Why generative grammar needs innate building blocks in practice: An open response to José-Luis Mendívil-Giró
Martin Haspelmath
March 2020
 

Dear José-Luis, in your recent “open letter” (Mendívil-Giró, March 2020, https://ling.auf.net/lingbuzz/005075), you advance the idea that generative grammar (GG) and functional linguistics have “different objects of study”, contrasting with some recent writings of mine (e.g. Haspelmath 2020), where I say that the possibility of non-innatist explanations calls the practice of relying on innate building blocks (an innate grammar blueprint or “universal grammar”) into question. In this response to your letter, I first point out that social grammars (“externalizations”) and mental grammars (“I-languages”) have not been, and cannot be, separated in practice by linguists. There are no specific methods for studying mental knowledge systems, and not only generative grammarians, but many other linguists strive for mental reality. There is no reason to think that some approaches in linguistics are less “hypothetico-deductive” than others. Like many other linguists, you are not very clear about the goal of GG in your open letter: Is it to describe particular I-languages, or is it to identify what is necessarily common to all I-languages? If the latter, then we would somehow have to subtract the culture-specific aspects of individual I-languages, but there is no clear method for doing so. In practice, GG must assume that the building blocks that it posits are innate, because formal notions like DP, CP, X-bar theory, c-command are thought to be identical across all languages, and there is no other causal mechanism for such identity. Innateness of such highly specific concepts is inherently unlikely, so the idea that languages can only be compared on the basis of uniform building blocks seems to be wrong. This does not mean that Chomsky’s 21st century ideas are wrong, because they no longer rely on a rich innate grammar blueprint. So you may be right that GG in this newer sense can do without innate building blocks, but then much of the practice of current generative grammarians cannot be on the right track.
Format: [ pdf ]
Reference: lingbuzz/005084
(please use that when you cite this article)
Published in: also on my blog: https://dlc.hypotheses.org/2324
keywords: generative grammar, universal grammar, language universals, language description, typology, syntax
Downloaded:219 times

 

[ edit this article | back to article list ]