At first sight, there seems to be a blatant contradiction between Bernard Williams’s two central theories: his view about thick concepts and his existence-internalism about reasons (many others have hinted at this too – Gibbard and Scanlon, for instance). I want to quickly sketch what that tension is. I mainly wanted to know if anyone’s seen or can come up with an interpretation of Williams's theories that dissolves the contradiction. I’ll suggest a couple, but I’m not very happy with them.
Reasons first. According to Williams’s internalism, A has a reason to phi only if A could reach the conclusion to phi by a sound deliberative route from the motivations he already has. The subjective motivational set should be understood broadly, and the sound deliberative route means correcting beliefs, making motivations more coherent, using imagination, and the like.
Here’s what Williams says about the famous husband-wife case: “Suppose, for instance, I think someone (I use ‘ought’ in an unspecified sense) ought to be nicer to his wife. I say, ‘You have a reason to be nicer to her’. He says, ‘What reason?’ I say, ‘Because she is your wife.’ He says – and he is a very hard case – ‘I don’t care. Don’t you understand? I really do not care.’ I try various things on him, and try to involve him in this business; and I find that he really is a hard case: there is nothing in his motivational set that gives him a reason to be nicer to his wife as things are. There are many things I can say about or to this man: that he is ungrateful, inconsiderate, hard, sexist, nasty, selfish, brutal, and many other disadvantageous things.”
So, the theory of reasons does not ascribe any reason to this man to be nice to his wife. However, all the thick ethical concepts still apply to him. And, here is what Williams says about thick concepts: “…there are enough left in our own: coward, lie, brutality, gratitude, and so forth. They are characteristically related to reasons for action. If a concept of this kind applies, this often provides someone with a reason for action, though that reason need not be decisive reason and may be outweighed by other reasons… Of course, exactly what reason for action is provided, and to whom, depends on the situation, in ways that may well be governed by this and by other ethical concepts, but some general connection with action is clear enough. We may say, summarily, that such concepts are ‘action-guiding.’”
It thus looks like, whilst the internal reasons theory states that the husband does not have a reason to be nicer, the thick concepts account does ascribe such a reason to him. This is because his act is brutal and ‘brutality’ is one of the action-guiding concepts that are characteristically related to reasons for actions. If a concept of this kind applies, this often provides someone with a reason for action… So, according to one theory, the husband does not have a reason to be nice, and according to the other he has.
There’s also another inconsistency. According to the first view, only elements in his motivational set provide reasons, whereas according to the second, that a thick concept applies provides reasons. These things (what’s in a motivational set and that a concept applies) seem to belong to different categories.
So, the question is, is there any way to understand both of these theories together consistently?
First proposal: we could emphasise the word ‘often’ in Williams’s account of thick concepts. Their applicability only often provides reasons, but perhaps not always – maybe not in the cases in which the agent to whom the concept applies does not have the suitable elements in his motivational set to have reasons. I worry about this view for two reasons: first, it seriously undermines the ‘action-guidingness’ of thick concepts. If actions are guided by reasons, then, according to this reading, thick concepts seem to be guiding only very specific set of agents with suitable motivations. In fact, it could be that, if no one had the suitable motivations, thick concepts would still apply to them, but no one would have reasons to act accordingly. Thus, no actions would be guided by the thick concepts. This means to make the relation between thick concepts and reasons too contingent. Second, it does not seem to get rid of the second inconsistency related to what provides reasons: elements in motivational set or the applicability of thick concepts.
Second proposal: When Williams talks about reasons in the thick concepts passage, he is talking about external reasons. On this view, it is part of the content of thick-concept claims that the agent to whom the concept is applied has an external reason to act, or it is a presupposition that he has such a reason. Williams’s work on reasons then shows that, all external reason statements are either false (there are no external reasons) or meaningless. This would mean that he has revealed that thick-concept statements are also either all false or meaningless (given that part of their content is false or meaningless, or their presuppositions are).
This interpretation would fit Williams’s Socratic conclusion ‘that, in ethics, reflection can destroy knowledge.’ It only now turns out that Williams underwent that destructive reflection himself. The two views would now be theoretically consistent however. The thick concept theory would only hold that, when we apply thick concepts, we assign external reasons to the husband (as Williams does himself when he uses these concepts), whilst the internal reasons theory would only deny that he does not have internal reasons. Of course, the internal reasons view would also deny that the agent has external reasons but this would make it no less true that we assign them. Not completely happy with this solution, so I would be very interested to hear other takes on this.
Hi Jussi, my suggestion is as follows (perhaps a variant of your first proposal): when a thick concept applies it supplies a reason to an agent *if it is that agent’s concept*. What it is for a thick concept to be your concept is (in part) for your motivational set to be set up so that there is a route to your being appropriately motivated by that concept applying. What we get then is a more precise and well-motivated view as to when the application of a thick concept supplies a reason (filling out the ‘often’). It allows us to say that the reasons are present in all the restricted range of cases without being external reasons.
This also fits with the idea that reflection destroys knowledge. By reflecting, we lose our thick concepts, and in so doing we lose our reasons.
By the way, I’m not sure that Williams thinks that all external reasons claims are false or meaningless, because I think he wants to make an exception for moral reasons. Wish I could remember what I read that suggested that though…
Thanks Daniel. That sounds like a fairly plausible reading and close to what I had in mind but better. That might require even more than my reading for the application to provide a reason. Not only does the agent have to have suitable motivation but also the specific concept in question that applies to him. It does avoid the contradiction but maybe now Williams’s claims that the applicability of these concepts provides characteristically reasons should be highly qualified – only if both the speaker and the agent share the same concepts. This would mean that when we apply these terms to a community (Smith’s Australians…) that do not share our concepts our thick concepts talk doesn’t even begin to ascribe reasons to them. I guess the worry here is that the action-guidingness of these concepts is left fairly weak. Of course, that’s just a substantial disagreement.
The last point you make is interesting. If you can find the passage, let me know. Of course, the thick concept reasons probably would often be moral reasons so that would take the tension away too. Yet, it would ill seem some of the statements in the reasons papers.
Hello Jussi and Daniel.
Daniel’s suggestion seems along the right lines, but can’t be quite right. I take it that a person cannot have a thick ethical concept unless that person in fact sees the point of the concept in a way which motivates them. If so, it would not be enough for the person’s motivational set to be set up in a way that they could come to see its point. For such a person may not yet see its point.
On this reading a thick ethical concept can provide reasons for some people who don’t possess the concept – it does so for people who don’t possess the concept but could come to do so. A person could come to do so in the appropriate sense if there is the right kind of route from the agents motivational set to the concept.
And that also suggests a way in which an agent who has a concept can persuade another agent who does not have it to acquire it by appealing to her rationality. A person who has the concept can persuade an person who does not to acquire it if the second has the appropriate motivational set. The first can show her that she has a reason to acquire to concept and be guided by it.
Jussi, I’m not convinced that the speaker needing to have the concept counts as a restriction. If the speaker doesn’t have the concept, then the question can’t even arise for her of whether the application of the concept is reason-giving.
I agree that the action-guidingness of thick concepts ends up being weak, but this seems required not just by the idea that reflection destroys knowledge but also by the related truth in relativism stuff. It seems important to keep separate Williams’ weak view of the action-guidingness of thick concepts from the stronger views of some Wittgensteinians who say superficially similar things about thick concepts and reasons.
I agree that the question cannot arise *for her* whether the application of the concept is reason-giving. I was thinking about this from the third-person perspective though – from our view to the husband who is mistreating his wife. In that case, it seems like a further condition. We might find a motive from his motivational set but not the concept in him which we want to apply to him. On your view, even in this case the application of the concept would not provide him a reason (we are talking here about normative reasons and not motivating reasons).
In any case, maybe you are right and action-guidingness is always first-person perspectival for Williams which does seem to require the possession of the concept.
Hi Victor, thanks for the correction, I agree with what you say. My suggestion was too simple in exactly the way you explain.
One caveat – it seems quite tricky for Williams to hold that we can *rationally* persuade agents who don’t share our thick concepts to acquire them, because of the reflection destroying knowledge point. I interpret Williams as holding that if we reflected, we would lose all our thick concepts and be left with only the thin ones (see p.148 of ELP, though maybe it’s too crazy a view to attribute charitably). If Williams is wrong about reflection destroying knowledge, as I believe, then such rational persuasion seems fine. Even on Williams’ view, it looks like there can be *non-rational* persuasion, but then we would have to wonder whether it was the *right* kind of route to motivation (given the internal reasons view). It’s possible then that *for Williams* the set of thick concepts an agent possesses is no narrower than the set such that the relevant route to motivation exists, even though I agree that the former category really is narrower.
Thanks Victor and Daniel. This is helpful. I guess one question that arises is whether the antecedent motivations are a necessary condition for being able to come to see the point of the new thick concept as Victor suggests, or whether seeing the point of the concept could come first in conversion and the motivations follow in suite as McDowell suggests. If the latter were a possibility, then the thick concepts would apply to the husband, the applicability provide reasons for the husband, and this would still contradict the internal reasons theory.
Hi Jussi,
I think this is only a problem for Williams given one or both of these assumptions:
1. If a concept applies to something, the thing it applies to must itself have the concept.
2. If a thick concept’s applicability provides reasons, it must provide reasons to the thing it applies to.
But both are pretty clearly false. (Quarks don’t know physics; beautiful and cozy buildings don’t, sadly, have reason to maintain themselves.)
So the fact that you or I could say all sorts of disadvantageous things about the husband doesn’t entail that he be able to say them about himself (or at any rate not in the sincere and natural way we could); similarly, the fact that these concepts apply to the husband from our perspective doesn’t imply that their applicability provides reasons for him–Williams just needs this to provide reasons for us, which of course it does. (In sincerely saying these things about him, we evince motivations to respond to him in all kinds of disadvantageous ways.)
Hi Jussi,
I don’t think that Williams is inconsistent in either of the ways you suggest.
(1) As you point out yourself, he’s careful to qualify the connection between thick concepts and reasons, twice: “characteristic” and “often”. So all he needs to say to be consistent here is that people often/characteristically have the motivational set to make these concepts reason-giving. You object that this makes the action-guidingness of thick concepts too contingent, but Williams can deny this without inconsistency (or, I think, implausibility).
(2) Your second charge of inconsistency turns on the claim that Williams thinks that (internal) reasons are items in one’s motivational set. I think that’s false; his view is that reasons DEPEND on items in one’s motivational set. If you have a desire for gin in your motivational set, then the fact that the stuff in the bottle is gin is a reason to drink from it, in virtue of your desire.
Finally, I don’t think Daniel can be right about Williams exempting moral reasons. His skepticism about external reasons is aimed precisely at moral reasons. One proviso: in ‘Internal Reasons and the Obscurity of Blame’, he allows that reasons claims may make themselves true (i.e. bootstrap themselves), through the blame that they imply.
Benjamin,
I wanted to take an issue with this:
“similarly, the fact that these concepts apply to the husband from our perspective doesn’t imply that their applicability provides reasons for him–Williams just needs this to provide reasons for us, which of course it does. (In sincerely saying these things about him, we evince motivations to respond to him in all kinds of disadvantageous ways.)”
and especially the first part of the claim. I did think about something like this but then found it very implausible. The idea would be that, because the word ‘brutal’ applies to the actions of the agent, I have a reason to better treat my wife. This sound false to me. I don’t even have a wife.
What you say in the brackets seems one step too removed. Isn’t using these terms itself a disadventagous way of responding?
Steve,
well, there is an alternative reading of ‘characteristic’ and ‘often’. He could have also had in mind Dancyan underminers for the reasons which are ascribed. So, take a case in which the husband and wife practice karate together as a hobby they both enjoy. In that case too, we could say that the husband is treating her wife in a brutal way even if we don’t assign him a reason not to do so. So, what could be going on in terms of the text could be this sort of external underminers.
I would like to know more about how Williams could deny the overt contingency of the other readings action-guidingness of thick concepts. When I think of some of the terms on Williams’s list – inconsiderate, selfish, ungrateful – it seems to me that actually very few people have in their motivational sets things that would move them to act less in these ways. So, when we apply these terms, there’d rarely be reasons for people to act less in these ways.
What I find implausible about this picture, is that our use of the thick concepts would be neutral about whether others have reasons – this would be all down to luck. But, then it seems like these concepts are not much of a criticism of the husband’s actions. This is why I find the external reasons as part of the content more plausible. I don’t yet see anything in the text that contradicts this reading either.
About the second point, well, he says that ‘there is nothing in his motivational set that gives him a reason to be nicer to his wife as things are’. This suggests that, if the motivations were different, then the husband would have something in his motivational set that would give him reasons. It often seems that Williams is a psychologist about reasons: consider this statement on page 103 of Internal and External reasons:
A member of S, D, will not give A a reason for phying if either the existence of D is dependent on false belief, or A’s belief in the relevance of phying to the satisfaction of D is false.
Here it really sounds like desires give reasons and not only that reasons depend on desires. This fits his idea that reasons must be able to explain actions. But, I agree that Williams is not consistent on this.
Jussi,
Why think the action-guidingness of “brutal” (even when applied to cruel husbands) must always involve the treatment of wives, in particular? Why can’t I just have reasons to do things like distrust the husband, avoid helping him, maybe help the wife in certain ways, etc.?
That’s a good point. Of course, I don’t want to deny that you have reasons to do those things when the husband mistreats his wife.
Maybe I don’t like the idea that the reasons in question are our – the speakers’ – reasons because, as Williams says, we say it *to* the husband that he is being brutal. So, we are trying to guide *his* actions and not our own. Of course, in the quote, Williams says that the applicability gives *someone* reasons. Maybe he did really have in mind that it is the speakers who are given reasons by the applicability of the terms. But, for some reason, that doesn’t fit what intuitively is going on in the situation.
If anyone’s more interested in the topic, the grapevine has just send me a paper by Ulrike Heuer called ‘Thick Concepts and Internal Reasons’ in which she argues in much more length and detail that these two Williams’s views are in tension. This paper is forthcoming in a collection of papers on Williams, but I’m sure she will be more than happy to send the paper to anyone who’s curious.
Jussi,
Saying that desires give an agent reasons isn’t the same as saying that desires are reasons. (If I give you a gift, it doesn’t follow that I am a gift!) In the first part of IER, Williams is careful to avoid what he calls “specific statements” of reasons. Unfortunately, Williams seems only to provide specific statements of reasons when he is talking about alleged external reasons, so it is hard to find direct evidence either way. But I don’t know of any good evidence for your less charitable interpretation.
Very few people have motivational dispositions that would be triggered by ‘inconsiderate’, ‘selfish’, and ‘ungrateful’? I strongly disagree. Who among us isn’t averse to being thought of in those terms? (Answer to rhetorical question: only a few!) The view that you find implausible here is the view I think is correct. There are people (e.g. Randians) who are indifferent about(or even proud of) being selfish, for example. I myself am totally unconcerned about being impious and blasphemous. The Muggletonians were proud of being irrational…
Hi Steve,
I admit that ‘gives a reason’ is ambiguous. Sometimes it is natural to read as an identity thesis. The fact that someone is in pain gives me a reason to help them. It is hard to read this in any other way that the fact that he is in pain is a reason to help them. I know that when desires give a reason this is sometimes understood in a different way. But, it is hard to read that different sense of ‘give’ to other contexts. But, I agree – he doesn’t seem to commit either way (maybe the act explanation stuff is where the commitment might be).
I give you selfish. Ungrateful might be a different story. The issue though is that we’d have to do some sort of empirical investigation to solve the contingent question of how many people have these reasons. Ulrike’s paper is very good on this point.
I suggest that the reason that it is hard to read “that someone is in pain gives me a reason to help them” otherwise than as an identity claim is: charity. We already know that that fact is a reason to help them. But then this same reason doesn’t apply in the case of “desires give us reasons to act”. Or rather, it applies in the opposite direction.
I offer an interpretation of Williams that reconciles the act explanation stuff with a nonpsychologistic account of reasons in a 2009 paper in Philosophers’ Imprint, “The Obscurity of Internal Reasons”.
When you say that “the issue” is that empirical investigation is needed, you make it sound as if your objection to Williams is that he failed to conduct any surveys in order to support his empirical claims. But your original objection (made without empirical support!) was that Williams’ account is implausible because it makes the action-guidingness of these concepts too contingent.
Thanks Steve. I’ll have a look at that paper. I don’t think you are being very charitable to me or Ulrike. The problem is not that Williams failed to do empirical investigation – seriously think I would charge him of that?
The original objection was that the reading of Williams that is being discussed here makes the action-guidingness a contingent matter. This is a modal question – is that there are reasons entailed but the applicability of the thick concepts (necessity) or does the applicability of these concepts leave the question of reasons open (contingent). I take it that modal question are and can be made a priori without empirical support. If we go for the latter, then it becomes an empirical question who has reasons – and it is this that seems implausible.
I find it more plausible that the speakers who use thick concepts assign external reasons to the agents whose actions are cruel or brutal. It isn’t open for them whether those agents have reasons not to do these acts – in the sense that that is not a possibility they’d need to consider (with the exception of Dancyan worries). The fact that the right kind of motivations happen to be common in our community doesn’t really address this modal question. And, so far, I haven’t seen anything in Williams’ text that contradicts this reading either. So, I’m too trying to read him charitably.
ah, sorry, I misunderstood your point when you wrote “the issue..” Not trying to be uncharitable! I agree that the use of thick concepts implies external reasons, and unlike Williams I accept the existence of external reasons (they just don’t matter…)
no worries. Loved your paper though! Brilliant.
Thanks! Did you really read it already? Wow!