I will begin by stating three theses which I present in this paper. The first is that it is not profitable for us at present to do moral philosophy; that should be laid aside at any rate until we have an adequate philosophy of psychology, in which we are conspicuously lacking. The second is that the concepts of obligation, and duty‑-moral obligation and moral duty, that is to say‑-and of what is morally right and wrong, and of the moral sense of "ought," ought to be jettisoned if this is psychologically possible; because they are survivals, or derivatives from survivals, from an earlier conception of ethics which no longer generally survives, and are only harmful without it.I think the second thesis is usually taken to mean that we should jettison the concepts she mentions. But what she actually says is that we should jettison them if this is psychologically possible. It's possible to read her as implying that we ought not to jettison them if we cannot do so. Cannot psychologically, that is. What does that mean? Well, an analysis of 'psychologically' might well belong to the philosophy of psychology, along with analyses of intention and other more familiar concepts in the philosophy of psychology.
Why might this matter? If this is what she means then her thinking seems to be in line with Wittgenstein's on secondary sense. One difference between nonsense and secondary sense is that the latter is psychologically necessary. (Is that right?) (Am I just regurgitating Reshef's ideas here? If so, are they properly digested?) It would also mean that she is usually misread. And, since I have wanted to defend the use of the concept of moral rights but reject the concept of moral obligation, it suggests that I should maybe re-think my position to be sure it's really coherent. If one is justified as either metaphorical or secondary, why can't the other be?
I don't mean that Anscombe is secretly hinting at some big idea about secondary sense. I think she does believe we should get rid of the (non-theistic version of the) notion of moral obligation. But I wonder whether she's making some acknowledgement here of a Wittgensteinian alternative to her view. This alternative would involve using words like 'ought' in a moral sense even though their use in this sense primarily belongs in another context. It would be like calling Tuesday lean or saying that someone calculated in her head. It would not be a metaphorical use of words, because no other words would do. It's this nothing-else-will-do part that makes it psychologically impossible to give up the words in question. Or that's what I'm suggesting.
True, Anscombe says that this use of words is "only harmful" without the primary context, but if giving it up is impossible then perhaps it's the best option available to us.