Using the features of translated language to investigate translation expertise: a corpus–based study

Abstract

The study reported on set out to test the hypothesis that linguistic operationalisations of the features of translated language will demonstrate signficant differences in the work of experienced and inexperienced translators. A custom–built comparable English corpus was used, comprising three subcorpora: translations produced by experienced translators, translations by inexperienced translators, and non–translated texts. A number of linguistic operationalisations were selected for three of the features of translated language: explicitation, simplification and normalisation. The differences in these linguistic features in the three subcorpora were analysed, using parametric or non–parametric ANOVA, and T–tests or Mann–Whitney U–tests as post–hoc tests where applicable. The findings of the study indicate substantial (though not unqualified) support for the hypothesis. It is argued that experience–related variation in register sensitivity, language competence, awareness of written language conventions and sensitivity to translation norms are the main factors contributing to expertise.