Maximizing Liberty: Retribution, Responsibility, and the Mentor State

Ronald Bailey of Reason magazine believes that morality premised on retribution still holds in a deterministic universe, and that without it we are headed for an oppressive therapeutic state.  But determinism and the absence of contra-causal free will undercut retribution, and a retributive conception of morality isn't necessary to block therapeutic interventions that unjustly compromise personal liberty.  Instead, liberty is maximized by a state in which morality is understood as consequentialist, not retributive, and which takes both mental health and moral virtue as goals of enlightened social policy. 

Retributivism in need of reasons

Although we didn’t end up talking to one another, Reason’s science editor Ronald Bailey and I attended the recent Our Brains and Us conference at MIT. Bailey has posted two interesting pieces about the conference, one of which, Prozac Justice, addresses the issue of moral responsibility in a largely deterministic universe.[1] He takes the naturalistic, science-based view, which I share, that people aren’t uncaused causers, and that if they were we couldn’t hold them responsible. He quotes William James on why having contra-causal free will can't help ground responsibility: "If a 'free' act be a sheer novelty that comes not from me, the previous me, but ex nihilo, and simply tacks itself on to me, how can I, the previous I, be responsible?".[2] Nevertheless, Bailey believes that in a deterministic universe retributive theories of justice still apply, and that “moral agents deserve punishment (retribution) when they interfere with the protected freedom of others.”

In a world in which all behavior is understood to be fully caused, what justifies retribution? In contrast to other rationales for punishment, retributive theories hold that an offender deeply merits punishment whether or not it serves any benefit to society, the victim, or himself. Retributivists believe that even if punishment entails no consequential goods, such as deterrence, public safety, or rehabilitation, it should still be imposed.[3] The mental and physical health of a morally responsible offender, his willingness to conform to social norms, his skills, talents, and potential for leading a productive life – none of these is of concern to retributivists, for whom the point of punishment is not to reform or rehabilitate, but to inflict suffering and deprivation. In actual practice, the retributive rationale for punishment often requires that such suffering consist in denying offenders much, if any, chance at improving themselves or their lot in life. The proper measure of deprivation entails that prisons shouldn’t be country clubs, so hard time is in order, not rehabilitative amenities. The net effect of exacting retribution, of course, is that inmates often return in worse shape than at the start of their sentences, and the cycle of crime, victimization and punishment continues. As a society, we end up paying dearly for retribution, but such seems the price of moral responsibility. Or is it?

In a deterministic universe, we understand that a criminal’s career is not a matter of an unconditioned personal choice, but fully a function of a complex set of conditions, genetic and environmental, that interact to produce the offender and his proclivities. Had we been in his shoes in all respects, we too would have followed the same path, since there is no freely willing self that could have done otherwise as causality unfolds. There is no kernel of independent moral agency that can intervene in this unfolding – we are not, as philosopher Daniel Dennett puts it, “moral levitators” that rise above circumstances in our choices, including choices to rob, rape, or kill.

Now, the question that naturally arises from such considerations is, why do moral agents in a deterministic universe deserve to suffer for their offenses? Why, if their character, motives, and actions are the result of conditions which they didn’t choose, and over which they couldn’t exert control, should they be deprived of amenities that would make them better persons, and instead be exposed to conditions that will very likely make them worse? Remember, the retributivist can’t answer this question by appealing to the salutary effects of the offender’s suffering, for instance in deterring others or the offender himself from further crimes. For that’s an appeal to consequences, and the essence of retributivism is that consequences are immaterial. What’s material is suffering and deprivation, strictly as just deserts.

So the retributivist still faces the question of why, if an offender’s pain and deprivation serves no consequential benefit, it should be imposed. If we lived in a universe in which moral agents were in some sense causal exceptions to nature, or were in some sense self-caused, then this question poses little difficulty. In such a universe, moral agents originate their behavior in some deep super-natural sense; they contribute something ultimately theirs over and above the determinism of the rest of nature, so they are ultimately responsible for their actions. Given such originative responsibility, it’s intuitively the case that they deeply deserve punishment. After all, were it not for their contra-causal choice, and that alone, no harm would have occurred.

But this is not the universe we live in, Bailey agrees, so this kind of ultimate responsibility and desert don’t apply. So the retributive suffering of the offender must be justified on some other basis, and in a deterministic universe, what is this?

For retributivists who accept that we don’t have contra-causal free will, there is no further justification forthcoming. They can’t appeal to the deeply deserving, freely willing self who could have done otherwise as life unfolds. Nor can they appeal to consequences. They can only stipulate what might be justly called the bare retributive intuition: people simply deserve to suffer, perhaps even die, for their fully caused harmful behavior, and that’s that. In a recent weblog exchange, philosopher John Martin Fischer, for whom freedom is what Bailey has in mind – acting on our own intentions and reasons without being coerced – said:

I believe that the harsh treatment which is part of punishment must “fit” or “match” the crime, and that in virtue of FREELY harming someone else, the criminal DESERVES this sort of treatment. I do not know how to prove or even further explain what seems to me to be a rock-bottom intuition, but of course I would welcome help from anyone!! (original emphasis)

And in conversation with me after his presentation at the Our Brains and Us conference, USC philosopher Michael H. Shapiro agreed that for compatibilists such as himself and Fischer (those for whom moral responsibility is compatible with determinism), there is no further basis for the retributive intuition. And he agreed that this left compatibilist retributivists vulnerable to criticism.

For why should we credit this intuition, and indeed act on it, as we now do? Many, in fact most of us, might feel strongly the intuitive justice of meting out just deserts, but of course that doesn’t justify doing so. Intuitions, without further defense, can’t stand as justifications on their own, at least not in a society in which reasons, not feelings, are thought necessary to gain assent for policy. We might deeply want killers to die for their crimes, or to rot in a hole for life, but wanting, as such, is again not adequate reason to impose such punishment. Fulfilling our wants must find justification outside their mere expression, otherwise the whole edifice of moral discourse is for naught.

It is exactly such discourse that Fischer invites in his call for help, since he well recognizes his position is untenable. He, Bailey, Moore, Shapiro, and other compatibilist retributivists can’t stand on intuition alone but must offer reasons that justify their version of moral agency, one that takes the offender’s suffering as an intrinsic good. We await those reasons.

Meanwhile, an alternative version, also compatible with determinism and lack of contra-causal free will, has more going for it. It says we needn’t suppose that, to be held responsible, offenders deserve to suffer for their sins for no purpose. Rather, it says that in holding moral agents responsible, we impose sanctions and restraints that are necessary to protect society, that reduce the likelihood of their re-offending, that restore victims’ property and dignity to the extent possible, and that deter others. The function of moral responsibility on this picture is consequentialist, in that holding rational, uncoerced agents responsible, we are not required to impose suffering beyond what is actually needed to serve some agreed-upon purpose. And suffering beyond the burden of a stiff fine, victim restitution, community service, or loss of liberty in non-punitive confinement, is not required for most purposes, and indeed might be counterproductive in further damaging and alienating the offender. In any case, moral agents on this view deserve punishment only in the limited practical sense that sanctions actually achieve some benefit.[4] This is to naturalize desert.

It might be argued that one benefit accruing from punishment is to satisfy the retributive desire for the offender’s suffering. But those urging this as a consequential benefit can obviously not appeal to the intrinsic deservedness of suffering to justify that satisfaction. Rather, they have to weigh the satisfactions afforded those who desire retribution against other goods. All things considered, it is difficult to reflectively endorse the desire to see another suffer, especially when it conflicts (as it often does) with other social goods, such as returning an offender back to society in better, not worse, condition. And the morally corrosive effects of indulging retributive appetites - of taking pleasure in another’s pain - certainly counts against such indulgence.

Consequentialist morality, distributed responsibility

Bailey worries that abandoning the retributive principle as the centerpiece of moral responsibility might lead us to a therapeutic state, in which offenders – thought of as sick, not morally culpable – could be held indefinitely on grounds that they aren’t quite cured. This is a legitimate concern, but we need not resort to retributivism to forestall the abuse of a medical approach to mentally ill offenders. Without retribution, moral agents and morality still exist, since we can (and must) distinguish those rational, uncoerced agents that intentionally break the law from agents that did so out of duress or insanity. The former are capable of taking the prospect of punishment into account, the latter not, and this is why we punish only the former. This capacity grounds what it means to be a moral agent on a consequentialist understanding of morality: the prospect of rewards and sanctions helps to guide behavior of those whose choices can be shaped by the rational anticipation of consequences.

This notion of moral agency is all that’s needed to block abuse of therapeutic interventions. If an offender’s behavior was not driven by a diagnosable defect of cognition or impulse control – that is, if he acted as a morally responsible agent – then he’s not a candidate for a cure, but rather for moral instruction. He ain’t broke, so don’t fix him, guide him. If, however, we find that defects played a significant role, then we’re dealing with someone who needs not moral instruction, but help regaining those capacities that would make him a moral agent. And we detain him (non-punitively and curatively, of course) only so long as he needs fixing.[5]

In theory, this is more or less how things stand in our current system, but of course the behavioral and brain sciences (to use the name of a leading journal) increasingly reveal the causal story behind the behavior of both moral agents and the mentally ill. The understanding that both categories of human beings are fully caused to be as they are, and act as they do, means that we can’t any longer characterize moral failures as failures of a causally autonomous free will, but rather the failure of family and social systems to properly guide behavior (and indeed sometimes to actively warp it in pernicious directions). This is exactly analogous to the failure of neural wiring, for instance as damaged by endogenous mental disease or head trauma, to allow individuals to be adequately responsive to social guidance and thus be moral agents. Just as we don’t blame the mentally ill for neural defects beyond their control, we can’t hold moral agents responsible, should they transgress, for the biological and social conditions that made them who they are. That responsibility, on a scientific understanding of ourselves, is distributed among those conditions.

This crucial insight reinforces the consequentialist intuition that people don’t deserve punishment in any deeper respect than what’s necessary to guide behavior, and it’s precisely this intuition that counteracts the retributive impulse. Instead of focusing all our retaliatory energies on the offender, we can (and should) target factors which create morally defective agents: environmental factors such as family, school, and community, and biological factors that might be amenable to control, such as predispositions to conduct disorder.[6] This means that the role of punitive sanctions after the fact of an offense might be cut back considerably, in favor of social policies and interventions that give moral agents the sort of positive inclinations which will keep them out of trouble with the law in the first place. The need for the deterrent threat of punishment will diminish as more moral agents become responsible, law-abiding citizens instead of alienated, anti-social predators.

Liberty and the mentor state

Libertarians like Bailey might object that this scenario adds insult to injury by expanding the therapeutic state, which treats diagnosable illnesses, into what might be called a mentor state, which seeks to treat and prevent moral defects among its citizens. Wouldn’t such a state’s policies and interventions constitute excessive intrusions on people’s liberty to exercise their “power of choice” as Bailey calls it? Not at all. Exercising choice, as Bailey acknowledges, isn’t a matter of being free from conditioning circumstances, since we are always and everywhere fully caused. Instead, it’s a matter of doing what we want, unconstrained by coercion or mental illness. Social policies and interventions which create healthy families and communities, which rehabilitate inmates, and which therefore foster benign moral sensibilities, are not coercive, but formative. They would simply ameliorate the often punitive, morally corrupting, and criminogenic conditions that now hold sway in too many families, communities and prisons.

We now know for sure, and Bailey agrees, that toxic circumstances, not free will, create bad apples, so there’s no excuse not to prevent them so long as individual freedoms aren’t compromised. And there’s no reason to suppose that preventing obvious moral defects (those that lead to abuse and criminality) need involve imposing a monolithic, narrow conception of moral virtue, the scenario sometimes invoked by conservatives and libertarians to inspire fear of liberal do-gooders. No, there’s still plenty of latitude for difference of opinion on what constitutes good behavior, the good life, and the good society, even after our worst impulses are attenuated. And taxpayer expense is no objection, since it’s at least arguable (and in some cases proven) that preventing crime and addiction costs the same or less than the crime, addiction, victimization, adjudication, punishment, rehabilitation, parole supervision and recidivism that would otherwise result.[7]

The question that libertarians must consider is: which state, our current laissez-faire disciplinarian state, or a mentor state, most infringes on freedom of choice, defined as the personal liberty to do as one wishes? It’s no contest. Coercive social control, which intervenes after the fact of misconduct, and depends primarily on retributively justified confinement with little or no rehabilitative amenities, reduces liberty far more than do social policies which encourage citizens to develop proclivities for making good choices in advance of potential misconduct. Such proclivities, not being harmful, need no threat or practice of punishment to channel. So, without any compromise of liberty up front (remember, ameliorative social programs aren’t coercive), we end up with better moral agents, less need for punishment, and thus an increase in liberty overall.

Those retributivists who rather like the imposition of just deserts, and those libertarians who are reflexively ideological in opposing public investment in character development, won’t cotton to this proposal. But having dispensed with the freely willing contra-causal agent, we can’t suppose there’s intrinsic moral worth to suffering, and this blocks retributivism. Further, seeing that we are always fully caused means that creating conditions that foster virtue - conceived as the absence of obvious moral defects - can no longer be characterized as coercive or manipulative. This puts the burden of proof on laissez-fairists to justify their hit or miss system of moral enculturation. Retributivists must offer us good reasons to impose punishment that’s divorced from achieving consequential benefits, and libertarians must say why they prefer after-the-fact coercive control to before-the-fact formation of virtuous citizens. Although we’re far from a therapeutic mentor state in which both mental health and responsible behavior are accepted as worthy goals of enlightened social policies, giving up the myth of contra-causal freedom reveals that such a state is no threat to liberty, but rather its champion.

TWC May, 2005


[1] The other is Minds on Brains.

[2] University of California at Riverside philosopher John Martin Fisher expresses this point nicely: “If causal determinism is false, then it would seem that everything about the past and the laws of nature – including all the facts about me (my character, values, personality, and so forth) – leaves it open which path will be selected. If this is so, it is hard to see how I select the path, how I make the relevant difference.” From “So what’s the problem,” Philosopher’s Magazine, 30, 2nd quarter 2005, p. 50.

[3] Law professor Michael Moore is an archetypical retributivist in this respect, see Against Retribution.

[4] For an excellent paper defending a consequentialist position on punishment see Joshua Greene and Jonathan Cohen, “For the Law, Neuroscience Changes Nothing and Everything,” Philosophical Transactions of the Royal Society of London (Science B), 359 (2004): 1775–1785.

[5] Of course there are many cases of criminal behavior involving both mental illness and well-formed intentions to act which present vexed dilemmas in how to interpolate medicine and criminal justice. Offenders might need fixing and moral guidance, and the line between them may not be clear. But the default humanitarian principle in all cases would be to minimize punitive approaches to both moral guidance and healing mental illness, in whatever combination necessary. Only a very strong, well-documented countervailing social need to impose suffering could override this principle.

[6] See for instance Caspi, et al, Role of genotype in the cycle of violence in maltreated children, Science, V 297, 2002.

[7] See for instance Youth Violence: A Report of the Surgeon General, ch. 5 on cost effectiveness; the National Center on Addiction and Substance Use report “Shoveling up: the impact of substance abuse on state budgets”, p. iii; and the Rand Corporation study “Controlling cocaine: supply vs. demand programs”.