Donald Davidson, back in the mid-80’s, provided an argument that there are no languages, and no such thing as specifically linguistic competence. Drawing attention to successful communicative interaction despite malaprops and new coinages, he writes: “I conclude that there is no such thing as a language… There is therefore no such thing to be learned, mastered, or born with. We must give up the idea of a clearly defined shared structure which language-users acquire and then apply to cases. …I think, we should give up the attempt to illuminate how we communicate by appeal to conventions” (p. 265)
I’ve long thought the argument a bit of a mess. A first point is that the picture he attacks is by no means standard. It definitely isn’t that of, say, Chomsky or Fodor. The more usual picture was best explained and popularized by Sperber and Wilson: yes, speaker and hearer know codes/conventions of language which more or less overlap, but knowing the code is (almost?) never sufficient even for arriving at the literal content of the speech act. In Chomsky’s terms, linguistic competence, the I-language, never suffices for comprehension even of perfectly ordinary, literal talk – the latter being a performance phenomenon. Malaprops et al. provide no evidence against this more modest role for shared, learned languages: they merely reinforce a main tenet of the view, namely that semantics should not be in the business of trying to provide interpretations of speakers. (Lots of pragmatics is required for that.)
The second point relates to this. Davidson’s inference seems to be…
P1: If so-called linguistic competence is to play a role in talk exchanges then: a) it is systematic/compositional; b) it is shared; c) it is learned in advance; d) it suffices for understanding literal utterance meaning.
P2: Nothing meets all of (a)-(d).
C: So-called “linguistic competence” does not play a role in talk exchanges.
In “A Nice Derangement of Epitaphs”, Davidson spends all his time on P2. But P2 is not that controversial, at least within standard linguistics. Indeed, as I say, P2 is a consequence of the Chomsky’s own competence/performance distinction! (It’s like: “Knowledge of grammar alone won’t get you the facts about which sentences sound good”. Fair enough. But we don’t conclude that knowledge of grammar therefore does not exist.)
What’s really dubious, I think, is P1. In particular, (d): why should the fact that the “code” is not sufficient entail that it isn’t necessary? (Compare: Why should we infer from the insufficiency of one cause that it doesn’t play a role in achieving the effect?) But without P1 there’s no support for C.