Predicting and Using a Pragmatic Component of Lexical Aspect of Simple Past Verbal Tenses for Improving English-to-French Machine Translation

Linguistic Issues in Language Technology – LiLT Predicting and Using a Pragmatic Component of Lexical Aspect of Simple Past Verbal Tenses for Improving English-to-French Machine Translation Sharid Loáiciga & Cristina Grisot Published by CSLI Publications LiLT volume 13, issue 3 August 2016 Predicting and Using a Pragmatic Component of Lexical Aspect of Simple Past Verbal Tenses for Improving English-to-French Machine Translation Sharid Loáiciga, University of Geneva & Cristina Grisot, University of Geneva & University of Neuchâtel Abstract This paper proposes a method for improving the results of a statistical Machine Translation system using boundedness, a pragmatic component of the verbal phrase’s lexical aspect. First, the paper presents manual and automatic annotation experiments for lexical aspect in English- French parallel corpora. It will be shown that this aspectual property is identified and classified with ease both by humans and by automatic systems. Second, Statistical Machine Translation experiments using the boundedness annotations are presented. These experiments show that the information regarding lexical aspect is useful to improve the output of a Machine Translation system in terms of better choices of verbal tenses in the target language, as well as better lexical choices. Ulti- mately, this work aims at providing a method for the automatic an- notation of data with boundedness information and at contributing to Machine Translation by taking into account linguistic data. 1 LiLT Volume 13, Issue 3, August 2016. Predicting and Using a Pragmatic Component of Lexical Aspect. Copyright c ⃝ 2016, CSLI Publications. 2 / LiLT volume 13, issue 3 August 2016 1 Introduction This paper aims at improving the results of Statistical Machine Trans- lation (SMT) systems with respect to verbal tenses. Verbal tenses are the primary linguistic source of temporal reference, i.e. the localization of events and states in time. Verbal tenses express temporal location of eventualities1 with respect to the moment of speech and with respect to each other. For example, the past perfect form in sentence (1) locates the two eventualities before S (‘moment of speech’), hence in the past, and the eventuality of hard working before the moment when another eventuality, the son’s disappearance, occurs. (1) He had worked hard the day when his son disappeared. The choice of the verbal tense depends on fine-grained temporal interpretations of the utterance in context. In this paper, we show the translation of verbal tenses can be improved at the sentence-level when we make use of a pragmatic component of the verbal phrase’s lexical aspect, i.e. boundedness, and the way in which it is perceived by human speakers. We have investigated the usefulness of boundedness for correctly translating the English Simple Past (SP) into French using a SMT sys- tem. The four most frequently used translations of the SP in French are: passé composé (PC), imparfait (IMP), passé simple (PS) and présent (PRES). Loáiciga et al. (2014) showed that this mapping has a skewed distribution in favour of the PC translation. Since SMT systems favour the most frequent translation and make little use of context or mean- …

Content truncated to 500 words