Formulaic expressions in scientific writing: A study of the borderline between familiar and novel utterances

This study investigates a key problem for applied linguistics; i.e., the analysis of formulaic utterances. The problem is in identifying the 'borderline' between formulaic and non-formulaic expressions along the recently popularized 'continuum' between novel expressions and fixed idioms. Although such a continuum promises to be quite useful in applied linguistics research and second language pedagogy, at this stage our understanding of it would appear to be simply notional, that is, intuited from a folk metalinguistic level of observation. It has become clear from previous investigations that a multi-functional approach is needed if we are to understand what the 'substance' of this continuum consists of, and thereby why the continuum is intuitively satisfying. There have been useful attempts to account for the continuum which implicate interrelated syntactic, lexical, pragmatic, and social criteria. However, no organic, constitutive relationship among these functional levels of analysis has been proposed which accounts for our feeling the relative strength of idiomaticity, formulaicness, or lexicalization upon the utterance of any particular expression. This study includes the aforementioned levels of analysis in an investigation of the aforementioned borderline in order to (1) develop an operationalizable theoretical account of the essential features of formulaicness and (2) test this account by targeting selected matrix verbs used in making knowledge claims in a corpus of journal reports from four scientific discourse communities (cognitive psychology, applied anthropology, ethology, and generative linguistics).