This thesis approaches the problem of cross-linguistic authorship attribution in the context of written English by native Japanese authors. To this end, a hybrid qualitative and quantitative...Show moreThis thesis approaches the problem of cross-linguistic authorship attribution in the context of written English by native Japanese authors. To this end, a hybrid qualitative and quantitative analysis of a learner corpus is conducted, making use of a novel automatic error classifier ERAA. It was found that in a comparison with authors from nine other Asian countries, language errors that native Japanese authors tend to produce in written English that were found in earlier literature appear to be unique to the Japanese authors within a given writing theme. Following statistical analysis, it is furthermore suggested that there exist types of L2 language errors that are unaffected by the surrounding writing theme, while other types are affected. Finally, it is concluded that the employed hybrid analysis method is viable to efficiently conduct large-scale corpus research to study a variety of L2 linguistic traits.Show less
