Gibbard is Not an Expressivist

This may be old news to some, but new to me. It’s true. Evidence to follow.

First, background. I agreed to do a grad readings course on normativity for a student, and, long story short, there’s now 8 in the group. And we’ve started with Wedgwood’s The Nature of Normativity. So some of the below is a bit second-hand.

So, the evidence.

Let expressivism be the view that “the meaning of normative statements, and of the sentences that are used to make those statements, is given in terms of the type of mental state that the statements made by uttering those sentences express” (Wedgwood, p. 35). As Ralph notes, a rival theory is the truth-conditional view. The rival view doesn’t need to be denied, because it can be endorsed as correct in some non-fundamental fashion. What is required for expressivism is that the “fundamental explanation of the [normative] term’s meaning … should be given in strictly psychologistic terms…” (Wedgwood, p. 37).

One fundamental problem for such a view is the Frege-Geach problem of what to do with normative claims embedded in conditionals and in suppositional reasoning more generally. Gibbard claims to have a solution to this problem in terms of the actions and other mental plans that a given normative judgement “disagrees with”.

Gibbard’s solution to the F/G problem employs the notions of a plan, including contingency plans and hypothetical plans, as well as the notion of the mental state of disagreeing with, rejecting, or ruling out, an action or mental state. All this culminates with the notion of a hyperplan, and the theory is sophisticated and powerful. Very interesting.

But note: “According to Gibbard, the content of a normative judgment is determined by the actions or other mental states that it disagrees with.” (Wedgwood, p. 45). First, this notion of disagreeing with is not a psychological notion: people disagree with things, in virtue of attitudes of disagreement, but attitudes don’t disagree with each other in any sense required for being a psychological state. They bear a number of types of negative relations to other attitudes, but disagreement isn’t one of them, if disagreement is a pyschological state. Attitudes can conflict with each other, be logically inconsistent with each other, be epistemically in tension with each other, etc. But any claim about attitudes disagreeing with each other, charitably interpreted, can only come out true when it is replaced with something that is not, at bottom, a psychological state.

There is such a thing as an attitude of disagreement. Some people tell me I display it regularly. But my attitude of endorsing that 2+1=3 is not an attitude of disagreeing with the claim that 2+1=4. I do take the attitude of disagreeing with the latter claim, but that attitude is a different attitude than the former and is logically independent of it: neither requires the other. Ditto for normative attitudes, however they are understood: the attitude expressed by saying “we ought to do more to prevent hunger” is not the attitude of disagreeing with the attitude expressed by “we needn’t be concerned very much about the plight of the poor.” I have the former attitude and I also have the attitude of disagreeing with the attitude expressed by the latter. But thye are logically distinct.

So, if we want some understanding of a notion of disagreement so that the attitude expressed by “we ought to do more for the hungry” disagrees with the attitude expressed by “we needn’t care much about the plight of others”, the understanding in question will be of something that is not itself an attitude. In fact, it appears to be somewhere in the neighborhood of a transmission principle that generates warrant for taking an attitude of disagreement. The presence of such warrant, however, implies no actual attitude of disagreement.

One might think that resorting to talk of types of attitudes, rather than tokens, can save the claim that such a view is a version of expressivism. I don’t think that will help. We individuate types in terms of their contents, and an attitude type. To give a theory that refers to types, then, requires an independent account of content. But that is just what expressivism meant to give us in the first place, in terms of attitudes. So, I think, talk of types at this point won’t rescue the claim that Gibbard’s view is a version of expressivism. It will only make it uninformative until the appeal to content for types of attitudes is explained.

None of this is an argument against Gibbard’s view, of course. His view might still survive scrutiny–on that issue, I’m not commenting here. But he claimed to be an expressivist nearly 20 years ago at least, and it is interesting to me to note that his most recent views don’t count as that anymore.


Comments

Gibbard is Not an Expressivist — 8 Comments

  1. I would have thought that most contemporary expressivists want to give an account of the meaning of normative sentences in terms of the psychological states (“normative judgments”) that they conventionally express, but to do so contemporary expressivists try to develop an account of the content of these psychological states which is not parasitic on an antecedent account of the meaning of the sentences which express them.

    As I read him, Gibbard’s appeal to disagreement serves the latter function of an explanation of the content of the normative judgments in his theory. Depending on what’s meant by “fundamental”, he can still say that the fundamental explanation of the meaning of normative sentences is in terms of the psychological states expressed, but he cannot say that the fundamental explanation of the content of these psychological states is itself in terms of psychological states expressed.

    You seem to assume that he needs to be able to do that in order to count as an “expressivist”. But I doubt any expressivist ever wanted to do that. Most early expressivists had very little to say about the content of the psychological states they thought were expressed by normative sentences, which is what made their theories so flat footed as accounts of the meaning of normative sentences. In contrast, by appealing to disagreement, Gibbard is able to develop something much more explanatory.

    Of course, it’s certainly fair to question the plausibility of the expressivists’ “dog-legged” strategy of explaining meanings of sentences in terms of psychological states expressed and then giving a fundamentally different explanation of the content of these psychological states, especially one which deploys the notion of “disagreement” which is itself in need of elucidation. However, I think Gibbard’s use of this strategy is still recognizably “expressivist”, even if he doesn’t explain disagreement in psychologistic terms.

  2. Matthew, I’m not sure we’re on the same page. The argument above is that the notion of disagreement in question isn’t a psychological one and isn’t part of the content of the psychological states in question.

  3. A certain kind of cognitivist will agree with Gibbard that the meaning of normative sentences is to be explained in terms of the psychological state they express, but he will argue that these psychological states are beliefs. That doesn’t, however, mean that this cognitivist has to explain the contents of these beliefs in terms of some psychological notion. He could just do it in terms of truth conditions.

    I think the notion of disagreement is playing an explanatory role for Gibbard analogous to the notion of truth is for this cognitivist. Neither has to be a psychological notion in order for the explanation of the meaning of normative sentences to be psychologistic.

  4. If so, Matthew, then Wedgwood’s account of expressivism needs to be changed, since he requires that the fundamental explanation of the meaning of normative judgments be given in strictly psychological terms, including embedded as well as non-embedded contexts.

  5. Hi, Jon. I don’t see any reason to think that Gibbard is not an expressivist. I do think there is an interesting interpretive question about whether the quasi-technical notion of ‘disagreement’ that Gibbard uses in Thinking How to Live is itself a purely psychological notion or not, but on both ways of reading him, there’s no reason to doubt that the view is expressivist.

    I take expressivism to be the view that we characterize the meaning of a sentence, ‘P’, by saying what it is to think that P. This is how Gibbard himself characterizes what expressivism is. So long as his account works by assigning each sentence ‘P’ to what it is to think that P, therefore, it can count as expressivist. So what’s the challenge to that?

    You seem to say that if disagreement is not a purely psychological notion, then if Gibbard’s account appeals to disagreement, then it can’t be expressivist. But of course, even if disagreement is not a purely psychological notion, we could still use it to pick out purely psychological states. On the way I’ve been inclined to read Thinking How to Live, this is exactly how it works.

    On this view, we can think of disagreement as a relation between mental state types – the relation that holds between two mental state types M1 and M2 when A’s being in M1 and B’s being in M2 suffices for A and B to count as disagreeing with one another. And then we use the sets of hyperplans that Gibbard’s semantics associates with each sentence to characterize the corresponding state of mind as that state of mind, whatever it is, such that it disagrees with all and only the hyperplans that aren’t in that set.

    In other words, the notion of disagreement only plays a role in a sort of complex description that we give of states of mind. No problem, here. (See chapter 7 of my Noncognitivism in Ethics for more on this reading of Thinking How to Live.)

    On the other hand, there’s also a natural way of reading Gibbard (which I now believe to be his preferred way) on which he really does want to say that disagreement, rather than simply being a relation between mental state types, is itself a kind of attitude. On this view, ‘disagreement’ really does pick out a type of purely psychological attitude.

    Now, you and I may think that if A believes that P and B believes that ~P, then they count as disagreeing because they believe inconsistent things, but on this interpretation, Gibbard holds instead that they count as believing inconsistent things because one disagrees with the very thought the other accepts – so the explanation goes the other way around.

    This latter view not only gives us a wholly psychological characterization of the mental state type expressed by complex sentences, it does so without even needing to appeal to non-psychological notions as part of the description. It carries a commitment that you may believe is false, but then the interesting claim about Gibbard is that you think one of his commitments is false, not that he doesn’t really hold the view that he himself coined the term for.

  6. Mark, that’s very helpful. As I wrote to Matthew, I think the fundamental issue here is what makes a view count as expressivist in the first place. This is boring taxonomy, of course, but still maybe worth thinking about… šŸ™‚

    I actually like your take on Gibbard, much more than your second alternative that you think Gibbard actually maintains. But on the issue of your take on what counts as expressivism, I wonder if it is too weak. You write,

    I take expressivism to be the view that we characterize the meaning of a sentence, ā€˜Pā€™, by saying what it is to think that P. This is how Gibbard himself characterizes what expressivism is. So long as his account works by assigning each sentence ā€˜Pā€™ to what it is to think that P, therefore, it can count as expressivist.

    So, I’m thinking Frege here, and we can characterize his entire semantics in terms of telling us what a sentence means by talking about what it is to think that p. I know what you wrote is supposed to be a shorthand description of a much more detailed account, and so I mention Frege only to see how to sort these things better. And in particular, do you think it gets sorted in terms of Wedgwood’s account, that the “fundamental explanation…should be in strictly psychological terms”? On your two readings of Gibbard, the second one seems to honor this requirement more than the first… Yes? No?

  7. I would put it this way. There are ways of living up to the basic idea of the expressivist about how to give a theory of meaning that are interestingly expressivist, and there are ways of doing so that are uninterestingly so. Any theorist whatsoever whose semantics assigns sentences to propositions or sets of truth-conditions or the like and holds that for all values of ‘P’, believing that P is just a matter of standing in the belief relation to the proposition (or truth-condition) that P, can just affix the belief relation to the propositions assigned by her theory in order to associate sentences with mental state types. It’s true that this wouldn’t be very interesting.

    Interesting expressivist views – and again I think it’s clear that on either of my readings Gibbard falls into this camp – hold that not all beliefs consist in the same belief relation to different contents. For Gibbard, for example, beliefs about what you ought to do consist in planning states, and beliefs about what you have done consist in ‘prosaic’ belief states. So I agree that to be interesting, an expressivist view needs to do a little bit more than the official characterization of what makes it expressivist. But I think Gibbard clearly does that.

    In fact, I actually think things are yet more complicated, because I think there are interesting expressivist views which allow that there is some level of description at which all belief-states consist in a single belief relation toward different contents – but which also hold that there is also a level of description at which they consist in different attitudes. But this complication is more complicated than is going to come through in a blog post.

  8. Thanks, Mark, that is very helpful. It shows what I suspected, which is that it is very hard to say exactly what expressivism is. And, of course, the label isn’t really important anyway–the view, on either of your two readings, is very deep and interesting.

Leave a Reply

Your email address will not be published. Required fields are marked *