-- The Determinism and Freedom Philosophy Website --

Here is what is sure to be a good piece, since it comes from a good and strong philosopher. Its title tells all -- or anyway a lot. But because of the distractions of your editor's  philosophical life, in particular a conference in a schloss in Bavaria after a conference in the gentler hills of Idaho, it takes its place without any words of proper introduction. Indeed without having been properly read. It takes time to read good philosophy. This will be put right. 

In a recent article Gary Watson instructively distinguishes two faces or aspects of responsibility. The first is the self-disclosing sense, which is concerned centrally with aretaic or excellence-relevant evaluations of agents. An agent is responsible for an action in this respect when it is an action that is inescapably the agent’s own, if, as a declaration of her adopted ends, it expresses what the agent is about, her identity as an agent. An action for which the agent is responsible in this sense expresses what the agent is ready to stand up for, to defend, to affirm, to answer for. (1996: 233-4) . The second face of responsibility has perhaps had a more explicit role in debates about free will — it concerns control and accountability. Watson argues that when one is skeptical about the second "accountability" face, one need not also be skeptical about responsibility as self-disclosure. I agree, and in my view, this helps us see why maintaining that determinism precludes accountability need not also commit one to the view that determinism precludes responsibility in a way that threatens meaning in life. Part of the reason for this is that when responsibility as accountability is undermined, less of what we deem valuable needs to be relinquished than often believed. But in addition, it turns out that the kind of accountability precluded by determinism is not nearly as important to what is most significant in human life as is responsibility as self-disclosure. Indeed, it may be that an unfortunate fusing of these two notions underlies the concern that if determinism imperils accountability, it also threatens what most fundamentally makes our lives meaningful. 

There is one notion of responsibility as accountability that, to my mind, is acutely threatened by general causal features of reality such as determinism, and indeed, certain varieties of indeterminism. Let’s call it strong accountability -- it is not the only notion of responsibility as accountability. For an agent to be strongly accountable for an action is for it to belong to the agent in such a way that she would deserve blame if the action were morally wrong, and she would deserve credit or perhaps praise if it were morally exemplary. The desert at issue here is basic in the sense that the agent, to be strongly accountable, would deserve the blame or credit just by virtue of having performed the action, and not, for example, by way of consequentialist considerations. This characterization leaves room for an agent's being strongly accountable for an action even if she does not deserve blame, credit, or praise for it -- if, for example, the action is morally indifferent. 

To judge a person strongly accountable does not essentially involve having a reactive attitude towards her. Rather, I think that to make a judgment of this sort is most fundamentally to make a factual claim about kind of control an agent has. To defend this position adequately would involve turning back a non-cognitivist position on judgments about moral responsibility, a task I will not undertake. But here are two considerations in favor of my view. First, judging a person strongly accountable for an action that is morally indifferent, or for an action that is not morally indifferent but generally expected, like feeding and clothing one’s children, need not be accompanied by any discernible attitude. Second, it seems possible to imagine rational but emotionless beings who yet have a deep concern for right and wrong, and who believe that agents are strongly accountable. Such beings would believe wrongdoers to be strongly accountable without having any reactive attitudes, like indignation or moral resentment, towards them. It is of course consistent with the view that judgments about moral responsibility are factual that in us such judgments are typically accompanied by reactive attitudes. 

Furthermore, I think that this notion of moral responsibility applies primarily to decisions. The view that responsibility for decisions is especially important is driven by the sense that responsibility is fundamentally a matter of a kind of control, a kind of control agents would have primarily over their decisions, in conjunction with the fact that decisions are causally prior to consequences of decisions. Intuitions about "moral luck" cases support this view. Suppose two agents, A and B, are psychologically identical and each makes the decision to shoot an innocent person, and then carries out the decision. However, A's bullet does not reach the intended victim because it hits a bird instead, whereas B's bullet kills him. A common intuition here is that A and B are equally blameworthy in some especially important respect, an intuition captured by the notion that responsibility for decisions is especially important. 

Now strong accountability differs from another kind of accountability, one which has not been most fundamentally at issue in the debate about determinism and moral responsibility. This notion is the notion of the legitimacy of demanding that agents explain how their decisions accord with the moral point of view, and that they consider what their decisions reveal about their moral character and dispositions (Bok, 1998). Incompatibilists have not maintained that determinism precludes the legitimacy of the demand to explain whether one’s actions accord with the best moral reasons, and to assess what one’s behavior reveals about their moral rationality. Making these demands of agents might be justified by its effectiveness in improving the agent morally -- we humans are manifestly susceptible to being causally influenced by admonition of this sort. Incompatibilists have not felt that this notion of accountability is threatened by determinism -- let us call it weak accountability. The capacity to be responsive to reasons is crucial here. To my mind reasons responsiveness is clearly required for weak accountability, even if it is controversial that it be the key component for securing strong accountability.

The central thesis of the position I have defended (Pereboom 1995, 2001) is that we do not have the sort of free will required for strong accountability. In this respect I am allied with Spinoza, Priestly, Holbach, and more recently, Galen Strawson and Ted Honderich. My argument for this claim has the following structure: An agent’s moral responsibility for an action depends primarily on a kind of control exercised in the action’s actual causal history, and not on the existence of alternative possibilities. Absent agent causation, indeterministic causal histories pose no less of a threat to strong accountability than do deterministic histories, and a generalization argument from manipulation cases shows that deterministic histories indeed undermine strong accountability. Agent causation is a coherent possibility, but it is not credible given our best physical theories. Consequently, no position which affirms the sort of free will required for strong accountability is left standing. 

I reject an alternative-possibilities type of incompatibilism, and accept instead a type of incompatibilism that ascribes the more significant role to an action’s causal history. My view is that an agent’s accountability for an action is explained not by the existence of alternative possibilities available to her, but rather by the action’s having a causal history of a sort that allows the agent to be the source of her action in a specific way. Following Ted Honderich (1988: 194-206) and Robert Kane (1996: 35), the crucial condition emphasizes that an agent must be the origin of her action in a particular way. According to my version of this condition, if an agent is accountable for her decision to perform an action, then the production of this decision must be something over which the agent has control, and an agent is not strongly accountable for the decision if it is ultimately produced by a source over which the she has no control. 

The grounding for this kind of incompatibilism includes the argument that certain Frankfurt-style cases rule out the notion that having alternative possibilities explains an agent’s responsibility for action (Frankfurt 1969), and the argument that a deterministic causal history would make it impossible for the agent to be the source of her action in the way required. The best strategy for establishing the latter claim involves devising manipulation cases in which the agent is covertly induced to perform an action by some external cause, and for that reason is not responsible for her action, and then generalizing to absence of accountability in more ordinary deterministic cases. I contend that no relevant and principled difference can distinguish an action that results from accountability-undermining manipulation from an action that has a more ordinary deterministic causal history (Taylor 1974: 43-4; Kane 1996: 65-71). Moreover, exclusively event-causal indeterministic histories are no less threatening to accountability than deterministic histories, and since deterministic causal histories undermine strong accountability, so do such event-causal indeterministic histories (Clarke 1997). If the crucial indeterministic events were appropriately produced by a randomizing manipulator, then one would have the intuition that the agent is not strongly accountable (van Inwagen 1983: 132-4, Mele 1999: 277). But there is no relevant and principled difference between the manipulated action and one that is indeterministic in a more ordinary way. Among available models for agency, to my mind only agent causation allows for strong accountability, but simply because it builds into the agent, as a primitive power, the capacity to be a source of action that is required for strong accountability. The agent-causation model is coherent as far as we can tell, but given evidence from our best scientific theories, it is not credible that we are in fact agent-causes. We are therefore left with the view that we do not have free will of the kind required for strong accountability.

I also believe that followers of Strawson would be mistaken to think that the priority of practice would insulate strong accountability attributions from scientific or metaphysical challenges. I contend that the best way to develop this point is by what Wallace calls a generalization strategy -- arguing from ordinarily accepted excuses or exemptions to the claim that determinism, for example, rules out moral responsibility. The excuses and exemptions that form the basis of this sort of argument would have to be widely accepted, so that they are plausibly features internal to the practice of holding people strongly accountable. The kinds of exemptions that I exploit are due to manipulation. It is also a feature of our practice of holding people strongly accountable that if no relevant moral difference is to be found between agents in two situations, then if one agent is legitimately exempted from moral responsibility, so is the other. And no relevant moral difference can be found between agents in the manipulation cases and agents in ordinary deterministic situations. So it is the practice itself, in particular central rules governing the practice, that makes it the case that "universal determinism" is relevant to responsibility after all. 

According to the view that Strawson develops, the practices that surround holding people morally responsible are insulated from general metaphysical claims or scientific discoveries. There are two ways to view this insulation. On the one hand, the reason to accept this insulation could be practical. We might need to hold that moral responsibility cannot be undermined by a general scientific discovery because our capacity to live meaningful and fulfilled lives would be severely hindered if we held otherwise and the relevant sort of scientific discovery were made. I have no quarrel with the practical legitimacy of endorsing the insulation view for this sort of reason. But what we would need to investigate is whether abandoning the view that we are responsible would indeed have such bad consequences. On the other hand, one might think that we have epistemic, and not only practical reasons for regarding our beliefs about moral responsibility as insulated from general scientific discoveries. Is this really so? I would be uncomfortable with a similar claim about religious practice. Some have argued that in the light of the importance of religious belief, we have more than just practical reason for regarding it as insulated from scientific discovery, but this view strikes few as especially attractive. 

Perhaps there is position that accommodates the notion that our investment in our self-conception as morally responsible has epistemic force without embracing full-fledged insulationism. John Fischer, in his review essay in Ethics, might at first appear to endorse insulationist perspective. He says:

I believe that we — you and I and most adult human beings — are morally responsible (at least much of the time) for our behavior. Further, I do not think that this very important and basic belief should be "held hostage" to esoteric scientific doctrines. For example, if I were to wake up tomorrow and read in the Los Angeles Times that scientists have decisively proved that causal determinism is true, I would not have any inclination to stop thinking of myself, my family and friends, and human beings in general as morally responsible. The precise form of the equations that describe the universe, and whether or not they correspond to universal generalizations, are not the sorts of thing that should be relevant to our most basic views of ourselves (as morally responsible agents and thus apt targets of reactive attitudes). (Fischer, 1999: 129) But he then develops the specifically epistemic force of these remarks.  Our reactive attitudes should not be held hostage to an esoteric scientific discovery of the kind in question. That is, the reactive attitudes, and our view of ourselves as morally responsible agents, should be resilient in a certain sense. This resiliency idea is a major motivation for my acceptance of semi-compatibilism. It is part of the background against which I evaluate the complicated debates pertaining to Frankfurt-type cases, and it makes me more inclined to conclude that such cases do indeed establish that alternative possibilities are not required for moral responsibility. It also influences my evaluation of the question whether causal determinism in itself and apart from considerations pertinent to alternative possibilities rules out moral responsibility. (Fischer, 1999: 129) What Fischer says here strongly suggests a reflective equilibrium approach, according to which a belief one has can legitimately have an effect on how one regards relevantly related arguments and evidence. Accordingly, a belief in moral responsibility can legitimately exert some force on how one evaluates various arguments that in some way bear on this belief. This general approach is attractive and plausible, with a few conditions. First, the picture should not be seen as revealing that one has epistemic warrant for a belief simply because it is a belief one has, but showing only that a belief has epistemic warrant insofar as it can be integrated with other beliefs to play a genuine explanatory role. Secondly, merely wanting a belief to be true cannot all by itself give it epistemic warrant. Not that the belief in moral responsibility runs afoul of these requirements, but they should function as reason to exercise caution in the degree of epistemic warrant one grants the belief in moral responsibility.

So how much weight should the belief in moral responsibility carry in the reflective equilibrium procedure? This is very hard to say, partly because it’s difficult to separate the desire for the belief to be true from epistemic warrant. But here’s an interesting test case. Against agent-causal libertarians I’ve contended that if were morally responsible agent causes, then it would almost have to be that microphysical events in the underlying constitution of freely deciding agents are no longer governed by the laws of quantum physics as we know them (2001, Chapter 3). In response, several people have argued (in conversation) that our belief in moral responsibility, given the reflective equilibrium procedure, would give us reason to believe that events in the brain indeed not governed by the laws of quantum mechanics. But I doubt that there are many physicists who would grant that our belief in moral responsibility could provide significant epistemic reason, if any reason at all, to doubt that the quantum mechanics. Indeed, quantum mechanics is an especially well-confirmed theory, but perhaps reflection should make us careful about how much epistemic reason-giving force we should more generally confer on the belief in moral responsibility in a reflective-equilibrium procedure.

In my view, determinism threatens only one kind of moral responsibility: strong accountability. The question to ask at this point is this: how important is this aspect of our ordinary self-conception to living meaningful and fulfilled human lives? So, first of all, rejecting strong accountability demands giving up our ordinary view of ourselves as blameworthy for immoral actions and praiseworthy for those that are morally exemplary. One might think that this would result in a significant loss in legitimate procedures to deal with wrongdoing. However, it is possible to achieve moral reform and education by methods that do not presuppose that wrongdoers are blameworthy, and in ordinary situations such practices could arguably be as successful as those that do. Instead of treating people as if they were deserving of blame, we can draw upon moral admonishment and encouragement, which presuppose only that the offender has done wrong. These methods can effectively communicate a sense of what is right and result in beneficial reform. Similarly, rather than treating oneself as blameworthy, one could admonish oneself for one's wrongdoing, and resolve to avoid similar behavior in the future. 

But what resources would we have for dealing with criminal behavior? Here rejection of strong accountability would appear to be a disadvantage. A retributivist justification for criminal punishment would clearly be ruled out, for it assumes that we deserve blame or pain or deprivation just for performing an immoral action, while hard incompatibilism denies this claim. We would therefore need to give up on retributivism -- one of the most naturally compelling ways for justifying criminal punishment. 

By contrast, the moral education theory of punishment is not imperilled by the exclusion of strong accountability specifically. Still, without significant empirical evidence that punishment of criminals would bring about moral education, it would be wrong to punish them for the sake of achieving this goal. In general, it is morally wrong to harm someone in order to realize some good if there is insufficient evidence that the harm can produce the good. Moreover, even if we knew that punishment could be effective in moral education, we should prefer non-punitive methods for producing this result -- whether or not we are strongly accountable. 

Although the two most prominent deterrence theories are not challenged by the absence of strong accountability in particular, they are questionable on other grounds. The utilitarian version is dubious for well-known reasons -- it would at times demand punishing the innocent, in some circumstances it would prescribe punishment that is unduly severe, and it would authorize using people merely as means. I contend that the type of deterrence theory that justifies punishment on the basis of the right to harm in self-defense is also objectionable (Farrell 1985: 38-60). The right to harm in self defense applies in a situation where someone poses an immediate danger, and then one can only inflict what one would reasonably believe to be the minimum harm required to prevent harm. But a threat that one could justifiably make and carry out to protect against someone who is immediately dangerous cannot legitimately be carried out against a criminal in custody, even if he would be dangerous if released. For the minimum harm required to secure protection from someone who is immediately dangerous is typically much more severe than the minimum harm required to secure protection from a criminal in custody. 

A view that would work draws an analogy between crime prevention and quarantine. Ferdinand Schoeman (1979) argues that if we have the right to quarantine people who are carriers of severe communicable diseases to protect society, then we also have the right to isolate the criminally dangerous to protect society. Schoeman’s claim is true independently of any legitimate attribution of strong accountability. If a child is infected with the Ebola virus because it has been passed on to her at birth by her parent, quarantine may nevertheless be justified. By analogy, suppose that someone poses a known danger to society by having demonstrated a sufficiently strong tendency to commit murder. Even if he is not in general a strongly accountable agent, society would nevertheless seem to have as much right to detain him as it does to quarantine a carrier of a deadly communicable disease who is not responsible for being a carrier.

One must note, however, that it would be morally wrong to treat carriers of a disease more severely than is required to defuse the threat to society. Similarly, given the quarantine model, it would be wrong to treat those with violent criminal tendencies more harshly than is needed to remove the danger to society. In addition, just as moderately dangerous diseases may only license measures less intrusive than quarantine, so tendencies to moderately serious crimes may only justify responses less intrusive than detention. Shoplifting, for example, may warrant merely some degree of monitoring. Furthermore, I suspect that a theory modelled on quarantine would never justify criminal punishment of the sort whose legitimacy is most in doubt, such as the death penalty or confinement in the worst prisons in our society. Moreover, it would require a degree of concern for the rehabilitation and well-being of the criminal that would decisively alter current policy. Just as society has a duty to try to cure the diseased it quarantines, so it would have a duty to attempt to rehabilitate the criminals it detains.. When rehabilitation is impossible, and if the protection of society were to demand indefinite confinement, there would be no justification for taking measures that aim only to make the criminal’s life miserable.

Would it be practically impossible for us to live without a conception of ourselves as praiseworthy for achieving what makes our lives fulfilled, happy, satisfactory, or worthwhile -- for realizing what Honderich has called our life-hopes? (Honderich 1988: 382) Honderich argues that there is an aspect of these life-hopes that is undermined by determinism, but that nevertheless determinism leaves them largely intact. I agree with this type of position, and develop it in the following way. It is not unreasonable to object that life-hopes involve an aspiration for praiseworthiness, which the rejection of strong accountability would undercut. For life-hopes are aspirations for achievement, and because it cannot be that one has an achievement for which one is not also praiseworthy, giving up praiseworthiness would deprive us of life-hopes altogether. However, achievement and life-hopes are not obviously connected to praiseworthiness in the way this objection supposes. If an agent hopes for a success in some endeavor, and if she accomplishes what she hoped for, intuitively this outcome can be her achievement even if she is not praiseworthy for it -- although the sense in which it is her achievement may be diminished. If an agent hopes that her efforts as a teacher will result in well-educated children, and they do, it seems clear that she achieved what she hoped for, even if because of the truth of hard incompatibilism she is not praiseworthy for her efforts. 

Furthermore, on might think that rejection of strong accountability would instill an attitude of resignation to whatever one’s behavioral dispositions together with environmental conditions hold in store. But this isn’t clearly true. Given that we lack knowledge of how our futures will turn out, we can still reasonably hope for success in achieving what we want most even if we turn out to be creatures of our environments and our dispositions. It may sometimes be crucial that we lack complete knowledge of our environments and dispositions. Suppose that there is some disposition that an agent reasonably believes might be an obstacle to realizing a life-hope. However, because he does not know whether this disposition will in fact function this way, it remains epistemically possible for him that he has a further disposition that will allow him to transcend the potential obstacle. For example, suppose that someone aspires to become a successful clinical psychologist, but is concerned that his irritability will stand in the way. He does not know whether his irritability will in fact frustrate his life-hope, since it is epistemically possible for him that he will overcome this problem, perhaps due to a to a disposition for resolute self-discipline. As a result, he might reasonably hope that he will overcome his irritability and succeed in his aspiration. If he in fact does overcome his problem and becomes a successful clinical psychologist, his achievement will not be as robust as one might naturally have believed, but it will be his achievement in a substantial sense nevertheless.

But how significant is the aspect of our life-hopes that we must forgo if we were to reject strong accountability? Saul Smilansky argues that although determinism leaves room for a limited foundation for the sense of self-worth that derives from achievement or virtue, the hard determinist’s perspective can nevertheless be "extremely damaging to our view of ourselves, to our sense of achievement, worth, and self-respect," and in response we should foster the illusion that we have free will (1997: 94, cf . Smilansky 2000). I agree with Smilansky that there is a type of self-respect that presupposes an incompatibilist foundation, and that it would be undermined if had to abandon strong accountability. The sort of moral worth that can be retained absent strong accountability differs from the ordinary conception. Without strong accountability moral accomplishments would not genuinely be an agent's own in a sense strong enough to sustain judgments of fundamentally deserved credit or praise. But the notion moral worth that accompanies responsibility as self disclosure can be retained. Agents can enjoy moral worth by virtue of being ready to affirm, to act on, and to stand up for moral values. One can therefore respect oneself as an agent that is morally worthy in this way. What will nevertheless be missing is that we can respect ourselves for being the undetermined originators of this aretaic moral worth, and that we can deserve praise simply for being such originators. I do question, however, whether Smilansky is right about how damaging it would be for us to find that we must give up this sort of self respect, and thus whether his move to fostering the illusion of free will would be justified.

One should first note that our sense of self-worth, our sense that we are valuable and that are lives are worth living, is to a non-trivial extent due to factors that are not produced by our volitions at all, let alone by free will. People place great value, both in others and in themselves, on beauty, intelligence, and native athletic ability, none of which are produced voluntarily. However, we also value voluntary efforts in the service moral ends, especially when they express fundamental commitments on the part of an agent. But how much does it matter to us that the voluntary efforts are also freely willed? In my view, Smilansky overestimates how much we care. 

Consider the formation of moral character. It is not implausible that good moral character is to a large extent the function of upbringing, and furthermore, the belief that this is so is common in our society. Parents typically regard themselves to have failed if their children turn out to be immoral, and many take great care to raise their children to prevent this result. Accordingly, people often come to believe that they have a good moral character largely because they were brought up with parental love and skill. But I suspect that hardly anyone who comes to this realization experiences dismay because of it. We tend not at all to be dispirited upon coming to understand that our moral character is not our own doing, and that we deserve at best diminished respect for having this character. Rather, we feel fortunate and thankful for the upbringing we have enjoyed, and not that something significant has been lost. 

Moreover, people typically do not become dispirited when they come to believe that success in a career depends very much on one’s upbringing, opportunities in one’s society, the assistance of colleagues, and good fortune. Realizations of this sort frequently give rise to a sense of thankfulness, and almost never, if at all, to dismay. Why then should we suppose that for this type of reason we would generally become dispirited were we to relinquish strong accountability? We would then give up the view that character and accomplishments are due to originating free will and that we therefore deserve respect, but given our response to the more commonplace beliefs in external determination, we have little reason to think that we would be overcome with dismay. But suppose that there are people who would become disheartened even upon coming to believe that moral character is largely due to upbringing. Then would it be justified or even desirable for them to sustain the illusion that they nevertheless deserve respect for producing their moral character? Most people are capable of facing the truth without incurring much loss, and those for whom it would be painful will typically have the psychological resources to cope with the new understanding. 

Does rejecting strong accountability threaten interpersonal relationships? P. F. Strawson (1962) contends that the justification for claims of blameworthiness and praiseworthiness terminates in the system of human reactive attitudes, and because moral responsibility has this kind of foundation, the truth or falsity of universal determinism is irrelevant to whether we are justified in regarding agents morally responsible. These reactive attitudes, such as indignation, gratitude, forgiveness, and love, are required for the kinds of interpersonal relationships that make our lives meaningful, and so even if we could give up the attitudes -- and Strawson believes that this is impossible -- we would never have practical reason to do so. Accordingly, we would never have practical reason to give up on moral responsibility. On the other hand, if universal determinism did threaten the reactive attitudes, we would face the prospect of the "objective attitude," a cold and calculating stance towards others that would undermine the possibility of meaningful personal relationships. 

Strawson is clearly right to believe that an objective attitude would destroy relationships, but I deny that we would adopt this stance or that it would be appropriate if we came to believe universal determinism and it did pose a threat to the reactive attitudes. In my conception, relinquishing strong accountability, in particular, would indeed undermine some of the reactive attitudes. For some of these attitudes, such as indignation, for example, presuppose that the person who is the object of the attitude is strongly accountable. I claim, however, that the reactive attitudes that we would want to retain either would not be threatened or else have analogues or aspects that would not have false presuppositions. The complex of attitudes that would survive by no means amount to Strawson's objective attitude, and they would be sufficient to sustain good relationships.

Arguably no attitude is more important for good personal relationships than love, but for Strawson love is one of the attitudes that would be threatened. Consider first whether loving someone requires that she have free will in the sense required for strong accountability. Parents love their children rarely, if ever, because they possess this sort of free will, or because they choose to do what is right by free will, or because they deserve to be loved because of their freely willed choices. Moreover, when adults love each other, it is also seldom, if at all, for these kinds of reasons. Undoubtedly the kinds of reasons we have for loving someone are complex. Considerations such as intelligence, appearance, style, and resemblance to others in one’s personal history all might have a part. But let us suppose that moral character and action are especially important in occasioning, enriching, and sustaining love. Here it is important to see that denying strong attributability does not imperil self-disclosing responsibility for moral action. One’s actions can yet reveal that morality is what one most fundamentally stands for. So even if there is a significant feature of love that is a deserved response to moral character and action, it is unlikely that love would be undermined if one came to believe that these moral qualities did not come about through freely willed decision. For responsibility for one’s moral action in the self-disclosing sense is loveable whether or not one is in addition deserving of praise for them, and I suspect that in loving others for their moral goodness we care much more about self-disclosing responsibility than about strong accountability. 

One might argue, however, that we nevertheless desire to be loved by others as a result of their free will. Against this, it is clear that parents' love for their children -- a paradigmatic sort of love -- is often produced independently of the parents' will. Kane endorses this last claim, and a similar view about romantic love, but he nevertheless argues that a certain type of love we want would be endangered if we knew that there were factors beyond the lover's control that determined it. He says:

There is a kind of love we desire from others -- parents, children (when they are old enough), spouses, lovers and friends -- whose significance is diminished... by the thought that they are determined to love us entirely by instinct or circumstances beyond their control or not entirely up to them... To be loved by others in this desired sense requires that the ultimate source of others' love lies in their own wills. (Kane, 1996, p. 88; cf. Anglin, 1991). The plausibility of Kane’s view might perhaps be enhanced by reflecting on how you would react were you to discover that someone you love was causally determined by a benevolent manipulator to have the love she has for you.

Leaving aside free will for a moment, in which sorts of cases does the will intuitively play a role in generating love for another at all? When the intensity of an intimate relationship is waning, people sometimes make a decision to try to make it succeed, and to attempt to regain the type of relationship they once had. Or when one is housed in a dormitory or barracks with someone one didn’t select, one might choose to make the relationship work. Or when one's marriage is arranged by parents, one may decide to do whatever one can to love one's spouse. 

But first, in such situations we might desire that another person make a decision to love, but it is not clear that we have reason to want the decision to be freely willed in the sense required for strong accountability. A decision to love on the part of another might greatly enhance one's personal life, but it is not at all obvious what value the decision's being free and thus praiseworthy would add. Secondly, while in circumstances of these kinds we might desire that someone else make a decision to love, we would typically prefer the situation in which the love was not mediated by a decision. This is true not only for romantic attachments, but also for friendships and for relationships between parents and children. 

Perhaps the will plays a significant role in maintaining love over an extended period. Kierkegaard suggests that a marital relationship ideally involves a commitment that is continuously renewed (Kierkegaard, 1971). Such a commitment involves a decision to devote oneself to another, and thus, in his view, a marital relationship ideally involves a continuously repeated decision. Indeed, many of us might very much desire a relationship with this sort of voluntary aspect. But again, it is difficult to see what is to be added by these continuously repeated decisions being freely willed in the sense required for strong accountability, as opposed to, say, expressing what the agent really stands for. It might well be desirable for each participant that the other make these decisions. But that the participants should in addition be praiseworthy for these choices seems hardly relevant.

Finally, suppose Kane's view could be defended, and we do have a desire for love that is freely willed, or free in the sense required for moral responsibility. If we indeed desire freely willed love, then we desire a kind of love whose possibility hard incompatibilism denies. Still, the possibilities for love that remain are surely sufficient for good interpersonal relationships. If we can aspire to the sort of love parents typically have towards children, or the kind romantic lovers ideally have towards one another, or the type shared by friends who are immediately attracted to one another, and whose relationship is deepened by their interactions, and each of whom loves the other for their self-disclosing responsibility for moral action, that then the possibility of fulfillment in interpersonal relationships is far from undermined. Finally, of all the attitudes that Strawson thinks might be imperilled by a belief in universal determinism, love is surely the most crucial for our relationships. If the types of love important for mature human relationships can survive, as I have argued, then the threat that the rejection of strong accountability poses to such relationships has been largely defused.

One might contend that abandoning strong accountability threatens the self-directed attitudes of guilt and repentance, which are arguably also essential to good interpersonal relationships. There is much at stake here, one could argue, since these attitudes are not only necessary for maintaining good relationships for agents prone to wrongdoing, but are also required for sustaining their moral integrity. Without guilt and repentance, such an agent would not only be incapable of restoring relationships damaged because he has done wrong, but he would also be kept from restoring his moral integrity. For other than the attitudes of guilt and repentance we have no psychological mechanisms that can play these roles. But abandoning strong accountability would seem to jeopardize guilt because it essentially involves a belief that one is blameworthy for something one has done. And if guilt is undermined, the attitude of repentance might also be threatened, for it could well be that feeling guilty is required for motivating repentance. However, suppose that you perpetrate some wrongdoing, but because you have rejected strong accountability, you deny that you are blameworthy. Instead, you agree that you have done wrong, you feel sad that you were the agent of wrongdoing, you deeply regret what you have done (Waller 1990). Also, because you are committed to doing what is right and to moral advancement, you resolve to forbear from wrongdoing of this kind in the future, and you seek the help of others in sustaining your resolve. It would appear that only weak accountability is required here — strong accountability need not enter in.

Gratitude is also crucial to good personal relationships, but this attitude might well presuppose that the person to whom one is grateful is strongly accountable for a beneficial act, and for this reason gratitude would be threatened. Still, certain aspects of this attitude would be unaffected, and these aspects can play the role gratitude as a whole has in good relationships. Gratitude involves, first of all, thankfulness towards someone who has acted beneficially. True, being thankful toward someone often involves the belief that she is praiseworthy for an action. But at the same time one can also be thankful to a pet or a small child for some kindness, even though in these cases one does not believe that the agent is strongly accountable. Even more, one can be thankful to a friend who beneficent actions proceed from deeply held commitments. Given hard incompatibilism, the aspect of thankfulness could be retained even if the presupposition of praiseworthiness is rejected. Gratitude also typically involves joy occasioned by the beneficent act of another. But hard incompatibilism fully harmonizes with being joyful and expressing joy when others are considerate or generous in one's behalf. Such expression of joy can bring about the sense of harmony and goodwill often brought about by gratitude, and so in this respect, hard incompatibilism is not at a disadvantage.

Relinquishing strong accountability, therefore, does not seem to endanger either interpersonal relationships after all. It might well jeopardize some attitudes that typically have a role in these domains, but there will typically be enough left over to provide what is needed. And love -- the attitude most essential to good personal relationships -- is not clearly threatened at all.

More generally, it appears that living genuinely meaningful lives is compatible with relinquishing strong accountability -- what incompatibilists have typically thought to be threatened in a deterministic universe. For much of what we care most about in life, if it is dependent on a notion of responsibility, can be secured by responsibility as self-disclosure or by weak accountability. By keeping in mind the distinctions among the aspects of responsibility, we can live in accord with a consistent conception of ourselves as agents whose actions are ultimately produced by factors beyond our control, and therefore are not their originators, but who can be deeply committed to moral values and perform actions that express this commitment, and who can be responsive to self-examination in accord with a confrontation with reasons for moral action. With this sort of life secure, I think we are capable of absorbing with equanimity the losses incurred by abandoning our conception of ourselves as strongly accountable. 


Anglin, W. S. (1991). Free Will and the Christian Faith (Oxford: Oxford University Press). 

Clarke, Randolph (1997). "On the Possibility of Rational Free Action," Philosophical Studies 88, pp. 37-57.

Farrell, Daniel M (1985). "The Justification of General Deterrence," The Philosophical Review 104, pp. 367-94.

Fischer, John Martin (1999). "Recent Work on Moral Responsibility," Ethics 110, pp. 93-139.

Frankfurt, Harry G (1969). "Alternate Possibilities and Moral Responsibility", Journal of Philosophy 66, pp. 829-39.

Honderich, Ted (1988). A Theory of Determinism (Oxford: Oxford University Press).

Kane, Robert (1996). The Significance of Free Will (New York: Oxford University Press).

Kierkegaard, Sören (1971). Either/Or, v. 2, tr. Walter Lowrie, (Princeton: Princeton University Press).

Mele, Alfred (1999). "Ultimate Responsibility and Dumb Luck," Social Philosophy and Policy 16, pp. 274-93.

Pereboom, Derk (2001). Living Without Free Will (Cambridge: Cambridge University Press).

Schoeman, Ferdinand D. (1979). "On Incapacitating the Dangerous," American Philosophical Quarterly 16.

Smilansky, Saul (1997). "Can a Determinist Help Herself?," C.H. Manekin and M. Kellner, eds. Freedom and Moral Responsibility: General and Jewish Perspectives (College Park: University of Maryland Press), pp. 85-98.

Smilansky, Saul (2000). Free Will and Illusion (Oxford: Oxford University Press).

Strawson, Galen (1986). Freedom and Belief (Oxford: Oxford University Press).

Strawson, Peter F. (1962). "Freedom and Resentment," Proceedings of the British Academy 48, pp. 1-25.

Taylor, Richard (1974). Metaphysics, 4th ed. (Englewood Cliffs: Prentice-Hall). 

van Inwagen, Peter (1983). An Essay On Free Will, (Oxford: Oxford University Press).

Waller, Bruce (1990). Freedom Without Responsibility, (Philadelphia: Temple University Press).

Watson, Gary (1996). "The Two Faces of Responsibility," Philosophical Topics 24, pp. 227-48.


BACK  Det & Free front page
BACK  T.H.'s front page