On the relation between acceptability and frequency
The aim of this work is to lay out the relation between the two main sources of linguistic evidence, acceptability judgments and frequency of occurrence in spontaneous speech. Tying acceptability to grammar and frequency to usage, the empirical relation between these data sources is seen as a manifestation of the underlying relation between grammar and usage. Contrasting with the functionalist view, I argue in favor of a clear distinction between grammar and usage, and contrasting with the standard generative view, I propose that both grammar and usage are constitutive parts of human syntax. They are related in a very specific way and their interplay is crucial in syntactic variation and the process of grammatical change. A systematic experimental study has been conducted in which acceptability and frequency data are obtained from the same subjects and on the same linguistic phenomenon: which is the well-known interaction between the semantic role of the wh-element and the preferred position of the subject (preverbal vs. postverbal) in Spanish. The results show that acceptability is not a sufficient but a necessary condition for usage. Constructions that are acceptable but not or hardly used are called latent constructions. It is suggested that latent constructions can occur as an intermediate step in the process of syntactic change.