Translation Quality Assessment: An Argumentation-Centred Approach

Appendix One: Model assessment

APPENDIX ONE MODEL ASSESSMENT To givethe reader a comprehensive picture ofthe results of an ARTRAQ assessment, I show how an evaluator might mark up the translation of the passage and complete the grids for text 2 (see 5.2.1). TARGET TEXT (Para. 3)The advantage of the previous (T—mistranslation) project is that it provides estimates for the distance travelled that fluctuate with the years instead of being considered constant values, as it (L—usage) seems tobe the case currently with this model. However, the data is not available according to the exact age of the vehicles, but rather according to age groups: 2years and under, 3-5years,6-8 years, 9years and over. The possibility, by using Bayesian analysis tools developed by Ms Nathalie Boucher, (L—syntax) of producing series equivalent to (T—critical defect: grounds of argument schema misconstrued) the provincial or regional scale (Maritimes, Quebec, Ontario, Prairies, British Columbia) should be studied for the next agreement. (Para. 4) In its review of input variables, the OEE indicates that the CVS (1999-) could constitute a new source of data for this variable. The survey is supposed, in effect (T—mistranslation), to be an important source of data for estimating distance travelled. We (T—major defect bordering on critical: author's narrative strategy misconstrued, undermining force of grounds and claim) note that the survey covers more than the requirements (T—mistranslation) for the present variable and consequently is also assumed to be a good sourcefor the variableestHTrkPVDT (EstimatedHeavy Truck per Vehicle Distance Travelled). The CVSsurvey was designed to estimate distances travelled by (T—omission) various categoriesof road vehicles, including light, medium and heavy trucks. In the 154 TRANSLATION QUALITY ASSESSMENT travel logbook for light vehicles (carsand trucks), we (T—narrative strategy defect repeated) asked our (defect repeated) respondent to specify vehicle use for each trip (question 7ofthe 2000 version) and one ofthe uses on the list refers to work use ("Driving as part of the job"). Distance estimates for light trucks can therefore be segmented according to the type of use (personal or commercial), making it possible to specifically (L—redundant) target commercial distance for the purposes of the present (L—usage) variable. The log book for higher mass (L—terminology) trucks (medium and heavy), used by default for commercial purposes, also shows the distance travelled for each of the vehicles selected for each of their trips, during the survey period. The two weight categories of the vehicles retained (L—gallicism) for sample stratification, 10 000-33 000 Ibs and over 33 000 Ibs, then make it possible to produce separate estimates for medium trucks (present variable) and heavy trucks (variable estHTrkPVDT). An estimate of the total commercial distance for each of the two categories of trucks can thus be obtained. ARTRAQ Grid Core Parameters Element Argument schema Prepositional functions/ conjunctives/other inference indicators Arguments Narrative strategy Translation assessment Grounds and claim inaccuratelyrendered Accurately rendered overall Accurately rendered overall Inaccurately rendered Since the translation is for information purposes, the evaluator will apply the information standard. The above grid summarizes the results of the evaluator's work in assessing how well the translator has preserved the content and APPENDIX ONE 155 structure of the argumentation featuresin ST. This part ofthe ARTRAQ process will have served to identify most, if not all, weaknesses in transfer of message. In selecting parameters for the weighted ARTRAQ grid, the evaluator will take into account the fact that the subject is specialized and based on scientific logic and that the translation is primarily for informationpurposes. Accordingly,the completed grid might look like this: Weighted ARTRAQ Grid Parameter Argument schema Propositional functions/ conjunctives/ other inference indicators Narrative strategy Typography Terminology Total Weight (/10) 3 2 2 1 2 10 Minimum Requirement (10)30 (8)16 (6)12 (6)6 (8)16 80 Quality (/10) 0 8 0 10 8 Rationale Misinterpretation of schema elements One error Narrative strategy misconstrued No errors One error Score (/100) 0 16 0 10 16 42% Given the weightings assigned to the various parameters, a scoreof 80% (30% for argument schema and 50% for selected parameters) would constitute the minimum requirement for this translation to meet the information standard. In the end,the evaluator finds it to be substandard on two counts: (1)failure to preserve the argument schema fully (sufficient in itself to warrant a substandard rating),and (2)failure to preserve the narrativestrategy. 156 TRANSLATION QUALITYASSESSMENT...