Authorial stance in academic english: native and non-native academic speaker writers' use of stance devices (modal verbs) in research articles

Academic writing has recently been conceptualized as not necessarily a purely objective, faceless, dry, convention-bound monolithic entity, involving a distant, convoluted, and impersonal prose, devoid of writer involvement, unfolding in a direct manner leading to an inescapable truth, as once used to be seen. In tandem with this relatively new conceptualization, ?linguists have become increasingly interested in the linguistic mechanisms used by speakers and writers to convey their personal evaluation and assessments? (Biber, 2004: 107; 2006:97).This interest has found its reflection in a number of studies, labeled differently: ?attitude? and ?modality? (Halliday, 1994), ?evaluation? (Hunston,1994; Hunston and Thompson 2000), ?intensity? (Labov, 1984), ?affect? (Ochs,1989), ?evidentiality? (Chafe, 1986; Chafe and Nichols, 1986), ?hedging? (Homes, 1988; Hyland, 1996, 1998), and ?stance? (Barton, 1993), Beach and Anson, 1992), (Biber and Finegan, 1988, 1989), (Bieber et al. 1999; Biber 2004), (Conrad & Biber, 2000).Stance, be it the writer stance or the speaker stance, involves, amongst other things: ?the communication of assessments and value judgments concerning the described situation by appeal to evidence (evidentiality), assessment of the degree of likelihood concerning the described situation (epistemic modality), and arguments regarding the necessity or desirability of the situation obtaining (deontic modality) Biber et al. (1999: 966, 972).This study was conducted to analyze authorial stance in research articles. Specifically, it was designed to investigate how non-native English speaker academic writers (Turks) and (Spaniards) and native English speaker academic writers (American) express their stance in their research articles published in a refereed international journal by focusing on only eight modal verbs; ?can?, ?could?, ?may?, ?might?, ?will?, ?would?, ?should?, or ?must?. The data for this study originated from 45 research articles written by non-native and native speaker academic writers. The research articles were published in the American Social Behavior and Personality between 1993 and 2007. All of the modal verbs in the different rhetorical sections ?Abstract?, ?Introduction?, ?Methodology?, ?Results?, and ?Conclusion? of the research articles were identified by Wordsmith Tools, a computer program, and manually by the researcher. The modal verbs used in direct quotation were excluded from the study. Then, the functions of the modal verbs were independently determined by the researcher himself and a native speaker, and the Spearman Coefficient test was run to ascertain inter-rater reliability. As the final step, Varbrul, a multivariate analysis program, was used for statistical analysis.The analyses of the data indicated the existence of both similarities and also differences. The findings demonstrated that both native and non-native writers are well aware of the functions of the rhetorical sections of the research article. It was observed that the distribution of the modal verbs across the different sections of the research articles bears some similarities as well as some differences, slight though they are. The results demonstrated that writer stance in the research article is partly governed by the global discourse community of the research article and its conventions as well as the local discourse community of the writer and its conventions.A couple of implications were drawn from the study and some suggestions were made regarding research article writing, academic and advanced writing courses, and teaching modal verbs, by considering all these from a totally different perspective.