Many of our ordinary moral judgments and practices presuppose that there can be moral luck—cases in which two agents differ in blameworthiness despite the differences between them being just a matter of luck. Cases of moral luck can be sorted into categories based on what was outside of the agent’s control, such as what results from their actions (resultant luck), what circumstances they’re embedded in (circumstantial luck), and even how they’re constituted (constitutive luck), among others.
A common example of moral luck is that of two assassins, both of which shoot at their targets in exactly the same way but only one of which is successful (perhaps a bird intercepted the bullet of one of the assassins). It seems that a person who actually committed murder is more blameworthy than a person who merely attempted (and failed) to commit murder. And we treat the cases differently in practical ways, applying harsher sanctions to those who bring about worse results, even if those results are partly a matter of luck.
Some philosophers think moral luck is impossible and claim that in cases of apparent moral luck the blameworthiness of the two agents is in fact the same. (Another response is to become a skeptic about blameworthiness, but I will set aside that position here.) One might think that the successful and unsuccessful assassins are equally blameworthy but that they are responsible for a different number of things (only one of them is responsible for the death of a person, while both are responsible for the attempt).
In other words, and to use terminology from Michael Zimmerman (2002), who defends the view, one might claim that an agent’s degree of blameworthiness can remain constant while the scope of her blameworthiness varies. Combined with the idea that if the result of someone’s action is a matter of luck then one would be just as blameworthy whether or not the result occurred, this view promises to do away with resultant luck, and the same approach can be applied to circumstantial luck as well.
But consider the following response to constitutive luck. Sometimes we make choices or take up courses of action that affect how we are constituted later on. When we’re very young, we have no say over what we are like, but typical adults have contributed to the formation of their characters. To borrow an example from Alfred Mele (2006: 171): Chuck might cultivate his amorality by torturing animals until he no longer is squeamish about inflicting pain on others, with the result that he later acts from his bad character by killing his neighbor. And when someone acts from a character over which they had some control, that person isn’t total constitutively lucky in performing that action. If this is correct, it seems that we are able to mitigate our constitutive luck over time.
It turns out that this very natural response to constitutive luck is inconsistent with the response to resultant (and circumstantial) luck sketched above, for the way in which agents shape their characters is itself subject to other kinds of moral luck. Suppose that Chuck’s way of making himself morally worse is by shooting animals and then watching them suffer. The connection between Chuck’s actions (his shooting at animals) and the affect they have on his character depends on whether certain results obtain (that the animals are shot), so whether Chuck succeeds in worsening his character is a matter of resultant luck. (Similarly, whether he is able to worsen his character will depend on being in circumstances that permit his character formation, and so whether Chuck mitigates his constitutive luck is also a matter of circumstantial luck.)
If this is right, then we face a dilemma: we must either deny that constitutive luck may be mitigated in the plausible wat that I sketched above, or we must accept that there is resultant (and circumstantial) luck. Either way, we’re stuck with moral luck. On one horn of the dilemma, we allow constitutive luck (and more of it than is intuitive, since this horn rejects the mitigation proposal); on the other horn, we must accept resultant and circumstantial luck. I find it rather implausible to reject the common-sense picture of the mitigation of constitutive luck in order to maintain that there is no resultant or circumstantial luck, so my own inclination is to accept the dilemma’s second horn.
One might try to escape the dilemma by applying the response to resultant (and circumstantial) luck sketched above to constitutive luck as well. Applied to Chuck, this would be to say that if Chuck performed an action because of his constitution (over which he had no control) he would be just as blameworthy as if he had some other constitution and had not performed the action.
A problem for this approach, however, is that it fails to take into account our moral judgments about cases of manipulation. Suppose that, without your knowledge, a team of neuroscientists tinker with your brain while you’re asleep, somehow manipulating you into having a bad character just like Chuck’s. Suppose further that, just like Chuck, you act from this bad character and kill your neighbor. According to the above approach, you and Chuck must be equally blameworthy for killing your neighbors, but surely you are surely less blameworthy (if blameworthy at all) for killing your neighbor than Chuck is for killing his neighbor, despite the difference being a matter of constitutive luck. Additionally, according to this approach, even though you haven’t been manipulated, because you would have acted differently had your constitution been different (and since you had no control over this), you are just as blameworthy as Chuck (and the manipulated-you) despite never doing any killing, and that would be a strange result.
Given the incredibly counterintuitive implications of applying this approach to constitutive luck, it would be better, I conclude, to admit that we face the dilemma introduced above and thus that moral luck is inescapable.
Works Cited
Mele, Alfred. 2006. Free Will and Luck. New York: Oxford University Press.
Zimmerman, Michael. 2002. “Taking Luck Seriously,” The Journal of Philosophy 99: 553-576.
Hey Taylor — thanks for the post!
If I’m understanding correctly, your claim is this: if we want to deal with the threat of constitutive luck by saying that we can slowly get rid of it over time through our efforts to form our own characters, then we have to accept that someone can be blameworthy for some of the things they do later in life, even though their being blameworthy for those things depended on their getting somewhat lucky (that their efforts to form themselves were successful). So, this response to constitutive luck requires admitting the existence of resultant luck.
I wonder: does the same argument apply generally to accounts of moral responsibility that appeal to tracing? The idea would be: if you say that it’s possible for someone to be blameworthy for what they do despite lacking control over what they’re doing because their blameworthiness can be traced back to a time when they DID have control and could reasonably foresee that they would end up doing this terrible thing, then you have to admit that someone can be blameworthy for some of the things they uncontrollably do, even though their being blameworthy for those things depended on their getting somewhat lucky (that their earlier controlled actions led to these blameworthy effects).
Or, to put it more succinctly: do tracing theorists have to admit the existence of resultant luck?
Neal,
Thanks for your comment! Yes, my idea is that 1) the threat from constitutive luck can be dealt with by saying that we can git rid of (or, more likely, mitigate) constitutive luck, but 2) this required admitting the existence of resultant (and circumstantial) luck.
I like your suggestion about the implications for tracing accounts of moral responsibility. A natural way to account for the blameworthiness of a drunk driver who injures a pedestrian while driving drunk is by saying that, though the agent lacked control at the time of the injury, the agent was blameworthy for getting drunk (and not taking precautions) and we can trace the agent’s blameworthiness for causing the injury to the earlier action over which the agent had control. But, of course, it was not under the agent’s control whether there was a pedestrian at just such a place and time, and so it is apparently a matter of resultant luck whether the agent is blameworthy for injuring the pedestrian. Given that I’m already inclined to say that constitutive luck can be diminished but only in a way that requires resultant luck, I think the tracing theorist should say the same here (that tracing requires admitting resultant luck).
That said, one could say that, when tracing happens, the scope of blameworthiness increases but the degree of blameworthiness doesn’t. (This would be to take the line Zimmerman does in the paper I mentioned.) This would imply that the drunk driver who injures the pedestrian is blameworthy for doing so (and because she is blameworthy for getting drunk), whereas a parallel drunk driver who (luckily) avoids pedestrians is not blameworthy for injuring anyone, but also the two drunk drivers would count as blameworthy to the same degree despite being blameworthy for a different number of things. I don’t think that this is the right thing to say about this case, and I see my dilemma as a way of trying to motivate accepting resultant luck in this sort of case instead.
Hi Taylor, thank you for this interesting post! I have a question about the Chuck case and manipulation mentioned in the penultimate paragraph. I am curious whether you’d be happy to say that there are some ways of spelling out the manipulation case such that Chuck and Manipulated-You are equally blameworthy. Is the kind of case you are describing (where the Manipulated-You does the killing), a case in which the killing happens before the Manipulated-You starts cultivating the newly acquired character in the relevant ways? Does it make any difference to you whether the Manipulated-You does the killing as some sort of instant agent or the Manipulated-You does the killing after a long period of “soul searching” and embracing and purposefully cultivating values acquired through manipulation?
You also mention that according to the approach you discuss, “even though you haven’t been manipulated, because you would have acted differently had your constitution been different (and since you had no control over this), you are just as blameworthy as Chuck (and the manipulated-you) despite never doing any killing.” I am wondering if you think this is necessarily part of this approach and if so why. (I might be tempted to treat the Manipulated-You killing case and Original-You no killing case asymmetrically).
Marcela,
Thanks for your comments! Yes, in the case in my penultimate paragraph, I was imagining that Manipulated-You goes and does the killing right away but satisfies any non-historical conditions on moral responsibility that you like (e.g., Frankfurt’s conditions). So Manipulated-You doesn’t embark on a process of character-formation, but they have a second-order desire for their first-order desire to kill to be effective, they wholeheartedly endorse the values expressed in the killing, etc. In my view, given that Manipulated-You is just as constitutively lucky as a relevantly similar instant agent would be, there would be no responsibility-level difference between them. The more time they are given to reflect on their values, though, and the more opportunities they have to shape their characters, the less constitutively lucky they are and thus the more morally responsible they are able to be.
Your other comment was about whether we should treat Manipulated-You (who kills) and Original-You (who doesn’t kill) as equally blameworthy. I think that taking the scope/degree response to constitutive luck implies that they are equally blameworthy, since whether the agent kills is a matter of constitutive luck. According to the scope/degree response, then, while Manipulated-You is blameworthy for an extra thing (the killing), Manipulated-You is blameworthy to the same degree as is Original-You (who never even considers the killing). In my view, that is a very strange result (and I do think it is necessarily part of the scope/degree approach, applied to constitutive luck), and so, like you, I want to treat the cases asymmetrically and reject that approach. But that means that there’s no way to escape the dilemma, and so we must admit the reality of moral luck.