A corpus-based study of the verb observar in English-Spanish translations of biomedical research articles
Ian A. Williams
Universidad de Cantabria, Spain
This paper describes an empirical contextual study of the English verbs and syntactic resources translated by observar in an extensive corpus of biomedical research articles. Quantitative analysis showed that the frequency of observar was significantly higher in the Spanish translations than in the comparable Spanish language original articles (360 vs. 162 instances; P < 0·001). Qualitative analysis of the Spanish native texts provided a linguistic profile for the verb. This profile was then used in a contextual study to assess the appropriateness of the 360 instances found in the translated texts. The results indicate that observar is only a natural translation equivalent for ‘observe’. For other verbs, an awareness both of the rhetorical, syntactic and collocational restrictions and of the range of alternative choices will allow translators to select appropriate lexical items and avoid excessive repetition of observar, thus creating a more varied target text.
In the last decade, increasing use has been made of corpora for studying the phenomena of translation and translation behaviour. A number of analysts have indicated that a corpus-based approach is suitable for investigating what might be universals of translation (Baker 1995, 1996; Laviosa 2002). Simplification, explicitation and normalisation have been proposed as possible universals that could be tested for in corpora. Although large general translation corpora will throw light on these theoretical questions, the compilation of specialised corpora consisting of both a parallel component—i.e., pairs of source language (SL) and target language (TL) texts—and a comparable component—i.e., similar native language [ p. 86 ](NL) texts in the TL—can provide valuable quantitative data and qualitative information on specific registers and genres for trainee and practising translators in specialised fields (Williams 2004; 2006). With this corpus design, analysis of the NL texts establishes the TL norms and reference values for the linguistic variables under study, comparison of the SL and TL texts informs of the behaviour of practising translators, and comparison of the TL texts with the comparable NL texts brings out both typical and idiosyncratic features of the translation products.