Integrating corpus and genre approaches: phraseology and voice across EAP genres.

This chapter deals with a study of personal pronouns in context and aims at revealing trends in popular and research articles written by historians. Focusing on two pronouns that turn out to be positive and negative keywords when comparing popular and research writing, the analysis has focused first on distinguishing averred from attributed forms and then on patterns of semantic preference and semantic sequence. Corpus and genre have played a key role in approaches to English for Academic Purposes (EAP) over the past 20 years. The interplay between the two notions, far from leading to contradictory methods, has proved extremely fruitful both from a descriptive and from a pedagogic point of view (Swales 2002). The integration of tools that can be related to the two notions has provided excellent means for the analysis of language variation across genres, cultures and disciplines. Swales’s (1990) seminal book on genre analysis initiated a vast area of EAP studies combining descriptive and pedagogic interest in the role of genre-based studies (Bhatia 1993, 2004; Berkenkotter/Huckin 1995; Johns 1997, 2002; Hyland 2000, 2006; Paltridge 2001; Swales 2004). Genre, defined as a class of communicative events with a particular purpose recognized by a discourse community (Swales 1990), rapidly became the focus of analysis and the key organizing principle of pedagogic programmes. Genres are staged, goal-oriented social processes in Hallidayan terms (cf. Halliday/Matthiesen 2004, Martin/Rose 2007). This means that they are described in terms of sequences of moves and steps (representing their generic structure) with elements of variation: there can be optional moves, a flexible order, and patterns of embedding and repetition. The fact that they are seen as social processes accounts for their dynamism (change across different contexts, in response to new media or to changes in the cultural and disciplinary background) and for the major role played by intertextual reference to other texts. Corpus approaches to language analysis (e.g. Sinclair 1991, 2004; Stubbs 1996, 2001; Tognini Bonelli 2001; Hunston 2002) have greatly contributed to register studies, as “a corpus is a collection of naturally occurring language text, chosen to characterize a state or a variety of a language” (Sinclair 1991: 171). Specialized corpora can thus be used to describe the practices of particular discourse communities (Biber et al. 1998; Ghadessy et al. 2001; Gavioli 2005; Connor/Upton 2004; Biber 2006). Wordlists reveal the range and frequency of occurrence of language items; concordances allow the study of specific items in their lexico-grammatical, semantic and pragmatic environment; phraseology contributes to the exploration of systematic relations between text and form (Sinclair 2004). Genre studies have rapidly taken advantage of the potential of corpus linguistics (e.g. Hyland 1998; Bondi 1999) and have become firmly rooted in a growing body of literature focusing on the integration of corpus and discourse approaches (Partington et al. 2004; Baker 2006, Ädel/Reppen 2008). This means relating textual practices to language choice, so that statements about genre can be supported with reference to data and, on the other hand, corpus data are not only described but also interpreted in terms of textual structure and social action (Bondi 2008). Attention to frequency and patterns highlights the existence of systematic relations in texts on a functional or semantic basis. The whole process can be seen as a form of interaction between the analyst and different types of data and methodological tools. Moving beyond the corpus-based vs corpus-driven distinction (Tognini-Bonelli 1993), but still keeping the distinction in mind as a heuristic tool, corpus work can be