Using computerized corpus analysis to investigate the textlinguistic discourse moves of a genre

Recently there has been a growing interest in and recognition of the value of specialized corpora, such as learner corpora [Granger, S. (1998). The computer learner corpus: a versatile new source of data for SLA research. In S. Granger, Learner English on computer (pp. 3-18). New York: Longman], in facilitating discourse analysis. Despite this trend, most corpus-based analyses have centered on the lexico-grammatical patterning of texts with less regard for functional and rhetorical, textlinguistic aspects [Flowerdew, L. (1998). Corpus linguistic techniques applied to textlinguistics. System, 26, 541-552]. The goals of this study were: (1) to demonstrate the efficacy of a multi-level analysis of a genre-specific learner corpus that included both a hand-tagged moves-analysis coupled with a computerized analysis of lexico-grammatical features of texts; and (2) to show how a pragmatic concept such as politeness can be operationalized to allow for computer generated counts of linguistic features related to that concept. In this study of politeness strategies used by Americans, Finns, and Belgians in a learner corpus of letters of application, we found that Americans as a group tended to be much more patterned, even formulaic, in their politeness strategies. The Belgians, on the other hand, showed more individuality in their letters with the Finns exhibiting both traits to lesser degrees. In this paper we argue for a textlinguistic approach that considers the special features of genre-specific corpora. (C) 2001 The American University. Published by Elsevier Science Ltd. All rights reserved.