By In Discussions, Ethics Discussions at PEA Soup, Moral Psychology, Normative Ethics, Practical Rationality Comments (41)

Ethics Discussion at PEA Soup: Abe Roth’s “Intention, Expectation, and Promissory Obligation,” with a critical précis by Sarah Stroud

Welcome to what we expect will be a very interesting and productive discussion of Abe Roth‘s “Intention, Expectation, and Promissory Obligation.” The paper is published in the most recent edition of Ethics and is available through open access here. Sarah Stroud has kindly agreed to contribute a critical précis, and it appears immediately below. Please join in the discussion!

Roth’s interesting and wide-ranging article covers a lot of ground. It may be helpful if I disaggregate some of the bricks in the wall, sequentially identifying what I take to be the important themes and moves in the account.

Giving the promisee a more important role. Roth’s opening move seems to me of interest even independent of what he later does with it. His suggestion is that we should beef up the distinctly secondary role which has always been assigned to the promisee in the creation of a promissory obligation binding on the promisor. Roth notes that some promises may in fact be unwelcome to the promisee. If promissory obligation were generated automatically, even when the promise is unwelcome, this would in effect harness “the resources of moral motivation, obligation, blame, sanction, etc. associated with promising … in a way that works against the promisee” (p. 89). Surely it would be more natural if promisees had available to them “a mechanism to block the generation of an obligation when the promise is unwelcome” (ibid.).

Roth maintains that we do in fact possess such a mechanism, namely rejecting a promise (of which accepting a promise is the mirror image). Granting promisees the normative power to accept or reject a promise gives promisees some control over which promissory obligations are generated, a desirable result. And Roth argues that it would not work as well to make the creation of promissory obligation dependent, not on an act of acceptance or rejection by the promisee, but merely on the promisee’s psychological states (such as whether he wants the promisor to ϕ). Once we focus on the promisee’s option to reject or accept a promise, we will see (says Roth) that rejection of a promise blocks the generation of a promissory obligation, while acceptance of a promise “help[s] to ensure that a promissory obligation is generated” (p. 89). This picture thus “gives a more intimate and active role for the promisee in generating promissory obligation” (p. 110); it portrays promissory obligation as a joint creation of the promisor and the promisee, rather than a solo act by the promisor.

Understanding the acceptance of a promise in terms of an intention that the promisor ϕ. Roth’s next move is to offer a novel account of what constitutes acceptance of a promise. He proposes that for B to accept A’s promise to ϕ is for B to (i) intend A’s ϕing and (ii) communicate that intention to A. Roth notes that this makes acceptance of a promise partly but not wholly an “internal” matter. Clause (i) of the proposal refers to an “internal” psychological state of B’s; but since clause (ii) of the proposal adds an “external” or public component, B’s psychological states alone do not suffice for acceptance. In this sense Roth’s proposal could be considered a hybrid model of acceptance: it is partly an “internal” mental state and partly a public act.

All the same, Roth’s subsequent analysis and arguments turn mostly on clause (i). (This is reflected in the name he gives his proposal: Accepting as Intending.) And while Roth thinks clause (ii) may in the end drop out as redundant (p. 93), he insists on the necessity of (i): saying (for instance) that you accept a promise does not constitute your accepting it, if your speech act is not in fact accompanied by an intention on your part that your promisor ϕ. That is insincere acceptance, says Roth, which is not a subspecies of acceptance and does not generate promissory obligation (n. 17, n. 33). This worried me a bit: I wondered how well the “internal” component of Roth’s hybrid conception squares with his argument that due to epistemic and other difficulties, “internal” psychological states are ill-suited to be the basis for promissory obligation. Roth’s hybrid conception of promise acceptance also seems to contrast with the dominant understanding of normative powers in general—even though the former would appear to be a species of the latter, if it does the normative work Roth assigns to it. Exercising a normative power is typically understood, I take it, as something we do, via “performative” speech or other public acts. I would not have thought it is generally conceived as a hybrid of what we do and what we are thinking “internally.” This would have been an interesting point to address.

Any worries one might have about (i), though, must be balanced against the arguments Roth offers for the idea that accepting a promise can best be understood in terms of intending the promisor’s ϕing. 1) Just as acceptance by a promisee settles a certain matter that prior to acceptance was up in the air, a salient aspect of intention as a mental state is that it too settles a practical question that previously was open. 2) It is possible for acceptance of a promise to lead to fulfillment of the promise via an intuitively deviant route. The similarity with the problem of “deviant causal chains” should strike us. 3) The acceptance of promises is subject to norms, e.g. of consistency, to which intentions (but not other conative or desiderative states) are also subject. One can’t, Roth says, reasonably and consistently accept two promises which are inconsistent (in the sense that A’s fulfilling his promise is incompatible with C’s fulfilling hers). If accepting promises involved intending their fulfillment, we would have an elegant explanation of this fact, since it is well known that one cannot reasonably and consistently intend two incompatible things.

Roth gives a fourth reason in favor which I’ll mention later. He also spends some time on why it is not problematic to think that someone can intend a different person’s action. We should not, he argues, retreat to a weaker content for the intention involved in promise acceptance.

Cognitivism about intention as entailing non-evidential warrant for belief. By general consensus, while one can perhaps desire to do something which one knows or believes one cannot do, one cannot intend to do such a thing. Many philosophers have concluded from observations such as these that intention involves belief: if I intend to ϕ, I thereby believe that I will ϕ. (Contrapositively: if I don’t believe I will ϕ, I cannot be said to intend to ϕ.) Roth calls this idea Cognitivism about Intentions (although some writers use that label to refer to a much stronger thesis, according to which intention somehow reduces to or is merely a species of belief). Roth points out a little-remarked-upon implication of cognitivism about intention: one which (again) I take to be of interest independent of what Roth does with it. When I decide to ϕ, and thus form an intention to ϕ, I thereby come to believe that I will ϕ, since intention (according to cognitivism) implies belief. But notice that I do not form the belief in question on the basis of evidence. Nor does the warrant for my belief derive from evidence: rather, in such cases my belief “has a warranted status by default” (p. 109). Cognitivism about intention thus seems to imply that some of our beliefs, at least, possess a non-evidential form of warrant.

Putting the above points together to respond to a well-known worry about Scanlon’s account of promissory obligation. On T. M. Scanlon’s account, the key trigger for promissory obligation is A’s leading B to expect that A will ϕ. (Some further conditions are also required.) Let us suppose that A leads B to expect this by, specifically, promising (as opposed to via some other route). Still, it may seem mysterious how exactly A’s promising to ϕ leads B to expect that A will ϕ, and thus how promissory obligation can ever get triggered. If A can only be expected to ϕ because or insofar as she is obligated to ϕ, but she cannot be obligated to ϕ unless and until B expects her to ϕ, a worrying circularity looms. Much ink has already been spilled (notably, by Kolodny and Wallace) about this problem; Roth points out that the constellation of views he has put on the table offers a novel response to it. Suppose A promises B that she’ll ϕ. As we have seen, A is not yet under a promissory obligation to ϕ: that will become the case only if and when B accepts A’s promise. Suppose he does. Then, as per Accepting as Intending, B now intends that A ϕ. Since—by Cognitivism about Intentions—intention involves belief, B now believes that A will ϕ, albeit not on evidential grounds. We could summarize this process by saying that A has led B to expect that she’ll ϕ. (She initiated the whole process, after all, by promising.) Thus Scanlon’s theory will predict that promissory obligation has indeed been triggered, as of course it has.

I want to close by mentioning a fifth theme which runs through various of Roth’s arguments: “directed” obligation and the special status of the promisee. One of the distinctive features of promissory obligation is that it seems to be a “directed” obligation, with respect to which the promisee is in a special position. A promisor is not merely under an obligation to ϕ; she owes it, specifically, to the promisee to ϕ. If she doesn’t ϕ, she does not merely act wrongly; she wrongs the promisee. And only the promisee can release the promisor from her obligation. Roth thinks the considerations his theory puts forward—and especially his emphasis on acceptance by the promisee—can explain the above features, and more generally the “directed” character, of promissory obligation. The promisee, says Roth, is in a special position to object to non-performance because of his power of acceptance, which as we have seen gives him effective authority to settle (to the extent an intention ever settles) whether the promisor ϕs. I found Roth’s explanations of this idea rather dark and hard to follow, although this may just reflect how difficult it is to explain or illuminate the source of directed obligation. What is clear is that Roth thinks acceptance by the promisee is necessary for the generation of a “directed” promissory obligation. Absent (or prior to) acceptance, he suggests, any residual obligations around the penumbra of a promise would not be “directed” (see, e.g., n. 9, n. 18). This is a surprising claim in relation to the literature on promising, since everyone seems to think promissory obligation is directed, while not everyone thinks promisee acceptance is necessary for promissory obligation, and fewer still give it the prominent role Roth does.

Love this idea? Nominate it for the Annual PEA Soup Awards!

, ,

41 Responses to Ethics Discussion at PEA Soup: Abe Roth’s “Intention, Expectation, and Promissory Obligation,” with a critical précis by Sarah Stroud

  1. Well, I guess I’ll get the conversation started!

    I have a couple of concerns at the very start – with some of the motivation provided for taking promisees to have a more central role in the formation of a promise. I might tackle the epistemic part of this later. But for now, I want to ask about the autonomy component.
    Roth suggests that because a promisor is performing the content of the promise on our behalf, then we, as promisees, ought to have some choice as to whether this relation gets going in the first place – as opposed to merely an opportunity to end the relation (through release) or block it (through rejection). After all, we might bear some responsibility for what they do, or take on a duty to release them from the promise.

    Yet, think about gift-giving. Imagine that you give me a gift that is too big, and I have a duty to return it to you, since the sacrifice was too great. Or imagine the gift is a real fur coat, and I want you to know that gifts like that won’t be acceptable to me, so I give it back to you, or block the gift. It seems that being put in a place where we need to exercise this discretion is no great hit to our autonomy. So why would it be such a hit when it comes to promises?

    Also, someone might do something on my behalf without me incurring responsibility for what they do. Imagine that next week I am ill and cannot go to the polls. You were not planning to vote, but suspect that I was planning to vote for a particular candidate. So, you go and vote for that candidate on my behalf. You leave a message on my phone, telling me as much. And I hear it, but don’t want to bother calling you back – because it would take too much effort, or because I soon forget (or am bombarded with such offers – as a parallel to one of your complaints against Shiffrin).

    It seems to me that you are as responsible for having voted for that candidate as anyone else who voted for the candidate. I bear no responsibility. If we had made a joint decision that you would so vote, then maybe I would. However, it is possible for you to vote on my behalf without me bearing responsibility.

    Perhaps promises – when they have not been rejected, but are successful (Shiffrin’s view) could allow for circumstances like this, where a promise is performed on a promisee’s behalf, but if the promisee played no role in the formation of the promise (and didn’t have the time/take the time to consider whether she wanted to reject it) perhaps she bears no responsibility.

    What do you think, Abe?

  2. Abe,

    I enjoyed the paper. I’m interested in your thoughts on the following.

    You say that “something very much like the demand for means-end coherence is present for the acceptance of promises — again, just what we’d expect if acceptance involves intending what’s promise.” As you point out, if I accept your promise to phi, and subsequently learn that I must psi in order for you to phi, “[i]t would seem that I am under some demand to (intend to) psi, or else to release [you] from the obligation” to phi.

    That doesn’t look to me like an objectionable result, at least not at first blush. But there’s an odd result that might be lurking in the neighborhood. Here’s what I’m thinking:

    There is a plausible principle regarding the transmission of reasons between, on the one hand, the actions one intends and, on the other hand, the actions that promote the performance of the actions one intends, so that, very roughly: If I intend to phi, and psi-ing promotes phi-ing, then there is a (pro-tanto) reason for me to intend to psi. (For what it’s worth, a corresponding principle regarding the actions that frustrate, rather than promote, the actions one intends, seems equally plausible.)

    Suppose that’s right (it might not be, but I think it’s near enough the correct principle for present purposes). Suppose I accept your promise to phi. So, I intend you to phi. Now suppose I discover, not that my psi-ing is a necessary means to your phi-ing, but that my psi-ing would promote your phi-ing. By the principle above, there’s a reason for me to intend to psi. Plausibly (though there could be reasonable disagreement over this), the weight of the reason there is for me to intend to psi depends at least on two things: what we can call (i) the importance of your phi-ing and (ii) the degree to which my psi-ing promotes your phi-ing. And whether, all things considered, psi-ing is what I should intend to do will depend, in turn, on the weight of the other reasons there are for me to intend to do other things incompatible with psi-ing.

    Even so, there’s no principled reason why there shouldn’t be a case where things turn out like this: You promise (me) you’ll phi, and I accept. I discover that, although my psi-ing isn’t necessary for your phi-ing, it makes your phi-ing a whole lot easier: that is, it promotes your phi-ing to a very high degree. Moreover, your phi-ing is suitably important so that the reason there therefore is for me to psi is sufficiently strong that, although there are actions other than psi-ing that I have compelling reason to do (things to do with my own, non-promise-related ends, say), these reasons are all outweighed by the reason there is for me to psi that there is on the basis of the fact that psi-ing would promote your phi-ing.

    That does look odd to me, though maybe it doesn’t look odd to you, or to others who have no doubt thought much more carefully about promises than I have. Thoughts?

  3. Nathaniel, I’m trying to get clear on the objection. Is this a fair characterization of what you mean (with examples)?

    Imagine that I have one colleague who doesn’t ever commit to doing service work, and he finally agrees to pick up the speaker from the airport. It matters to everyone else in the dept, including me, that he keep his word.

    Claim A: It seems to me that I am rationally required to buy the ticket for the speaker, such that the speaker is at the airport to be picked up, if I intend for my colleague to pick him up. This is what Abe’s consistency claim entails.

    Now say that there is always terrible traffic on the way to the airport – something we all (the promisees and the promisor) know, and most of us have experienced, when the colleague makes his promise. I know that it will be easier for him to keep his promise, and he’ll be more likely to do it, if I go with him, so that he can drive in the car-pooling lane and make it a faster trip. But this would mean me losing the whole afternoon – and its not my turn, etc.

    Claim B: Surely, I am not rationally required to take seriously the idea (let alone give it priority) of going along for the ride, to make my colleague’s fulfillment of the promise easier or more likely – even if I care a lot about my colleague’s promise.

    I think that both A and B can be true. Are you suggesting that Abe’s mechanism for explaining Claim A would carry over to the other case, and preclude Claim B?

  4. Abe Roth says:

    Hi Hallie and Nate. Thanks for the comments. Nate, I have to get back to you after my class. As for Hallie, here are some initial reactions.

    Gift giving is interesting. You might think that gift giving will establish or reinforce a relationship and that, moreover, the recipient incurs some sort of obligation e.g. to reciprocate. I’m not necessarily endorsing that picture, but let’s just go with it for the sake of argument. Then, on such a view, it seems that gift giving could be even more of a hit on the recipient’s autonomy than the case of promising (in that it’s the recipient who incurs an obligation to reciprocate), and so even more in need of some sort of acceptance condition. But there doesn’t seem to be anything like a requirement of prior permission needed to give someone a gift. But, on second thought, it seems that gift giving is a transaction that requires uptake. Acceptance is literally built into the process. The suggestion is that your friend didn’t manage to give you a gift, which you then returned. You’re preventing the gifting in the first place. The case you describe, then, is one with an acceptance condition.

    I wonder about the case of someone unilaterally doing something on my behalf. I agree that in the voting case you describe, you are not responsible for the vote your friend has cast. But it’s not so much the worry about being responsible for what the friend does. He’s just doing something he wants to do; it’s his decision. He’s not making a commitment to you, binding himself. So if some difficulty comes up, he can change his mind. Whereas, if he made a promise to you, then he is obligated to you; and now you are more involved because you might feel the need to release him if somehow it was harder for him to vote – e.g. hours-long waits at the polling booth as a part of a tactic of voter suppression. And this responsibility might not be something you want to take on.

    I hope I’m getting at your concerns.


  5. Abe Roth says:

    Thanks to Sarah for a précis that makes perspicuous a number of my central points, and for raising good challenges. Let me try to address some of her concerns.

    1. It’s true that I want to raise the profile of the promisee and that one of the considerations in favor of doing so is that it addresses an epistemic worry about how the promisor is to know whether the promise is welcome. I figure that the active involvement of the promisee in generating the obligation by accepting it would be the best way to address this concern. That’s not to say that it is foolproof. Acceptance as I understand it is hybrid. It involves not only an overt communicative act but also an internal mental component of intention. Sarah worries that the latter component will generate the same epistemic worries I was seeking to avoid in emphasizing acceptance in the first place. It’s true that sometimes the promisee might accept, but would be better off declining and might really want to say no; the promisee may even have the relevant intention but her heart might not be in it. So I don’t think that the acceptance condition will necessarily handle all cases. But I don’t think that taking seriously the epistemic concerns requires us to provide a mechanism for infallible epistemic access to the will of the promisee.

    2. Given the role I assign to acceptance, I think it’s fair for Sarah to say that I see it as involving the exercise of a normative power. This is not to say that I’m committed to so-called normative power views that hold that one has the power directly to will into existence certain forms of moral obligation without, for example, invoking a convention or generating expectations in others. Although I am open to the possibility of such views (e.g. that of Seana Shiffrin), I was trying to remain neutral on that matter. In any case, Sarah worries that the exercise of a normative power must be a public act, not a hybrid of “what we do and what we are thinking internally,” which is how I think of acceptances.

    I take it that the worry is not so much the presence of a mental component per se. If an utterance of something like “I promise to ϕ” is going to be a performative that exercises a normative power, it had better be intentional act, and that will require the appropriate internal component. The worry, rather, seems to be that one cannot intentionally say ‘I promise’ while secretly harboring the intention not to exercise the relevant normative power of placing oneself under an obligation; at least, this will not succeed in cancelling one’s obligation. The public utterance (when accepted) suffices for incurring the obligation. That may be right for the case of making a promise; we want to protect promisees from insincere promising. But given the role of promising to provide assurance to promisees, it’s not as clear to me that we need a similar safeguard against insincere acceptance.

    I do need to address the issue about directed obligations. I’ll come back to that later.

    Hope this helps,


  6. Hi Abe – thanks for your response. I’m satisfied with those explanations.
    To pick up on your comment directly above, I’d like to make a point about your jointly-sufficient conditions for the acceptance of a promise: (i) the public, communicated part, and (ii) the internal part. (As you just said, there is reason for insincere promises to count when gauging the success of a promise, but there isn’t any reason for insincere acceptance to count.)
    Now, it seems to me that this part of your proposal can resolve Scanlon’s circularity problem – even to Killodny and Wallace’s satisfaction – without any need for the use of intentions. It can supply the evidentiary basis that K&W want. (Though, I like the way your intentions account avoids the need for this.)
    Maybe the acceptance of a promise comes in two parts – it is a process. The public, communicated part gets rendered first. B communicates to the promisor, A, that B accepts the offer to form an expectation – which means that A is required to act on the assumption that B expects A to ϕ. That is, B’s communicated acceptance of the invitation (even if he does not yet form an expectation that A ϕ) should be enough to ground a moral requirement that A “take himself” to be under promissory obligation – since, for all he knows, B will now form an expectation.(Being required to “take himself” to be under promissory obligation to ϕ is another way of A being required to ϕ.)
    Once A is required to “take himself” to be under promissory obligation, then B has a real reason to actually believe that A will ϕ. Thus, B can rationally form the expectation, and fulfill the internal criterion. Now, A really is under promissory obligation. This would mean that the real promissory obligation only begins directly after A receives the communicated component of B’s acceptance.
    So – my thought is just: the two-component account of promissory acceptance is enough, by itself, to resolve the circularity problem. Let me know if/how I’m wrong.

  7. wesbuc says:

    As pointed out above it is kinda weird to intend other people’s actions (or for that matter, their thoughts, or feelings, or dispositions, so on, the things we typically make promises with each other to change). If there is going to be an internal requirement for felicitous obligations, why not simply desire?

  8. Jack Woods says:

    I wrote out a long polemical screed here, but I guess I’ll just invite Abe to say more about why the internal aspect of intending is even prima facie for my acceptance of a promise giving rise to promissory obligations. This seems a complete non-starter, for many of the reasons Sarah outlines above, as well as general views about how performative utterances work. In my own work on promises, I take a general social convention line, but one which requires only that we don’t publicly reject the promise (since that’s a power given us by the social promissory conventions we have internalized).

    I suppose I can be brought around to thinking that not rejecting a promise is communication of my intention that the promisee keep it, though I’ll note that it does seem to me that if you promise me that you’ll do something, and I say

    Okay, sure, though I could give a shit whether you keep that promise or not

    then you’re still obliged. It’s just that then I’d be a real jerk to actually blame you for failing to keep the promise. But that’s a normal aspect of social sanctions—they can be legitimate, even if applying them makes you a bit of a jerk.

    (I really like the idea that accepting and rejecting promises is crucial for them to give rise to promissory obligations. This strikes me as right, at least of most of our promissory conventions.)

  9. Hallie,

    That might be an example, depending on the details of the case as you’re imagining it. In a slogan, the worry is that if my accepting your promise to me to X is a matter of my intending that you X, and quite generally (as seems plausible) intending gives us pro-tanto reason to intend to do that which promotes that which we intend, then accepting a promise is *way* more rationally committal than it might have seemed.

    (I didn’t use a case like this in my initial question, but here’s another way to bring out my sense that there’s something strange going on: suppose there is more than one agent to whom you might promise to X, both of whom will accept your promise, that the reasons in favor of promising to one as opposed to the other are on a par, and that you can only promise one of them to X. Then you have a reason to promise to the person who has the least on her plate that you’ll X (after all, then she’ll be available to do — what her reasons will then require of her! — that which helps you keep your promise, i.e., X).)

  10. Ted Hinchman says:

    I’m very sympathetic to Abe’s approach, but I too have written a long screed explaining why I disagree in the particulars. Since I’m probably more sympathetic to Abe’s approach than most of Abe’s critics will be, it may be useful if I try to say at some length where I depart from it. Anyone reading should feel free to skip around in what follows, looking for items of interest.

    My principal worry concerns the relation between promising and planning agency. On Roth’s account, a promissory relation presupposes the exercise of species of planning agency on both sides of the promissory relation: A, the promisor, invites promissory trust by representing herself as intending to @ [= ‘phi’]; and B, the promisee, accepts the invitation by himself intending that A @. My worry is that this is too strong on the promisee’s side, since I don’t see why B qua promisee need have an actual intention that A @. But the way I would develop the worry amounts to a friendly criticism of Roth’s proposal, since I don’t regard the worry as undermining Roth’s core emphasis on the bipolarity of promissory planning agency.

    I’ll develop the worry in a moment. Let me first briefly describe my preferred alternative conception of the promisee’s planning agency. On the view that I’m developing (e.g. in a paper soon to appear in an OUP collection on trust), B as promisee must provide uptake – lest the would-be promise ‘misfire,’ in Austin’s sense – by recognizing that A as promisor intends to give him, as promisee, a kind of planning reason: a reason to perform (or at least not to omit performing) acts that depend on A’s keeping her promise. So B’s planning agency is in play insofar as B enters into the illocutionary relation. But it is not in play through B’s actually having any particular intention, and especially not through his having any intention directed at A’s agency. What B must do, to institute the promissory illocution on his end, is recognize that A intends to give him a planning reason simply through issuing the promise. (I’m somewhat simplifying the quasi-Gricean reason-giving structure here. The reason itself would be grounded in A’s status as relevantly trustworthy.) So there is that much of an acceptance condition on promising, and the promisee’s acceptance does manifest a species of planning agency. But it need not amount to the species of planning agency wherein B would intend that A perform the promised act.

    My principal reason for preferring this alternative view of the promisee’s planning agency is that I believe it gives a satisfying explanation of how we can have promissory normative powers. (In brief, A incurs a promissory obligation not through a sheer act of will but through entering into this reason-giving relation – a reason-giving relation that derives neither from expectations nor from broader social practices but instead rests on A’s presumption to do for B what B could in principle have done for himself by forming an intention to perform the act himself. Big issues in play here.) But I also think that Roth’s proposal confronts some more mundane problems that my alternative avoids.

    The chief mundane problem is that Roth’s approach makes promising too similar to shared intention. I actually agree with Roth’s account as applied to the promissory dimension of follow-through on a shared intention. But what’s plausible in that application doesn’t generalize beyond that sort of case to the familiar sorts of case that give point to the concept of promising. In these central cases, there is no shared intention or project in play, and there is therefore no point in regarding the promisee as needing to ‘settle’ whether any performance will go forward. All we need in these central cases is illocutionart uptake sufficient to ensure that the would-be promissory speech act doesn’t misfire.

    I’ll begin with the shared-intention cases. It is an important feature of a shared intention that parties to it rely on each other to perform distinct components of the coordinated act. When my spouse and I make dinner together, for example, we divide our labor through a series of implicit and explicit agreements: ‘Shall I chop these carrots?’ I ask, as a way of offering to chop. ‘Sure,’ she replies, ‘you do that while I wash the rice.’ ‘Okay,’ I say, thereby launching follow-through on our shared intention to make dinner together, an intention that (if one wants to imagine it as future-directed) has shaped our planning agency over the course of the day. The division of labor generates interpersonal obligations that look a lot like promises. If I get distracted and fail to do what we’ve agreed is my part in the joint performance, my spouse may criticize me in terms that looks just like criticism for failure to keep a promise: ‘Hey, you said you were going to chop the carrots,’ she can say; ‘I was counting on you to actually do it!’

    I see no reason to deny that these are promises – though it is important to see that they are promises of a special sort. Since the promises arise in the context of a shared intention, it is plausible to hold that each promissory offer to perform must be accepted by the promisee in a way that crucially involves the promisee’s intending that the promisor pull this bit of his weight in the joint performance. It is plausible to regard the promisor as proposing and the promisee as disposing here because the entire performance is structured by the need to maintain a sufficient degree of reciprocity and mutuality – lest the intention to make dinner together degenerate into a different performance: she makes dinner while he ‘helps out.’ (Compare: A and B go for a walk together, versus B ‘keeps A company’ during part of what amount to two distinct ambulations.)

    So that’s one sort of case, for which I think Roth’s account works quite well. At the other extreme is a very different sort of case, in which a would-be promise fails to come off – and the question now is why. Imagine this promissory misfire: you’re at the gym, absentmindedly watching a stranger struggle to lift the weight before her. She sees you watching, returns your gaze, and addresses you with what appears to be a promise: ‘I’m going to rest for one minute, and then I’ll lift it. You just watch – I promise.’ (I think of such cases as ‘paternalistic promises,’ since the would-be promisor is trying to exploit the institution of promising for an extra reason to do something that she would like to do for self-regarding reasons – just as we might use the legal system to provide a legal reason to do something (e.g. wear a seatbelt) that one has independently sufficient reason to do.) Thus addressed, you needn’t buy into the illocutionary relation – and our question is what it is for you to refuse. Does your refusal lie along the lines of Roth’s approach, in your failure to intend that this would-be promisor lift the weight? I think the problem is more fundamental: there is no realistically conceivable respect in which this would-be promisor could give you a planning reason. It is in that way that the would-be promise is idle as a promise, an Austinian ‘misfire’: not that it doesn’t lead you to intend that this person lift the weight, but more fundamentally that this person’s invitation to trust completely fails to engage any realistic extension of your planning agency. She appears to be inviting you to rely on her to lift the weight, but the invitation can only leave you puzzled what form your reliance is supposed to take. You cannot recognize your planning needs as in any way engaged by the promise.

    We might contrast such a misfiring because purely paternalistic promise with a promise that can come off despite containing a paternalistic element. As a teacher, I’m occasionally addressed by a student ‘promising’ to study harder for the next exam, and I recognize that the primary point of such a promise is to provide the student with a paternalistic reason to study harder. Does the promise thereby misfire? Not necessarily, since I do planfully care about my students’ academic success. So a student can promise me that she’ll study harder, and I can enter into this promissory relation by recognizing how it engages my planning needs – e.g. my need to keep the course going with new material rather than slowing it down to give remedial help to struggling students. But this case may appear to resemble a case of shared intention, since the promise appears to be rest on the division of labor that informs this joint undertaking that I share with my students. (Does it matter if we shift from a student to one of my own children promising to study harder? That can obviously give me a planning reason (e.g. to cease looking into hiring a tutor for the child), but one might counter that the planning agency must be likewise understood against the implicit joint performance that we call my ‘raising’ this child.) What of a case in which there is no such joint undertaking in view?

    Many paradigmatic instances of promising unfold with no hint of any shared intention, project, or undertaking. If my estranged brother calls out of the blue to ask for a loan, promising to pay me back on the 1st of the next month, I can accept by recognizing that he intends to give me (and would, if relevantly trustworthy, succeed in giving me) planning reasons to perform acts that depend on his paying me back on the 1st. And that’s it: I give him the money, and he is now obligated to pay me back on or before the 1st. I have no idea why he needs the money, and he has no idea why I need the loan repaid by the 1st. I love and trust him, but we’re otherwise estranged. Other than giving him the money, as an expression of how I understand his speech act (viz, as aiming to give me a reason to count on him), I need not initiate any joint or shared performance. I am not yet actually doing any planning with regard to his promise. I merely recognize his intention to give me reasons to plan. When actual planning comes — if it does — our estrangement will ensure that my planning will all be on my end, about my actions, as I make adjustments to my finances on the assumption that his word (backed by relevant trustworthiness) gives me a reason to expect repayment on the 1st.

    Okay, I’ll stop now — this is already too long for a blog comment, and I’ve only pasted part of my reply… But I think I’ve said enough to show how even this sympathetic reader was provoked into a series of what I hope will prove to be productive disagreements with this very stimulating paper. Thanks, Abe!

  11. Ted Hinchman says:

    Yikes, what happened to the paragraphs? That was supposed to be 11 paragraphs! I don’t appear to be able to edit….

  12. Abe Roth says:

    Sorry for taking this long to reply, Nate. It’s good to press me on this. You’re right that my understanding of acceptance involves rational commitment, roughly to the extent that intentions involve rational commitment. That shouldn’t be too surprising given that there’s something incoherent about accepting mutually incompatible promises from two different parties – especially if the point of promising is to receive assurance about the things being promised.

    I think that the main reason why folks might balk at this point is that rational commitment could very well require taking steps to facilitate the promisor’s ϕ-ing. But the intuition is that promising often is meant to relieve the promisee of that sort of responsibility. This comes out in Hallie’s example of the colleague who shirks his service duties. Hallie cares very much about him living up to his promise to pick up the speaker. In accepting his promise to do this, must Hallie go to great lengths to make it likely that he’ll come through – to the point of accompanying him to take advantage of the carpool lane? That does seem like a bad result. After all, the point of the promise was that it would relieve his colleagues from having to do as much service. So this is the sort of case where it is clear that there is an implicit element to the promise – namely, that the colleague would take care of the matter on his own; and if Hallie were to have to accompany him, then he really wouldn’t be keeping his promise. In other situations where it’s important for me that you keep a promise, but where it’s not a matter of saving me from extra work, then it seems reasonable to think that I am under some rational pressure to take some steps to help you keep the promise.

  13. David Sobel says:

    Sorry Ted, I tried to add an extra line between paragraphs to fix it but it had no effect. Will look into this.

  14. Abe Roth says:

    Hallie, I will probably have to think more about your suggestion that we can solve the circularity problem without invoking intentions. But here are some initial thoughts. On your proposal, B accepts the promise by first communicating acceptance. Then you say that A now must ‘take himself’ to be under an obligation. Why is this? For Scanlon, we need expectation in the promisee in order for Principle F to apply. And for all A knows, there is no expectation yet – since A is not yet under the obligation. Remember, we’re thinking that the promissory obligation was supposed to bring about expectation. Taking oneself to be under the obligation is not the same as promissory obligation, so it’s not clear that B is yet warranted to expect that A will perform. So maybe there’s some other principle that’s relevant here, one that applies when, for all you know, someone may have formed an expectation that you will ϕ. Scanlon has formulated other principles concerning the creation of expectations regarding your actions, but Kolodny and Wallace have argued, rightly I think, that those other principles don’t actually require you to act to conform to those expectations. So, the applicability of those other principles won’t serve as evidence for the expectation. K&W, of course, will invoke a convention at this point; triggering some convention and the the possibility of associated sanctions would indeed provide evidence for the expectation. But Scanlon wanted an alternative to conventionalism.

    Hmm…that was a bit compressed. The basic worry is seeing exactly why one should “take oneself to be under an obligation” and what exactly this is supposed to mean.

    Hope this helps. I’m probably going to have to think about this some more.

  15. Abe Roth says:

    Hi Webuc (Wesley?)

    Yes, intending another’s actions sounds a bit odd; so does acting directly on another’s intentions. But I assure you that it sounds less odd the more you think it and say it. And I’ve thought this way a fair bit because it is my favored way of thinking about what goes on in shared agency. In any case, you ask why not have desire as the internal component of acceptance. Part of the thought was that the rational requirements on intentions (like consistency and coherence) seem to me to match those on promissory acceptance (despite Nate’s concerns). It also seems to me that in accepting a promise to ϕ one is in effect settling the matter of whether the promisor will ϕ. That sounds more like intention than desire. (If it looks like a duck, walks like a duck, swims like a duck, and quacks like a duck, then it’s probably an intention.) And it connects with the fundamental point of promising – namely, to provide assurance that one will act in a certain way.

  16. Abe Roth says:

    I meant to put up these remarks about directedness earlier; they might bear on Ted’s remarks, which I have yet to address.

    My discussion in the paper of directed obligation is rather compressed and it is not surprising that Sarah would have found it hard to follow. So let me try to say a bit more here. Some find it natural to describe promissory obligation as directed or owed to the promisee; it’s an obligation-to-the-promisee to ϕ – say, to do some shopping. But what does this mean? It’s not so much that the promisee stands to benefit from the shopping. After all, it might be that the promise is to do the shopping not for the promisee but someone else, e.g. the promissee’s homebound neighbors, in which case the neighbors are the ones that really stand to benefit from the shopping. Rather, the directedness of my obligation reflects the special normative status the promisee has with respect to my obligation. There are several markers that are evidence for this special status. First, as as Michael Thompson pointed out, if I fail to do the shopping (without any good excuse) I not only act wrongly; I wrong the promisee. Second, though anyone might judge me for not living up to the promise, the promisee would seem to be in a special position to criticize me and to urge me to follow through. I have some responsibility to do the shopping, but I am more specifically responsible or accountable to the promisee to do it. And this is related to the third marker, which is that only the promisee can exercise a normative power by releasing me from the obligation. In these ways, the directedness of the obligation reflects the distinctive status of the promisee with respect to the promissory obligation, something not had by third parties.

    An account of directed obligation must articulate what it is about the promisee that distinguishes her from third parties and which underlies these several markers of directedness. I see the directedness of obligation fundamentally as a manifestation of agency. More specifically, it’s a matter of the authority one exercises in deciding or settling some matter by forming an intention. (It’s not so much the causal power that is also involved in intention; that won’t do the work I need it to do.) Often this sort of authority is exercised over oneself, or one’s future self. But sometimes this sort of authority over what one does can be had by others, for example in shared agency, or cases where one acts at the behest of another. If you have this authority over me and issue an intention for me to act on – commanding me, if you will – then all the markers for directedness seem to be present. First, you are susceptible to a distinctive form of wronging when for no good reason I fail to comply. Perhaps wronging is too strong a term. But there is something similar to it insofar as it is directed. There is a sense in which I let you down (just as I let myself down when I fail to act on my own intentions). Second, as someone who issued the command, you are in a special position to criticize me on my performance or lack thereof. I have a general responsibility to do what I’ve been urged to do, but I am more specifically accountable to you. What you say to me on the matter is part of the exercise of your authority; whereas, what anyone else says to get me to comply would be advisory at best. Third, you have the normative power to rescind the command; no one else does. (In the paper I tried to emphasize that such markers of directionality are absent if, lacking this authority, you seek merely to cause me to ϕ.)

    So, the proposal in the paper is to account for the directedness of promissory obligation in this way. Mere promissory acceptance is not doing the work here. It is, rather, what underlies that acceptance: the promisee’s legitimate form of (limited) agential authority over what the promisor does.

  17. Daniel Star says:

    Hi Abe. I wanted to press you to say more about why we should think that the beliefs that accompany intentions are formed on the basis of insufficient evidence, hence epistemically unjustified (a claim you make in section VI of your paper). As Sarah points out, you define cognitivism in a fairly weak way, as simply the view that when an agent intends to X that agent believes that they will X (or that it is likely they will X), and not the view that intentions just are a special class of beliefs. The weak view, at least, doesn’t seem to require that the relevant beliefs are based on insufficient evidence. As you mention in the paper, the (weak) cognitivist can here appeal to evidence of a fairly general kind. You are unmoved by this response, but I don’t see why. In the middle of winter, I may find myself lacking any evidence about the weather tomorrow, in particular, yet still be justified in believing it will be cold tomorrow. Similarly, when I now form the intention to empty the dishwasher tomorrow morning, I am justified in believing I will do so not because of any evidence concerning tomorrow morning, in particular, but because I know that when I intend to empty the dishwasher, I empty the dishwasher. You say that such evidence can not be enough to explain why it is appropriate to form particular intentions, and this seems true, but if we don’t view the beliefs that accompany intentions as themselves intentions, it doesn’t seem like the beliefs in question should be viewed as unjustified.

  18. Hi Abe,

    This is a really rich and thought-provoking paper – I’ve read through it a couple of times now, and feel I am still grappling with it at a very basic level. The paper shines a spotlight on the nature and role of the recipient’s act of acceptance – something I wholeheartedly agree we should be paying more attention to when we think about promising. And it offers a novel and compelling account, appealing to some familiar – if not always widely shared – ideas in philosophy of action. But to say only this is to undersell the significance of the “Accepting as Intending” thesis which, the more I think about it, the more I suspect really challenges us to think quite differently about other, more “familiar” features of promising, too.

    One way in which I’m beginning to think this might be so is that I see more tension between what you’ve argued and Scanlon’s Principle F than you evidently do. It seems to me that features of Acceptance as Intending may count against thinking Principle F, or anything at all like it, can actually illuminate the ground of promissory obligation. Let me try to say a little more about why I think this.

    I’m with you in thinking that the recipient’s reasons for acceptance are “non-evidential” – I think you’re right to see this as the key to solving the circularity (aka Can’t Get Going) problem. Moreover, I think we must say that promisor, through her performance, provides her audience with some new reason for forming the requisite attitude (according to you, an intention which, assuming Cognitivism, involves a belief), though not by providing her audience with *evidence* about what she will do. I also think this fits the phenomenology, so to speak: one doesn’t ordinarily decide whether to take someone’s word by considering one’s evidence that she will act as promised (which is not to say that one does it by ignoring one’s evidence, either). Another bonus: it serves to unite promising with other species of word-giving, at least if one is attracted to certain views about the non-evidential nature of the reasons for belief provided through testimony.

    But of course one *can* choose to treat the promiser’s performance as evidence of what she is likely to do – nothing prevents interpreting the promisors behavior in this way. And, in cases where it seems good evidence, one may then go on to form the expectation (i.e. belief) for that reason, rather than coming to intend the promiser’s action for the non-evidential sort of reason you have in mind. The expectation based on treating the promiser’s act as evidence of what she will do would seem sufficient to trigger Principle F. Yet I think that forming such an expectation doesn’t, on your view, constitute acceptance. (And I agree, it shouldn’t…) Rather, in order to accept your word, and thereby play my part in producing your obligation, I must intend your action for the very special, non-evidential sort of reason you take yourself to be able to offer me.

    Here’s a different way of making what I think is the same point. On page 98 in discussing directedness, you write “B’s special position for objecting is at least in part a reflection of B’s (authorized) intending of A’s Φ-ing.” The parenthetical “authorizing” seems crucial. A is bound (indeed, bound to B in particular) not just because she successfully leads B to intend her action any old way that is effective; it is rather because she does so in a way that constitutes the authorizing of that very intention and, in so authorizing it, she comes to stand in some special moral relation to her audience – one she does not stand in to others, even if she plays the very same causal role in producing otherwise indistinguishable attitudes for them.

    I hope that makes at least a little sense. I should stress that I’m very sympathetic to much of what’s motivating you, as well as to many of the key moves you make in the paper (I think I can vie with Ted for the title of “least critical critic”…). The point about the tension with Principle F is not meant to put pressure on your view, but rather to suggest that perhaps the view has broader implications than you let on: it points away from Scanlon’s way of thinking about the ground of obligation (and, I suspect, towards another way – though that part will have to wait).

  19. Wesley (Wesbuc):

    One thing that helped me understand the idea of “intending another person’s action”, after my frist read of the paper, was to think about cases of volitional coercion – where I create a circumstance, by taking away a conjunct of options that is your first choice, wherein you are rationally compelled to do what I want you to do. For instance, if I threaten to kill your family if you do not burgle a casino, then I am appealing to your rationality – that you will choose your best option (the burglary) from your remaining options. It seems fair to say that I intend for you to burgle.

    After making the threat to you, and fully expecting your compliance, it would make sense for me the answer the following question affirmatively, “Do you intend for the casino to be burgled?” It might be more helpful to imagine that my buddy wants to know my plans for when I am going to flee the country. “Do you intend for the casino to be burgled BEFORE you leave the country?” Yes. Yes I do. Because I need to transfer the money into my account before leaving.

    This was helpful for me in understanding Abe’s point, for the following reason. I could easily conceive of accepting a promise as intending for someone else to act for the same reason as, in coercion, I intend someone else to act. When I have made a promise, I have participated in making it the rational thing for them to do (in virtue of the obligation that they have incurred).
    Of course, Abe says he cares that the promisor is not just a puppet of the promisee. In coercion cases, there is a strong sense in which the coercee is just a puppet. But in promising cases, the promisor invites the promisee to restrict his reasonable options.

  20. Abe Roth says:


    There are some aspects of promising, such as the directedness of the obligations, that are not addressed by conventionalist approaches, at least as standardly understood. That’s not to say that there are no such conventions. And, depending on the details of the convention, there may be different ways in which the convention-based obligation may be triggered. On some versions, those obligations might be triggered without the promisee having an internal element of intention. So, you might be right about the triggering conditions for the convention based obligations; but that need not be the case for what I take to be promissory obligations.

    That being said, I wonder whether something like the social convention line might be wedded to my conception of intention so as to account for the aspects of promissory obligation that are missed by the convention account. The thought would be that the convention provides a sort of social scaffolding that enables one individual to settle what another does, which is what I take to be happening in promissory exchanges. Given my view, I don’t think that this social scaffolding is necessary for all cases. But I don’t see why such a social scaffolding cannot also effect something similar (though it may be tricky to work out the details). In the paper, I suggested that my story about the promisee exercising agency with respect to the promisor can be wedded to the normative powers view of promissory obligation, and that it can also be attached to expectation based approach of Scanlon. So, what I’m suggesting is that I think that it can contribute to a conventionalist approach as well. In that case, one might trigger the convention by accepting the promise. And then the promisor would have an obligation to perform. But the obligation wouldn’t be directed to the promisee unless the promisee really takes seriously (read: accepts sincerely) the agency she can now wield with respect to the promisor’s ϕ-ing.

  21. Abe – this is a response to your last reply to me.

    General Clarification:
    When I said that the promisor should “take himself to be under obligation” after the communicated acceptance, I meant this: Once the promisee has communicated acceptance, than whether or not the promisee has actually accepted the promise (on the inside) is not something for the promisor to adjudicate. As long as it is possible that the promisee has internally accepted, then the promisor must do ϕ, just in case.

    Imagine an email correspondence.

    Step 1:
    You email me, promising to pick up my kid from school if I wish. Of course, upon getting the email, I don’t expect you to pick up my kid yet, since I haven’t accepted, and there isn’t yet promissory obligation.
    Step 2:
    I email you back saying “yes, please.” I don’t yet expect, at the time I’m writing, that you’ll ϕ. After all, you haven’t gotten my communication.
    Step 3 – your end:
    When you do get my email, you know that I couldn’t have expected anything AS I was writing. However, you don’t know what I expect, now that I have sent the email. It is possible that I expect you to ϕ; you know that I know that you know that it is possible that I expect you to ϕ. This means that, from your standpoint, it is possible that you are under promissory obligation. (You might never know if you are truly under promissory obligation as long as something internal to me is a necessary condition of the obligation). But, in this weak epistemic position, you must ϕ, to play it safe – and take me at my word.
    Step 3 – my end:
    As soon as I send my email (if I believe you will receive it) then I assume that you will take yourself to be bound to ϕ, for the reasons just articulated.
    Step 4:
    In light of Step 3 – my end, I DO form an expectation that you will ϕ, and so the promise actually completes.

    I guess my thought was: by separating out the components of “acceptance”, you make it possible to have an intermediate step of obligation, grounded in moral risk-aversion.

  22. Abe Roth says:

    Hallie –
    I think that the case of coercing someone to rob a casino could be one where the coercer intends the casino to be robbed. But, as you say, it’s not perfectly analogous to what I’m envisioning for promising, as you can imagine. One difference is that the coerced casino robber might not regard the coercer as having the relevant authority over him. Thus, if he figures out a way to avoid doing the heist and also saving his family, he might do so. He wouldn’t owe it to the coercer to follow through. The coercer uses the threat on the family to mimic aspects of agential authority over the coerced, but that isn’t a real instance of this authority. At least, that’s what I hope to say.

  23. Jack Woods says:

    Right, interesting. Thanks Abe, that’s helpful! I definitely think a portion of your story could be usefully tack-welded to some promising conventionalist stories. Though I do think it would put pressure on the internal aspect of acceptance. But I like the though that conventions partially structure the normative landscape here.

    Even on the flatter sort of conventionalism I favor, there’s a kind of conventional structure favoring the internal condition on acceptance. It involves the legitimacy of blame; it might be that if it comes out that the promisee insincerely accepted the promise, then that exculpates the promiser from otherwise warranted blame. Interesting. It’s a bit “spoils to the actual facts” for these kind of pictures.

    Anyways, thanks. Lots to think about.

    FYI: for a conventionalist account that takes seriously directedness, you might be interested in my forthcoming OSNE paper on the normativity of promissary obligation. It’s in the same volume as Hallie’s paper arguing against uniform explanations of promissory obligations.

  24. Jack Woods says:

    This deletion of spaces between comments is wicked annoying. Is that fixable?

  25. wesbuc says:

    Hey Abe, you say that part of the motivation to use promisee intention over the more natural sounding desire was that it provides assurance that one will act in a certain way or that it in effect settles the matter. But what assurance or settling does my intention someone else will do something actually give me? They could just end up doing whatever they want. Those theoretical aspects of promising seem much more likely brought about by the promiser having the intention, and perhaps that the promisee desire that they do.

  26. Abe Roth says:

    Ted, here are some initial reactions to your post. I’m sorry I haven’t managed to address everything you raised. Apologies for the somewhat disjoint character of this response.

    You’re right to think that I’m invoking something like an interpersonal form of Bratman-esque planning agency. You think that the picture works well enough for what goes on in cases of joint action. I’m happy to agree with that – although I prefer not to talk of directed or contralateral commitments between participants as promises. But you don’t think that my picture works for the sorts of cases where there is no ongoing joint activity. One important difference that I see between the case of shared agency and the case of promising is that in the latter there is an important asymmetry: it’s the promisor who has the promissory obligation to the promisee, and not the other way around. In the case of joint action or shared agency there is a more symmetrical pattern of commitments. This is why I don’t think you can in a straightforward way apply a theory of joint agency to the case of promising (e.g. the way Gilbert does). Whereas, I think that the idea of one person issuing an intention to the other and the latter acting directly on the intention will be more promising for a theory of promising.

    What am I to make of the odd case where the person in the gym promises to me that she’ll lift the weight next time. I could not care less about whether she succeeds and don’t buy into the promise. What exactly makes this a special problem for me? My planning needs are, as you say, not in any way engaged by the promise. So I decline and don’t will her to do it. Is the worry that it’s possible for me to accept the promise even though I don’t care at all? But sometimes I can will things that I don’t care much about, just to settle the matter. But, having done so, it would be problematic for me to act in ways that conflict with the intention. So I can’t for example accept the promise from the person in the gym and then prevent her from lifting the weights, or accept someone else’s promise to stop her.

    You agree that promisee uptake is needed for promissory obligation to ϕ to be triggered, but for you this only involves B recognizing (and presumably accepting?) that A in promising to ϕ intends to give B a planning reason. Once the obligation is in place, then B has a reason to take for granted that ϕ will occur. That is, B can go on to do things that depend on ϕ-ing. I’m not sure what the force of this planning reason is supposed to be. On the one hand, it’s described as a reason. But on the other, it doesn’t seem to have prima facie normative force but something more conclusive; after all, B can presume on its occurrence in deciding to do other things that depend on it. So it seems to have a normative force of an intention. Moreover, it seems to be something that is, in the end, put in place by B. Finally, I’m going to presume that you agree that, barring some significant change in circumstance, A cannot unilaterally take back this planning reason; it’s normally B that can give it back, so to speak. I think we can agree on these things, even if you want to resist calling it anything like an intention. So it would be good to know exactly how it falls short of intention. Why would talk of intention be too strong here?

  27. Heath White says:

    Hi Abe,

    Here are two cases. I wonder how your view would handle them.

    1. Manny and Tito are boxers who will compete for the championship. They are each dating Lola, and each promises her that he will win the match and use the winnings to take her on a nice vacation. I think she can accept both promises, because *her* planning is not conflicted no matter who wins. However, both promises cannot be fulfilled. Does that make for a problem for your view?

    2. Zacchaeus is a Roman tax farmer who has been extorting money from his fellow Jews to get rich. However, he repents, and then vows in public never to act like that again, and to repay anyone he has overcharged. To whom are such public promises made? The listeners? Every member of the public, listening or not? It doesn’t quite sit right with me to say that everyone in the social group intends Z to keep his promise, or has to undergo a conscious mental act of acceptance. Anyway, how should we understand public promises?

    Thanks for any insight you can give.

  28. Ted Hinchman says:

    Hey everyone, I just discovered that when you select (e.g. with command-A) the text in these comments, the paragraph structure emerges. That makes everything (including my earlier too-long comment) a lot more readable! (Thanks for your reply, Abe. I’ll get back to you about it and related issues later this afternoon.)

  29. Abe Roth says:

    Hi Daniel. Sorry it’s taken me some time to get to your question. As I see it, the weaker form of cognitivism doesn’t identify intention with a kind of belief. It’s weaker as a thesis in the philosophy of mind, I suppose. But that doesn’t make it much if at all weaker in the sense of being committed to a less strong epistemic thesis. If I decide and intend to ϕ, that will require that I believe that I will ϕ, even if that belief is not identified with the intention. Background beliefs are relevant here, such as the belief that I am capable of ϕ-ing in certain circumstances, and that those circumstances will be in place at the appropriate time. If I can’t rely on these background beliefs, then it’s likely I won’t be able to ϕ. But even if these background beliefs are in place, they don’t suffice for me to infer that I will. That’s because I’m supposing (in contrast to beliefs about the weather) that the ϕ-ing will depend on my decision or intention – something like your example of emptying the dishwasher. At this point, you suggest that I can just form the intention and that will yield evidence that I will empty the washer, given the background belief that I usually empty the washer when I intend to. But why was I allowed to form the intention to empty the washer in the first place? The belief was not yet in place. The intention is the evidence we need for the belief, but the belief is needed in order to have the intention. So the belief cannot be based on the intention.

  30. Ted Hinchman says:

    Thanks again for your reply to my earlier comment, Abe. Here’s how I would continue the conversation. It leads quickly into deep issues about the normative structure of intention that I won’t be able to pursue here very far. (I’ll add dashes between paragraphs; perhaps that will help the paragraphs appear more clearly.)

    (1) My tactic in my earlier comment was to contrast your proposal, ‘Accepting as Intending,’ with another, ‘Accepting as (Illocutionary) Uptake,’ according to which it is sufficient for B to count as accepting A’s promise to @ that B understands that A intends that her speech act make available to B a special sort of reason for B to perform acts that depend on A’s @ing (viz. a reason grounded in A’s status as trustworthy in promising B to @). It is part of what it is for B to realize this species of understanding that he understand that the illocution grants him the normative power that you emphasize: he is in position to let A off the promissory hook. (And there are other respects in which this understanding is distinctively directed or second-personal, including, most crucially, both parties’ understanding of the distinctive nature of this planning reason.) So part of what I’m doing is proposing an alternative explanation of that normative power, an explanation that does not involve anything like B’s intending that A @.

    Each explanation proposes an internal condition on acceptance. I hold that there is no promise if A’s speech act fails to bring about the right sort of promissory understanding between A and B. And each explanation treats the internal condition as a manifestation of planning agency. The difference is that I don’t hold that B’s acceptance need involve any actual planning – including any plan or intention that A follow through on the promise. I think it’s enough for B to manifest an understanding that A intends to give him this promissory planning reason. And in any case that does not already involve shared intention I think it’s a mistake to ascribe to B an intention that A follow through on the promise.

    (2) What I said about the gym case in my earlier comment was designed to compare these contrasting explanations. Roth: B fails to accept A’s promise to lift the weight because A fails to form an intention that A lift the weight. Hinchman: B fails to accept A’s promise to lift the weight because B fails to understand how A’s promise could engage B’s planning agency: it is just not at all clear how A could be inviting B to count on her to lift the weight.

    In your reply above, you endorse my explanans but then treat it as a reason to imagine that B will fail to intend that A lift the weight. My question is: why not rest with the observation that B can see no engagement with his planning needs? You answer by arguing that when there is acceptance, as here we agree there is not, the acceptance takes the form of the promisee’s intending that the promisor perform, not merely that he fail to see how her promise engages his planning needs. My reply then is twofold: (a) Why do we need that extra bit – B’s actual intention, beyond B’s recognition of A’s intention to engage his planning needs? (b) Aren’t there plenty of easy cases in which the promisee doesn’t in any obvious respect appear to be doing any planning *for* the promisor? I intended my estranged brother case as illustration of (b). But I concede that this isn’t a knock-down objection, since you think there’s good reason to attribute the other-direct intention to the lending brother, despite what I’m claiming are the appearances. So pursuing (b) won’t settle the issue.

    (3) The real issue is (a), and in the paper you have a lot to say in response to (a) that I haven’t yet tried to engage. There’s a lot to discuss. Let me begin with a clarification of my proposal. Then I’ll say some preliminary things about the normativity of intention, and about how the normative structure of promising might appear to run in parallel.

    (4) On my account, the planning reason that a trustworthy promisor gives her promisee is not a conclusive or sufficient reason to do any particular thing. But it is a consideration that contributes its rational weight to other reasons that the promisee will have to act, and it may well be that this contribution proves decisive. So, for example, B may treat the consideration that his brother has promised to pay back the loan on the 1st as a reason for him, B, to go ahead and plan on paying rent on the 1st in the usual way, despite being currently short of funds to cover it. Without the promise, B would not simply plan on paying rent without taking extra steps to ensure that he will have sufficient funds. But his brother’s promise gives him a reason to form that plan.

    The reason that the promise gives B – again, assuming the borrowing brother trustworthy – is thus normatively isomorphic to the reason that B might give himself to plan on making a rent payment by forming an intention to deposit a check on the 1st (say, the check he’s now carrying in his wallet). In each case, B would not form the payment plan without taking extra steps to ensure that he will have sufficient funds. But B’s trustworthy promise gives him a reason to form the payment plan, in a way that parallels how B’s own intention to make a deposit on the 1st may give him a reason now to form the payment plan. Neither of these reasons is on its own a conclusive reason to form the payment plan, but each may prove crucial to the rationality of that plan. And each reason seems grounded in the trustworthiness of the practical commitment: in one case, of the brother’s promise to repay the loan on the 1st; in the other case, of B’s own intention to make a deposit on the 1st.

    (5) Alas, I’ve run out of time and will have to postpone developing the deeper point that I wanted to pursue about the normativity of intention. The point concerns your argument in the first paragraph of section IV of the paper (pp. 101-2). I suspect that the dispute between us will turn on how to understand your appeal to ‘defeaters’ in this passage. Since I’m out of time, I’ll simply paste what I wrote in the margin of your paper at the top of page 102. Here goes:

    ‘This misconstrues the diachronic rationality of intention. An intention doesn’t simply commit the agent to act without redeliberating but to act from the non-(re)deliberative stance of trust. B doesn’t invite A’s trust here; in fact, A doesn’t act from trust in B at all. So Roth gets one important part of intention in play (how intentions settle what to do) without another (how the non-deliberative rationality of intention is the rationality of trust). The role of trust is revealed by cases wherein circumstances change such that acting on the intention would run contrary to the point of it, cases wherein mere tied-to-the-mast precommitment will generate action but not trustworthy intention. These changes aren’t merely ‘defeaters’ – what would count as such a change must register in the understanding that shapes the trust relation at the core of the promise. This key element in the intrapersonal relation characteristic of intention goes missing when we shift to B’s putative intention that A act.’

    Again, sorry to be cryptic – but that’s all I have time for…

  31. Abe Roth says:

    Hi Jorah. Thanks for the comments. Might there be some tension between my story about promissory uptake (Accepting as Intending) and Scanlon’s Principle F? It’s a great question. I had thought that my approach would help Scanlon address the Can’t Get Going/Circularity worry. And I also think that it helps to account for the directedness of promissory obligation – which some like Gilbert and Darwall also have found wanting in Scanlon’s picture (though I did not myself press that worry). But Accepting as Intending won’t be of much help if there’s a fundamental incompatibility between it and with Principle F. So are they incompatible?


    You point out that sometimes one has good evidence that a promisor will follow through and on that evidential basis form an expectation, which would trigger Principle F. And you’re right that I won’t count that as acceptance, if acceptance means the sort of thing that will account for some of the special features of promissory obligation, such as directedness. On my view, we need for the promisee to will the promisor’s performance. So the promisor is in some special way authorizing the promisee to will his performance. All this is right as a reading of me. You worry that the picture seems very different from that of gaining evidence to trigger Principle F. One concern (which I believe is secondary in the thinking underlying your question) is that the intention might not even be the sort of thing that would trigger F. In effect, this amounts to a challenge to the cognitivist picture of intention. If cognitivism is not an option, then perhaps we’d need to try to read Scanlon’s F or his use of ‘expectation’ in broad fashion, so that its triggers can be a heterogeneous lot, including belief based on evidence as well as some sort of attitude like intention understood non-cognitively and much more loosely tied to evidence.


    But I take it that a more significant issue underlying your question is this. Once we have on the table the picture of the promisor establishing a moral relationship wherein the promisee is authorized to will his performance, then why even bother with Principle F? Why isn’t the promissory obligation simply understood in terms of this moral authorizing relationship between promisor and promisee?


    Suppose you don’t have anything like Principle F at work here. Then it seems that you’d have the sort of directed or what I call contralateral commitment between the individuals. That happens in the case of shared agency; but I’m somewhat reluctant to think of those distinctive commitments between participants in shared agency as having to be moral commitments. I hesitate even to speak of obligations here, let alone anything like a promissory obligation. The problem is not so much that the commitments in shared agency lack directionality; they have that. What they lack is the force and the connection with assurance that I think we have in promising. And that’s what we get when we invoke the apparatus from Scanlon. But I am not absolutely committed to that apparatus. As I said in the paper, some normative powers sort of view might be possible. And we might even have some form of conventionalism here as well (see what I said in response to Jack Woods).


    I hope I’m not missing the thrust of your comments…

  32. Abe Roth says:

    Hallie, Regarding the Circularity problem – Okay, I see your point now. Considerations of moral risk aversion suggest that the Promisor really has a reason to ϕ. It’s an obligation to ϕ out considerations of risk aversion. And that is the basis of the promisee’s expectation. Once there’s the expectation, the promisor is then subject to the promissory obligation.


    I guess that my worry is that this strategy relies on a principle that’s too strong: if for all one knows conditions are in place for one to be obligated to ϕ, then one must ϕ, just in case one actually is obligated. What if I think that there’s a chance that I should regularly be looking in on my neighbor, but in fact I really should not be doing this since it would be overly intrusive. So am I now obligated to look in on the neighbor, just in case? Am I now under conflicting obligations, to look in on them and not to as well?


    It seems that perhaps there’s a weaker version of the moral risk aversion principle that might be better. Perhaps something like being careful about triggering conditions that incur obligations that you cannot keep. (I think this is a lesson learned from Kolodny and Wallace.)

  33. Abe Roth says:

    Wesley, The promisor DOES acquire the intention from the promisee once acceptance is communicated, much in the way that your prior decision/intention formation leads to your subsequently having the intention and acting on it.

  34. Abe Roth says:

    Heath –

    The case of the love triangle with the two boxers is funny. I can imagine many cases where planning considerations might matter. Boxer A wants to vacation in Paris and see the Louvre, and Boxer B wants to go to Westworld. Suppose she accepts both promises. They each are busy training, so Lola must make reservations. But these are expensive and non-refundable. Would she go ahead and pay for both trips? The loser wouldn’t have the funds to repay her. So her planning is affected insofar as her budget is affected. And also insofar as she has to plan both trips, which is a pain. That said, I take your point that there might be a case where it really makes no difference to Lola and she might accept both promises as a way of ensuring that she gets to go on at least one of the trips. But something like this can happen for intentions as well. Assuming that you can only marry one individual, it might be that one has the best chance of marrying one of A and B by courting A with the intention of marrying A and courting B with the intention of marrying B. Sometimes good planning tolerates limited forms of incoherence.


    I’m tempted to think what Zacchaeus is doing is vowing, where this is different from a promise. It’s public, but doesn’t require uptake and doesn’t necessarily relate the speaker and audience in the intimate fashion that we find in promising. The hearer doesn’t have to accept the vow in order for the speaker to be committed, and the hearer is not in a special position to release the speaker. So I don’t think there is any good reason to think that the hearer must intend anything about the speaker in this case, and I agree it sounds odd to think that the hearer does.


    If you insist that Zacchaeus is making a promise and not merely vowing, then I’d have to think more about how this would work. Probably the most efficient way would be for the promise to be made to the entire public considered as a group. That at least would make more tractable how it is that there might be uptake and an intention on the part of the group for Z to follow through. I don’t have much more yet on the matter. It’s an interesting issue to think about.

  35. wesbuc says:

    Thanks Abe, I see what you are trying to do and appreciate the connections to joint agency. But the promisor could just as well acquire the intention to act in lots of other ways, including from communication of promisee desire that the action occur. And its promisor intention that ultimately seems to be driving a lot of the features of promises you are talking about. So I still don’t see the reason to prefer promisee intention over this more natural candidate.

    With this point in mind, I also had trouble following your rejection of (i) and (ii) in the more modest proposal section of potential promisee intentions involved. Assurance, settling and consistency can be gotten from promisor intention, which can be brought about in lots of ways potentially. So the fact (i) or (ii) don’t alone guarantee those things is no strike against them being accurate requirements on the promisee side.

  36. Abe Roth says:

    Thanks, Ted. I think your further comments will help me clarify what it is that I’m claiming about acceptance. I agree that when one accepts a promise one doesn’t usually think of oneself as willing or intending the promisor’s performance, especially when this is not done in the context of ongoing joint action. However, some of the features of the obligation and the relationship between promisor and promisee suggest to me that it is theoretically useful to see acceptance as amounting to intending promisor performance. So even if the promisee thinks what she’s doing as some form of illocutionary uptake (for some reason this promisee has a handle on Austinian terminology!), my suggestion is that given what this uptake accomplishes, it is useful to think of it as an intention regarding what the promisor will do.

    Regarding the gym case, you say that B fails to see how A’s lifting the weight engages with his (B’s) planning needs. You seem to object to my reaction, which is that B just doesn’t intend A’s lifting the weight. You ask why we need to go beyond what your propose and have the extra bit about the intention (or, more accurately, the extra bit about not having the intention). Again, it would be odd if B actually were to think in terms of his intention regarding A’s weight lifting. There is no reason to think that B thinks of what he’s doing in just those terms; that’s what makes my proposal somewhat surprising. At most B will think that he won’t accept the promise. More likely he’ll just shrug and move on. My point, again, is that if we see acceptance as intending promisor action, that illuminates a number of issues about promising of interest for philosophers. But I’m not wedded to thinking that the promisee thinks of herself in these terms. In particular, I’m not committed to thinking that in this gym case when B withholds acceptance that she’s thinking of what she’s doing as refraining from intending that this person lift the weights.

    Regarding (4), you say the promisor provides a reason but it is not conclusive or sufficient reason for the promisee to do anything. That sounds right. But that’s compatible with the thought that the reason given should be sufficient or conclusive for thinking that the promisor will do what was promised. Which is what we want if promises are supposed to provide assurance.

    Finally, I think that we are probably not so far apart concerning the issue of trust and the role of defeaters. But probably we should wait until your paper comes out before we engage on that.

  37. Juan S. Piñeros Glasscock says:

    Hi Abe,

    Thank you for such a rich and interesting paper! Like others, I’m sympathetic to the general approach, but have some criticisms of the details. But first a friendly comment that bears on your dispute with Ted: one of the reasons I think it’s really important to theorize about the acceptance conditions on promises is because of cases where a victim doesn’t want to accept the promise of the person who wronged her. In an example that Steve Darwall appealed to in conversation at some point, a rape victim might refuse to accept the promise not to rape of the man who raped her. I think one of the nice features of your analysis is that it gives an attractive account for how to understand this sort of case: certainly we don’t want to say that the victim has anything against the man not raping again, but then, why the refusal? One way to get this on your account is something this (there are a few ways of working out this out): she might not want to engage in a shared project with him by providing an intention on the basis of which he might act. On the other hand, it seems hard to account for the case on an uptake view like the one Ted is suggesting. If I’m understanding the view correctly, so long as she appropriately recognizes the speech act (where appropriately includes the ordinary conditions on uptake, like common knowledge perhaps), she accepts the promise. But then we are forced to say that the victim cannot recognize in the relevant way that the rapist is promising something to her while rejecting this promise, given that the former is sufficient for uptake, and, on the view, for acceptance. But this seems like the wrong result.
    Now for criticisms: my central concern is with something that others seem to be worried about as well, which is that on your view sincere acceptance=acceptance (simpliciter). To test the view we might consider a case like the following: A promises to B to @ and B accepts, not because B thinks A will @, but precisely because B thinks A will *fail* to @, and B knows that A will owe compensation to B for her failure to @ (imagine @ to be ‘pay back’ and the compensation to be accrual of high interests). If I understand your view correctly, these cases are in some sense impossible: B’s acceptance of this promise, if she intends for A to fail rather than to succeed, would fail to elicit the obligations of the promise (and, therefore, the obligation to compensate). But this seems to me the wrong result: it seems like the promise succeeds, however despicably B is acting in accepting it. Compare the case of giving an order: a general might give a soldier an order to @ with the intention that he fail (perhaps he wants to teach the soldier a lesson). Again, we might say the order is not sincere, but it seems wrong to say there was no order even if we think intention is necessary for sincere ordering (I don’t think it’s plausible to say this is a pretense order, along the lines of your suggestion in sec.IX). So why not distinguish between successful and sincere promise acceptance, and take intention as a constituent of the latter but not the former?

    Another case that might create difficulties for your view: a mother doesn’t want (never mind intend) her daughter to take a job in a foreign country, fearing that she might not come back. Suppose the daughter promises to come back, and the mother accepts the promise. Since taking the foreign job is a requirement for coming back, and since one must intend the necessary means to one’s intended ends, it seems on your view that mother could only accept the promise by intending for her daughter to take the foreign job. But this seems wrong: it seems the mother could accept the promise without intending or even wanting her to take the job.

    One last worry: it’s important for your view that intentions have ‘inertia’ in Bratman’s sense—they are resistant to reconsideration. Thus, if I intend to not drink tonight, and I feel like drinking once I’m at the bar, that desire should not, if I’m rational, reopen deliberation as to whether to drink: the intention should settle that I won’t. The problem here is that acceptance of a promise doesn’t seem to have this feature: suppose you promise me that you won’t drink tonight (so that I can drink and you can drive), but then while at the bar I see that you really want to drink. I may then release you of your promise, and then you’re free to drink. But if accepting the promise required me to form the intention for you not to drink, it seems I would be irrational to reopen deliberation about the matter simply on the basis that you want to drink (just as it would have been irrational for you to do so in the case where you intend it). But I don’t see anything irrational in this behaviour.
    Last, a couple of references to work on the question of beliefs about what we will do and evidence that might be helpful in pursuing this question: I think Sarah Paul’s “How We Know What We Intend” is the best non-cognitivist treatment of practical knowledge. She works out an idea along the line suggested by Daniel Star, that your intention provides the evidence for your belief about what you are doing/will do (her focus is on the progressive doing, but if I remember correctly she extends the account, or, at any rate, the extension would be straightforward), and takes on the circularity worry that you raise for it in response to Daniel, to my mind, successfully. On the other hand, Berislav Marušić has a series of articles arguing against evidentialism on the basis of what he calls cases of “difficult action” (e.g. quitting smoking), where it seems you can set to do something even though it seems really difficult to do, and so the evidence tells against your success. One of the articles is “Promising Against the Evidence”, which I haven’t read, but it might be useful to look at in considering the relation between evidence and beliefs formed on the basis of intending.

    Again, thank you for a great paper, and I look forward to hearing what you think!


  38. Ted Hinchman says:

    Thanks for the latest round, Abe. This may be the narcissism of small differences, but I actually think we’re very far apart on the role of trust in intention. (There’s a longish footnote on this difference, citing your 2004 PR paper, that I’ve for years believed was published in “Receptivity and the Will” (Noûs 2009), but I now see that I deleted the paragraph that included this footnote to get the word count down just before submitting the final version!) It’s too much, and for most readers too tangential, to pursue here. On the other hand, I don’t want to leave my comments simply incomplete in light of Friday evening’s interruption. So I’ll try to finish that thought at least, which will also generate a first-pass reply to the last of your latest comments. I’ll end with a reply to the challenge posed in the first paragraph of Juan’s recent comment.

    Here, for easy reference, the passage from your paper that I had in mind: “The proposal is that the promisee’s acceptance settles the practical matter much in the way that one’s prior intentions and decisions settle what one later does. When the promisor acts to fulfill the promise, he doesn’t re-open the question whether to @. Absent defeaters, reopening the question would be incompatible with the relevant sense of control had by the promisee. Instead, the promisor acts in a way that takes for granted what was settled for him to do when the promisee accepted.” (101-2)

    A crucial part of the argument for Accepting as Intending lies in the claim that the promisee can in this way ‘settle’ what the promisor will do. The claim isn’t, of course, that the promisee, B, can control what the promisor, A, does directly – as if A were B’s puppet. The claim is that A can follow through on B’s communicated intention in the way that B’s later self follows through on a remembered intention formed by B’s earlier self. So the question becomes: can we coherently conceive of these normative relations – one interpersonal, one intrapersonal – as running thus in parallel? I do not believe we can, for a reason that I’ll now spell out.

    This reason articulates a key dimension of the diachronic rationality of intrapersonal intention. An intrapersonal intention doesn’t simply commit the agent to act without redeliberating but to act from the non-(re)deliberative stance of trust in his intending self. The role of trust in intention is revealed especially sharply by cases wherein circumstances change such that acting on the intention would run contrary to the point of it. In these cases, mere tied-to-the-mast precommitment will generate action but trustworthy intention will not. One might naively assume that this is simply a matter of the intending self’s expectations about the circumstances of action: if B’s expectations about his future circumstances at t are verified at t, then B will non-(re)deliberatively follow through on his intention to @ at t, but if his expectations are not verified, he will need to redeliberate (perhaps reaffirming the intention, perhaps not). But that’s too simple: in any realistic case, some of B’s earlier expectations about his circumstances will be falsified at t though that falsification makes no difference whatsoever to the rationality of non-(re)deliberatively following through on his intention to @ at t. In this way, an executed intention manifests an ongoing understanding not merely that the earlier intending self is trustworthy but how it is trustworthy: specifically, of how the intending self is trustworthy through its grasp of this distinction between which of its expectations about the circumstances of action are relevant to follow-through on the intention and which are not.

    So how could B manifest that understanding *for A*? From the other side, how could A get that understanding *from B*? I think there’s a problem here that derives from the fact that this species of understanding is never fully explicit. It is never fully explicit because B has no set list of expectations about the circumstances of action that he could enumerate and sort into ‘relevant’ (= if it’s falsified, he ought to redeliberate) and ‘irrelevant.’ In a purely intrapersonal case, B’s ability to sort such expectations manifests his implicit grasp of the point of his intention: what he’s broadly up to in this instance of planning agency, how this plan fits into B’s other plans, etc. It is not at all clear how B could pass this understanding to A.

    The promissory agreement between A and B itself institutes a shared understanding of the point of the promise, and that grasp will inform how A sorts his expectations into ‘relevant’ and ‘irrelevant’ going forward. (I assume his promise sincere. If it is not, then this is a matter of how he will represent himself as understanding the promise). But here’s the key point: A cannot simply receive this understanding from B. I lack space to make a full argument here, but I hope I’ve said enough to suggest how that argument would go. One implication of the argument would be this: because *each* party must act from her or his own understanding of the promissory agreement, B does not exert any command-like or broadly pre-commissive authority over A’s follow-through. Here’s another way to put the implication, in line with your attempt, Abe, to bring your version of the authority view into the service of Scanlon’s expectation view of promissory obligation (I don’t call the latter an ‘assurance’ view because I don’t regard what Scanlon discusses as genuinely second-personal assurances): B, as promisee, has no right to expect simply that A will execute the promise. We can interpret the implication as a point addressing the nature of the promisee’s authority over the promisor, or as a point addressing the content of the promisee’s expectation about the promisor’s follow-through on the promise.

    Your reformulation of the expectation view of promissory obligation thus reveals why the expectation view cannot work: it would be manifest a misunderstanding of the promise for the promisee straightforwardly or simply to expect execution from the promisor, just as it would reveal misunderstanding of one’s own intention straightforwardly or simply to expect performance from your own later self. What the promisee expects is that the promisor will ‘keep’ the promise, i.e. remain ‘true’ or ‘faithful’ to it – just as, in order to form an intention, one must expect that one will remain constant in the intention. Fidelity in a promise and constancy in intention are matters of ongoing self-governance, including crucially a grasp of the distinction between relevant and irrelevant expectations that informs the implicit point of the promise or intention. In each case, this is not the same as expecting that the committed party will execute the commitment. The latter expectation leaves out an element crucial to understanding the commitment as a promise or an intention: an understanding of how follow-through is mediated by the pertinent interpersonal or intrapersonal trust relation. The uptake condition thus shapes not merely what we naturally call ‘acceptance’ of the promise but the content of the expectation thereby generated. The content of the expectation represents follow-through as mediated by a shared grasp of the illocutionary relation instituted through the promisee’s acceptance of the promisor’s invitation.

    We naturally oversimplify in everyday life and talk as if we simply expect to do what we intend to do and as if we simply expect a promisor to do as she promises. But responses to cases reveal that this is not really the content of these expectations. What we expect is that the committed party will remain faithful to or constant in the commitment, where such faithfulness or constancy is fully compatible, in the right context, with failure to execute the commitment. In each case, if circumstances change in relevant ways, fidelity or constancy may positively preclude execution. So the problem that B cannot simply transmit an understanding of the point of the promise to A, as if communication with A could substitute for one’s later self’s memory of an intention, rests on a more fundamental problem: neither a promise nor an intention generates an expectation of positive performance.

    Let me close with a clarification designed to fend off Juan’s worry about unwelcome promises in the first paragraph of his comment immediately above. In the case borrowed from Darwall, a sexual predator (A) ‘promises’ his victim (B) not to rape, and she rejects the promise. (Do I have this right? The description that Juan gives doesn’t quite make sense, since the promise must occur before any actual rape.) The worry is that B can apparently ‘recognize A’s speech act’ and thereby, putatively on my view, count as ‘accepting’ the promise while nonetheless going on to reject it. But this appearance is misleading, since a lot more is involved in recognizing A’s speech act than is conveyed by the unelaborated phrase, ‘recognizing that A intends to be making a promise.’ On the version of the uptake view that I’m proposing, what B would recognize in accepting the promise is A’s intention to make a reason available to B specifically through A’s trustworthy responsiveness to a subset of B’s ongoing planning needs. B would fail to recognize such an intention in A insofar as B fails to view A as even a candidate for trustworthiness vis-a-vis B’s planning needs, and that would presumably describe B’s stance toward A in any realistic rendition of the present case. In the gym case, B rejects A’s promise because B doesn’t regard A as having any accurate conception of B’s planning needs. In this sexual predation case, B rejects A’s promise because B doesn’t regard A as even a candidate for trustworthiness insofar as A represents himself as serving B’s planning needs. In each case the promise is rejected through B’s failure to recognize an illocutionary intention. My version of the uptake view emphasizes this normative complexity informing recognition of an illocutionary intention, in promising as well as in testimony, advice, and other illocutions of assurance.

  39. Abe Roth says:

    Hi Juan. Thank you for the thoughtful comments and for helping in reply to Ted. I suppose Ted would insist that in cases where the promise is unwelcome the understanding on the part of the promisee of what’s on offer won’t count as acceptance so long as it is accompanied by explicit rejection. But then you might press Ted and say that sometimes the promisee wouldn’t even want to be involved with the promisor even to the extent of explicitly rejecting the offer (as in the case I mentioned of offers that are not even worthy of a response).
    Let me turn to your concerns. One might, you suggest, accept a promise but will for the promisor to fail, or not care whether he succeeds. This doesn’t seem to sit well with Accepting as Intending. I wanted to include intention in acceptance so as to handle a number of the normative features of the promissory relationship. If I allow that we can have acceptance without intention, then my story would not account for those normative features in those cases. So that’s why I was motivated to insist on sincerity for genuine acceptance. It doesn’t strike me as that implausible to say that someone who accepts a promise just to toy with the promisor with the aim of preventing him from fulfilling the obligation would really succeed in generating the obligation. If the promisor were to recognize that the promisee had this in mind, I think it would be fair for him to think that he’s not really under any obligation and never was.
    Your intuition might differ. An alternative response would be for me to allow that insincere acceptance does generate obligation, but that it also involves promisee intention. An insincere acceptance would involve promisee intention, but for the wrong kind of reason: the promisee is able, in effect, to intend the promisor’s ϕ-ing, but not on the basis of considerations in favor of ϕ and being assured about it. Something like this can happen with ordinary individual intentions – although there are limits to this as the Toxin Puzzle suggests.
    Turn to your case of the mother and daughter. The mother accepts the daughter’s promise to come back. Since returning requires that the daughter depart, you suggest that if the mother intends the daughter’s return, she must intend the daughter’s departure as well; otherwise the return cannot happen. But the daughter’s departure is not a means to her coming back. The departure is a presumed condition for the willed return. In anticipating a storm and hoping that the displaced will find shelter quickly, I am not also hoping that those individuals will be displaced in the first place.

    Turning to the last issue: I used to think that a promisee can release the promisor for no reason at all. But having seen the light and recognized Acceptance as Intending, it strikes me that I was wrong. Intentions shouldn’t be revised arbitrarily; that would undermine their point in settling matters. And this holds for promissory release as well. Unless the promise is trivial, reconsideration and revision of the acceptance must be legitimately triggered by something; at least, if I revise for no reason at all, this cannot be something that happens habitually. And in your case of the promise of the designated driver, it strikes me as irrational for me to release you from the promise. Perhaps if you are banging your head on the table and begging me, I would. Or if someone else is available to drive us home, I would. But if the point of your promise was that you were the designated driver, and no new considerations come up, then it strikes me as irrational to release you for no reason at all.
    Thanks so much for the references. I know that Sarah Paul is inspired in part by Grice’s view, and yes it seems that Daniel’s remarks earlier are also in this vein. Maybe I was too quick in my reply to Daniel. If he were to hold a non-cognitivist view of intention, one that didn’t require a belief, then if one is able to form such an intention, it might then serve as evidence upon which one might base an expectation or belief – so long as there is a regularity between my forming an intention to ϕ and my subsequent ϕ-ing. That would address the circularity worry in the individual case. But I don’t think that this sort of approach gets right how the agent can settle some matter by forming an intention by deciding. That’s not to say, however, that I would go so far as Berislav as to think that practical reasons count as reasons for belief. But, yes, I certainly do owe more of a story about why a cognitivism that involves some form of non-evidential (but also non-practical) warrant is to be preferred. Some day….

  40. Abe Roth says:

    I see that I posted a reply to Juan without noticing Ted’s comment. I’ll have to get to this tomorrow…

  41. Juan S. Piñeros Glasscock says:

    Thank you, Ted and Abe, for your generous replies! I want to start by clarifying the case I had in mind, as Ted is right that it was very unclear from my initial post (I apologize). The case is supposed to be one where the violator makes a promise to his previous victim not to engage in such behaviour *again* (e.g. as might happen when he is being tried in court, once he has repented), and where the victim rejects the promise.
    In response to Ted: Thank you! That was really helpful! I was thinking uptake would be of the ordinary sort required for e.g. the success of an assertion in an Austinian view (where what’s required is something like the audience’s (or a possible audience’s on some views) recognition of the type of speech act in the right way (with R-intentions and what not). The conditions on uptake on your view are stronger, I now see, and so might avoid the objection, but that will depend on how we understand the notions of trustworthiness vis-a-vis someone’s planning needs. For instance, I think the victim might recognize that A genuinely intends to not rape again, might believe on the basis of his sincere promising that he won’t do so (and so plan accordingly, e.g. by writing a letter to the judge asking for a lighter sentence), and still refuse to accept the promise for something like the following reason: she doesn’t want to be part of a shared scheme of which A is a part. I can imagine different ways of working out the ideas of trustworthiness in relation to planning needs in ways in which this fails to count as meeting the uptake condition and others in which it succeeds (and my intuitions about the case once it is fleshed out in this way are more tenous). But I have your paper now, so I look forward to reading how you work out these views, and will hopefully get back to you on this in the near future!)
    In response to Abe: I think our intuitions do differ with respect to the initial case. I guess one reason is that, like Shiffrin and Markovits, I’m inclined to see contracts as tightly connected with promises (indeed as legal promises), and it doesn’t seem like failure to have the relevant intention would be good grounds to render a contract null. Of course, this could be for specific reasons to do with a contracts’ legal nature, but I think that it’s not, and that an overt saying of the appropriate type, especially when explicit (like ‘I accept your promise, here’s what you owe me if you don’t fulfill it) can be sufficient for acceptance regardless of lack of intention. (I’ll have to think more about your proposal to think that insincere acceptance involves intending for the wrong reasons; but as I was imagining the case, there isn’t an intention for the @ing at all, so I’m not sure that helps.)
    I think your response to the daughter-mother case is satisfying, though I wonder if the case could be modified to take account of the distinction between conditions and means: e.g. Alexander the Great promises Olympias that he will come back from war *victorious*; Olympias doesn’t want (or intend) him to go to war, but it seem like going to war is a means (and not a mere condition) to come back victorious (thus: ‘I went to war in order to come back victorious’ might be something Alexander says truthfully). Still, she might accept the promise without intending him to go to war.
    I’m glad that our pretheoretic intuitions agreed on the point about release (and maybe I will eventually come to your view!). I guess I was envisioning the case as something like this: I just see that you want a beer, and I feel like being nice to you, so I release you of the promise so that you drink tonight. And there doesn’t seem to be anything irrational about that. But I agree that being in the habit of releasing promisors for no reason at all seems wrong. I wonder, though, whether there isn’t a middle view, where one can’t release for just any (or no) reason, but where the reasons need not be as strong as those required to revisit an intention. In any case, one thing your comment made me realize is that the notion of release is in need of theorization as well (it seems even more undertheorized than acceptance), and one advantage of your view is that it yields interesting constraints on rational release.

    I think going forward I would be curious to know how widespread my intuitions about these cases are. Of course, even if the ones I have were well supported by ordinary standards for judging intuitions about cases, the weight we give to them would have to be assessed in relation to all the other advantages of your view.