METM09 presentation Thread: Original research
Arabic and computer-aided translation: linguistic problems and a novel solution
Barbara Quaranta – Campobasso, Italy
Interest in Arabic is increasing due to current economic and geopolitical events. Although computer-aided translation (CAT) tools may simplify the translation process, they are not commonly used by translators working with Arabic. Moreover, since CAT tools were initially developed for Western languages, they are not completely compatible with Arabic.
This presentation draws on a 3-year study of the use of CAT tools for Arabic. Two sales contracts in Arabic were experimentally translated using the SDL Trados 2007 CAT tool suite. First, a generic sales contract form was translated in order to experiment with alignment, creation of a translation memory and term database (termbase), and setting of segmentation rules. Then, an actual stipulated contract was translated using the previously created termbase (glossary) and translation memory.
Two kinds of problems were found during translation. One relates to formatting differences between Arabic and Western languages, including font type and size and punctuation. The other concerns certain grammatical and morphological features of Arabic, particularly the use of prefixes and suffixes that prevent the software from recognising terms and segments. With this knowledge and based on morphological analysis of terms in the contracts (performed by an Egyptian researcher), a computer application was developed in Microsoft Access to solve the most limiting problems in term recognition; the application is compatible with all CAT tools that support the creation of termbases from .xls and .xml files.
This presentation will explain the distinctive features of Arabic that hinder the use of CAT tools, illustrate the problems encountered during translation using SDL Trados, and show how the new application improves the performance of this software. This study, although at a preliminary stage, suggests that the development of CAT tools for this language should involve both morphological analysis and statistical survey of Arabic corpora to improve term recognition.
Barbara Quaranta has a degree in translation and interpretation for specific purposes and an MA in technical and scientific translation in English and Arabic from -œS. Pio V- University in Rome. She has been studying the use of CAT tools with Arabic for three years and, at Cairo University, conducted research on how morphological analysis can help optimize software for computer-aided translation of Arabic. For further information, please visit the Web site www.barbaraquaranta.it.
