ONLINE TRANSLATION OF PROVERBS BETWEEN AVAILABILITY AND ACCURACY Sami Abdullah Hamdi, English Language Center, Jazan University, Jazan, Kingdom Of Saudi Arabia Shamdi@Jazanu.Edu.Sa Kazuhiko Nakae, Department of Foreign Languages, Kansai Gaidai University, Osaka, Japan Nakae@Kansaigaidai.Ac.Jp Mohamed Okasha, English Language Center, Jazan University, Jazan, Kingdom Of Saudi Arabia Mohamed_Okasha20@Yahoo.Com Abstract The use of proverbs in our daily communication, whether spoken or written, becomes almost a natural tendency. Even when communicating with others who speak different languages people try to converse with others using translated or equivalent proverbs in their communication. Online machine translation systems/ tools provided by many companies and institutions have been useful means to translate proverbs as well as phrases and sentences. Although the usefulness of proverbs is clear and useful in linguistic communication, the accuracy of translated proverbs and how to judge accuracy of the translated proverbs remains a critical issue. Some translated proverbs are semantically accurate and understandable but syntactically inaccurate. Others are syntactically accurate but semantically inaccurate, that is to say the truly intended meanings cannot be understandable. It is an ideal that we can retrieve the equivalent proverbs suitable in the specific contexts in other languages through online machine translation but it is almost infeasible for the present stage. The target of this research is to examine the accuracy of translated proverbs and to evaluate a set of the most popular online machine translation tools. This kind of research method will lead to the deeper understanding of the relative distance between two cultures in which the two languages are spoken. In this paper we will deal with the translation from English into Arabic, because these two languages frequently use proverbs in the daily conversation and writings. Keywords: Online machine translation; proverbs translation; English; Arabic I. Introduction The evolution of machine translation (MT) in the 1950s impacted the study of language and opened a new window to translate languages via machine which led to the emergence of computational linguistics as a new field of study. Generally speaking, human translators were engaged in MT process [5] through editing assistance. Large companies and organizations start working on the primary translation systems with extensive post editing as the quality of the outcomes wasn't that satisfactory. The MT process has been facilitated later by translation aids such as computational database of terminology, word processors, and translation memories which improved MT task in terms of speed and quality. Technically speaking although there have been various computational paradigms to perform MT task, most of MT systems (especially commercial ones) follow rule based MT systems which employ [2] rules derived from linguistic knowledge. However, phrase based statistical machine translation (PB-SMT) becomes the state of the art approach in MT research which uses phrasal and lexical alignments via statistical models working on the parallel corpus. With the increasing needs to translate languages through machine, individuals start becoming interested in exploiting this service which was restricted to big organizations at first. In 1990s online MT came to the scene as a chargeable service where some companies initiated this like Systran and Fujitsu then other companies followed. However, the emergence of Bablefish in 1997 as a free online MT tool achieved a notable popularity among users and brought MT to the next level. This popularity of online MT helped raising the awareness of the public encouraging them to have online MT tools in their PCs. What distinguishes online MT and attracts users is that it is free and