This article reviews literature on constructional change and its integration with grammaticalisation theory. It discusses a number of issues that are relevant to work on grammaticalisation and addresses several questions which fall out from a constructional view of grammaticalisation. It explores a constructional approach to grammaticalisation and the question of how constructions change and evaluates whether construction grammar is useful to linguists examining grammaticalisation changes to lexical items. It also considers the relationship between constructional change and the grammaticalisation of lexical items.

This paper reexamines the development of the it-cleft construction from the perspective of grammaticalization theory. In a previous study, Ball (1991, 1994) finds that the it-cleft was initially restricted to NP foci, with the relative clause expressing presupposed information that is already known to the hearer/reader. However, it is well known that the modern day it-cleft also permits a range of non-NP foci and that the relative clause is no longer restricted to presenting information that is necessarily known to the intended audience. Using data from the Penn Parsed Corpora of Historical English, I argue that synchronic variation in the it-cleft can be understood as a consequence of gradual constructional emergence. Non-NP focus it-clefts are shown to originate by extension from the existing NP focus it-cleft, resulting in a more schematic it-cleft construction. In line with Lambrecht’s (1994) theorizing, I provide evidence that the development of the ‘informative-presupposition’ (IP) it-cleft involves grammaticalization, whereby a sense of presupposition gradually becomes associated with the construction as a whole and is no longer predictable from the meaning of its parts. Discussing the implications for the present day it-cleft, this paper provides insight into how the gradual grammaticalization of constructions, and mismatch in particular, intersects with synchronic gradience (as understood by Croft 2007).

No abstract

No abstract

Recent research in construction grammar has been marked by increasing efforts to create constructicons: detailed inventories of form-meaning pairs to describe the grammar of a given language, following the principles of construction grammar. This paper describes proposals for building a new constructicon of English, based on the combination of the COBUILD Grammar Patterns and the semantic frames of FrameNet. In this case study, the valency information from FrameNet was automatically matched to the verb patterns of COBUILD, in order to identify the frames that each pattern is associated with. We find that the automatic procedure must be complemented by a good deal of manual annotation. We examine the “V that” pattern in particular, illustrating how the frame information can be used to describe this pattern in terms of constructions.