Translation error type in medical and legal student translations: impact of dictionary, CAT tool and corpus use

Abstract This paper complements a previous study on the impact of the use of monolingual original corpora in the target language consisting of texts produced by native speakers (hereafter MOC) on translation quality (error rate) (cf. Lambrechts & Verplaetse, forthcoming). This earlier study showed a limited positive effect of MOC in combination with other resources such as translation memories (TMs) on student translations of business and legal text fragments. For the current study a total of 20 students took part in two translation experiments. The students were asked to translate two text fragments from English into Dutch in the medical domain (6 students) and the legal domain (14 students) using various translation aids, viz. a bilingual dictionary, a specialized TM and a MOC compiled by the researchers. Error typologies were used in order to determine the type and the number of adequacy errors (errors relating to the transfer between source and target text) and acceptability errors (target text errors, independent of the source text). In the current study we found that MOC use mostly leads to a decrease in the number of acceptability errors, confirming the findings of the previous study. A high number of MOC consultations decreases the error rate to some extent in the legal translations, but no considerable positive effect was seen in the smaller group with medical translations. This may indicate that text type may influence the effective use and impact of MOC to decrease the error rate. In addition, also MOC size and MOC quality may influence the effective use of MOC and its impact on translation quality. In addition to the primary aim of gauging the impact of the use of MOC on translation quality in terms of error rate for our selected domains and text types, the current paper also wants to showcase a methodology for both translation training practice as well as replication analyses with other language pairs, other domains (and hence) MOCs, and different translation student populations, if available for specialized domains.

Jaar van publicatie:2019

Toegankelijkheid:Open