Structural analogy and universal grammar

Abstract

For much of the 20th century, and sporadically in earlier periods, many grammarians drew, more or less explicitly, structural parallels between the syntax and the phonology. Here I discuss one explicit instance of this, embodied in what I have called elsewhere the structural analogy assumption (SAA), a less extreme version than the isomorphism advocated by e.g. Louis Hjelmslev. This interpretation of the SAA takes analogy between levels to be unmarked, and attempts to limit in a principled way the divergences in structure that can be exhibited by different levels of representation. Such syntax-phonology analogies have been argued, notably by Phil Carr, to be incompatible with the hypothesis concerning universal grammar (UG) that sees it as strongly innate and that on these grounds excludes phonology from UG; and the viability of particular proposed analogies has as a consequence been called into question. Here, in response, I question the soundness of the alleged motivations for such a view of UG, and argue, on both substantive and formal grounds, for the appropriateness of anticipating pervasive and detailed analogies between the syntax and phonology, as well as briefly considering further the basis for limitations on analogy. In the course of this consideration I expose what I take to be an inappropriate intra-syntactic analogy, the attribution, by the X′ theory of syntax, of parallel structures to the constructions projected by different categories.