Blog Coordinator

« What in the World is Weakness of Will? | Main | What Are Omissions? »



Feed You can follow this conversation by subscribing to the comment feed for this post.

The two cases look relevantly similar to me. In both cases John looks to be clearly responsible (blameworthy) for his lazy choice, at least. There was no impairment to his free agency here to excuse him, so in that - "agential" - sense, he is surely responsible for what he did (or didn't do).

Perhaps there's another sense in which we might say that he isn't responsible for the outcome, in virtue of the other agents waiting in the wings. In particular, if our concern is to "dole out responsibility" for the bad outcome here, then it doesn't seem appropriate to fix on John as the problem. (Though again, this seems equally true of both cases.)

Which of these is the intended interpretation?

You know my answer, I suspect. I am surprised by your claim that most people would think that John isn't responsible in case 1. I haven't seen an outpouring of dissent with regard to Penned in Sharkes in Responsibility and Control (to jog your memories: John is walking along the beach when he notices a child drowning. John is a good swimmer, and (psychologically and physically) capable of rescuing the child. But John chooses to ignore the child’s plight. Unbeknownst to John, an evil marine biologist has penned up a number of hungry sharks; were John to attempt a rescue, the marine biologist would release the sharks and John would be devoured).

Perhaps, though, you understand the strong man as an actual feature of the case, not a counterfactual? In any case, may I suggest a poll? Is John responsible for the omission in cases 1, 2, and 3 (where 3 is the Fischer and Ravizza case)?

I like these cases. I've never thought about Frankfurt-style cases in terms of omissions.

My initial reaction is to hold John responsible in both cases. In the first case he decides not to bother saving the child. He's lucky that intervention ends up not to be necessary. But he's still responsible for his character, or something like that. A drowning child doesn't even motivate him to try to save the child in case 1 or to even consider saving the child in case 2.

If we modify the cases such that the child drowns, my guess is that most people would hold John responsible for not even trying to save the child in case 1, and, while there would be some divide on case 2, John doesn't even give serious thought to saving the child, and he's responsible for that, whatever that amounts to.

That said, my intuitions are usually atypical and I haven't given much thought to omissions. I could be in a minority here.

At the very least, this set of cases is very cool because to the extent that John is responsible, if at all, it's for something other than an action.

Anyone out there who thinks there IS a difference between case 1 and 2 in regards to John's responsibility?

"Most people, I think, will say that in case 1 John isn't responsible for not saving the child. There's disagreement about case 2."

Randy, is your question about 2 whether the individual is responsible for *not saving the child* or for that or any other omission in the story (eg, for not "giving serious thought" to saving the child")?

We can fill in the cases so that, in 1, John is responsible for deciding not to bother, and he's responsible for not even trying to save the child; and so that, in 2, John is responsible for not even giving serious thought to saving the child. My question is whether, in each case, John is responsible for not saving the child.

Neil, sure, a poll. But you must give reasons for your answers (limit 140 characters).

By the way, I don't think John is responsible for not saving the child in Penned in Sharks. The case is structurally similar to Flat Tire (Resp. & Control, p. 126). Imagine that the mafiosi would have stopped you from robbing the driver by unleashing attack dogs. Surely your responsibility for not committing the robbery doesn't hinge on whether they would have shot you or sicced the dogs on you.

John isn't responsible for failing to save the child in case 1. There's no relevant difference between cases 1 and 2. So he isn't responsible for failing to save the child in case 2. Convincing, right?

And for what it's worth, I'm with Randy on Penned in Sharks: not responsible.

I used to think that the two cases were different but now I think they're the same. Omission cases are complicated in ways that commission cases are not, which might explain the differing intuitions. As Fritz hints, in most cases of omission there is clearly something that the agent is responsible for and the agent had alternative possibilities wrt to it. For instance, in the second case, John is responsible for not trying to save the child (or for not even giving it serious thought, as Fritz suggests). Thus, he is blameworthy for something, and it is something that is related to the death of the child. I don't see a corresponding blameworthiness in the commission case.

Suppose we make it explicit that John can't even try to save the child? Better: Suppose we change the example. As it is now, John can try to give it serious thought ("Had John given serious thought to saving the child ...") but suppose the device prevented even this from happening. (Don't ask me how!) Then things would be just as unsettled as in the commission case.

Thanks for this interesting post, Randy.
Dan: if I may be so obnoxious, check out an extended discussion of Frankfurt-style omissions cases in my (and Mark Ravizza's) book, Responsibility and Control.

Also, about Penned In Sharks, my view is that we shouldn't just consult our unreflective intuitions about such a case. Mark Ravizza and I certainly conceded that, if that's all one does, one won't distinguish Sharks from Penned in Sharks. But our view was (and still is) that one has to consider a whole range of cases, and one has to be consistent across this range. That's why we came to the conclusion we did about Penned In Sharks. Note: just about any approach to these cases will have to bite some bullet or another.

Having said all of that, I admit that the bullet is uncomfortable, and recent work by Philip Swenson (unpublished) and Matt Talbert (unpublished) have caused me some indigestion in this regard. So I honestly am not sure Mark and I were right.

Okay Randy, my reasons. John has given us what I take to be the best explanation of how FSCs work. The agent is responsible because in the actual sequence the state of affairs is sensitive to his behavior. Black's intervention is counterfactual, not actual sequence. The lesson here is a straightforward extension of the Martin/Lewis analysis of finkish dispositions.

We ought to think of omissions in precisely the same way. The strong man (case 1), the neuroscientist (case 2) and the sharks are all alternative sequence obstacles, not actual sequence. In the actual sequence, the relevant states of affairs are sensitive to the agents' behavior. Verdict: the agent is responsible.

Anyone who dissents must give us a better story wrt FSCs then John has done. Couple of supporting points:
1. There ought to be a presumption in favor of treating actions and omissions alike, given that standard justifications for distinguishing between them (eg. that causes are necessarily positive forces) are apparently weak (see Jonathan Schaffer).
2. There ought to be a presumption in favor of John's view, on the grounds that he's John Martin Fischer.

"My question is whether, in each case, John is responsible for not saving the child."

He's not responsible for that in either case. He may well be responsible for other things in one or both case, depending on how further details are filled in.

No bullets to bite I hope.

I come to this question with the benefit of having just studied Fischer & Ravizza’s “Responsibility for Omissions” (1998, repr 2006). I am sorry to report that I still cannot fathom the intuition that would excuse John (1), provided that we accept that he had a public or Samaritan duty to perform a rescue. The only thing that would excuse John’s failure to accomplish a seemingly easy rescue is a determined effort on his part to rescue the child that was actually thwarted by someone or something he could not overpower.

Were John (1) a California Beach Lifeguard—they are sworn peace officers—he could face criminal and civil liability for such an egregious dereliction. I try to imagine this lifeguard offering a “strong man” defense: if I had bothered to try to rescue the child, I would been thwarted. No jury would be sympathetic. Notice that there are already settled cases in which we hold someone responsible for failing to do his duty though he could not have done it. The drunken lifeguard who passes out minutes before a child drowns off his station could not have saved that child, but he will be held responsible for failing to save the child. Suppose John(1) had in fact paid the strong man to restrain any would-be rescuers including himself. Then, for that reason, he could not have saved the child, but as in the lifeguard case, his disability or incapacitation would be self-inflicted and very clearly culpable. So we already accept that “could not have performed his duty” is not sufficient to excuse, where the incapacity is self-caused, and also where there is no credible attempt to do his duty.

I don't see a difference between case 1 and case 2. I agree with Philoponus that John has a duty to attempt to rescue the child in both cases. I, for one, am not comfortable with consequentialist theories of ethics, as it is all too easy to know what is good or bad with 20/20 hindsight. (Even this I doubt in many cases.) The problem for John is what does he know in the moment. The neglect of his duty is immoral in the moment because, if I dare say, he is in an immoral state of mind. We shouldn't base our ethics on hypotheticals where we can manipulate the future (as in this thought experiment). We need a theory of ethics that we can actually live with and perform without having to know all the consequences of our actions. After all, we are limited beings with limited knowledge, and to expect more than we are capable of is too much to ask.

About the RESPONSIBLE IN BOTH CASES view: Consider a variation of case 1, in which, instead of the strong man, what stands between John and the struggling child is an impenetrable glass wall (so high he can't get over it, so wide he can't get 'round it, etc.). It's as if John is watching the child drown in an aquarium.

What's common to this case and the original case 1 is that, no matter how hard John might have tried, he wouldn't have been able to save the child.

The glass wall, like the strong man, doesn't actually interact with John. But I can't see how that makes John blameworthy for not saving the child.

One thing more to notice about case 1 (as well as this variant): John's not trying to save the child doesn't cause the child's death. Accept, if you like, that omissions can be causes. Still, this one doesn't cause that outcome. (Whether the omission takes place makes no difference to whether the outcome ensues.)

About the NOT RESPONSIBLE IN EITHER CASE view: I think this might be right. I'm afraid I have a paper in press arguing for a difference between the two cases (with respect to responsibility for not saving the child), but I'm no longer confident about that view.

Here's one difference between case 1 and case 2: In the latter, but not the former, if John had decided to save the child, he would have been able to do that; nothing at all would have stood in his way.

Further, in case 2 John decided on his own not to save the child, and omitted on his own to decide to save the child; nothing in fact made him do either of these things.

Do these features give us enough to say that in case 2 John is responsible for not saving the child? I'm not sure that they do. After all, another feature of case 2 is that John couldn't have decided to save the child!


You say that John's failing to try doesn't cause the child's death in case 1. But what about in case 2? Is John's not trying to save the child a cause of the child's death in that case? If so, wouldn't that be an important difference between the two cases?

In so far as we seem to be taking sides, I will side with Dan H. and suggest that we should hold John responsible for not trying to save the child. There certainly is a relevant difference between trying to save the child and failing (and he would fail in all the mentioned cases) and not trying to save the child.

Take away the thought experiments (interesting though they are) and simply remember concrete examples of someone trying and failing, i.e. a firefighter during the 9/11 attacks trying to save someone and not succeeding. Had the firefighter not even tried I think we would have been critical of his/her omission as a failure to perform up to what was promised by being a firefighter. I think that this 'duty' to act carries over to a non-professional who has the skills to intervene when doing so might help someone, but fails to utilize them thereby allowing someone to suffer or die who could have been helped. The example of the people watching the person bleed to death after he save a women from harm comes to mind as an example.

I think this is ultimately a question of where we focus when we in fact assign responsibility and I have a tendency to think we should focus on the mental states of the agents being evaluated.

Anyway, interesting post.

I suppose that if not-doings can be causes, then, as Justin suggests, in case 2 John's not trying to save the child is a cause of the child's death. So perhaps this is a further difference between cases 1 and 2. But does it matter, since in case 2 John can't try to save the child?

John A.: In case 2, if John had tried to save the child, he would have been able to; nothing at all would have prevented him.

We can grant that, in case 1, John is blameworthy for NOT TRYING TO SAVE THE CHILD. The question is: is he also responsible for NOT SAVING THE CHILD. This isn't a question about whether there's something he's to blame for, but about what he's to blame for.


Wouldn't John at least have to form the intention to attempt to save the child in case 2? Otherwise, what cue would be given to the neuroscientist for him to know to mess with John's brainstates. I still think John being fully aware of the drowning child and not even beginning to form the intention of saving her, that this in itself is immoral. Although, only John and the neuroscientist are aware of this immorality, these two being the only ones with access to or knowledge of John's brain states. It all begins with the right intentions...

However, I do now see how case 2 is different from case 1. In case 1 John can actually attempt the rescue after forming the intention. In case 2 he can still form the intention, but he can never attempt the rescue, as he would be thwarted by the neuroscientist before he could act on the intention.

Frankfurt discussed his distinction of "personal failures" from "impersonal failures" in response to PvI's example of the unknown-dead-phone-not-used-to-call-police for crimes observed but ultimately unreported (in Frankfurt's The Importance of What We Care About)--his argument is that close counterfactual worlds with dead phones tend to lessen responsibility (impersonal failure), whereas other (personal) situations of failure due to simple neglect enforce responsibility even though unneeded counterfactual forces were present that would have acted to ensure the failure. (I responded in a Sorites piece that close-worlds analysis of Frankfurt's counterfactual claims seems to equivocate on what "failure" means, potentially blurring metaphysical and evaluative uses. It could have been a clearer criticism but I stand by it.)

In any case Frankfurt might have by his distinction of personal and impersonal failures one out here: as proposed by the original two scenarios on this thread there is always an intentional agent--the strong person or the neuroscientist--who is obviously intent--and I mean INTENT--to bring about the drowning. By almost anyone's assessment this must lessen John's accountability--he is a pawn in some grander scheme whereby the controller intentionally is responsible. This makes the two scenarios above more like the problem of evil in the philosophy of religion than matters of mere mortal accountability. This is why I think that thought-experiments about manipulation are somewhat better stationed to pump intuitions about FW and responsibility if the manipulative circumstances are non-intentioned. But sure as shootin'--if that condition is satisfied, that will just start the luck pumps a-pumpin'. So I wonder if manipulation arguments are either way lose-lose in terms of getting to what Frankfurt calls personal failure.

So what was your reason for the view in your in press paper, Randy? Here's one stab: the counterfactual dependencies between John's deciding and the event are the same in both the actual sequence and the counterfactual sequence in 2; whereas the chain of counterfactual dependencies are different in 1.

Aren't "NOT TRYING TO SAVE THE CHILD" and "NOT SAVING THE CHILD" the same omission in these stories? It's not like he tried to save the child and failed (perhaps for reasons that leave him morally culpable). He simply did not act to save the child, and whether we describe that as "not trying to save the child" or "not saving the child" is largely a matter (I think) of how high we rate his chance of success had he tried. Is it even possible that he could be morally responsible for the omission under one description, and not morally responsible for the same omission under a different description?

I've always thought that Frankfurt's basic insight was that the actual sequence rather than counterfactual possibilities is what's important for ascriptions of responsibility.

So, if you want to know whether an individual is responsible, and for what, simply look to what actually took place, despite what conditions may have existed that constrained counterfactual possibilities.

Since one could omit to do something (though guaranteed to by a intervener poised to intervene), just as one can act in a particular way (though guaranteed to in the usual Frankfurtian way), I would think the original question (whether there are Frankfurt cases for omission) ought to be answered in the affirmative.

I should note that I agree there are interesting puzzles raised by cases of omission. I happen to think these are raised mostly by figuring out the relation between omissions, causation, and responsibility - and likely has more to do with general metaphysics than anything peculiar to moral responsibility.

Still, if we follow Frankfurt's basic insight, in both Case 1 and 2 (and the variations I've seen), there will be something John is blameworthy for. And it's unclear to me why I should wonder about John's blameworthiness for not saving the child when it's clear enough he can be blameworthy for deciding not to try to save the child. Certainly, of the two possible objects of responsibility, deciding not to even try to help evinces a disrespect to the child that doesn't accompanying all failures to save (for instance, when he's prevented or when he simply fails because the water's too rough or the child has stopped breathing for too long).

Randy (if I may)
I must have misunderstood case 2: I thought that John could save the child, in the sense that he has the knowledge and skills and seemingly the ability to do so, but chose not to act to save the child. Had he decided to save the child the scientist would 'have changed his mind.' When I said he would fail in both cases it was relative to had John chosen to act to save the child he would have been thwarted in his attempt by outside forces. Is this not what makes these cases Frankfurt-Style omissions?

I agree "This isn't a question about whether there's something he's to blame for, but about what he's to blame for." To be clearer, I hope, I think that what we hold John responsible for in both cases is his failure to act to save the child, not the outcomes of his attempt. This would certainly seem to be the case if, in both cases, there were no outside forces that would act to stop him from saving the child. We can imagine a case where John acts to save the child but fails, he drowns too. I think that it would be misusing language to blame John for the death of the child if he failed for some reason outside his control - hence why the strongman and scientist in your cases are not relevant in how we assign responsibility and blame. In your cases, I think that the reason we assign responsibility and blame is just because John, in both cases, decides not to act. It seems to me that the information after the 1st sentence in each case is irrelevant to our assigning responsibility and blame to John. This information would only become relevant if John tried and failed because he was stopped by forces outside of his control.

I think that we use responsibility as a criteria for assigning blame (or praise, as the case may be)based not on outcomes, but on the intentions and/or actions of the agent. There seems to be empirical data to support this, e.g., Knobe's experiments dealing with the CEO's decision to improve profits while harming/helping the environment. The two cases you present could be the subject of an experiment and possibly should be because there are some empirical claims being made, e.g., "Most people, I think, will say that in case 1 John isn't responsible for not saving the child. There's disagreement about case 2."

Anyhow, interesting post and discussion.

Mark: in case 1, John could have tried to save the child, but he couldn't have saved the child. That's a reason not to identify these two omissions.

Matt: as the cases are set up, it seems that John is responsible for not saving the child just in case he's responsible for the child's death. The latter might matter; there might be a question of legal liability. It seems to me to matter morally, too. It might matter to John.

Neil: The rationale I offer in the paper is pretty much what I said in an earlier post here: in case 2, John decides on his own not to save the child, he omits on his own to decide to save the child, and had he decided to save the child he'd have been able to do it. It's this last feature that's missing in case 1. That's what seemed to me to make the difference.

Matt--I agree about there being a tension between what FS-omission examples are supposed to show by emphasis on the actual sequence and the use of counterfactuals, but Frankfurt in the book cited above clearly wishes to clarify the importance of that sequence by counterfactual evaluation. The overall strategy is familiar enough: Frankfurt wants to pinpoint personal failure (as opposed to impersonal) by something like a counterfacual analysis used to assess dispositions. The case involves a failure (say safe driving) by looking left in an overdetermined situation enforcing that action where the external overdetermination wasn't needed. I recall Frankfurt saying something like "If x had not turned his eyes left, he would not have failed." That captures at least one sense that the failure was personal--it was due to x's action and x's action alone in the actual sequence (I believe he said the failure was "constituted" by x's action alone). However, the use of the counterfactual introduces questions of how we are to understand the "not failing" under counterfactual circumstances. If overdetermination still exists in the closest world, then it is simply false that x wouldn't have failed (in *some* sense--thus my concern whether the evaluative sense of "fail" gets mixed up here with the metaphysical action sense of "fail"). If x's counterfactual action is evaluated in a more ordinary world without overdetermination, then it is true (and more clearly in both the evaluative and metaphysical senses of "fail") but seems to rely on something like an availability of PAP (compared with the actual world) to allow that. So that tension I alluded to above is definitely there due to the introduction of the counterfactual to explicate the sense of personal failure. (I just overcame my inherent laziness enough to look up the page number in TIOWWCA--it's appropriately enough p. 101.)

Good stuff, all. This is for John Alexander, Mark Young, and maybe Noah. My own hesitance to conclude that John is responsible for not saving the child is rooted in the intuitive pull of "ought implies can". I agree that it is appropriate to blame John for something in these cases... but not for not saving the child. After all, he couldn't have saved the child. To blame him for not doing something he could not have done would be inappropriate given OIC.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan