In the Oxford Handbook of Linguistic Analysis, 2015, Chapter 2, “The Adaptive Approach to Grammar”, by T. Givón, the author says that Joseph Greenberg’s attempt to characterize the universal aspect of language was more moderate than Chomsky’s, which he calls comparatively extreme.

Many fine linguists, especially those who followed the structuralist dogma of arbitrariness… expressed strong doubts about language universals… Others, like Chomsky, have militated for an extreme version of universality and innateness, by extracting from the vast and varied phenomenology of language features a few sufficiently abstract ones that are then said to be shared by all human languages…

A more balanced empirical approach to the problem, perhaps best exemplified in the works of Joseph Greenberg… adopts a middle-ground biological perspective, whereby both variation and universals are acknowledged. Thus, specific features of both phonology and syntax may vary considerably across languages, and the aggregation of such variation may lead to a, seemingly, staggering cross-language diversity. But within each functional-structural domain, the variation is severely constrained — say, five to seven major types of structures that usually code the same communicative function. And the constraints on variation are mediated by general adaptive principles… Language universals are not a set of concrete traits found in all languages but rather a set of general principles that control development.

(p. 59)

This sounds to me like a misreading/misrepresentation of Chomsky, where what is being attributed to Greenberg is more or less exactly what Chomsky’s vision of language is, at least as laid out in Chapter 1 of The Minimalist Program: a theory of a system with some variational parameters which account for the seeming observable variation in human language.

Am I wrong?

Is Greenberg more “moderate” than Chomsky in some essential way?

How so?

  • 1
    I like this question because I have been thinking about it too for a while. Thanks for bringing it up!
    – Yili Xia
    Commented Aug 1, 2023 at 9:39

1 Answer 1


Greenberg's approach is completely different from Chomsky's. The two differ in what objects are being investigated: data-patterns analyzed in terms of a set of "language types" (G) versus fundamental causal mechanisms of the mind, viz the theory of rules and representations (C). The term "universal" means different things: for G it refers to "statistically significant properties" and for C it refers to "possible vs impossible".

It is, however, correct that over time, Chomsky has dialed back his stance on how much is "universal". Earlier Chomsky saw language acquisition as an essentially impossible task without massive assistance from UG, whereas the current theory is that in fact a lot can be learned from experience, and factors outside of grammatical theory explain typological patterns. Of course, Givon over-states the difference by his rhetoric (I admit that "over-states" is prejudicial in the same way that his "militates" is prejudicial). His faith in this concept of "language type" as a primitive is not well justified, in my opinion.

Insofar as Greenberg and his program is not widely practiced, we can't really say how his notion of "language type" compares with contemporary generative analysis.

  • 1
    Greenberg's analyses have been crucial in African linguistics, but widely criticized in Amerind linguistics, beyond establishing the order of the last two migrations, of Eskimo-Aleut and Na-Dene. That's an interesting difference, and one wonders how much of it is due to Greenberg and how much to the natures of his audiences.
    – jlawler
    Commented Jul 31, 2023 at 16:21
  • 2
    As for the African thing, I think this is more a "shoulders of giants" effect, that the most popular ideas were already gaining currency. There has since been a significant deconstruction of his phyla, for example Khoisan is gone, various aspects of Niger-Kordofanian are in serious doubt (e.g. the Kadugli languages are not Kordofanian, they are Nilo-Saharan)., and Dogon is in question as a member of NK.
    – user6726
    Commented Jul 31, 2023 at 17:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.