Using Finite State Transducers for Helping Foreign Language Learning

Abstract

The interest and demand to foreign language learning are increased tremendously along with the globalization and freedom of movement in the world. Today, the technological developments allow the creation of supportive materials for foreign language learners. However, the language acquisition between languages with high typological differences still poses challenges for this area and the learning task it self. This paper introduces our preliminary study for building an educational application to help foreign language learning between Turkish and English. The paper presents the use of finite state technology for building a Turkish word synthesis system (which allows to choose word-related features among predefined grammatical affix categories such as tense, modality and polarity etc...) and a wordlevel translation system between the languages in focus. The developed system is observed to outperform the popular online translation systems for word-level translation in terms of grammatically correct outputs.