Michael Cholbi’s recent post on the “murderer at the door” case brought up the issue of duties to self, which reminded me of a paper idea I’ve been meaning to develop, and which I thought I’d run by this crowd first.

In contrast to current popular morality in the U.S., most normative ethical theories (besides, perhaps, relativism) contain duties to oneself. For agent-neutral consequentialists, I might count for only one, but I at least have that much moral significance. Kant offers duties of self-improvement, among others. Perfectionists, obviously, highly value duties to self. And so on. But this is one area where I actually identify with current pop morality in the U.S.: I’m not a big fan of duties to self. Kant famously raised an objection to duties to oneself (and replied to it), but I want to try to bolster the objection here.

Kant puts the objection in terms of a tension between the self’s being passively constrained (duty) with the self actively doing the constraining (not duty), but I want to put it slightly differently. (Kant’s discussion is at The Metaphysics of Morals, 417-418.) It seems that with duties to others, such as the duty to keep one’s promise, the other can always release the agent from the duty. So, if Dave has promised to send me a video, I can release him from his obligation simply by saying “Don’t worry about it.” Similarly, if Doug has a duty of beneficence to give me $100, I can, simply by so choosing, release him from this duty. So the question is: Is there any rationale for saying we can’t do the same thing with duties to self?

Kant tried to get around the problem by distinguishing between the noumenal self and the sensible self, but, first, that’s not too helpful for naturalistically-minded folks like myself, and, second, I fail to see why that doesn’t just push back the question to the noumenal self. (I find Kant’s actual discussion a bit, um, obscure on this one.)

Agent-neutral consequentialists might point to the value that would be produced by not releasing myself from my putative duty to self. But this would equally hold for not releasing Dave from his obligation to send me the video. So either I can’t release Dave from his obligation, which seems counterintuitive (and unnecessary for the consequentialist since there’s value in the simple fact of my exercising my free choice to release him as I see fit) or I can, in which case I should also be able to release myself from duties to self as I see fit.

And that argument holds for any moral theory. Either I can’t release Dave from his obligation to fulfill his promise, which seems counterintuitive, or I can, which seems to suggest that I should also be able to release myself from any putative duties I “owe” to myself. (Of course there may be extraordinary cases where agent A can’t release agent B from an A-regarding duty, so in such extraordinary cases the parallel with duties-to-self wouldn’t hold. But – granting there are such cases – let’s put those to the side here.)

So there seems to be a presumptive case for the claim that (1) any agent can, at her discretion, release herself from any putative self-regarding duty. But, (2) if it is purely a matter of discretion whether an agent performs some action, then that action is not actually obligatory, so (3) there are no duties to self.

Any thoughts?

11 Replies to “It’s Easy Being Me: Duties to Self

  1. Josh, I really like your argument here. In an exchange of ideas with Allen Habib a few months ago, I had a very similar thought. It went like this.
    Let’s say that when x has a duty to y, and hence y has a right against duty (I’m assuming their correlation here), x is “the dutied” and y is “the righted.” Let’s also define “schmuty” as something that is just like a duty, with this one exception: whereas with duty only the righted can release the dutied from her duty, with schmuty both the righted and the schmutied can release the schmutied from her schmuty.
    Intuitively, it seems that the this distinguishing feature of schmuties is sufficient to make them *not* a special case of duties. That is, it is an essential property of duties that they cannot be undone, or released from, by anyone other than the righted.

  2. Josh,
    I’m not sure how far (neologism warning) ‘releaseability’ extends, even with respect to duties to others. In fact, I think that unreleaseable duties are more ordinary than “extraordinary” as you put it. For instance, it’s hard for me to imagine how a child could release a parent of the parent’s duties of welfare toward the child. Likewise, it’s difficult to stomach the idea that I could release you from your duty not to kill me (though there was that unusual case in, I believe, Germany where a guy answered a classified ad seeking someone to be killed and eaten for the sexual gratification of the advertisor – tough case for libertarians!). If I did “release” you from that duty (if, e.g., I’m the only carrier of the Ebola virus and I recognize the wisdom of your killing me in order to prevent the spread of the disease), then that wouldn’t be so much releasing you from the duty but showing that what you thought was your duty isn’t really your duty in this circumstance. Promising seems to me to be a special case, precisely because promissory obligations are created by acts undertaken by the parties, and are therefore not ‘natural’ duties, so to speak. That’s why it seems plausible to think that one cannot meaningfully owe oneself a promissory obligation. But I don’t know that the promising example generalizes well. The best example for duties to self that I can think of is the duty not to place oneself in perpetual servitude. Is it not only imprudent but actually unjust to oneself to knowingly enter into a relationship of perpetual enslavement?

  3. Josh,
    I’ve been working on this issue so I’ll take a swing at this. One point here about Kant: he counts some duties as ‘imperfect’; if you have imperfect duties, then you have a class of duties for which there are no reciprocal rights (my duty to help others does not give anyone a right to my help, for instance; likewise, my duty of self-development gives me no right over myself.) However, he thinks there are ‘perfect’ duties to yourself, so that would leave some puzzles about those. Another point is that we should just deny the kind of voluntarism that generates the puzzle. That, I take it, is at the heart of the way that Uriah and Michael want to go: it is not the case that if you owe a duty to someone, she can release you from it. Kant himself appears to be attracted to the sort of voluntarism that generates the problem (as in the passage you cite), but his own view in fact is opposed to it. Moral duties to ourselves and others only require that there be some non-desire-based reasons or principles requiring one to treat oneself and others in certain ways. One’s own humanity, for instance, is a reason to treat oneself in certain ways but not others.
    Another way to look at it is this: Our duty in any particular case is determined by the moral principles covering the circumstances. Think of these on a par with a game: I have to give you your Queen back when your pawn reaches the end of the board. Here, I owe you a queen, but it’s not as if you can ‘release’ me from that; it’s just what the rules of chess require (well, ok you could say ‘just leave my pawn’ I suppose, but you get the point). So I have a duty to you just in case there is some moral principle that says so. And whether you can release me from that duty or not at will depends on whether that is a part of the principle that determines my duty to you in the first place–not on some element of the very idea of having a duty to someone.

  4. Josh,
    You argue as follows: “(1) any agent can, at her discretion, release herself from any putative self-regarding duty. But, (2) if it is purely a matter of discretion whether an agent performs some action, then that action is not actually obligatory, so (3) there are no duties to self.” I think that this argument is invalid. I’ll give the following putative counterexample. Consider the putative duty to safeguard one’s own life, a duty to self. Now suppose that Bob is a diabetic who negligently fails to take his insulin with him on an extended and solo hiking trip, and, consequently, he dies. This act of negligence on his part is clearly prohibited by the putative duty he has to safeguard his own life. And since this is not a case of intentional suicide but a case of negligence, we can suppose that it was never Bob’s intention to release himself from the duty to safeguard his own life. Of course, premise (1) is nevertheless true, for Bob could have, at his discretion, released himself from this duty. But, in fact, he didn’t. Premise (2) is also true for if it is purely a matter of discretion whether an agent performs some action (e.g., refraining from jumping of a building with the intention of killing oneself), then that action is not actually obligatory. True enough. But (3) doesn’t follow from (1) and (2). For there could still be a duty to safeguard one’s life, a duty that prohibits acts of negligence that put one’s own life at risk.

  5. Great points. I can see I’ve got my work cut out for me.
    Doug, let me make sure I’ve got your case right. Your duty is “to safeguard one’s own life,” and in this instance Bob negligently failed to comply with his duty. Also, you (unlike Michael and Robert) are willing to grant that (1) Bob could have released himself from this duty (but didn’t) and that (2) if Bob had done so, then safeguarding his own life would no longer have been obligatory. However, since as a matter of fact Bob did not release himself, he was still obligated, i.e., he had a duty to self. Is this right? If so, I think you and I are reading premise (2) differently. You’re taking the conditional to be something of a causal relationship where the truth of the antecedent depends on empirical facts: if,as a matter of empirical fact, A releases himself from a duty to perform some action, then it ceases to be obligatory. By contrast, I’m reading it as a logical (and modal) conditional, where the truth of the antecedent is not subject to empirical truth: if A *can* release himself from a duty to perform some action, then it is not, and never was, a duty. So as long as you grant that Bob could have released himself from the “duty” to preserve himself, then my version of (2) means that he had no real duty to preserve himself. That is, it’s not a counterexample, unless, of course, I’m misreading you.
    Robert and Michael, by contrast, suggested that we cannot release ourselves from duties to self, though I gather for different reasons. My argument for the “releasability thesis” for duties to self was little more than an analogy with the releasability feature of duties to others. Michael, you deny that we can release agents from duties to others (thereby undermining the analogy). I guess to dispute this we’d just have to appeal to our intuitions about possible cases, as you do. First, though, I’d suggest ruling out the child-parent case as applicable, for ordinarily we think that agents can release other agents from duties (here, noncontroversially, think of promises), and children of course do not have full-fledged agency. But what about the other case, the duty not to kill? It seems to me that we *can* release another from this duty, e.g., in cases of assisted suicide. I take it, Michael, that you’d say something like, “This isn’t a case of a purely discretionary release-of-duty; rather, the ‘release’ is nothing more than a pointing out that the duty is overridden in these circumstances.” (Right? That’s how I’m reading the ebola case.) But that seems to me like a distortion of what actually happens. What overrides the duty to not assist in the killing of others is not that the other is dying a painful death, but that the other has calculated that she wants to die and is asking for my help. At least that’s my “pre-theoretical” intuition about the case. And, if the desideratum is to figure out whether unreleasable other-regarding duties are more ordinary than extraordinary, I guess I’d want to see more cases where it seems that other-regarding duties are supposed to be unreleasable.
    Perhaps, though, Michael has something in mind closer to what Robert means, which if I’m following (and here I’m also being informed by what he said over on Uriah’s post on the topic at Desert Landscapes) is something like this. If a duty to others can be overridden by the dutied’s release, this is because such a duty will contain a release-override as one of the circumstances in which it will be rendered inapplicable, just like “saving a life” will be one of the circumstances that is an overriding condition contained in the duty to keep an ordinary promise. So the duty itself isn’t overridden; rather, the content of the duty specifies when it is and is not applicable, and that (whole) duty is laid down by some more fundamental principle and as such (i.e., as containing all possible “escape” conditions within the duty) the duty itself cannot be overridden by anything, including the “release” of the dutied.
    Is that right (or close)? If so, the debate comes down to this: when a duty is rendered inapplicable, is this because the duty itself is overridden, or it it because the duty contains “escape” clauses that do not render the duty overridden but instead render the course of action prescribed in the duty unnecessary to perform? Frankly, I’m not sure how to answer this question, though I would say that while we now have an articulation of how there could be duties to others, we don’t yet have a defense of such duties. And, it seems to me that the basis for the original analogy is still somewhat compelling: our (or my) ordinary intuitions are that we can actually release each other from duties we owe to each other, not merely that we’re still obligated by such duties but that when the dutied expresses a putative ‘release’ claim what’s really happening is that the dutied is merely pointing out that the actions the duties prescribe aren’t necessary in those particular circumstances.

  6. Josh, one other thought on the topic: I think there is widespread acknowledgement in ordinary practice that we do have duties to ourselves. In fact it is among the most pedestrian of our moral judgments: We think we can ‘let ourselves down’ that we ‘owe it to ourselves’ to do this or that, we berate ourselves for having failed ourselves in this or that way, that we had a right to expect more from ourselves and so on. We think this of others as well: She owes it to herself to tell him what she thinks; if she doesn’t apply herself, she’ll only be letting herself down, and so on. Naturally, this can easily slop over into moralizing (in the pejorative sense). But if you consider how you regard yourself and those close to you, even occasionally students, you’ll conclude pretty quickly that you think in these terms quite often: at least you *think* we have obligations to ourselves. That may be mistaken, but it is a part of ordinary morality.

  7. It’s true that Josh’s argument and mine presupposed the following principle:
    (*) An agent can always release herself from what appears to be a duty to herself.
    And I guess this is not as obvious as I had thought. Robert has an interesting example with the chess game: you cannot release me from my duty to give you the queen in accordance with the rules of chess.
    But perhaps one could reply as follows. Yes, you *can* release me from my duty to give you the queen (in discordance with the rules of chess) – at the price of not playing chess anymore. So the question now becomes: is a situation where the moral system dishes out duties so that the only way to release someone from her duty is at the price of not “playing morality” anymore. If so, then a person may indeed have moral duties to herself, in the sense that although she can release herself from the duty, doing so would entail her not being moral (or her giving up on morality, which comes to the same thing).
    I must say that I’m having difficulty imagining a self-duty dished out in this way by the rules of morality. For that matter, I’m having difficulty imagining a duty to *others* that is like that. In acordance with the rules of morality, I have a duty not to torture innocent people – perhaps their humanity commands that. But suppose an innocent person begs me to torture her, for whatever reason (that’s their business). Would I be flaunting the rule of morality or unduly disregarding their humanity by obliging? My intuition says No. But I’m open to persuasion on all this.
    There’s another interesting issue that Robert brings up in the last post, namely, is it really the case that pop morality says there are no duties to oneself? We talk of owing something to ourselves, and as Allen Habib pointed out to me, there are such things as pledging to work out twice a week or making a new-year resolution to stop smoking this coming year. On the face of it, these appear to commit the folk to existence of mechanisms for creating duties to oneself.
    But there is still a question whether we should take these at face value. My inclination is to say that these are really mechanisms for “psyching” oneself into being very serious about some self-regarding matter. What is not part of the picture is that by making a pledge or a resolution one is also *incurring* an obligation. But anyway this is clearly open to debate.

  8. Josh responds to Michael that the inability of children to release their parents from duties of care is explained by the fact that children are not full-fledged agents. But surely Michael’s point can be made just as well by cases of full-fledged agents: siblings’ or spouses’ duties to one another, e.g., or (adult) children’s duties to their parents.
    In fact, I’m not sure that there is any case in which B’s duty to A can be cancelled merely by A’s decision to release B. Take the case in which A decides she wants to die, and so releases B from his duty not to kill. Josh writes, “What overrides the duty to not assist in the killing of others is . . . that the other has calculated that she wants to die and is asking for my help.” But surely this is not nearly enough. Suppose B thinks that A has not calculated correctly because she is missing some crucial information about her diagnosis. (Perhaps she isn’t really dying at all.) Or suppose that B has reason to believe that A is suffering from a temporary depression and will feel differently about the situation later. In those cases B’s duty not to kill A still holds, despite A’s release. So it is the objective facts about A’s interests, and not A’s decision, that determine B’s duties. (However, because people tend in most circumstances to be the best judges of their own interests, we do take those decisions seriously. But mostly, their role is to serve as evidence about the objective facts.)
    (Thus I disagree with Uriah when he writes that, if one were to choose to torture someone who begged to be tortured “for whatever reason,” one would not be violating morality or disregarding that person’s humanity. The reason matters: if she can show that it truly is in her interest to be tortured (this of course is rather hard to imagine) then, and only then, may I oblige.)
    This even seems to apply to promissory obligations. Suppose Bob promises to give Ann a place to live after Ann leaves her abusive husband. Ann leaves her husband but then, not wanting to impose on Bob’s hospitality, tells him he needn’t follow through on his promise. (She is one of those guilt-ridden people who has trouble accepting kindness of any sort from anyone.) This decision is not in Ann’s best interest: she has nowhere else to go, and there is a real risk that, if she doesn’t stay with Bob, she will end up going back to her husband through sheer necessity. I would think that Bob is obligated to insist that Ann stay with him (he can’t force her, of course, but he can try to persuade her) despite Ann’s decision to release him. Cases like this are unusual, because most people are good at estimating what is in their interest, and are better pursuers of their interests than Ann is. But they seem to show that even in the ordinary promissory cases what is doing the work, where someone’s duty is annulled, is not the fact that the promisee has released the promisor, but rather the objective facts about the promisee’s needs and interests.

  9. I think I mostly agree with Troy’s last comment, though in the promising case, it seems that we have two duties, one based on a promise that has nothing to do with Ann’s needs, and the other based on Ann’s needs.
    Here’s an analogy that may help regarding the releasability issue: One thing that has worried some about S. Africa’s reconcilliation project is that,in effect, the particular victims involved in the attrocities committed there, as well as the new regime, did not have the standing to ‘release’ the perpetrators from the justice they so richly deserve. True, there are powerful pragmatic considerations that support the amnesty programs (a thought that has Kant rolling over). But one feels something of an outrage that these killers are allowed to walk the streets. If you find this thought compelling, and I do, then it makes sense that duties we have to one another cannot be nullified simply by our sayso. And if that is the case, then the idea that you can’t have duties to yourself can’t be argued for on the grounds that the object of a duty can release the agent at will.

  10. One more thought (pardon my bloggorhea). There is another issue behind our discussion: just what does it mean to have a duty *to* someone (as opposed to merely *regarding* someone). I may promise my neighbor that I’ll watch her kids. My duty is *to* her, but *regarding* her kids. In this case, my duty is *to* her because she, not her children, has a right to my performance. The moral principle of promise keeping, when spelled out, specifies that she has this right, or so I think. More broadly, who or what you think can and cannot have a duty *to* depends in large part on who or what you think can and cannot have a right. Who or whatever can have a right to someone’s performance can generate an obligation to her. She (or it) comes to generate the obligation to her by coming to possess the right to the performance. How she comes to possess a right to the performance is a further issue, and depends on your favored moral theory.

  11. Robert’s right to point out that we often do say things like “we owe it to ourselves to X.” I’m not so sure, though, that these are moral duties, rather than prudential duties (or counsels). Either way, moreover, I think we ordinarily tend to take a more casual approach to such duties than to other-regarding duties. For example, if I break a promise to a friend to help her with some project, she might rightly be angry with me. But if I break a promise to myself to tackle some project, it seems natural to say that both I and the friend would take more of an “Oh, well” assessment of my broken promise. If this is right, I think it goes some length to vindicating Uriah’s armchair psychological diagnosis of what happens with putative duties to self.
    Furthermore, let’s grant for a moment that there really are duties to self, and that when the agent releases herself from such a duty, she is going extra-moral (like the chess player who lets the opponent keep the queen). I’m not sure that even this means she is “not being moral.” Doug’s got a nice paper (linked to here: http://peasoup.typepad.com/peasoup/2004/06/why_most_moral_.html) arguing that non-moral reasons can “counterbalance” moral reasons, and that morality itself must make room for this possibility. So, if he’s right, it would not be a case of not being moral if I go with my all-things-considered reasons at the expense of my moral reasons to act.
    Troy thinks that the child-parent case can be equally captured by, say, a spouse-spouse case. I don’t agree, since I took it that the reason we think it’s odd for the child to release his parent from a duty to the child is precisely that the child is incapable of making such a decision. Maybe there are other cases not involving children that are more directly on target, but those would need to be spelled out.
    But I do find Troy’s case of assisted suicide more compelling, i.e., that we wouldn’t think that a patient had really released an agent from the duty not to kill if the patient’s reasons for wanting to die were wacky or ill-informed. Yet couldn’t the “release” model shore up a hole by saying that for a patient to legitimately release an agent from a duty to the patient, the patient must be rational, or well-informed, or whatever idealized conditions seem appropriate? If, then, the patient makes such an idealized decision, doesn’t that release the agent from the duty to not assist in killing her?
    It’s interesting, with the Bob and Ann case, that Bob can only repeatedly insist that, but not force, Ann to stay with him, as he promised. That seems to suggest to me that in the end the duty of keeping his promise is a wholly discretionary matter for Ann. So Bob actually seems to have two duties (slightly different ones than Robert pointed to, I think): (1) To keep his promise, unless Ann releases him; (2) To convince Ann not to release him, as one instantiation of a duty of beneficence, unless Ann releases him from that, too. (E.g., “No, really, Bob, I appreciate your concern, but I’d rather try to go it on my own for now”). At which point the right thing to do would be for Bob to lay off, right?
    The torture case – like the S. Africa case – is much harder, I think. Here I do have an intuition that the victim cannot release the agent from the duty. And it is in these very dire cases where, as Troy suggests, the content of the reasons for releasing the agent, not merely the patient’s say-so, seem relevant. These are the kind of “extraordinary” cases I wanted to put to the side initially. But I do think the “release” side of the debate can say a couple of things here. First, as extraordinary, they might not bear on more ordinary cases. Second, I don’t think the content of the reason is the only thing that matters: the release still matters. So if there are reasons to let the patient be tortured, but the patient doesn’t see those reasons, that fact of non-release might override any independent reasons to release. But I’m not sure what to think about these extreme cases.

Comments are closed.