Skip to main content
Contribution to Book
Error patterns and automatic L1 identification
Approaching Transfer through Text Classification: Explorations in the Detection-based Approach (2012)
  • Yves Bestgen
  • Sylviane Granger, Université catholique de Louvain
  • Jennifer Thewissen
Abstract
The strong version of the Contrastive Analysis Hypothesis rested on the ‘difference equals
difficulty principle.’ That is to say, learner errors can be predicted on the basis of differences
between their mother tongue (L1) and the target language (L2). Although this strong version
has been abandoned, research in error analysis and more recently, computer-aided error
analysis, has demonstrated that a sizable proportion of errors are L1-induced . It would thus seem reasonable to use error patterns as a clue to L1
identification. However, a survey of the literature shows that errors have rarely been used as a
profiling criterion for this purpose, although they are one of the standard variables used in
other types of stylometric studies, notably those that aim to determine learners’ proficiency
level. The aim of this chapter is to
investigate the extent to which error patterns can be relied on for automatic L1 identification. p.p1 {margin: 0.0px 0.0px 0.0px 0.0px; font: 11.5px Times}
In section 1, we describe the corpus data and error
annotating system used. Section 2 gives an outline of the statistical methods employed for the
automatic L1 identification and presents the classification results. In section 3 we tackle the
possible impact of proficiency level differences on the results. Section 4 provides a finegrained
analysis of the most discriminatory error patterns, while Section 5 concludes our
chapter and suggests avenues for future research.
Disciplines
Publication Date
2012
Citation Information
Yves Bestgen, Sylviane Granger and Jennifer Thewissen. "Error patterns and automatic L1 identification" Approaching Transfer through Text Classification: Explorations in the Detection-based Approach (2012) p. 154 - 177
Available at: http://works.bepress.com/yvesbestgen/13/