UG or not UG: Where is Recursion?

Abstract

The operation Merge, applying to two syntactic objects to produce\n\t\t\t\t a third and instantiating the property of recursion, has been a fundamental\n\t\t\t\t and largely uncontroversial feature in the development of the Minimalist\n\t\t\t\t Programme. In early formulations, such as Chomsky (1995a, b), Merge is\n\t\t\t\t cited as a feature of the human language faculty that illustrates virtual\n\t\t\t\t conceptual necessity, and it is an examination of this characterisation that\n\t\t\t\t stimulates the concerns addressed here, where we argue that neither of the\n\t\t\t\t familiar routes (satisfaction of interface conditions or computational\n\t\t\t\t economy) provides a justification for the conceptually necessary status of\n\t\t\t\t Merge. A third route, via considerations of ‘languages as such,’ a notion that\n\t\t\t\t includes human and artificial languages, may provide the required\n\t\t\t\t justification, but, as Chomsky (1980) urges, the study of ‘languages as such’\n\t\t\t\t is unlikely to yield empirically interesting results. Specifically, this route to\n\t\t\t\t justification will not locate Merge in UG if the content of UG is an empirical\n\t\t\t\t matter. This conclusion is damaging to the view (Hauser et al., 2002) that the\n\t\t\t\t emergence of recursion (and Merge) is the single development crucial to the\n\t\t\t\t evolution of language, an empirical proposal, albeit in a different discourse,\n\t\t\t\t that firmly places Merge in UG.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 15 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.