Honderich, A Theory of Determinism. 17. For a recent criticism of real self views see Kasper Lippert-Rasmussen, Identification and Responsibility, Ethical Theory and Moral Practice, 6 (2003), pp Arpaly develops the argument in On Acting Rationally Against One s Best Judgement, Ethics, 110 (2000), pp and in Unprincipled Virtue (Oxford University Press 2003). For discussion of Arpaly s book see the forthcoming symposium in Philosophical Studies. 348

7 on any of the existing real self accounts, as all of those accounts require some sort of conscious endorsement of an action for it to count as a responsible action. Arpaly and Schroeder urge that desires as well as beliefs are important in assigning moral responsibility, and propose what they call a whole self view. On this view, an agent is morally responsible when the agent s desires and beliefs are well-integrated. On their analysis of the Huckleberry Finn case, Huck s act of helping Jim to escape is better integrated than his beliefs about the right thing to do. Intellectual Process Views There is another family of views that I shall call intellectual process views, because they all claim that there is some sort of intellectual process that the agent must have gone through in acting in order to count as responsible. 19 Susan Wolf argues for what she calls a reason view. Wolf complains that real self views do not capture the deep sense of responsibility. 20 Wolf proposes the view that an agent is responsible when she has the ability to act on the basis of right reason (the Good and the True, as Wolf calls it). Wolf insists that she is not using a libertarian account of ability, and is not using a metaphysically ambitious sense of the True and the Good the crux of her view is that an agent is responsible when she has the capacity to be governed by reasons. More recently, Jay Wallace, Hilary Bok and John Fischer and Mark Ravizza have all argued for intellectual process views, though beyond that structural similarity their views are very different. Wallace develops a sophisticated Strawsonian view about the meta-question, according to which our practice of engaging in reactive attitudes defines responsibility, but at the same time, our reactive attitudes have cognitive content. On Wallace s view, we appropriately hold people responsible when it is fair for us to do so. Thus Wallace s answer to the meta-question and his answer to the normative questions are bound up with each other what it is fair for us to hold people responsible for depends on what sort of excuses and exemptions we think there are. 21 At this point, the intellectual process view comes in, and Wallace provides an argument for why responsiveness to reasons is so important. Wallace argues that the excuses and exemptions we accept depend on the absence of intention and the capacity for normative control respectively. Wallace points out that on Strawson s approach, the reactive attitudes are directed towards the attitudes and feelings of other people, but this doesn t explain enough: we need an explanation of why we care about the attitudes and feelings of others. On Wallace s picture, the stance of holding people responsible depends on believing 19. These are often referred to as historicist views, because they use the history of the act to determine whether the agent is morally responsible. Fischer and Ravizza classify them as externalist views (Fischer and Ravizza, Responsibility and Control ). Elsewhere Fischer refers to these as reasons responsiveness accounts ( Recent Work, p. 127). 20. Susan Wolf, Freedom within Reason (Oxford University Press, 1990). 21. This discussion does not do justice to Wallace s complex and subtle defence of his position; see particularly Ch. 4 of Responsibility and the Moral Sentiments. 349

8 that they have violated an obligation. Obligations can only be violated intentionally so we hold people responsible only when they have acted intentionally. Thus, if on a particular occasion someone has not acted intentionally, or if they lack the capacity to act intentionally at all, we do not hold them responsible. 22 Hilary Bok s approach to the meta-question is quite different she is a Kantian compatibilist. Bok starts with a distinction between two standpoints: the practical and the theoretical. Her basic strategy is to show that moral responsibility is in the realm of the practical, whereas the truth of determinism is only relevant from the theoretical standpoint. Bok argues that in practical (as opposed to theoretical) reasoning, we must use the compatibilist conception of possibility just because of our epistemic position. In other words, we must think of the alternatives in a given situation as being those that we would perform if we chose to perform them. Bok argues that it doesn t matter that we in fact have one alternative only that is only relevant from the point of view of the theoretical standpoint, which seeks to describe the world. From the practical standpoint we are interested in determining the will, and so our alternatives from that point of view are relative to our epistemic position. Bok allows for correction of factual errors in her account of compatibilist possibility (for example, if my belief that my bicycle is still chained to the railings outside is false, cycling home is not one of my alternatives), but argues that lack of knowledge of the metaphysical unavailability of some of these alternatives does not count as a factual error. Bok argues that the knowledge of which alternative is determined is in principle unavailable from the practical point of view. 23 So on Bok s view the practical point of view inevitably and innocuously involves using the traditional compatibilist account of available alternatives I can do whatever I would do if I chose to. Bok s view can be thought of as an intellectual process view because for her what is crucial is that we choose one action rather than another we are responsible for actions that we choose. However, Bok s view is unusual in that her argument for why choice matters is essentially forward looking. Bok s explanation of why choice matters is Kantian in spirit: we must have a set of standards by which we evaluate our reasons this is a condition of engaging in practical reason. So we cannot help but have an interest in improving the effectiveness of our wills i.e. our choices. Thus we hold ourselves responsible for our choices. 24 Fischer and Ravizza s recent book is the latest formulation of a view they have been defending over a period of two decades. They defend a historicist 22. For discussions of Wallace s book see the symposium in Philosophy and Phenomenological Research, 64 (2002). 23. In the interests of space I shall leave aside the complex argument she makes for this claim. 24. For other (broadly) intellectual process views see Keith Lehrer Metamind (Oxford University Press 1990); Eleonore Stump, Sanctification and Free Will, The Journal of Philosophy, 85 (1988), pp ; Philip Pettit and Michael Smith, Freedom in Belief and Desire, Journal of Philosophy, 93 (1996), pp ; Michael Smith, A Theory of Freedom and Responsibility in Garrett Cullity and Berys Gaut (eds.), Ethics and Practical Reason (Oxford University Press 1997); For a recent criticism of intellectual process views see Manuel Vargas, The Trouble with Tracing, forthcoming in Midwest Studies in Philosophy, 29 (2005). 350

9 view (intellectual process view) with a broadly Strawsonian compatibilist background. Much of the book is devoted to solving rather technical problems, and I shall not discuss those aspects of it here. The central idea is that what is necessary for responsibility is not regulative control, which would require alternate possibilities, but guidance control. An agent has guidance control when she is reasons-responsive, and takes responsibility for her reasonsresponsiveness. In this way, she makes her actions her own. Fischer and Ravizza s accounts of exactly what reasons-responsiveness consists in, and what it is to take responsibility for the mechanism are long and complex, informed as they are by criticisms of previous formulations of the view. I shall mention only a couple of the pertinent aspects of the theory. Fischer and Ravizza make a distinction between internalist and externalist theories a distinction that broadly maps onto the distinction between real self views and intellectual process views. The point of the new terminology is to highlight what they take to be two big advantages of their theory: that the agent s history is taken into account, in so far as an agent can be held responsible for an action so long as we can trace it back to a qualifying decision made by the agent. On a view like Frankfurt s all that matters is the state the agent is in at the moment of acting. But an externalist view can take into account how the agent got into that state. The other advantage is that on the Fischer/Ravizza view, an agent is responsible only when she is responding to the world in certain ways. Reasons-responsiveness has a factive element to be reasons-responsive, one must be responsive to the reasons that there are. 25 Our Concept of Moral Responsibility Much of the literature on responsibility focuses on the question of whether the view is really compatible with determinism. Until very recently, less attention has been paid to the issue of what our concept of moral responsibility really involves. Watson s Two Faces of Responsibility In a rich and suggestive article, Gary Watson suggests that there are two faces to our concept of responsibility. 26 Watson characterises real self views (or, self disclosure views, as he calls them) as being concerned with attributing the action to the agent, rather then with holding the agent accountable, or engaging in blame behaviour towards her. According to Watson, attributability represents one aspect of our concept of responsibility, and an important one. Accountability the practice of holding people to standards and blaming them if they fail represents another face of our concept of responsibility. Watson s article arose out of his response to Susan Wolf s book, and he is at least partly concerned to defend his own version of a self-disclosure view 25. For discussions of Fischer and Ravizza s book see the symposia in Philosophy and Phenomenological Research, 61 (2000), and in The Journal of Ethics, 6 (2002). 26. Two Faces of Responsibility, in Agency and Answerability. 351

10 from her criticisms. One very interesting strand of his article is his defence of attributability against Wolf s criticism that it is a superficial notion of responsibility. Watson argues that attributing an act to an agent, as opposed to attributing an effect to a cause, does capture something deep and important. The very nature of agency is what makes attributing an act to an agent important an agent, unlike an object or an automaton expresses her moral capacities and commitments in her actions. As Watson puts it: The point of speaking of the real self is not metaphysical, to penetrate to one s ontological center; what is in question is an individual s fundamental evaluative orientation. Because aretaic appraisals implicate one s practical identity, they have ethical depth in an obvious sense. 27 Watson points out that it is the accountability aspect of moral responsibility that gives rise to problems about determinism, and he offers a diagnosis of why this is the case. Accountability, unlike attributability is wrapped up with the practice of blaming people, where blame is not just a judgement, but behaviour, and crucially, behaviour that harms people. Thus a fairness issue arises it is not fair to inflict harms on people for behaviour that they could not avoid. So the notion of avoidability is central to the accountability face of responsibility, and, of course, avoidability seems to involve genuine alternatives, which in turn require the falsity of determinism. As I have presented the argument so far it might appear that Watson is simply defending a real self view over an intellectual process view. However, Watson s point is that both aspects of responsibility are part of our concept, and grasping this point will help us to explain our ambivalence towards certain cases. In an earlier article Watson examines our attitude to cases where someone acts extremely badly, and seems to be evil through and through. 28 Watson describes the murderer Robert Harris, whose acts are undeniably horrible. However, Harris s life story is so horrible that it almost beggars belief. Watson points out that we are torn in this case on the one hand we blame Harris, but on the other we excuse him. In Two Faces Watson suggests that we are shifting between attributing a murderer s act to him (and making a blame judgement in that sense), and holding the murderer accountable for his acts (and hence judging blame behaviour to be fair). So perhaps our concept of responsibility is in need of revision. Revisionism The idea that maybe we should be revisionist about responsibility is pursued by Manuel Vargas in a series of recent papers. 29 Vargas argues that theories of responsibility should not attempt to accommodate all of our pre-theoretical intuitions about responsibility. He entertains the possibility of two different sorts of tension in our intuitions. First, we have some intuitions that support 27. Agency and Answerability, p. 271f. 28. Responsibility and the Limits of Evil, in Agency and Answerability. 29. See On the Importance of History for Responsible Agency and The Revisionist s Guide to Responsibility both forthcoming in Philosophical Studies, and Responsibility and the Aims of Theory: Strawson and Revisionism, Pacific Philosophical Quarterly, 85 (2004), pp

11 a historicist (intellectual process) view, but some that do not. Vargas argues that rather than trying to squeeze our disparate intuitions into an ill-fitting theoretical framework, we should accept that some of our intuitions might have to be abandoned. Second, Vargas points out that some of our intuitions about responsibility may be incompatibilist, in which case a compatibilist should admit this, and admit that not all of our intuitions can be accommodated by compatibilism. Vargas s point is not that there are no revisionist theories he admits that many existing compatibilist theories are more or less revisionist. His point is rather that compatibilists should be more explicit about their revisionist methodology. Convergence A different moral to the story is that real self views and intellectual process views are getting closer and closer to one another. To make a cartoon sketch: real self views seem problematic because they don t capture the deep sense of responsibility what Watson calls accountability. So a real self view can be improved by adding more about what it is for an act to spring from the agent s real self, and crucially, why it is important that an act spring from the agent s real self. 30 Watson s remarks about practical identity, and Bratman s interpretation of identification might be interpreted in this light. The more we say about why the real self is important, the closer we get to saying something about the process the agent goes through in acting. On the other hand, intellectual process views seem problematic because they don t say enough about what makes the act the agent s own so they can be improved by adding more about what makes an act an agent s own, which of course brings them closer to real self views. Fischer and Ravizza s explanation of what it is for an agent to own her action, and their talk of taking responsibility, for example, are reminiscent of Frankfurt. The literature about moral responsibility has been very preoccupied with the threat from determinism. Less work has been done on what I have called the normative question and less still on the implications of different theories of responsibility for normative ethical theories. However, there is progress in philosophy, and it might be possible at this stage to trust that compatibilism in one form or another is successful, and concentrate on the normative question. A different survey would seek to bring together the disparate work on responsibility that has taken place outside the free will debate, and see how far this project has progressed. 30. Clearly Arpaly and Schroeder are exceptions to this, and indeed are moving in the opposite direction. 353