This year’s conference season is over (for me at least), and I will now once again, I hope, be able to devote a few shreds of surplus attention to keeping my posts here a bit more regular. The latest conference I went to was in Vienna (where I always like to go anyway); the last day of it was on the Berggasse right next to where Freud’s office used to be (and a Freud museum now is). I’m sure that someone somewhere must have remarked on the irony that the Berggasse is the continuation of the Schwarzspanierstraße, where Beethoven died — in the building Otto Weininger sought out to commit suicide in 75 years later. (Freud, by the way, unlike Wittgenstein, was apparently unimpressed by Geschlecht und Charakter.)
One thing that came up a number of times at this very interesting conference, organized by Christian Damböck (together with Meike Werner and Günther Sandner), was Carnap’s “non-cognitivism.” The word was used in a number of different ways, which I found very confusing. I propose that when talking about Carnap, at least, we stick to what Carnap himself meant by it, which seems especially appropriate since, as far as I can tell, he actually introduced the term. At the beginning of his reply to Kaplan in the Schilpp volume (his only extended exposition of a logic of normative statements), he explains why he intends to avoid the widely-used term “emotivism,” and characteristically proposes an ugly but less tendentious neologism in its place:
My own conception of value statements belongs to the general kind which is customarily labeled “emotivism”. However, this term is appropriate only if understood in the wide sense in which Stevenson speaks of “emotive meanings”. He warns explicitly (pp. 59f.) that his term does not refer to momentary emotions in the ordinary sense, but rather to attitudes. However, since the term “emotivism” is sometimes associated by critics with too narrow an interpretation which today is rejected by most of the adherents to the conception. . ., it is perhaps preferable to use a more general term, e.g. “non-cognitivism (with respect to value statements)”. (Schilpp volume, pp. 999-1000)
Is anyone aware of an earlier use of the term “non-cognitivism”? I’m not. If anyone who reads this knows of an earlier instance, please let me know (keeping in mind that Carnap wrote the above in 1957 or so, six years before it was published, and circulated it widely in mimeograph) — and I will correct this post accordingly.
And what does Carnap mean by it? He gives two formulations of the “thesis of non-cognitivism”: first, in a “very weak” form:
If a statement on values or valuations is interpreted neither as factual nor as analytic (or contradictory), then it is non-cognitive; that is to say, it is devoid of cognitive meaning, and therefore the distinction between truth and falsity is not applicable to it. (p. 999)
After a page or two in which he explains (as above) why he avoids the term “emotive” and “emotivism,” and introduces the term “optatives” (for what Hare calls “imperatives”), he gives a stronger version of the thesis: “There are pure optatives.” (p. 1001).
He goes on in the following pages to explain that he means this to say that pure optatives can be introduced into constructed languages (not that they must be, of course); he takes no position on the question whether a vague or approximate version of the distinction between normative and descriptive sentences can be found in ordinary language. He does go to some pains to stress that he means his thesis to obtain even if it is not (as e.g. it is presented in Hare) an appropriate explication of any distinction in ordinary language.
This thesis is simply Hume’s point about the impossibility of deriving an “ought” from an “is.” Moore’s “naturalistic fallacy” is just the assertion of that possibility. In Hume and Moore (and of course in Carnap) this simple logical claim is combined with other elements. I propose, though, that we restrict the application of the term “non-cognitivism” to just this minimal core, and find other words to describe the elements it is combined with by these and other authors.
At the Vienna conference last week there was talk, for instance, of “democratic non-cognitivism” and “authoritarian non-cognitivism,” among other extravagances. Not surprisingly, in the discussion someone said that he completely rejected non-cognitivism, which he called absurd. When I asked him later what his objection was to the simple minimal Humean claim of the underivability of ought from is (i.e. to what Carnap meant by non-cognitivism), he said, “no objection whatever — that part I have no problem with.” So I suggest we try to keep things clear and simple, to avoid pointless disputes.