JOHN FISCHER: FREE WILL AND MORAL RESPONSIBILITY
-- The Determinism and Freedom Philosophy Website --
The subject of determinism and freedom has certainly been bound up with moral philosophy, among other things, and it continues to be. It is bound up with reasoning about what you ought to do, and holding people morally responsible for their actions, and the principle that someone can be judged only for doing or not doing what in fact he could do -- 'ought' implies 'can'. Does this commit you to believing in a freedom that you cannot have if determism is true? Or should we give up that seemingly strong or even indubitable principle? John Fischer, who also has another paper on this website, is a veteran of these philosophical contests, and teaches at the University of California at Riverside .
Much has been written recently about free will and moral responsibility. In this paper I will focus on the relationship between free will, on the one hand, and various notions that fall under the rubric of “morality,” broadly construed, on the other: deliberation and practical reasoning, moral responsibility, and ethical notions such as “ought,” “right,” “wrong,” “good,” and “bad.” I shall begin by laying out a natural understanding of freedom of the will. Next I develop some challenges to the common-sense view that we have this sort of freedom. I will go on to explore the implications of this challenge for deliberation, moral responsibility, and the central ethical notions.
II. Free Will and the Challenge from Causal Determinism
We naturally think of ourselves—“normal” adult human beings—as “free.” That is, we take it that we have a certain distinctive sort of control. I shall use “free will” (or “freedom of the will”) as an umbrella term to refer to the sort of freedom or control we presuppose that we human beings possess, and that is connected in important ways to ascriptions of moral responsibility. As I shall be employing the term, “free will” need not entail that we have a special faculty of the will, but only that we have a certain kind of freedom or control. But what is this freedom?
It is extremely natural and plausible to think that the typical adult human being has freedom in the sense that we often (although perhaps not always) have the freedom to choose or refrain from choosing a particular course of action (where “course of action” can refer to an omission as well as an action, narrowly construed), and to undertake or refrain from undertaking this course of action. That is, we take it that we often (although perhaps not invariably) have “alternative possibilities”: although we actually choose and undertake a particular course of action, we had it in our power (or “could have”) chosen and undertaken a different course of action. Of course, we recognize that sometimes we are “coerced” or “compelled” to choose or act as we do; and some individuals never have control over their choices and actions (because of significant mental illness, brain damage, and so forth). But we assume that the typical adult human being at least sometimes has more than one available path. That is, we assume, in Borges’ phrase, that the future is a garden of forking paths.
But there are various skeptical worries or challenges to the intuitive notion that we have free will in the sense that involves alternative possibilities. One of the most important such challenges comes from the doctrine of causal determinism. Causal determinism is the thesis that every event (and thus every choice and bit of behavior) is deterministically caused by some event in the past; thus, every choice and bit of behavior is the result of a casual chain, each link in which is deterministically caused by some prior link (until one gets to the beginning, if there is a beginning). More specifically, one can say that causal determinism is the doctrine that a complete statement of the laws of nature and a complete description of the temporally nonrelational or “genuine” facts about the world at some time T entail every truth about the world after T. That is, if causal determinism is true, then the past and the natural laws entail a unique present and future path for the world. Note further that if someone had available to her the description of the past and the statement of the laws, she could with certainty state what happens in the present and what will happen in the future. But it does not follow from the truth of the metaphysical doctrine of causal determinism that anyone actually has access to the relevant truths about the universe or its laws.
I contend that no human being currently knows whether or not the doctrine of causal determinism obtains. Certain physicists believe that the study of physical phenomena at the micro-level renders it very plausible that causal determinism is false (and thus that “indeterminism” is true.) Note, again, that indeterminism is a metaphysical, rather than an epistemic doctrine; that is, causal indeterminism posits indeterminacies in nature, not just incompleteness in our understanding of nature. But other physicists (and philosophers) cling to the view that causal determinism is true, and that what appear currently to be genuine metaphysical indeterminacies reflect mere inadequacies in our knowledge of the world.
Since we cannot be certain at this point that causal determinism is false, it is perhaps worthwhile to think about what would follow, if it turned out that causal determinism is true. It is troubling that there is a very potent argument, employing ingredients from common sense, which appears to show that if causal determinism indeed turned out to be true, then no human being would have free will in the sense that involves alternative possibilities. The argument appears to show that the future is not a garden of forking paths, on the assumption of causal determinism. The following is an informal and intuitive presentation of the argument.
Suppose I make some ordinary choice C at time T2. If causal determinism is true, then the total state of the universe at T1 together with the laws of nature entail that I make C at T2. Thus, it was a necessary condition of my making a different choice at T2 that either the state of the universe at T1 have been different from what it actually was, or some proposition that expressed a natural law would not have expressed a natural law. But, intuitively, I cannot—do not have it in my power—at any time so to behave that the past would have been different from the way it actually was. And, similarly, I cannot at any time determine which propositions express the natural laws. Intuitively, the past and the natural laws are “fixed” and not “up to me.” It seems to follow from the above ingredients that I could not have chosen otherwise than C at T2, if causal determinism turned out to be true.
Here is a slightly different way of presenting basically the same argument. As I suggested above, intuitively the past and the laws of nature are fixed and out of my control. The future is a garden of forking paths: the paths into the future extend from a given past, holding the laws of nature fixed. So, one might say that my freedom is the freedom to extend the actual (or given) past, holding fixed the laws of nature. Assume, again, the truth of causal determinism, and that I make choice C at T2. It follows from the assumption of causal determinism that the state of the world at T1 together with the laws of nature entail that I make choice C at T2. So in all possible worlds with the same laws as the actual world in which the past is just as it actually is, I make choice C at T2, Thus, it is logically impossible that my making some other choice C* at T2 be an extension of the given past, holding fixed the natural laws. It is evident, then, that if causal determinism is true, I cannot make any other choice than the one I actually make.
The above argument, suitably regimented and refined, appears to be generalizable to show that if it turns out that causal determinism is true, then no human being has the sort of free will that involves alternative possibilities—freedom to choose or do otherwise, or the power to select one path the world will take, from among various paths that are “genuinely” or “really” open. This argument for “incompatibilism”—the incompatibility of causal determinism and (in this instance) the sort of free will that involves alternative possibilities—has been the focal point of much discussion. Although the argument is controversial, here I shall not explore the ways in which it can be resisted. Rather, I shall assume that the argument is sound and explore the implications of this assumption. As I proceed in this paper, I shall focus on the question of what would follow, in terms of “morality,” broadly construed, if we in fact lack the sort of free will that involves alternative possibilities. Also, I shall consider whether there are features of causal determinism which would threaten morality, apart from its ruling out free will (in the sense that involves alternative possibilities).
III. Deliberation and Practical Reasoning
III.1. Taylor and van Inwagen. One of the most central aspects of human “persons” is that we can engage in significant deliberation and practical reasoning. In deliberating, we consider and weigh reasons for (and against) various courses of action. We seek to “figure out what is best to do,” and to act in accordance with this sort of judgment about what is best, all-things-considered. We of course are fallible in our judgments, and certainly we sometimes fail to act in accordance with our judgment about what is best to do, all-things-considered. But in any case the process of deliberation (or practical reasoning) involves identifying and weighing reasons with an eye to figuring out what we have sufficient reason to do.
Some philosophers have argued that it is a conceptual truth that I cannot engage in deliberation, if I do not believe that I have free will, in the sense that involves alternative possibilities. After pointing out that I can deliberate only about my own behavior (and not the behavior of another), that I can deliberate only about future things (rather than present or past things), and that I cannot deliberate about what I already know that I am going to do, Richard Taylor adds, “And, finally, I cannot deliberate about what to do, even though I may not know what I am going to do, unless I believe that it is up to me what I am going to do.” He goes on to argue that the relevant notion of “up to me” is incompatible with causal determinism; on this notion, an act’s being “up to me” implies that it up to me whether or not I do it.
I am not convinced by Taylor that I would not or could not engage in deliberation, if I believed that causal determinism were true and thus that I have it in my power only to choose to do (and to do) what I actually choose to do (and do). As long as I do not know what I will in fact choose, it seems that there is a perfectly reasonable point to deliberation; after all, I still need to figure out what I have sufficient reason to do and to seek to act in accordance with this judgment. This purpose of deliberation would not disappear, in a world in which I knew that it is not “up to me” (in the sense that involves alternative possibilities in which the actual past and natural laws are held fixed) what I will choose. Note that it may still be true, even in a causally deterministic world, that in a particular context I would choose a course of action if and only if I were to judge it best. Further, it does not follow simply from causal determinism that there is some special sort of obstacle to my choosing a particular course of action; causal determinism does not entail that I have some kind of phobia or compulsion that would rule out my choosing a certain sort of action. And if one insists that it is a conceptual truth that my process of weighing reasons would not count as “deliberation,” then so be it: call it “deliberation*” or simply “figuring out what it would be best to do,” and there can be a clear point to such activities even in a world in which I know that I have only one path that is genuinely available into the future.
Peter van Inwagen holds a view which is similar to, but slightly different from, Taylor’s. On van Inwagen’s account, an agent who believes that he does not have free will (in the sense of alternative possibilities) can deliberate, but in so doing he would be contradicting himself. Van Inwagen says, “In my view, if someone deliberates about whether to do A or to do B, it follows that his behavior manifests a belief that it is possible for him to do A—that he can do A, that he has it within his power to do A—and that it is possible for him to do B.” Thus, an individual who sincerely believes that he lacks free will (understood as above) would be contradicting himself in deliberating—he would be holding an inconsistent set of beliefs. Whereas this is not impossible, it is certainly undesirable; for example, holding inconsistent beliefs guarantees that at least one of one’s beliefs is false.
But I am not convinced that van Inwagen is correct to say that deliberation manifests the belief in free will (understood as above). He says:
Anyone who doubts that this is indeed the case may find it instructive to imagine that he is in a room with two doors and that he believes one of the doors to be unlocked and the other to be locked and impassable, though he has no idea which is which; let him then attempt to imagine himself deliberating about which door to leave by.
I agree that it would be odd to think that I could deliberate about which door actually (or “successfully”) to open. But surely in such a case I could deliberate about which door to choose to open. That is, I could weigh reasons and come to a judgment about which door it would be best to seek to open, and I could form an intention—choose—to act in accordance with my judgment. There is not the same intuitive oddness about saying that I could deliberate about which door to choose to push against that there is to saying that I could deliberate about which door “to leave by.” It is important to note that van Inwagen does not purport to be offering an argument for the contention that anyone who deliberates must believe that he has alternative possibilities, apart from his invocation of the example of the alleged oddness of deliberating about which door to leave by.
But van Inwagen may reply that the apparent lack of oddness in supposing that I could deliberate about which door (say) to choose to open stems precisely from the fact that I can suppose that I am able either to choose to open door A or choose to open door B. I am not so sure, however, that this is the explanation of the asymmetry in our intuitions between deliberating about which door to open and deliberating about which door to choose to open. Suppose I do in fact choose to open door A. Now if causal determinism is true and the argument for the incompatibility of causal determinism and free will (understood as involving alternative possibilities) is sound, then it turns out that, unbeknownst to me, just prior to my choice I did not have it in my power to choose to open door B. Further, it seems to me that I could know that causal determinism is true and that the incompatibilist’s argument is sound, and thus that whichever choice I make is the only one I actually can make. This knowledge does not eliminate the point of deliberation (the need to figure out which door it would be best to choose to open); and I do not have any hesitation in supposing that, even with the knowledge that whatever door I choose will be the only door I in fact can choose to open, I can deliberate about which door to choose to open. Thus I do not believe that the asymmetry in our intuitions between deliberating about which door to open and deliberating about which door to choose to open stems from an asymmetry in our beliefs about alternative possibilities.
In a causally deterministic world (and given the incompatibilistic argument), every choice and action would be such that, if I make it (or perform it), I could not have made another choice (or performed another action.) But it seems to me that there could still be a perfectly reasonable point to deliberation, and that I need not contradict myself in accepting the truth of causal determinism, the soundness of the argument for incompatibilism, but nevertheless deliberating. All that is required is that I have an interest in figuring out what I have sufficient reason to choose, and that I do not know which course of action I will in fact choose to take (and take). Further, van Inwagen has not produced an example in which it is obvious that this yields an odd result.
III.2 Searle. John Searle has argued for a point related to the claims of Taylor and van Inwagen, but it is slightly different. Searle’s contention is that there would be no point to practical reasoning or deliberation, if I knew that causal determinism were true. Searle says:
The gap can be given two equivalent descriptions, one forward-looking, one backward. Forward: the gap is that feature of our conscious decision making and acting where we sense alternative future decisions and actions as causally open to us. Backward: the gap is that feature of conscious decision making and acting whereby the reasons preceding the decisions and the actions are not experienced by the agent as setting causally sufficient conditions for the decisions and actions. As far as our conscious experiences are concerned, the gap occurs when the beliefs, desires, and other reasons are not experienced as causally sufficient conditions for a decision (the formation of a prior intention…)
Searle goes on to state:
I am advancing three theses here.
1. We have experiences of the gap of the sort I have described.
2. We have to presuppose the gap. We have to presuppose that the psychological antecedents of many of our decisions and actions do not set causally sufficient conditions for those decisions and actions.
3. In normal conscious life one cannot avoid choosing and deciding.
Here is the argument for 2 and 3: If I really thought that the beliefs and desires were sufficient to cause the action then I could just sit back and watch the action unfold in the same way as I do when I sit back and watch the action unfold on a movie screen. But I cannot do that when I am engaging in rational decision making and acting. I have to presuppose that the antecedent set of psychological conditions was not causally sufficient. Furthermore, here is an additional argument for point 3: even if I became convinced of the falsity of the thesis of the gap, all the same I would still have to engage in actions and thus exercise my own freedom no matter what…
For example, there is a kind of practical inconsistency in maintaining the following two theses:
1. I am now trying to make up my mind whom to vote for in the next election.
2. I take the existing psychological causes operating on me right now to be causally sufficient to determine whom I am going to vote for.
The inconsistency comes out in the fact that if I really believe 2, then there seems no point in making the effort involved in 1. The situation would be like taking a pill that I am sure will cure my headache by itself, and then trying to add some further psychological effort to the effects of the pill. If I really believe the pill is enough, then the rational thing to do is to sit back and let it take effect.
In discussing Searle’s view, I would first point out that when Searle first introduces the notion of a “gap,” it is a point about our experiences. Recall that he says, for instance, “the gap is that feature of our conscious decision making and acting where we sense alternative future decisions and actions as causally open to us.” But he goes on to say, “We have to presuppose the gap. We have to presuppose that the psychological antecedents of many of our decisions and actions do not set causally sufficient conditions for those decisions and actions.” If the second sentence of the latter quotation is intended as exegetical, then “the gap” is now thought to be not so much a feature of our phenomenology, but of the objective relationship between our mental states. This is somewhat confusing. From now on, I will take “the gap thesis” to be the claim that both our experience and the objective reality of the relationship between our mental states is indeterministic.
I believe that Searle’s view is incorrect. Note that Leibniz describes what is essentially this view as the “Lazy Sophism”:
This … demolishes… what the ancients [the Stoics, perhaps following Cicero] called the ‘Lazy Sophism,’ which ended in a decision to do nothing: for (people would say) if what I ask is to happen it will happen even though I should do nothing; and if it is not to happen it will never happen, no matter what trouble I take to achieve it… But the answer is quite ready: the effect being certain, the cause that shall produce it is certain also; and if the effect comes about it will be by virtue of a proportionate cause. Thus your laziness perchance will bring it about that you will obtain naught of what you desire, and that you will fall into those misfortunes which you would by acting with care have avoided. We see, therefore, that the connexion of causes with effects, far from causing an unendurable fatality, provides rather a means of obviating it.
It seems to me that Searle’s view about deliberation falls prey to the same objections as the views of Taylor and van Inwagen. I believe that there would be a clear point to deliberation and practical reasoning, even if I were to reject the gap: I would still have an interest in—and deeply care about—figuring out what I have reason to do, and seeking to act accordingly. Even if the gap thesis is false, and antecedent psychological states are causally sufficient for my decision, and I know this, it does not follow that I know what decision I will make and what action I will perform. Hence, insofar as I care about acting in accordance with what I have all-things-considered reason to do, there is a clear point to engaging in deliberation.
Recall that Searle says that there is a practical inconsistency in maintaining the following:
1. I am now trying to make up my mind whom to vote for in the next election.
2. I take the existing psychological causes operating on me right now to be causally sufficient to determine whom I am going to vote for.
He says holding these two theses would be like “taking a pill that I am sure will cure my headache by itself, and then trying to add some further psychological effort to the effects of the pill.” But in Searle’s analogy you know that the pill will cure your headache; in contrast, I am not assumed to know whom I will vote for in the next election. If I did know whom I would vote for, I agree that the point of making up my mind would appear to vanish.
Suppose I know that my decision about the next election is causally determined by my current configuration of mental states (desires, beliefs, and so forth). Still, I can also know that my decision will depend on my practical reasoning in the following sense: if I were to judge it best, all-things-considered, to vote for Candidate A, I would vote for Candidate A; but if were to to judge it best, all-things-considered, to vote for Candidate B, I would vote for Candidate B. Further, I can know that nothing distorts or impairs my practical reasoning—my ability to recognize the reasons there are, and to weigh them with an eye to making an all-things-considered judgment as to what is best. That is, nothing in the doctrine of causal determinism entails that the counterfactuals (that specify the relevant sort of dependency) are false, and nothing in this doctrine entails that I have any special sort of impairment of my capacity to engage in practical reasoning—certain phobias, compulsions, mental illnesses, and so forth. And, finally, nothing in the doctrine of causal determinism entails that I do not care about choosing and acting in accordance with my judgment about what is best to do. So there is a clear point to deliberation, even if I believe that antecedent mental states are causally sufficient for my decision.
Imagine, to make the point dramatically, that there are two doors in front of you, and you must choose which door to open. You know that behind door 1 is a million dollars, and behind door 2 is a den of rattlesnakes. Imagine, further, that you know that causal determinism is true, that causal determinism rules out alternative possibilities, and that causal determinism in itself does not entail that one has any physical paralysis or impairment of the human capacity for practical reasoning (no intense phobias, compulsions, paranoid schizophrenia, and so forth). Would Searle really not deliberate? What would he do—flip a coin, act arbitrarily, or what? Would he simply “sit back and watch the action unfold?” It would seem perfectly reasonable (at the very least) to take into consideration what is behind the doors, and to choose and act accordingly. Having collected the million dollars, you might pause to reflect that it turns out that that was the only thing you could have done (as long as this thought would not unduly delay the celebration!).
Searle admits that it is conceivable that our experience of indeterminism does not map onto the reality of the brain (and that the neuro-biological events are causally deterministic); but he argues against this as follows:
This result, however, is intellectually very unsatisfying, because, in a word, it is a modified form of epiphenomenalism. It says that the psychological processes of rational decision making do not really matter. The entire system is deterministic at the bottom level, and the idea that the top level has an element of freedom is simply a systematic illusion. …The thesis is epiphenomenalistic in this respect: there is a feature of our conscious lives, rational decision making and trying to carry out the decision, where we experience the gap and we experience the processes as making a causal difference to our behavior, but they do not in fact make any difference. The bodily movements were going to be exactly the same regardless of how these processes occurred.
Maybe that is how it will turn out, but if so, the hypothesis seems to me to run against everything we know about evolution. It would have the consequence that the incredibly elaborate, complex, sensitive, and—above all—biologically expensive system of human and animal conscious rational decision making would actually make no difference whatever to the life and survival of the organisms. Epiphenomenalism is a possible thesis, but it is absolutely incredible, and if we seriously accepted it, it would make a change in our worldview, that is, in our conception of our relations to the world, more radical than any previous change, including the Copernican Revolution, Einsteinian relativity theory, and quantum mechanics.
Why would [the hypothesis under consideration] render consciousness any more epiphenomenal than any other higher-level feature of a physical system? After all, the solidity of the piston in the car engine is entirely explained by the behavior of the molecules but that does not render solidity epiphenomenal. The difference is this: the essential characteristics of solidity matter to the performance of the engine, but the essential characteristic of conscious decision making, the experience of the gap, would not matter in the least to the performance of the agent. The bodily movements would have been the same, regardless of the experiences of the gap.
I have argued above that we do not need to presuppose a “gap” of the sort to which Searle is referring in order to engage in practical reasoning. I have suggested that practical reasoning may require an “epistemic gap”—it may be necessary that we not know exactly what we will choose and do, in order for there to be a point to practical reasoning (and deliberation). Searle’s gap then is not an “essential characteristic of conscious decision making.” And the epistemic gap clearly would make a difference: if it didn’t exist, it may well not be reasonable to deliberate, and so my bodily movements might be quite different.
Note that, on the view of practical reasoning I am suggesting, psychological processes of rational decision making do matter in a straightforward sense: if my deliberations had gone differently (and had thus issued in a different judgment as to what is best, all-things-considered), then my decisions and bodily movements would have been different. It is not the case that the bodily movements are going to be exactly the same, regardless of how my deliberations go. This surely is the important point about the causal efficacy of practical reasoning. So “the incredibly elaborate, complex, sensitive, and—above all—biologically expensive system of human and animal conscious rational decision making” does make a difference to the life and survival of organisms. Surely what is evolutionarily important in our capacities for practical reasoning is a certain capacity to recognize and respond to reasons; it seems bizarre to suppose that what is crucial to our survival—and the crowning glory of evolution--is the experience of the causal insufficiency of our mental states! If there is a gap here at all, it is in Searle’s argument.
III.4. Kantian approaches. I further contend that I can at the same time (or from the same “perspective”) acknowledge both that my choice is causally determined (and thus that I have but one path genuinely available to me) and deliberate about which choice to make. That is, when I am engaged in practical reasoning and deliberation, I can continue to believe, and to acknowledge, that I am causally determined and thus not free. This follows from the fact that the theses I acknowledge are metaphysical contentions the truth of which can leave an epistemic gap and from the distinctive purpose of practical reasoning. I can thus accept that the characteristic purposes of theoretical and practical reasoning diverge, while maintaining that an agent engaged in practical reasoning can in fact continue to hold such deliverances of theoretical reasoning as that he is causally determined and thus not free (in the sense of possessing alternative possibilities, construed incompatibilistically).
My view here is in stark contrast with the “neo-Kantian” two-perspective approach developed by such philosophers as Hilary Bok and Christine Korsgaard. For example, Hilary Bok says:
If, when we engage in practical reasoning, we must regard ourselves as standing in the order of reasons rather than the order of causes, and if those orders are distinguished from one another by the relations of necessity to which they appeal, when we engage in practical reasoning we will not regard ourselves as subject to the same sort of necessity appealed to by theoretical reason. Theoretical necessity is causal: one object acts on another, thereby rendering some change in the latter necessary. To see oneself as necessitated in this way is to see oneself as passive: acted on rather than acting.
Further, she says:
Because we regard ourselves as subject not to causal but to rational necessity, when we engage in practical reasoning we regard ourselves not as the passive object of external forces but as determining our own conduct; not as acted on by things outside us but as choosing for reasons that we are free to accept or reject. And we regard these choices not as events that might simply befall us and with which we might or might not identify but as necessarily our own. For these reasons, as Christine Korsgaard writes, ‘[a]t the moment of decision, you must regard yourself as the author of your action.’
But whereas I agree that at the moment of decision, one must in some suitable sense see oneself as the author of one’s decision, I do not think that it follows that one must at that moment believe (either occurrently or dispositionally) that one is not causally determined. I certainly do not think that it should be accepted as uncontroversial that it follows from my choice’s being causally determined that I am not the author of it or that I am merely passive with respect to it—these claims require argumentation, as there are ways of seeking to explain authorship and the difference between activity and passivity that are consistent with causal determinism.
Additionally, the quotations from Korsgaard and Bok raise the vexing issue of the relationship between their notion of “regarding” and the more ordinary notion of “believing.” With respect to this issue, consider the following passage from Bok:
Insofar as regarding our choices as caused involves regarding them as determined by antecedent events, we cannot regard ourselves as caused to choose as we do when we engage in practical reasoning. [Here Bok inserts a footnote pointing us to Kant, Grounding for the Metaphysics of Morals, p. 50 (Ak. 448); and Korsaard, “Morality as Freedom,” pp. 162-3.] This is not because we believe we are not caused to choose as we do, but because when we engage in practical reasoning, we are concerned with another form of determination.
Bok, however, faces the following dilemma. When we engage in practical reasoning, either we do in fact believe that we are not causally determined or we do not so believe. If we do, then it is obvious that a belief we have from the practical perspective can come into direct conflict with a belief we could have from the theoretical perspective. But it is a central feature of Bok’s approach that the two perspectives cannot conflict in this way; the claim that the two perspectives cannot conflict is essential for Bok’s project of showing freedom to be compatible with causal determinism.
Thus it seems as if Bok must say that, when we engage in practical reasoning, we do not believe (even dispositionally) that we are not causally determined. (I suppose the picture here is that, when one takes up the practical perspective, one does not believe in either causal determinism or its denial—one fails to form either of these beliefs.) But this leaves the notion of “regarding” somewhat mysterious; it seems as if from the practical perspective we regard ourselves as not subject to causal necessity but we do not believe we are not subject to causal necessity.
But if regarding is prized apart from believing in this way, what exactly is it to regard ourselves as not subject to causal necessity? Further, I find it unattractive to suppose that from the practical perspective I cannot have (even dispositionally) a belief such as that causal determinism is false. Of course, the mere fact that, when engaged in practical reasoning, I am “concerned with” another form of necessitation does not entail that I do not—perhaps dispositionally—believe that I am in fact subject to causal necessitation. After all, when I am “concerned with” the leaking plumbing in my house, it does not follow that I do not believe (perhaps dispositionally) that the house is painted white (or that George Washington was the first President of the United States). If I do in fact have the belief that causal determinism is false, then why should I be precluded from having access (even dispositionally) to this belief when I take up the practical perspective? On this picture, the practical perspective is epistemically partitioned off from the rest of the agent in a puzzling way. The resulting compartmentalization is very unattractive, and, as I have suggested above, unnecessary.
IV. Moral Responsibility
IV. 1. The Concept of Moral Responsibility. Some philosophers have argued that if we lacked free will (in the sense that involves alternative possibilities), then we could not legitimately be considered morally responsible agents. There are, of course, different accounts of the concept of moral responsibility, as well as its conditions of application. I will simply sketch three views about the concept (or “nature”) of moral responsibility; an elaboration of these accounts is beyond the scope of this paper.
On the first view about the nature of moral responsibility, an agent’s moral responsibility consists in his or her being an appropriate candidate for ascriptions of certain ethical predicates, such as “good,” “bad,” “courageous,” “charitable,” “dastardly,” “cruel,” and so forth. The view is often put in terms of a metaphor; on this approach, an agent is morally responsible insofar as he has a “moral ledger.” The ascription of moral predicates corresponds to making marks on the ledger.
A second view contends that when an agent is morally responsible for some behavior, it would not be inappropriate to expect the agent to provide an explanation of the behavior in question. On this view, when the agent is morally responsible in this sense, it follows that he has a moral ledger; but it is the expectation that the agent can provide a certain sort of explanation that is the essence of moral responsibility.
A third sort of account of the nature of moral responsibility follows Peter Strawson. On this view, roughly speaking, an individual is morally responsible for some behavior in virtue of being an apt target for one of the “reactive attitudes” on the basis of the behavior. According to Strawson, the reactive attitudes include gratitude, indignation, resentment, love, respect, and forgiveness, and they manifest our involvement with other human beings in distinctively interpersonal relationships. There are various versions of the “Strawsonian” approach to the concept of moral responsibility.
In what follows I shall not take a stand on the correct account of the concept of moral responsibility. I shall simply speak of moral responsibility and let the reader fill in her favorite account of its nature. No matter what particular account of the concept of moral responsibility one accepts, it is clear that if it turned out that human beings lacked free will, there would be a deep and disturbing challenge to the idea that we are in fact morally responsible.
IV. 2. The Principle of Alternative Possibilities (PAP) and the Frankfurt-type Examples. As I suggested above, we naturally think that the future is a garden of forking paths—that we at least at some important points in our lives have more than one path branching into the future. If this intuitive picture turned out to be false, then it would seem that we could not legitimately be held morally responsible for our behavior. After all, if I don’t have free will in a sense that involves alternative possibilities, then I have to choose (and do) what I actually choose (and do). And if I have to choose what I do in fact choose, then presumably I am compelled so to choose, and cannot fairly be considered morally responsible for my choice. It is very plausible, then, to accept something like the “Principle of Alternative Possibilities” (PAP), according to which an agent is morally responsible for (say) an action only if he could have done otherwise. If (PAP) is true, then moral responsibility requires free will (in the sense that involves alternative possibilities); and if causal determinism rules out such alternative possibilities, it would thereby rule out moral responsibility.
Peter van Inwagen gives a particularly pointed defense of (PAP):
If we do not have free will, then there is no such thing as moral responsibility. This proposition, one might think, certainly deserves to be a commonplace. If someone charges you with, say, lying, and if you can convince him that it was simply not within your power not to lie, then it would seem that you have done all that is necessary to absolve yourself of responsibility for lying.
… without free will there is no moral responsibility: if moral responsibility exists, then someone is morally responsible for something he has done or for something he has left undone; to be morally responsible for some act or failure to act is at least to be able to have acted otherwise, whatever else it may involve; to be able to have acted otherwise is to have free will. Therefore, if moral responsibility exists, someone has free will. Therefore, if no one has free will, moral responsibility does not exist.
Whereas (PAP) might appear to be an obvious truth, it has been questioned by some philosophers. These philosophers contend (in one way or another) that what matters for moral responsibility is how the relevant choice or action is brought about, not whether the agent has alternative possibilities available to him. In contemporary philosophy, Harry Frankfurt has helped to focus the case against (PAP) with a set of examples with a characteristic structure. These examples contain failsafe mechanisms that (allegedly) both make it the case that the agent has no (relevant) alternative possibilities and also play no role in the agent’s actual choice and action. Frankfurt says that if something plays no role in the agent’s choice and action, then it cannot be relevant to his moral responsibility; thus, it would follow that the mechanisms in question both make it the case that the agent has no alternative possibilities and do not thereby threaten the agent’s moral responsibility.
Here is a version of my favorite “Frankfurt-type case.” Jones is in a voting booth deliberating about whether to vote for the Democrat or the Republican. After weighing reasons and deliberating in the “normal” way, he chooses to vote for the Democrat. Unbeknownst to him, Black, a neurosurgeon with Democratic sympathies, has implanted a device in Jones’s brain which monitors Jones’s brain activities. If he is about to choose to vote Democratic, the device does not intervene. If, however, Jones is about to choose to vote Republican, the device triggers an intervention which involves electronic stimulation of the brain sufficient to produce a choice to vote for the Democrat and an actual vote for the Democrat.
Now one might ask how the device can tell whether Jones is about to choose to vote Republican or Democratic. Frankfurt himself did not say much about this difficult problem, except that “Black is an excellent judge of such things.” We can however add a “prior sign” to the case as follows. If Jones is about to choose at T2 to vote for the Democrat at T3, he shows some involuntary sign—say a blush, a furrowed brow, or a neurological pattern in his brain readable by some sort of “neuroscope”—at T1. If it detects this, Black’s device does not intervene. But if Jones is about to choose at T2 to vote Republican at T3, he shows a different involuntary sign at T1. This would trigger Black’s device to intervene and cause Jones to choose at T2 to vote for the Democrat and actually to vote for the Democrat at T3.
It seems that Black’s device is precisely the kind of failsafe device described above: it plays no role in Jones’s deliberations, choice, or action, and yet its presence renders it true that Jones could not have done otherwise than choose and vote Democratic. Indeed, it seems that in this case Jones freely chooses to vote Democratic, freely does so, and can be considered morally responsible for his choice and action, even though he does not have alternative possibilities (given the presence of Black’s device). This suggests that there is a kind of freedom or control—corresponding to choosing and acting freely--that does not require alternative possibilities, and that this sort of control (and not the alternative-possibilities control) is the freedom-relevant condition necessary for moral responsibility. There seem to be two kinds of freedom or control, and the Frankfurt-type examples help us to prize them apart. It appears, then, that we have a counterexample to (PAP).
IV.3.A Dilemma for the Frankfurt-type Examples. The suggestion (emerging from the Frankfurt-type examples) that moral responsibility does not require free will in the sense that involves alternative possibilities has not been entirely irresistible. In fact, a huge literature has developed surrounding these examples. Consider the following dilemma in response to the Frankfurt-type examples. Notice that in the typical presentation of the examples (as above) it is not made explicit whether causal determinism obtains. So suppose first that causal determinism obtains in the example. Now it would seem question-begging to conclude straightforwardly from the example that Jones is morally responsible for voting for the Democrat; after all, the issue of whether causal determinism is compatible with moral responsibility is in dispute. But if it is assumed that causal determinism is false, and specifically that there is no deterministic relationship between the prior sign at T1 and Jones’s subsequent choice at T2, then Jones would appear to have free will at (or just prior to) T2: he can at least begin to choose to vote for the Democrat. After all, given the prior sign and the laws of nature, it does not follow that Jones will choose at T2 to vote for the Democrat (on the current assumption of causal indeterminism). So, the proponent of the dilemma says that either Jones is not morally responsible or there are alternative possibilities for Jones: one does not have a single context in which it is both true that Jones has no alternative possibilities and is morally responsible for his choice and action.
This is indeed a worrisome challenge to the conclusion that I (and others) draw from the Frankfurt-type examples—that (PAP) is false. Elsewhere I have presented a strategy of response to the dilemma. Here I wish briefly to sketch this response, and then to consider an important objection to it.
IV.4. A Response to the Dilemma. First consider the possibility that causal determinism is false (in the relevant way). Various philosophers have proposed that one can construct versions of the Frankfurt-type cases in which the agent is morally responsible and yet there are no alternatives at all, or at least no robust alternatives. I think it is promising that such an example can be constructed, although I shall not attempt to defend this possibility here.
Suppose that causal determinism is true. That is, suppose that there is indeed a causally deterministic relationship between the sign exhibited at T1 and Jones’s choice to vote for the Democrat at T2. Now it follows, given the argument for the incompatibility of causal determinism and the sort of control that involves alternative possibilities, that Jones does not have the power at T2 to refrain from choosing to vote Democratic at T2. It is not my strategy, however, simply to claim that Jones is obviously morally responsible for his choice; I agree that this would not be dialectically kosher.
Rather, I begin by suggesting that the fact that Black’s device would intervene and ensure that Jones would choose to vote for the Democrat (and indeed vote for the Democrat), if Jones had shown a different sign at T1, does not in itself show that Jones is not morally responsible for his actual choice (if he is in fact not morally responsible). That is, I am not supposing at this point that Jones is morally responsible for his actual choice at T2 to vote for the Democrat. Rather, I am saying that the fact that he cannot do otherwise does not in itself (and apart from indicating or pointing to some other fact) make it the case that Jones is not morally responsible for his choice at T2.
It seems evident to me that the fact that Black’s device would intervene in the counterfactual scenario and ensure that Jones choose to vote for the Democrat is irrelevant to the “grounding” of Jones’s actual moral responsibility for choosing to vote for the Democrat (and actually doing so). Something grounds moral responsibility, in the sense in question, insofar as it explains (or helps to explain) why the agent is morally responsible, apart from simply being an indicator of something else that in fact explains the agent’s moral responsibility. Black’s counterfactual intervention does not make any difference as to Jones’s moral responsibility; if Black’s device were “subtracted” from the example (to use Frankfurt’s phrase), this would not change my assessment of Jones’s moral responsibility in any way. Thus, I think that the example renders it plausible (although it does not decisively establish) that Jones’s lack of alternative possibilities is irrelevant to the grounding of Jones’s moral responsibility.
It is important to be a bit more careful here. I have claimed that consideration of the example of Jones (a typical Frankfurt-type case) should first elicit the intuition that the fact that there is a failsafe device present that would intervene in the counterfactual scenario is irrelevant to the grounding of Jones’s moral responsibility. My contention is that this then suggests that even if Jones had no alternative possibilities at all, this would be irrelevant to the grounding of his moral responsibility. It would then follow that in a causally deterministic world, in which it is assumed that Jones has no alternative possibilities at all, his lack of alternative possibilities would be irrelevant to the grounding of his moral responsibility. That is, his lack of alternative possibilities cannot in itself and apart from indicating something else explain why Jones is not morally responsible, if Jones is in fact not morally responsible.
In my view, this then is the moral of the Frankfurt-type cases. They suggest that alternative possibilities are irrelevant to the grounding of moral responsibility. Thus they are an important step along the way toward arguing that causal determinism is compatible with moral responsibility. Of course, someone might say that alternative possibilities are a necessary condition for moral responsibility because their presence indicates some other factor (perhaps causal indeterminism in the actual sequence) which must be present for there to be moral responsibility. This is a perfectly reasonable position, which can then be addressed; I shall briefly discuss this maneuver below. But it does not diminish the importance of the moral of the Frankfurt-type cases; in my view, once one establishes that alternative possibilities are irrelevant to the grounding of moral responsibility, it is considerably easier to argue that causal determinism (and the lack of alternative possibilities) is compatible with moral responsibility.
IV.5. A Recent Objection and a Further Reply. Before I address the contention that alternative possibilities indicate some other factor that grounds moral responsibility, I wish to consider a recent objection to my strategy for dealing with the “deterministic” horn of the dilemmatic attack on the Frankfurt-type examples. My contention is that Black’s presence and counterfactual intervention is irrelevant to the grounding of moral responsibility. But someone might grant this, while insisting that it is not pertinent, since it is not Black’s counterfactual intervention, but the condition of the world at T1 (including the sign Jones exhibits) that makes it true that Jones does not have it in his power at T2 to choose to vote for the Democrat. If it is the condition of the world at T1 that makes it true that Jones cannot at T2 choose to vote for the Democrat, then it is not so obvious that what makes it the case that Jones cannot at T2 choose otherwise is irrelevant to the grounding of Jones’ moral responsibility. This line of attack has been developed by Stewart Goetz.
[The Frankfurt-style example] creates the appearance that it is Black’s device, which is in the alternative sequence of events, that makes it the case that Jones is not free to choose otherwise. This appearance is illusory because without the obtaining of causal determinism in the actual sequence of events, the device cannot prevent Jones from making an alternative choice, and with causal determinism in the actual sequence of events it is not the device that prevents Jones from making an alternative choice. In short, if Jones is not free to choose otherwise, it is because of the occurrence of causal determinism in the actual sequence of events and not because of Black’s device in the alternative sequence.
Goetz goes on to say:
[Fischer’s strategy] requires the truth of causal determinism in order to create the illusion that it is the presence of something in the alternative sequence of events (e.g., Black’s device) that makes it the case that Jones is not free to choose otherwise. It is only through this illusion that one is tempted to endorse the conclusion of the first step of Fischer’s argument, which is that the lack of alternative possibilities is not sufficient for the lack of moral responsibility, and, thereby, Jones might be morally responsible even though he is not free to choose otherwise. Once this illusion is exposed, one’s initial conviction that the lack of an alternative choice is sufficient for the lack of moral responsibility is vindicated.
Goetz’s point could be put as follows. What really makes it the case that Jones cannot choose otherwise at T2 is the prior state of the world together with the laws of nature. In other words, what makes it the case that Jones lacks an alternative possibility at T2 is causal determination in the actual sequence. So it is quite beside the point that Black’s counterfactual intervention is irrelevant to the grounding of Jones’s moral responsibility; after all, it is not Black’s counterfactual intervention that makes it the case that Jones cannot choose otherwise at T2. Thus we do not have a case in which the fact that the agent could not have chosen (or done) otherwise is irrelevant to the grounding of his moral responsibility.
Frankfurt-type scenarios are cases in which an action is “causally overdetermined.” The overdetermination is considered “pre-emptive,” rather than “simultaneous.” In simultaneous overdetermination, two causal sequences both operate and actually issue in the overdetermined action (or event). In pre-emptive overdetermination, some event is actually caused in a certain way, and would have been caused in a different way, had the actual causal sequence not taken place. So, in the Frankfurt-type scenario presented above, Black’s device is part of what makes it the case that Jones’s choice at T2 is pre-emptively overdetermined.
What is also of note is that in these scenarios Jones’s inability to choose (or do) otherwise is also overdetermined. (I would describe the overdetermination here as simultaneous, rather than pre-emptive; but is it delicate, as the thing in question is a fact—a modal fact—rather than a concrete action or event.) It is not only the act that is overdetermined; it is also the agent’s lack of alternative possibilities. So, my response to Goetz is as follows.
In the Frankfurt-type scenario, two causes make it the case that Jones is unable to choose otherwise at T2: the prior condition of the world (together with the laws of nature) and Black’s counterfactual intervention. What the examples show is that the mere fact that Jones is unable to choose otherwise does not in itself establish that Jones is not morally responsible for his choice. This is because Black’s counterfactual intervention is one of the factors that make it the case that Jones is unable to choose otherwise at T2, and yet it is irrelevant to the grounding of Jones’s moral responsibility. Considering this factor (the counterfactual intervention), and bracketing any other factor that might make it the case that Jones is unable to choose otherwise at T2, it seems to me that Jones may well be morally responsible for his action. The mere fact that he lacks alternative possibilities, then, cannot in itself be the reason that Jones is not morally responsible, if indeed he is not morally responsible.
Now of course it is also true that the prior condition of the world together with the natural laws makes it the case that Jones lacks alternative possibilities. But, given that the mere fact of lacking alternative possibilities does not in itself rule out moral responsibility, why should this way of lacking alternative possibilities rule out moral responsibility? Why exactly should the significance of causal determination be that it rules out moral responsibility? This is exactly what the Frankfurt-type examples call into question.
In an interesting passage, Goetz states:
The proponent of PAP thinks that the lack of the freedom to choose otherwise does not by itself explain the absence of moral responsibility. This is because he believes that when this lack obtains, its obtaining is itself explained by, and can only be explained by, the occurrence of causal determinism in the actual sequence of events. What the advocate of PAP believes, then, is that when an agent is not morally responsible because he is not free to choose otherwise, he lacks moral responsibility not simply because he is not free to choose otherwise but because he is not free to choose otherwise because of causal determinism.
Precisely this move is made by Derk Pereboom. But why exactly does it matter that causal determination rules out alternative possibilities? If the mere fact of the lack of alternative possibilities does not in itself rule out moral responsibility, why would a particular way of expunging alternative possibilities rule out moral responsibility? Granted that causal determination is a certain way of taking away alternative possibilities. Why should it be thought that causal determination threatens moral responsibility in virtue of constituting a way of ruling out alternative possibilities? The Frankfurt-type examples, then, suggest that one needs to look in a different direction if one seeks to argue that causal determination rules out moral responsibility.
The dialectic could be put somewhat differently. There can be two different ways in which some factor renders an agent unable to choose or do otherwise (or eliminates alternative possibilities). In one way, the factor does not play a role in the actual sequence; it does not flow through the actual course of events. In another way, the factor does flow through the actual sequence. The Frankfurt-type scenarios are all cases in which the ability-undermining factor does not play a role in the actual sequence leading to the relevant choice and action. So it seems unfair to extrapolate from the Frankfurt-type cases to the other sort of cases; that is, even if we are inclined to say that the agent is morally responsible in the Frankfurt-type cases, it would not follow that the agent is morally responsible in a causally deterministic world.
To reply. I grant that the Frankfurt-type examples are not decisive. That is, they do not provide examples that would absolutely and uncontroversially decide the issue about the relationship between causal determinism and moral responsibility. But I certainly do not believe that it is reasonable to expect such examples here—or in any contentious area of philosophy! And, as I argued above, the Frankfurt-type examples suggest that if causal determination is indeed problematic, it is not so in virtue of flowing through the actual sequence and thereby ruling out alternative possibilities.
IV.6. Source Incompatibilism. So far I have been primarily concerned with the issue of whether alternative possibilities are relevant to the grounding of moral responsibility. As I pointed out above, even if they are irrelevant to the grounding of responsibility, they may nevertheless be relevant to responsibility as a sign of something else which in fact grounds moral responsibility. Many years ago I emphasized that the mere fact that the Frankfurt-type examples show that alternative possibilities are not required for moral responsibility does not in itself show that causal determinism is compatible with moral responsibility. I pointed out that causal determinism is a thesis about the “actual sequence,” and thus that it does not follow from the falsity of (PAP) that causal determinism is compatible with moral responsibility. As I put it:
Both the compatibilist and the incompatibilist alike can unite in conceding that enough information is encoded in the actual sequence to ground our responsibility attributions; as philosophers we need to decode this information and see whether it is consistent with deterministic causation.
In my subsequent work I have explored various ways in which it might be thought that causal determination in the actual sequence rules out moral responsibility. I have in the end concluded that causal determination in the actual sequence does not rule out moral responsibility.
Other philosophers have disagreed, contending that causal determination in the actual sequence rules out moral responsibility “directly” (and not in virtue of expunging alternative possibilities) . Robert Kane has argued that in order to be morally responsible, we have to meet a condition of “ultimacy,” according to which the “causal buck must stop here”; that is, we cannot be mere intermediate links in a causally deterministic sequence which begins prior to our births. Similarly, Laura Ekstrom argues that the past and laws “push” us into our choices and actions, if causal determinism is true. She thus argues that causal determination in the actual sequence is incompatible with moral responsibility. Additionally, although Derk Pereboom believes that versions of the Frankfurt-type examples successfully show that (PAP) is false, he nevertheless defends the following principle:
An action is free in the sense required for moral responsibility only if it is not produced by a deterministic process that traces back to causal factors beyond the agent’s control.
I believe that none of the arguments purporting to show that causal determination in the actual sequence rules out moral responsibility is particularly strong, although this of course is a highly contentious matter. One of the difficulties is to see how to argue for the incompatibility claim; after all, Pereboom’s principle seems to be a simple restatement of incompatibilism about causal determinism and moral responsibility, not an argument for it. In any case, I would contend that the Frankfurt-type cases at least help us to make progress toward defending the compatibility of causal determinism and moral responsibility insofar as they help us to take a very important first step: they render it plausible (although they do not decisively establish) that the sort of free will that involves alternative possibilities does not ground attributions of moral responsibility, that is, it does not in itself and apart from indicating some other factor explain why we are morally responsible, if we are in fact morally responsible.
V. Ethical Judgments
V.1. Judgments of Deontic Morality. Above I pointed out that there are various accounts of the concept of moral responsibility. On the “ledger view,” if an agent is morally responsible, then he has a moral ledger—the marks correspond to various sorts of moral judgments. Some philosophers hold that these judgments include claims about what the agent ought and ought not to do, and what is right or wrong for the individual to do. These philosophers thus connect moral responsibility tightly to the appropriateness of judgments about ought, ought not, right, and wrong. Peter van Inwagen appears to make this sort of connection in the continuation of a passage quoted above:
If someone charges you with, say, lying, and if you can convince him that it was simply not within your power not to lie, then it would seem that you have done all that is necessary to absolve yourself of responsibility for lying. Your accuser cannot say, ‘I concede it was not within your power not to lie; none the less you ought not to have lied’.
On this sort of approach to moral responsibility, if (contrary to van Inwagen) one has successfully defended the compatibility of causal determinism (and the lack of the sort of free will that involves alternative possibilities) with moral responsibility, one has thereby defended the compatibility of causal determinism (and the lack of free will) with judgments employing “ought,” “ought not,” “right,” and “wrong.” But one might accept an alternative account of the concept of moral responsibility, or even a ledger view according to which the relevant “marks” correspond to (say) “goodness” and “badness,” but not “ought,” “ought not,” and so forth. If one accepted (say) a Strawsonian account of moral responsibility (or the sort of ledger view just sketched), it might be that causal determinism is compatible with moral responsibility but not with judgments employing “ought,” “ought not,” “right,” and “wrong.” This is precisely the view held by Ishtiyaque Haji. Haji accepts the conclusion of the Frankfurt-type cases that moral responsibility does not require alternative possibilities and further that it is compatible with causal determinism; but he rejects the contention that causal determinism is compatible with judgments employing “obligation,” “ought,” “ought not,” “right,” and “wrong.” (Following Haji, let us call the latter “judgments of deontic morality.”) On this sort of view, the Strawsonian “reactive attitudes” are prized apart from the judgments of deontic morality, and whereas the former are compatible with causal determinism, the latter are not. (Note that Haji distinguishes judgments pertaining to notions such “goodness” and “badness” from the judgments of deontic morality; he is willing to concede that the former sorts of judgments are entirely compatible with causal determinism.)
Why might one think that the judgments of deontic morality are incompatible with causal determinism? I will treat “ought not” and “wrong” as interchangeable, and “ought” and “obligatory” as interchangeable. I shall lay out the argument with respect to “wrong.” It will be easy to see how to construct parallel arguments for the other judgments of deontic morality. Here is a simple version of the argument:
1. Suppose some individual, John, does something morally wrong.
2. If John’s Xing was wrong, then he ought to have done something else instead.
3. If John ought to have done something else instead, then he could have done something else instead.
4. So John could have done something else instead.
5. But if causal determinism is true, then John could not have done anything other than he actually did.
6. So, if causal determinism is true, it cannot be the case that John’s Xing was wrong.
This is a potent and disturbing argument. I have sought to argue that causal determinism is compatible with moral responsibility. This result would be considerably less interesting if causal determinism were nevertheless incompatible with the central judgments of deontic morality. There are however various ways of seeking to block the conclusion of the argument. I have discussed the rejection of premise 2 elsewhere. Here however I shall focus on the rejection of premise 3 (and thus the rejection of the ought-implies-can maxim (henceforth, “the Maxim”).
V.2 Copp’s Defense of the Maxim. I believe that there are Frankfurt-type omissions cases that are relevantly similar to Frankfurt-type cases with respect to actions. That is, there are cases in which an agent is morally responsible for not Xing, although he cannot in fact X. Some of these are cases in which an agent is blameworthy for not Xing and yet he cannot X. In fact, I believe that anyone who accepts the Frankfurt-type action cases must accept that there are such omissions-cases. Further, the basic intuitions elicited by the Frankfurt-type cases conflict with the Maxim and cast doubt on its intuitive plausibility. Although this certainly does not decisively refute the maxim, it does suggest that it is not ad hoc for anyone who accepts that the Frankfurt-type cases show that moral responsibility does not require alternative possibilities to reject the Maxim.
But rejection of the Maxim comes at a steep price. In the most detailed, sustained, and penetrating discussion of the motivation for the Maxim of which I am aware, David Copp contends that it is preferable to preserve the Maxim than to reject PAP on the basis of the Frankfurt-type cases (or on any other basis). Copp presents two primary arguments on behalf of the Maxim, and it will be useful to discuss each of them.
Copp contends that there is a conflict between the interpretation of the Frankfurt-type cases according to which they show that moral responsibility does not require alternative possibilities and the Maxim. According to Copp, “… [this] is not a conflict between intuition and a recherché theoretical proposition. It is a conflict among intuitions.” Copp
The most basic motivation for the Maxim, it seems to me, begins with the thought that it would be unfair to expect a person to do something, or to demand or require that she do it, if she lacked the ability to do it. This thought is about what we might call ‘agent-requirements,’ which arise in cases in which an authoritative agent requires someone under her authority or jurisdiction to do something. An example might be a situation in which a boss requires an employee to do something that the employee lacks the ability to do. A supervisor at the post office might demand that a mail carrier cook a soufflé for everyone in the post office in the next five minutes when the mail carrier does not even know what a soufflé is. We can imagine many similar cases, including cases in which a parent expects a child to do something she cannot do, or a teacher requires a student to do something she cannot do, or a sergeant requires a recruit to do something she cannot do. The intuition is that agent-requirements of this kind are morally unfair when the person of whom the demand is made lacks the ability to comply.
Copp goes on to claim that although the intution elicited above is about “agent-requirements” rather than “moral requirements,” he contends that a similar point applies to moral requirements. As Copp puts it, “if there would be unfairness in the latter case [the mere agent-requirement case], then there is surely a kind of unfairness in the moral requirement in the former case even if there is no agent who is being unfair.” So the first argument in favor of the Maxim is the intuition that it would be unfair to morally require someone to do something, if he cannot do the thing in question.
Copp’s second argument in favor of the Maxim is based on meta-ethical considerations about the “point” of moral requirements. Copp says:
The heart of the argument is roughly as follows: any moral theory must somehow account for, or make room for, the intuition that there is a point to requiring an action, namely, crudely, to get it done. Clearly, moreover, an action will not be done if the prospective actor cannot perform it.
… The argument can be summarized as follows. If an agent is morally required to do A in a particular situation, then all other options she faces are morally ruled out. If the agent cannot do A, then doing A is not among her options. Hence, if an agent is morally required to do A but cannot do A, then all of her options are morally ruled out. But information that an agent is morally required to do something provides her with guidance among her options by distinguishing between options that are morally ruled out and options that are not morally ruled out. If all of an agent’s options are morally ruled out by a moral requirement, then information about the requirement cannot provide her with such guidance. Given then that moral requirements have a characteristic relevance to our decisions, by distinguishing between options that are morally ruled out and options that are not morally ruled out, it follows that if a person cannot do A, it is not the case that she is morally required to do A. That is, the Maxim follows from the intuition about the relevance of moral requirements to decision-making.
Copp thus offers two strategies for motivating the Maxim: the fairness argument and the argument from the relevance of moral requirements to decision-making. He thus points out that if one favors one’s intuition that the Frankfurt-type examples show (albeit not decisively) that PAP is to be rejected, then one must give up strong intuitions about fairness and the relationship between morality and practical reasoning. Copp thinks that giving up these latter intuitions would be too steep a price to pay. In accepting PAP, however, Copp admits that his view might be open to incompatibilist worries; if causal determinism turned out to be true along with PAP, then there emerges the danger that no one could legitimately be accountable (blameworthy) for what they do. In the end, Copp concludes that “any adequate analysis of the ability to act must be compatibilist. It must be such that the ability to do something other than what one actually does is compatible with determinism.”
Let us first consider Copp’s argument from fairness. More specifically, the contention is that it would be unfair to hold someone blameworthy for failing to do X, if he could not do X. In order to bolster this judgment, Copp invokes an example in which it does seem unfair to require a mail carrier to cook a soufflé for everyone in the post office in five minutes. But I would reply that it is crucial to distinguish two importantly different sorts of omissions: “simple” and “complex” omissions. If one focuses solely on complex omissions, it does indeed seem as if it would be unfair to hold an individual blameworthy for failing to do X, if he is unable to do X. But my intuitions about simple omissions are quite different.
Consider an example offered by Harry Frankfurt:
Imagine that a person—call him ‘Stanley’—deliberately keeps himself very still. He refrains, for some reason, from moving his body at all. … suppose that here is someone with a powerful interest in having Stanley refrain from making any deliberate movements, who arranges things in such a way that Stanley will be stricken with general paralysis if he shows any inclination to move. Nonetheless, Stanley may keep himself still quite on his own altogether independently of this person’s schemes. Why should Stanley not be morally responsible for keeping still, in that case, just as much as if there had been nothing to prevent him from moving had he chosen to do so?
I agree with Frankfurt here. And surely Stanley could be considered blameworthy, should something morally important hang on his moving his body rather than keeping still.
Stanley’s not moving his body, or refraining from moving, is a “simple omission”: the omission is entirely constituted by his failure to move his body. There are many more such omissions, and in these cases it is plausible that the agents are indeed morally responsible—and potentially morally blameworthy—although they could not have refrained from keeping still.
I do not have any “proof” of my contention that in simple omissions, an agent can be blameworthy for failing to do X, even though he could not have done X. It seems to me however that this is a completely reasonable intuition, shared by many philosophers and supported by a range of examples. I agree however that it seems upon initial consideration that in cases of complex omissions, an agent cannot be blameworthy for failing to do X, unless he can in fact do X.
My purpose here is simply to suggest that there are cases in which an agent can legitimately be considered blameworthy for failing to do X, although he could not have done X. I would contend that Copp fails to see this because he focuses entirely on a proper subset of cases—the complex omissions. If I am correct, then the argument from fairness is vitiated—there are cases in which it would not be unfair to blame someone for failing to do something he could not do (and never could do).
Now I suppose someone could say that because it is obviously unfair to blame someone for his failure in the complex omissions cases, we should conclude that it would also be unfair to blame the agent in the simple omissions cases. But I think that this gets the dialectic wrong: we are supposed to be generating general principles by reference to intuitions about all of the relevant cases. It would seem inappropriate to generate such a principle based on a proper subset of the cases, and then apply it to all of the cases, even when it does not seem to yield the correct results in all of the cases.
What would be ideal is a theory that explains exactly why the agents are indeed morally responsible in the simple omissions cases and not morally responsible in the complex omissions cases (discussed above), Such an explanation would obviously not invoke the notion of inability to do otherwise, lest it lead to implausible results in the simple omissions cases. I (and my co-author) have offered just such a theory of moral responsibility for omissions; on this approach, moral responsibility is associated with freedom (or control), but not the sort of freedom (or control) that involves alternative possibilities. Quite apart from whether this theory is adequate, my point here is that Copp has not really motivated the central claim of the argument for fairness: he has relied on only a proper subset of the relevant data.
Copp’s second argument in favor of the Maxim pertains to the role of moral requirements in guiding action. I agree that moral requirements play a distinctive and important role in guiding our practical reasoning (and, thus, our behavior). But, as above in our discussion of practical reasoning and deliberation, it is crucial to distinguish between genuine metaphysical possibilities and possibilities, for all the agent knows, or “epistemic possibility.” An epistemic possibility is not ruled out by the agent’s knowledge. Indispensable to the proper analysis of deliberation and also the Frankfurt-type examples is the fact that one’s metaphysical possibilities (the paths that are genuinely available to one) may diverge from one’s epistemic possibilities (that paths that are, for all one knows, available to one). I would contend that moral requirements rule out certain of the courses of action that are, for all we know, open to us—certain epistemic possibilities.
Recall Copp’s argument, which begins as follows:
If an agent is morally required to do A in a particular situation, then all other options she faces are morally ruled out. If the agent cannot do A, then doing A is not among her options. Hence, if an agent is morally required to do A but cannot do A, then all of her options are morally ruled out. But information that an agent is morally required to do something provides her with guidance among her options by distinguishing between options that are morally ruled out and options that are not morally ruled out.
Given the distinction between the two different kinds of possibilities, the argument becomes:
Given an agent is morally required to do A in a particular situation, then all other epistemic options she faces are morally ruled out. If the agent cannot do A, then doing A is not among her metaphysical options. Hence, if an agent is morally required to do A but cannot do A, then all of her options are morally ruled out. But information that an agent is morally required to do something provides her with guidance among her options by distinguishing between options that are morally ruled out and options that are not morally ruled out.
It is evident where the problems lie. The conclusion that if an agent is morally required to do A but cannot do A, then all of her options are morally ruled out, infelicitously elides the distinction between epistemic and metaphysical options. From the mere fact that an agent lacks a certain metaphysical option it does not follow that she lacks the corresponding epistemic option. So, from the mere fact that an agent in fact cannot do A, it does not follow that she knows that he cannot do A. Thus, all that follows from the moral requirement and the metaphysical fact is that all of the agent’s epistemic alternatives are ruled out, except A. But there is nothing problematic about this; and now the moral requirement can have its distinctive role in guiding deliberation and action. Moral requirements insert themselves into the space of epistemic possibilities, not directly into the space of metaphysical possibilities.
I conclude that despite his noteworthy efforts, David Copp has not successfully presented a compelling motivation for the Maxim. If we reject the Maxim, we can reject PAP. And we are thus not pushed toward a compatibilist account of freedom; as I explained above, a compatibilist must say that we are free either to “change” the past or the natural laws. That is, the compatibilist must deny that our freedom is the freedom to extend the given past, holding the laws of nature fixed. But this is quite implausible.
Causal determinism threatens our intuitive and natural view of ourselves as having free will in the sense that involves genuinely available alternative possibilities. It threatens the common sense view that the future is a garden of forking paths. For all we know, causal determinism might turn out to be true. In this paper I have explored the question of what would be lost in a world without free will of this sort. Would there still be a point to deliberation and practical reasoning? Could there be moral responsibility and ethical judgments?
The discovery that causal determinism is true would significantly alter our picture of ourselves: in my view, giving up the view that the future is a garden of forking paths is a major change, with important resonances in the way we understand and couch our deliberation, moral responsibility, and ethical judgments. But I do not believe that we would need entirely to jettison any of these aspects of our moral lives. I believe that deliberation, moral responsibility, and judgments of deontic morality are compatible with causal determinism and the lack of free will (in the sense involving alternative possibilities, understood as above).
Other philosophers are not so sanguine, and there is a bewildering distribution of views on these issues. As we have seen, Peter van Inwagen is a philosopher who believes that causal determinism and the lack of free will would rule out both moral responsibility and judgments of deontic morality, as well as rendering us inconsistent every time we deliberate. Thus, van Inwagen and I represent, as it were, “corner positions.” There are various “in-between” views. Ishiyaque Haji contends that causal determinism and the lack of alternative possibilities are completely compatible with robust moral responsibility, but not judgments of deontic morality. In contrast, Derk Pereboom is willing to concede that robust moral responsibility (involving, say, reactive attitudes such as indignation and resentment) does not require free will in the sense that involves alternative possibilities, but he insists that causal determinism indeed rules out such moral responsibility. Nevertheless, he believes that causal determinism is compatible with judgments of deontic morality. Similarly, Saul Smilansky holds that causal determinism rules out robust moral responsibility, but not the judgments of deontic morality. Both Pereboom and Smilansky argue that although causal determinism would rule out robust moral responsibility, it still leaves room for something akin to moral responsibility—something which is significant and valuable. It also leaves room for various ethical judgments. So whereas Haji thinks that the more significant threat from causal determinism is to the judgments of deontic morality, Pereboom and Smilansky argue quite the opposite. In contrast, both Van Inwagen and I view the threats from causal determinism as equal in strength (although we come to opposite conclusions).
A recurrent theme has been the difference between an agent’s epistemic possibilities and metaphysical possibilities, given the truth of causal determinism. This disparity is crucial to understanding practical reasoning and deliberation in a causally deterministic world. It is also an indispensable ingredient in the description of the Frankfurt-type examples. Additionally, the non-identity of these two sets of possibilities explains how moral requirements can play their signature role of guiding action, even in a causally deterministic world. On my view, the collapse of these two sets into one—the set of metaphysical possibilities—would be a dramatic as the collapse of the wave pocket in quantum mechanics. My view is the opposite of the famous Biblical contention that the truth shall make us free. But this is really not surprising: if I genuinely knew all my future choices and behavior, then it would seem to me that I could just sit back and let the future unroll.
John Searle writes:
Suppose you go into a restaurant, and the waiter brings you the menu. You have a choice between, let’s say, veal chops and spaghetti; you cannot say: ‘Look, I am a determinist, che sara, sara. I will just wait and see what I order! I will wait to see what my beliefs and desires cause.’
Given the fact that the sets of metaphysical and epistemic possibilities are not identical, no determinist need reason in the indicated way. But if we collapse the sets into one, “che sara sara” would not be inappropriate, or out of tune.
This article is a somewhat expanded version of one that is forthcoming in David Copp's Handbook on Ethical Theory (Oxford University Press, 2004). It had a lot of footntes, but technology has prevented their being available to us.
HOME to Det & Free front page
HOME to T.H. front page