Towards a radically usage-based approach to constructional attrition

Abstract

Given that it has to date been dominated by a strand of research one could call “constructionist grammaticalization theory” (Noël 2013), it is no surprise that in the steadily emerging field of “diachronic construction grammar” (cf. Barðdal et al. 2015) most attention has focused on “constructionalization” (Rostila 2004; Noël 2007; Traugott & Trousdale 2013), i.e. on how new constructions come about in the “constructicon” (Jurafsky 1992), the pool of lexico-morphosyntactic resources which lay people and linguists alike call “a language” and whose contents those considered to speak it share to varying degrees. The opposite phenomenon, the disappearance of constructions from that pool, which Colleman and Noël (2012) call “constructional attrition”, has received much less interest. Traugott and Trousdale (2013), who call it “obsolescence”, do make mention of it as one kind of “post-constructionalization constructional change” but, in stark contrast to the “usage-based” explanation they attempt to give of constructionalization, they stop at describing it as a “falling out of use” of constructions and say nothing about cognition. Ideally, a usage-based account should have as much to say about speakers’ knowledge as about use, however, given that, as Langacker (1987: 494) defines it, the usage-based model implies a conception of language in which “[s]ubstantial importance is given to the actual use of the linguistic system and a speaker’s knowledge of this use” (my emphasis). Because the term “usage-based” has for many become so diluted to be synonymous with a “corpus-based” methodology, Noël (2016) adopted the expanded name “radically usage-based” to refer to an approach to diachronic construction grammar that takes seriously the “cognitive commitment” to psychological reality of cognitive linguistics (Evans 2016) — cf. Hilpert’s (2018) “open” question of whether this is a requirement in diachronic construction grammar. Such an approach replaces a language perspective with a speaker perspective and takes as a given that “no two members of a speech community have identical linguistic knowledge” (Schmid 2015: 4). There is already radically usage-based research which, as a methodological consequence, makes use of idiolectal historical corpora in the investigation of constructionalization (see, e.g., Schmid & Mantlik 2015, De Smet 2016, Petré 2016). Taking off from a more orthodox corpus-based description of a case of constructional attrition in the history of English, the present paper addresses the questions of why research on constructional attrition should turn to such corpora as well and what answers this should bring.