Does moral responsibility require the ability to do otherwise? For example, must one have been able to refrain from an evil deed if one is to be appropriately blamed for it? The answer turns on the truth of a familiar principle:
(PAP) If S is blameworthy for doing X, S must have been able to do otherwise than X.
The traditional view is that (PAP) is true; Frankfurt argued that it was false, with a form of example which is still widely discussed. I’m going to argue for Frankfurt’s conclusion in a way that has nothing to do with Frankfurt-style examples. I’d be interested in feedback.
Blaming (or punishing) someone for failing to live up to a moral standard is a special case of a more general phenomenon. There are many cases where there is some kind of requirement, someone fails to live up to it, and negative consequences are imposed as a result. It is instructive to look at how we view “couldn’t have done otherwise” in these other cases.
Case 1. I give a class a logic exam. One student turns in a blank blue book. I ask him why. He says that he couldn’t do any of the problems. I ask him to elaborate. (Case 1a) He says, “because I haven’t understood anything in this class.” (Case 1b) He says, truly, “because I was sitting at the back of the row and never got a test. I tried to get your attention but you were playing Pacman.”
I take it that in Case 1a, the fact that the student couldn’t do the problems is not an excuse. He just fails. In Case 1b, the fact that the student couldn’t do the problems is a valid excuse. He should get a makeup test. What is going on?
Here is my view. I administer the test in order to see whether students have a certain level of knowledge and skill in logic. Ordinarily, their performance on a logic test is pretty good evidence of their knowledge and skill. In case 1b, this evidence is defeated: the fact that a student without a test cannot complete logic problems is no evidence that they lack knowledge or skill in logic. In case 1a, however, the evidence is not defeated: the fact that a student who doesn’t understand logic cannot complete logic problems is excellent evidence that they lack knowledge and skill in logic.
Case 2. A boss tells his secretary to type up a business letter. She replies that she is unable to. Case 2a: She is unable because she doesn’t know the format of a standard business letter. Case 2b: she is unable because the office computer is down.
In case 2b, the explanation of the secretary’s inability to type the letter defeats any inference from “can’t type a business letter” to “is a lousy secretary.” In case 2a, however, inability is no excuse, precisely because this inference is intact.
Now turn to the case of moral responsibility and blame. Suppose I promise to pick you up at the airport, but I don’t show. When you call me angrily, I explain that I couldn’t come. Case 3a: I couldn’t come because I was feeling incredibly lazy and couldn’t drag myself out of bed. (Alternatively: I couldn’t bring myself to come because I have developed a deep hatred for you.) Case 3b: I couldn’t come because my car had two flat tires and they had blocked off the main road to the airport.
It seems to me that case 3a is vastly different from case 3b. My psychological inability to keep my promise in case 3a is no kind of excuse, while the various physical obstacles in case 3b are a good excuse for not keeping my promise. Here is an explanation of the difference. Moral standards are a set of standards for good (enough) “quality of will.” What we care about is the quality of the relationships in which we stand, which has to do with the attitudes others take toward us. Failure to abide by behavioral standards like promise-keeping is generally good evidence that one does not have the proper, good-relationship-constituting attitudes toward another. However, this evidence can be defeated, and it is defeated in case 3b: I didn’t keep my promise, not because my will toward you is bad, but because it was physically impossible. In case 3a, however, the evidence is not defeated. The explanation of why it was impossible to keep my promise reinforces and confirms the initial impression, that my will toward you is bad. So “couldn’t have done otherwise,” in this case, is no excuse.
Thinking about moral responsibility in these terms gives us a theoretical argument against (PAP). Blame, as a negative response to the violation of a moral standard, is an instance of a larger family of similar cases. Denying this is, I think, a fairly heavy theoretical cost. If we accept it, however, we eliminate even the expectation that (PAP) will be true. For then we should also accept that there is a reason for holding each other to moral standards, that is, a point to the endeavor, just as there is a point to giving logic exams and to making sure secretaries can type letters. Whatever that point is for moral responsibility, there will be some kinds of inability to do otherwise which do nothing at all to excuse behavior that violates moral standards. (I have suggested a Quality of Will account, but that detail is inessential.) They will be the cases where inability to do otherwise is explained by an absence of whatever features the practice of holding each other responsible is trying to ensure. So we have a good theoretical argument that “couldn’t have done otherwise” is not a universal excuse, and therefore that (PAP) is false.
This is a really neat line of thought, and it opens up rich lines for investigation.
Some rambling thoughts…
1. My own inclination is to say that blame is different from negative assessment in the other cases, in part precisely because the excuses available are different.
2. Here’s an interesting related case. Brilliant student takes a calculus test. She gets a B because she didn’t bother to study. Struggling student takes a calculus test. He gets a B because he studied harder than anyone else. The grades are perfectly fair, because we’re not evaluating the moral worth of their actions. But naturally we want to praise the struggling student, because he couldn’t do any better, and we want to criticize the brilliant student, because she could have done better. So as soon as the mode of evaluation moves closer to the moral sphere, the ability to do otherwise seems to start to count for more.
3. What about the suggestion that we’re evaluating the quality of the person. That’s not quite right on exams. I may know that a student knows all the material, but if she deliberately gives all the wrong answers in order to send a message to her parents that she should change majors, I still need to give her an F. So while the reason why I chose the exam I did was because of a high correlation between doing well on the exam and knowing the material, nonetheless in the final evaluation the actual performance trumps the knowledge.
4. I think that in the moral sphere, blame attaches in the first instance to particular actions. We praise uncharacteristic good deeds and punish uncharacteristic bad deeds, barring special circumstances.
Or consider repeat offenders. After a bunch of repetitions, of course we have very good inductive data as to what the person’s character is. But we punish each action, and the legal system carefully counts up the offenses.
5. Notice, too, that we accept excuses that admit a person to have a seriously flawed moral character. For instance, brainwashing is an obviously good excuse, even if there is no question that the brainwashed person has a seriously flawed moral character as a result of the brainwashing. This is disanalogous to the nonmoral cases. If a calculus student had a serious head injury as a child and hence is incapable of abstract mathematics, she still needs to be given an F (the F needs to be given with as much kindness as one can muster without causing needless embarrassment).
6. I am inclined to think that the various ways of criticizing an action that moral blame is a special case are related to moral blame as follows. In moral blame, we criticize the action on balance in the circumstances. In the other forms of blame, we criticize an aspect of the action.
The mediocre student who studied really hard for a B in calculus did things that were on balance right and praiseworthy, though in respect of mathematical content some of them were wrong. The brilliant student who got the B acted in ways that were not only mathematically faulty but on balance faulty. It is not surprising, then, that when we look at the on balance evaluation of the action we accept a wider class of excuses.
Hi Heath,
I agree with you that praise and blame has to do with quality of will, but I think there’s a problem with the cases you’re discussing, which is that they all have to do with a failure to do X rather than a doing of X. Will this leave defenders of PAP some room to deal with your broader theoretical argument? It’s hard to say, but I think I can say something in defense of PAP that’s consistent with your broader approach to blame and responsibility. Here’s the argument:
1. S is blameworthy for doing X only if S did X.
2. X is a doing of S’s only if S could have done otherwise.
C. S is blameworthy for doing X only if S could have done otherwise.
The first premise seems trivial. The defense of the second premise will rest on considerations about the nature of action and the relationship between action and exercises of two-way powers (e.g., Steward’s account from her recent book). Since _that_ debate seems to have little obvious connection to debates about whether the quality of will account captures something important about blame, it doesn’t look like the truth of (2) will require abandoning that account. And if that’s right, it looks like there’s an account that upholds PAP that’s consistent with the quality of will account and an attack on PAP will require further attacks against the picture of agency that links action with the exercise of a two-way power.
Interesting stuff. Can I ask for clarification on a couple of points? First, I’d be interested to hear more about the “theoretical cost” that you think must be paid if we reject the claim that moral blame is of a piece with the other sorts of evaluation that you describe. Second, could you say more about why, if we do accept this claim, it’s inevitable that “there will be some kinds of inability to do otherwise which do nothing at all to excuse behavior that violates moral standards”? Putting on my (unfamiliar, uncomfortable, and unaccountably itchy) Kantian hat, for instance, I might respond to your case 3a by simply denying that the factors you cite could make you incapable of picking me up. So I would be denying that there are sources of inability that fail to excuse in that case, on the grounds that there are no sources of inability at all.
Thanks, folks, for the input.
Clayton,
I haven’t read Steward’s book, but your premise 2 looks deeply question-begging to me.
I would argue this way. The denial of PAP is part of a broader argument for the coherence of compatibilism about moral responsibility. Two-way powers either are or are not compatible with determinism. If they are, then the compatibilist doesn’t need to deny PAP. This is the classical compatibilist strategy and I suppose I am not a fan. Be that as it may … if two-way powers are not compatible with determinism, then it might turn out that no one had ever done anything. That would be an extremely counterintuitive result. I would say it counts against 2.
Dale,
The theoretical cost I have in mind is this: there are all sorts of evaluation which have the basic structure of moral responsibility, meaning that there are some standards and people impose positive or negative consequences on those who (fail to) meet them, in backward-looking fashion. If moral responsibility is held to be somehow very different from all these other cases, that is prima facie special pleading and requires at least some theoretical explanation.
I said that there will be some “inevitable” failures of PAP because if any set of standards is in place to ensure goal G, then inability to perform G will ipso facto not be a sort of inability we will regard as an excuse for not meeting the standards. I suppose one could hold that such inabilities were metaphysically impossible—we just could not have wills that were psychologically determined in blameworthy ways. That sounds to my ears rather ad hoc, along the lines of a metaphysical necessity to knowing how to type business letters or do logic proofs, and these days even many libertarians would say we can be psychologically determined, only we have to freely choose our way into it.
Alex,
On 1: see my “theoretical cost” comments to Dale above.
On 2: I would say we would praise the excellence of Brilliant Student’s mathematical performance, and praise the excellence of Struggling Student’s moral performance. That fits the rest of what I want to say.
On 3 and 4: I think what we care about, ultimately, is states of the person, or something like dispositions. But what we evaluate and respond to, in both moral and mathematical cases, is overt actions or performances.
On 5: One consequence of my view is that brainwashing etc. is less of an excuse than is sometimes thought. (Let me add: is sometimes thought, in philosophy seminars; I think the folk are less susceptible to it in real cases. For example, Watson makes a good case that Robert Harris was deeply morally warped by his childhood, but there’s not much inclination in the legal system, nor even among his fellow Death Row inmates, not to blame him for his crimes.) Once we see the reason that psychological determinism to evil is no excuse, we will also see that it doesn’t matter how one got that way, just as it doesn’t matter, for testing purposes, whether you can’t do calculus because you didn’t study or because you have brain damage.
That said, if there were some way to easily change a person for the better, then it might make sense to go ahead and do that instead of blaming or punishing them. For example, the victims of Frankfurt interveners are often described as having their choices controlled at the flip of a switch. Presumably, they can be fixed at the flip of a switch too. If that were so, then it would make sense to just flip a switch rather than go through the blaming-and-punishing routine.
On 6: I would bet that your “on balance,” closely analyzed, will boil down to an evaluation of the agent’s will and nothing else. For example, suppose the struggling calculus student is from another country. They are not properly prepared for the material and can’t answer any of the questions on the exam. They also conduct themselves in ways which come across as both disrespectful and disgusting in this country, though they would be appropriate in their home country. Maybe they have even had some kind of postmodern education and think of calculus as performance art with mathematical symbols. In short, they fail to meet every conceivable standard except … they tried! Really hard! Which is to say, their will was good, and that was the only thing that was any good.
Heath,
Thanks for the post. Here are some thoughts:
As others have suggested, there’s an important difference between sheer evaluation and having the reactive attitudes associated with holding moral responsible. We can evaluate a logic student as not understanding logic or a secretary as incompetent without having or thinking it appropriate to have reactive attitudes toward them. If student 1a has tried hard but doesn’t get it, we evaluate but don’t hold responsible. Likewise with the secretary. We can fire secretaries and fail students without making anything like a judgment of moral responsibility.
I think you’re on the right track in thinking that our practices of holding responsible are largely to do with tracking (and influencing!) quality of will. But I also think that people would rightly reject the claim that it was impossible for me to pick you up because I’m too lazy. People would not (and should not) agree that it was psychologically impossible for me to pick you up because I’m lazy (I also don’t think they would agree that it was metaphysically impossible, but they could agree with this while allowing that it was psychologically possible, and insisting that the latter is what matters).
It would take extraordinary circumstances for normal people to agree that it was *impossible* for one person to pick another up from the airport due sheerly to their laziness or other failures of character or will. And in the extreme circumstances in which they would allow that it was impossible, I would expect their tendency to hold morally responsible would attenuate or disappear.
In general, I think it’s very important when constructing cases like this to be very careful about stipulating things that people would not in fact agree to in real life, and then using intuitions about moral responsibility to get philosophical leverage. People might well hold you morally responsible for your failure of quality of will, but they would not grant that it was (psychologically) impossible for you to pick up your friend. They would say something like, “Yes you could, you just have to not be so damn lazy.”
In my view, the fact that people won’t count your laziness as an excuse is intimately bound up with their unwillingness to grant that your laziness makes it impossible for you to do things like pick people up. And that is most fundamentally because your laziness and quality of will can be greatly affected by your own and others’ reactive attitudes toward it and you, whereas the road being blocked and the performance of students and secretaries *who try hard but nevertheless fail* cannot be.
It is very tempting here to say that these people would fail to make the crucial distinction between whether something is really psychologically impossible vs. whether it’s instrumentally valuable to act as if it were. But since I have already gone on too long, I will only say that I think this is a deep though ubiquitous mistake. The *general standards* for (as opposed to individual judgments of) psychological impossibility are properly responsive to normative considerations.
Eric,
It seems you have two points. The first is that there need be no reactive attitudes associated with non-moral evaluations. I agree, but I also think reactive attitudes are only contingently associated with moral evaluations (it’s an empirical fact about human psychology, explicable in evolutionary terms). There are a number of ways to argue this point, but quickly: (i) the traditional view of legal judgments is that they are a kind of moral judgment, but it is better if the judge and jury lack reactive attitudes toward the defendant. (ii) Both Buddhism and Christianity, rightly understood, don’t suggest ceasing to hold others responsible for their bad actions, but they do suggest ceasing to do so with attitudes of resentment, contempt, indignation, etc. (iii) I often have (what we can euphemistically call) reactive attitudes toward my children’s bad behavior, but I strive to hold these in check. They are a hindrance, not constitutive, to appropriately holding my children responsible for their behavior. (iv) Consider a race of Vulcans …
The second point is that the folk won’t grant claims of psychological impossibility in cases of laziness, hatred, etc. As an empirical matter, I’m doubtful, both that the claims are always false and that that the folk would never grant them. However, let that slide. If what you say is true, it is at best contingently true. Surely there is some other possible species, much like us, who could suffer from these sorts of psychological determinism. Then, I would argue, this sort of determinism would be no excuse for their behavior, since what people care about is standing in good relationships and their problem is precisely an inability to stand in good relationships. So there is still an argument that (PAP) is not a conceptual truth.
Heath, thanks for your reply.
I think it’s important to distinguish between holding morally responsible and moral evaluation, and between holding morally responsible and holding responsible in other ways, including legally (the traditional view is, I believe, clearly mistaken; it is, I think, clearly possible to hold legally responsible without holding morally responsible). I also don’t think your reactive attitudes toward your children entail that you are holding them morally responsible, though perhaps you are. I think it’s at least good an interpretation of those Buddhist and Christian traditions to see them as opposed to rather than attempting to improve the practice of holding morally responsible (a la Pereboom).
I was only meaning to address the specific topic of holding *morally responsible*, which I think there are excellent reasons to think is intimately bound up with reactive attitudes. Even if this weren’t so, I think there are other features of holding morally responsible that are strongly in tension with an acceptance that the person could have done otherwise. However, I don’t think this is a conceptual truth, but a human psychological one, grounded in deep facts about our socio-motivational psychologies. I also don’t think the interest of the PAP lies in whether it’s a conceptual truth or not. What’s interesting to me is understanding why people are prone to accept it and normatively evaluating whether and in what sense we should accept it.
And of course there is the issue of diachronic evaluation, which I think is utterly pervasive and too-often overlooked. People might accept that it was impossible for you to pick up your friend given your actual psychology at the time, but not accept that you couldn’t have done something in the past to make that not so. And in the sense that matters (to them, for their practices of holding responsible), that implies that it was possible for you to have done otherwise than you did.
correction:
“strongly in tension with an acceptance that the person could *not* have done otherwise”