ABSTRACT
Although there is a vast body of literature on interlanguage (for an overview, see Ellis 1985, 1994), as well as a burgeoning interest in the assessment of text quality (HampLyons 1991; Adam 1998; Coulthard 1994a; Ferrari and Manzotti 1998; House 1997; Reboul and Moeschler 1998), the evaluation of professional and semi-professional nonnative speaker writings, as published in journals, books and Internet archives, is only now beginning to attract scholarly attention. This chapter aims to make a small contribution to this new area of research; the largest part of the chapter is taken up with what Granger (1998a:12-13) has termed a ‘contrastive interlanguage analysis’, meaning a comparison of native and non-native speaker performance. This kind of analysis is in tune with the broad thrust of research on interlanguage, which recently began to avail itself of purposebuilt learner corpora (Greenbaum 1992; Granger 1993, 1998a, 1998b; Lindner 1994; Granger and Tyson 1996; Ringbom 1996, 1998; Hyland and Milton 1997; Lorenz 1997; Milton 1999).1 While early error analysis focused on learner difficulties in expressing referential meaning, interlanguage research can now give proper consideration to textual and pragmatic aspects as well as to the representation of particular target-language features in non-native discourse. Section 2.1 discusses issues associated with the analysis of interlanguage in greater detail, the main argument being that the advent of large corpora allows us to describe error and deviation in interlanguage with unprecedented precision. This is illustrated through the analysis of textual inadequacies in a sample text by a German writer of English.
