Syntactic change: A minimalist approach to grammaticalization (review)

A quarter of a century ago, David Lightfoot began to explore the diachronic implications of adopting a formal syntactic framework of the type elaborated in those decades by Noam Chomsky and his associates. Over this span of years, Lightfoot’s work has devastatingly criticized approaches to syntactic change (e.g. those based on notions such as ‘drift’) that cannot be reduced to the inescapable and most basic reality of language, namely true cognitive objects (grammars) present in the mind/brain of individual speakers. This groundbreaking work has provided us with an important demarcation criterion between scientific investigation of diachronic syntax, based on the study of the objects above (essentially I-languages, in Chomsky’s (1986) terms), and more unrealistic and less safe idealizations whose explanatory apparatus assumes the existence of such dubious entities as genetic memory. The tightening of scientific requirements imposed by this fundamental line of reasoning has had the effect of forcing scholars to rethink the status of many classical phenomena that seem anyway to have some salience in historical linguistics, most notably the concept of derivation of one language from another, that of apparently (uni)directional change, if not the very notion of E-language (again in Chomsky’s (1986) sense), as traditionally used in historical linguistics. Some of the work of the last years is thus attempting to reconstruct the useful part of the diachronic apparatus within the new, more severe epistemological bounds made unavoidable by Lightfoot’s work and more generally by the rise of the generative biolinguis-tic program. Keenan’s (1996, 2002) hypothesis of inertia and its interaction with even the classical notion of etymology in Longobardi 2001, the attempt to elaborate a historical-comparative method from parametric syntax (Guardiano & Longobardi 2005), securing the very notion of historical relatedness in the cognitive framework, the explicit formulation of a ‘logical problem of language change’ (Clark & Roberts 1993), and the collection of controversial essays on syntactic reconstruction put together by Ferraresi and Goldbach (2008) all bear witness to this trend.

Syntactic change targets precisely a much debated concept of formal and less formal historical linguistics, going back at least to Meillet (1912) and definitely in need of serious reconsideration in the biolinguistic framework, namely the concept of grammaticalization. Roberts and Roussou aim at providing an account of grammaticalization within the minimalist framework, supporting the analysis with a number of relevant case studies. Grammaticalization is intended as a diachronic change that involves ‘the creation of new functional material, either through the reanalysis of existing functional material or through the reanalysis of lexical material’ (2). The authors claim that this process is epiphenomenal, in that it can be reduced to a regular parametric change, which leads to structural simplification. The related questions that they address in order to give an exhaustive theoretical account of the issue are: (i) the reason why syntactic change in general is so common in languages, given that according to the minimalist program languages are perfect systems, which in principle should not vary over time; (ii) the nature and the inventory of functional categories, whose definition plays a fundamental role in the analysis of the process; and (iii) the apparent conflict between a descriptive approach to grammaticalization, based on the empirical evidence that the process follows clear paths of development in languages, and the explanatory adequacy of the minimalist model, which does not necessarily suggest privileged directions of shift in parametric change.

R&R start by pointing out that the study of linguistic change within the minimalist program poses a logical problem, which derives from hypotheses about the process of language acquisition. According to the principles-and-parameters theory (PPT), the points of crosslinguistic variation in syntax lie essentially in the assignment of different parametric values in different languages. Language acquisition precisely consists of setting these values in accordance with the empirical evidence that the learner is exposed to, that is, with the trigger experience. Thus the process is idealized as deterministic...