Résumé
Joint work with Dalina Kallulli, University of Vienna.
Since the 1980s, the generative theory of cross-linguistic variation has relied on the notion of parameter of Universal Grammar, according to which universal properties of language are allowed to vary in highly restricted ways (e.g. all languages have verbs, subjects and objects, but these elements may be ordered in different ways in different languages, with associated preferences and implications for other properties).
But can parameters be reduced to a more primitive linguistic notion? A recent conjecture is that parametric variation reduces to features of grammatical elements, known as functional heads; this has become known as the Borer-Chomsky Conjecture, owing to its progenitors. However, while we have a reasonable extensional idea of what such formal features are (e.g. Person, Number, Case, Tense, Aspect, Mood, Negation, …), though such lists are always incomplete, we have no adequate intensional definition to date. This arguably undermines the BCC, since without such a definition, strictly speaking no predictions can be made about the limits to cross-linguistic variation. We present an alternative to the BCC which is based on the twin notions of Redundancy and Deficiency (Onea et al 2023). We argue, on the basis of evidence from word order and silent subjects, that this approach is both more empirically adequate and more restrictive than the BCC approach, as well as providing the desired intensional definition of parameters.
