Grammar of Binding in the languages of the world: Innate or learned?

Peter Cole,Gabriella Hermon,Yanti
DOI: https://doi.org/10.1016/j.cognition.2015.04.005
IF: 4.011
2015-08-01
Cognition
Abstract:Languages around the world often appear to manifest nearly identical grammatical properties, but, at the same time, the grammatical differences can also be great, sometimes even seeming to support Joos's (1958) claim that "languages can differ from each other without limit and in unpredictable way" (p. 96). This state of affairs provides a puzzle for both nativist approaches to language like Generative Grammar that posit a fixed "Universal Grammar", and for approaches that minimize the contribution of innate grammatical structure. We approach this puzzling state of affairs by looking at one area of grammar, "Binding", the system of local and long distance anaphoric elements in a language. This is an area of grammar that has long been central to the Generative approach to language structure. We compare the anaphoric systems found in "familiar" (European-like) languages that contain dedicated classes of bound and free anaphors (pronouns and reflexives) with the anaphoric systems in endangered Austronesian languages of Indonesia, languages in which there is overlap or no distinction between pronouns and reflexives (Peranakan Javanese and Jambi Malay). What is of special interest about Jambi anaphora is not only that conservative dialects of Jambi Malay do not distinguish between pronouns and reflexives, but that Jambi anaphora appear to constitute a live snapshot of a unitary class of anaphora in the process of grammaticalization as a distinct system of pronouns and reflexives. We argue that the facts of Jambi anaphora cannot be explained by theories positing a Universal Grammar of Binding. Thus, these facts provide evidence that complex grammatical systems like Binding cannot be innate. Our results from Austronesian languages are confirmed by data from signed and creole languages. Our conclusion is that the human language learning capacity must include the ability to model the full complexity found in the syntax of the world's languages. From the perspective of child language acquisition, these conclusions suggest that Universal Grammar does not provide a general solution to the problem of poverty of the stimulus, and the solution to that problem must reside at least in part in special properties of the grammar construction tools available to the language learner rather than simply in a fixed set of grammatical rules hard wired into the brains of speakers.
psychology, experimental
What problem does this paper attempt to address?