The Author (“A”) raises important issues about criminal justice, science, and traditional conceptions of responsibility, reaching the correct conclusion that “to remain credible and to become more effective, the concepts applied in the law must evolve so as to become grossly consistent with a contemporary scientific understanding of human nature.”

The illusion of conscious will

Using citations from psychology literature, A spends a good deal of time undermining the thesis that conscious control is all that it’s cracked up to be, and suggests instead that actions are largely shaped by processes outside conscious awareness. A recent book by Daniel Wegner at Harvard, The Illusion of Conscious Will, could profitably be cited with regard to this thesis. But this section is in a way unnecessary for A’s larger point (and so could be shortened or dropped), since even if consciousness isn’t overrated and does play a crucial role in behavior (and it seems it does, all things considered), determinism applies to consciousness as well: our rational deliberative processes are in no way uncaused. If this is true, then her main point against the traditional, contra-causal notion of agency and responsibility still holds.; We don’t have to discount the role of consciousness and rationality in the least in order to show that contra-causal free will doesn’t exist and that responsibility isn’t ultimate, and that therefore moral and legal notions that depend on ultimate responsibility are, as she rightly says, unfounded. But further, even though we don’t have free will, the exercise of our capacity for conscious rationality is essential for the control of higher level behavior, and agents having that capacity are those who can be justifiably held accountable using moral frameworks and criminal codes compatible with determinism (see table/discussion below).

Compatibilism

A says that “…there is a clear prima facie conflict between the traditional account of responsibility and the thesis of determinism with respect to human action. If one contends that the two are compatible, some logical strategy for reconciling them must be offered. Compatibilists attempt to accomplish this, but in doing so, they seem to fundamentally compromise the traditional meaning of responsibility.  Indeed, the present author maintains that no satisfactory compatibilist strategy has yet been explicated.”  Clearly, there can be no satisfactory way of reconciling the traditional notion of responsibility – that which depends on contra-causal freedom (free will) – with determinism, or with the further important insight of A’s that “the absence of determinism may present a greater threat to the traditional notion of responsibility than does determinism.  If our thoughts and actions are not the inevitable products of prior conditions, it would seem they must occur by chance, i.e., randomly.”  But there exist varieties of freedom (not being coerced) and agenthood (being rational) that are compatible with determinism and that support a notion of responsibility that can justify some, but (importantly) not all, of our criminal justice practices.   (Such freedom and agenthood may well not support categorically retributive punishment, as A crucially recognizes).  Thus compatibilism is really the project of showing how, absent the incoherent notion of libertarian agency, a sufficiently robust and fair notion of moral responsibility can be salvaged.  Owen Flanagan, in The Problem of the Soul, calls this “neo-compatibilism” (p. 127).   The difficulty with traditional compatibilism, in my view, is that compatibilists fudged or simply failed to address the obvious problem (that A raises) that the sorts of freedom compatible with determinism do not support moral responsibility practices that depend on intuitions about contra-causal agency.  Therefore their claims about having reconciled determinism and responsibility ring very hollow.  The neo-compatibilists recognize that if we take science seriously with regard to human nature, our notions of responsibility, agency, and even the aims of the law should change, as A importantly suggests (see my Science and Freedom for a brief take on this). 

Libertarianism and “could have done otherwise.”

A could usefully cite the recent work of Robert Kane (The Significance of Free Will) and Daniel Dennett (Freedom Evolves), since Dennett carries out a detailed critique of Kane’s advanced brand of libertarianism in  chapter 4 of Freedom Evolves.  In this critique, Dennett points out that there are two conceptions of “could have done otherwise” (CHDO).  In one, the “narrow” conception (CHDOn), we think of the agent acting in the exact same situation as the original, and see that on a deterministic, causal view, the agent could not have done otherwise, since her motives and external determinants would have been precisely the same.  On the “wide” conception of CHDO (CHDOw), we imagine the agent in somewhat different circumstances from the actual situation (that is, in a counterfactual situation), and see that for most agents of normal rationality and flexibility (see Freedom Evolves), the agent may well have done otherwise or at least had the capacity to have done otherwise had they been in the changed scenario.  (Example: the thief, not being a truly compulsive kleptomaniac, may well have not stolen the purse had a policeman been present, and so we properly say he CHDOw.  Whereas a real kleptomaniac would have stolen the purse even with the police officer present, in which case we properly say it’s not the case he CHDOw.)  The significance of CHDOw is that since most people could have done otherwise in a somewhat different situation than what actually transpired, then it makes sense to hold them accountable as a way of creating a different set of conditions the next time a similar situation comes up: they will have a newly installed set of expectations of possible consequences that will serve to shape behavior (“guide goodness,” as Stephen Morse puts it in his article "Reason, results and criminal responsibility," Illinois Law Review).   

The significance of CHDOn is the important point A makes: given determinism (or more broadly, the incoherence of contra-causal, libertarian free will), then backwards-looking, desert-based punishment that serves no consequential purpose – that is, retributive punishment – is ruled out.  So, it’s important for A to carefully distinguish these two senses of CHDO since they are both relevant to untangling our intuitions about responsibility, rationality, and agenthood.

On the traditional notion of responsibility

A could profit by citing Galen (not to be confused with Peter) Strawson’s work on free will and moral responsibility, e.g., as set out for instance in Freedom and Belief or in his entry on moral responsibility in the Routledge Encyclopedia of Philosophy at http://www.rep.routledge.com/article/V014SECT4 (part of a larger entry on free will).  He talks of the impossibility of “ultimate” responsibility, the sort of responsibility that depends on being a self-caused self (causa sui) and that supports ultimate credit and blame.   But there is another, non-ultimate sort of responsibility that we can justly ascribe to rational, uncoerced agents, namely the responsibility that involves holding them responsible as a practical means to shape moral behavior, or “guide goodness” (Dennett talks about this in Elbow Room and Freedom Evolves).  So, there are two sorts of responsibility: ultimate (backwards-looking) vs. practical (instrumental, forward-looking), based on two different conceptions of agency (contra-causal vs. rational/uncoerced), that support two different aims of the law (retributive vs. guiding goodness).  And these paired conceptions are respectively evoked by the two different senses of “could have done otherwise” that agents are thought capable of: CHDOn vs. CHDOw.  Here’s how these might line up:

 A says “man’s ability to reason and make genuine choices creates his moral capacity”.   Using the table above, we can now see that this statement runs together two very different criteria for two different sorts of responsibility (moral capacity).  Reason supports responsibility as practical accountability (being held responsible), while making “genuine choices” (by which A means contra-causal choices) supports the responsibility of ultimate credit and blame.  A seems to want, in this paper, to discount rationality since she somehow thinks it supports the traditional contra-causal conception of moral agency, but this is not the case: rationality supports a useful, valid conception of moral agency that we can and must apply to guide goodness without supposing that retributive justifications for punishment still hold.   Our rational processes probably are what we imagine them to be, that is, they most likely are necessary for executive control over one’s action, so reason and rationality are not illusory, and they support crucial ascriptions of naturalized moral agency.  They just don’t support the sort of moral agency that the traditional notion of contra-causal, supernatural freedom (libertarian free will) supports.  

Similarly, A says “Thus future research may reveal a variety of factors that further encroach upon the space within which human beings can be considered to exercise moral agency.”  But what sort of moral agency does A have in mind?  If it’s traditional moral agency, that involving ultimate responsibility, then this statement is true.  If she means moral agency in terms of being able to hold rational, uncoerced agents accountable for purposes of guiding goodness, it’s not clear that it’s true.  So my recommendation is that A clearly distinguish and label the sorts of responsibility and moral agency she has in mind, so that the reader always knows when she means traditional, contra-causal agency and the related sort of ultimate responsibility, as opposed to the neo-compatibilist sort of agency and responsibility that survives under science and naturalism.   

Similarly A says:  “If conscious thought (e.g., practical reasoning) is determined, it cannot possibly confer executive control to the individual over his or her desires, choices, and behaviors.  That is, it does not enable the individual to do otherwise and thus cannot serve as the foundation for holding individuals responsible.”  The equivocations/conflations in this passage can be disambiguated using the table above, and we can thus see that practical reasoning can serve as a foundation for holding individuals responsible within a neo-compatibilist framework.    

Coercion

I think A is wrong to suppose that the “distinction between behavior that is coerced by external forces and that which is caused by natural, internal forces is ultimately arbitrary and has no logical significance vis-à-vis responsibility”.   Rational agents, when coerced against their will, shouldn’t be held responsible, since they have no proclivity to wrongful action that needs to be countered by sanctions.  On the other hand, if they offend by reasons of innate character and uncoerced desires, then it makes sense to hold them responsible, in the sense of forward-looking, guiding goodness responsibility (see table and discussion above).  So even though all behavior is caused, it makes all the difference what the causes of behavior are.   Even though persons are fully caused, we don’t disappear as identifiable agents, in which case it’s fair to say we often do reach decisions on our own, in contrast to being coerced to decide (p. 43).  This issue comes up again on page 60:  “…as there would be no place for the traditional concepts of blame and responsibility, the current focus on defendants’ intentions and subjective reasons for their actions would largely or completely disappear.”  This seems wrong, since even in a deterministic world, it’s vital to know the reasons and causes behind a crime, since this might well help us develop a productive response to the offender.  

Voluntariness

A states that “once we accept that choice and will are themselves the inevitable products of prior conditions, we can no longer regard these as entities that make us free.  The actual nature of these phenomena simply cannot support a meaningful concept of voluntariness or the traditional notion of responsibility.”  This is wrong, since the concept of a voluntary act has nothing to do with its being undetermined, but rather involves the causal chain behind it, a causal chain which differs systematically from that involved in involuntary behavior.  See, for instance, p. 33 of Wegner’s The Illusion of Conscious Will where he clearly distinguishes the voluntary from involuntary action within a deterministic context. 

Conclusions

I think A is correct that “a very different model of responsibility might be retained as an instrument of social utility”, namely the neo-compatibilist, practical, accountability-based responsibility that leverages rationality as a way to guide goodness.  Therefore A is wrong to say, a sentence or two later, that “People would neither be regarded as responsible or not responsible”.  They would be regarded as responsible or not depending on their capacity for rationality and freedom from coercion. 

A is correct, and importantly so, to state that “The unintelligibility of the traditional, phenomenologically-based concept of responsibility would be recognized, and retribution could no longer serve as a justification for punishment.  The only appropriate considerations behind any legal response to an individual’s behavior would involve the impact of that response on the future behavior and welfare of the offender, the victim, and other members of the society.  This would represent a fundamental paradigm shift.  It is a forward-looking stance, i.e., a utilitarian or consequentialist approach, which stands in stark contrast to the current backward-looking focus of the law.”  It’s just that A needs to realize that she need not, and should not, attack the rationality requirement for moral agency, the criterion that Morse and Moore rightly defend.  What A sees, that Morse and Moore do not, is that dropping contra-causal free will undermines retributive justifications for punishment, while what Morse and Moore see, but A does not, is that rationality is a perfectly legitimate criterion (indeed the only possible criterion) for the sort moral agency that justifies sanctions consistent with the forward-looking aims of the law: deterrence, rehabilitation, incapacitation, restoration, reconciliation, and prevention.  Her paradigm shift in criminal justice is to be welcomed, but it must be argued for clearly, taking into account the distinctions that exist, within a deterministic world, between the rational and irrational, coerced and uncoerced, voluntary and involuntary.  Pragmatism isn’t necessary to replace “the loss of such foundational concepts as responsibility” (p. 59), rather our concept of responsibility will become more pragmatic, an “instrument of social utility” as she says earlier. 

Re justice as problem solving: A correctly says that “…discarding a retributivist model and adopting a purely problem-solving approach should encourage the proliferation of creative responses to illegal behavior that go well beyond current conceptions of containment (i.e., incarceration) and treatment (e.g., drug rehabilitation).  She might want to mention the various “problem-solving courts” initiatives, for instance see Problem-Solving Justice.  

Difficulties for consequentialism.  A might want to anticipate or at least acknowledge deontological objections to her consequentialist position for the law, for instance as raised by the president’s bioethics committee at http://www.bioethics.gov/transcripts/march03/session3.html:

PROF. SANDEL:  It depends what the consequences would be.  There's the famous hypothetical of a heinous crime that's been committed.  The entire town is outraged.  They're about to go on a rampage against the neighborhood from where they think the criminal came.  The sheriff doesn't know who committed the crime, but to prevent this terrible cost and rampage goes and takes the town drunk from jail, announces that he's the criminal and hangs him.

DR. PINKER: Yes.  That would be the utilitarian calculus of preventing the -- and your question is why?

I mean, I assume that the assumption is that we all consider that to be a bad thing to do, and the question is why.  Is that what you're asking?

PROF. SANDEL:  Yeah.

DR. PINKER: It's a good question.  I would concede that it's the weak point of this analysis, and I haven't thought that case through in enough detail to answer you, but I suspect it would be state that as a policy, and that is not just what does the sheriff do on the spur of the moment, but you write it down.  What should sheriffs do in general?  What should law enforcement officers do in general?  Look at the policy and see how much unnecessary suffering does it cause as opposed to the alternatives.  I'm not going to bluff and say that it will come out the same way, but I have a hunch that it would.  But granted it's a valid objection.