Conhecimento experto em tradução: orientação e revisão em tarefas tradutórias executadas por pesquisadores expertos não-tradutores

Abstract

This thesis reports on an ongoing project EXPERT@ Expert knowledge in translation: modeling peak performance (CNPq 479340/2006-4), developed at LETRA (Laboratory for Experimentation in Translation), at Faculdade de Letras, UFMG. It approaches the initialorientation and end revision phases, as well as online orientation and online revision pauses (and respective external and internal supports) from data pertaining to two translation tasks carried out under experimental conditions by four medicine expert researchers. Drawing onthe data collected and partially analyzed by Silva and Pagano (2007), this study probes the impact of domain knowledge (SCARDAMALIA; BEREITER, 1991) upon orientation (MACHADO; ALVES, 2007) and revision (BATISTA; ALVES, 2007) in the performance by these expert non-translators for the accomplishment of a translation task. Data collection, as described in detail in Silva & Pagano, consisted of: (i) a semi-structured interview, (ii) two translation tasks of research article introductions from Portuguese into English (one of them from a domain-specific text and the other from an adjacent domain), and (iii) retrospective recall protocols collected right after the completion of each translation task. The tasks were recorded by means of a keylogging software (Translog©) and a screen-logging software (Camtasia©), complemented with observational notes by the researchers conducting the experiment (PACTE, 2005). Data analysis is partly based on the methodology in Machado and Alves (2007) and Batista and Alves (2007), working on pauses equal to or longer than five seconds in Translog© log files. The dependent variables in the study, related to the notionof durability (ALVES; GONÇALVES, 2007), were: pauses (number of occurrences, time duration per occurrence, and total time duration, alongside means and standard-deviation, as well as maximal and minimum values) (ALVES, 2005), recursiveness (BUCHWEITZ;ALVES, 2006) correlated to pauses (number of occurrences due to spell-checking, attempted and discarded solutions, and revision of implemented solutions), and examination of the types of external and/or internal support resorted to for each pause. The results point out expert nontranslators' characteristics that corroborate issues discussed in Silva and Pagano (2007) concerning domain knowledge impact on the number of occurrences and the time duration of online orientation pauses. The results also suggest tendencies among these subjects that are comparable to the subjects investigated in both Machado and Alves (2007) and Batista and Alves (2007), those being related to the phases of the translation process and, more specifically, to the internal and/or external supports correlated to each online orientation and online revision pause. Moreover, for the expert non-translators under scrutiny recursiveness (especially due to spellchecking or prompt substitution of attempted solutions) was proved to correlate to orientation (rather than revision) pauses found in the translation process. Some ofthe findingsparticularly time spent for initial orientation, the variety of internal/external supports related to online orientation pauses, and the number of pauses in the end revision phasesingle out one the subjects in the study, also singled out in Silva and Pagano (2007),suggesting a more durable task and a higher level of expert performance on his part.