What corpora can and cannot tell us about cognition: The case of prototypicality
(2008) Cognitive Corpus Linguistics: Current issues in theory and methodology — Location: Freiburg im Breisgau (12.October.2008)
Files
No attached file found for this publication.
Details
- Authors
- Abstract
- Since their advent on the linguistic scene, corpora have been widely used to describe language. Linguists have repeatedly shown that corpus-based analyses can reveal facts about language which would not be accessible on the basis of intuition alone, and the descriptive function of corpora, therefore, need no longer be demonstrated. Some linguists, however, have suggested that corpora can also be used to say something about cognition. The idea is that, since the human mind is a black box which cannot be accessed directly, it has to be studied indirectly, via some of its manifestations. Language, as one type of systematic behaviour of humans, has been presented as a “relatively privileged (…) ‘source of information’ on conceptualization” (Pederson & Nuyts 1997: 4), because it “explicitly encodes and transmits conceptual information” (ibid.). But the relation between conceptualisation/cognition and language is far from clear. While some underline the links that exist between corpora and cognitive processes (cf. Reali & Christiansen [2007], who claim that “processing of relative clauses is made easier by frequency of occurrence”), others argue that such links are not systematic (Schmitt et al. [2004], for example, show that recurrent clusters identified on the basis of corpora are not necessarily “psycholinguistically valid, that is, stored as holistic units in the minds of proficient speakers”). In this discussion of the link between corpora and cognition, the focus will be on the issue of cognitive salience (understood here as what is most accessible in the mind), and more precisely the following question: can linguistic frequency (as attested in corpora) be used as a shortcut to cognitive salience? Results from my own and others’ research will be reported which, by comparing corpus and experimental data, suggest that the (sometimes unspoken) assumption that frequent items are somehow cognitively salient, and more generally that corpora are “psycholinguistically valid”, may have to be called into question.
- Affiliations
Citations
Gilquin, G. (2008). What corpora can and cannot tell us about cognition: The case of prototypicality. Cognitive Corpus Linguistics: Current issues in theory and methodology, Freiburg im Breisgau. https://hdl.handle.net/2078.5/158687