JESP Discussion on McHugh & Way’s “All Reasons are Fundamentally for Attitudes”

Welcome to our latest Journal of Ethics and Social Philosophy discussion! Our featured article is “All Reasons are Fundamentally for Attitudes,” by Conor McHugh and Jonathan Way.

What follows is a critical précis by Keshav Singh. Please join him, our authors, and additional discussants (Hille Paakkunainen, Eva Schmidt, and Daniel Wodak) in the comments below.

Critical Précis by Keshav Singh

Summary

In “All Reasons are Fundamentally for Attitudes,” Conor McHugh and Jonathan Way argue, unsurprisingly, that all reasons are fundamentally for attitudes. As with all their work on reasons and normativity, both co- and individually authored, this is a fruitful and thought-provoking paper, with much in it that I agree with.

Conor and Jonathan (hereafter C&J) begin by noting that in normative theorizing, reasons for action are often taken as the paradigmatic or fundamental case, which leads to a variety of puzzles about how there could be reasons for attitudes. C&J’s aim is to turn this common assumption on its head, arguing that reasons for actions are in fact the non-paradigmatic and non-fundamental case.

On the alternative approach they suggest, reasons for action are derivative from reasons for intention. This is what they call the Special Hypothesis. As they see it, reasons for action are the primary obstacle to acceptance of what they call the General Hypothesis, which is that all normative reasons are fundamentally for attitudes. In arguing for the Special Hypothesis, they aim to remove this obstacle, thus clearing the path for the General Hypothesis.

C&J spend most of the paper arguing for the Special Hypothesis. In particular, they work with this version of it:

Special Hypothesis, Simple Version: For p (a fact) to be a reason to φ (an act type) is for p to be a reason to intend to φ. For p to be a reason to notφ is for p to be a reason to intend to not-φ. (153)

C&J first provide some initial motivation for the Special Hypothesis with the claim that you are criticizable with respect to reasons to act only insofar as you are criticizable with respect to reasons to intend. Take Jack and Jill, two agents who each have conclusive reason to φ, but fail to φ. Jack fails to φ because he fails to even intend to φ, whereas Jill intends to φ, and thereby tries her best to φ, but fails anyway. Here, only Jack is rationally criticizable with respect to his intentions and with respect to his actions; the Special Hypothesis gives the correct verdict.

Following this initial motivation, C&J present two main arguments for the Special Hypothesis. Each of these arguments relies on a different assumption about reasons. The first argument relies on the assumption that reasons are subject to a response constraint: they must be capable of being responded to. The second argument relies on the assumption that reasons are potential premises of good reasoning. C&J argue that as long as we accept these assumptions, we should accept the Special Hypothesis.

Here’s how C&J use the response constraint to support the Special Hypothesis. The basic idea is this: if p is a reason for you to φ, then you can respond to p by φ-ing on that basis. But in order to φ on the basis of p, you must, in φ-ing, execute an intention held on the basis of p. In general, responding to some consideration involves treating it as a normative reason. So, in φ-ing on the basis of p, you not only treat p as a reason to φ, but also as a reason to intend to φ.

The next step is this: some consideration is a normative reason if and only if it’s appropriate to treat it that way. So, if p is a reason to φ, then its appropriate to treat p as a reason to φ. But in doing so, you also treat p as a reason to intend to φ, so it must also be appropriate to treat p as a reason to intend to φ. And if it’s appropriate to treat p as a reason to intend to φ, then p is a reason to intend to φ. Thus, if p is a reason to φ, then it’s a reason to intend to φ.

The final step in the argument is that the Special Hypothesis provides a good explanation of the above conclusion: all reasons to act are also reasons to intend because to be a reason to act just is to be a reason to intend. Moreover, as C&J point out, competitors to the Special Hypothesis, like the view that reasons for intention are instrumental to reasons for action, are incompatible with that conclusion and so can’t explain it at all.

C&J’s second main argument for the Special Hypothesis uses the assumption that reasons are potential premises of good reasoning to argue that it is reasons for intention, not reasons for action, that play an analogous role in practical reasoning to the role played by reasons for belief in theoretical reasoning. In theoretical reasoning, one forms a belief by reasoning about the object of one’s belief: a proposition. Reasons for belief, then, are reasons having to do with the believed proposition.

As C&J argue, the analogue in the practical case would be this: in practical reasoning, one forms an intention by reasoning about the object of one’s intention: an action. Reasons for intention, then, are reasons having to do with the intended action. The reasons that figure in practical reasoning aren’t fundamentally reasons for action any more than the reasons that figure in theoretical reasoning are fundamentally reasons for propositions. 

This, according to C&J, gives us a nice, unified picture of how reasons figure in reasoning: a reason for a response is a premise in good reasoning about the object of the response that then concludes in that response. The view that practical reasoning fundamentally involves reasons for intention fits better with both the Special Hypothesis and this unified picture. By contrast, C&J argue, the view that practical reasoning fundamentally involves reasons for action would yield a strangely disunified picture of the role of reasons in reasoning.

After presenting these two arguments for the Special Hypothesis, C&J move to considering and replying to objections, which I won’t have space to recount. They then conclude by returning our attention to the General Hypothesis, which they take to have been adequately supported by establishing the Special Hypothesis. As a reminder, the General Hypothesis is the hypothesis that all normative reasons are fundamentally reasons for attitudes.

C&J mention two major upshots of the General Hypothesis. The first is that once we see reasons for attitudes, not actions, as the paradigmatic and fundamental case of normative reasons, the common view that normative reasons are grounded in value no longer seems plausible, because paradigmatic reasons for attitudes don’t have to do with value-promotion. It instead starts to seem more plausible that reasons are grounded in fittingness, or their role in good reasoning.

The second upshot is that it no longer seems plausible that there is any general connection between reasons and voluntary control. Rather, the general connection between reasons and control is one between reasons and what we might call attitudinal control – a form of nonvoluntary control in which we revise our attitudes directly in response to reasons. The voluntary control we have over actions, then, is the outlier, not the norm.

Recognizing these upshots of the General Hypothesis, C&J conclude, allows us to see how, if we stop taking reasons for action to be the paradigmatic case, many of the puzzles about reasons for attitudes vanish.

 

Commentary

I’ll start by emphasizing just how much I agree with here. I agree that the case of reasons for action is the main obstacle to the General Hypothesis. I agree that the General Hypothesis is true. And I think C&J are absolutely right about its upshots. The view that reasons are grounded in value, and the illusion of puzzles about voluntary control over attitudes, have both been mistakes brought on by the focus on reasons for action.

I also can’t resist mentioning that C&J’s views are very congenial to something I’ve argued for in a forthcoming paper of my own. In that paper, (“What’s in an Aim?” forthcoming in Oxford Studies in Metaethics), I argue that starting with reasons for attitudes instead of reasons for action yields a much more promising version of constitutivism about normative reasons.

According to the view I defend there, reasons for attitudes are grounded in the constitutive correctness conditions of attitudes, which are generated by how those attitudes constitutively represent their objects. Normativity is fundamentally about attitudes because normativity is fundamentally about how we represent the world. Or, as C&J write at the end of their paper, “Attitudes constitute our orientation toward the world, in which we can get things right or wrong, be justified or not. Our actions change the world in more or less beneficial ways—but it is only insofar as they express or embody our orientation toward the world that they fall within the realm of reason and normativity” (170). I think this is absolutely right as well.

Despite agreeing with C&J about so much, I’m not quite convinced that the Special Hypothesis is the best route to the General Hypothesis. This is partly due to some of the objections to C&J’s arguments for the Special Hypothesis that I haven’t discussed, which will hopefully come up in further discussion. But I want to discuss a different, broader reason, which is that it’s not clear to me that there isn’t a much more direct route, from the nature of intentional action to the General Hypothesis.

At the end of section 1 of their paper, C&J consider another alternative to the Special Hypothesis, according to which reasons for intentional action are prior to both reasons for action (intentional or not) and reasons for intention. As they say, they mostly ignore this view because they don’t know of any defenses of it. In a footnote, they mention that this view might appeal to some defenders of the view that intentional action is prior to intention, such as Thompson, McDowell, and Setiya.

Interestingly, however, they don’t mention the progenitor of that sort of view, whose influence is a large a part of why philosophers like Thompson, McDowell, and Setiya defend it: Anscombe. And while Anscombe didn’t write using the terminology of normative reasons, I think taking Anscombe’s views about intention and intentional action seriously yields essentially the alternative to the Special Hypothesis that C&J mention before setting it aside.

Why am I bringing up Anscombe here? Because I think the kind of view of intentional action I attribute to Anscombe has the potential to provide a much more direct argument for the General Hypothesis than C&J provide – one that does not require thinking of intention as in the first instance a mental state that plays a generative role in action.

As I interpret Anscombe, what makes an action intentional is not that, in acting, one executes an intention, understood as a separate entity. Rather, what makes action intentional is that it’s a case of doing something where what you’re doing is partly constituted by your agent-level representation of what you’re doing. If this is right, then intentional action is itself partly constituted by an attitude that takes your action itself as its object. Indeed, I argue in the aforementioned paper of mine that such representations are the source of normative reasons for action.

If reasons for action are fundamentally reasons for intentional action, and what makes action intentional its being partly constituted by a certain attitude, then that is already enough to establish that reasons for actions are fundamentally reasons for attitudes. We don’t have to commit to the view that there is some separate attitude of intention, distinct from the action itself, from which reasons for action are derivative.

The point can be put another way that doesn’t rely on anything quite so Anscombean. Plausibly, whatever your view of intentional action, there is something about the action itself that makes it intentional. And that something can’t be solely behavioral – it must be something attitudinal. Thus, intentional action must be partly constituted by some kind of attitude in virtue of which it is intentional. My thought here that this is all we need to establish that reasons for intentional action are fundamentally reasons for attitudes – we don’t need to establish the stronger claim that reasons for action are fundamentally reasons for intention in the sense that requires C&J’s two arguments.

To be clear, I’m not claiming that whatever this attitude is that’s constitutive of action isn’t itself a form of intention – we might call it, as some do, intention in action. What I am claiming is that it seems to me that C&J, with their two arguments for the Special Hypothesis, argue for something stronger than what they need to establish the General Hypothesis. They argue that reasons for action are derivative from reasons for intention understood as a distinct existence, which I think is a much more difficult claim to argue for.

One final way of putting it is that perhaps we should instead think of reasons for action as derivative from reasons for a particular form of intention, understood as whatever attitude it is that is partly constitutive of intentional action, and in virtue of which intentional action is intentional. As far as I can tell, this is sufficient to establish the General Hypothesis without making either of the arguments C&J make, and without making direct use of the two assumptions from which their arguments proceed. And in doing so, we might be able to avoid some of the thornier objections C&J grapple with in their paper.

Ultimately, this is not so much a criticism of what C&J argue for as a suggestion that perhaps there is a significantly shorter route to their conclusion – a route that becomes available once we think about intentional action less like normative theorists tend to and more like action theorists (especially Anscombeans) do. An interesting upshot of this is that, far from being a problem for the General Hypothesis, the view that intentional action is prior to intention might actually make the General Hypothesis exceedingly easy to establish. Given that the truth of the General Hypothesis is what C&J are ultimately concerned with, my hope is that this is a welcome upshot.

Though there is much more to say, I’ll leave it there for now. Thanks to Conor and Jonathan for this great paper. I really enjoyed thinking about it, and look forward to further discussion.

26 Replies to “JESP Discussion on McHugh & Way’s “All Reasons are Fundamentally for Attitudes”

  1. I’ll just start off the discussion, then. Thanks to Jonathan and Conor for a really nice and well-written paper – I don’t agree with the conclusion you’re trying to establish, but as ever it’s really hard to find ways to cause trouble for your argument!

    Conor and Jonathan set up this nice parallelism between theoretical and practical reasoning:

    theoretical reasoning:
    – aim to settle what is true/form a belief
    – consider objects of belief (=propositions), not beliefs themselves
    – reasons support beliefs

    practical reasoning:
    – aim to settle on a course of action/form an intention
    – consider object of intention (= action), not intentions themselves
    – reasons support intentions

    Here are two worries for this parallelism:

    1. As Conor and Jonathan concede in fn 22, we can also form doxastic attitudes by way of higher-order reasoning. They go on to say that in such cases we primarily reason to higher-order attitudes. But I think higher-order considerations become very prominent in reasoning once we take into account reasoning towards suspension of judgment (which I take it is not a higher-order attitude). Often enough, in trying to settle a question, I realize that I don’t have enough evidence, or don’t know enough about the matter, and so I suspend. The case of suspension shows that theoretical reasoning is more messy than Conor and Jonathan’s neat presentation suggests, I believe, and thereby makes it harder to carry the neat picture over to practical reasoning.

    2. Here’s a version of the “it’s actions that matter, not intentions” worry that I drew from the parallelism. If the point of practical reasoning really is to form an intention, then I should be able to form the intention and call it good. But clearly (at least to me!) the point of practical reasoning is to act, and if I merely form the intention, practical reasoning hasn’t done everything it should. Think of it in terms of the parallelism with belief: There, it’s indeed good enough to form a belief, and nothing else seems to be missing. My suspicion is that, again, reasoning is messier than Conor and Jonathan take it to be. Our conception of practical and theoretical reasoning should do justice to this.

    I’m curious to see what you think of these worries!

  2. Many thanks to Keshav for his excellent précis and comments, and to Matt King for organising this discussion. In this post we will briefly reply to Keshav’s main critical point.
    If all reasons are fundamentally for attitudes (our ‘General Hypothesis’), then reasons for action must be fundamentally reasons for some attitude. We argue that they are fundamentally reasons for intention (the ‘Special Hypothesis’). Keshav agrees with the General Hypothesis, but queries whether ours is the best way to defend it. Keshav suggests that we should instead adopt an Anscombean view of intentional action, on which intentionally acting is partly constituted by the attitude of representing yourself as doing something. Reasons for action are fundamentally reasons for intentional action, and since intentional action is partly constituted by this attitude, reasons for action are fundamentally reasons for an attitude. So there is no need to take on the burden of arguing for the Special Hypothesis as we do.
    We welcome this exploration of an alternative route to defending the General Hypothesis. Indeed, in the paper we leave open the possibility that reasons for action might be reducible to reasons for intentional action, so Keshav has, if he is right, helpfully filled in a lacuna in our argument.
    As a clarification, we don’t take our arguments to turn on any particular view of the relation between intention and intentional action. We take them to be compatible with the Anscombean view Keshav proposes, as far as we understand it.
    Our main point here, though, is that we are not (yet!) convinced that Keshav’s proposal offers a quicker route to the General Hypothesis than what we offer in the paper. Keshav says that, if reasons for action are fundamentally reasons for intentional action, then, given the Anscombean view of intentional action, reasons for action are fundamentally reasons for attitudes. But how is the claim that reasons for action are fundamentally reasons for intentional action established? This claim seems as much in need of argument as our Special Hypothesis.
    To illustrate, take the action of bringing your book to work. This is something you could do intentionally, or not – you might do it unintentionally because, unbeknownst to you, your book is in your bag. The claim is that a reason to bring your book to work is fundamentally a reason to intentionally bring your book to work. This isn’t obvious. After all, the reason might derive from some benefit of bringing your book to work – you’ll have it for the reading group, say – that will come about regardless of whether you bring the book intentionally.
    Nor does the claim seem to follow from the Anscombean view of intentional action. As we understand it, that’s a view about what makes an instance of bringing your book to work count as intentionally doing so. It doesn’t make intentionally bringing your book (the type) prior to bringing your book , nor does it make reasons for the latter fundamentally reasons for the former.
    If all of this is right, then there is no quick route to the General Hypothesis here: like the Special Hypothesis, the claim that reasons for action are fundamentally reasons for intentional action is one that requires argument. We suggest in the paper that our arguments for the Special Hypothesis could be adapted to support this claim. Of course, if someone has better arguments, that’s all to the good!
    Thanks again to Keshav and to Matt. We really look forward to continuing the discussion.

  3. Thanks, Eva, for this excellent comment. Your thought is that both theoretical and practical reasoning are messier than our ‘parallelism’ argument suggests. In one sense, I agree. We were thinking somewhat schematically about core cases of theoretical and practical reasoning, in order to bring out what we take to be the structural parallels. I think these basic parallels remain, and they remain significant for the nature of reasons, even when we take into account the complications you point out.
    On your two specific worries:
    1. It seems right that reasoning to suspension of judgment can often feature higher-order considerations, e.g. about what you know or what your evidence supports (though I’m not sure that it *must* do so). And, as we concede, reasoning towards belief can feature higher-order considerations too. But the premises of good reasoning directly to belief must bear on the object of belief – on whether q, say. And you can reason well to belief without thinking about higher-order considerations at all – you can just think something of the form ‘p, and if p then q, so, q’. So it still seems that the considerations that play the role of reasons in reasoning towards belief are ones that bear on its object and that normatively support the response, believing.
    2. I certainly agree that there’s a connection between practical reasoning and action, but I think our view can do justice to it. It seems clear that the point (in some sense) of intending is to act, so in reasoning towards intention you aim to form an attitude whose aim is to act. So if you engage in practical reasoning but you don’t end up acting, an aim has not been fulfilled. Still, I don’t think that means that, as you put it, practical reasoning hasn’t done everything it should. Instead we can say that, once you form an intention, practical reasoning has done its bit; it’s then up to the intention, so to speak, to get itself executed. (Here I’m ignoring the view that the conclusion of practical reasoning is an intentional action, which raises somewhat different issues as we touch on in the exchange with Keshav.)
    Thanks again for the comment. Do these points speak to your worries?

  4. Wonderful paper, Conor and Jonathan. And a great commentary too, Keshav.

    I had a question similar to the “different objection to the Special Hypothesis” raised by Ulrike Heuer’s examples, that you discuss on pp.163-164.

    According to the Special Hypothesis, “For p (a fact) to be a reason to Φ (an act type) is for p to be a reason to intend to Φ. For p to be a reason to not-Φ is for p to be a reason to intend to not-Φ.” (p. 153).

    Here’s a worry about this, particularly about the part concerning reasons to not-Φ. Intuitively, when there are reasons for me to not kill my neighbor, to not maim my cousin, to not torture anyone, and so on, these reasons are usually decisive reasons of requiring strength, making it the case that I’m under an authoritative demand not to do these things. But also intuitively, I’m satisfying the demands imposed by all of these reasons simply because I’m not doing the relevant things; and I’m not doing them simply because it would never even occur to me to do any of these things. In order to satisfy the demands imposed by these reasons, it doesn’t seem that I need, in addition, to think about all of these various things and form and have all of these intentions – intentions to not kill my neighbor, to not maim my cousin, etc. But according to the Special Hypothesis, it seems that I would have to have all of these intentions in order to satisfy the demands of the relevant reasons.
    On pp. 163-164, you discuss what seems like a similar objection. You say that whereas you “ought not to go into space without a spacesuit and you ought not kill … it is okay not to intend these things.” And this seems to suggest that there are some reasons to (not) Φ that are not reasons to intend to (not) Φ.

    In response, you appeal to the costs of the deliberation and conscious attention that would, or may (p.163), be required for forming the relevant intentions. If I understand you correctly, the idea is that reasons to not-Φ can still be reasons (presumably of the same strength) to intend to not-Φ, compatibly with the idea that the reasons require you to not-Φ but don’t require you to intend to not-Φ, so long as the costs of forming the relevant intentions somehow prevent the reasons to intend from generating a requirement to intend. And of course, this has to happen without there being similar costs of deliberation and conscious attention that prevent the reasons to not-Φ—for example, the reasons to not kill—from generating a requirement to not kill. We want there to be requirements to not kill, but at least often, no requirements to intend to not kill.

    One way to try to make sense of this would be the following: even though the reasons to not kill, and so the reasons to intend to not kill, have requiring strength, there are opposing reasons of “justifying” strength, to do with the costs of deliberation and attention, that prevent the initial reasons to intend to not kill from generating an actual requirement to intend. This would be akin to how moral requiring reasons to keep a promise can be prevented from generating an actual moral requirement to keep a promise by prudential reasons of justifying strength, when breaking the promise would be especially beneficial to oneself, or keeping it especially costly, as in one of Douglas Portmore’s examples in his Commonsense Consequentialism. I’m not sure this type of move is what you have in mind, though, given what you say at the bottom of p.163 and the top of p. 164. I found the quick mention of “other possibilities” there a little hard to parse, however, specifically the idea that the facts about the costs of deliberation and attention might “affect the overall status of intention indirectly by providing reasons to form no intention on the matter” (p.164). If the initial reasons to intend to not kill have requiring strength, as it seems they would have to if the reasons to not kill also have requiring strength, then precisely how are these requiring reasons prevented from generating a requirement to intend, if not by something like opposing “justifying” reasons not to intend?

    I’m also wondering about the following. Suppose the costs of deliberation and conscious attention are enough to somehow prevent the reasons to intend to not kill from generating a requirement to intend to not kill. Then why wouldn’t similar costs of deliberation and conscious attention—for someone with a seriously flawed character who had to think carefully in order to prevent themselves from killing in some circumstance—be enough to prevent the reasons not to kill from generating a requirement not to kill? After all, presumably the reasons not to kill and the reasons to intend not to kill must have the same (kind and degree) of strength, if the Special Hypothesis is true. (Right?) And presumably the costs of deliberation and conscious attention can be similar in both kinds of cases, too, for some people. If so, then it’s hard to see why the costs of deliberation and conscious attention might be enough to prevent the reasons to intend to not kill from generating a requirement to so intend, without the similar costs of deliberation and conscious attention being enough to prevent the similar-strength reasons to not kill from generating a requirement to not kill.

    One might suggest that the reason for the disparity—that the costs of deliberation are enough to prevent a requirement to intend to not kill, but are not enough to prevent a requirement to not kill—is that the costs of failing to not-kill are themselves a lot higher than the costs of failing to merely intend to not-kill. If you fail to not-kill, you have killed someone. Someone is dead, and by your doing. Whereas if you merely fail to intend to not-kill, nothing very bad has yet happened. But I’m not sure this retort is available to you. The massive costs of failing to not-kill are presumably strong reasons to not-kill. By the Special Hypothesis, they should be (presumably equally strong) reasons to intend to not-kill. If they are equally strong reasons to intend to not-kill, then the costs of deliberation and conscious attention should not be enough to make it okay to fail to intend to not-kill, any more than they are enough to make it okay to fail to not-kill.

    I hope this is all reasonably clear. Thanks in advance for any responses! And thank you for the great paper.

  5. Thanks for your response, Conor, that’s very helpful. As to my second concern (something is left to do for practical reasoning after an intention is formed), I guess I would have put the part where intention gets itself executed still under the header of “practical reasoning”. But I must admit that I can’t provide any good reason why one should do it that way, and not, as you suggest, cut off practical reasoning after an intention is formed.
    As to my first concern, I want to insist that reasoning to belief is no more THE central kind of theoretical reasoning than is reasoning to suspension. And if reasoning to suspension is equally central, and standardly involves higher-order considerations, the neat picture isn’t there to be carried over to practical reasoning.

    As a further point, I wanted to support Keshav’s suggestion above that Anscombe is the go-to philosopher for a position that takes reasons for intentional action as fundamental. (Even though I’m not sure she would endorse his claim that it is a kind of attitude that makes for intentional action.)
    I was wondering though whether this account can really deliver the General Hypothesis, that all reasons are fundamentally reasons for attitudes. As I understand that claim, it implies that its being a reason for an attitude will explain why a particular reason to act is (also) a reason for an action. Does this last claim follow from Keshav’s suggestions? He argues the following: Since what makes an action intentional is that it is partly constituted by an attitude (a representation of what one is doing), so the reason to perform that intentional action must fundamentally be a reason for that attitude. (I hope I understood you correctly, Keshav!)
    Now my question is whether this last claim really follows. Couldn’t we say: This intentional action is intentional because of its constitutive attitude; but the reason for this intentional action does not count in its favor because it counts in favor of its constitutive attitude? Why not maintain that reasons for intentional actions are explanatorily basic despite the constitutive claim? Here is an example: That it’s raining is a reason to put on my raincoat. Putting on my raincoat is an intentional action because it involves my agent-level representation of my putting on my raincoat. But nonetheless, on the Anscombean picture, that it’s raining is fundamentally a reason to perform that intentional action, and not fundamentally a reason for me to represent myself as putting on my raincoat.

    Or am I missing something? Looking forward to hearing your thoughts on this point.

  6. Thanks for the great discussion so far, all!

    I have some other thoughts about the paper I didn’t include in my original commentary, but first let me respond to the points about my alternative sort of proposal from Conor and Jonathan, and from Eva.

    C&J point out that it’s not enough to suggest that reasons for intentional action are fundamentally reasons for whatever attitude is constitutive of intentional action. We still need to show that reasons for action are fundamentally reasons for intentional action, and it’s not clear this will be any simpler than establishing the Special Hypothesis. This is a fair point. I was thinking the view that reasons for action are fundamentally reasons for intentional action traverses a smaller gap than the the view that reasons for action are fundamentally reasons for intention. But I may have overstated how much of a difference there is. Moreover, I didn’t explain some of the problems I think it avoids to do things in the way I suggested. So let me try to do a little better here.

    With their book example, C&J argue that a reason to bring one’s book to work might derive fundamentally from some benefit of bringing your book to work, such as that you’ll have it for the reading group. And you could comply with this reason by unintentionally bringing the book to work. One thing we could say here is that you only have a reason to bring your book to work insofar as you have a reason for that which it’s a means to: having it for the reading group. And presumably you intend to have it for the reading group. So we could say that the reason to bring your book is not fundamentally a reason to intentionally bring your book, but rather a reason to intentionally have it for the reading group.

    Another thing we could say is something similar to what C&J say about reasons, e.g., to be able to speak French. As they point out, it’s quite plausible to reinterpret reasons to be able to speak French as reasons to bring it about that you’re able to speak French. As they put it, “this reinterpretation is natural because reasons are for responding to.” I agree. We could not respond to reasons to be able to speak French except by bringing it about that we are able to speak French.

    Similarly, it’s quite plausible to reinterpret reasons to bring it about that you’re able to speak French as reasons to intentionally bring it about that you’re able to speak French. For we could not respond to reasons to bring it about that we are able to speak French except by intentionally bringing it about that we are able to speak French. Therefore, if all reasons for outcomes are fundamentally reasons for action, then by the same sort of reasoning, all reasons for actions are fundamentally reasons for intentional action.

    Of course, this relies on something like the response constraint on reasons (which I accept). And it’s not totally dissimilar to C&J’s first argument for the Special Hypothesis. So, they’re right that something in the vicinity of what they’ve argued might be necessary to establish the alternative I’ve suggested. Nevertheless, I think my way of going is still significantly more straightforward than their argument from the response to the conclusion that reasons for action are fundamentally reasons for intention in that it seems to eliminate several steps. So while C&J might be right that my suggestion still relies on some of the same machinery they do, like the response constraint, it might nevertheless provide a shorter and more straightforward route to the General Hypothesis.

    Moreover, it seems like going my way might avoid some of the issues raised for C&J’s arguments for the Special Hypothesis. I agree with Eva that there’s something odd about saying that practical reasoning is done once one has formed an intention. The view that practical reasoning concludes in an intentional action seems better to me. So perhaps we should take intentional action, not intention, to play the role in practical reasoning that belief plays in theoretical reasoning. My alternative would allow us to do that.

    Now to try to address Eva’s point that, even if there’s an attitude that makes intentional action intentional, reasons for intentional action may still be basic rather than being fundamentally reasons for that attitude. As Eva writes, “Putting on my raincoat is an intentional action because it involves my agent-level representation of my putting on my raincoat. But nonetheless, on the Anscombean picture, that it’s raining is fundamentally a reason to perform that intentional action, and not fundamentally a reason for me to represent myself as putting on my raincoat.”

    I agree that it’s odd to say I fundamentally have a reason to represent myself as putting on my raincoat, where that’s an attitude that can be divorced from putting on my raincoat. What I failed to make clear in my initial comments is that I think a large benefit of the Anscombean picture (as I interpret it) is that this representation of what one is doing that is constitutive of the intentional action is also inseparable from it. Not only is the action only intentional in virtue of being constituted by this attitude, but this attitude is only the attitude it is in virtue of constituting an intentional action. It’s not a conjunction of two parts, “action” and “intention” that are stuck together in the right way. So perhaps the better thing to say, following Sergio Tenenbaum in his recent book on instrumental rationality, is that intentional action should itself be thought of as an attitude in the relevant sense. So, the reason is after all fundamentally a reason to intentionally put on your raincoat, where intentionally putting on your raincoat is both an action and an attitude.

    I think this would still get us the General Hypothesis. And it respects C&J’s crucial upshot that “[a]ttitudes constitute our orientation toward the world, in which we can get things right or wrong, be justified or not. Our actions change the world in more or less beneficial ways—but it is only insofar as they express or embody our orientation toward the world that they fall within the realm of reason and normativity” (I love the way they put this). If intentional action is itself, in the relevant sense, an attitude, then it fits with this picture of the nature of reason and normativity. What’s interesting about intentional actions is that they at once both constitute and embody our orientation toward the world. I think that’s a very nicely Anscombean way of thinking about things.

  7. Thanks, Hille, for the great questions. As you say, we deny that the Special Hypothesis entails that the overall status of acting and intending (i.e. their status as required, permitted, etc.) must be the same. The basic thought is that the overall status of a response is determined by more than just the reasons for it. How exactly it is determined is a very tricky issue, as your comments bring out. I think the view we had in mind in the sentence you quote from p. 164 was that the overall status of an option can depend indirectly on the reasons for other options, for instance because they are incompatible with it, rather than depending only on the reasons for and against the option in question. If so, then reasons for forming no intention might be relevant to the overall status of intending. Perhaps by making it permissible to form no intention, they thereby make intending merely permitted where it would otherwise be required. But, since forming no intention is compatible with not killing, these reasons aren’t relevant to the overall status of not killing.
    Perhaps this suggestion might help with the very good question you raise towards the end of your post – about why, if costs of deliberation can make it the case that intending to not kill is merely permitted rather than required, similar costs of not-killing can’t make it the case that not killing is merely permitted rather than required. On the view suggested, such costs affect the overall status of intending by being reasons to form no intention, which is incompatible with intending. If they affect the overall status of not killing it must be by being reasons for something else – say, for killing. But we don’t seem obliged to say that they must be strong enough, as reasons to kill, to make killing permitted, just because they are strong enough, as reasons to form no intention, to make forming no intention permitted.
    All of that said, we are in fact more sympathetic to another response to the objection – the second “other possibility” we mention in the passages you cite. That is to say that you in fact ought to intend not to kill, but you are not criticisable for lacking this intention, because you will not intend unless you deliberate, and you ought not deliberate. What our intuitions are picking up on is this difference in criticisability between lacking the intention and failing to act (i.e. killing). This is the line we take in our forthcoming book. This allows for a more straightforward treatment of how reasons determine overall status, and your questions make me slightly more confident that we are right to go this way!

  8. Thanks, Eva, for the further thoughts. Re. theoretical reasoning, I take your point that reasoning to suspension is not obviously less central to it (whatever that amounts to) than reasoning to belief. We could, perhaps, have put our points more cautiously, in terms of reasoning to belief rather than theoretical reasoning generally (though I don’t mean to suggest that this would make them uncontentious). Reasoning to belief is the reasoning that, I take it, most directly reveals something about the nature of reasons for belief.

  9. Keshav, thanks for these very helpful further thoughts. I now see more clearly why you were thinking your proposal offers a shorter route to the General Hypothesis. As you anticipate, the line of thought you sketch strikes me as pretty close to the first argument we offer, based on the response constraint. My suspicion is that, once fully spelled out, it wouldn’t be significantly more straightforward than our argument. After all, even if responding to a reason to act requires intentionally acting, it doesn’t immediately follow that it is a reason for intentionally acting, far less that it is fundamentally one. But I’m open to persuasion on this!
    As for regarding intentional action as an attitude, for what it’s worth (and speaking just for myself) I find that very attractive. I would be tempted to regard token intentional actions as identical to token present-directed intentions, in which case the view we defend in the paper might end up being close to what you’re suggesting. But in the paper we tried to make as few contentious assumptions as we could about intention, intentional action, and their relations.

  10. Conor: I too like the view that token intentional actions are identical to (or at least constituted by) token present-directed intentions, but I can see how that might be contentious. Even if my alternative is ultimately no easier to argue for, it might bolster the case for the General Hypothesis. The way you and Jonathan frame things, with intention as a distinct mental state, is going to be more compelling to Davidsonians. My alternative is probably going to be more compelling to Anscombeans. So, this is proof of concept, perhaps, that one can argue from the response constraint to the General Hypothesis regardless of one’s action-theoretic allegiances.

    At the end of the day, I basically agree with the thrust of what you and Jonathan are arguing. And it makes the metaethical theory I’m developing, which I referenced in my comments, much easier to defend. So I’m all for it, pesky details aside.

    On another note, I had a thought about some of the issues raised by Heuer’s cases and in Hille’s comments. The Special Hypothesis says, “For p (a fact) to be a reason to Φ (an act type) is for p to be a reason to intend to Φ. For p to be a reason to not-Φ is for p to be a reason to intend to not-Φ.” As Hille points out, the latter part of the Special Hypothesis raises this issue of having decisive reasons to intend not to do all these crazy things that would never occur to us, massively overgenerating intentions. And I don’t find appeals to the costs of deliberation, or the idea that you ought to have these intentions but are blameless for lacking them, very convincing.

    I’m wondering if we can instead get out of this problem by replacing “reasons to not-Φ” with (the to-my-ear more natural) “reasons not to Φ” in the Special Hypothesis. Let’s say we replaced the Special Hypothesis with this variation: “For p (a fact) to be a reason to Φ (an act type) is for p to be a reason to intentionally Φ. For p to be a reason to not to Φ is for p to be a reason not to intend to Φ.” This variation of the Special Hypothesis would avoid over-generating reasons for intention, and so seemingly avoid the problem without relying on an error theory of our intuitions that it’s not the case that, say ought to intend not to kill whenever we ought not kill.

    Is there some downside to modifying the Special Hypothesis in this way, or some way in which it fails to avoid the problem that I’m overlooking? There are at least no obvious issues to me. It’s not even obvious to me that the idea of distinctive “reasons to not-Φ,” as opposed to the more natural “reasons not to Φ,” makes sense.

  11. Sorry, the variation on the Special Hypothesis should say: “For p (a fact) to be a reason to Φ (an act type) is for p to be a reason to intend to Φ. For p to be a reason to not to Φ is for p to be a reason not to intend to Φ.” It won’t let me edit my comment.

  12. Thank you, Keshav, for clarifying your Anscombean picture, that really helped me. To think of intentional action itself as an attitude is intriguing. Do you develop this idea in the paper you mentioned in your critical précis?

    I was wondering about another possible kind of counterexample to the Special Hypothesis, but not the General Hypothesis, inspired by Hursthouse’s “Arational Actions”. I take it that we can have emotions for good reasons, and that some of our actions express our emotions or are caused by them. For example, a subject cries because he’s grieving the death of his father, or hugs his child because he loves her. A similar case, which may or may not relate to emotions, is when someone, with the unfortunate character trait of Schadenfreude, laughs in response to the mishap of another, who trips over his own feet.

    Taking the grief and love cases, the subject responds appropriately to his reasons (that his father died; that his child is wonderful) by having these emotions. So these emotions are had for good reasons. And I think it’s plausible that this carries over to the actions performed by the subject because of his emotions. He has good reasons to cry, and cries for these reasons; and he has good reasons to hug his child, and hugs her for these reasons. At the same time, the way I am imagining these cases, these actions are not produced by forming an intention; instead, they’re simply expressions of the subject’s emotions. Further, it strikes me as wrong to say that the fact that his father died is a reason to intend to cry, and it’s also a bit weird to say that about the hugging case, although maybe less so. If all this is true, there are some action types for which it is not true that to be a reason for them is a reason to intend these actions.

    Similar things go for the Schadenfreude case. We can think of laughing as an expression of amusement, which is probably again a kind of attitude. I take it that the fact that the other person tripped over his own feet is funny, and so a reason to be amused and to laugh (even though it’s probably outweighed by reasons of kindness not to laugh). Here’s my evidence: Say I laugh about the other’s mishap, and then he asks me, “why are you laughing?” I would probably say, “for no reason”, but that would be a lie. So there is a reason for which I laugh; but again it seems that this is not a reason to intend to laugh, and I don’t intend to laugh for that reason, but just do.

    One thing that’s interesting about these potential counterexamples is that they do not undermine the General Hypothesis, as far as I can see. We can still hold on to the idea that the reasons to cry, hug, or laugh are reasons for these expressive actions only because they are reasons for the corresponding attitudes. But the counterexamples do break the link between reasons to act and reasons to intend.

    Looking forward to your responses!

  13. Hi all, sorry to be a bit late to the discussion. Thanks to everyone for the really interesting comments and questions, and thanks especially to Keshav for the excellent commentary. Lots to think about!

    Keshav, you asked whether we might reformulate the Special Hypothesis’ claim about reasons not to act, as follows: for p to be a reason not to φ is for p to be a reason not to intend to φ.

    One worry about this is that there do seem to be considerations which intuitively count against intending to φ which don’t count against φ-ing. For instance, if I expect to get more information that’s relevant to the decision about whether to φ, and don’t need to decide yet, then this seems to support making no decision now, and thus not intending to φ, but doesn’t speak either way about whether to φ. (Mark Schroeder gives examples like this in ‘The Ubiquity of State-Given Reasons’.)

    (By the way, I’m not sure anything much turns on putting things in terms of reasons not to φ vs reasons to not-φ. Either way we get the substantive question of whether reasons not to φ/reasons to not-φ are to be understood in terms of reasons not to have an intention or reasons to have an intention not to act. But I might be missing something.)

    Eva, those are very interesting cases. I guess I don’t think it’s too odd to say the person has reasons to intend to hug the child. It’s perhaps a bit stranger to say the person has reason to laugh, and maybe stranger still to say that they have reasons to cry. That might be because laughing and crying seem like the kinds of things which can, but need not be, actions, and it’s less clear that the people here have reasons to perform actions of laughing and crying.

    Maybe it’s also worth noting that on our view, even if the people here ought to laugh, cry, and hug, it doesn’t necessarily follow that they ought to intend to do so. So we can agree that nothing need be amiss if the people here do these things without intentions. Of course, understood in this way these cases are a nice different sort of way of bringing out the worry that Heuer’s cases also raise, so part of this will turn on whether what we say about those cases works out.

  14. Thanks, all, for the really interesting and helpful further thoughts and examples. Two brief points in reply to Keshav’s last post:
    – I didn’t take us to be assuming anything controversial about intention being a ‘distinct’ mental state. I think I’m missing something, or not fully understanding what you mean by that, or both. But, again, I’m very happy to accept the constructive suggestion for framing things in a way that might seem more congenial to Anscombeans.
    – One observation about your proposed revision of the Special Hypothesis is that it no longer offers a unified account of reasons for acting and reasons for omitting – it says that reasons to act are reasons to intend to act, but that reasons to omit are reasons not to intend to act. I don’t know if that’s a problem.

  15. Jonathan: You wrote “One worry about this is that there do seem to be considerations which intuitively count against intending to φ which don’t count against φ-ing.” Can you explain to me a bit further why this is a worry for the amendment I proposed? Even if for p to be a reason not to φ is for p to be a reason not to intend to φ, this doesn’t seem to rule reasons not to intend to φ that aren’t reasons not to φ. That would seem to be ruled out by the opposite direction of analysis, not the one I proposed. But I might be getting confused.

    As for whether anything much turns on in terms of reasons not to φ vs reasons to not-φ: I wasn’t thinking it affected whether or not there’s a substantive question there. I was thinking it might help deal with the cases brought up by Hille – cases where the Special Hypothesis as you’ve formulated it seems to entail that we have reasons to intend not to do all sorts of crazy things that would never occur to us to do. And like Hille, I’m not totally satisfied with responses that appeal to the costs of deliberation, or the idea that you ought to have these intentions but are blameless for lacking them. So I was trying to suggest an alternative that might deal with the problem more neatly.

    Conor: I think whether my proposed revision no longer offers a unified account depends on whether there is anything to “reasons for omitting” over and above reasons against action. It’s not immediately clear to me why there would be. And if there isn’t, it still seems like a perfectly unified picture to me: reasons to φ are fundamentally reasons to intend to φ, and reasons not to φ are fundamentally reasons not to intend to φ. And this seems like it would help with the cases Hille raised, as well as some of the cases you discussed from Heuer. It’s possible it would bring up some different problem, but it’s not obvious to me what that would be.

    Eva: I wasn’t yet thinking of things that way, so I didn’t develop the idea that intentional action is itself an attitude in the paper I mentioned. I actually got the idea from Sergio Tenenbaum’s recent book “Rational Powers in Action,” where he argues that intentional action is the relevant attitude that figures in principles of instrumental rationality.

  16. Thanks so much for your response to my initial question, Conor – and sorry about the slow response on my end. What you say helps a lot. I see better now how the first option you helpfully spell out might work. I am worried about the second option, though – and thanks for making clear that that’s the option you prefer.

    On this option (p.164), I ought to intend to not kill my neighbor, to not maim my cousin, to not steal from the old lady across the street, etc.: one required intention per each of the innumerable things that I have decisive requiring reasons to not do. You say that I’m not criticizeable for lacking these innumerable intentions I ought to have, because I ought not to deliberate about all of these innumerable topics. And this, I gather, is supposed to take the intuitive sting away from the idea that we really ought to have all of these innumerable intentions, in order to satisfy the demands that our reasons place on us. It’s true that we really ought to have all these intentions to satisfy the demands that our reasons place on us, but that’s not so counterintuitive, I take it is the thought is (?), when we’re routinely not criticizeable for failing to meet these demands.

    I have three worries about this line of thinking.

    First, it seems to buy into a type of extremely widespread deontic dilemma, where whatever you do, you fail to do something you ought to do. If you have the relevant intentions, you’ve (in all likelihood?) deliberated about the topic of the intention, which you ought not to have done. If you don’t have the relevant intentions, then you lack intentions you ought to have. I’m fine with the idea that *sometimes* the normative landscape is tragic, and that in such instances you are excused from blame or criticism. But it seems more troubling to posit an extremely widespread sort of deontic dilemma, which routinely excuses people from blame or criticism for failing to meet authoritative obligations that, supposedly, they are nonetheless under. If such dilemmas are very widespread, and the very structure of the theory (rather than people’s circumstances) produces routine excuses for failing to meet obligations that nonetheless supposedly hold according to that same theory, it all rather makes me think that something has gone wrong with the theory. The major obligations it posits—or at least, a big chunk of them, namely, the ones concerning intentions not to act in all the innumerable ways we ought not to act—are largely toothless.

    Second, and related: it strikes me as particularly odd if the theory produces the result that we are so largely excused from criticism for failures to *intend* as we ought, but not likewise largely excused from criticism for failures to not-act as we ought—even though the obligations and reasons to intend are supposed to be the fundamental and explanatory primary ones, and when it is our intentions rather than “mere actions” (or omissions) that express our take on the world in virtue of which we are subject to normative demands in the first place. (By the way, like Keshav, I also love the way you end your essay: Our actions change the world in more or less beneficial ways—but it is only insofar as they express or embody our orientation toward the world that they fall within the realm of reason and normativity”. There seems to be *something* deeply right about that remark.) According to your view, it now seems that at the primary site of practical normative assessment, namely, the intention, we are routinely excused from one type of violation of obligation (failing to intend to not-Φ); but at the derivative, secondary site (the not-killing, the not-maiming, etc.) we are not likewise routinely excused. This seems odd to me. It’s somewhat akin to holding the following view: I have a decisive reason not to eat factory-farmed meat because, and solely because, I have a decisive reason not to contribute to the suffering of factory-farmed animals. I am somehow routinely excused from criticism for failures to abide by my reason not to contribute to the suffering of factory-farmed animals. But I am not likewise routinely excused from criticism for failures to abide by my reason not to eat factory-farmed meat. Seems rather odd and surprising. (It’s possible that this attempt to illustrate the oddity introduces disanalogies that muddy the waters. But I hope you get the general point I’m trying to drive at.)

    Third, and going back in part to the last set of thoughts in my initial question: If I understand things correctly, it’s the costs of deliberation that are supposed to routinely excuse from criticism for failures to intend to not-Φ. I’m not sure why similar costs of deliberating about not-killing or not-maiming (for some people) wouldn’t, on your view, produce excuses for them for failures to not-kill or not-maim. Yet the idea that the cost of deliberation is enough to excuse not-killing or not-maiming seems quite counterintuitive. Again, you might try saying in response that the cost of deliberation isn’t enough to excuse failures to not-kill or not-maim, because the fact that a person will suffer or die if you maim or kill is an extremely strong reason to not-kill and not-maim; and any excuse that could possibly work would also have to be extremely strong. But by the Special Hypothesis, again, the fact that a person will suffer or die if you maim or kill is a reason – and presumably an equally strong reason – to intend to not-kill and to intend to not-maim, if it is a reason to not-kill and not-maim. By parity of reasoning, then, any excuse for failures to abide by these strong reasons to intend would likewise have to be extremely strong. So how can the costs of deliberation be enough to excuse from failures to abide by the (strong) reasons to intend to not-kill, to not-maim, etc., if equally high costs of deliberation are not enough to excuse from failures to abide by the (equally) strong reasons to not-kill, to not-maim, etc.?

    Thanks for any further thoughts you or Jonathan may have about these topics – and sorry if you have already addressed these things in your book and I either failed to appreciate it or didn’t get to that part of the manuscript – I confess I didn’t yet read the entirety.

    Btw, I was also wondering about reformulating the Special Hypothesis along the way that Keshav suggests, to get around some of these problems. But I do see the issue you raise with that, Jonathan.

  17. Thanks, Hille, for these very helpful and challenging further comments. You express three worries for the view that, in the relevant cases, you ought to have the intention but are not criticisable for lacking it because you will only have it if you deliberate, which you ought not do. I see the force of all three worries, but I’ll briefly indicate some possible lines of response.
    1. First worry: the view entails widespread dilemmas, in which you ought to intend, ought not deliberate, but will not intend unless you deliberate. One point here is that if, as many of us think, there are no ‘wrong-kind’ reasons for attitudes, dilemmas of this general sort could be quite commonplace – for instance, incentives for attitudes might often make it the case that you ought to bring it about that you have attitudes you ought not have. Perhaps more importantly, though, the dilemmas in question won’t undermine deliberation, in the sense that if you are deliberating about, say, whether to kill, then your deliberation can permissibly conclude in one and only one way – with an intention not to kill. The fact that you ought not be deliberating in the first place doesn’t change that.
    Second worry: the view entails that you are routinely excused (not criticisable) for not intending in these cases, but you are not routinely excused for acting/not-acting, even though reasons for intending are more fundamental. I wonder if some of what we say in section 3.1 of the paper might help with this – in particular the point that the fundamentality of reasons for intention doesn’t entail that intentions are more important than actions. Note also that the view doesn’t entail that there are any cases where you are critcisable for acting/not-acting but not criticisable for your (lack of) intention – on the contrary, we’re committed to there being no such cases.
    Third worry: if the costs of deliberation excuse you from criticism for not intending, why don’t they similarly excuse you from criticism for acting/not-acting? I suggest: because you won’t intend as you ought unless you deliberate, whereas the same is not true of acting/not-acting as you ought. For example, you can (and presumably will!) not-kill even if you don’t deliberate.
    These are just indications of possible responses and they don’t address all aspects of the worries, which I think are good ones. Thanks again for taking the time to set them out.

  18. Thanks Keshav. I think maybe we’re not thinking about analyses in the same way or something (or else I’m missing something). I was taking it that when to be F is to be G, then something is F iff it’s G. So if to be a reason not to φ is to be a reason not to intend to φ, then all reasons not to intend to φ are reasons not to φ. That’s why it seemed to be a problem that things like the costs of deliberation can count against intending to φ without counting against φ-ing.

    Thanks very much for your further thoughts, Hille. I certainly agree that the results of our view are surprising here. As Conor says, I do think we can mitigate this but some oddness certainly remains.

    Just to say a touch more on your third worry. It does seem to me that we can explain why the costs of deliberation impact the normative status of deliberating about whether to kill without doing the same for killing. Whether you ought to deliberate depends on the reasons for and against deliberating. In the cases at issue, there is no need to deliberate – so nothing really counts in favour of deliberating – and costs to doing so. Thus the reasons against deliberating win out. Whether you ought to kill depends on the reasons for and against killing, and obviously the latter win out there. In other words, the reasons against killing aren’t, in these cases, reasons for deliberating, and so don’t bear on whether you ought to deliberate.
    By the way, it might also be worth noting that one way to try and avoid these results would be to appeal to a different picture of how reasons determine oughts. I’m not sure exactly how that would go, but just wanted to note that the problem here isn’t coming just from the Special Hypothesis.

  19. Thanks, Jonathan. That clears up my confusion.

    I think I defaulted to thinking of the conditional instead of the biconditional because the conditional is all you need to establish the General Hypothesis. You only need to show that all reasons for action are fundamentally some sort of reasons for intention, which is compatible with there being reasons for intention that aren’t reasons for action. So now I’m wondering, why not commit only to the weaker claim, and leave open the possibility that there are reasons for/against intending that aren’t reasons for/against action?

    At some points in the paper, it seems like you’re already amenable to this. You mention the possibility of reasons against intending to φ that are not reasons against φ-ing in response to Heuer’s cases. And you say that if wrong-kind reasons are genuine reasons, the Special Hypothesis can be easily amended. So why not say the same thing with regard to the kinds of state-given reasons against intention Schroeder talks about, insofar as those are genuine reasons? Wouldn’t that get you out of more of the counterexamples, as I’ve suggested, while still getting you what you need for the General Hypothesis?

    It also occurs to me that the Anscombean way of going, where reasons for action are analyzed in terms of reasons for intentional action/intention-in-action, might also avoid this problem. Because it seems like Schroeder’s cases of reasons against intention are reasons against future-directed intention, not intention-in-action.

  20. Thanks Keshav, that’s a good question. As you say, we are open in the paper to taking reasons bearing on action to be a restricted class of reasons bearing on intention. So the question will be what restriction is needed here. That’s to say, we’ll need to fill in the formula: to be a reason not to φ is to be a reason, of kind K, not to intend to φ. To avoid the problem cases, we’ll need kind K to exclude things like the costs of deliberation. But then I worry that the remaining reasons not to intend to φ will just be the reasons to intend not to φ. If so, we’re back to where we started.
    Thanks also for the suggestion that the Anscombean approach might help. That’s pretty interesting, and I’ll need to think about it. One quick concern is that if the idea is that reasons not to φ are reasons not to intend-in-action to φ, then whenever you’re required not to φ, you’ll also be required not to intend-in-action to φ. But in some of these cases, you’re just permitted not to have the intention, not required not to do so.

  21. Thanks, Jonathan.

    Let’s say we formulated reasons of kind K as something like “reasons that bear on the correctness of the intention.” It seems like costs of deliberation don’t bear on whether the intention is correct, so they’re ruled out. But it seems like the fact that more information is forthcoming does bear on the correctness of the intention, in roughly the same way higher-order evidence that more evidence is forthcoming bears on the correctness of a belief. I haven’t thought this through in that much detail, but wouldn’t this (a) exclude the costs of deliberation and (b) include reasons not to intend to φ that are not reasons to intend not to φ?

    With regard to the Anscombean approach: in what cases would it be required not to φ but permissible to intend-in-action to φ? I’m having trouble imagining how such cases would work, because it seems like intending-in-action to φ would entail that you φ.

  22. I found this piece to be extraordinarily compelling upon the first read, but upon further examination I found a peculiar stipulation that is not accounted for in regard to the response constraint. If it is the case that as C&J claim that the response constraint in regard to actions tells us that, “The response constraint tells us that a reason for acting is a reason that can be acted on.” (McHugh et al. 157). Then, how does the response constraint relate to reasons for intention? It seems as though it must also be the case that reasons for intending are reasons that can be intended upon. But, it seems that not all reasons that can be acted upon are also reasons that can be intended upon which poses a severe problem to the Special Hypothesis if correct. Do C&J have a response to this concern?

  23. *Continued*
    It seems that when we undergo the practical reasoning process we tend to have reasons that speak both in favor and against forming an intention, reasons for and against certain act types. But, when considering an act type it seems as though the agent considers the reasons that support both that act type and the inverse of that act type. Surely both act types have reasons that speak in favor of them, but in some instances it seems as though the reasons that speak in favor of an act type can never lead to the formation of the intention.
    Take for example a dispute one may have with a neighbor. The neighbor has cut a tree down and the tree has fallen on your house. You take this act to be a reason to cause significant harm to your neighbor’s property, the reason speaks in favor of an action that is within the agent’s power to perform, an action that can be acted on. This would seem to make this reason a reason for action. But, this same reason could never lead yourself to actually intentionally damage your neighbor’s property in retaliation because you have other reasons that speak against forming the intention to do so, therefore though it is a reason to form the intention it is not a reason that can be intended upon. That is if I am correct in assuming it to be appropriate that the response constraint applies both to reasons for action and reasons for intention, which seems as it should be the case.

  24. Hi Adam, thanks for the comments and sorry for the slow reply. You’re worried that the response constraint, according to which reasons can be responded to, might not apply to reasons for intention (or perhaps that, if it does, then some reasons for action are not reasons for intention). I don’t think I’ve quite understood the basis for the worry, though. In your example, the damage to your house presumably is a consideration you could respond to by intending to damage your neighbour’s property. It’s just that you wouldn’t do so, given the stronger reasons not to have that intention. Similarly, it’s a consideration you could, but would not, respond to by damaging your neighbour’s property. So I don’t see a difference between the intention and the action, in that respect. Am I missing something?

  25. Hello Dr. McHugh, sorry for the lack of clarity. But, I believe my worry is founded mostly upon the nature of intentions not necessarily being the same as that of actions. While actions are necessarily concerned with the present, intentions can be future directed. In such, it seems as though there is a distinction to be made between reasons that speak in favor of present directed intentions versus those that speak in favor of future directed intentions. It seems as though the special hypothesis is correct in regard to present directed intentions, but to claim reasons for action are necessarily reasons for intention seems to take away from the stronger normative value that reasons for future directed intentions seem to hold.
    For instance, in the law there is a distinction between first-degree murder and second-degree murder. While second-degree murder is intentional, it is not premeditated unlike first-degree murder which is premeditated. My suggestion being that the Special Hypothesis may be revised to, “Reasons for action are fundamentally reasons for present-directed intention.” This revision would serve to preserve the notion that reasons for future directed intention tend to carry more weight than those that speak in favor of present directed intention, while also allowing for the primary aim of the Special Hypothesis to ring true.
    I apologize that my first serious of comments were somewhat unclear, they were not yet a finalized thought.

  26. Hi Adam, thanks a lot for this further comment. That’s an interesting point about the distinctive significance, at least sometimes, of prior intentions. My initial reaction is that, rather than casting doubt on the Special Hypothesis, this might have something to do with what prior intentions, in particular, reveal about the agent. But I’d like to think more about it. Thanks again.

Comments are closed.