Modal Markers as Potential Sources of Distortion in Translated Medical Abstracts

Abstract

which has for English title "Sources of distortion in translated systematic review abstracts. A comparison of conventional human translation and postedited machine translation."

Post-Editing Neural MT in Medical LSP: Lexico-Grammatical Patterns and Distortion in the Communication of Specialized Knowledge

“…Table 1 presents token counts for each sub-corpus as well as the translation coefficient, i.e., the token count ratio, for the two French-language sub-corpora. The coefficient is somewhat higher than the usually quoted expansion rate for this language pair, i.e., 15-20% [22] (p. 159), but similar to what was previously observed for these texts [19].…”

“…For positive lexico-grammatical patterns signaling potential effectiveness of interventions, the distortion in translation arises specifically from more affirmative translation choices for certain markers occurring within these patterns, mainly modal auxiliaries and evidential verbs, a discussion of which can be found in [19]. For example, in the context of a positive lexico-grammatical pattern where potential effectiveness is signaled through the modal auxiliary may (i.e., intervention may be effective), a typical biased translation in French consists in using the verb pouvoir in the indicative mood (peut), expressing inherent potential, instead of the more hypothetical conditional mood (pourrait) [19]. The potential for positive bias due to the use of the indicative form in this context was also highlighted by some of the domain experts having reviewed previous translations of Cochrane abstracts, one of whom even stated that the indicative mood was "never" an adequate translation for the auxiliary may(Original comment in French: 'pourrait et jamais peut'.…”

“…They have a high number of variable lexical parameters (i.e., interventions, criteria of effectiveness) and are therefore not easily constrained for searching in corpus. These patterns are, however, built around specific modal markers, some of which were previously observed to be subject to biased translation [19]. These markers were used as anchor points in searching for positive patterns.…”

Representation of interactional metadiscourse in translated and native English: A corpus-assisted study

“…Translation trainees have also been found to omit or modify hedges more frequently in their translations of English newspaper commentaries [ 39 ] and to use fewer stance markers in their translations compared to non-translated texts [ 40 ]. In addition, modal markers with a hedging reference are underrepresented in medical abstracts translated from English into French [ 41 ]. However, not all studies have found a significant difference in the use of metadiscourse markers between translated and non-translated texts.…”