On behalf of Carolina Sartorio, who's writing a paper on a new kind of luck, I'm posting, with her permission, her core motivating cases in order to see what people's intuitions are.  In each case, I am giving Carolina's report of what our intuitions likely are, but I do not include her analyses.  Do you share the basic intuitions?  Why or why not?

CASE 1: I want an explosion E to occur. I have good reason to believe that
pressing button A will trigger an explosive that will result in E. I press A and
E occurs.

CASE 2: This time three buttons (A, B, and C) need to be pressed for the
explosive to be triggered and thus for E to occur. Two other responsible
agents independently press B and C while I press A (each of us knew about
the buttons, had good reason to believe that the others would press the other
buttons, and acted with the intention that E occurs). E occurs.

We tend to think: (1) I am significantly less responsible for E in CASE 2 than in CASE 1.

Compare, in particular, CASE 2 with:

CASE 3: Again, there are three buttons and three fully informed and
responsible agents, but this time pressing any of the buttons would be
sufficient to bring about E. I press A, and the two other agents press B and C.
E occurs.

In contrast with CASE 2, here we tend to think that each of us is “fully” responsible
(as responsible as we would have been if we had acted on our own).

But now consider one last scenario:

CASE 4: As in CASE 2, three buttons need to be pressed for E to occur.
However, this time I have good reason to believe that the other two buttons,
B and C, will be automatically pressed in the next few seconds, as a result of
some purely mechanical, irreversible process. I press A and E occurs.

It seems to me that we want to say:(2) I am, at best, slightly less responsible for E in CASE 4 than in CASE 1.

## 14 Replies to “Sartorio’s Intuition Pumps about Responsibility”

1. My own intuition is that A, B and C are each fully responsible in CASE 2, since each had complete power to prevent the explosion; and that there is weak but not vacuous argument that each is slightly less than fully responsible in CASE 3, if each had good reason to believe the others would press their buttons.
Interestingly, this depends on an impression that the explosion is a bad thing; if instead we talked about starting a life-giving IV, I’d have the significantly-less-responsible intuition in CASE 2. Furthermore, the argument for less than full responsibility in CASE 3 suddenly seems stronger.
I have no consistent defense of these intuitions!

2. David Sobel says:

My quick intuitive reaction agrees with all the predicted reactions.

3. Jussi Suikkanen says:

Hi
I would prefer if the cases were described in a more uniform way. The description of CASE 1 leaves it open whether the pushing of the button really causes E even if I have good reasons to believe this whereas in CASE 2 the causal connection is made explicit.
It would also help to clarify whether we are talking about causal responsibility or moral responsibility. Also, as Yarrow points out, it would help here if it was specified what explotion E is and what it caused. Is it something good or bad or neutral? Do people die, for example?
I think this might make a difference for the intuitions. Imagine that E kills 10 people and that this was known in advance. We can then ask was the agent responsible for killing ten people. In this case, I would want to say that the agent is equally blameworthy for the deaths in all cases. I don’t think it is an excusing factor here that other people too contributed to the outcome. That doesn’t make the others any less blameworthy though.

4. PH Wong says:

I do not really share the intuition that the agent is less (morally?) responsible in CASE 4 than in CASE 1. Perhaps, it is just a matter of description – but does having extra causal steps (i.e. buttons B & C) really matter if there are obvious causal relations between buttons A, B and C? Say, If one provides a finer description of CASE 1, then probably there are various intermediate “mechanical” stages/steps from A to E as well, does it then render CASE 1 the same as CASE 4?

5. Andrew Sepielli says:

I find it difficult to form intuitions involving concepts that, IMHO, fit uneasily into moral theories that I find plausible, and I’m afraid “moral responsibility” is one of those concepts. Nonetheless, re: cases 1 and 2 — I suppose I’m inclined to say that I’m not less responsible in 2 than in 1 (I reasonably and truly believed that I’d cause E given the background conditions in both cases, and insofar as E is bad, I’m equally culpable in both), but perhaps that I bear less of *the* responsibility.
My judgment about 2 is influenced by my judgment about 4, but also by my judgment about cases where other agents act non-contemporaneously. Suppose the buttons are on an ancient Roman doomsday machine, and I have good reason to believe that B and C have already been pressed by Caligula and Nero. Then I can’t see a difference in how responsible I am between this case and 2, nor between this case and 1.

6. Ben says:

To echo Andrew, I’m not sure I know what “responsibility” per se is, so my intuitions about it are much hazier than my intuitions about legal responsibility, blameworthiness, and so on. But when I flesh out the cases to bring in these other things, the asymmetry seems to me to either go away or become uninteresting.
For instance, suppse I wanted the explosion to occur because it will destroy Jones’s beloved vintage Ford, and I know I’ll be really amused by both the spectacular destruction of the vehicle and Jones’s ensuing sorrow and outrage. (And suppose in Case 2 all three of us want this.)
Now, it seems pretty clear to me that (for example) it would be just as fitting for Jones to resent me in both cases, to just the same degree. The only difference is that in the second case it would also be fitting for Jones to resent the other two agents just as much. If there’s an intuitive asymmetry at all, I’d think it has to do with the fact that it’s easy to imagine that in the second case, I might be subject to some kind of mob mentality or coordination pressure that makes my action less purely an expression of malice. But if we bracket these possibilities, the sense of asymmetry goes away.
The same goes for other ways of fleshing out “responsibility,” like desert. Suppose in the first case that justice demands that I be sentenced to six years in jail. Obviously, I do not get to lighten the burden by bringing other people into the crime; in the second case, we get six years apiece, not two.
In fact, the only dimension of responsibility that occurs to me as plausibly exhibiting an asymmetry is the obligation to compensate. If, in addition to jail time, justice demanded that Jones be compensated financially for emotional damages and the loss of his car, it’s obvious that in Case 2 I would only have to pay a third as much. Maybe this helps explain away the intuition.

7. Carolina Sartorio says:

Hi all,
Thanks to Dave for posting this and to all who have shared their reactions and thoughts!
What I’m thinking is this. Resultant moral luck scenarios seem to suggest that responsibility tracks, among other things, one’s *actual* causal contribution. Thus, for example, if I want to cause the death of my enemy by shooting him but a bird gets in the way and deflects the bullet, I am not responsible for his death (even if the death occurs for other reasons) because I don’t causally contribute to it. This is so regardless of how much I wanted to cause the death, or how much I wanted it to happen.
But then, if responsibility tracks one’s actual causal contribution, and if my contribution is considerably *less significant* in Case 2 than in Case 1 (or Case 3), as I am assuming, it would appear that I’m less responsible in Case 2 than in Case 1 or 3. Again, this is so regardless of what I wanted to be the case. Perhaps I wished I could have been in Case 1 or Case 3 (I’m that evil!), but the fact of the matter is that I am not, and this seems to make a difference to my responsibility.
At any rate, it seems to me that this is a quite natural (although perhaps misguided!) take on these cases. Of course, if one were to simply reject the existence of resultant luck (a possibility that I want to leave open at this point), then there would be no motivation for sticking with this distinction. But otherwise it seems that the motivation is there.
As I see it, the puzzle arises when we move to Case 4. For in that case we just don’t feel any strong motivation to judge that the agent is (much) less responsible. But this is puzzling because, of course, the only difference between Case 2 and Case 4 concerns whether other moral agents are in charge of the other buttons, and it seems crazy to think that this could matter to my responsibility.
So there is a puzzle, I think, and I’m not sure what’s the best way to resolve it. But I think it helps to bring out some important questions about moral luck and the nature of responsibility.

8. I too find it difficult to have clear intuitions about the cases just in terms of the notion of responsibility. If we’re talking about my responsibility for E as E’s attributability to me, then I’d say that, in Case 1, the result E is more attributable to me than in Case 2. At least, this is so if we read the cases in such a way that, in 1, I play a greater part in bringing E about than I do in Case 2. (Of course, the action of pressing the button A is no less attributable to me in 2 than in 1.) In Case 3, E is overdetermined, and I am unsure about what to say about attributability of E to me here. For Case 4, I’d say that E is just as attributable to me as it was in Case 2. I guess this indicates that I think that attributability tracks one’s actual causal contribution, in some roughly proportional way.
If we’re talking about my responsibility for E as my accountability for E’s occurrence, however, then I’d say that I’m equally accountable in all the cases. At least, this is so given that I actually play the expected part, in each case, in bringing about E. This is because, in each case, what happened was what it was reasonable for me to believe would happen given my causal contribution to bringing about E — a contribution that I actually manage to make.
Does this mean that accountability doesn’t track the magnitude of one’s actual causal contribution in any way? Not exactly. Suppose I’m one of three snipers attempting to assassinate X. If I miss while the others hit their target (perhaps I hit a bird that flies in between me and the target, as in Sartorio’s scenario in her response above), then, even if the others’ bullets are enough to kill X, I’m not accountable for X’s death, nor is the death attributable to me. (Though I am accountable for a particular action that is attributable to me — attempted murder.) My accountability for X’s death here does seem responsive to luck — the kind of luck involved in me (thankfully) failing to play any part in bringing about X’s death. If I fail to play any part in bringing about X’s death, then I’m not accountable for X’s death (though I might be fully accountable for some related action, such as attempting to bring about X’s death).
If this is right, then there is a way in which accountability is responsive to actual causal contributions. But notice that the following two theses are compatible: (a) if I make no causal contribution to bringing about an outcome, then I am not accountable for that outcome; and (b) if I make a partial (and intentional) causal contribution to bringing about an outcome that I have good reason to think will be brought about given this contribution, then I am fully accountable for that outcome. So it doesn’t follow from the lack of accountability in sniper/bird style cases that accountability tracks the magnitude of one’s causal contribution in a proportional way.

9. Like Jussi I got a bit confused by the way the cases are described. In case 1, I believe that pressing A will cause E. I press and E occurs. But it makes a big difference whether E occurs, contrary to what I believed, causally independently of my pressing A (1a) or whether, as I believed, E is caused by my pressing A (1b). In 1a I am not responsible for E. In 1b I am fully responsible for E. (Assuming no relevant exempting or excusing conditions are complicating matters.) In case 4 again there is ambiguity between 4a where I press A and E then occurs independently and 4b where E occurs as a result of my pressing A. In 4a as in 1a I am surely just not responsible.
I don’t think there’s much difference between 1b, 2, 3 and 4b. It seems to me an important feature of shared responsibility that sharing responsibility is not at all like sharing your lunch. If I share my lunch with you I have less lunch. But sharing responsibility is not at all like that. Where responsibility is shared it is not diluted. If I single handedly beat you to death, I am fully responsible for your death. If me and my thuggish friend Mike together beat you to death, then Mike and I are each fully responsible for your death. Does it make a difference if the injuries we each respectively inflict would each have sufficed alone to kill you – analogous to case 3 – or if alone either would not have sufficed but they suffice together – as with case 2. I can’t see how, assuming in the 3-type case they are inflicted simultaneously. (Of course in case 3 if I press button A after B and C have been pressed when the explosion has already occurred I am not responsible at all. That is just like case 1a.) Does it make a difference if we are engaged in joint agency – having conspired to beat you to death – or independently (we have both independently decided to beat you to death and happened to show up at the same time). Again I don’t really see why it would. In all these cases (except the 1a type case) I think I am fully responsible for your death. As to 4, again full responsibility seems involved. If you happen to be poorly today and that serves as a causal enabling condition for me to beat you to death because, had you been well, you could easily have fought me off, that hardly diminishes my responsibility. Or if I am only able to push you off the cliff because of the strong wind that happens to be blowing, again no difference. In 4 the pressing of B and C is surely just like your illness or the wind.
(I might stress that in the preceding paragraph I fall into my frequent habit of using “you” as an indefinite pronoun. I hope no one reading it will suppose me to be engaged in airing murderous fantasies about him or herself personally!)
I’m not sure I fully understand the notion of magnitude of causal contribution.

10. I’ll report that I don’t really have the asymmetrical reactions to the various cases. Also, like Ben, I did think that duties of reparation might be less in joint causation cases just because reparation requires only removing the harm and three people can jointly do that more easily than one. (It might be worth thinking about whether one has a duty to compensate more if one of the others does not contribute.) Causation is always operative within a set of background conditions that enable one’s actions to have their effects and I’m not sure that the predictable actions of others are much different than nonhuman enablers.

11. Carolina Sartorio says:

Thanks to all again!
I find it a bit surprising that so many of you find the notion of “diluted” responsibility so implausible. Imagine the following case: I am evil and hear, together with other evil people, about starving children who are in urgent need of food supplies. I send a batch of poisoned food really hoping to cause those children’s deaths. The other evil people do the same. In the first case, I get “lucky” (given what I want) and the children end up eating all and only the poisoned food that I send and die as a result. In the second case, only a poisoned breadcrumb in the food I send makes it to them; the children die mostly as a result of the poisoned food that the other people have sent (still, given that my breadcrumb makes some contribution, I also causally contribute to their deaths). Isn’t it natural to believe that I’m more blameworthy for their deaths in the first case than in the second case? Assuming that we grant that my responsibility can be “preempted” when factors outside my control result in my not making a causal contribution that I was hoping to make, isn’t it natural to believe that a big chunk of my responsibility can also be “preempted” when factors outside my control result in my making a much lesser contribution than the one I was hoping to make?
If it helps (and the first comment by Yarrow in this thread suggests that it might, for some people), switch to a “good” case. Imagine this time that I am a good person (yeah!) and that I (together with other good people) send a batch of good food hoping to prevent the starving children’s deaths. In the first case, I’m lucky and the children survive because of all the food I sent. In the second case, I’m not as lucky and the only contribution I make is given by one (good) breadcrumb. Isn’t it natural to believe that I’m more praiseworthy in the first case than in the second case?
Note, in particular, that claiming that I’m fully praiseworthy in the second case would seem to make the possibility of resultant luck even *more* puzzling than it already was. For note: How implausible does it seem to suggest that I can go from being not at all praiseworthy to being fully praiseworthy with just one breadcrumb? It seems much less implausible to suggest that there are gradations of responsibility that track gradations of causal contribution. (Again, I’m not arguing that this is the right view of responsibility, but only that it’s a quite natural extension of the standard concept of resultant luck and, in particular, of the thought that responsibility tracks actual causal contributions.)

12. I’d concede that sometimes it may make a difference whether contributions are independent or a matter of joint agency Suppose Clara is a schoolgirl who is driven to suicide by the abusive bullying of her 50 classmates. But suppose what each classmate does is, considered in isolation, fairly minor, little acts of spiteful unkindness. It is their cumulative effect that is so destructively intolerable. One way this might happen is that they have conspired together to drive her to suicide and each does their disagreeable small bit. Another is that each just sets out to do a small unkindness intending no more harm than a small unkindness alone would normally occasion. In the first case I think each is fully responsible for Clara’s death and there is no dilution of responsibility. In the second clearly each is blameworthy to a lesser degree.
But “less blameworthy” is ambiguous. I am less blameworthy if I steal your pencil than I am if I kill you. But of course that is not because I have a lesser degree of responsibility but just because what I do is less wrong. In that way there can be degrees of blameworthiness with there being degrees of responsibility. In the case where Clara is the victim of 50 independent bits of small unkindness, what each classmate does is less wrong but they are each fully responsible for doing that. Each is fully responsible for the small part they play but playing that small part, while nasty, is less nasty, and so less blameworthy than had they intended to kill her.
Carolina’s poisoned breadcrumb case is like this with an added complication. Here we might imagine each classmate has resolved single-handedly to drive Clara to suicide by, say, sending her 50 spiteful emails. But in every case 49 of the emails go astray due to some technical glitch and only one gets through so the actual harm each does is the same as in the previous case. Here why not simply say this? Each is fully responsible for their attempt to make Clara kill herself just as they would have been had their email been the only one received. And each is fully responsible for their actual contribution to Clara’s death. Though the making of that actual contribution is less blameworthy, because less wrong, than what they would have done had they single-handedly driven Clara to her death as they each planned.

13. Carolina Sartorio says:

Thanks, Jimmy, for your interesting thoughts. I agree with you that in your Clara case each classmate is fully responsible for their actual contribution, and that this is less than they had intended, and that this is a sense in which they are each less blameworthy than they would have been otherwise.
But this doesn’t eliminate the question: What about the outcome of Clara’s death (or of Clara’s killing herself)? How responsible is each classmate for *that*? That’s the question I’m interested in. And note that one possibility is, indeed, to say that each classmate is less than fully responsible for that outcome, precisely *because* the actual contribution they make (and for which they are, as you say, fully responsible) is less than it would have been otherwise.

14. I haven’t read all of the discussion, but contrary to what was expected, the only case in which I thought that the agent might be less responsible for E than case 1 was case 3. This was the only case in which the agent’s contribution was superfluous.