Journal of
ENGLISH for
ACADEMIC
PURPOSES
Journal of English for Academic Purposes 6 (2007) 336–355
United States, totaling 1.4 million words.
ARTICLE IN PRESS
www.elsevier.com/locate/jeap
1475-1585/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jeap.2007.09.004
Tel.: +1 619 594 7114.
E-mail address: ecsomay@mail.sdsu.eduResults indicate that students in American classrooms take more turns than teachers. However,
over 80% of student turns contain fewer than 30 words at a time. When comparing linguistic
differences manifested in longer turns, it is apparent that teachers use significantly more linguistic
features associated with a contextual, directive orientation in Engineering, Business, and in
Education, and in five of the six disciplines examined (except Engineering) students use mostly
linguistic features associated with ‘personalized framing’ elements. Patterns of variation are also
present across levels of instruction.
r 2007 Elsevier Ltd. All rights reserved.
Keywords: English for academic purposes; Corpus linguistics; Spoken discourse analysis; Classroom interaction;
Language variation; University classroom contextA corpus-based look at linguistic variation in
classroom interaction: Teacher talk versus student
talk in American University classes
Eniko Csomay
San Diego State University, Linguistics and Asian/Middle Eastern Languages, MC 7727,
San Diego, CA 92182-7727, USA
Abstract
Studies on classroom interaction have typically focused on relationships between turn-taking
patterns and some larger unit of analysis of varying length and nature. However, two questions still
left unanswered are how teachers talk differently from students in general and how linguistic
variation between two participants might relate to differences in discipline or level of instruction.
The present study takes a corpus-based approach to explore differences between ‘teacher talk’ and
‘student talk’ analyzing a large collection of spoken texts collected in American university
classrooms. The corpus comprises 196 transcribed sessions recorded at five universities across the

ARTICLE IN PRESS1. Background
Over the past three decades, most classroom interaction studies (Cazden, 2001) have
been carried out in elementary school settings (e.g., cultural differences in interactional
sequences by Poole, 2005) or in language classrooms (Bygate, 1988), focusing mainly on
interactional patterns as they relate to larger functional units of analysis (e.g.,
Initiation–Response–Feedback models by Sinclair & Coulthard, 1975). These studies are
invaluable as they provide us with detailed analyses of interaction in the classroom;
however, large-scale investigations into patterns of language use by teachers and students
are still missing. Additionally, while a number of studies have discussed student writing in
academic contexts from multiple perspectives (e.g., Hyland, 2002; Johns, 1997), little
attention has been given to the characterization of participants’ patterns of language use in
university class sessions.
1.1. University classroom talk
Linguistic studies investigating academic lectures have focused on one or a few class
sessions at a time to ‘‘reflect the general interest in the lexical, rhetorical, and topical
structures of discourse’’ (Csomay, in press). For example, numerous studies have reported
on the varying functions of individual linguistic and lexical items in lectures (e.g., lexical
phrases by Nattinger & DeCarrico, 1992; idioms by Simpson & Mendis, 2003; reflexivity
by Mauranen, 2001; evaluative adjectives by Swales & Burke, 2003; pronouns by Fortanet,
2004). As for the rhetorical perspective, most studies were carried out in the 90s,
and reported on disciplinary differences as they relate to variation in discourse
organization (e.g., Dudley-Evans, 1994; Young, 1994). After examining the topical
structure of a few lectures, Hansen (1994) concluded that topic shifts can be identified most
reliably through tracing those discourse markers that signal topic shifts; others have
looked for discourse markers and their functional variants in a corpus of lectures
(Camiciottoli Crawford, 2004). Although these studies investigate university classroom
discourse from different perspectives, nearly all of them focus on lexical items and their
functions, while neglecting individual grammatical features or a group of grammatical
features and their functional correlates. Additionally, they mostly investigate teacher-led
monologic class sessions.
In order to provide comprehensive descriptions of linguistic variation in university
classrooms, computational methods and quantitative analyses are needed to carry out
large-scale investigations. Corpus-based approaches to research provide this framework.
They explore linguistic variation among spoken and written registers in the academic
context (Biber, 2006) and have successfully described patterns of co-occurring linguistic
features in university settings in general (e.g., Biber, Conrad, Reppen, Byrd, & Helt, 2002;
Biber et al., 2004) and in the classroom in particular (Csomay, 2005). Supported by this
methodology, other studies also show that North American university classes display
linguistic features of both face-to-face conversation and written academic prose (Csomay,
2000, 2006).
Investigating class sessions further, corpus-based studies include both monologic and
interactive types of classes and report on the relationship between language use and
degrees of interactivity. For example, Csomay (2002) tracks varying patterns of turn taking
E. Csomay / Journal of English for Academic Purposes 6 (2007) 336–355 337in class sessions to identify degrees of interactivity and describes linguistic variation in