Compatibilism Old and New

 Richard Hudelson

University of Wisconsin Superior

 

            In Problems of Ethics, first published in 1931, Moritz Schlick offered a discussion of the "so-called problem of the freedom of the will."  In fact, Schlick argued, the problem was a "pseudo-problem" arising from some elementary confusions about the nature of freedom and responsibility.  Schlick is clear about the centrality of the question of responsibility.  He tells us that it is concerning "responsibility" that "misunderstanding arises" and that "the concept of responsibility constitutes our theme."[1]  In attempting to answer the question of what we mean by saying that someone is responsible, Schlick proposes to focus on the practice of punishment as a model for ascriptions of responsibility. Concerning punishment, Schlick has this to say:

"What is punishment, actually?  The view still often expressed, that it is a natural retaliation for past wrong, ought no longer to be defended in cultivated society; for the opinion that an increase in sorrow can be "made good again" by further sorrow is altogether barbarous.  Certainly the origin of punishment may lie in an impulse of retaliation or vengeance; but what is such an impulse except the instinctive desire to destroy the cause of the deed to be avenged, by the destruction of or injury to the malefactor?  Punishment is concerned only with the institution of causes, of motives of conduct, and this alone is its meaning.  Punishment is an educative measure, and as such is a means to the formation of motives, which are in part to prevent the wrongdoer from repeating the act (reformation) and in part to prevent others from committing a similar act (intimidation).  Analogously, in the case of reward we are concerned with an incentive."[2]

For present purposes, let us leave aside the apparent conflations of retaliation, vengeance, retribution, and restitution.  For now, what is interesting about this passage is its thoroughly consequentialist framework.  What justifies punishment, in Schlick's view, is it consequences in preventing like actions in the future.  "Punishment is concerned only [my emphasis] with the institution of causes, of motives of conduct."  "This alone [my emphasis] is its meaning."

            Moritz Schlick was one of the leading figures in the Vienna Circle School of logical positivism.  His consequentialist compatibilism was shared by fellow logical positivists like A.J. Ayer and Otto Neurath.  But it also represented a much broader and older current within philosophy, reaching back to the 19th century positivism of J.S. Mill and even further back to 18th century empiricists like Jeremy Bentham and David Hume.  Underlying this whole tradition is a metaphysics of naturalism which rejects postulation of any order of reality transcending the natural universe.  Opposed to this naturalism were the transcendental idealism of Immanuel Kant and the schools of German Idealism and Neo-Kantianism that dominated academic philosophy in the 19th century.  In one way or another, Kant and the Idealists and Neo-Kantians who followed him placed limits on the naturalistic perspective, insisting on the necessity of a metaphysical realm transcending the natural order.  Further, in practical reason, in our ability to discern right from wrong and freely act on that discernment, Kant and the critics of naturalism claimed to find signs pointing to this transcendent metaphysical realm.

            For some 200 years, from the work of Hume into the second half of the 20th century, this opposition between naturalism and transcendental metaphysics was intertwined with an opposition between consequentialist and deontological standpoints in ethics.  Consequentialists saw the rightness of an act as determined by the consequences of the act to promote or retard the total amount of happiness in the world.  Deontologists emphasized notions of moral desert, retribution and moral rights that could not be reduced to consequentialist terms.  Deontologists like Kant saw such notions as directly rooted in the dignity of human beings as inhabitants of the metaphysical realm transcending the natural order.  Naturalists, like the logical positivist, Otto Neurath, regarded the central concepts of deontological ethics as "remnants of metaphysics" to be replaced by "a thoroughly empirical `felicitology' (Felicitologie), on a behaviorist foundation, which would take the place of traditional ethics."[3]

            The account of moral responsibility offered by Moritz Schlick must be understood within this context.  His characterizations of retributive theories of punishment as "no longer to be defended in cultivated society" and as "barbarous" are inextricably intertwined with his rejection of supernaturalism in metaphysics.  So too is his consequentialist analysis of punishment.  If punishment is concerned "only" with the education of wrongdoers and the deterrence of others, if "this alone is the meaning of punishment," then there is no need for deontological notions like desert and retribution that find their home in a non-naturalist landscape.  Nor is there any conflict with determinism in this consequentialist perspective.  As Schlick himself emphasizes, on the consequentialist view, punishment itself "is concerned only with the institution of causes."

            By the middle of the 20th century, this consequentialist compatibilism had established itself as the dominant view in academic philosophy in the English speaking world.  But, developments in the second half of the 20th century have made this particular form of compatibilism less attractive.  In particular, there has been a large shift in ethics away from utilitarian consequentialism toward a more "Kantian" deontological perspective, a perspective that puts more emphasis on notions of human dignity, human rights, desert and retributive theories of punishment.[4]  The revival of deontological theories in ethics, based in part on a widespread conviction of the inadequacy of consequentialism, necessitates a reconsideration of compatibilist reconciliations of determinism and moral responsibility.

            On the surface it might appear that all that is required is minor surgery on the compatibilist position.  It might appear that all that is required is the excision of the consequentialist theory of responsibility and its replacement with a deontologically enriched account of responsibility.  But this ignores the centrality of the consequentialist account of responsibility to the plausibility of compatibilism.  For, there seems to be little doubt that determinism is compatible with deterrent and rehabilitative theories of punishment and with accounts of moral responsibility built around such consequentialist theories.  Punishment for purposes of deterrence, rehabilitation, or incapacitation presents no apparent conflict with determinism or metaphysical naturalism.   But it is much less clear that determinism is compatible with retributive theories of punishment or the deontologically enriched accounts of responsibility that go with them.  Can we say that a person deserves punishment as retribution for past actions and at the same time hold that a deterministic naturalism is true?

            The shift from consequentialist to deontological approaches in ethics has important consequences for the problem of determinism and moral responsibility.  The significance of this shift has not always been appreciated.  Textbooks in philosophy continue to include consequentialist versions of compatibilism from Mill, Schlick, Ayer, Hook, or others, without including any essay critical of their crucial reliance on consequentialist moral theory.  And, in philosophy of mind and other areas of philosophy where philosophers continue to defend a deterministic and naturalistic metaphysics, appeals can be found to consequentialist compatibilism to avoid apparent threats to human moral responsibility, appeals which are flatly contradictory to the now dominant currents in ethics.   Consider, for example, the analysis of responsibility offered by Daniel Dennett in his widely acclaimed book on determinism and free will published in 1984:

Why then do we want to hold people—ourselves included—responsible?  "By holding someone responsible and acting accordingly we may cause him to shed an undesirable trait, and this is useful regardless of whether the trait is of his making."  (Gomberg, 1978, p. 208).  Once again, the utility of a certain measure of arbitrariness is made visible.  Instead of investigating, endlessly, in an attempt to discover whether or not a particular trait is of someone's making—instead of trying to assay exactly to what degree a particular self is self-made—we simply hold people responsible for their conduct (within limits we take care not to examine too closely).  And we are rewarded for adopting this strategy by the higher proportion of "responsible" behavior we thereby inculcate.[5] 

Like Schlick, Dennett here follows the well-trod path of consequentialist compatibilism favored by philosophers committed to metaphysical naturalism.  Now I have no quarrel with what Dennett has to say here.  My point is just that such a teleological conception of responsibility will not satisfy the requirements of the deontologically enriched notions of desert currently popular in ethics.

            Within the specialized literature on determinism and responsibility the debate over compatibilism continues more or less informed by the deontological revival in ethics.  Aware of the multiple senses of "compatibilism," Ted Honderich has prominently argued for the incompatibility of determinism with notions of retribution and desert found within deontological ethical theory while recognizing the compatibility of determinism with consequentialist ethical theory.[6]  Like the logical positivists and their empiricist predecessors, Honderich favors abandoning core deontological concepts, going so far as to argue that the truth of determinism "requires that we change our social institutions and practices in so far as they are owed to our image of origination."[7] This puts Honderich squarely at odds with the deontological turn in ethics so prominent in the last quarter of the twentieth century, though nowhere is this noted.   A number of other philosophers have agreed with Honderich in thinking that deterministic naturalism cannot be reconciled with deontological notions of desert and retribution.  Among them are Daniel Dennett and Bruce Waller, philosophers convinced of the truth of deterministic naturalism, and Robert Kane and Robert Nozick, philosophers who appeal to non-deterministic accounts of human action.[8]  But again, participants in these discussions have not situated the current debates over compatibilism in reference to the changed landscape in ethics.

 

            In The Metaphysics of Free Will (Cambridge, MA: Blackwell, 1994), John Martin Fischer provides an intelligent, informed and careful discussion of recent work on the problem of determinism and moral responsibility that does not rely on the consequentialist analysis of responsibility.[9]  While I shall not attempt to do justice to the whole of this work, I would like to mention two aspects of it.

            Fischer defends a position that he calls "semi-compatibilism."  The semi-compatibilist holds that while determinism may be incompatible with the freedom to do otherwise, still determinism is compatible with moral responsibility.   Central to Fischer's position is his rejection of the claim that moral responsibility requires the freedom to do otherwise.  His rejection of that claim relies upon the kind of counterexamples first introduced by Harry Frankfurt.  Fischer gives us this version of the Frankfurt-type counterexamples:

Imagine, if you will, that Black is a quite nifty (and even generally nice) neurosurgeon.  But in performing an operation on Jones to remove a brain tumor, Black inserts a mechanism into Jones' brain which enables Black to monitor and control Jones' activities.  Jones, meanwhile, knows nothing of this.  Black exercises this control through a sophisticated computer which he has programmed so that, among other things, it monitors Jones' voting behavior.  If Jones were to show any inclination to vote for Bush, then the computer, through the mechanism in Jones' brain, intervenes to assure that he actually decides to vote for Clinton and does so vote.  But if Jones decides on his own to vote for Clinton, the computer does nothing but continue to monitor--without affecting--the goings-on in Jones' head.  Suppose that Jones decides to vote for Clinton on his own, just as he would have if Black had not inserted the mechanism into his head.  It seems, upon first thinking about this case, that Jones can be held morally responsible for his choice and act of voting for Clinton, although he could not have chosen otherwise and he could not have done otherwise.[10]

Such Frankfurt-type counter-examples are supposed to show that we are mistaken in thinking that even a fundamentally deontological notion of moral responsibility requires the ability to have chosen otherwise. 

            Many philosophers have found such counter-examples unpersuasive.  They argue that we are inclined to hold Jones responsible in such a case because if we trace the actual history of his voting for Clinton backward in time we find that this vote did in fact follow from a choice Jones made in a setting where an alternative path was possible for him.  To be sure, had Jones initiated this alternative path, the mechanism installed by Black would have intervened to close this option for Jones.  But, nonetheless, there remains, even in the Frankfurt-type cases, this moment of choice where Jones does face open alternative paths.  There remains a flicker of freedom in the Frankfurt-type cases that grounds our attribution of responsibility.[11]  Though acknowledging the "undeniable appeal of the flicker of freedom strategy," Fischer is not persuaded by it and stands by the Frankfurt-type cases as successful counter-examples to the view that responsibility requires alternative possibilities.  Fischer argues that, while there does remain a flicker of freedom in such cases, "this alternative possibility is not sufficiently robust to ground the relevant attributions of moral responsibility."  He holds that "it is highly implausible to suppose that it is in virtue of the existence of such an alternative possibility that Jones is morally responsible for what he does."[12]

            According to the flicker of freedom strategy, what makes Jones responsible for voting for Clinton is that his vote was the end result of a path of action chosen by Jones at a moment of choice in which Jones confronted genuinely open alternative pathways.  While it is true that had Jones chosen otherwise his choice would have been ineffectual, overruled by the mechanism in his brain, in fact the actual outcome was the terminus of a path of action initiated by Jones's free choice.  Against this view, Fischer argues that it is possible to reconstruct the example in such a way as to undercut the flicker of freedom strategy:

Suppose we again consider the version of the Jones and Black case in which Black can be alerted to Jones' future inclination to vote for Bush by the presence of some involuntary sign, such as a blush or twitch or even a complex neurophysiological pattern.  So if Jones were (say) to blush red, then Black could intervene prior to Jones' doing anything freely and ensure that Jones indeed votes for Clinton.  Here the "triggering event" (i.e., what would trigger the intervention of Black) is not any sort of initiating action, and thus cannot be said to be freely done.  Again, precisely as above, this sort of triggering event appears to be not sufficiently robust to ground responsibility ascriptions. [13]

But surely the reply to this is that what makes Jones responsible for voting for Clinton is not the involuntary blush prior to Jones' taking the pathway leading to voting for Clinton but rather Jones' choice of the pathway leading to voting to Clinton rather than the alternative pathway open to him at the moment of choice subsequent in time to the involuntary blush. 

            Fischer goes on to argue for the view that in order for a person to be responsible for an action it must be the case that the mechanisms leading up to an agent's performance of that action are to at least some extent responsive to possible reasons for the non-performance of that action.[14]   Fischer is probably right in thinking that such a reason-responsiveness condition is a necessary condition for moral responsibility.  It is probably also true that some sort of condition requiring that the agent identify with the action is a necessary condition for the agent's being morally responsible for that action.[15]  Failure to satisfy the identification condition, the reason-responsiveness condition or the non-coercion condition is normally behind real world difficulties in determining whether or not an agent is rightly to be held responsible for an action.  But that questions regarding the satisfaction of these conditions is commonly behind practical questions of responsibility does not show the non-relevance of a contra-causal alternative possibilities condition for ascription of responsibility.  The assumption that human beings do have some such contra causal freedom is deeply embedded in our traditional legal and religious conceptions of what we are as human beings.  It may well be that in practice we take this metaphysical condition for granted and concern ourselves only with failure of satisfaction of one or more of these other requirements.  But none of these other necessary conditions obviates alternative possibilities as also a requirement for a deontologically adequate account of moral responsibility.  Fischer relies upon the Frankfurt-type cases to undermine the alternative possibilities requirement and here his argument is not persuasive.

            There is, however, a second consideration underlying Fischer's semi-compatibilism.  I would like to conclude with some remarks on this second consideration.  Fischer opens The Metaphysics of Free Will with some comments on our intuitive sense of what is it to be a person.  He rightly points to the importance of being a proper subject for ascriptions of moral responsibility as a part of what we think it is to be a person.  He goes on to consider the possibility that a consortium of scientists were to conclude that the world is deterministic:  "Under these circumstances, I do not believe we would have any inclination to give up the reactive attitudes [pertaining to desert]."[16]  And, Fischer says, this reluctance to abandon the reactive attitudes of ascriptions of moral responsibility is not a mere psychological incapacity, "Rather, I am making a normative point.  I am saying that, upon due reflection, it just does not seem appropriate or plausible to think that we should abandon our view of ourselves as persons, if it turned out that the consortium of scientists were right."[17]

            Now it is certainly a sound maxim in philosophy as well as science that we should be very hesitant to adopt positions that contradict central and well-established parts of our conception of the world.  Such an appeal to conservativism in theory construction can be employed, as Fischer does, to support a compatibilist position regarding the metaphysics of free will.  But in this case the appeal to conservativism is something of a two edged sword.  For, as was pointed out above, the conviction that determinism and moral responsibility are incompatible is itself a historically well-entrenched part of our conception of the world.  Indeed, in the course of the roughly two hundred years within which conceptions of determinism grounded in modern science have been a part of philosophy, the consensus has been that strongly deontological conceptions of responsibility conflict with universal causality.  Historically, the key to compatibilism has been replacement of deontological by consequentialist conceptions of responsibility.  This consequentialism lies at the heart of the compatibilist tradition of Hume, Mill, Schlick and a host of others.  The semi-compatibilism of Fischer and other current compatibilists exists within a radically different philosophical environment.  Because of the revolution in ethics, it cannot appeal to such a consequentialist account of responsibility.  This, of course, is not to say that the new compatibilism is mistaken.  But it does make the appeal to the maxim of conservativism somewhat less convincing.

            For some two hundred years a consensus of sorts has generally prevailed among all sides engaged in the debates about determinism and moral responsibility.  All sides had agreed that deontological notions of desert and retribution were in conflict with deterministic naturalism.  Libertarians, accepting the fundamental validity of deontological categories in ethics, rejected naturalistic metaphysics.  Compatibilists, accepting metaphysical naturalism, embraced a consequentialist framework in ethics.  The last quarter of the twentieth century saw a revival of deontological ethics without a concomitant rejection of naturalism in metaphysics.  The result has been a kind of schizophrenic development in philosophy with ethics going in one direction and metaphysics in another.  To be sure, it may well prove possible to construct a compatibilism that can reconcile these seemingly conflicting tendencies.  It may prove possible to reconstruct compatibilism without recourse to consequentialist accounts of morality.  But the task facing the compatibilist, who must now reconcile naturalistic determinism with a fully deontological notion of desert, is much more difficult that the task facing compatibilists of the last 200 years.  The new compatibilism will have to be radically different from the old.

 

Notes