Constructicons are bound to change: constructional change in a radically usage-based perspective

Abstract

This paper explores some of the implications of a radically usage-based diachronic construction grammar approach for a theory of grammatical change. In the introductory chapter to the recent edited collection so entitled “diachronic construction grammar” is succinctly characterized as “the historical study of constructions” (Barðdal & Gildea 2015: 42). It has also been described as a field of work in linguistics that addresses linguistic change from the perspective of construction grammar (slightly adapted from Traugott & Trousdale 2013: 39). In other words, diachronic construction grammar is constructionist historical linguistics. Alternatively, switching round the object of study and the approach, one could characterize it as historical constructionist linguistics, i.e. as a field of linguistics which looks at the evolution of the constructional resources of a language, i.e. of “constructicons”. In “radically usage-based” diachronic construction grammar, however, languages are abstractions, to the extent that in terms of the locus of language these constructicons can only be assumed to exist at the idiolectal level, as part of the speaker/hearer’s communicative resources, and even there they are never fixed but always in flux. The radicalness of this approach to historical linguistics, therefore, resides in that it takes seriously the distinction between the individual’s “internal” linguistic system, “structures posited by the analyst as a claim about mental structure and operation” (Kemmer & Barlow 2000: x), and the “external” linguistic system, i.e. descriptions of the conventionalized linguistic system, “hypothesized structures derived by the analyst from observation of linguistic data, with no expectation that such structures are cognitively instantiated” (ibid.). Change operates on internal systems and indirectly results in changed external systems. In historical linguistics this distinction needs to be crucially made to arrive at plausible explications of how change comes about and the failure to do so, even in work that declares itself to be usage-based, has hampered progress in this area, e.g. with regard to the question of whether change is gradual or abrupt (cf. Nørgård-Sørensen & Heltoft 2015: 268-9). Omitting to distinguish between the results of change observed in usage data and the changes proper that are likely to have occurred in the innovators’ constructicons may lead to mistaken statements of paths of change of the kind “form x with function ‘y’ changed into form x' with function ‘z’”. Traditionally, historical linguistics indeed tends to be predominantly semasiological and to favour polysemy above homonymy. The recognition that the loci of change are speaker/hearers’ internal systems invites a more holistic approach which considers “the synchronic system of grammar that is part of the speaker’s acquired knowledge” (Fischer 2008: 338), both from a semasiological and an onomasiological perspective (Van de Velde 2010), and which is amenable both to homonymy and to the likelihood of “multiple sources” for innovations (cf. Hendery 2013; De Smet et al. 2013). Making reference to some of the implications of a radically usage-based diachronic construction grammar and drawing on usage data from a diachronic corpus and several text archives, this paper will present the development of the “participant-external necessity”/epistemic be bound to construction. Though it is tempting from a semasiological polysemy perspective to interpret the data to be supportive of the claim that it evolved from the deontic be bound to construction as a conventionalized implicature, a broadening of one’s outlook to include other likely ingredients of the innovators’ constructicons suggests an altogether different (alternative or complementary) scenario.