S. Wolf, Freedom Within Reason. (Oxford: Oxford University Press, 1990) pp. 3-22.
The Dilemma of Autonomy (In Which the Problems of Responsibility and Free Will Are Presented)
To be accorded the status of a responsible being is to be regarded as an appropriate object of a certain range of attitudes and judgments and as a legitimate participant in a certain range of practices. The range of attitudes I have in mind includes pride and shame, gratitude and resentment, respect and contempt. The range of judgments includes the judgment that one is worthy of respect or contempt, that one ought to be proud or ashamed, and so on. And the range of practices includes praising and blaming, forgiving, excusing, rewarding, and punishing according to rules designed to make these practices expressions of the above sorts of attitudes and judgments.
It is a deep and essential feature of life in modern Western society that normal human beings who have reached some level of maturity regard themselves and one another as responsible beings. That is, if people did not regard themselves and one another as responsible beings, life would be unrecognizably different from what it actually is. But the concept of responsibility is a mysterious one which tends, on examination, to become increasingly opaque and to threaten variously to be incoherent or impossible or universally inapplicable. Thus there is a philosophical problem of responsibility and, connected to it, a philosophical problem of free will, understanding free will to be that relation to one's will which is necessary in order for one's actions (as well as one's character and life insofar as they are governable by one's will) to be "up to oneself" in the way that is necessary for responsibility. We can express the problem of responsibility in the form of the question "How, if at all, is responsibility possible?" And we can express the problem of free will in the form of the question "What must our relation to our wills be," or better, perhaps, "What kind of beings must we be if we are ever to be responsible for the results of our wills?"
This book is unabashedly devoted to solving these problems, though to put it that way suggests an incredible hubris on the part of the author, and might also mislead the reader into thinking that the book is intended to put these problems, once and for all, to rest. It would be just as accurate to describe the book's aim as to provide a way of understanding--or, if you like, interpreting--these problems. In other words, this book is an attempt to put these problems in a new light, a light that shows their connections to other problems and that directs our way of thinking about them along paths that have long been unexplored.
But the very existence of problems that can be expressed in the form of the questions above is not self-evident. These questions may appear to be alternatively trivial or unintelligible. For the troublesome sense of mystery or opacity that attaches to the concept of responsibility is not a universal or a natural affliction. In the absence of intrusive philosophical thought, we are comfortable with the attitudes and practices associated with responsibility and with the patterns of thought by which these attitudes and practices are formed, defended, and withdrawn. To the question "How, if at all, is responsibility possible?" one might be inclined to reply "Why shouldn't it be possible?'And to the question "What kinds of beings must we be if we are ever to be responsible for the results of our wills?" one might be inclined to answer "Why, we must be precisely the kinds of beings that we are!"
Setting Up the Problem(s): The Dilemma of Autonomy
The acquisition of tendencies to take the attitudes, form the judgments, engage in the practices that are associated with responsibility, and to do so in accordance with the relevant patterns of thought, occurs, for members of contemporary Western culture, as a part of our natural development. These tendencies arise as part and parcel of the enlargement and enrichment of our consciousness of the world and of our repertoire of responses to it that constitute our growth from psychological infancy to psychological maturity. As the sense of well-being and its opposite and the feelings of pleasure and displeasure associated with distinct individuals and objects that we experience as infants are gradually differentiated into kinds, we begin to have identifiable feelings of pride, shame, and guilt, to feel admiration for (some) others, to be vulnerable to having our feelings hurt, to feel gratitude, resentment, and eventually respect, contempt, and indignation. As we develop these feelings and attitudes, we learn to discriminate between occasions in which these feelings and attitudes are and are not appropriate. We learn to feel pride and shame about some aspects of ourselves but not others, to feel gratitude and resentment toward some individuals but not others, and to have these feelings on the basis of some but not other features and actions of the relevant individuals. In conjunction with the expansion of our range of attitudes and feelings and the increase in our ability to direct them, there is an increasing tendency on the part of others to take the attitudes associated with responsibility toward us, and to regard us as appropriate objects of these attitudes. Thus, the development of our tendencies to form such attitudes toward ourselves and others is at one with the development of distinctive kinds of relationships and forms of interaction with the individuals toward whom we take these attitudes and who take these attitudes toward us.
In this way, we acquire senses of ourselves and (some) others as responsible beings, as beings with a distinctive kind of status in the world, whose relations to each other and to the world at large are different in kind from the relations that non-responsible beings are capable of forming and having. Although it is notoriously difficult to say in what, precisely, this distinctiveness consists, certain kinds of imagery and modes of expression seem peculiarly apt in our attempts to elaborate it-the imagery of depth and power, and the tendency to put the difference between responsible and non-responsible beings in terms that suggest that the former have a special kind of significance. We take responsible beings more seriously than we take non-responsible ones-we treat them as persons, and not as objects. We credit them and hold them to blame for things in a way that suggests that we attribute events and qualities to them more deeply than we do to others. These beings seem to have a special kind of power or control over what they do and who they are, in virtue of which they are owed a special attitude of respect and are endowed with a special kind of value.
It seems natural and reasonable for us generally to grant ourselves and creatures relevantly like us the status of responsibility and not to grant this status to beings who differ from us in salient ways. Moreover, it seems natural and reasonable for us to distinguish aspects of our behavior and personality for which we apparently are responsible from those for which we apparently are not. By examining the patterns of our attributions of responsibility, we are able to formulate conditions of responsibility, features of an agent and of her relation to some event, object, or property that are required if the attitudes and judgments connected to responsibility are to be deemed appropriate and justifiable. For example, it seems necessary that an agent be capable of having thoughts of a certain kind and that she be able to govern her behavior in light of these thoughts, and it seems necessary that the agent's behavior be free of certain constraints or interferences that may be imposed by drugs or hypnosis or deception. But when our examination goes beyond the mere cataloging of the diverse and relatively specific conditions of responsibility that we are apt to invoke to explain our responses to agents and events in concrete situations, when we try to find more general features in our patterns of attributing responsibility that will explain and organize these specific conditions in a systematic or enlightening way, the concept of responsibility begins to seem problematic. For the attempt to understand why the specific and diverse conditions of responsibility are conditions of responsibility can lead us to formulate a more general condition, to which all these more specific ones may be understood to point, which casts doubt upon the appropriateness of our attributions of responsibility even in what formerly seemed to be our most central cases.' In other words, we can be led to believe that there is a condition of responsibility that it is doubtful any of us ever satisfy, a condition that may seem impossible to satisfy and that may imply that the concept of responsibility is incoherent.
This is one route that may lead us to ask the question "How, if at all, is responsibility possible?" A positive answer to this question would then seem to have to take one of two possible forms: Either it would show us that the apparently impossible condition of responsibility is not a condition of responsibility after all, or it would show us that this apparently impossible condition is one that it is possible, even plausible, to think we regularly satisfy.
Let us see what this apparently impossible condition is, and what considerations might lead us to accept it. It is obvious enough that one condition of responsibility involves the possession of a will. In other words, only an agent who has a will-that is, who has desires, goals, or purposes and the ability to control her behavior in accordance with them-can be responsible for anything at all. A volcanic mountain is not responsible for erupting, a defective tire is not responsible for skidding, because these objects, and inanimate objects generally, cannot be said to have control over their characteristics or behavior. Unlike agents that have wills, these individuals, if they can be said to be agents at all, cannot direct their own actions or choose what effects they will produce. Moreover, an agent that has a will can be responsible only for things that are related to her in such a way that they fall, so to speak, within the sphere of influence of her will. Thus an agent is not responsible for the color of her eyes or her susceptibility to sunburn. She is not responsible for her silence if she has been gagged or for her hand-prints on a wall into which she has been pushed. More generally, an agent can be responsible only for events and properties that stand in a relation to her such that her will is or could have been effective in determining the existence of these events or properties.
But clearly the possession of a potentially effective will is not enough to warrant attributions of responsibility. Dogs and cats, young children, the insane, and severely mentally retarded adults have potentially effective wills, and yet we do not regard them as responsible beings. Though these individuals have a kind of control of their behavior, they cannot control their behavior along the right lines. That is, they cannot control their behavior with respect to the features that would be relevant to judgments of responsibility. One would not, for example, blame the neighbor's dog for eating the brownies one baked for a picnic in the way in which one might blame a house-mate for doing the same thing. One would not credit a child who, in playing with finger-paints, produced a beautiful picture in the same way that one would credit a more mature artist who produced an equally beautiful painting. When we judge an agent responsible for some event (or object or character trait), our judgment is made against a background of imagined alternatives that could have but did not occur and against a background of reasons that did or could have influenced the agent in question to do one thing rather than another. When we judge an agent responsible for some event (or object or character trait), we might say, we judge her responsible for that event under a certain description, with some feature or features more or less distinctly in mind that give that event its positive, neutral, or negative value. In order for an agent to be responsible for an event or object or character trait, the agent's will must be not only potentially effective in determining the event's occurrence or the object or trait's existence but also, as it were, potentially intelligent: That is, it must be possible for the features that might weigh in favor of or against that event in relevant ways to enter into the determination or the content of the agent's will. We do not regard the dog or the child in the examples above as responsible for their actions because the dog and the child could not possibly be influenced by the relevant moral or aesthetic considerations. Whole classes of beings may be generically excluded from the set of individuals who might deserve the status of responsibility if they lack basic intellectual capacities that are necessary for the possession of a suitably intelligent will. In addition, individuals who are generally capable of comprehending the relevant sorts of features may be exempt from responsibility for particular actions if, owing to unavoidable ignorance or inexperience, for example, they are incapable of recognizing features that are relevant to a particular situation on a particular occasion.
It is a condition of responsibility, then, that an agent possess a potentially effective and relevantly intelligent will (that is, a will whose content can be informed and governed by the relevant considerations). This condition accounts for many of our intuitions concerning the appropriate attributions of responsibility in a way that coheres with the meaning responsibility and the attitudes connected to it have for us. It makes sense that we should take agents who can control their behavior in intelligent ways more seriously than other agents; it makes sense that we should regard the features of the world that fall within the sphere of their intelligent control as being more deeply attributable to them; and it makes sense that we should regard these features as saying more about them, as being deeper or more significant indications of these agent's selves.
A further survey of our patterns of attribution of responsibility, however, suggests that the condition formulated above is still not a sufficient condition of responsibility. We can easily imagine cases of individuals who have potentially effective and relevantly intelligent wills but who nonetheless do not seem responsible for the actions that are a result of their wills. A person who, under hypnosis, has been instructed to obey the commands of her hypnotizer without question, may be subsequently told by her hypnotizer to treat a third person with the utmost kindness and consideration. When the hypnotized agent obeys, she must utilize her ability to recognize kindness and consideration in shaping her behavior. But though she acts kindly and considerately on purpose, she does not deserve the gratitude of the beneficiary of her acts. A person, threatened with violence to himself or others who are dear to him, may betray a confidence fully aware of the considerations that weigh against his act. He acts intentionally and knows what he is doing, but, if the threat is serious enough, he is not blameworthy for his action. Finally, a kleptomaniac who steals a piece of jewelry from a store intentionally performs an action she knows to be criminal and wrong. Indeed, the fact that her action has these properties may be essential to her motivation to perform it. Yet she is not responsible for stealing-the action is not deeply attributable to her.
At one level, the explanations for why the agents in these examples are not responsible for their actions appear quite diverse and unrelated. In the hypnosis example, we might focus on the fact that although the agent acts on a desire in her-a desire to act kindly and considerately-there is a sense in which the desire is not "really" hers; it has been implanted in her by the hypnotist. The motivating desire of the man in the coercion example, however--the desire to protect himself or a loved one from harm--is not a desire that in ordinary circumstances would cast doubt on the man's responsibility. Here the problem lies in the imposition of a situation that manipulates the agent by means of his desires. The situation is such that given the agent's desires, which may be unobjectionable in themselves, the agent cannot help acting on these desires in a particular way. The case of the kleptomaniac differs from the others in not involving external manipulation at all. Yet like the hypnotized woman, the kleptomaniac seems to have a desire she cannot help having, and like the coerced man, the kleptomaniac seems compelled to act on her desire in a particular way.
The agents in these cases are not responsible for their actions despite the fact that their wills are effective and relevantly informed. The agents act intentionally and in light of the features of their actions that make those actions alternatively good or bad. Although at one level the explanations for why these various agents are nonetheless exempt from responsibility may appear diverse, at another level they may seem fundamentally the same. For in each case the problem is neither with the effectiveness nor with the content of the agent's will. It is rather with the source of the agent's will-with the fact that the agent is not in control of what the content of his or her will will be. The agents in these cases seem to be mere vehicles of change in the world rather than initiators of it. The control they have of their behavior seems to be only control of an intermediate kind. Though their behavior results from the content of their wills, the content of their wills results from something else, and this something else is such that the agents are powerless to choose or resist it.
It seems, then, that in addition to the requirement that the agent have control over her behavior (that she have a potentially effective will) and the requirement that she have control along the right lines (a relevantly intelligent will), there is a requirement that the agent's control be ultimate-her will must be determined by her self, and her self must not, in turn, be determined by anything external to itself. This last condition I shall call, after Kant, the requirement of autonomy.
This condition, like the others, seems to cohere with the meaning responsibility has for us. It makes sense that beings who can purposefully initiate change should have a different status in the world from that of those who merely execute it. It makes sense that such beings should have a special significance, for they are sources of value (and dis-value) rather than mere carriers of it. And the imagery of depth that naturally colors our attempts to describe the difference between responsible and non-responsible beings seems readily accounted for as an expression of the perception, however dim, that the control responsible beings exhibit is ultimate control. By contrast with the ultimate control of autonomous beings, the control exerted by beings whose behavior is controlled by their wills, but whose wills are controlled by something else, seems merely superficial.
At first glance, the condition of autonomy may seem no more problematic than the other conditions of responsibility. That is, it may seem to be a condition that, like the others, we satisfy most of the time. If we speak occasionally of finding ourselves with desires that are not our own, desires that move us but with which we do not or cannot identify, we do so, presumably, by contrast to a more normal state of affairs in which the desires that provide the basis for our actions are wholly and comfortably our own. And if we sometimes describe situations in which, although we act intentionally, we have no choice but to perform the actions we do, we contrast this with more typical situations in which, it seems, we do have a choice. That is, most of the time, we seem free to act in whatever way we please. We choose to do some things rather than others, and nothing makes us choose.
A closer look at ourselves and our actions, however, may make us doubt our own apparent autonomy and may suggest that the expressions that look like claims of autonomous agency are really just figures of speech. For although few of our desires are implanted in us in as obvious and unnatural a way as that in which a hypnotist implants a desire in her subject, neither do they arise out of nothing external to ourselves. My desire for a pastry is clearly a result of the smells wafting from the bakery as I walk past; my desire for a new sweater can be traced to a magazine advertisement that caught my eye. A passionate speech makes me want to write my congressman; a letter from a friend makes me want to give her a call. It seems natural that I should have these desires in these situations-they cohere with my other desires and with my general character in ways in which desires that are implanted by hypnotists may not. But the source of these desires is no less external than the source of the desires I might be hypnotized to have, and if I identify with the former but not with the latter desires, it is not because the former desires are up to me.
Similarly, it seems that, although situations are rarely imposed on us in so artificial and manipulative a way as that in which an armed criminal may coerce his victim, situations that arise in less objectionable ways push us no less firmly to act one way rather than another. Zero-degree weather makes me turn up the heat; an empty refrigerator makes me go to the store. An upcoming tenure decision makes an assistant professor write articles for publication; a child's illness makes a father leave work early to take his daughter to the doctor.
These observations suggest a picture of ourselves as creatures whose desires are a result of some combination of our heredity and environment, who try to satisfy our desires as well as we can by acting as our situations demand. But if our desires are a result of heredity and environment, they come from something external to ourselves. And if, in conjunction with our desires, the situations in which we find ourselves dictate which actions we will ultimately decide to perform, then our behavior is completely explained by forces that originate outside of ourselves. This picture seems incompatible with the satisfaction of the condition of autonomy. This condition, which we were led to accept in our effort to explain why the agents in a few exceptional cases were not responsible for their actions, now threatens to exclude all human agents in all situations from responsibility. The cases of hypnosis and coercion now seem exceptional only in being cases in which the agents' lack of autonomy is dramatically evident. But if a lack of autonomy that is dramatically evident excludes agents from responsibility, a lack of autonomy that is less easily perceived will exclude agents as well.
In light of the apparently drastic implications the conclusion that we are not autonomous beings would have, it would be foolish to accept this conclusion too easily. The remarks made above suggest that the claim that we are autonomous beings is not so obviously or so generally true as it might formerly have seemed. But they do not imply that autonomy is impossible. Perhaps the examples of "ordinary desires" and "ordinary situations" were taken from an insufficiently wide range of experiences. If we are not responsible for as large a portion of our behavior and personality as we ordinarily think, this does not imply that we are responsible for none of our actions or character at all.3 Or, perhaps, the description of ordinary human action above is simplistic in insidious ways, omitting some elements that are crucial to a realization of autonomy.
Unfortunately, attempts to construct an alternative picture of human action, or to revise the picture presented above in a way that will justify our apparent assumption that we typically act as autonomous agents, are likely to make the condition of autonomy seem even more puzzling than before. For if the agent's control of her actions seems superficial when we add, in accordance with the picture above, that what control she exhibits is itself controlled by forces external to the agent herself, her control seems no less superficial if we remove these external forces and imagine that the agent's control is controlled by nothing at all. According to the problematic picture above, the agent's will is not wholly or deeply her own because the content of her will is completely determined by forces, people, and events external to herself. But if the content of the agent's will is not so determined-if her having the will she does is instead, in part, a result of random events, or if it is a matter of brute, inexplicable fact-this hardly seems to make her will more wholly or deeply her own. Indeed, recalling the case of the kleptomaniac, it may seem irrelevant whether the agent's will is controlled by something else or by nothing at all.
Autonomy, then, requires that the content of an agent's will (which, we may assume, determines the agent's behavior) be up to the agent herself, and this is opposed not only to its being up to anything else, but also to its not being up to anything at all. But now the concept of an autonomous agent may seem to be an impossible one. For it seems that about any agent and any act whatsoever we can ask for an explanation of why that agent performs that act. And though we may begin to answer this question in terms of features internal to the agent, we can always press beyond these beginnings and ask why the agent possesses these features. If this question in turn is answered by reference to still other features internal to the agent, we can press further and ask why the agent possesses these additional features. Eventually, we will reach a set of features that must be explained by facts external to the agent, or our explanation will simply come to an end, with the understanding that the agent's possessing these features is either a random occurrence or a brute, inexplicable fact. An agent, for example, performs some action because she wants to perform it, and she wants to perform it because she wants something else to which the action in question is perceived as a means. But why does she want that something else? Perhaps because the pursuit of that goal offers the best chance of satisfactorily realizing her complex system of values. But then we may ask why she has that particular system of values If, on the one hand, we can answer this question by describing the agent's heredity, her upbringing, her most significant recent experiences, and so on, then the agent seems to fit the problematic model of the non-autonomous agent. The agent acts in accordance with her values, but her values are a result of forces external to herself. If, on the other hand, we have no answer of this sort, it seems that the response to the question "Why does the agent have this system of values?" is simply "She just does have it" or "This is just what she is like." But the agent seems equally "stuck with" her set of values on this picture as on the former one; she is thrown into the world, as it were, complete with an identity she did not choose.
In order for an agent to be autonomous, it seems, not only must the agent's behavior be governable by her self, her self must in turn be governable by her self-her deeper self, if you like-and this must in turn be governable by her (still deeper?) self, ad infinitum. If there are forces behind the agent, so to speak, making the agent what she is, then her control of her behavior is only intermediate, and therefore superficial. But if there are no forces behind the agent making the agent what she is, then her identity seems to be arbitrary. The ability to act in accordance with a nature or a character-or, for that matter, with an uncharacteristic motive-that one simply finds in oneself as an arbitrary given, seems equally to constitute a merely superficial species of control. But this would seem to exhaust not just the empirical but the logically possible alternatives: Either something is behind the agent, making the agent what she is, or nothing is. The idea of an autonomous agent appears to be the idea of a prime mover unmoved whose self can endlessly account for itself and for the behavior that it intentionally exhibits or allows. But this idea seems incoherent or, at any rate, logically impossible.
The dilemma that was earlier sketched only schematically can now be given a more concrete form: The condition of autonomy seems at once impossible and necessary for responsibility. If there is to be any hope for a positive solution to the problem of responsibility, we must find a way to resist the conclusion that the condition is as it seems. Either this apparently necessary condition must be shown not to be a necessary condition of responsibility after all, or this apparently impossible condition must be shown to be one that, despite appearances, it is possible, and even plausible, to think we regularly satisfy.
Defending the Problem as a Problem: The Metaphysical Stance
I hope that, in light of the dilemma of autonomy, the problems of responsibility and free will, as I earlier characterized them, will seem meaningful and compelling. As I have already declared, my central aims in this book are to provide insight and some degree of relief from these problems. The attempt to achieve these aims will proceed through an examination of the dilemma of autonomy and alternative responses to it. Before I embark on this project, however, there are some general doubts that need to be addressed. For in developing a context that gives life to the problems of responsibility and free will I reveal an assumption that is indicative of a controversial philosophical stance. The assumption is that there is a fact of the matter as to whether we are free and responsible beings, a fact we might discover and about which we might be wrong. From a certain perspective this assumption seems mistaken, and philosophical inquiry into the concepts of responsibility and free will, insofar as it is infected by this assumption, may seem misguided from the start. Those who make this assumption may be said to take a metaphysical stance toward free will and responsibility. Those who oppose it may be said alternatively to take a pragmatic stance toward these concepts.
We may characterize the difference between these two stances by reference to two different kinds of responses we might have when, in the course of our experience, we find ourselves wondering whether we should hold a particular person responsible for a particular act. Sometimes the question of whether we should hold a person responsible for an act seems to demand an answer to a prior question of whether she is, in fact, responsible for it. But, in other cases, there seems to be no prior question to be asked-all the facts are in, so to speak, concerning what the person did and what circumstances and psychological events led to her doing it-and allthat is left for us, apparently, is the choice of whether, in light of these facts, we want to hold her accountable. The metaphysical stance relies on our disposition to respond to cases in the first way, to think that the answer to whether an individual is an appropriate object of certain attitudes and judgments, or a legitimate participant in certain practices, is somehow given to us, even if obscurely, by the nature of the individual and her relation to the action in question. The pragmatic stance, on the other hand, takes the second sort of response to be the more philosophically revealing, for such a response reflects an acute awareness that the attitudes, judgments, and practices whose appropriateness is in question are, after all, our attitudes, judgments, and practices-and shouldn't we be able to decide what to count as appropriate and legitimate instances of these in whatever ways we please?
From a pragmatic perspective, then, it appears that whatever facts are relevant to the justification of the activities connected to responsibility are relevant only because we choose to make them relevant, because, in other words, we set up roles that assign these facts a certain weight.
To be sure, the pragmatist will say, we must abide by the rules we establish unless we are prepared, in some general way, to revise them. So that if it is one of our rules that a person is not to be held responsible for actions she is unaware of performing, we must acknowledge that the fact of her ignorance, a fact of which we may be unaware, puts a limit on which judgments and attitudes will turn out to be justifiable. But which facts shall limit the justifiability of our attitudes, judgments, and practices depends on the rules that regulate the formation of these attitudes, judgments, and practices, and which rules regulate the formation of these attitudes, and so on, depends on us.
From the pragmatic perspective, there can be facts of the matter that constrain our attitudes, judgments, and actions insofar as they turn on the question of whether a person in a particular instance is to be accorded the status of a responsible being, but there can be no fact of the matter the discovery of which might undermine the justifiability of these practices as a whole. For if we are committed to a way of life that involves according most adult human beings the status of responsibility--and it seems we are so committed--then we are committed to setting up rules that govern the justifiable attributions of this status in such a way as to make these rules fit--that is, justify--at least the core cases of according this status to normal adults acting in normal situations. If the rules we set up seem to backfire on ourselves, so that by our own rules it appears that none of us are ever responsible, we should not take this as a discovery that we are not, in fact, responsible beings, but rather as a signal that we must revise or clarify our ideas about the conditions of responsibility.
The metaphysical stance toward free will and responsibility assumes there is a fixed fact of the matter about what it is to be a responsible being, though the fact is obscure and philosophical investigation, involving discovery and analysis, is required to bring it to light. Until this investigation is satisfactorily completed, according to this view, it is an open question whether anyone actually is responsible and indeed whether the concept of responsibility is a coherent one at all. According to the pragmatic perspective, however, this view is fundamentally wrong. What is fixed is that we are responsible beings-that is, most of us are, most of the time-and this is fixed not by a fact we happen to know to be true but by a commitment we have to making it true. The philosophical inquiry that is required, on this view, is, in the last analysis, not a theoretical investigation at all, not ultimately a matter of discovery. Rather, it is a practical inquiry involving forward-looking considerations, the ultimate goal of which is to arrive at the best system of rules to govern the practices associated with responsibility in ways most satisfactory to our needs.
There is a truth embodied in the pragmatic perspective which I hope will be, by the end of this book, better understood and given its full weight. But it is wrongly identified if it is understood to imply that we reject the metaphysical question of free will-the question that treats the claim that we have never been and can never be responsible as, at the outset, an epistemological possibility. It is a mistake to treat the question "Under what conditions do people deserve to be held accountable for their actions?" as ultimately and solely dependent on the practical questions of what conditions for desert will contribute to maximally meaningful and rewarding lives and what conditions for desert will most effectively encourage people to be morally good. For although the rules governing the complex of practices that are conceptually held together by their connection to the status of responsibility may not be given to us by "the nature of things" independent of our dispositions and choice, neither are they as flexible as to tolerate the sorts of revision that the adoption of the pragmatic approach would require. However we explain the development of this complex of practices, the development has already taken place. These practices, and the concepts, attitudes, and judgments that form essential parts of them, have established meanings, however obscure, and it is with these meanings that these practices and their components play the deep and pervasive role in our lives that they do.
Even attempts to revise or refine the rules of games like chess and baseball-examples, presumably as pure as can be, of activities whose rules are up to us to determine-are constrained by the requirement that certain elements be preserved, for without them the games in question would no longer be those games. It is an essential feature of the complex of practices associated with responsibility that they be subject to a justification involving reference to the status of responsibility. And it is an essential feature of the concept of responsibility involved that it be one that applies or fails to apply to a person as, in a certain sense, a matter of fact. Thus, the proposal to treat the question of whether we are ever responsible as something other than a matter of fact must ultimately be seen as a proposal not to deal with the problem of responsibility at all. Rather, it is a proposal to look at ourselves and our lives in such a way that the problem of responsibility loses its significance, a proposal that would reassign the significance it formerly had for us to other more tractable problems.
The aptness of this criticism is most clearly evident when it is leveled against philosophers like Moritz Schlick, for whom "the question of responsibility is the question: "Who, in a given case, is to be punished?" (or, as he later adds, rewarded). This, as a practical question, amounts to a call to develop the system of rewards and punishments that most effectively encourages people to act well rather than badly. Obviously, basing a decision to punish someone solely on the consideration that by doing so we will better succeed in influencing that person and other members of society to act in morally preferable ways in the future does not provide us with an interpretation of the concept of responsibility. Rather, this decision ignores the question of the agent's responsibility or deems the question to be irrelevant. This approach to the question of punishment and reward would have us understand these practices as species of moral manipulation, no different in kind from the strategies we use to toilet-train children or to tame circus animals. It disregards the difference in meaning we attach to praise and blame when these are expressions of admiration and indignation from the meaning overtly similar activities might have when they are motivated simply by the desire to reinforce or inhibit the kind of behavior in question. Thus, it leaves no room for any importance for the question of whether the agent is responsible. On this view, it does not matter whether the person deserves to be punished-if it would be useful to punish him, then we should do so whether or not his punishment is deserved.
Schlick's view, insofar as it is offered as a solution to the problem of responsibility, is naive and simplistic because it fails to recognize that the concept of responsibility is connected to the practices of reward and punishment only by way of and in the company of this concept's connection to subtler, less overt practices that involve attitudes, for example, of admiration and indignation, and judgments of agents' deserts. But one may adopt Schlick's basically pragmatic approach to the concept of responsibility without losing sight of the complexities of the role of this concept in shaping and defining our practices. If one searches for forward-looking reasons for maintaining not only our practices of reward and punishment but also the attitudes and judgments that form a basis for distinguishing between kinds of reward and punishment, one will find them. Just think how impoverished our lives would be if we ceased to take such attitudes and form such judgments-how much thinner would be our sense of self-worth if we ceased to feel pride, shame, and, guilt, and how much shallower our interpersonal relationships if we ceased to admire, respect, feel gratitude, resentment, and indignation toward one another.b Since there are forward-looking reasons for preserving the distinctiveness of these attitudes and judgments, it would seem that we can ask how best to preserve them, how best, that is, to set up the rules for these attitudes and judgments so as to encourage us to form them in ways that optimally enhance our sense of the meaning, value and richness of our lives. This more sophisticated form of a pragmatic approach less evidently fails to address the problem of responsibility that is connected to the justification of our practices-but it fails nonetheless.
For the sense that there is a fact of the matter as to whether a person is responsible for his actions, a sense that pragmatic approaches to responsibility either fail to notice or dismiss as confused, is an essential feature not only of the justification of a particular kind of punishment or reward but also of the justification of the attitudes and judgments this punishment or reward might express. We can make the sense in which the question of one's responsibility must be a matter of fact somewhat less abstract if we focus for a moment on a single type of attitude or judgment. Let us focus on resentment.
Clearly, we resent someone only if we think the person has done us some harm or done something that would have been expected to do us some harm. But it is essential to the nature of resentment that it can be deserved or undeserved, appropriate or inappropriate, where the conditions of desert or appropriateness go beyond the establishment of the fact that the person really did perform the action in question and that the action really did, or would have been expected to do, some harm to us. We do not, in general, resent someone who causes us a harm if her doing so was accidental or if she acted in ignorance or if her action was coerced or if she was under hypnosis. Our resentment is conditioned on the belief that in acting as she did she was expressing an attitude toward us of contempt or unconcern-or, at any rate, on the belief that in acting as she did she was failing to express an attitude of respect or consideration that we think she should have had. Moreover, our resenting her rests on the belief that her acting in a way that expresses this attitude toward us (or that reveals the absence of a different attitude toward us) was up to her. That is, we believe she had a kind of control over what she was doing that would have allowed her to take the attitude that we feel was our due, an attitude that, if she had taken it, would have led her to act differently. But there is a fact of the matter as to whether she did have that kind of control of her actions, as to whether the action that offended us was truly up to her in the sense required for her action to reveal an attitude toward us that we regard as insulting.
If we do not care about whether she had the relevant kind of control of her actions, then we do not care whether our resenting her is appropriate. But we cannot be said to be in a state of resenting her at all if we do not believe that our resentment may or may not be appropriate. And if we never cared whether our resentments were appropriate, then resentment would lose its distinctness from anger or annoyance or frustration that the world was not behaving as we would like it to behave.
The suggestion that we can decide what shall count as appropriate or inappropriate occasions for resentment would, if taken seriously, ultimately destroy our ability to feel resentment, as distinguished from other negative attitudes. That is, this suggestion would ultimately undermine the possibility of forming the very attitude it tries to legitimate. More generally, the suggestion that we take a pragmatic approach to the set of attitudes and practices associated with the concept of responsibility would, if taken seriously, ultimately undermine the whole complex of attitudes and practices whose preservation this approach is designed intentionally to protect.
The metaphysical stance is appropriate then, because the concepts of free will and responsibility that are already firmly established are intrinsically metaphysical concepts. That is, it is an essential criterion for the correct application of these concepts that they be subject to the demands of justification by the facts. Since we have a substantial interest in living in accordance with the facts, we have a substantial interest in finding out whether, in thinking of and treating ourselves and one another as free and responsible beings, we are living in a way that is justified by the facts or, at any rate, not in contradiction to them. And because the thought, or perhaps the un-thought assumption, that we are free and responsible beings plays so important a role in the way we think about ourselves and our relation to the world, the discovery of whether we are (internally) justified in having this thought or making this assumption will presumably lead to a significantly deeper understanding of ourselves.
If the answer to this question is positive, then it is possible that we will gain something more. For a positive answer to the metaphysical questions of free will and responsibility will tell us not only that we are justified but why. That is, it will tell us what features of ourselves are essential to our being free and responsible beings, and this may give us reason to care about these features in a special way and to arrange our lives and our values in such a way as to preserve and promote these features.
But it might turn out that the answer to this question is negative. In other words, we might be forced to conclude that we cannot coherently and justifiably continue to think of ourselves as free and responsible beings. In that case, it will be of great interest to know what deep-seated tendencies in human nature have encouraged and allowed us to be mistaken about so important a matter for so long. But it will also be necessary to come to terms with that conclusion. If we are not free and responsible beings, what then? This last question must be answered from a pragmatic perspective. It is, after all, a practical question. Thus, even from a metaphysical standpoint, the pragmatic perspective must remain available as, so to speak, a last resort. Indeed, there is reason to think that some of the proponents of the pragmatic approach have adopted it precisely because they believe that negative answers to the metaphysical questions of responsibility and free will are unavoidable and because, in light of this, they are impatient with philosophers who continue to seek a more positive conclusion.
This book will not even attempt to refute the view that negative answers to the metaphysical questions of responsibility and free will are unavoidable. To the extent that this book offers answers to these questions, they are answers that will not guarantee that we are, in the full metaphysical sense of these concepts, free and responsible beings. In fact, if the views in this book are correct, there can be no such guarantee. Nonetheless, the book addresses the metaphysical questions in a spirit that takes pessimism about them to be unwarranted, and in the hope that patience, in philosophy as elsewhere, will have its own proverbial rewards.