Benefits of corpus data

1. Leech (1992) argues that the corpus is a more powerful methodology from the point of view of the scientific method, as it is open to objective verification of results

2. Is language production really a poor reflection of language competence as Chomsky really argued? Labov (1969) showed that "the great majority of utterances in all contexts are grammatical". We are not saying that all sentences in a corpus are grammatically acceptable, but it seems probable that the Chomsky's (1968: 88) claim that performance data is 'degenerate' is an exaggeration (see Ingram 1989: 223 for further criticisms of this view).

3. Quantitative data is of use to linguistics. For example, Svartvik's (1966) study of passivisation used quantitative data extracted from a corpora. Elsewhere, all successful approachs to automated part-of-speech analysis reply on quantitative data from corpora. The proof of the pudding is in the eating.

4. Abercrombie's observations that corpus research is time-consuming, expensive and error-prone are no longer applicable thanks to the development of powerful computers and software which is able to perform complex calculations in seconds, without error.

   ---

   Back to Chomsky re-examined