tv.um.es

This paper tries to bring together two strands of thought: the Idiom principle described by Sinclair (1991) and the notion of an individual grammar proposed by Hopper (1987, 1998) and Hoey (2005).
Sinclair notes that, in English, multi-word-units form meaningful non-compositional lexical items; he also makes very clear that such phrases allow for variation and he sees it as a prevalent form of language construction. By contrast, Erman and Warren (2000) found that more than a half (around 55%) of a text would consist of prefabricated language, and only 45% single word choices. A possible reason for open choice principle (cf. Sinclair 1991; Erman and Warren 2000) might be explained by the theory of how grammar emerges based on each producer’s (individual and collective) experience and use of language (as described by Hopper and Hoey).
Given that Erman and Warren’s study is now 20 years old, and given it may ignore Sinclair’s dictum that the idiom principle allows for multi-dimensional variation, a new look on fresh data might be useful. As O’Keefe et al. (2007) pointed out, teaching (and, by extension) understanding of a language has to move away from single words and should take into account highly frequent chunks instead. Furthermore, an investigation looking at the proportion of phraseology in texts
appeared in a more recent study by Colson (2017).
Looking at both spoken (BNC 2014) and written corpora (various sub-corpora of the BNC-written), differences in two dimensions have been found. Firstly, online spoken production is, indeed, far more formulaic than edited published text. Secondly, while a broad general tendency can be detected in line with the idiom principle as described by Sinclair, individual choices become apparent where variation appears dependent on co-text (cf. Sheela and Kuperman, 2016) and the producer’s own background. It could therefore be predicted that idiomaticity is far more common than older studies suggest.