Using collocation analysis to reveal the construction of minority groups: The case of refugees, asylum seekers and immigrants in the UK press

Refugees, asylum seekers, and immigrants (henceforth RASIM) coming into the UK have attracted increased press attention (Greenslade, 2005). As their representation in the press can construct their identity (Duffy and Rowden, 2005: 6, in Greenslade, 2005: 7), the discourses surrounding these groups have been the focus of linguistic studies (e.g. ter Wal, 2002). This paper reports on the ESRC funded project, â Representation of refugees and asylum seekers in UK newspapers 1996-2005â. Although the project combines critical discourse analysis and corpus linguistics approaches, the paper aims to exemplify the contribution of corpus research to discourse analysis (cf. Koller & Mautner, 2004; Orpin, 2005; Sotillo & Wang-Gempp, 2004). The study used a corpus of 140 million words, comprising 175,000 articles from 15 UK newspapers, spanning 1996-2005 (see Gabrielatos, 2007). Specifically, the paper reports on the collocational analysis, which adopted the methodology in Baker & McEnery (2005) and McEnery (2006). An added methodological notion, akin to key keywords (Scott, 2004: 115), is that of consistent collocates, i.e. words which are collocates in at least seven out of the ten annual sub-corpora. Collocates can contribute to âa semantic analysis of a wordâ (Sinclair, 1991: 115-116). Also, as â they can convey messages implicitly and even be at odds with an overt statementâ (Hunston, 2002: 109), they are a suitable vehicle for the discoursal presentation of a group (Baker, 2006). The analysis also makes use of the related notions of semantic prosody (Louw, 1993: 157), semantic preference (Stubbs, 2001: 65), and discourse prosody (ibid.: 65-66). The examination of collocation patterns has revealed systematic semantic associations, which map onto the CDA notions of topos (Reisigl & Wodak, 2001: 74â76) and topic (Sedlak, 2001: 129-130), as well as metaphors commonly employed in racist discourse (van der Valk (2000: 234). Arguably, these patterns reveal elements of the underlying discourses relating to RASIM.