Grammar, Text, and Ideology: Computer-assisted Methods in the Linguistics of Representation

M. Stubbs

Applied Linguistics · 1994 · 98 citations · 0 references

Concepts

TL;DR

The background highlights that attributing events and knowledge to agents and the need for detailed semantic analyses of verb classes are key issues in using corpus data for stylistic analysis. The purpose is to argue for a relationship between syntactic, textual, and ideological analysis and to outline the descriptive methods needed for text analysis. The authors analyze language in two school textbooks (80,000 and 30,000 words) and a 1‑million‑word corpus, focusing on causativity in ergative constructions and modality in projecting that‑clauses. The findings reveal statistically significant differences in syntactic pattern distribution between the two books, interpreted as evidence of differing ideological stances.

Abstract

This article analyses the use of language in: – two school textbooks of 80,000 and 30,000 words; – a corpus of written English of 1 million words. Two main linguistic features are analysed: – the expression of causativity in ergative constructions; – the expression of modality in 'projecting' that-clauses. These concern, respectively, whether events and knowledge are attributed to agents. The analyses show large (statistically significant) differences in the distribution of syntactic patterns in the two books, and these differences are discussed as evidence of the different ideological stances expressed in the books. But the use of corpus data for the stylistic analysis of long texts is shown to require more detailed semantic analyses of verb classes than are currently available. The main argument concerns the relationship between syntactic, textual, and ideological analysis, and the descriptive methods required in text analysis.