Kill Your Darlings: Bernard Williams' "Moral Luck"
The end of the decade's got me thinking about what I've done in the last ten years, what I've enjoyed, and what I've read. In that light, I decided to go back and re-read what I've always thought of as one of my favorite papers: Bernard Williams' "Moral Luck" (you can find a pdf here--it starts on page 20).
Although I still like the paper quite a bit, I found myself more critical of it this time around than I have been in the past, and, given that apparently I have nothing better to do, I decided to put my thoughts down. In all likelihood, this will serve as lecture notes the next time I teach the paper.
I've given a brief summary of the setup of the paper and the relevant bits that I'm criticizing, but I don't go into the whole thing. In fact, I stop fairly quickly into the essay (I don't even get to Anna Karenina or the agent regret stuff). I might do a follow up post that covers the second part, but what I've written below is very long as it stands. I've also included little grey "justify it" boxes that make explicit some of the stuff that's either in the background or that is important, but tangential to the central arguments or exposition of the post. Those who might not be familiar with Williams' essay or with philosophy might find them useful--those who are familiar with both can skip them (although they do provide some insight into what I think is going in some tricky parts). Finally, those who are very familiar with Williams' essay can feel free to just jump down to the criticism sections and see if those make sense. Okay, here we go.
The paper itself, like most of Williams' work, is a difficult one and Williams does little to help his reader in some of the trickier parts. The ostensible thesis, however, is clear: the idea that moral value is immune to luck is a mistaken one.
Justify It Box 1: Why care whether moral value is immune to luck?
Here's the brief argument that's lurking in the background in favor of this mistaken view. We know that the good things in life can be stripped away by contingent forces: a tornado can destroy your gorgeous house; a dip in the market can cost you your excellent job; a car accident can take away your beautiful family; a clump of mutating cells can ruin your perfect health; and so on. We tend to think that we can plan ahead to prevent such things or, at the very least, to mitigate their effects when they happen, but, realistically, we know that there's nothing that can be done to prevent all (or even most) such cases.
The problem, then is that if the good life is simply a matter of attaining the things in life that are good, and if those things are themselves vulnerable to luck, then it seems that the good life itself is a matter of luck. This, in turn, tends to bother people quite a bit since, if true, it suggests that one is more or less impotent with respect to how well one's life goes. Arguably, only a person who is utterly indifferent to one's well-being, one's projects, and one's life is completely unbothered by this prospect and there are very few people like that. For most human beings the effect on luck will have some troubling effect and its mitigation will be of some concern.
All that being said, the problem can be avoided altogether if first, there is something of value that is immune to luck, and second, if that thing of value is itself something that we can attain and which is substantial enough to form a life around. After all, if there were something of value that is immune to luck but which we couldn't attain, then our predicament wouldn't be improved (indeed, knowing about it might make it even worse!). And, if there were some such thing that we could attain, but which were of such negligible value around which someone couldn't plausibly form a good life, then we're still in the same boat ("Yes, it's true, my kids are dead, but I finally saw the perfect shade of red and that can't be taken away from me" doesn't paint a picture of the good life to me even if we assume that seeing the perfect shade of red is of some value).
Many people have held that both of the requirements needed to solve the problem are to be found in moral value. The idea that morality is of substantial (if not supreme) value is neither an uncommon, nor an implausible view. The question, then, is whether it is immune to luck as well. Some philosophers--notably Kant, but one can make an argument for Plato as well--have argued that it is. Very simply put, for Kant, morality is a matter of exercising one's rational nature in the proper way and such an exercise of one's rational nature is not vulnerable to contingencies of luck. Furthermore, since all people have a rational nature that operates in the same way and to the same degree, all people regardless of circumstance are able to be moral, and hence, to live the good life; regardless of what happens, as long as one remains a person and has a rational nature, one can be moral, and, hence, one can live the good life.
If all this is right, then the problem that the good life can be take away from us due to luck falls to the side. At the very least, we can have a very good life by being moral, and, furthermore, if moral value is the supreme value, then we can get the best life by being moral.
Williams' primary concern in the part of the paper that I want to talk about isn't about showing that whether one acts or is even able to act morally is a factor of luck (that's more Nagel's aim). Rather, he wants to first focus on the role of justification and the conditions under which one can be justified in taking a course of action. It's only near the end of the paper that he links together justification with morality. Williams' stated reason for going this route is that the Kantian view mixes together the notions of rationality, morality, justification, and supreme value in such a way that it has the consequence that whether one is justified in acting cannot be a matter of luck (see below for more detail).
Justify It Box 2: Why does Williams think the Kantian view links morality with justification?
This is one of the many frustrating places in which one wishes Williams would explain what he has in mind. I'm not a Kant scholar, and, admittedly, my understanding of Kant is not stellar. However, I believe what Williams has in mind here is something like the following: one figures out the right thing to do by by testing the maxim on which one acts for conformity with the Categorical Imperative. If the maxim passes the test, then acting on it is the right thing to do, and if it is the right thing to do, then, clearly, one is justified in acting upon that maxim since one is always justified in doing the right thing. In this respect, one's justification for doing x and x being the right thing coincide. Crucially, whether a maxim is in conformity with the Categorical Imperative is entirely a matter of certain of its logical properties and structure, and as such, is a matter that's entirely independent of anything about the agent herself, or anything that happens or might happen to her. In other words, the maxim fails or passes its test in the rational realm where luck is not a factor. Just as it can't be matter of luck that 2 + 2 = 4, so it can't be matter of luck whether a maxim is in conformity with the Categorical Imperative.
If this is the right picture, then it follows that it also can't be a matter of luck whether one is justified in acting.
The strategy then is to show that whether one is justified can indeed be a matter of luck. If he's able to show that successfully, and if justification, rationality, and morality are as tightly bound as the the Kantian picture presents them to be, then it seems to follow that rationality and morality will also be vulnerable to luck.
With respect to his methodology, Williams is mercifully clear:
My procedure in general will be to invite reflection about how to think and feel about some rather less usual situations, in the light of an appeal to how we—many people—tend to think and feel about other more usual situations, not in terms of substantive moral opinions or 'intuitions' but in terms of the experience of those kinds of situation.
So, he's going to present us with some unusual cases which will pose problems for the Kantian view of justification he's attacking, and he'll argue that we should judge these cases on the basis of more ordinary cases in which our judgements, are, presumably, in agreement (and not because we have some theoretical commitment that makes us have those judgments). To that end he gives us the first of two cases:
Modeled after the real Gauguin, our Gauguin is a creative artist who is considering foregoing certain real and urgent moral commitments in order to live a kind of life that he believes will let him pursue his art. More specifically, we can imagine that he is a married man with a spouse and children who need his support to have a minimally decent life, that he thinks he could be a fantastic painter if he didn't have to support his family and could devote himself entire to his art, and that it's impossible for him to achieve success as a painter while supporting his family. He is faced with a choice between two incompatible lives: one in which he supports his family and decisively fails to be a painter, and one in which he tries to succeed as a painter but in which he decisively fails to support his family. Furthermore, he's someone who sees the demands of morality as important--he realizes that if he were to abandon his family he would be committing a serious moral wrong--but he doesn't think that these demands are decisive.
We are to imagine that with all this in mind, our Gauguin decides to pursue the life that he holds will allow him to be a great painter and leaves his family. Crucially, in making this decision, he does not know whether he will be successful as a painter or not.
What are we to make of this case?
Williams' claim is that in Gauguin's case the only thing that could justify this choice is his success in his endeavor. It's clear that if he fails in becoming a great painter, then he not only did the morally wrong thing in abandoning his family, but, crucially, that he was not justified in abandoning them when he did. Simply put, there is nothing to be said in favor of his being justified in making his decision the way he made it. However, Williams claims that if he's successful, then there's at least some basis for thinking that he was justified in doing what he did--a basis which, to reiterate, he would not have if he fails.
To put the matter another way, if he's successful, then in response to being confronted with the claim that he was not justified in leaving his family, he can reply with something that has some pull on us; namely, "Yes, but look at the greatness I've achieved which would not have been achieved otherwise!" This might not be sufficient to completely exculpate him from being a total bastard (certainly not to his family!), but we're supposed to have the intuition that it's at least something which has some weight and which speaks in favor of his action.
Before we delve into whether Williams is right in making these claims it's important to make note of two points: the first is that given how the scenario is set up, Gauguin does not know prior to making his decision whether he would end up being successful. In fact, he can't know since whether he is successful is a matter of whether he makes the very decision that he's deliberating about. Given that this is the case, it's impossible that his actual success can factor in as a justification for trying to succeed--there simply is no actual success at that point but, at best, only the possibility of success. In other words, the justification of his success is something that can only occur retroactively. Crucially, this is one of the places where the wedge between justification and morality can be inserted.
The second point is that what matters in whether Gauguin is justified is not simply whether he succeeds or fails, but the source of the failure. Williams is explicit that a kind of external failure (e.g. a heavy crate crushing Gauguin's hands while en route to Tahiti, preventing him from ever painting) would not be sufficient to demonstrate that he was unjustified in leaving. In such a case and others like it, the question is, I believe, essentially supposed to remain an open one since the claim that he could have succeeded had this event not occurred is likewise still open. What decisively shuts the door on Gauguin not being justified, then, is a kind of intrinsic failure of the project; as Williams puts it, it has to be the case not that the project fails but that he fails at the project.
Crucially, whether one Gauguin is the kind of man who really can be a great painter is a matter of what Williams calls 'intrinsic luck'. As he puts it: "it is not merely luck that he is such a man, but luck relative to the deliberations that went into his decision, that he turns out to be such a man: he might (epistemically) not have been. That is what sets the problem." (pg. 25)
Justify It Box 3: What sets the problem?!
I admit, I find Williams' claim about "luck relative to the deliberations that went into his decision" to be puzzling. The best I can make out of it is that he thinks luck factors doubly in Guaguin's case. In the first place it's a matter of luck whether Gauguin just is the kind of person that can be a great painter (see my brief criticism of that claim below); and in the second place that it's a matter of luck that his very decision hangs on luck about his ability to be a great painter. That is, he's (un)lucky insofar as whether he's justified is a matter of something he can only know post facto.
I suppose this can be made sense of though if I'm correct, then, the latter strikes me as a kind of luck of circumstance and the former a kind of luck of character (loosely put--again, see my criticism below). Here, there may be some overlap with moral luck in Nagel's sense. Specifically, I'm thinking of a case in which a person might fail to do the right thing by virtue of the fact that they have the dual bad luck of being in a bad situation for which they're internally badly disposed to handle. Thus, one might imagine being unlucky in being the first responder to a car-crash and being unlucky in being naturally too cowardly to provide the necessary help. Why the situational kind of luck should be an intrinsic kind of luck, though, is a mystery to me. It seems more accurate to me that situational luck is extrinsic even if the situation in which I find myself is one in which intrinsic factors weigh heavily.
Alternatively, I suppose Williams could mean that the intrinsic factors are such that they cause one to be in a certain situation which they wouldn't be in if but for certain matters of luck about who they are. In Gauguin's case, the situation is the very one of deliberating about leaving his family and it's caused by the fact that he currently doesn't know whether he can cut it as a great painter. That is, if he could know that he wasn't going to be a great painter, then he would never be in the situation in which he has to deliberate about something he can't be ex ante justified. But he isn't and in that respect he's unlucky.
Again, I just find this stuff really hard to parse.
Williams' stress on intrinsic luck here makes up my first major criticism. Personally, I find the notion of intrinsic luck--especially with regard to something like artistic talent--to be pretty suspect if only because it seems to entail that things like artistic talent are just ingrained features of the person. Williams seems to hold the view that whether one is a great painter (destined to be one?) is some kind of intrinsic feature that one is either graced with it or not as a matter of luck and which one discovers through the course of their life. Not only is this not very g r o w t h m i n d s e t oriented (joke) but I also find it to be weirdly essentialist in nature. Maybe it's something that comes of his broadly Nietzschiean commitments about human nature and determinism, but I see no reason to hold that view without something extra to back it up.
In any case, we can set the question of whether being a great painter is written in the stars and focus on a different matter. In particular, there's tension between Williams' claim that the only thing that could justify Gauguin is his success and the claim that he's not unjustified if his project fails as a result of extrinsic luck.
It's true, the two claims are not obviously inconsistent since not being unjustified does not amount to being justified so it remains possible that the only way to be justified is to succeed. However, the tension can be brought out if we reflect on what it must mean for him to not be unjustified. Williams' direct quote is: "Irreducibly, luck of this kind affects whether he will be justified or not, since if it strikes, he will not be justified. But it is too external for it to unjustify him, something which only his failure as a painter can do." (pg. 25) The only way I can make sense of this claim is to hold, as I do above, that the question of whether Gauguin is justified is theoretically still open. It's not actually open since Williams admits that Gauguin has failed and that because of that failure he can't be justified. But it's still in some sense open since he's not unjustified.
That all seems fine and well, but why wouldn't he be unjustified? Suppose someone says to Gauguin "You really didn't make it as an artist cause of all that hand crushing business, huh? In light of that it looks like your leaving your family was completely unjustified." What would be said in response to make the point that he's not unjustified? The most obvious answer seems to be that he's not unjustified because he could have been a great painter if but for the hand crushing. I think this is right, but it looks to me that this means that the possibility of having been a great painter serves as some justification or some reason for claiming that one is justified. Granted, it's not enough of a justification to count as justifying his action completely (although we should remember that even success doesn't do this), but it's enough to push him from being unjustified into not being unjustified.
Justify It Box 4: Aren't you assuming some stuff?
I explicitly assume two things. First, I assume that the burden of proof is on Gauguin (or Williams) to explain why he's not unjustified. That is, I assume that the abandoning of his family is pro tanto sufficient to hold that he's completely unjustified unless he can offer some justification to push in the other direction. This seems like a reasonable assumption that (to stick with Williams' own methodology) most ordinary people make. Second, I also assume that the shift from unjustified to justified is a matter of offering pieces of justification or reasons that speak in favor of one being justified (and vice versa--to move from being justified to being unjustified is to be presented with reasons that speak against one being justified).
Putting those two assumptions together with the claim that Gauguin is not unjustified because he could have been a great painter amounts to the claim that the possibility of being a great painter is itself a bit of justification for his decision (if, indeed, that is what Gauguin would say in response to his accuser--maybe I'm just completely wrong about that). So, it turns out that it's not just his actual success that serves as some justification, but also his potential success or the possibility thereof that does too.
This might seem like small potatoes since all it means is that Gauguin now has another piece of justification available to him. However, I think it actually causes big problems for Williams. Let's assume that Williams is right that if Gauguin were actually successful as a great painter then he would have some justification for leaving his family (more on that below) and that he can't know if he's justified ex ante. Even if this is true, it's simply not true that the justification regarding the possibility of being a great painter can't be known before making the decision. Indeed, that can be reasonably estimated and on that basis we can judge whether Gauguin is or is not justified in leaving.
Before I get into how we do this (the procedure will be familiar), I should say something about what kind of possibility we're talking about when we say that the possibility of being a great painter is some justification for Gauguin. Clearly, it's not some kind of logical, metaphysical, or nomological possibility that's at play--if that were the case, then there would be justification for almost anything. There would, for example, be some justification for putting drain cleaner in your coffee because there's some possible world in which it doesn't kill you. No sane person would take this as any kind of justification. Rather, what we're talking about is a kind of counterfactual possibility tied to a probability of success; i.e. if the crate hadn't crushed his hands he would have had a decent chance of succeeding at being a painter or there would have been at least some likelihood of success.
Here, I think Williams is unfairly benefiting from the fact that we know that the real Gauguin really was successful as a painter and that we implicitly smuggle that in. Consider what happens if we mess with the likelihood of success in setting up the example. Suppose we stipulate that our Gauguin has never put brush to canvas in his life, or that instead of painting he wants to leave his family so he could train and beat the world record for the 100 meter dash despite never having run a day in his life. In other words, imagine that his aspirations are entirely unrelated to anything in his life that indicates he would have any success in achieving what he sets out to do--indeed, we might even point to factors that indicate that he's highly unlikely to succeed ("Gauguin, my dude, beating a world record requires a lifetime of training that you just never had.")
In these cases it makes sense to think that even if he can't know whether he'll be successful in beating the world record he has pretty damn good reason to think that he won't be able to. This also seems like pretty damn good reason to think that he's not justified in leaving and not because of any moral considerations, but precisely because he's very likely to fail. Indeed, we can also have reason to think that he'll fail not because of some external reasons, but precisely because the evidence points to the fact that he is not the kind of person who can do the thing he sets out to do. That is, we can argue that he's likely to experience a kind of intrinsic failure. Most importantly of all, we can figure this out before he sets out to make his decision--there's just not the kind of uncertainty that could only be closed by finding out later whether one is ultimately successful or not.
As stated above, this procedure is a familiar one and we do this kind of reasoning all the time. I might get the idea that I should abandon my studies to pursue a career in country music. In trying to figure out whether I would be justified in making this decision you might very well ask me if, for example, I know how to play an instrument or sing, whether I understand or even like the music, whether I have any ins into the business, and so on. In other words, you might reasonably ask whether I have any reason to think that I'll succeed in this. When you find out that the answer to all of these questions is 'no' you might reasonably say that I'm not at all justified in doing this (and you would be right!).
All this is to say that Williams' argument relies on the existence of cases in which the only justification that needs to factor into whether one should do something is available only post facto. I've been arguing that even in the Gauguin example this isn't the case and that there's always some other evidence available prior to making the decision that can settle the question of whether he's justified in taking off or not.
Nevertheless, Williams could take my comments aboard and say two things. First, he could claim that an even more schematized version of Gauguin's case could be made in which we really can't say that he's unjustified prior to his making a decision. We might suppose, for example, that the odds are perfectly even that he could be successful or a failure, or that the situation is so ambiguous that it's just impossible to confidently say one way or the other. I won't have much to say about this other than that I suspect such cases would be extremely rare; if moral luck regarding justification can only appear at those fringes then I'm not terribly worried.
More importantly, however, he can still insist that even if Gauguin makes his decision to leave while being completely unjustified, if he were to nevertheless succeed he would still have some kind of justification available to him. And he could argue that this is really the important thing to note in this highly schematized example.
(It should be noted that the role of luck here has changed. It now strikes me that in the case that our Gauguin succeeds the luck is precisely in his success. But I'll set that to the side.)
The question, then becomes whether a success like Gauguin's really is any justification for having done what he did. Or more broadly, whether post facto justification makes sense. I don't think it does and I think that it's pretty easy to see why by looking at cases of negligence. This is my second objection.
Suppose, for example, that I, never having shot a gun in my life, come to think that I can become the world's greatest marksman. As my first attempt to do this I take an apple and place it on my infant son's head (don't worry, I don't really have any kids) and shoot the apple off. Prior to shooting it, most people, I hope, would think that I am not justified in taking the risk of missing the apple and killing my son (or of even exposing him to that kind of danger). Suppose, furthermore, that I actually hit the apple and my son remains unharmed. In that case I was successful and we can even say that my success is some evidence, however slight, that I am the world's greatest marksman (after all, I did this on the first try with no practice!). However, I don't think anyone would think that my success provides any post facto justification for having done what I did.
I might try to defend myself by saying that I have something to say in my defense (viz. that I shot the apple) which I wouldn't have if I had failed and killed the boy since, in that case, nothing would have justified me. But this strikes me as unconvincing. That I didn't kill the boy or that I succeeded in shooting the apple does not, in this case, serve as any kind of post-facto justification at all. We would certainly be right to say that I was lucky, but the central claim of Williams' point is that this kind of luck can act as justification for having made the decision to shoot in the first place. And I simply can't see why that should be the case.
Crucially, the example with my son, while extreme, is relevantly analogous to Gauguin's case since I have opted for a choice that neglects pressing moral claims on me by my family in order to do something that exposes that family to serious risk. What's under discussion is whether success under these conditions provides any justification for deciding in the way I did--Williams argues that the Gauguin case does and I insist that in the case with my son it adds absolutely nothing.
Justify It Box 5: But aren't there relevant difference here?
It's true, my example is not fully analogous since one might argue that nothing is gained by my putting my son at risk for my stupid dream of being the world's greatest sharpshooter and succeeding while something is gained by Gauguin putting his family at risk and succeeding at being a painter. One might also say that the value of art is such that success in that domain does offer post facto justification while success in gun-play does not and that this is where the real difference lies.
This might be right, but this doesn't strike me as something that Williams would take up. In the first place, this line of thought is almost explicitly consequentialist and Williams was no consequentialist. But even if he would avail himself to this kind of reasoning, I still think one could put pressure on the earlier claim that this kind of justification can only be known post facto since that simply isn't true if one takes up this line. That is, it's not the case that we don't know anything about the value of art such that it can't factor into our deliberations ex ante; it's not like we discover upon my successfully shooting the apple off my son's head that this stunt wasn't valuable or in Gauguin's case that art is valuable.
Maybe one would argue that at this point we're at loggerheads--Williams claims one thing, and I claim another. However, I think more could be said in my defense. In particular, it seems to me that Williams is relying on the fact that people tend to think that if something has been accomplished, then its having been accomplished speaks in favor of it having been done in the first place. But it really does matter whether this way of thinking about the matter--regardless of whether it's widespread--is a good way of thinking. And I just don't see don't see how it could be. Aside from the example I gave, one can construct countless others: the fact that I didn't kill anyone driving drunk last night does not mean that I was justified in driving drunk in the first place; the fact that I successfully pulled out all my teeth does not mean that I was justified in pulling them out; the fact that the palace coup succeeded is not a reason to think that it was justified; etc. Let's call the taking of a success in action as speaking in favor of the doing of that action the 'fait accompli' fallacy. Given the number of examples that one can generate to demonstrate that this way of thinking is pretty bad, it seems to me that the onus is on Williams to show that the when it is applied to the case of Gauguin this kind of reasoning is not fallacious. In other words, it must be shown that Gauguin's success really does provide any kind of justification at all without just assuming that.
Now, the obvious rejoinder here is that, again, there's something that can be said if Gauguin is successful that he wouldn't be able to say if he failed. This is true, but it won't cut it. If that 'something' that can be said is not a piece of justification (but instead a bit of fallacious reasoning) then the fact that it can be said is irrelevant. Likewise, the fact that one can affirm the consequent in the course of an argument is not a reason to think that that person is arguing well. Crucially, one can't simply assume that whatever is offered in response to such cases is justification.
Still, it is true that at least sometimes people do commit the fait accompli fallacy and that we do say things like "well, everything worked out alright in the end!" to at least attempt to justify their actions. Why do they do this? I think one reason is to stop criticisms about what could have gone wrong in a situation--it's no use yelling at me for taking a risk in doing x since it's all over and what could have gone wrong is in the past now. Clearly, if the thing does go wrong, then one can't say that one can't be criticized about what could have happened since it actually happened. This, I think, is the 'something' that can be added only if one is successful in what they're doing. Crucially, this is a move to stop a criticism, but the move itself does not stop criticism by showing that one was or is now justified, but one that relies on showing that further criticism would be fruitless. I think Williams confuses this bit.
Okay, let's wrap this up. I've attacked two of Williams' claims regarding justification in the Gauguin case: the first focused on the claim that in certain cases justification for one's actions can only appear retroactively on the basis of whether one ends up being lucky in certain respects. I argued that even in those cases it's just not true that the only justification is available retroactively and that we have pretty good ways of judging whether someone like Gauguin is justified on the basis of how likely he is to succeed in whatever he wants to do. I think we very frequently judge whether someone is justified in taking on a course of action in precisely this way. The second claim I attacked was the claim that being able to say something upon succeeding in an unjustified course of action provides some kind of justification for having done that action. I argued that this is implausible and that it rests on a bit of fallacious reasoning (what I called the 'fait accompli' fallacy). Given that there are plenty of easily constructed counter-examples in which success in taking an action adds nothing to whether one was justified in taking that action in the first place, the burden is to show why it should do this in Gauguin's case. Finally, I closed with a short argument about what I think is actually going on when people make this fallacy.
Okay, I didn't actually touch any of the stuff about morality or anything about agent regret. I might come back and do a follow up post at some point if I get bored again.
Justify it Box 6: Why are you writing this? Don't you have other work to do?
Yeah...I don't know why I do these things.