Abstract
Most existing corpus-based analyses focusing on the contrasts between source texts and translations are either purely lexical or are constrained by sentence boundaries (e.g. Johansson and Hofland 2000, Schmied and Fink 2000, Frankenberg-Garcia 2005, Johansson 2007, Pérez Blanco 2009 and Saldanha 2011). Probably the main reason underlying this limitation is the fact that, in parallel corpora, texts are usually segmented at the level of the sentence due to the relative ease with which sentence boundaries can be identified automatically. The alignment of source texts and translations is then usually carried out such that whenever there is not a one-to-one correspondence between source-text and translation segments, they are aligned either on a one-to-many or on a many-to-one basis, blurring out the details of what happens when source-text sentence boundaries are changed in translation. The aim of the present exploratory study is to take a closer look at sentences that translators split apart and join together. Using COMPARA, a parallel, bidirectional corpus of Portuguese and English fiction of 3 million words whose alignment was manually post-edited so as to differentiate between one-to-one, one-to-many, one-to-part and one-to-zero alignment, it was possible to analyse from a quantitative perspective over 90 thousand source-text sentences and their corresponding text in translation. The idea was to find out more about the extent to which sentences are split apart and joined together, whether sentence splitting is more common than sentence joining, and whether translation direction can affect this. A closer look was also taken at a sample of over one thousand segments involving sentence joining and splitting, so as to find out more about how translators in both language directions go about splitting and joining sentences. From a theoretical perspective, the analysis supports the ideas of explicitation and normalization in both Portuguese to English and English to Portuguese translation. From a practical perspective, there are implications for translator education and the development of CAT tools. References Frankenberg-Garcia, A. (2005) ‘A corpus-based study of loan words in original and translated texts’. In P. Danielsson & M. Wagenmakers (eds.) Proceedings of the Corpus Linguistics 2005 conference, Birmingham, UK, 14-17 July. Available online at http://www.corpus.bham.ac.uk/ pclc/index.shtml. Johansson, S. 2007. Seeing through Multilingual Corpora: On the use of corpora in contrastive studies. Amsterdam & Philadelphia. John Benjamins. Johansson, S. and Hofland, K. 2000. “The English-Norwegian Parallel Corpus: current work and new directions”. In In P. Botley, T. McEnery and A. Wilson (Eds.), 134-147. Pérez Blanco, M. (2009) ‘Translating stance adverbials from English into Spanish: a corpus- based study’. International Journal of Translation, 21, 41-55. Saldanha, G. 2011. “Translator style. Methodological considerations”. The Translator, 17(1), 25-50. Schmied, J. and Fink, B. 2000. “Corpus-based contrastive lexicology: the case of with and its German equivalents”. In P. Botley, T. McEnery and A. Wilson (Eds.), 157-176.