Three and a half years ago I had the pleasure of doing a guest post here on PEA Soup. I'm still working on the same stuff, so I thought I'd start my non-guest posting with an update. There's quite a few things going on in the paper, so I'll just try to hit the highlights here in the hopes that it generates discussion. (I'll also be talking about this stuff at RoME in August, for those of you who will be there.)
I think that expressivism is incompatible with the possibility of expressing certain kinds of negated normative judgements that further express nihilism or nihilistic doubt. And I think that because nihilism is a view that we must at least be able to entertain, this is a serious problem for expressivism.
Consider four things that might be true of you:
(a) You think that murdering is wrong.
(b) You do not think that murdering is wrong.
(c) You think that murdering is not wrong.
(d) You think that not murdering is wrong.
Cognitivists can make sense of these states of mind just by replacing "think" with "believe." Expressivists have a harder time of things. For one, it's not clear that they can explain why (a) and (c) are inconsistent. This is (a part of) "the negation problem" for expressivism. In Being For, Mark Schroeder argues that expressivists need to accept his "biforcated attitude semantics" (or something formally equivalent) on which all declarative sentences express a general non-cognitive attitude: BEING FOR . I'll assume here that he's right about this. Biforcated reinterpretations of (a)-(d) look like this:
(e) You are FOR blaming for murdering.
(f) You are not FOR blaming for murdering.
(g) You are FOR not blaming for murdering.
(h) You are FOR blaming for not murdering.
Now, suppose that a normative nihlist—someone who has no substantive normative committments—asserts "murdering is not wrong." It seems most natural to understand him in line with (c). For the expressivist, that means understanding him in line with (g). But that won't work, because BEING FOR something means having a substantive normative committment, and he doesn't have any of those. (I think it's clear that BEING FOR has to involve substantive normative committments in order to play the relevant role in the semantics, but obviously this is a point someone might want to push me on.)
Alternatively, we might interpret the nihlist in keeping with (f) instead. But this raises problems: Aguably, it's not possible to directly express the absence of an attitude. Even if it is possible, this isn't enough to solve the negation problem. After all, if I believe something and you merely don't, we do not therefore disagree. So expressivsts would have to explain why BEING FOR something and not BEING FOR it are inconsistent in the relevant sense.
So far as I can tell, there aren't any other (plausible) ways of interpreting our nihilist. So expressivists have to bite the bullet. One thing they might say is that the bullet is small because nihilism is coherent. But even if it is incoherent in the sense of self-contradictory (which I don't think it is), it still seems to be a view that we can entertain (lots of people have, philosophers and non-philosophers alike). And if it's a view we can entertain, then expressivists would still need to offer a plausible interpretation of the nihilist's claim, and I've just said that they can't.
The last move (that I can see) open to expressivists is to claim that because nihilism is a metanormative view, nihilistic claims aren't something they need to be able to make sense of any more than claims like "murdering has the Platonic property of goodness." Now, this might be a move the expressivist can make. But I don't think it's an appealing one, because I think that entertaining nihilism requires only one metanormative committment: the idea that normative facts and properties might be "objective"—be "out there" in the world. (An analogy: Even if I am certain I am having a computer-screen-like experience now, I can still doubt that there's a computer in front of me. That's because my thinking about computers allows for the possibility (even the likelihood) that the subject of my thought is "out there.") Since I think that such a "weak objectivism" about normative thought and language is extraordinarily plausible, I think that denying it (as expressivists have to in order to assert that nihilism is in competition with their view and thus not something they need to capture in their semantics) is an unacceptable price to pay.