Multilingual Grammar-Based Language Training: Computational Methods and Tools

Abstract

Living in the communication era we face an ever-increasing demand for communication. Considering the human languages role as the major means of communication, has inspired the author to do this thesis work, which we believe is an effort towards solving the challenge of human communication. The main goal of this thesis is introducing a teaching/learning aid - which we have called the GF Translation Quiz or in short the Quiz - to be used in training human languages in general and specifically targeting their grammatical aspects. The Quiz is designed as a web application, that makes it also available on mobile phone platforms - such as iPhone and Android. It may be applied either as an aid in a language learning methodology or as a stand alone exercise tool for self training. The Quiz is intended to help teachers and education supervisors by reducing the burden of creating exercises and exam questions, and at the same time help learners by providing them with instructive feedback. About the background many efforts have been done in this mutual area between computer science and linguistics, among which we have focused on GF (Grammatical Framework) and its unique conceptual approach towards translation and multilinguality. Thus this thesis work relies heavily on GF from many aspects, i.e. most characteristics of the Quiz application, i.e. grammatical precision, support for multilinguality, and coverage of natural languages in small fragments are provided by GF. From another point of view the Quiz application works as an interface that makes GF capabilities available to ordinary non-specialist users.