Thoughts/propositions as motivating reasons

38 Replies to “Thoughts/propositions as motivating reasons”

1. Clayton, nice post! I imagine I might be one of the opponents you have in mind (especially with the implicature story.)
   I favour the following line of thought (which I take to be Dancyan). When we’re talking about motivating reasons, we’re talking about intensional contents. Intensional contents can be things that don’t exist. So motivating reasons in the illusory case are still facts, but they are nonexistent facts. (Just as, when I’m afraid of Dracula, the thing I’m afraid of is a vampire, albeit a nonexistent one. Dracula is the reason I’m hiding under my bed!)
   The linguistic counterevidence to this, as you note, is that then (1) wouldn’t entail (2), but it seems to. I take your point about the problem for the implicature story (although I still think I can get a reading of (1) on which it doesn’t entail (2)), but how about the following variation?
   Because ‘reasons’ explanations are in general (i.e. other than in the motivating reasons case) factive, and because there’s a close connection and easy confusion between the reasons FOR WHICH you do A and the reasons WHY you do A, we have a linguistic convention that you insert ‘she believes’ (more transparently, ‘as she believes’) when you take her reasons to be illusory. Then you’re right, it’s not merely an implicature, but a matter of semantic convention, that asserting (1) commits you to (2). However, it remains the case that philosophically we should say that people can be motivated by reasons/facts that don’t exist.
   What’s wrong with that story?

2. Hey Steve,
   Thanks for the comments. I’m a bit tired, so I can’t promise this will be on track. I’m also pretty sympathetic to everything that Dancy says. I think our only disagreement is about that hole he digs himself. You think it’s a non-existent hole, I think it’s probably real, but now I’m worried that my reason for writing the post is that there’s a non-existent hole that Dancy has dug himself.
   I have to admit that I’m not sure how to test or understand your proposal. (Not that this is your fault. The fault is entirely with me.) Is the thought that (1) as stated might entail (2), but really that doesn’t matter so much because what we really _mean_ to say when we specify an agent’s motivating reasons is something close to (1), but not (1) itself? I think when we spoke about this last time, that was something you were sympathetic to. So, maybe (1), properly written, is something like ‘She voted for Bill for the reason that, as she believes, Charlie is a crook’?
   I guess I do say things like, ‘She voted for Bill because, as she sees it, Charlie is a crook’ and it’s not obvious at first whether I’m on the hook for saying that Charlie is a crook. The problem for me is that that sounds a lot like, ‘She voted for Bill for the reason that she thinks Charlie is a crook’ and ‘Her reason for voting as she did was that she thought Charlie was a crook’. I don’t want to say that these are false, but I’d rather not be forced to say that this is all we can say.
   I wonder if it’s bad to say ‘Alice voted for Bill for the reason that, as she believes, Charlie is a crook but she didn’t vote for Bill because Charlie is a crook’? That seems bad to me, but I wonder if that’s evidence against the view I’m imagining my opponent (you?) might take.

3. I’m suggesting that semantically (1) might entail (2), but that (1) is only semantically proper for the case when the motivating reason is veridical rather than illusory. In the illusory case, the semantics requires some sort of disclaimer (like ‘as she believes’).
   The last sentence you consider doesn’t sound bad to me, although my intuitions are probably influenced by theory. However, ‘she didn’t vote for Bill because Charlie is a crook’ is ambiguous, I think. If you fix on the purely explanatory reading (she wasn’t caused to vote for Bill by Charlie’s being a crook), then I think this sentence is good. But if you fix on the motivating reason reading (she didn’t vote for Bill for the normative reason that Charlie is a crook), then the sentence is bad, because it reads as follows:
   “Alice voted for Bill because she took it to be a normative reason that (as she believed!) Charlie is a crook. But she didn’t vote for Bill because she took the (actual!) fact that Charlie is a crook to be a normative reason.”
   There are two things wrong with this claim. First, the two sentences contradict each other about the cause of Alice’s voting for Bill. Second, the two sentences are in at least pragmatic contradiction about whether Charlie is a crook.
   I’m not sure how you’d test this proposal. I’d support it by an argument to the best explanation. It’s just very natural (as you agree) to think that by ‘the reasons that motivate people’ we mean those things that they take to be normative reasons and are thereby motivated by. And it just seems wrong that in the illusory cases, what people take to be their reasons, and are motivated by, are their mental states. So it looks like we need some other explanation of why we say ‘Her reason was that she believed…’ Blocking factivity then looks like a natural candidate.
   Perhaps the best reason for skepticism about this explanation is that we need an explanation of why motivating reasons talk would be factive. For what it’s worth, here’s the story I like. Semantically, ‘reason’ just means ‘explanation’, and ‘explanation’ is factive about both explanans and explanandum; i.e. p ‘explains’ q only if p and q. But talk about ‘S’s reason for doing A’ is a contraction or idiom, which can be expanded as ‘S’s reason/explanation for why to do A’ (or: Why doing A is good/what she ought to do). (Note that although ‘the explanation’ is factive, ‘her explanation’ (or ‘the explanation she accepted’) is not. ‘Her explanation for p’ can simply mean: what she accepts as/gives as an explanation for p.)
   So why would ‘The reason for which S did A’ have a presupposition of factivity of explanans/ reason? I suggest it’s because of the contraction. Superficially, it looks like the explanandum here is ‘S did A’, which is factual. So it looks like we’ve got a regular case of explanation, in which case the explanans (the reason) also has to be factual. So the factivity of ‘reason’/’explanation’ semantically requires a fact.
   However, the real content of this thought is revealed by the following:
   ‘p is the reason/explanation [that S accepted, for why to do A], for which S did A.’
   Here, (1) the explanandum is ‘why to do A’, rather than ‘why S did A’, and (2) it’s only a putative, rather than a real explanation, so there’s no factivity.
   In short, the proposal is that the insertion of ‘she believes’ is explained by a confusion about the real form of the sentence (due to the contraction/idiom), and the semantic rule that ‘reason’/’explanation’ is factive. (Actually, that’s a bit different from the explanation I started out with! Here the psychologization is not a disclaimer, but a confused attempt to satisfy the factivity requirement. So now I’ve offered you two theories.)
   I fear I may have only deepened your puzzlement about my proposal…

4. Interesting. I’m with Steve – I’ve actually tried to argue elsewhere that all reason-statements are intensional contexts even if I do hold that reasons are facts.
   Anyway, it seems to me that:
   (2) Alice voted for Bill because Charlie is a crook
   is ambiguous. It can either mean:
   (2c) That Charlie is a crook caused Alice to vote for Bill
   or
   (2u) That Charlie is a crook enables us to understand why Alice voted for Bill.
   Same ambiguity can be found from (2′).
   That is, there are causal explanations and reasons explanations/rationalisations. We use the term ‘because’ to express both. It is true that (2c) is factive and has existential import. It does entail (3). Yet, (2u) can be an intensional context just like (1). We can understand people’s actions as rational in the light of facts that didn’t obtain but only were thought to obtain. And, if we understand (1) to only entail (2u), then it doesn’t entail (3) and the facticity problem doesn’t arise. And, it’s clear that (1) doesn’t (2c) in any direct way.
   So, I think that you are right about the evidence that (1) entails (2) but once we disambiguate (2) we can block the inference to (3). And, the same goes for (1′). In fact, same kind of ambiguity between different senses of ‘explanation’ can be found from (4).

5. Hey, where are the people who agree with me?
   Steve wrote, “I’m suggesting that semantically (1) might entail (2), but that (1) is only semantically proper for the case when the motivating reason is veridical rather than illusory. In the illusory case, the semantics requires some sort of disclaimer (like ‘as she believes’).”
   Okay, I think I understand the suggestion better. (Sorry, I was really tired last night and drifting in and out of consciousness as I tried to type out my response.) I’m sympathetic to the “global” motivation of blocking factivity (it takes care of a problem that arises for the view that is clearly right!), but part of my interest here is to develop a factivity preserving response to illusion cases that allows me to be Dancyan. (It’s part of the series of papers in which I argue that various famous British philosophers have made tiny little mistakes and the bash their critics over their heads for defending hopeless internalist views that throw out the baby with the bathwater and then sets the house on fire.)
   So, you’re settled view seems to be this:
   “‘p is the reason/explanation [that S accepted, for why to do A], for which S did A.’
   Here, (1) the explanandum is ‘why to do A’, rather than ‘why S did A’, and (2) it’s only a putative, rather than a real explanation, so there’s no factivity.”
   That’s pretty good. So, the thought is that in citing an agent’s motivating reason(s), we describe the explanation that they’d offer, and this captures both the idea that we have to get their perspective on things right and we aren’t on the hook for saying that their (putative) explanations are factually correct.
   So, we can paraphrase (i) as (ii):
   (i) Alice voted for Bill for the reason that Charlie is a crook.
   (ii) That Charlie is a crook is the explanation Alice would offer for why she voted for Bill.
   There’s nothing wrong with appending to (ii) claims to the effect that Charlie isn’t a crook or claims to the effect that I don’t believe/know that Charlie is a crook, so (i) is fine if Charlie isn’t a crook. Is that the thought?
   Hey Jussi,
   I think I can agree that when we’re talking about reasons, we’re talking about intensional contexts _and_ say that reasons are always facts. When we ascribe knowledge, we can’t negate the propositions that we say the agent knows but we’re dealing with an intensional context. I take it that your main point is not that there cannot be intensional contexts with factivity but that ‘because’ and ‘explains’ statements can be taken in different ways. I agree, btw, that (1) doesn’t entail your (2c), but I worry a bit because I think (1) entails (2u) and I still have the residual worry that (2u) is true only if Charlie is a crook.
   Consider: I don’t know if Charlie is a crook. My guess is that he isn’t. That Charlie is a crook, however, enables us to understand why Alice voted for Bill.
   I feel like I know why Orly Taitz says the things that she does. She’s crazy. She believes that Obama isn’t a citizen. Those facts aren’t unrelated. I don’t think I can say: That Obama has a fake birth certificate enables us to understand why Orly is in court.
   I think we can offer explanatory reasons that explain why she’s there, but one way out of this mess is to distinguish explanatory and motivating reasons and say that when someone has lost their grip on reality, we cannot specify any fact that serves as the their reason for doing what they do, we know that, and that’s why we shift into talk of explanatory reasons.
   Maybe I should have posted after the presentation. You two are making me worried.

6. Hi,
   thanks. The first point was that we can talk about facts inside the intensional contexts without any resulting facticity as such. So, it can be true that ‘I am afraid that this plane will crash’. In this case, I’m not afraid of the proposition that this plane will crash, but rather the state of affairs or fact that this plane will crash which may or may not come to obtain. Of course, you are right that inside the intensional knowledge contexts fact talk is factive but this is because of knowledge being factive.
   Anyway, not to worry. That’s a good response. I’m starting to think that ‘understanding’ is ambiguous too in the same way as ‘because’ and ‘explanations’ are. ‘That Obama has a fake birth certificate enables us to understand why Orly is in court’ is easy to hear as a factive false causal statement. I’m starting to think that maybe there’s no simple way to disambiguate (2) that itself wouldn’t be ambiguous between the non-factive motivating reasons and the factive causes.
   This wouldn’t be altogether satisfactory but it would allow one to accept (1); deny that it entails (2) in the causal sense of ‘because’ that would entail (3); and accept that perhaps there is another sense of ‘because’ expressing reasons-explanations which, in effect, gives us just a restatement of (1) – which is thus obviously entailed by (1) but which does not entail (3). I know this still needs for instance an explanation of how the non-factive reasons work.
   Maybe they help us undertand the agent’s actions in terms of the normative reasons which the agent would have had, if the facts she thought to obtain would have really obtained. This would help to maintain the basic idea of Dancy that motivating reasons and normative reasons belong to the same ontological category.

7. We three seem largely in agreement, surprisingly. As you say, where are the nay-sayers? Some small points:
   Clayton: you’ve basically got the view right. But (ii) isn’t an accurate paraphrase. Rather, it should be
   (ii’) That Charlie is a crook is the explanation Alice accepted for why to vote for Bill, for which she voted for Bill.
   The explanandum is ‘Why vote for Bill?’, not ‘Why did Alice vote for Bill’. Also, note that it’s not the explanation Alice would offer, but the explanation that she accepted. (The difference between her rationalization and her reason).
   Jussi: I agree with Clayton on this; I don’t think that ‘That Obama has a fake birth certificate enables us to understand why Orly is in court’ has any acceptable (i.e. nonfactive) reading.
   It might be worth noting that the idea that motivating and normative reasons are of the same ontological category goes back at least to Williams (1979), who insists on it, and perhaps Davidson (1963), if not further. As I see things, the heresy starts with Smith, whose distinction is widely accepted until Dancy. (Smith cites Nagel’s 1970 distinction between explanatory and normative reasons, but as I understand Nagel, his ‘explanatory reasons’ are the reasons WHY someone acts, not the reasons FOR WHICH they act, and he wants to keep the reasons that motivate us in the same ontological category as the normative reasons).

8. Hi Clayton,
   What would you say if someone responded by saying that 1 doesn’t entail 2, but instead 1 entails the following?
   (2a) Alice voted for Bill because “Charlie is a crook.”
   which might be intended to convey (or perhaps equivalent to)
   (2b) Alice voted for Bill because, as she sees it, Charlie is a crook.
   I don’t think it’s unlikely that we “hear” 2 when spoken as 2a, which could then be glossed as 2b.

9. Hi Clayton,
   (1) Who are you thinking of when you say that, “those who defend such views tend to eschew the psychologized redescription of (1)-(4) that makes reference to facts about mental states and it denies that motivating reasons are facts.”
   (2) Couldn’t you embrace a disjunctive view, and just say, naturally enough, that 1-4 are only appropriately asserted in the veridical case? That *seems* to be at least one way to avoid the advice in your last sentence – but maybe I am not getting how your argument is meant to fit into the larger debate.

10. Also. You end with this: “Either go more radically towards a kind of externalist, factualist account of motivating reasons or embrace the more pscyhologized version that rewrites (1)-(4) to make reference to mental states because of arguments from error.”
    Why can’t I just rewrite 1 and keep 3-4 as they are (while denying that 1 entails them)?

11. Hey Brad,
    I was thinking of those who think of motivating reasons as thoughts/propositions rather than facts or the mental states that have those propositions as their contents.
    My feelings towards disjunctivism are mixed. I think Dancy credits Williams with the idea that the form an explanation of an action takes shouldn’t depend upon whether the agent’s beliefs are correct. So, if possible, I’d want to say that we shouldn’t say that one agent’s reason was the fact that p and another agent’s reason was just the belief that p or the thought that p when there’s no mental difference that distinguishes these subjects. That’s not an argument against the disjunctive view. My worry about the disjunctivist views are always that it seems in talking about motivating reasons, we’re talking about things like a basis for the agent’s action and I think the basis won’t be different without a mental difference.
    As for your suggestion, you could rewrite (1) in such a way that there’s no appearance of factivity, say, by making reference to a psychological state. I raise this issue because I’m trying to develop a paper where we resist that move and I offer a response on Dancy’s behalf to arguments from error that goes a bit more externalist than he seems to want to in Practical Reality. Here, I’m just looking to see if there’s linguistic evidence that suggests that he’d have to go further and do away with his idea of non-factive explanations. One view (the view I’m toying with) describes the agent’s motivating reasons as what the agent saw/knew that led the agent to act. If acting for a reason is a kind of achievement (i.e., the agent has successfully identified the nature of her practical situation and responded as she sees fit), the thing to say about the case of error is not the disjunctive thing, not the psychologized thing, and not the non-factive fact thing. The thing to say is that cases of error aren’t cases we use motivating reasons to describe. Instead, we offer explanatory reasons, reasons that explain why the agent did what she did. Those can be psychological states, but they aren’t motivating reasons and so we can accommodate the idea that motivating and normative reasons belong to the same ontological category.
    John,
    It’s an interesting suggestion. I’m not quite sure what to say at the moment. I don’t understand (2a), but (2b) is interesting. I’ll give it some thought. My initial reaction is that the “as she sees it” locution is more natural in cases where the reason we specify has a kind of evaluative component to it. It’s weird to my ears to say things like, “He drank the contents of the glass because, as she saw it, it contained gin” but not to say, “He asked her to dance because, as he saw it, she was the cat’s pajamas”. Anyway, my ears are very tired right now.

12. Clayton,
    2a is just supposed to be ordinary quotation. You’d quote someone as saying her reason was “Charlie is a crook,” and so voted for Bill because “Charlie is a crook.” (We could similarly characterize the content of a thought.) That shields us from any entailment or commitment.

13. Clayton,
    I take it that the “motivating reasons are psychological states” view is driven by a desire to provide a unified account of reasons-explanations, some of which include false beliefs on the part of the agent. So, for example, Alice believes Charlie is a crook and that’s why she votes for Bill. In some cases Alice is right about Charlie and in other cases Alice is wrong. But the thought is that in either case, we should explain Alice’s action the same way.
    One way to put this is as an inconsistent triad:
    1. When S “acts on” (fill in the story) true beliefs, the motivating reasons for S’s action are facts.
    2. When S “acts on” false beliefs, the motivating reasons for S’s actions are not facts.
    3. The motivating reasons for S’s actions, whether her beliefs are true or false, are always of the same ontological type.
    The “motivating=psychological” crowd gives up 1. You seem to want to give up 2. Disjunctivists give up 3. To defend your view, you just have to give a persuasive story about what’s going on in the case of acting on false beliefs. Dancy has a story, which is terribly unpersuasive in my mind.

14. Hi Clayton,
    Thanks.
    First, I was wondering who, specifically, you had in mind. Does Smith, e.g., “eschew” the re-written version of 1? But not a big deal.
    Second, I was interested to hear you say this: “…[it] seems in talking about motivating reasons, we’re talking about things like a basis for the agent’s action and I think the basis won’t be different without a mental difference.”
    I see why it makes you nervous about catching disjunctivitus, but it also seems to raise a worry for your view. Imagine, e.g., that P and S have mental states with type-identical content (etc) – the idea is that there is no mental difference between them.
    But one of P’s beliefs is false, while S’s analogous belief is true, and these beliefs are central to the explanation of why they act. Now it looks like your account will say they have different bases for action – P has am existing fact, while S has a non-existent fact – even though their is no mental difference between them.
    This suggests that your slogan tells in favor of some mental state view.
    Hope that rough description is enough to get the worry across…
    Finally, those comments about the larger dialectic help. Did not understand before that your main target was Dancy’s non-factive explanation view or that you were tempted to dramatically redefine motivating reasons!

15. Hey Heath,
    I tried to develop an account of the false belief case that’s an alternative to Dancy’s but one of the difficulties I ran into were people who didn’t think there was anything wrong with what Dancy already said. It’s hard to know how to develop that strategy until I know what people think about the logical relations between statements like my (1)-(4) above. The thought I had was that if you really bought into Dancy’s way of doing things (MR are facts, you shouldn’t say that they are facts only when the agent’s beliefs are correct, MR aren’t (just) facts about a subject’s non-factive mental states, MR are the considerations in light of which the subject acts) you shouldn’t try to say that there are non-factive and successful explanations of an agent’s actions in the case of belief in terms of MRs, you should say that in the case of error there are no MRs available to do the explaining and say that the thing to be explained by MR is a proposition that is true only in the ‘good’ case. If acting for a reason is a kind of achievement or success, you get pretty much everything Dancy wants. The view: if an agent’s (relevant) beliefs are correct, we can say that the agent A’d for the reason that p; if an agent’s (relevant) beliefs are mistaken, we cannot say that there’s any reason that was the agent’s reason for A-ing, but we can say that the agent A’d because the agent thought such and such. If we already have a distinction between reasons why/explanatory reasons on the one hand and motivating reasons on the other, this is how I’d use it.
    Yes, the suggestion faces a very serious objection, but I suspect that the objection will in its own way assume that motivating and normative reasons don’t belong to the same ontological category. I don’t think I quite understand why MR and NR have to come from the same ontological category, but the view I describe is I think the best way to respect this constraint and if you abandon this constraint, then I see nothing wrong with saying that MR are mental states and NR are facts.
    Hey Brad,
    I didn’t want to mention any names, they might come find me! I’ve read some papers where people want to reject Dancy’s view that MR are facts but seem to want to say with Dancy that we don’t need to rewrite (1). Instead, we just say that MR are propositions. If the logical relations between (1) and the rest are what I think they are, this propositionalist view is not a good alternative to the psychologist’s view.
    I’ve come clean some about what I’d say about arguments from error. I try to “flip” the disjunctivist strategy. Rather than say that the same phenomena takes place in the good and bad cases and then offer different explanations, I think the positions Dancy stakes out leads more naturally to a view on which your actions can be explained by reference to motivating reasons only when there are facts that you respond to. When we know that there isn’t or don’t know if there is, I think we offer explanatory reasons rather than motivating reasons to explain why you did what you did. Dancy accepts (I think) that reasons-why can be psychological states.
    Anyway, my main target is really someone who wants to reject Dancy’s view without embracing the sort of view that Smith defends, but I do think there are some minor problems with Dancy’s view.
    If I ever get this worked out, I’ll post a link to something in the case anyone cares to take a look.
    Final thought. Suppose Audrey ran down the hall for the reason that Bob was chasing her. Coop says as much and he told me that. I’m skeptical so I ask her.
    Me: Coop told me that you ran down the hall for the reason that Bob was chasing you. I asked him if he really knew that and he said he did.
    Audrey: He did, I told him.
    Me: But, Bob wasn’t there.
    Audrey: Oh, I know that now but I didn’t know that at the time.
    It seems _very_ weird to me to say that someone could say truthfully, ‘He knew that the reason for which she A’d was that p’ when it’s known that p is false or when it’s not known that p is true. I don’t see why this would be if at the time of action the agent believed p and motivating reasons aren’t facts. Audrey should say that Coop didn’t know what he’s supposed to know, he just thought he did.

16. Clayton,
    FWIW, I agree with your views about (1)-(4) and I think this
    acting for a reason is a kind of achievement or success
    is a great way to put it. Good luck!

17. Consider:
    1. Socrates was executed for corrupting the young.
    2. Socrates was executed because he corrupted the young.
    3. Socrates corrupted the young.
    Most folks believe that (2) entails (3), because ‘P because Q’ (corner quotes) entails ‘P and Q’ (ditto). Most folks believe that (1) neither entails nor excludes (3): It gives the reason those who executed Socrates had for executing him, and the author of (1) may or may not believe that Socrates did corrupt the young.
    The distinction between (1) and (2), on this understanding of them, is a point of syntax in Greek. Both would be expressed with the Greek for “Socrates was executed” followed by a participial phrase expressing “corrupting the young” and with the participle “corrupting” agreeing with “Socrates. They would then be distinguished by the use of one particle to express (1) and another particle to express (2). (Greek particles are something like interjections in English, but more frequent and much more important, as this example makes plain.)
    FWIW, the grammar books sometimes express the difference between the two particles by saying that with the first, “there is no implication that the cause assigned is one in which the speaker believes.”
    Hope this helps.

18. Charles,
    Thanks for that, very interesting.
    Suppose I say:
    (1) Bill was sentenced for trespassing.
    I think (1) might be consistent with saying:
    (1+) Bill was sentenced for trespassing, but he was innocent of the charge.
    If (1) and (1+) are consistent, I wouldn’t say that (1) entailed:
    (2) Bill was sentenced because he trespassed.
    I’d say that it was something like this:
    (3) Bill was sentenced because he was convicted of trespassing.
    (3) doesn’t entail (2).
    What about:
    (4) Bill was sentenced for the reason that he trespassed.
    It’s here that I confess my intuitions get a bit murky. I wouldn’t want to say that (1) entails (4) (wouldn’t want to deny it, either). I’m fence sitting on this one because it’s hard for me to hear (4) in such a way that it’s fine to say:
    (5) Bill was sentenced for the reason that he trespassed, but he was innocent of all charges.
    I don’t think there’s anything wrong with:
    (6) Bill was sentenced for the reason that he was believed to have trespassed, but he was innocent of all charges.
    I wonder if there’s much difference between (1+) and (6). If not, I wonder if the defectiveness of (5) is a reason to think that there’s an important difference between (4) and (1).

19. Clayton,
    Interesting post. (Did I come find you?)
    Just a couple thoughts.
    “for the reason that” can be used for purely explanatory reasons (reasons which aren’t also motivating reasons). A google search confirms this. So for instance:
    “It is a physical impossibility during such a short session for Congress to give attention to much general legislation for the reason that it requires practically all of the time to dispose of the regular appropriation bills.”
    It can be used for purely explanatory reasons even when it follows something of the form “S Aed”. Consider:
    “Bob flunked the test for the reason that he came into the classroom only in the last 10 minutes of the test period.”
    An aside here: if you google “for the reason that,” you’ll find that the grammar police hate it. Replace it with ‘because’, they say.
    Rather than thinking of motivating reasons as canonically formulated “S Aed for the reason that p,” maybe we should prefer this:
    (1) S’s reason for A-ing was that p.
    Now consider:
    (2) S A-ed because p.
    Abstractly, these do sound equivalent and they sound like the imply
    (3) p.
    But when we consider examples, it’s a little tricky — especially when you look at negations. After all, if 1 (or 2) entails p, then not-p should entail not-(1) (and not-(2)).
    But consider a case in which the judge excludes a confession, and explains himself this way:
    “My reason for excluding the confession was that it was obtained under police duress.”
    Later on, matters are cleared up, and becomes widely known that the confession wasn’t obtained under police duress. The judge, I take, won’t say:
    “Well, what I thought was my reason wasn’t really my reason.”
    And the judge certainly won’t agree that he had no reason for doing what he did. The natural thing for the judge to do is to reaffirm his reason-claim adding a qualification about how, though the confession wasn’t obtained under duress, he reasonably thought it was at the time. (And this isn’t the same as saying his reason was some fact about his beliefs.)
    I think the same works for 2 here. The judge won’t say “I was wrong when I said that I excluded the confession because it was obtained under duress”. Nor would he say “I didn’t realize my reason was something different than I thought it was.” The judge would do best, I think, to reaffirm his earlier statement but with the relevant qualification.
    By the way, I think you get this phenomenon with a number of propositional attitude verbs/verb phrases that are often classified as factive (‘regret’, ‘am happy that’), but not with others (‘know’).
    I don’t know enough about the relevant literature in linguistics and philosophy of language, but I would think one possible hypothesis is that ‘S’s reason for A-ing was that p’ defeasibly presupposes that p. The defeasible presupposition would have to be pretty robust, though. When you and your conversational partners know p is false, you can’t ever away with a bare use of *My reason for A-ing was that p*.

20. Um, I believe I’m one of the ‘people’ who think that motivating reasons are propositions, rather than facts or psychological states.
    I agree with pretty much everything that Matt said this morning; when we think that P, we report someone’s motivating reason as having been that P, and when we don’t think that P, we report her motivating reason as having been that she thought that P. But we don’t change our mind about what her reason was, when we change our mind about whether P, and two people who disagree about whether P can agree about what her reason was. (I discuss this in note 23 to chapter 1 of Slaves of the Passions.)
    So I think everyone should agree that reporting the motivating reason as having been the belief and reporting it as having been the content of that belief are just different ways of reporting the same thing, and that whether we believe that P merely determines which way we will report it. Consequently, if I don’t believe that P, I won’t accept Clayton’s ‘canonical’ formulation of the motivating reasons claim. That doesn’t mean that I deny that the motivating reason is a false proposition in such a case; on the contrary, I think it is – I just wouldn’t put it in ordinary English by saying, ‘her reason was that P’, because I would expect to be understood as having presupposed that P.
    Similarly, I think that ‘wants’ relates agent to propositions, so in particular I think that if Sue wants Dan to die, then what is happening is that Sue is related by the wants relation to the proposition that Dan dies. But I wouldn’t put this in ordinary English by saying that what Sue wants is the proposition that Dan dies. In this case, what goes wrong is that when ‘wants’ takes an NP argument as in ‘Sue wants NP’, it generates the reading, ‘Sue wants to have NP’. It doesn’t show that ‘wants’ doesn’t always relate agents to propositions; it just shows that you can’t relate an agent to a proposition in the relevant way just by saying that she wants it.
    I think a different but similar thing is going on in the motivating reasons case. The fact that when I don’t believe that P, I can’t aptly describe Sue’s reason for phi-ing by saying that her reason for phi-ing was that P, doesn’t show that her having the motivating reason for phi-ing that she did doesn’t consist in her bearing the had-as-her-motivating-reason relationship to the proposition that P. It just shows something about the further commitments of saying ‘her motivating reason was that P’.

21. Maybe I should have named names so I could pick my fights more carefully. My main motivation in this project is to try to defend the view that normative reasons are facts from those who side with Dancy in thinking that normative and motivating reasons belong to the same ontological category but then argue from cases of error that normative reasons are either the contents of propositional attitudes or the attitudes themselves. So, I think my main target might be people who accept Matt’s picture but I’m not sure if my main target in this project is Mark. Lucky me, I get to say things that offend the sensibilities of lots of people.
    Matt,
    You wrote, “And the judge certainly won’t agree that he had no reason for doing what he did. The natural thing for the judge to do is to reaffirm his reason-claim adding a qualification about how, though the confession wasn’t obtained under duress, he reasonably thought it was at the time. (And this isn’t the same as saying his reason was some fact about his beliefs.)”
    I’m not sure that’s right. “Well, it turns out I didn’t have a reason to exclude that confession after all. I thought I had a reason to exclude the confession because there was good evidence that it was coerced, it just turns out I was wrong” is a very natural thing to say.
    Speaking of the judge from the 3rd person perspective, consider the exchange:
    A: The judge excluded the confession because it was coerced.
    C: How do you know that?
    A: I know that because that’s what the judge told me.
    C: But, I just read in the paper that the confession wasn’t coerced.
    A: That’s correct, the confession wasn’t coerced. Still, strictly speaking, the judge excluded the confession because it was coerced.
    I take it that on your view, A doesn’t speak falsely at any point and (is this right?) A’s last remark isn’t defective. It seems weird (to me) to think, ‘I know the judge A’d because q’ doesn’t entail q. I can’t tell if you
    (i) disagree;
    (ii) think the ‘knows’ does something funny here;
    (iii) want to say distinguish ‘because’ and ‘explains’ statements from ‘for the reason’ statements.
    I suspect that it’s (i), but then I think this is a fair question to ask. If we’re reporting what someone believes or says, appending statements that say that you don’t believe what they said/believed or deny what they said/believed strikes nobody as odd. If we’re reporting what someone knows, appending statements that say that you don’t believe what you say they know or that you deny what you say they know, this strikes everyone as being odd. I think there are lots of us who think it’s odd to say:
    (1) He did it for the reason that p but I don’t know if p/but ~p.
    (2) He did it because p, but I don’t know if p/but ~p.
    It seems what generates the defectiveness in these cases is precisely the thing you suggest cancels the implication or presupposition that p. Isn’t the apparent defectiveness evidence that this is more like the ‘knows’ case than the ‘believes’ case? (Of course, it can be evidence that is overridden by arguments for a theory on which these things turn out to be correct, but I’m just trying to better understand your take on things.)
    Hey Mark,
    I can see that my plan on trying this stuff out before we meet this weekend has failed. You wrote:
    “I agree with pretty much everything that Matt said this morning; when we think that P, we report someone’s motivating reason as having been that P, and when we don’t think that P, we report her motivating reason as having been that she thought that P. But we don’t change our mind about what her reason was, when we change our mind about whether P, and two people who disagree about whether P can agree about what her reason was. (I discuss this in note 23 to chapter 1 of Slaves of the Passions.)”
    I agree that we don’t change our minds about what her reason was. (I think disjunctivists try to pull this shift off, and I think it fails.) So, I think we agree on a few points. If the speaker doesn’t believe p, the speaker won’t say:
    (3) She did it because p;
    (4) She did it for the reason that p.
    If the speaker _does_ believe p, the speaker will say (3) and (4). And, it’s crazy to think that in learning that ~p, the speaker will say “Oh, I thought she had been motivated by the facts, but it turns out she was motivated by a thought instead”. So far, we’re agreed. One thing to say, in learning that ~p is that there wasn’t a motivating reason and there’s not going to be a true explanation of the deed in terms of her motivating reasons. That’s not the way that you go, clearly, but I have a question about your overall view. Am I right to assume that you don’t buy into the idea that motivating and normative reasons belong to the same ontological category?
    One question I’m interested in has to do with the logic of motivating reason ascriptions. Another has to do with the ontology of normative reasons. I think that the best approach for someone who insists that normative reasons are facts and motivating reasons belong to the same ontological category as normative reasons is to say in the case of error that there’s no true ‘because’ statement that specifies the agent’s motivating reasons. Rather than redescribe the agent’s motivating reasons, we try to explain what brought it about that the agent behaved as she did.
    At any rate, I want to note that I’m not defending the view that says that agents in the good and bad case have motivating reasons that are different. I’m trying to argue that someone who likes most of what Dancy says should just say that in the bad case, the agent fails to act for a reason because if you buy into everything that he says, you should think of acting for a reason as a kind of achievement.

22. Clayton,
    Suppose ‘S’s reason for Aing was that p’ entails ‘p’. Then ‘not-p’ will entail ‘It’s not the case that S’s reason for Aing was that p’. It’s then a good idea to “check the negations,” as DeRose says, when asking about entailments.
    Consider ‘know’. Suppose I say ‘I know p’. Later I find out that not-p. I won’t continue to insist “I did know p, it just turns out that p is false,” and certainly I won’t go in for the nearly incoherent “I did know that, as I thought at the time, p.” I’ll say, “I guess I didn’t know that p, after all.” This is the predicted pattern for an entailment. So all is well.
    Compare ‘am happy that’. Suppose I said, “I’m happy that Gore won Florida,” after hearing this announced election night 2000. The next day I learn he didn’t win Florida. I take it that I won’t say “I guess it turns out I wasn’t happy that Gore won Florida.” I won’t think my earlier claim was false. Of course, I won’t just reaffirm as is. My reaffirmation will have to involve a qualification: “I was happy that, as we all thought at the time, Gore won Florida.” (You get a similar pattern for ‘regret’ – I think DeRose mentions this somewhere.)
    There is that difference, then, between ‘know’ and ‘am happy that’. Do you agree about the difference? I’m not asking whether you agree that ‘know’ is factive and ‘am happy that’ isn’t, but only about the different data we get when we check the negations in these cases.
    I think the negation data with ‘S’s reason for A-ing was that p’ is like that of ‘am happy that’. When we find out later that not-p, we don’t find it correct to say “Well, I guess my reason for A-ing wasn’t that p’. Maybe we can at least agree that the case for entailment is stronger with ‘knows’ than with either ‘am happy that’ or ‘S’s reason for A-ing was that p’.
    Concerning your imagined conversation:
    A: The judge excluded the confession because it was coerced.
    C: How do you know that?
    A: I know that because that’s what the judge told me.
    C: But, I just read in the paper that the confession wasn’t coerced.
    A: That’s correct, the confession wasn’t coerced. Still, strictly speaking, the judge excluded the confession because it was coerced.
    The first statement is defective. A shouldn’t be putting it this way if he believes the confession wasn’t coerced. He should say “because he thought it was coerced” or “because, as he thought at the time, it was coerced.”
    As for the last statement, I incline to the view is that it is true. But I think the last line is still defective. In my earlier post, I didn’t say exactly where the qualification would have to go to render a reason-statement with a false that-clause non-defective. I do think it’s hard to make such a statement ok except by inserting the qualification directly into the sentence itself as a side comment. Compare ‘am happy that’. It’s hard to get away with “That’s correct, Gore didn’t win Florida. Still, strictly speaking, I was happy that he won Florida.” That’s bad. Also, when one doesn’t endorse a presupposition made in a previous utterance, it’s often more acceptable to say “what was said was strictly speaking true, even though p isn’t true” than it is to reaffirm what was said using the same sentence. A might say, non-defectively I think, “What the judge said was strictly speaking true, but of course he was wrong about how the confession was obtained.”
    There are analogues here with epithets and expressive uses of language. Suppose Bob is a Republican, supportive of Republicans in Congress. Olbermann says, “Boehner is a Rethug”. Bob is not going to want to assert the negation of this sentence, or to say that what Olbermann said is false. Nor is Bob going to assert the sentence as is, even with a preceding comment disowning the expressive element. It’s possible, though, to say something like “what Olbermann said was strictly speaking true, but can’t we do without the namecalling?”.
    Fantl and I do get into some of this stuff in the book in chapters 3, 4 and 6. But it’s not central to our project. If someone insists that only facts can be reasons, they can buy into the main conclusions of the book. If someone insists that only mental states can be reasons, the same goes for them. In ch. 3 we work out ways for the “reasons as mental states” people to buy into our project, and in ch. 4 we do the same for the “reasons as facts” people. Big tent.

23. “Am I right to assume that you don’t buy into the idea that motivating and normative reasons belong to the same ontological category?”
    Sorry, I should have put it this way. Am I right to assume that you think ascriptions of normative (objective) reasons are factive but ascriptions of motivating reasons aren’t?

24. Hi, Clayton.
    I’m looking forward to discussing this tomorrow and Saturday in Austin. Actually, I do subscribe to the ‘normative and motivating reasons belong to the same ontological category’ view – I think they’re both propositions – normative reasons only when true, and motivating reasons only when the objects of belief or perception (and some other conditions obtain). So if facts are just true propositions, then I agree that normative reasons are facts, but if facts are something other than true propositions, then I don’t agree with that part.
    One argument that normative reasons are facts as opposed to propositions is that it doesn’t make sense to say that the reason for Ryan to help Katie is the proposition that Katie needs help; rather, we say that the reason is that this proposition is true or the truth of this proposition. I think this is like the argument that I mentioned in my last comment, to the effect that desires can’t relate us to propositions, because we don’t desire propositions – we desire the things the propostions are about. I think it fails for the same reason I outlined above for the desire argument.
    Ah – now I see Clayton’s clarification of his question. Let’s make Matt’s distinction – call it the distinction between Factive1 and Factive2. ‘Knows’ is Factive1; ‘regrets’ and ‘is happy that’ are Factive2. Like Matt, I think that motivating reason ascriptions are Factive2. I think that what we commonly call normative reason ascriptions are Factive1. So I don’t think that acting for a motivating reason is the same kind of achievement as acting for what we commonly call a normative reason.
    I should add that once again, I agree with everything else that Matt said.

25. Mark and Matt,
    Thanks for the helpful comments. (Helpful like a punch in the face for someone studying karate, but helpful.)
    I’ve been hoping to get back to this today to say something useful, but I haven’t found the time to work it all out. I need to think about this factive1/2 distinction. I think I can agree that there’s a difference between ‘knows that’ and ‘happy that’, but I’m not sure what Matt thinks the difference comes to.
    Matt wrote:
    “Compare ‘am happy that’. Suppose I said, “I’m happy that Gore won Florida,” after hearing this announced election night 2000. The next day I learn he didn’t win Florida. I take it that I won’t say “I guess it turns out I wasn’t happy that Gore won Florida.” I won’t think my earlier claim was false. Of course, I won’t just reaffirm as is. My reaffirmation will have to involve a qualification: “I was happy that, as we all thought at the time, Gore won Florida.” (You get a similar pattern for ‘regret’ – I think DeRose mentions this somewhere.)”
    There’s a difference here between “knows” and “happy”. Matt says he wouldn’t say, “I guess it turns out I wasn’t happy Gore won Florida” and that seems right. Will he think his earlier claim was false? Don’t know. Will he think that it was true? That doesn’t seem right. If I say, “I’m happy because Gore won FL” but then later discover that Gore didn’t win FL, I can say that “I was happy because I believed Gore won FL” and that corrects a mistake I made earlier, it’s factive, and everything is fine.
    Happy, here, seems to work like “regret”, “shocked”, “despaired”, but I thought these were pretty typical cases of factive terms. So, I wonder if there’s a reliable test for factivity that would allow us to say that some terms that I’m tempted to label as factive aren’t but “knows” is. Could it be that the tests we’re using aren’t identifying necessary features of factive terms? That seems like a possibility.
    Alright, have to get to work. I’ll post if I can think of something. Thanks again, both of you.
    Mark,
    Sorry I botched your view in the original comment. So, I think on normative reasons, we’re not really going to fight about much. My main concern is that there’s a truth (a true proposition, a fact, or a state of affairs) that is required for our ascriptions of normative reasons. One worry is that this leads to an implausible view about motivating reasons, but I think we disagree on a lot but we agree that that worry isn’t that worrisome. (My way of responding to the worry may well be more problematic than the problem it addresses.)
    Have a safe trip, see you tomorrow.
    Best,
    Clayton

26. Just a quick follow up. I don’t know if I’ve done a worse job during q&a at a talk, but I hope I haven’t and won’t repeat that sort of performance again. Part of it was that I was just tired and couldn’t track the questions, part of it was that the questions seemed to be coming from all over the map (people seem to have very different views of what motivating reasons are supposed to be), part of it was that the positions available are complicated, but part of it is that I’m still not entirely sure what the parties to the disagreement want. That’s the biggest part of the problem I face now.
    Just to recap what seemed to be going on during q&a.
    On Dancy’s view:
    Good case:
    Mustard is running down the hall. Mustard knows that a murderer is chasing him.
    (1) Mustard’s reason for running is that the murderer is chasing him.
    Bad case:
    Mustard is in the same mental states but there’s no murderer.
    (2) Mustard’s reason for running is that the murderer is chasing him.
    My worry: (1) and (2) seem to entail that there’s a murderer to run from, but the bad case is a case without a murderer. I can’t make the appearance of entailment go away. (Thanks, btw, to those who had my back and seemed to agree that you can’t make the appearance of entailment go away because there’s an entailment.)
    Critic’s response: no it doesn’t. (2) can be true even if there’s no murderer.
    My worry: But (1) and (2) entail:
    (3) Mustard ran down the hall because the murderer was chasing him.
    Critic’s response: No, that entailment doesn’t hold. (3) is factive, but (1) and (2) aren’t.
    Me: The entailment does hold. I understand not wanting to say that the entailment holds because (3) looks factive and you don’t want to say that (1) and (2) are, but there’s an entailment there. [This part of the discussion isn’t getting us anywhere.]
    Critic’s response: Look, look, look, we don’t need to know the facts to know why someone acted. That’s what motivating reasons are for. But that’s what you’re denying. This is madness.
    Me: Hmmmmm. Part of me just wants to agree with this and give psychologized redescriptions of all the “reason for which” claims, but we’re trying to avoid this as long as we can. At any rate, on your view, (3) isn’t a consequence of (1) and (2). On your view, there will be a bunch of ‘because’ statements that are true of the bad case and the good case. (e.g., “Mustard ran because he thought there was a murderer after him” comes out true on everyone’s view and is true regardless of whether we’re describing the good case or the bad.) Whatever you want to plug in for ‘Mustard ran down the hall because X’ will be, by your lights, a satisfactory explanation of the agent’s action that tell you why the agent ran. These ‘because’ statements, by your lights, do not specify the reasons for which the agent ran. Once you know all the relevant ‘because’ statements and we agree that they are true, you then demand that I supply additional ‘reason for which’ statements. I don’t know what you want from me. You want some information that’s not contained in the ‘because’ statements _and_ whatever it is, it had better not be information that entails that the agent’s beliefs are correct. What’s that information? I have no idea. I think you’re asking the impossible.
    The following seems like a plausible view. In the bad case, we can’t say what reason the agent had to do what she did, we can’t say (truthfully) ‘S A’d for the reason that p’. However, we can say this in the good case. All the true ‘because’ statements we use to explain action in the bad case are true in the good case. Because we can agree that they are true, we should agree that we’ve given the reason why the agent did what she did. (Again, part of me thinks that once you have this, you have the motivating reasons that you’re looking for.)
    I can’t tell whether the problem with my view is the assumption that Dancy’s treatment of error cases is problematic or with the idea that I’m not able to say both that acting for a reason is a matter of successfully responding to the situation and explain the agent’s behavior in the bad case. If it’s the first thing, I should note that I’m not the only one who doesn’t like to say things like ‘He did it for the reason that p, but ~p’. If it’s the second thing, I think I have a story to tell about why agent’s do what they do in the bad case. It’s the very same psychological story that my opponents accept.

27. Hi Clayton. Isn’t (3) ambiguous between a causal and a non-causal reading? The causal reading says something like “That the murderer was chasing him caused Mustard to run down the hall.” It’s only the causal reading that seems factive to me. But the causal reading isn’t entailed by (2). It is entailed by (2*):
    (2*) The reason that Mustard ran down the hall is that the murderer was chasing him.
    Actually, (2*) also seems ambiguous to me between a causal reading and a non-causal one. The non-causal one is basically just (2). The causal reading of (2*) entails the causal reading of (3). And if you don’t think there is a non-causal reading of (3), then I guess I’ll just say that only the causal reading of (2*) entails (3).

28. Hey Jeremy,
    That’s an interesting suggestion. I don’t have any problem with a non-causal because, but I’m not entirely sure why the non-causal because would be non-factive. “9 is divisible by 3 because 9 is 3 squared” seems to be both non-causal and factive. “It was a good idea for him to run down the hall because the murderer was chasing him with an axe” seems to me to be both non-casual and factive. (Of course, this is consistent with the claim that there are other cases where the non-causal “because” statement isn’t factive in the explanans position.)
    I think the options are multiplying like bunnies. I have the motivating reason ascriptions aren’t factive and don’t entail “because” claims crowd and the motivating reason ascriptions aren’t factive and do entail “because” claims crowd. I wish you non-factive types weren’t getting along so well. If you’d fight amongst yourselves, it would make my life easier.
    Does the following strike you as odd?
    “Look, we both know that there was no one else there that night in the house with him. But, you asked me why he was running like that and I’ve told you. He was running like that because there was a murderer chasing him with an axe.”
    It strikes me as odd and some reason to worry about the hypothesis that there are non-factive and non-causal “because” statements where the explanandum is some proposition about an action.

29. You could say we “non-factives types” disagree among ourselves. You could also say that there are several different ways to block your argument!
    I think most non-factive types will agree that it is at best very odd to say “A did it because p” when we know that p is false. Agreed! I think most non-factive types will also insist that if you add a parenthetical or side comment within the sentence it becomes perfectly ok:
    “He was running because, as he thought at the time, a murderer was chasing him with an axe.”
    At least I’d say this was perfectly fine. I’d say that a similar phenomenon is found all over the place, as in:
    “I was happy that, as I thought at the time, Gore won Florida.”
    “My Republican friend was disappointed that, as we all thought at the time, Gore lost Florida.”
    These are perfectly ok but if I took out the parenthetical they would not be ok. (Ok-ness is not being understood to imply truth, obviously.)
    Next question: what to say about the *truth-values* of the original unqualified statements and of the qualified counterparts? We can ask this question about ‘S’s reason for A-ing is that p’ and about ‘S Aed because p’, when the latter is meant to be giving a motivating reason.
    Clayton, you seem not to be sure of my view of the truth-values here. It’s this: the qualified statements are true, and so are the unqualified ones.
    What’s the argument? Well, it’s pretty uncontroversial that the qualified versions — with the side comment included — are true. I’d further say in these cases the reason being specified is not a fact about the person’s mental states. It is specified by ‘p’. The reason is the false proposition. (All this would have to be argued for at length. One premise I’d lean on is this: if p is one’s motivating reason for A-ing, then a belief that p played an important causal role in bringing about one’s A-ing. If this premise is true, then one could argue that truths about one’s mental states — e.g., “I thought that such and such” — are not going to be motivating reasons in the typical “bad” cases, because beliefs *about* one’s thoughts in these cases typically don’t play the required causal role. Empirical premises come in here.) Suppose, then, that the qualified statements are true, and that in these statements, the reason which is specified is a false proposition. Then I’d say this is a good ground to think that, despite their oddness, the unqualified statements are true, too.
    We non-factives have to explain why, despite the unqualified statement is true, it is unassertable. We have that burden. Maybe this is our doom. However, I think there is some encouragement to be drawn from the consideration of similar phenomena (‘happy that’, ‘regrets’, the use of epithets, and etc.), which contrast with the behavior of the clearly factive ‘knows’.

30. Hey Matt,
    There may be several ways to block the argument, but there is disagreement within the non-factive camp as to whether there’s an entailment to ‘because’ claims. I think it’s worth working it out whether there’s such an entailment.
    I’m not sure what to say about the cases involving “happy”. A friend pointed out that Swanson had examples of non-factive factives, cases where you could felicitously say that you were, say, upset/happy/disappointed that p only to later discover that ~p. But, he then pointed out that this just points to a contrast with the ‘because’ case. There’s no felicitous cancellation with “because” but there is with these other expressions.
    You say that it’s uncontroversial that the qualified version of the “because” claims are true, but I’m not sure what you mean by that. I’m not alone in thinking that the “because” claims are factive and they remain firmly factive with the parenthetical remarks added. From our conversations, that seemed to be Dancy’s view. (That’s why he thinks the claims about reasons for which don’t entail the “because” statements I think they do.) Hornsby has a nice discussion of this in her paper, “Knowledge in Action”:
    “We know that there is a perfectly good factive explanation for the case where what the agent supposed was false: He Phi’d because he believed that p”. And it a very strange idea that explanations are ever non-factive. To many ears, “He Phi’d because, as he supposed, p” is true only if p. (One plausible account of “as x supposes” used parenthetically within a sentence s will treat it [as a sentence adverb such as “luckily” should arguably be treated] as conveying something about what is said in s without affecting its truth-conditions. If so, then, given that “p because q” requires the truth of p and of q, introducing a parenthetic “as x supposes” within it will not produce anything non-factive.)” (292).
    Maybe she’s wrong and I’m wrong, but surely it’s wrong to say there’s no controversy here.

31. Clayton, can we agree to this? There is such a thing as the proposition that S uses when deciding whether to phi — the proposition, belief in which causes S to phi. Then there’s the thing that causes S to phi. The thing that causes S to phi is, in the standard cases, presumably a mental state (like a belief accompanied by the relevant desire). And this mental state can have as its object the proposition that S uses when deciding whether to phi. That proposition can surely be false. And the proposition itself need not be a proposition about S’s beliefs. It seems useful to reserve a term to refer to that proposition. And “S’s reason for phi-ing” seems like it works nicely. Do you have a better term for it? Or does it not need a term because it’s not interesting to figure out what proposition that is, or what epistemic relation S has to bear to that proposition for it to be usable by S?
    When you talk about “the explanation for S’s phi-ing” or “the reason for which S phi’d” or “the thing because of which S phi’d” I think you can pick out either of these two: the proposition S used when deciding whether to (or, if you prefer, the proposition that was the object of the mental state that caused S to phi) or something standing in an appropriate causal relation to S’s phi-ing. So, your expressions seem ambiguous to me between a non-factive and a factive sense, the latter of which is causal.

32. Clayton, there are theories about why the unqualified statements are true which are consistent with a factive account. Here’s one. The qualified statements are true because they specify a *fact about a belief* as the subject’s reason.
    Or in the case of ‘because’, one might think that:
    ‘S Aed because, as he thought at the time, p’
    is true in the bad case because it asserts that S had as his motivating reason that *he thought that p*, which is a fact.
    So, I don’t see how the claim that the qualified versions are true rules out the factive view. I said in my previous comment that one needs to *argue* from the premise that they’re true to the falsity of the factive view.
    I don’t know the Swanson stuff you’re referring to, but I do feel a marked difference between these:
    Set-up: Bob didn’t have a heart-attack, but I thought he did, and I thought he died because of a heart attack. He did die. Later, I learn he died of some other cause. Consider:
    “Bob died because, as I thought at the time, he had a heart attack.”
    I think that the only way to make this statement come out true is to someone think of “as I thought at the time” as governing the whole because statement. If we read it as a side-comment just on ‘he had a heart attack’, it seems to me false.
    Set-up 2. Bob thinks a murderer is chasing him with an axe. Bob runs like hell. There is no murderer.
    Consider:
    Bob was running because, as he thought at the time, he was being chased by an axe murderer.
    or:
    Bob was running because he was being chased by an axe murderer — or so he thought at the time.
    These seem far better to me than in the heart-attack case. I’m just reporting how things seem to me….
    I said it’s “pretty uncontroversial” that that the qualified versions are true of the bad cases — is that too strong? Maybe someone would say that side-comment qualifications like ‘He did it because, as he thought at the time, p’ are false but that qualifications like ‘He did it because he thought at the time that p’ are true. That’s slicing things in an odd way, but ok. It sounds like Hornsby wants to go that way.
    I agree with Jeremy (big surprise). Jeremy picked out a relation — call it R — between persons and propositions that is relevant to the normative properties of action, agrees in all the good cases with a factive account of motivating reasons like yours, but doesn’t require the relevant proposition to be true. Regardless of one’s stance on the semantics of “S’s reason for doing A is that p” or “S did A because p”, shouldn’t one want to acknowledge this relation R and refer to it in giving an account of the normative properties of actions?

33. Just to be clear about the heart-attack case —
    “Bob died because, as I thought at the time, he had a heart attack.”
    The only true reading I can get for this (when he died due to some other cause) is one in which “as I thought at the time” comments on the because claim, as in:
    “Bob died, and I thought at the time that this was because he had a heart attack.”
    But in the second set-up, the reading involving a side-comment on ‘he was being chased by an axe murderer’ — and not on the because claim — seems true to me.

34. Hey Jeremy and Matt,
    Apologies for not getting back earlier, I’m getting packed for Chicago and trying to get some thoughts in order for the APA. My quick response to Jeremy is that even if the reason is a proposition rather than a fact, the reason ascription could be factive. The facts could be conditions under which the ascription is true just as you’d say (I think) that there are mental conditions that must obtain if the proposition in question is going to be the reason. So, if there’s evidence that the ascriptions are factive, that’s consistent with saying that the reason ascribed is a proposition. (Knowledge ascriptions are factive even if the object of the belief that constitutes knowledge is a proposition.)
    Matt,
    I’m not sure I’m tracking the example:
    “Set-up: Bob didn’t have a heart-attack, but I thought he did, and I thought he died because of a heart attack. He did die. Later, I learn he died of some other cause. Consider:
    “Bob died because, as I thought at the time, he had a heart attack.”
    I think that the only way to make this statement come out true is to someone think of “as I thought at the time” as governing the whole because statement. If we read it as a side-comment just on ‘he had a heart attack’, it seems to me false.”
    I can’t read this sentence in such a way that it could be true if Bob didn’t die because of a heart attack. I think you might disagree?
    You wrote:
    “Consider:
    Bob was running because, as he thought at the time, he was being chased by an axe murderer.
    or:
    Bob was running because he was being chased by an axe murderer — or so he thought at the time.”
    Again, I hear the first as factive. I can’t hear this as the sort of thing that could be true if there’s no axe wielding murderer.
    I don’t know if the second sentence was offered as evidence for the non-factivity of the first. I like “Bob was running because he was being chased by an axe murderer — or so he thought at the time”. However, I also like:
    (i) She knew that there was beer in the fridge — or so she thought at the time.
    (ii) There was beer in the fridge — or so I thought at the time.

35. “Regardless of one’s stance on the semantics of “S’s reason for doing A is that p” or “S did A because p”, shouldn’t one want to acknowledge this relation R and refer to it in giving an account of the normative properties of actions?”
    Yes, but I think I can do that. Such a relation is a necessary condition for the truth of various claims we’d offer to explain actions. Such a relation is part of what we’d focus on when evaluating the normative status of an action. Why can’t I agree with you on those points?

36. Clayton,
    I’m sorry to keep dragging on about this. But my argument wasn’t this: reasons are propositions and propositions can be false, so reason-ascriptions aren’t factive. The key premise is, rather, this: it’s an interesting question what epistemic relation S must bear to p for S to be legitimately caused to phi by a belief that p. For example, for S to be legitimately caused to phi by a belief that p, must p be true? I think not. I think that you think so. But I don’t see how that issue will be resolved by bickering over the semantics of “p is S’s reason”. If the answer to the former question is “no”, then I think there’s a perfectly good sense of “p is S’s reason” according to which p can be false. But the interesting question is the former one.

37. Hey Jeremy,
    I didn’t think you were using that argument, I just want to be clear that I don’t think that the claim that reason ascriptions are factive commits you to any particular claim about the ontology of the reasons. (This is a slip I made earlier when I tried to respond to Mark and said something I shouldn’t have.)
    At any rate, bickering over semantics is important! As a friend once said, people used to be burned at the stake because of semantic matters! Not ideal, I’ll grant.
    I think I can say lots of things given the ontology and semantics that I’m tempted towards about “legitimacy” that you’d want me to say. If “legitimate” is understood as something like “rational” or “reasonable”, I’d say (as I think you would) that the legitimacy of certain moves doesn’t depend upon what the facts are but what they seemed to be. If “legitimate” is understood as “permissible”, then I think we’d have a substantive disagreement. I don’t think it’s permissible to treat things that aren’t reasons as reasons for no reason. So, in one sense of legitimate, you’re right that I’d defend the view that it’s not legitimate to treat _that this is filled with gin_ as a reason to give the drink to a gin-desiring friend when it’s not gin but petrol.
    I might be missing this. At any rate, I appreciate that you are “dragging on about this”, I get a lot out of it. Off to Chicago!

38. Just a couple clarifications and then I’m done. I promise. Thanks, though, for the extended and interesting exchange!
    When I know Bob died due to some other condition, I won’t find
    (1) He died because he had a heart attack, or so I thought at the time.
    to be true unless I strain to hear it as:
    (2) He died and I thought at the time that this was because he had a heart attack.
    In (2) the “because” claim is embedded under “I thought.” What I was claiming is that in the second example I discussed, involving the person who thought he was being chased by an axe murderer, the claim
    (3) He was running because an axe murderer was chasing him, or so he thought at the time
    seems true to me not only when we take “he thought at the time” to govern the “because” but also if we took it to comment only on what follows the “because.”
    In the examples you gave such as:
    “She knew there was beer in the fridge, or so she thought at the time”
    I can’t sense a true reading in which “she thought at the time” comments only on “there is beer in the fridge”.
    But maybe we should forget for the moment about whether the qualified statements are true or not, or even whether they seem true. Just think about whether people say such things. When I think about election night 2000, I think I would say, if I were explaining the events to my kids, “I was happy because, as we all thought at the time, Gore had just won Florida.” The same for “I breathed a sigh of relief because, as we all thought at the time, Gore won Florida.” I don’t keep a journal, but I feel I say these sorts of things alot. Try: “I stopped checking Pea Soup because Clayton and I had wrapped up the exchange, or at least that’s what I thought!”
    Here’s a few things google searches turned up:
    “one evening when it was getting dark, I packed my sweater in a plastic bag because, as I thought, it was windy in the lake and the water was also cold.”
    [in the story it turns out the water had all dried up and all was fine]
    “I didn‘t really „trust“ my pregnany until I was way into the second trimester. Because then I was safe – or so I thought back then (ah, blissful ignorance).”
    “Right then and there, I started worrying about a sub attack as we couldn’t possibly have gotten up six decks to the life boats in case of emergency – or so I thought at the time”
    [I read this as specifying a motivating reason. As it happens, the sub was still at dock. So they could have gotten up six decks.]
    Ok, off to the APA. See you soon, Clayton.