The Use of Active and Passive Voice Constructions to Reveal Stance: A Corpus-Based Study on English Department Students' Academic Writing

ABSTRACT Shitadevi, Ida Ayu. 2013. The Use of Active and Passive Voice Constructions to Reveal Stance: A Corpus Based Study on English Department Students’ Academic Writing. Thesis, English Department, Faculty of Letters, State University of Malang. Advisor: NurenziaYannuar, S.S, M.A. Keywords: voice construction, stance, corpus study Academic writing is signified as impersonal and objective (Starfield, 2007); while the focus of the author is deemphasized. However, showing the role of author is part of gaining credits (Hyland, 2002) and showing the author is part of stance, which reflects author attitude, personal thought, and feeling within the text.Baratta (2009) suggested that passive voice can reveal the author stance. In other words, the voice preference constructed in the text becomes important in determining author position in his/her writing. Therefore, the core proposals of the study are to find what makes it different to express ideas in active and passive construction, and the implications of using the two constructions in academic writing. Active and passive voice construction are represented by the stance verbsproposed by Biber (2004). This corpus based study is a descriptive qualitative study. The study employs corpus software Ant.Conc 3.2.4w (Windows) 2011 to generate data corpus from the rough data of English Department students’ undergraduate theses in 2011, 2012, and 2013. The stance verbs used in this study is selected randomly, however they represents semantic features like attitudinal, certainty, non-factive, like-hood, cognitive, speech act, desire, and cause.Data collection stages are: (1) collecting the undergraduate theses submitted to State University of Malang in form of soft files; (2) converting the files into TXT format; (3) categorizing the TXT files into two: the year of submission and the major of study; (4) operating the corpus analysis software of AntConc 3.2.4w; (5) checking the Words box and typing the lexical item in Search Term box, then clicking Start to start the searching; and (6) creating a concordance data by selecting Concordance tab which generates a list of all occurrences of the selected word.Process of data analysis is started by copying the KWIC (Key Word In Context) results to document files, and then grouped them into that-complement clauses and to-complement clause. Each complement clause contains semantic classes that ease the process of interpretation. There are 2526 tokens of active voice constructions, which covers 64.8% of the data and 1375 tokens of passive voice constructions, which covers 35.2% of the data. Thus, the total occurrences of active and passive voice constructions under the verb stances suggest, expect, show, predict, report, believe, hope, allow is 3901 tokens. The high frequency of stance verb show, suggest, expect, and the least frequent of stance verb predict indicate the semantic features of general academic writing, particularly English thesis writing, are factual, communicative, showing attitudinal, and less likely to show prediction. The phenomenon of higher frequency of active voice construction suggest two implication of Indonesian author: that Indonesian authors show resemblance of English native writer, regardless their Indonesian seasoned mixing with English rhetoric style; and it is due to the determination of Indonesian authors to elevate their position internationally. *This article has been presented in International Conference on English Linguistics and Literature (ELITE) 2013 in Maulana Malik Ibrahim Malang University and published in ELITE proceedings