Moral Responsibility and Mental Illness

Suppose you discovered that someone has committed a horribly violent crime. And now suppose I tell you one additional fact about the person who performed this act: he or she is mentally ill. In fact, suppose I tell you that the reason he performed this act he is suffering from damage to a particular area of his brain. Would you still conclude that he could be morally responsible for what he had done?

At this point, you might be guessing that no one would hold an agent morally responsible in such a circumstance. After all, how could we hold someone morally responsible for behavior that was clearly the result of neurological illness? Surely, anyone would agree in such a case that the agent is not to blame for what he has done!

Guess again.  A new paper from the philosophers Eric Mandelbaum, David Ripley and Felipe De Brigard shows that people actually are willing to ascribe moral responsibility in cases like that one. In their study, subjects were randomly assigned to one of two conditions. Subjects in the 'abstract' condition received the following story:

Dennis has recently found out from his doctor that he has a neurological condition that has caused him to behave in certain ways. Were someone else to have this neurological condition then that person would have had to behave in the same ways as Dennis.

Just as you might expect, most subjects who received this story said that Dennis was not morally responsible for the behaviors he performs. But don't be too swift to assume that people with neurological conditions will get off the hook. Mandelbaum and colleagues also included a 'concrete' condition, in which subjects were told:

Dennis has recently found out from his doctor that he has a neurological condition that has, in the past, caused him to rape women. Were someone else to have this neurological condition then that person would have had to behave in the same ways as Dennis.

When the story is made more concrete in this way, people's intuitions change radically. They end up concluding that Dennis actually is morally responsible for what he'd done. 

It seems that, no matter how much we tell people about damage to an agent's brain, the impulse to blame will get the last word. It is as though people are thinking: 'Well, he does have a neurological condition... but then again, someone ended up getting raped.  We just can't let this go by without declaring at least one person to be morally responsible!'

Comments

[V. Alan White]

This is the best sort of contribution to FW research that X-philosophy makes: examining how specific circumstances tweak people's intuitions and responses about responsibility. Perhaps this kind of research will help us understand the evolution of US laws on insanity. Prior to Hinckley, such laws were based on tradition and case precedent (McNaughten, "irresistable impluse", etc.) as well as, in the 20th century, psychological research--collectively the foundation for the "Model Code", which was the law in the Hinckley trial but rescinded since because of public outcry. But also since Hinckley, the single biggest influence on state laws has been (in my humble opinion) the media publicizing such cases and the subsequent knee-jerk reaction of legislators. In evidence, I offer the fact that Montana, Idaho, Utah, and most recently Kansas have abolished the insanity defense entirely--and I'd also argue that such moves step back from the centrality of mens rea for responsibility to some other (more classic compatibilist? more FW nihilist?) position. X-philosophy may well help us understand why this is happening at least at a coarse-grained level, and also show us why laws have moved from being reflectively based on empirical data from top-down academic sources to being based on (unpredictably influenced) reactive attitudes of the masses bottom-up.

[tnadelhoffer]

The most important shortcoming of these otherwise interesting studies--one that I recently pointed out at the Central APA and one that has already been mentioned by all of the commentators on the thread about this paper over at the x-phi blog--is that the participants are asked merely to make judgments about "responsibility" rather than either "moral responsibility" or "blameworthiness." This makes it difficult to know which intuitions are getting pumped in the various cases. Are some participants thinking merely in terms of causal responsibility? Are participants' likely to focus on one kind of responsibility rather than another depending on the context and circumstances described in the scenarios? Given the experimental design, we simply don't know.

But even if we assume for the sake of argument that the participants were thinking in terms of moral responsibility, it is still an open question whether they are thinking in consequentialist forward-looking terms or backward-looking retributivistic terms. Indeed, this is my biggest gripe with nearly all of the studies on free will and moral responsibility that have been run so far. To my knowledge, the only researchers to try to disambiguate this for the participants are Nahmias, Coates, and Kvaran--whose paper (at least in my own humble estimation) is the benchmark for how these kinds of studies should be run!

That being said, there are several other worries that I have with the way the vignettes are worded: (a) the agents in most of the cases could have done otherwise, (b) the agents in most of the cases knew in advance that they were suffering from the disorder--thereby making room for tracing-related attributions of moral responsibility, (c) not all of the cases are really deterministic, and (d) the disorders are not described in sufficient detail. Given these problems, I think there is much more work to be done before we can reach any firm conclusions concerning folk intuitions concerning the relationship between blame and mental illness. I, for one, look forward to the results of the follow-up studies Eric and Filipe are sure to run. Of course, that's what experimental philosophers always say :)

[Blake]

The assertion that Utah has abolished the defense of mental illness is incorrect. See 76-5-205.5 and 305 Utah Code Ann. The mental illness must be such that the actor is acting under a delusion. Further, mental illness is still regarded as negating the mens rea necessary to form the intent for the crimes that require mens rea as an element.

I also strongly suspect from my work as a trial attorney that if the mental illness is described and diagnosed as leading to delusion so that the defendant does not appreciate the fact of what he or she is doing, juries will find the defendant no guilty because they lack the necessary mens rea. Further, the scope and degree of the delusion is relevant to the degree of diminished capacity that the jury may also consider. I think that the cases presented in the paper are simply misleading because they are so incomplete that readers fill in a lot of blanks and make assumptions about the mental state of the perp. Giving further details drastically alters the jury's assessment of guilt in my experience.

[Brian Boeninger]

The authors claim that their results (especially concerning Dennis, the character whose behavior is described as having been caused by a psychological or neurological condition) "suggest that our participants believe that people are still responsible even if a standing neurological or psychological condition causes their action." The authors offer as a possible explanation for this prima facie puzzling result the suggestion that "folk judgments of responsibility are strongly affected by the specificity and affective nature of the vignettes presented." This explanation, in turn, finds support in the work of Nichols and Knobe, as well as the authors' own study on the differential responsibility judgments participants reported for "generic" vs. "specific" descriptions of actions.

I share the worries about these studies raised by Nadelhoffer, and would like to add one more. Here is an alternative explanation for the results mentioned above. Participants rate Concrete-Act-Dennis more responsible than Abstract-Act-Dennis not because of an "impulse to blame" which they find themselves unable to shake despite recognizing the fact that poor Dennis found himself at the causal mercy of his "condition." Rather, participants don't believe that Dennis's specific action really was caused by the neurological/psychological condition. Sure, they understand the scenario; they just don't think it's plausible, and so they hold Dennis responsible. Perhaps participants (consciously or otherwise) reason as follows: "a condition that makes Dennis specifically rape women?? I don't know of anything like that - sounds ridiculous! If Dennis tried to excuse himself with that explanation, I wouldn't buy it." Perhaps participants don't (full-blown) disbelieve the scenario; they just have their doubts. And that would explain why their responsibility ratings aren't zero, but neither are they as high as in cases in which someone like Dennis suffers from no disorder. So, pace the authors, it might be that participants do not "believe that people are still responsible even if a standing neurological or psychological condition causes their action," because they don't really believe that those conditions cause the actions (or they don't "apply" that belief properly when they answer the study's questions).

Now of course as philosophers, we're perfectly used to employing thought experiments and trying, as best we are able, to maintain fidelity to the premises of the experiment, no matter how distant the possible world described happens to be. But most people aren't used to doing this. It is entirely typical, for instance, to witness one's undergraduate students not quite understand how to "do" a thought experiment. We all get frustrated when, to our expertly crafted scenarios, a student replies, "Well that would never happen, so...".

I would love to see future experiments/studies address this possible explanation (what we might call the "Nullification Response"). It might not be too much trouble: one could ask participants to answer separate questions regarding how plausible they took the imagined scenario to be. If there is a strong (negative) correlation between a participant's score on plausibility ratings and her score on responsibility ratings, this might be reason to suspect the "Nullification Response" or something like it. (Another way to describe the Nullification Response: participants are "context shifting." They might judge Dennis's action to be caused by the condition, but then answer the question about responsibility without really applying this judgment.)

[Anibal]

How these findings about the folk-knowledge of people regarding mental illness and its link to moral reponsibility (how a damage or lesion in a brian structure mediating the control of behaviour) could affect the legal liability of someone in the court, taking into account the american jury sistem (a jury is ussualy formed by ordinary people with folk knowledge either about the law and, of course, neuroscience)

[V. Alan White]

Blake (if I may)--you are right in the respect that in all four states I cited I believe SOME defense may be mounted on the basis of mental disease/defect to ameliorate guilt, but still all four states have abolished insanity as a separate defense. There can be no bifurcated trials, for example, where exculpation may be attained as in the Hinckley case. Further, I believe that the Supreme Court hasn't weighed in on whether abolishing insanity as a separate defense is constitutional.

Readers interested in a quick overview of the mish-mash of laws concerning insanity should check out:

http://www.pbs.org/wgbh/pages/frontline/shows/crime/trial/states.html

The contents are somewhat dated (e.g., Kansas' abolition of the insanity defense in 2005 isn't listed), but still it shows the remarkable diversity of how such cases are dealt with.

[Tamler Sommers]

I want to repeat what I said on the Experimental Philosophy Blog:

Monteresso et al

(http://www.informaworld.com/smpp/content~content=a784402380~db=all~order=page)

find that people are MUCH more likely to exonerate the agent when the explanation is physiological (excess of some hormone in the brain) rather than experiential (poor upbringing). In fact, they find that this variable matters more than the uniformity variable. In other words, even if only 25% of people who have the physiological condition perform the violent act, they are held more blameworthy than someone who has the kind of the upbringing that leads 90% of the people who have it to perform the violent act.

We actually discussed this paper a few years ago here:

http://gfp.typepad.com/the_garden_of_forking_pat/2005/04/explaining_away.html

Again, my question for the authors of the new paper is: what do they think accounts for the significant difference in results? One explanation is their use of 'responsible' instead of 'blameworthy,' and also that they don't link the crime to an appropriate punishment. Another, as many of have noted, is that the authors do not give any sense of how often the impairment actually causes someone to perform the immoral action. So it's not clear to what extent the subjects believe that the agents had the (compatibilst) ability to refrain from performing the act.

[Dave Ripley]

- Re: Boeninger's potential explanation of the results, according to which participants didn't fully buy the rapist scenario.

It seems to me, at least, that this is totally possible as a story about what happened here. In fact, we received some handwritten responses from participants that lend some support to this explanation. (We solicited but didn't require written explanations.) For example, in the concrete condition, we received (among others) the following responses: "It seems so impossible that he would have no choice", "Disease or no, he had to take conscious action". These do seem to be resisting the scenario in some way. No parallel responses came up in the abstract condition. I'm hesitant to draw strong conclusions from these responses, but I do think they provide some indication that the "Nullification Response" may be at work.

A key question arises from this explanation, though: why do participants want to nullify the scenario in the concrete case but not in the abstract case? One explanation is immediately available: that they're not familiar with any rape-causing condition, and such a condition seems implausible. This explanation would predict similar nullification, and so presumably similar responsibility judgments, in any concrete case that featured a similarly implausible or unfamiliar consequence of mental illness.

There's another possible explanation, though, one that may shed light on insanity-defense laws and the like: perhaps participants wanted to believe that Dennis is responsible for the rapes he committed, but---as the abstract case indicates---don't believe that people are responsible for actions caused by their illnesses. If one wanted to hold onto these two beliefs firmly enough, there could be only one conclusion: the rapes weren't caused by Dennis's illness.

If this is what caused the nullification, we'd expect to see similar nullification, and so similar responsibility judgments, whenever the action in question is one that makes participants want to believe that Dennis is responsible.

Given some ideas about what consequences of mental illness are unlikely or implausible, and some ideas about which actions participants might want to hold Dennis responsible for, we could try to get some evidence to decide between these explanations.

- But suppose the latter explanation is right, just for now. This would jibe with Blake's prediction about jury behavior, for one.

After all, if someone were able to convince a jury that Dennis's rapes were in fact caused by his illness, then members of the jury would feel pressure to give up one of the other two beliefs in question: i) that Dennis is responsible, or ii) that people aren't responsible for actions caused by their illnesses. If ii) is firmly ingrained enough, then i) presumably will go.

- This style of explanation also seems relevant to the points made by White above. Perhaps the removal of insanity defenses, where it happens, isn't a move away from requiring mens rea at all, but is instead this sort of inference before-the-fact: if someone does X they're responsible, if they're responsible it wasn't caused by their illness, so X wasn't caused by their illness, and no insanity defense should be allowed.

It might be an increase in belief in "if someone does X they're responsible", which could well be fed by media attention and reinforcement of kneejerk responses. But this explanation turns on people holding to mens rea (or something like it) as a condition for responsibility, rather than moving away from it.

- Re: Nadelhoffer's comments

In the study mentioned in the body of this post, we did ask participants whether Dennis is "morally responsible". In addition, in other versions of the studies cited in the paper, we asked "morally responsible" rather than "responsible" simpliciter, and didn't get differing responses. (We haven't run any studies to directly compare the effects of asking "responsible" versus "morally responsible"; it would be interesting to see what happens there.)

However, I'm not sure that even asking "morally responsible" guarantees that participants are thinking in terms of moral responsibility; it seems entirely possible that they're not, or not clearly, distinguishing varieties of responsibility in the first place, even when asked a specific question. This is especially true of experiments, like ours, that make no effort to elucidate what's meant by "morally responsible"; if we want to get clearer on what sort of responsibility participants are thinking about here, you're surely right that we should be clearer in our question.

Incidentally, I don't think that asking "blameworthy" helps; as far as I can tell, this is a word likely to be unfamiliar to participants. (A Google search for it seems suspiciously weighted towards dictionary definitions and experimental philosophy papers.) They may well be able to put "blame" and "worthy" together and figure it out, but just requiring them to do this could affect their responses. (In particular, I'd expect it to make them less confident, and so (?) give less extreme responses.) Asking whether the agent "deserves blame", as some have done, would be a better alternative, but presumably giving a fuller explanation of what we're after would be better still.

Given all this, we should certainly consider the possibility that participants (or at least some of them) are responding in terms of causal responsibility instead of moral responsibility. If participants in the experiments given in this paper are responding in terms of causal responsibility, that might well provide an alternate explanation for their medium-strength judgments of responsibility. They'd simply have to judge that the agents in question were causally responsible, at least somewhat.

It wouldn't, however, explain why that judgment seems to go away when the question is presented in the abstract. The abstract/concrete asymmetry mentioned above would still need an explanation, and the same sort of impulse to blame explanation still seems like a good option, even if we assume that participants are responding in terms of causal responsibility. All we'd need to suppose in addition is that participants believe that an agent can't be morally responsible unless they're causally responsible; then their desire to hold Dennis morally responsible would lead them to hold him causally responsible as well. Such a supposition isn't totally out of left field; Nahmias's "bypassing" explanations seem to use a principle something like this, as does Alicke's "culpable control" model of blame.

Another place where I think you're right: it would help to be clearer on the difference between forward-looking and backward-looking responsibility. Among our handwritten responses, we did get a few that seemed to clearly indicate forward-looking responses; I don't think we got any that clearly indicated backward-looking. Of course, I agree that the NCK study drew finer distinctions here that would be worth exploring in future work.

Your responses a) and c) are surely right, but I think (unless I misunderstand you) they're directed against an earlier version of the paper, in which we drew conclusions considerably stronger than we were entitled to about determinism, could-have-done-otherwise-ness, and the like. a) and c) are excellent reasons not to draw those conclusions, and I hope we've backed off.

Your response b) is also a very present worry about the studies in the paper (although less so about the abstract/concrete studies, since they already rule out Dennis's having known beforehand), but we ran follow-up studies that ruled out Dennis's having known about his disorder beforehand (these were the same studies in which we switched the wording to "morally responsible"), and got the same results. This leads us to suspect that b) isn't driving the intuitions in question.

Response d) opens up a whole world of future research possibilities: what descriptions of disorders affect responsibility judgments, and how do they do it? To get a better feel for folk intuitions in this area is an important project for future work.

I agree it's certainly too early to draw any firm conclusions about the factors that affect folk attributions of responsibility to the mentally ill, but hopefully we're taking steps towards a clearer picture.

[Joe O.]

I think that (d) is the 2-ton elephant in the room. I also agree that forward looking vs. backward looking responsibility is a pertinent consideration, but my comment is about (d)
It's not clear that a neurological disorder differs vividly in description from a personality disorder, for which, people could reasonably be held responsible. One might think that the folk aren't familiar with any distinction here, but that might also overestimating naivite. For example, I felt there was something neurologically wrong with certain bullies in high school, but I still thought they were responsible for their actions. I had, in a very rudimentary way attributed a personality disorder to them.

Given an accurate description of one's mental state given a personality disorder, such that it doesn't constitute compulsion, I can argue for moral responsibility.

Accurately describing a level of duress, or, as Blake notes, delusions that comes with the neurological disorder might change the results of these surveys drastically. Why? Maybe we should ask whether the subject of these vignettes is in control of his/her actions?

[Justin Coates]

Re: Study 2

Do you guys have any data regarding judgments of responsibility in cases (more or less) similar to the following:

Dennis and John have been friends for thirty years who always meet for a weekly walk. Dennis has been away on vacation for a month and so the friends have not been able to go on their walk until last week. On their walk last week they passed a jogger on their normal trail. Seemingly unprovoked, Dennis ran up to the jogger and punched him in the stomach multiple times. Shortly after this incident John asked Dennis why he attacked the jogger, and Dennis only replied that he was having a bad day and "really wanted to hit someone."

On a scale of 1-7, how responsible is Dennis for hitting the jogger?

I suspect that the mean judgment of responsibility would be much* higher than 3.8 in this scenario (since this is merely a way of asking whether we are typically responsible for our actions). If this suspicion is correct and the mean judgment of responsibility in this sort of case is significantly higher than 3.8, then people mitigate responsibility when certain sorts of psychological or neurological conditions are present. But now there doesn't look to be so much tension between how the folk respond and how philosophers theorize: these sorts of conditions do significantly mitigate responsibility.

Also, on page 10 you say that your scale ran from 1 to 7 with 1 being "not responsible" and 7 being "very responsible." If this is right, then among equally distributed responses, the mean would be 4.0. Then a mean of 3.8 would indicate that the folk are* moving towards judgments of non-responsibility. But this is just the sort of behavior our philosophical theorizing about these issues prescribes.

Moreover, Figure 2 starts at 0, but your scale starts at 1. This makes 3.5 look like the mean of equally distributed responses, and creates the illusion of 3.8 being on the 'responsible' side of things (if it's fair to carve things up like that), but as it's below 4.0, the mean judgment of 3.8 is actually on the 'non-responsible' side of things.

[Ken Batts]

How would people answer the following question:

"Dennis has recently found out from his doctor that he has a neurological condition that has caused him to commit violent crimes. Were someone else to have this neurological condition then that person would have had to commit the same crimes as Dennis.

Does society have the right to impose consequences on Dennis, up to and including lifelong incarceration, to protect itself?"

Equally interesting is how people would answer the following:

"Wouldn't society be best served by making the conditions of Dennis' incarceration appropriate for someone who has a health problem (quarantine, rehab, etc.), not for someone who must be punished because he has freely chosen to harm people?"

I know, it sounds like a push poll, but still it would be interesting to hear people's answers.

[Mark Smeltzer]

Anyone here watch the TV show Dexter?

The main character is a serial killer whom the audience is meant to feel inclined to sympathize with. The sympathy comes from both a horror thrust upon him in his early childhood and a sense that he's actually not that much different that some of us on our worst days.

Dexter typically kills someone each episode, and he ONLY kills people who "DESERVE" it. In other words, he follows a strict code of ethics.

In recent episodes, Dexter comes face to face with his long lost brother, Brian. As it turns out, Brian was also grossly affected by the same tradgedy and similarly became a serial killer. However, Brian kills people that seem to be innocent, and his approach is more malevolent.

The audience is left to sort out the moral delimma... Is Dexter evil? If not, is Brian? If so, why not Dexter? Are they both evil? If so, does their horrific past affect our right to blame them for their crimes?

Seems relevant.

[Ken Batts]

There's another question with I think a more obvious and useful answer: Does society have the right to stop Dexter and his brother from committing their crimes, regardless of the reasons they commit them? The question of blame seems irrelevant when public safety is so seriously threatened.

I sympathize with both Dexter and Brian, but would have no problem placing them permanently behind bars. They are criminally insane, which should be a good enough reason to keep them away from the rest of us.

[Linda Harris]

Mark: Perhaps the question is really about who is the greater stakeholder in the retribution process. That is, who benefits more from stuffing Dexter and Bro in a cage. Dexter, Brian, or society?

Just a thought: I wonder if Dexter will one day feel compelled to kill Brian?

For all: A very relevant article: Wasserstrom, Richard. "Punishment v. Rehabilitation", Philosophy of Punishment. Ed. Robert M. Baird and Stuart E Rosenbaum. New Work: Prometheus, 1988. p 57-65.

Pages 59-60 deal specifically with the issues of mental defect and crime.

[Ken Batts]

(Don't read this if you don't want to know what happened in the season finale of Dexter)

Linda: Dexter DID kill Brian! But he did so with very mixed feelings. On the one hand Brian fit the criteria for the type of person Dexter feels compelled to kill. But on the other hand Dexter knew Brian understood him and accepted him as he really was, more than the other people in his life ever could if they knew the truth about him.