Free Will and Capital Punishment

As an intern for the Public Defender, I have spent the last two weeks observing a capital murder trial.  It has been one of the most fascinating experiences of my life.  I sat just a few feet from the defendant.  The trial touched upon the human condition, free will, metaethics, religion, and the dark side of human nature. 

At the beginning of the trial, when I saw photos of the victim senselessly stabbed and lying in a pool of blood, I—who am otherwise quite progressive on issues of punishment—started to feel sympathetic towards the death penalty.  I almost felt glad that the defendant done this evil deed in the state of Virginia.  Free will skeptics, it seems, tend to also have progressive views on capital punishment.  As a consequentialist, however, I don’t recognize a deontological constraint upon executing criminals (nor do I recognize any deontological principle of retribution), although I doubt that such executions maximize the good.  But the Public Defender challenged my new sympathy when he raised the issue of free will during closing arguments and began to trace the history of the defendant’s life.  The Public Defender began a project, not unlike what Gary Watson accomplishes in Responsibility and the Limits of Evil, whereby a character we once regarded as evil becomes tragic.  As Spinoza wrote “I have made a ceaseless effort not to ridicule, not to bewail, not to scorn human actions, but to understand them.”  My sympathies realigned themselves towards their previous and more natural state.

The Public Defender urged that the defendant was “on a trajectory” since childhood to commit crime.  The murder, so the argument went, was fated and inevitable.  The prosecutor argued back, just as vigorously, that the defendant hadn’t been “on a trajectory.”  On the contrary, he made “choice after choice after choice.”  Clearly, the prosecutor regarded a “choice” and a “trajectory” as incompatible—and expected the jury to share these incompatibilist intuitions.  But, of course, if the amount of indeterminism in human nature and our local environment is negligible—and the prosecutor had no way of verifying that it isn’t nor did he feel the need to raise the issue—then the defendant had been “on a trajectory.”

Both sides cited the same chilling evidence in support of their arguments: twenty-eight years ago a psychologist gave the defendant—then just a thirteen year old boy—an evaluation and called his behavior “sadistic.”  The defendant had shot a little girl.  It was so long ago that she hadn’t yet received her doctorate and now had no conscious recollection of the event.  Twenty-eight years before he stabbed this woman to death, the psychologist had noted that the boy had a fascination with knives and “machetes.”  For the defense, this just proved its point about the defendant being a victim of fate.  For the prosecution, this was proof that he deserved to die.

It is truly remarkable how the same piece of evidence can be used to reach such opposite conclusions.  Yes, the defense was focusing upon the boy’s age while the prosecution was focusing on his “evil” nature.  But a better explanation, I think, is that the two sides had two different conceptions of moral responsibility: responsibility as social policy and responsibility as self-creation.  The fact that something evil isn’t responsible for being evil is just irrelevant to questions of responsibility as social policy.  Criminals don’t get a “get out of jail free” card because they aren’t causa-sui.  But the fact that something evil isn’t responsible for being evil is relevant to questions of responsibility as self-creation.  The prosecution seemed to regard moral responsibility as the former while the defense seemed to regard moral responsibility as the latter.

Clarence Darrow seemed to refer to responsibility as self-creation when he used this exact same argument (but much more explicitly) to defend Leopold and Loeb in their capital murder trial in 1924.  Darrow’s argument struck me as such a novelty, and so unpopular, that I was surprised to see it reappear in this capital murder trial.  But after all these decades, the same question still haunts us: how can we be morally responsible for our choices if we are all on a “trajectory”?

Darrow, however, presented this argument in its purest form.  The Public Defender argued that the defendant was mentally handicapped and had a cruel and impoverished childhood.  Similarly, Gary Watson argues in Responsibility and the Limits of Evil that the abuse Robert Alton Harris experienced as a child mitigates the contempt we feel for him now.  But Watson’s argument (if I remember correctly) might find support from at least three intuitions, and it is not clear which one of them is doing the work: Harris was abused, this abuse corrupted or damaged his mind and moral capacity, and Harris didn’t have free will in some sense—he was on a trajectory.  The first intuition pumps (to turn Dennett’s phrase into a verb) sympathy, the second intuitions pumps the intuitions behind the insanity defense, but the third intuition is philosophical. 

It is the only the third intuition that I think is relevant to the free will problem.  And it is the only intuition that Darrow had to work with because both Leopold and Loeb were privileged as children.  They were both intelligent enough to attend Ivy League universities.  Their crime seems all the more senseless and perverse, and Darrow’s argument all the more bold, because they weren’t abused as children and they had perfectly healthy minds—albeit, perhaps, unhealthy appetites.  Darrow’s case tests our belief in free will in a way that Watson’s example and the capital murder trial I observed do not.  Moral responsibility, on compatibilist (or semi-compatibilist) accounts, such as Fischer and Ravizza’s, necessitates having a mind that is not defective and responds to reasons.  So in cases where this mechanism is damaged or defective, the question of the sufficiency of such accounts never arises.  But in the Leopold and Loeb trial, however, this question did arise:

“No one knows what will be the fate of the child he gets or the child she bears; the fate of the child is the last thing they consider. This weary old world goes on, begetting, with birth and with living and with death; and all of it is blind from the beginning to the end. I do not know what it was that made these boys do this mad act, but I do know there is a reason for it. I know they did not beget themselves. I know that any one of an infinite number of causes reaching back to the beginning might be working out in these boys' minds, whom you are asked to hang in malice and in hatred and injustice, because someone in the past has sinned against them.”

Here Darrow, concerned with responsibility as self-creation (“they did not beget themselves”) seems to suppose a sort of Transfer Principle.  But Darrow doesn’t follow the principle to the conclusion that Galen Strawson reaches: according to Strawson, even the “someone in the past” cannot be morally responsible for his or her actions.  The meta-controller, in this case, is just as much a victim of fate as Leopold and Loeb were.  The only thing that might be morally responsible, in this sense, is whatever chooses the rules and initial conditions of the universe.  But such a choice presupposes that a chooser, and hence initial conditions, already exists.  So it is impossible for anyone to choose in this sense.  Yet we seem to have a natural intuition that our ordinary choices are somehow like such self-negating choices between metaphysically distinct universes.  Let’s call these choices*.  This helps to explain the prosecutor’s remark that “he wasn’t on a trajectory, he made choice after choice after choice.”  This also seems to be the intuition that libertarians have tried to capture by suggesting that we are “unmoved movers.” 

The previous post “Compatibilism and Modality” distinguishes between epistemically compatible choices and metaphysically compatible choices*.  I suspect that we naturally tend to mistake choices for choices* because we think that if we can contemplate an alternative universe then that universe must be open to us.  But when we use Spinoza’s and Nagel’s strategy of reflecting upon our fixed nature and environment and the fact that these will combine to produce just one future—even if we won’t yet know which future that is—we realize that the alternative universes we contemplate hang before us like fruit just out of reach.

This is the intuition which the Public Defender was trying to pump in the jury.  In his closing remarks, he reminded me—not the most sympathetic observer of religious behavior—of the good in Christianity.  The more zealous members of the religious right forget that Jesus defended prostitutes:

“Then he turned toward the woman and said to Simon, ‘Do you see this woman? I came into your house. You did not give me any water for my feet, but she wet my feet with her tears and wiped them with her hair. You did not give me a kiss, but this woman, from the time I entered, has not stopped kissing my feet. You did not put oil on my head, but she has poured perfume on my feet. Therefore, I tell you, her many sins have been forgiven—for she loved much. But he who has been forgiven little loves little.’”
Luke 7:44-47

At the beginning of the trial, I wanted this man to die.  I was confident that he had committed the crime—he was only caught because the victim fought back, leaving blood under her fingernails for the authorities to DNA test.  This was a man around whom I didn’t even feel safe when sitting several feet away.  I played scenarios in my mind of him jumping over the wooden separation between us and asked myself whether the guard would have enough time to tazer him before he reached me.  That is why I felt so struck by the image of the Public Defender, while addressing the judge, placing his hands on the defendant’s shoulders—not unlike how Jesus humbled himself by touching a prostitute.  The Public Defender told the story of a boy who lived in the ghetto and dropped out of school at age 15.  He told the story of a boy whose father disowned him.  This demon—something that commits evil in a vacuum and for no reason—became a human being.  With his eloquence, the Public Defender, like Clarence Darrow before him, saved his client from Death Row.  But even then the defendant never apologized.  He never admitted to committing the murder.  Although he had been greatly forgiven, he did not greatly love.  And so I am left with a greater dilemma than Simon faced.

Gardeners:
1. Which of the three possible intuitions do you think are doing work in Watson’s Responsibility and the Limits of Evil, and which do you think are relevant?
2. Do we confuse choices with choices* and, if so, is this confusion relevant to questions of moral responsibility (semi-compatibilists seem to say yes and no)?
3. How do you respond to those who are greatly forgiven but do not greatly love?

Comments

[Neil]

Nice post, Kip. But I don't agree that the conflict you describe is necessarily a conflict between incompatibilists (a hard determinist on one side, and a libertarian on the other). First, it seems that compatibilists and incompatibilists alike need to be able to give an account of someone being "on a trajectory". Even most libertarians think there are factors which "incline without necessitating" (and they may think that there are psychological dispositions which necessitate). Second, I think we can make sense of the conflict by ascribing versions of compatibilism to both sides. The prosecution might be adopting a version of attributionism (see Angela Smith's recent article in Ethics, and Scanlon's work on responsibility, for a defense). Roughly, an attributionist claims an agent is responsible for an action if it reflects their character. The defence might be invoking a normative competence view, like Susan Wolf's. Or they could be emphasising different aspects of Watson's 'Two Faces of Responsibility' (attributionism vs avoidability).

[Kip Werking]

I’m not yet familiar with Smith or Scanlon’s work, but I agree that there is an ambiguity between the notions of “being on a trajectory” here. This, I think, is related to the ambiguity I describe between the mitigating factors of sympathy, mental illness, and a lack of metaphysical freedom (or lack of self-creation). For example, if the defense was just relying upon sympathy and (tacitly) mental illness, then the Public Defender might very well have been defending compatibilism. But it is not clear that he was limiting himself to these intuitions.

Compatibilists can provide accounts of people not being “on a trajectory” even within a deterministic universe, just as Dennett stresses how “evitability” can exist within a deterministic universe. The problem, of course, is that such accounts do not seem to fully capture our intuitions about what being “on a trajectory” means. They seem, as Saul Smilanksy might say, “superficial.” There is an *obvious* sense in which everything in a deterministic universe is on a trajectory. But compatiblists are right to charge that this sense is not, and should not be, the sense relevant to moral responsibility. So the question becomes, when the prosecutor denied that the defendant was on a trajectory, was he just inconsistent (because he does share incompatibilist intuitions) or was he defending one of these compatibilist accounts of being on a trajectory, which strike many, including myself, as somewhat superficial, incomplete, and revisionary.

I also agree that we need to distinguish between attributionism and other notions of moral responsibility. I tried to do that when I distinguished between responsibility as social policy (attributionism) and responsibility as self-creation, but this account is probably inadequate. Attributionism seems consistent with hard compatibilism, and Watson’s other notion (avoidability) seems, as Derk Pereboom says to perhaps have “had a more explicit role in debates about free will.”

If I may digress:

The problem with attributionism, and hard compatibilism in general, is that it doesn’t seem to do full justice to what we mean when we hold people morally responsibility. And it fails to do this because it doesn’t put enough importance upon how agents acquire the attributes that they have. When we hold someone morally responsibility because they have some attribute, such as evilness, we seem to make a further claim A, which I might sketch as “you freely acquired your evilness and you, alone, are to be held accountable for being evil.” If hard compatibilists wish to keep this claim—which seems to be an essential aspect of how we practice moral responsibility—then they are just inconsistent. Evil people do not ultimately make themselves evil. Other things, such as their environment and heredity, can be held accountable. This leaves hard compatibilists another option: relinquishing this claim A. The problem here, however, is that this notion of moral responsibility seems quite foreign to our traditional notions of moral responsibility. Furthermore, this revised moral responsibility, which we might call MR-A (moral responsibility minus A), seems to be converging upon our notion of sickness or illness. If people freely acquire the ebola virus, we hold them responsible for doing so. But if they are infected through no fault of their own, we do not hold them responsible. The question is: why does this principle apply to things like the ebola virus but not to things like whatever it is about criminals’ brains that makes them criminal?

Hard compatibilists appeal, at this point, to the notion of a healthy mind. The idea is that if this mechanism is still healthy, then if it responds imperfectly to certain inputs, then we can still hold people MR. So the health of their “screening device” prevents them from being excused, even if unhealthy influences try to infect them. Consider this analogy:

A person comes to a software designer and says “I want a program that will usually give the right answer to the inputs I give it. But I want it to fail sometimes. And when it fails, I want to give it blame or praise signals to correct itself.” The software designer makes the mechanism and the person uses it. He gives it a large sequence of inputs and the program gives the right answer most of the time.

But there is an ambiguity here. If we consider the perspective of the software designer, the program performs flawlessly. The right answers and the small number of wrong answers are all, because they were specified, the right answers. It performs exactly according to specifications. From the perspective of the other person, who wanted a “faulty” mechanism, however, the program still gets a few answers “wrong.”

The point of this analogy is: how fair is it to hold the program accountable for how it performs? In one sense, we hold it accountable because it has an (imperfect) accountability function built into it. The program is designed so that when it misbehaves, we can send praise and blame signals. But, as Smilansky would say, this strikes us as a shallow sense of accountability. The program might very well say:

“I am a mechanical object, and I perform exactly according to specifications. It’s not fair to say that the answers I gave, and which you specified, are both right and wrong, and deserving of punishment. When you hold me accountable for my behavior, you seem to also hold me accountable for this faulty mechanism of mine which imperfectly computes your inputs. It works most of the time, but it fails sometimes, and this is what the designer asked for. It’s not fair to hold me accountable, and further hold me accountable for how this faulty mechanism performs, because I did not choose, create, or have any part in the selection of this mechanism. But if you say you hold me accountable for my behavior, but you do not hold me accountable for having the mechanism I do, then this seems to be unfair too. The behavior might be attributed to the mechanism, but can the mechanism be attributed to me?”

The program might continue with an even more impressive response:

“Most importantly of all, the inconsistency in your position will become apparent when you have the capability of correcting my behavior without sending me blame signals. Similarly, the police may one day intervene before crime occurs or give criminals ‘moral pills’ to correct their behavior. Instead of using this archaic and faulty ‘accountability’ mechanism which I have inherited, if you could just rewrite my code, then what reason would there be to hold me accountable for my behavior then, or ever? At that point, won’t it just seem as if I had been sick, and not evil?”

And so, as I hope this illustration helps to show, hard compatibilism has a revisionary character and doesn’t quite capture our intuitions about moral responsibility. In particular, our natural intuitions about moral responsibility involve a notion of self-creation and exclusive-blame (blame which cannot be attributed to both an agent and its designer) which hard compatibilism fails to capture.

The only sense, I think, in which it might be fair to hold a person accountable for the behavior that results from their mechanism is if the mechanism is perfect, and not just healthy. If you give me a car that works most of the time, but breaks down sometimes, then when it breaks down, I have the defense “I didn’t make or choose this car.” But if the car works perfectly, then I have no excuse. Similarly, if the “screening device” works flawlessly to screen out bad influences, then the agent truly wouldn’t have an excuse for bad behavior. But, of course, in that case, there would be no bad behavior! So there would be no need for an excuse. I’m not sure, however, how much I agree with even this response.

Finally: I forgot to mention some excellent resources for those interested in behavior therapy and progressive attitudes towards the intersection of philosophy and law. First, there is Slobogin’s fascinating “The Civilization of the Criminal Law” which suggests how the legal system could be changed to accommodate the implications of free will denial:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=600674

Then there is Ronald Bailey’s series of articles on punishment and behavior modification, from a libertarian and skeptical perspective:

http://www.reason.com/rb/rb060105.shtml -- Should We Cure Bad Behavior?
http://www.reason.com/rb/rb051805.shtml -- A Day at the Brain Spa
http://www.reason.com/rb/rb042005.shtml -- Prozac Justice
http://www.reason.com/rb/rb032305.shtml -- Bad Brains

Tom Clark has a fascinating critique of Bailey’s position here:

http://www.naturalism.org/maximizing_liberty.htm#_ftn3 -- Maximizing Liberty

Finally, for a brilliant exploration of the limits of hard compatibilism, and more specifically the limits of Fischer and Ravizza’s account, see:

Moderate Reasons-responsiveness, Moral Responsibility, and Manipulation by Todd R. Long in the recent MIT Freedom and Determinism

[Rob]

"...the defendant never apologized. He never admitted to committing the murder. Although he had been greatly forgiven, he did not greatly love. And so I am left with a greater dilemma than Simon faced."

But is forgiveness really possible without some form of acknowledgement on the part of its recipient that (1) looking back, he or she has indeed done something punishment-worthy and that (2) looking ahead, this acknowledgement will appropriately manifest itself in his or her future behavior? Doesn't this defendant rather stand, at least for the time being, outside the process of re-integration that is central to the operation of forgiveness?

[Kip Werking]

Rob, that is what the prosecutor argued.

At the risk of abusing language, however, I must say that I am not yet convinced that acknowledgement of error is necessary for forgiveness. We think of forgiveness as an achievement that overcomes an obstacle, such as rage or anger or resentment. So forgiving those who, like this defendant, do not acknowledge their error, refuse to apologize, and do not even want forgiveness seems to be especially pure, because such forgiveness is more difficult. Smilansky explores a related idea in "The Ethical Advantages of Hard Determinism." Certainly, if we hold being-causa-sui to be the free relevant condition for moral responsibility (as I tend to), then we cannot hold this defendant blameworthy, and so there is nothing to forgive.

[Paul Torek]

Do we confuse epistemically compatible choices with metaphysically compatible choices*? Before we answer this question, let's improve it. Remember the definition Kip gives: "statements S1 and S2 are epistemically compatible for a person p at time t only if nothing in p’s epistemic body of evidence at t rules out its being the case that both S1 and S2 are true."

Speaking of epistemic compatibility seriously understates the epistemic dimension of choice. We don't just have epistemic compatibility with various choices, we have what David Velleman calls epistemic freedom. Epistemic freedom is more than the mere ignorance which would usually be associated with Kip's definition. Rather, it is a license to believe any of various contrary propositions about one's future without fear of being wrong. As Velleman writes in http://www-personal.umich.edu/~velleman/Practical_Reflection/chapter5.html

"For there are two senses in which the future might be undetermined. On the one hand, there may be no particular way that the future is going to turn out -- or, at least, no way that's necessitated, under the laws of nature, by the present state of the world. In that case, the future would be metaphysically or causally open. On the other hand, there may be no particular way that we must describe the future as turning out, in order to describe it correctly -- or, at least, no way that's necessitated, under the laws of nature, by a correct description of the present state of the world. In that case, the future would be, as I put it, epistemically open. [...] For even if there is already something that we're going to do next, we would be correct in expecting ourselves to do something else, insofar as our alternative expectation would amount to an effective intention to perform the alternative action."

Now then, to the improved question: do we confuse epistemically free choices with metaphysically open choices*? Heck yeah! Just look at the extreme care which Velleman must use to convince the reader that the two are actually distinct.

Given that the two are easily confused, which is the mistake? Do we wrongly conclude that we do have free will, because we illegitimately infer free will from epistemic freedom? Or is the mistake to doubt that we have free will, because we worry that the future is not metaphysically open and illegitimately infer that we cannot then really be entitled to expect whichever of our future actions we choose? I strongly suspect the latter.

I like to compare our situation to the famous "moon illusion". The NASA website http://science.nasa.gov/headlines/y2002/24jun_moonillusion.htm describes it thus: "The moon illusion is a well-known trick of the eye: a low-hanging moon looks unusually big."

When I was living in Dearborn, Michigan, a city of many straight, flat, tree-lined streets, I had the privelege of experiencing the "moon illusion" in full force. From my perspective looking down the street, the moon looked to be the size of a continent.

Guess what? The moon IS the size of a continent.

The NASA website gets it backwards, insinuating that it is the on-horizon perception which misleads. Wrong. It is the cold, detached perspective on the moon, high in the sky with no perspectival clues, which yields an illusion. Similarly, I suggest, it is the cold, detached perspective on ourselves - mistakenly thinking that such a perspective gives us proper appreciation of the implications of the causal order - which misleads us about free will. Some parts of the causal order are best appreciated in personal context.

[Kip Werking]

Paul, that was another interesting comment.

You sound like Thomas Nagel. :)

[mv]

Kip-

This was a nice post and an interesting thread. I'm sorry I missed the conversation a few weeks ago when this was an active thread.