(1) Are there actually any (secular) deontologists? Or is deontology just something that a subset of professional philosophers continue to pretend to believe in because that was the team they picked when taking Ethics 101, it would be embarrassing to change their mind in public, and anyway they need to continue to pretend to believe in deontology in order to keep their jobs as bioethicists? (This thought brought to you courtesy of the "Treating Persons as Means" page on SEP, which gives the strong impression that the author can't find a way to make the concept intelligible but doesn't want to admit it.)
(2) >Insofar as one can distinguish narrowly “act-directed” from broader “state-directed” motivations, the latter have greater normative authority.
("More of a comment than a question...") This is the entire point of your disagreement with deontologists, so simply stating it like this appears to be begging the question. (This criticism applies more to this post than to your draft, which considers act-directed motivations at somewhat greater length, even if deontologists may still feel the draft doesn't provide adequate support for this claim.)
(More substantially, and with an actual question...) Nye, Plunkett & Ku seem to be arguing that deontology can be grounded in fittingness considerations — the words "fitting" and "fittingness" appear on practically every page of their paper. By contrast, you seem to use the language of fittingness here only very reluctantly: you don't use it at all in these posts, and in your draft you use it only when replying to them directly and quickly move on from talking about "what is fitting" to talking about "what matters" or "what one ought to do." To what extent do you think your divergence from Nye, Plunkett & Ku should be characterized in terms of divergent views on what is fitting, vs. divergent views on the normative force of fittingness considerations?
(3) >All we can say is that we ought to feel (irreparably) torn, which leaves us entirely lacking in practical normative guidance.
"We ought to feel irreparably torn, and lacking in practical normative guidance" seems to me an entirely accurate description of the human condition, quite independently of our views on deontology. Is there any reason to believe it is *wrong*, as opposed to just an unappealing conclusion for a certain variety of moral philosopher?
Incidentally, I think "ipso fact" on p. 22 of your draft should be "ipso facto." (An ipso facto typo?)
(1) About a third of philosophers, according to the PhilPapers survey -- https://survey2020.philpeople.org/survey/results/4890 -- or likely an outright majority if you include virtue ethicists (who seem likely to also endorse deontic constraints?)
(2) I don't think that's necessarily a point of disagreement at all. Deontologists can (and, I argue, should) offer a distinctively deontologically-flavoured account of the fitting "state-directed" motivations, such that, e.g., they should prefer the state of affairs in which Protagonist *doesn't* kill, even though this results in more killings overall. But yes, the actual arguments here appear in the paper -- this post is just offering a summary of the key results!
I'm actually very happy to formulate all the issues in terms of fittingness (I've a 2012 paper arguing that this is a good way to clarify ethical debates), I just refrain from doing so here because many find it a less familiar concept. That said, I'm not sure how best to characterize the heart of my disagreement with NP&K, because it partly seems to come down to their not taking sufficiently seriously their own Motivation-Action Principle. But perhaps some also comes down to a disagreement about *which* fittingness facts matter (and how): I say that the facts about preferability (or what it's fitting to care about, or desire) have a distinctive kind of normative force -- they warrant *caring* about -- in a way that doesn't necessarily hold true of other fittingness facts.
(3) Fair question. One thing to say is that, like nihilism, a view that offers no practical guidance may be safely ignored for practical purposes. In response to moral uncertainty, it seems that what we (meta-expectably) should do can only be influenced by views that yield directional answers of some sort.
(1) Interesting. Of course, I realize that a significant fraction of philosophers claim to be deontologists, but my vague (outsider's) assumption is that this is mainly because the field has a consensus that there is value in maintaining this tradition of thought — sort of like music departments continuing to hire people to study and teach baroque counterpoint, even if essentially nobody really listens to or makes that kind of music any more. I am strongly in favor of music departments continuing to study and teach baroque counterpoint, and I can accept that perhaps it is worth continuing to study and teach deontology for the same reason. (And Kant would have heard plenty of baroque counterpoint — perhaps there is a connection here?) But I'm not sure deontologists would appreciate me thinking about them in this way...
(2) >"they should prefer the state of affairs in which Protagonist *doesn't* kill"
I think a deontologist would reject this formulation. From NP&K (p. 24): "The non-consequentialist [...] can hold that it is fitting to value the survival of four individuals more than one’s own moral purity, but that what states it is fitting to value doesn’t always settle what to do."
Their claim is precisely that there can be practical reasons that fundamentally cannot be explained by preferences in terms of particular states of affairs. I am not quite sure how to make sense of this, since it seems to require that an "act" can be disentangled from a "state of affairs," but that at least seems to be what they are saying.
Anyway, we have a real live deontologist right here in this thread (!), so perhaps they can say whether this formulation makes sense to them.
Yes, I disagree with NP&K on this point. But deontologists in general are split. At least, many I've spoken to are happy to agree with me that (i) act-directed and state-directed preferences ideally ought to coincide, and (ii) what's distinctive of deontology is that the wrongness of certain violations is what explains what states we should prefer, rather than (as consequentialists would have it) desirable states of affairs explaining what acts are right and wrong.
Fair enough! Like I said, it's an outsider's view, informed mainly by my own incomprehension. (Well, also by Nietzsche, I suppose — but mainly by my own incomprehension.)
(I posted an earlier version of this comment, but then deleted because I realized I misunderstood something about your argument)
> So the only way to lack the preference demanded by (4) is to fail to much care about those five additional killings. I take that to be plainly morally unacceptable. So everyone must accept (4).
Disclaimer: I'm a consequentialist.
I think if we take deontology seriously, then we should indeed reject (4). The key difference between deontology and consequentialism is that in consequentialism, we care about the consequences, whereas in deontology, we care about the rules.
Regarding "One Killing to Prevent Five" vs "Failed Prevention", sure the consequences differ, but maybe we (as deontologists) don't care about the consequences; we care about the rules, and the rules were equally broken in both cases, in which case a self-consistent deontologist position would be to say "One Killing to Prevent Five = Failed Prevention".
Of course, I'm not sure any actual human who self-labels as a deontologist would truthfully advocate such a position -- and maybe that's your point -- but I'm talking about deontology as a theoretical "ideal" rather than how humans might adopt it in practice.
A decent form of deontology may give *priority* to "the rules" (as you put it) -- that's why they accept premise (3) -- but they can't ignore reasons of beneficence entirely. It's clearly immoral to *not care at all about whether people live or die*, even when all else is equal, as it is in (4).
I agree that there is a logically consistent view which rejects (4). But it isn't a decent one. Hence my conclusion that if we want a view that offers "adequate respect and concern for the five rescuable victims" then we must embrace consequentialism instead.
Hi Professor Y Chappell. I agree that no plausible deontological theory can completely ignore reasons of beneficence, but it seems to me that a common characteristic of deontological theories is that there must be at least some deontic constraints which no amount of reasons of beneficence can overrule. Otherwise is the theory in question really a deontological one? Do you think most deontologists think that there are no moral prohibitions which one’s duties of beneficence cannot outweigh? I have spoken with some professors in my department who seem to think that you have a lot of “discretion” when it comes to your duties of beneficence (what exactly this means, I don’t know, but it is supposed to entail that beneficence doesn’t require you to help the worst off or do the most good); I also worry that some deontologists would feel that you aren’t taking their views seriously if duties of beneficence can overrule all other duties. Fairly recently, I spoke to a mentor about Greaves and MacAskill’s “The Case for Strong Longtermism” and she felt as though the argument for Deontic Strong Longtermism just didn’t take deontology seriously, or that it was just a consequentialist argument with some deontological makeup. I would be grateful to hear your thoughts :)
Is this related to my argument? Philosophers typically distinguish between "moderate deontology" (which allows constraints to be overruled, just not easily) and "absolute deontology" (which says you shouldn't violate a constraint even if the fate of the world depended upon it). My sense is that moderate deontology is more commonly accepted these days, since absolutism just seems kind of crazy. But I don't think anything in my present argument depends upon taking a stand on this.
Right, I'm a consequentialist, and to me it's practically tautological that consequentialism is "right". Obviously, we care about consequences: we care about whether people live or die, and that's a consequence!
Still, I try to put an effort to understand other positions when I can.
I think you may have misunderstood. The idea that consequences matter -- that it matters whether people live or die -- is not in itself a distinctively consequentialist claim. That's a component of every sane view (even Rawls said it would be "irrational, crazy" to disregard consequences entirely). Consequentialism is the view that *only* consequences matter.
So the dispute with deontologists is not over whether consequences matter, but about whether *other* things (e.g. rights/constraints) matter in addition, and in a way that isn't reducible to their conduciveness to better consequences.
Edit: Mr. Bulldog has informed me that I have misinterpreted the article. I thought it was talking about the ideal preferability of acts. They say it's about world states. In that case, I think that this is trivial. Of course, the world would be "better" if we killed one to save five. The whole point is that some things are bad even if they make the world better. Premise 3 is thus false. A world
Original Comment:
> (1) Deontic constraint (for reductio): The protagonist acts wrongly in One Killing to Prevent Five, and ought instead to bring about the world of Five Killings.
What does it mean to “bring about the world of five killings”? I suppose you could say that simply doing nothing is “bringing about”, but that isn’t attributable to our Protagonist. Otherwise this is clearly false - killing five people is worse than Murdering one.
> If an agent can bring about W1 or W2, and it would be wrong for them to bring about W1 (but not W2), then W2 ≻ W1.
Makes sense. One would take action 2 over action 1.
> (3) Five Killings ≻ One Killing to Prevent Five. (from 1, 2)
Again, actively causing five killings to happen is worse than killing one, but not involving one and allowing five to occur is indeed the better act.
> One Killing to Prevent Five ≻≻ Failed Prevention. (premise)
Two things.
A. I doubt that it’s really “> >”. Presumably, there is some difference in having a successful act on the wrongness of murder, but you do the action either way, and the wrongness of that action dominates. One killing to prevent five is certainly not one whole “you kill someone” better than failed prevention.
B. There may actually be no difference at all. In both cases, you killed someone honestly believing (I assume) that you would save five people. The morality of your action doesn’t change based on the unknowable success of that act outside your control.
> (5) Failed Prevention ⪰ Six Killings. (premise)
Makes sense. If you kill someone for the heck of it you’re probably a bit worse then someone who tried to save five.
> (6) [Transitive Property]
Sadly I can’t dispute transitivity. :(
> (7) It is not the case that Five Killings ≻≻ Six Killings. (definition of ‘≻≻’)
Entirely the product of you being sneaky with your definitions. Five Killings *is* far more preferable to six killings to you, since *you* would be violating the deontology rule by committing that murder. Sure you can say that six generic killings are worse then five, but you absolutely cannot say that you doing the killing is weighted equally to any random killing. That denies the whole point of the assumed deontology premise.
The idea that one can "bring about" or "prefer" Failed Prevention seems paradoxical. As one cannot prefer "failing to achieve her end". Thus, this only makes sense from a third-person perspective, rather than the protagonist's perspective. However, if we take the third-person perspective, it is not the case that One Killing to Prevent Five is worse than Five Killings. Because there is nothing wrong with trying to bring about a world where someone else kills a person while more lives are saved. I think this secret shift of perspective renders this argument invalid.
The argument is meant to be run from a third-person perspective throughout. As I discuss in the full paper, one option is indeed to retreat to a "quiet deontology" where you want others to act in consequentialist-approved rather than deontology-approved ways.
But it's very surprising that deontologists would *have* to resort to this. Most would have antecedently assumed that they could coherently prefer that the agent in Transplant *not* murder an innocent person for their organs, etc. If you're now pro-murder (in the sense of preferring that the agent commit this murder in this case), it really narrows the distance with utilitarianism, and also casts doubt on the normative authority of the claim that it is "wrong" to kill the one. How could it be seriously wrong (in a way that matters) to do what an ideal agent must prefer you to do?
This just reinvents Batman's problem: should Batman kill Joker instead of giving him up to authorities, who put him in an asylum which he reliably escapes from to do more killings?
And it is well known that deontological answer - exemplified by Batman himself - is "no, he shouldn't". There is no paradox for deontologists, quite plausibly because they reject 7 by agent-relativity: _for Batman himself_ the world where they kill someone is indeed vastly worse than the world where they don't, while murders by others don't count as much.
(To put it mathematically, murder by yourself is M, murder by other is m, and m >> M for >>='vastly preferable', 0 > m, thus 0 >> M, thus 5m >> 5m + M.)
It all goes back to Copenhagen morality, where it's better to do nothing ;)
(It also seems true that it is _not_ true that 0 >> 5m in deontological approach, as @Nebu Pookins notes.)
I agree that works as a response to Scheffler's original "paradox of deontology". But in my version, the question is what *bystanders* should want. My argument shows that bystanders can't have distinctively "deontological" preferences. In your terms: *everyone else* has to want Batman to kill the Joker.
No? Deontological view is literally "everyone should prefer everyone to work off _their_ agent-relative function" (I think that's what your premise 2 negates, thus sneaking the conclusion in?).
(Just in case, I don't really _ascribe_ to that view.)
Who does "their" refer to? The assessor or the actor? I think the latter is a much better fit with deontological ethics. (The former seems more like a kind of egoism.)
Premise 2 *affirms* that bystanders should want other agents to act rightly (including not killing innocent ppl as a means, if deontology is true). The rest of the argument establishes that deontologists can't coherently maintain this (prima facie plausible) commitment.
If you abandon that premise, you're left with the result that bystanders have to have consequentialist preferences about kill-1-to-save-5 cases, i.e. hoping that the agent (wrongly) kills one as a means to doing more overall good. (Noting that a "bystander" is someone with no agent-relative grounds for bias towards any of the six potential victims.)
If we designate an action by agent x as a(x), and an agent-relative function of evaluation of actions by deontological commitments as f(x, a) (first argument is agent, second is action, just for specificity), then it's one of the axioms of deontology that Ax Ay f(x, a(y))=f(y, a(y)). That is, if it's wrong for Batman himself to kill someone, then it is ipso facto also wrong for anyone else to wish for Batman to kill someone. (Again, I don't ascribe to this view, but it is consistent.)
Well, it doesn't establish that. Specifically because most deontological f's would _not_ return 7.
Your definition is ambiguous. If >> means "more than one generic killing" (i.e. the difference is more than m), then Five Killings is indeed >> Six Killings. If >> means "more than one killing by the agent whose actions are evaluated" (i.e. the difference is more than M), then one of your previous steps (step 4, I think?) is false. You expect these two to mean the same, but they do not for a deontologist.
Having a huge aversion to murder no matter the circumstances probably saves a lot more lives than 5 whatever the calculation in some philosophy seminar.
Yes, of course, that's why utilitarians endorse such an aversion. If you want to distinguish utilitarian from deontological *theories*, you need to look beyond that issue. See: https://www.goodthoughts.blog/p/ethical-theory-and-practice
On many slippery slopes here. How do you decide in real life that the lives actually are saved. Your victims family may resist your murderous impulse and be willing to kill a lot of people to save them. If you conclude it is the right course of action the murderer should still be tried as a murderer.
One Killing to prevent Five + 1 killing > Failed Prevention
Five Killings ≻ One Killing to Prevent Five + 1 killing ≻≻ Failed Prevention ⪰ Six Killings.
I don't question
Failed Prevention ⪰ Six Killings, so I'll remove the part about Failed prevention
so we can shorten it to
Five Killings ≻ One Killing to Prevent Five + 1 killing > Six Killings.
to get the reductio I think we would need to move around so we get
Five Killings + 1 killing > Six Killings
but I don't see how we move it? I am partial to consequentialism already so I'm not asking for other arguments, just want to make sure I understand if there is a missing step that deontologist could retort to?
A little late to the party, but: You frame the rejection of (2) as a radical choice, but doesn't pretty much every non-consequentialist accept that there are "preferable" world states that it would nonetheless be unacceptable to bring about through some immoral act? I imagine everyone would "prefer" a world where five people lived over one. They just don't think that justifies using forced organ harvesting to get there. Alternatively, you could integrate some notion of "the fact that a bad decision has been made" into the nature of the world state, but then you'd be begging the question there - in that case, the deontologist *wouldn't* consider the world with [five people living through forced organ harvesting] better than the world with [five people dying through a refusal to engage in forced organ harvesting]. So either way, I guess I don't see how this would have much appeal to someone who isn't already a consequentialist. Am I missing something?
Acts are part of the world state, so it's the second reading. Except you've gotten it backwards -- I'm not begging the question, because I'm precisely *agreeing* that deontologists *wouldn't* consider the world of forced organ harvesting (W1) to be better. Given that it would be wrong to bring about, they instead prefer the alternative (W2). You clearly agree with this. So you agree that premise (2) is clearly true.
Okay, thank you for clarifying! I was indeed misreading it. I guess my only other comment, then, would be that it doesn't strike me as bizarre for a more agent-centered deontologist to find killing six to be much worse than killing five on the grounds that the former involves *two* people violating basic deontic restraints against killing. If I imagine a universe with four people (A, B, C, and D) then I would vastly prefer a world where A killed C and D while B did nothing over a world where A killed C and B killed D. I would probably even prefer a world where A killed B, C, and D over that. But my justification for it would probably teeter more towards a virtue ethics account than strict deontology. Either way, very interesting argument!
All of this adopts a non-deontological standpoint though, so is it really a reductio? It's assumed here that the concern of the deontologist is to try to bring about a world, or whatever. But the deontological question isn't "what world should we try to bring about?"
See point (III) in the OP. I'm not assuming this is "the concern" of the deontologist; just that their view must not be *inconsistent* with plausible claims about what we should prefer.
I mean obviously a deontologist is gonna reject P2 - they are already committed, by the very essence of what it means to be a deontologist, to the idea that something can be good yet still wrong.
P2 is compatible with the idea that something can be good yet still wrong. It just implies, in such a case, that you sometimes should prefer worlds with less value (because the greater-valued alternative could only be immorally obtained).
I cite Setiya, for example, who explicitly endorses such a principle in his recent paper defending deontic constraints. And I go on to explain how rejecting P2 would be extremely costly, effectively robbing constraints of their normative authority (since you'd then be hoping/preferring that people acted wrongly in these cases).
I'm gonna grant that there are SOME deontologists who would probably accept P2, but you know fully well that the vast vast majority of deontologists - from Kant to Ross - believe that normative reasons are agent-relative, e.g. that what matters is that YOU don't lie, not that fewest possible lies are being committed. So to just assume that this is false, even if people like Kant haven given loads of arguments for that kind of agent-relativity, just seems clearly question-begging.
Also, I don't see at all how this is supposed to rob constraints of their normative authority. If Kant's argument for the Categorical Imperative is right, then there is no mystery whatsoever why it would be obligatory for me personally not to do X, while still thinking that a world where X happens contains more value. Again: this only seems stupid if you are sceptical of Kantian deontology in the first place.
P2 is compatible with the claim that you shouldn't lie to minimize overall lying (just as it is compatible with the claim that you shouldn't kill to minimize overall killings). It simply implies that, in such a case, a morally ideal bystander shouldn't want you to act wrongly either.
Seems invalid? It's not at all clear how the sense of "is worse than" in premise 1 relates to preferability. A third party should surely prefer you to kill one indiscriminately (while no-one else is threatened) than to have six killings occur. Or, if you're talking about killing one indiscriminately *while five others are likewise being killed* then premise (2) is false. Six killings is not preferable to five. If you're equivocating between the two versions of "killing one indiscriminately", then you can't apply transitivity.
In terms of dispreferability, Killing 1 to save five ≤ killing one indiscriminately ≤ five people each killing one indiscriminately, from the standpoint of a third party. Thus, transivitively, a third party observer would evaluate 5 killings as worse than killing one to save five. (Here I used the opposite format that you did so if a ≤ b, that means that b is more dispreferable than A).
Do you mean "killing one indiscriminately (while no-one else is threatened)" or "killing one indiscriminately (while five others are likewise being killed)"? I'm assuming the former. But then constraints theorists like Setiya reject the idea that this is dispreferable to Killing 1 to save five. See fn 5 on p.11 of my paper.
(I agree that's already a significant bullet for deontologists to bite. But I also like to be able to show further problems even if we grant them this verdict.)
Okay, yeah, I think that mine loses its force if we accept that killing one indiscriminately is less bad than killing one to save five. But that seems really implausible. I think the two paradoxes aren't substitutes so much as different paradoxes.
Interesting response! But, to be clear, your position is not *just* that the constraint against killing as a means disables the lives saved from serving as reasons to prefer KOSF *over Five Killings*, but further, that even once the One has (wrongly) been killed, the very wrongness of that killing means that you no longer have *any* reason to prefer that the remaining lives be saved rather than not saved (at no *further* cost)?
That's a nice point -- it's true that my argument does depend on a kind of compositionality. I continue to find it very plausible, in that I think the comparative claim I made goes through even while bearing in mind the "once the One has been killed" part of the overall state of affairs (that is in common to both KOSF and Failed Prevention). We can explicitly bear this in mind, and still find it incredible to evaluate <This Bad Part + saving five lives> as *in no way preferable* to <This Bad Part + failed prevention where the five are killed>.
For example, if I imagine pausing time at the point where the One has just been killed, and asking any benevolent bystander, "Given what has just happened, what do you hope will happen from here?" (explicitly reminding them to evaluate the *whole* situations, not just the two futures in isolation, pretending that the past never happened), I think it would be outrageously indecent for them to be indifferent as to whether or not the five are saved.
But if you just don't share this (extremely strong) intuition of mine, there probably isn't much more that I can say. It's something I'm happy to take a premise, and expect that most others would share my sense that it is (very) costly to deny. But I take your point that one *could* deny this premise while still "caring" about the five in a broader sense (just not allowing that care to generate reasons in this particular situation). So, thanks for pointing that out.
I don't think it would have anything distinctive to say about it--my own reasons for being interested in contractualism have never really had anything to do with side constraints, and it may be that Richard and I would wind up agreeing that they are justified by pragmatic considerations rather than anything extremely deep. (And, to be clear, modifying contractualism to include animals is not original to me!)
I have three questions:
(1) Are there actually any (secular) deontologists? Or is deontology just something that a subset of professional philosophers continue to pretend to believe in because that was the team they picked when taking Ethics 101, it would be embarrassing to change their mind in public, and anyway they need to continue to pretend to believe in deontology in order to keep their jobs as bioethicists? (This thought brought to you courtesy of the "Treating Persons as Means" page on SEP, which gives the strong impression that the author can't find a way to make the concept intelligible but doesn't want to admit it.)
(2) >Insofar as one can distinguish narrowly “act-directed” from broader “state-directed” motivations, the latter have greater normative authority.
("More of a comment than a question...") This is the entire point of your disagreement with deontologists, so simply stating it like this appears to be begging the question. (This criticism applies more to this post than to your draft, which considers act-directed motivations at somewhat greater length, even if deontologists may still feel the draft doesn't provide adequate support for this claim.)
(More substantially, and with an actual question...) Nye, Plunkett & Ku seem to be arguing that deontology can be grounded in fittingness considerations — the words "fitting" and "fittingness" appear on practically every page of their paper. By contrast, you seem to use the language of fittingness here only very reluctantly: you don't use it at all in these posts, and in your draft you use it only when replying to them directly and quickly move on from talking about "what is fitting" to talking about "what matters" or "what one ought to do." To what extent do you think your divergence from Nye, Plunkett & Ku should be characterized in terms of divergent views on what is fitting, vs. divergent views on the normative force of fittingness considerations?
(3) >All we can say is that we ought to feel (irreparably) torn, which leaves us entirely lacking in practical normative guidance.
"We ought to feel irreparably torn, and lacking in practical normative guidance" seems to me an entirely accurate description of the human condition, quite independently of our views on deontology. Is there any reason to believe it is *wrong*, as opposed to just an unappealing conclusion for a certain variety of moral philosopher?
Incidentally, I think "ipso fact" on p. 22 of your draft should be "ipso facto." (An ipso facto typo?)
(1) About a third of philosophers, according to the PhilPapers survey -- https://survey2020.philpeople.org/survey/results/4890 -- or likely an outright majority if you include virtue ethicists (who seem likely to also endorse deontic constraints?)
(2) I don't think that's necessarily a point of disagreement at all. Deontologists can (and, I argue, should) offer a distinctively deontologically-flavoured account of the fitting "state-directed" motivations, such that, e.g., they should prefer the state of affairs in which Protagonist *doesn't* kill, even though this results in more killings overall. But yes, the actual arguments here appear in the paper -- this post is just offering a summary of the key results!
I'm actually very happy to formulate all the issues in terms of fittingness (I've a 2012 paper arguing that this is a good way to clarify ethical debates), I just refrain from doing so here because many find it a less familiar concept. That said, I'm not sure how best to characterize the heart of my disagreement with NP&K, because it partly seems to come down to their not taking sufficiently seriously their own Motivation-Action Principle. But perhaps some also comes down to a disagreement about *which* fittingness facts matter (and how): I say that the facts about preferability (or what it's fitting to care about, or desire) have a distinctive kind of normative force -- they warrant *caring* about -- in a way that doesn't necessarily hold true of other fittingness facts.
(3) Fair question. One thing to say is that, like nihilism, a view that offers no practical guidance may be safely ignored for practical purposes. In response to moral uncertainty, it seems that what we (meta-expectably) should do can only be influenced by views that yield directional answers of some sort.
Thanks for catching the typo!
(1) Interesting. Of course, I realize that a significant fraction of philosophers claim to be deontologists, but my vague (outsider's) assumption is that this is mainly because the field has a consensus that there is value in maintaining this tradition of thought — sort of like music departments continuing to hire people to study and teach baroque counterpoint, even if essentially nobody really listens to or makes that kind of music any more. I am strongly in favor of music departments continuing to study and teach baroque counterpoint, and I can accept that perhaps it is worth continuing to study and teach deontology for the same reason. (And Kant would have heard plenty of baroque counterpoint — perhaps there is a connection here?) But I'm not sure deontologists would appreciate me thinking about them in this way...
(2) >"they should prefer the state of affairs in which Protagonist *doesn't* kill"
I think a deontologist would reject this formulation. From NP&K (p. 24): "The non-consequentialist [...] can hold that it is fitting to value the survival of four individuals more than one’s own moral purity, but that what states it is fitting to value doesn’t always settle what to do."
Their claim is precisely that there can be practical reasons that fundamentally cannot be explained by preferences in terms of particular states of affairs. I am not quite sure how to make sense of this, since it seems to require that an "act" can be disentangled from a "state of affairs," but that at least seems to be what they are saying.
Anyway, we have a real live deontologist right here in this thread (!), so perhaps they can say whether this formulation makes sense to them.
Yes, I disagree with NP&K on this point. But deontologists in general are split. At least, many I've spoken to are happy to agree with me that (i) act-directed and state-directed preferences ideally ought to coincide, and (ii) what's distinctive of deontology is that the wrongness of certain violations is what explains what states we should prefer, rather than (as consequentialists would have it) desirable states of affairs explaining what acts are right and wrong.
No, your diagnosis under (1) isn't right at all.
Fair enough! Like I said, it's an outsider's view, informed mainly by my own incomprehension. (Well, also by Nietzsche, I suppose — but mainly by my own incomprehension.)
> Are there actually any (secular) deontologists?
Yes! :D
Dozens of us, I swear
Almost everybody acquiesces to a rule based system of criminal justice, so , for weakish definitions of deontology,almost everyone is a deontologist.
(I posted an earlier version of this comment, but then deleted because I realized I misunderstood something about your argument)
> So the only way to lack the preference demanded by (4) is to fail to much care about those five additional killings. I take that to be plainly morally unacceptable. So everyone must accept (4).
Disclaimer: I'm a consequentialist.
I think if we take deontology seriously, then we should indeed reject (4). The key difference between deontology and consequentialism is that in consequentialism, we care about the consequences, whereas in deontology, we care about the rules.
Regarding "One Killing to Prevent Five" vs "Failed Prevention", sure the consequences differ, but maybe we (as deontologists) don't care about the consequences; we care about the rules, and the rules were equally broken in both cases, in which case a self-consistent deontologist position would be to say "One Killing to Prevent Five = Failed Prevention".
Of course, I'm not sure any actual human who self-labels as a deontologist would truthfully advocate such a position -- and maybe that's your point -- but I'm talking about deontology as a theoretical "ideal" rather than how humans might adopt it in practice.
A decent form of deontology may give *priority* to "the rules" (as you put it) -- that's why they accept premise (3) -- but they can't ignore reasons of beneficence entirely. It's clearly immoral to *not care at all about whether people live or die*, even when all else is equal, as it is in (4).
I agree that there is a logically consistent view which rejects (4). But it isn't a decent one. Hence my conclusion that if we want a view that offers "adequate respect and concern for the five rescuable victims" then we must embrace consequentialism instead.
Hi Professor Y Chappell. I agree that no plausible deontological theory can completely ignore reasons of beneficence, but it seems to me that a common characteristic of deontological theories is that there must be at least some deontic constraints which no amount of reasons of beneficence can overrule. Otherwise is the theory in question really a deontological one? Do you think most deontologists think that there are no moral prohibitions which one’s duties of beneficence cannot outweigh? I have spoken with some professors in my department who seem to think that you have a lot of “discretion” when it comes to your duties of beneficence (what exactly this means, I don’t know, but it is supposed to entail that beneficence doesn’t require you to help the worst off or do the most good); I also worry that some deontologists would feel that you aren’t taking their views seriously if duties of beneficence can overrule all other duties. Fairly recently, I spoke to a mentor about Greaves and MacAskill’s “The Case for Strong Longtermism” and she felt as though the argument for Deontic Strong Longtermism just didn’t take deontology seriously, or that it was just a consequentialist argument with some deontological makeup. I would be grateful to hear your thoughts :)
Is this related to my argument? Philosophers typically distinguish between "moderate deontology" (which allows constraints to be overruled, just not easily) and "absolute deontology" (which says you shouldn't violate a constraint even if the fate of the world depended upon it). My sense is that moderate deontology is more commonly accepted these days, since absolutism just seems kind of crazy. But I don't think anything in my present argument depends upon taking a stand on this.
Right, I'm a consequentialist, and to me it's practically tautological that consequentialism is "right". Obviously, we care about consequences: we care about whether people live or die, and that's a consequence!
Still, I try to put an effort to understand other positions when I can.
I think you may have misunderstood. The idea that consequences matter -- that it matters whether people live or die -- is not in itself a distinctively consequentialist claim. That's a component of every sane view (even Rawls said it would be "irrational, crazy" to disregard consequences entirely). Consequentialism is the view that *only* consequences matter.
So the dispute with deontologists is not over whether consequences matter, but about whether *other* things (e.g. rights/constraints) matter in addition, and in a way that isn't reducible to their conduciveness to better consequences.
Other things, eg. rules and virtues, can be reducible to consequences in ways that significant differ, or conflict.
Edit: Mr. Bulldog has informed me that I have misinterpreted the article. I thought it was talking about the ideal preferability of acts. They say it's about world states. In that case, I think that this is trivial. Of course, the world would be "better" if we killed one to save five. The whole point is that some things are bad even if they make the world better. Premise 3 is thus false. A world
Original Comment:
> (1) Deontic constraint (for reductio): The protagonist acts wrongly in One Killing to Prevent Five, and ought instead to bring about the world of Five Killings.
What does it mean to “bring about the world of five killings”? I suppose you could say that simply doing nothing is “bringing about”, but that isn’t attributable to our Protagonist. Otherwise this is clearly false - killing five people is worse than Murdering one.
> If an agent can bring about W1 or W2, and it would be wrong for them to bring about W1 (but not W2), then W2 ≻ W1.
Makes sense. One would take action 2 over action 1.
> (3) Five Killings ≻ One Killing to Prevent Five. (from 1, 2)
Again, actively causing five killings to happen is worse than killing one, but not involving one and allowing five to occur is indeed the better act.
> One Killing to Prevent Five ≻≻ Failed Prevention. (premise)
Two things.
A. I doubt that it’s really “> >”. Presumably, there is some difference in having a successful act on the wrongness of murder, but you do the action either way, and the wrongness of that action dominates. One killing to prevent five is certainly not one whole “you kill someone” better than failed prevention.
B. There may actually be no difference at all. In both cases, you killed someone honestly believing (I assume) that you would save five people. The morality of your action doesn’t change based on the unknowable success of that act outside your control.
> (5) Failed Prevention ⪰ Six Killings. (premise)
Makes sense. If you kill someone for the heck of it you’re probably a bit worse then someone who tried to save five.
> (6) [Transitive Property]
Sadly I can’t dispute transitivity. :(
> (7) It is not the case that Five Killings ≻≻ Six Killings. (definition of ‘≻≻’)
Entirely the product of you being sneaky with your definitions. Five Killings *is* far more preferable to six killings to you, since *you* would be violating the deontology rule by committing that murder. Sure you can say that six generic killings are worse then five, but you absolutely cannot say that you doing the killing is weighted equally to any random killing. That denies the whole point of the assumed deontology premise.
The idea that one can "bring about" or "prefer" Failed Prevention seems paradoxical. As one cannot prefer "failing to achieve her end". Thus, this only makes sense from a third-person perspective, rather than the protagonist's perspective. However, if we take the third-person perspective, it is not the case that One Killing to Prevent Five is worse than Five Killings. Because there is nothing wrong with trying to bring about a world where someone else kills a person while more lives are saved. I think this secret shift of perspective renders this argument invalid.
The argument is meant to be run from a third-person perspective throughout. As I discuss in the full paper, one option is indeed to retreat to a "quiet deontology" where you want others to act in consequentialist-approved rather than deontology-approved ways.
But it's very surprising that deontologists would *have* to resort to this. Most would have antecedently assumed that they could coherently prefer that the agent in Transplant *not* murder an innocent person for their organs, etc. If you're now pro-murder (in the sense of preferring that the agent commit this murder in this case), it really narrows the distance with utilitarianism, and also casts doubt on the normative authority of the claim that it is "wrong" to kill the one. How could it be seriously wrong (in a way that matters) to do what an ideal agent must prefer you to do?
This just reinvents Batman's problem: should Batman kill Joker instead of giving him up to authorities, who put him in an asylum which he reliably escapes from to do more killings?
And it is well known that deontological answer - exemplified by Batman himself - is "no, he shouldn't". There is no paradox for deontologists, quite plausibly because they reject 7 by agent-relativity: _for Batman himself_ the world where they kill someone is indeed vastly worse than the world where they don't, while murders by others don't count as much.
(To put it mathematically, murder by yourself is M, murder by other is m, and m >> M for >>='vastly preferable', 0 > m, thus 0 >> M, thus 5m >> 5m + M.)
It all goes back to Copenhagen morality, where it's better to do nothing ;)
(It also seems true that it is _not_ true that 0 >> 5m in deontological approach, as @Nebu Pookins notes.)
I agree that works as a response to Scheffler's original "paradox of deontology". But in my version, the question is what *bystanders* should want. My argument shows that bystanders can't have distinctively "deontological" preferences. In your terms: *everyone else* has to want Batman to kill the Joker.
No? Deontological view is literally "everyone should prefer everyone to work off _their_ agent-relative function" (I think that's what your premise 2 negates, thus sneaking the conclusion in?).
(Just in case, I don't really _ascribe_ to that view.)
Who does "their" refer to? The assessor or the actor? I think the latter is a much better fit with deontological ethics. (The former seems more like a kind of egoism.)
Premise 2 *affirms* that bystanders should want other agents to act rightly (including not killing innocent ppl as a means, if deontology is true). The rest of the argument establishes that deontologists can't coherently maintain this (prima facie plausible) commitment.
If you abandon that premise, you're left with the result that bystanders have to have consequentialist preferences about kill-1-to-save-5 cases, i.e. hoping that the agent (wrongly) kills one as a means to doing more overall good. (Noting that a "bystander" is someone with no agent-relative grounds for bias towards any of the six potential victims.)
If we designate an action by agent x as a(x), and an agent-relative function of evaluation of actions by deontological commitments as f(x, a) (first argument is agent, second is action, just for specificity), then it's one of the axioms of deontology that Ax Ay f(x, a(y))=f(y, a(y)). That is, if it's wrong for Batman himself to kill someone, then it is ipso facto also wrong for anyone else to wish for Batman to kill someone. (Again, I don't ascribe to this view, but it is consistent.)
Well, it doesn't establish that. Specifically because most deontological f's would _not_ return 7.
re: your first paragraph, that's precisely what my premise 2 affirms.
(7) is true by definition.
Your definition is ambiguous. If >> means "more than one generic killing" (i.e. the difference is more than m), then Five Killings is indeed >> Six Killings. If >> means "more than one killing by the agent whose actions are evaluated" (i.e. the difference is more than M), then one of your previous steps (step 4, I think?) is false. You expect these two to mean the same, but they do not for a deontologist.
Having a huge aversion to murder no matter the circumstances probably saves a lot more lives than 5 whatever the calculation in some philosophy seminar.
Yes, of course, that's why utilitarians endorse such an aversion. If you want to distinguish utilitarian from deontological *theories*, you need to look beyond that issue. See: https://www.goodthoughts.blog/p/ethical-theory-and-practice
On many slippery slopes here. How do you decide in real life that the lives actually are saved. Your victims family may resist your murderous impulse and be willing to kill a lot of people to save them. If you conclude it is the right course of action the murderer should still be tried as a murderer.
This is ethical theory, not real-world advice.
How about killing 15 people to save 16?
so I get how w1 >> w2 entails
w1 + murder > w2
so we can exchange
One Killing to Prevent Five ≻≻ Failed Prevention
with
One Killing to prevent Five + 1 killing > Failed Prevention
Five Killings ≻ One Killing to Prevent Five + 1 killing ≻≻ Failed Prevention ⪰ Six Killings.
I don't question
Failed Prevention ⪰ Six Killings, so I'll remove the part about Failed prevention
so we can shorten it to
Five Killings ≻ One Killing to Prevent Five + 1 killing > Six Killings.
to get the reductio I think we would need to move around so we get
Five Killings + 1 killing > Six Killings
but I don't see how we move it? I am partial to consequentialism already so I'm not asking for other arguments, just want to make sure I understand if there is a missing step that deontologist could retort to?
When you add "+1 killing" to "One Killing to Prevent Five", you also need to add it to each prior link in the chain, i.e. to "Five Killings" too.
Realised just as I was going to bed. Exciting stuff, thanks for the response
Very clever.
A little late to the party, but: You frame the rejection of (2) as a radical choice, but doesn't pretty much every non-consequentialist accept that there are "preferable" world states that it would nonetheless be unacceptable to bring about through some immoral act? I imagine everyone would "prefer" a world where five people lived over one. They just don't think that justifies using forced organ harvesting to get there. Alternatively, you could integrate some notion of "the fact that a bad decision has been made" into the nature of the world state, but then you'd be begging the question there - in that case, the deontologist *wouldn't* consider the world with [five people living through forced organ harvesting] better than the world with [five people dying through a refusal to engage in forced organ harvesting]. So either way, I guess I don't see how this would have much appeal to someone who isn't already a consequentialist. Am I missing something?
Acts are part of the world state, so it's the second reading. Except you've gotten it backwards -- I'm not begging the question, because I'm precisely *agreeing* that deontologists *wouldn't* consider the world of forced organ harvesting (W1) to be better. Given that it would be wrong to bring about, they instead prefer the alternative (W2). You clearly agree with this. So you agree that premise (2) is clearly true.
Okay, thank you for clarifying! I was indeed misreading it. I guess my only other comment, then, would be that it doesn't strike me as bizarre for a more agent-centered deontologist to find killing six to be much worse than killing five on the grounds that the former involves *two* people violating basic deontic restraints against killing. If I imagine a universe with four people (A, B, C, and D) then I would vastly prefer a world where A killed C and D while B did nothing over a world where A killed C and B killed D. I would probably even prefer a world where A killed B, C, and D over that. But my justification for it would probably teeter more towards a virtue ethics account than strict deontology. Either way, very interesting argument!
All of this adopts a non-deontological standpoint though, so is it really a reductio? It's assumed here that the concern of the deontologist is to try to bring about a world, or whatever. But the deontological question isn't "what world should we try to bring about?"
See point (III) in the OP. I'm not assuming this is "the concern" of the deontologist; just that their view must not be *inconsistent* with plausible claims about what we should prefer.
I mean obviously a deontologist is gonna reject P2 - they are already committed, by the very essence of what it means to be a deontologist, to the idea that something can be good yet still wrong.
P2 is compatible with the idea that something can be good yet still wrong. It just implies, in such a case, that you sometimes should prefer worlds with less value (because the greater-valued alternative could only be immorally obtained).
I cite Setiya, for example, who explicitly endorses such a principle in his recent paper defending deontic constraints. And I go on to explain how rejecting P2 would be extremely costly, effectively robbing constraints of their normative authority (since you'd then be hoping/preferring that people acted wrongly in these cases).
I'm gonna grant that there are SOME deontologists who would probably accept P2, but you know fully well that the vast vast majority of deontologists - from Kant to Ross - believe that normative reasons are agent-relative, e.g. that what matters is that YOU don't lie, not that fewest possible lies are being committed. So to just assume that this is false, even if people like Kant haven given loads of arguments for that kind of agent-relativity, just seems clearly question-begging.
Also, I don't see at all how this is supposed to rob constraints of their normative authority. If Kant's argument for the Categorical Imperative is right, then there is no mystery whatsoever why it would be obligatory for me personally not to do X, while still thinking that a world where X happens contains more value. Again: this only seems stupid if you are sceptical of Kantian deontology in the first place.
P2 is compatible with the claim that you shouldn't lie to minimize overall lying (just as it is compatible with the claim that you shouldn't kill to minimize overall killings). It simply implies that, in such a case, a morally ideal bystander shouldn't want you to act wrongly either.
What do you think about this simpler version?
1 A perfectly moral third party shouldn't conclude that killing one to try unsuccessfully to save 5 is worse than killing one indiscriminately.
2 A perfectly moral third party should prefer you killing one indiscriminately to five people killing one each.
3 You should act in ways that perfectly moral third parties want you to act.
Therefore, by transitivity, you should kill one to prevent 5 killings -- it's preferrable by a third party.
Seems invalid? It's not at all clear how the sense of "is worse than" in premise 1 relates to preferability. A third party should surely prefer you to kill one indiscriminately (while no-one else is threatened) than to have six killings occur. Or, if you're talking about killing one indiscriminately *while five others are likewise being killed* then premise (2) is false. Six killings is not preferable to five. If you're equivocating between the two versions of "killing one indiscriminately", then you can't apply transitivity.
In terms of dispreferability, Killing 1 to save five ≤ killing one indiscriminately ≤ five people each killing one indiscriminately, from the standpoint of a third party. Thus, transivitively, a third party observer would evaluate 5 killings as worse than killing one to save five. (Here I used the opposite format that you did so if a ≤ b, that means that b is more dispreferable than A).
Do you mean "killing one indiscriminately (while no-one else is threatened)" or "killing one indiscriminately (while five others are likewise being killed)"? I'm assuming the former. But then constraints theorists like Setiya reject the idea that this is dispreferable to Killing 1 to save five. See fn 5 on p.11 of my paper.
(I agree that's already a significant bullet for deontologists to bite. But I also like to be able to show further problems even if we grant them this verdict.)
Okay, yeah, I think that mine loses its force if we accept that killing one indiscriminately is less bad than killing one to save five. But that seems really implausible. I think the two paradoxes aren't substitutes so much as different paradoxes.
Interesting response! But, to be clear, your position is not *just* that the constraint against killing as a means disables the lives saved from serving as reasons to prefer KOSF *over Five Killings*, but further, that even once the One has (wrongly) been killed, the very wrongness of that killing means that you no longer have *any* reason to prefer that the remaining lives be saved rather than not saved (at no *further* cost)?
That seems utterly incredible to me.
That's a nice point -- it's true that my argument does depend on a kind of compositionality. I continue to find it very plausible, in that I think the comparative claim I made goes through even while bearing in mind the "once the One has been killed" part of the overall state of affairs (that is in common to both KOSF and Failed Prevention). We can explicitly bear this in mind, and still find it incredible to evaluate <This Bad Part + saving five lives> as *in no way preferable* to <This Bad Part + failed prevention where the five are killed>.
For example, if I imagine pausing time at the point where the One has just been killed, and asking any benevolent bystander, "Given what has just happened, what do you hope will happen from here?" (explicitly reminding them to evaluate the *whole* situations, not just the two futures in isolation, pretending that the past never happened), I think it would be outrageously indecent for them to be indifferent as to whether or not the five are saved.
But if you just don't share this (extremely strong) intuition of mine, there probably isn't much more that I can say. It's something I'm happy to take a premise, and expect that most others would share my sense that it is (very) costly to deny. But I take your point that one *could* deny this premise while still "caring" about the five in a broader sense (just not allowing that care to generate reasons in this particular situation). So, thanks for pointing that out.
Yes, that definitely seems like a less bad way to go!
I don't think it would have anything distinctive to say about it--my own reasons for being interested in contractualism have never really had anything to do with side constraints, and it may be that Richard and I would wind up agreeing that they are justified by pragmatic considerations rather than anything extremely deep. (And, to be clear, modifying contractualism to include animals is not original to me!)