Viviane Possamai(University of Birmingham)
Sintagmatic associations in Medical articles and the development of a corpus-based tool to help translators
Our research is aimed at building a corpus-based tool to help translators, especially beginners, with the translation of Medical articles from Portuguese into English. The tool uses a database of previously identified associations and whenever a new text with similar associations is introduced, the tool suggests possible translations to those text extracts, retrieved from an aligned corpus. In order to build such database we need to identify and extract these associations beforehand from a corpus. To that end, we have used a corpus of Medical articles, in Portuguese, that was divided into four subcorpora that correspond to the sections of articles as it follows: Introduction (52,112 tokens), Methods (85,213 tokens), Results (85,822 tokens), Discussion (132,724 tokens). By using WordSmith Tools to make wordlists, calculate keyness and collocations we were able to find node words for each session. Currently we are working on a methodology to identify associations to those node words as well as on defining the nature of the event we are focusing on, where we are not completely sure they can be considered as collocations. Our research also draws attention to the fact that the specialized character of technical and scientific texts is due to the associations between all lexical items as a whole, and not only by some of them known as technical terms. We consider this to be a contribution from the point of view of translators to the fields of Terminology and Terminography.