Stereotype in Webkorpora: Strategien zur Suche in sehr großen Datenmengen

Uwe Quasthoff, Erla Hallsteinsdóttir


Research on stereotypes (as well as on similar constructs such as prejudice, image, identity, etc.) goes back around a century. Most research on stereotypes belongs to the field of psychological, social and cultural studies, with a strong focus on cognitive and socio-cultural phenomena. Since Uta Quasthoff’s pioneering analysis in 1973, there have been several linguistic studies of stereotypes, i. e. on linguistic aspects of how stereotypes are realized in language use. Linguistic research on stereotypes uses methods from disciplines such as sociolinguistics, ethnography, discourse and conversation analysis. Discursive psychology and cognitive linguistics and single studies on stereotypes have used corpus-linguistic methods to explore stereotypes in language use. Nevertheless, we state that no large-scale empirical studies have yet investigated the linguistic realization of stereotypes by using empirical data from large text corpora. In this article, we explore the potential of corpus linguistic approaches in the research on the occurrences of stereotypes in written language. By using an analysis of co-occurrences, we aim to describe the relation between associative semantic stereotypes and co-occurrences in a corpus. Our hypothesis is that the linguistic construction of stereotypes as their realization in language use reproduces stereotypic representations of cognitive conceptualizations and meaning patterns, thus, stereotypic meanings should be detectable in statistical patterns in large corpora.