Applying Tools and Techniques of Natural Language Processing to the Creation of Resources for Less Commonly Taught Languages

Abstract

This paper proposes that research results from the area of natural
language processing could effectively be applied to creating software
to facilitate the development oflanguage learning materials for
anynaturallanguage. We will suggest that a knowledge-elicitation
system called Boas, which was originally created to support a
machine-translation application, could be modified to support
language-learning ends. Boas leads a speaker of any natural Ianguage,
who is not necessarily trained in linguistics, through a series
of pedagogically-supported questionnaires, the responses to which
constitute a" profile" of the language. This profile includes morphological,
lexical and syntactic information. Once this structured
profile is created, it can feed into virtually any type of system,
including one to support language learning. Creating languagelearning
software using a system like this would be efficient in two
ways: first, it would exploit extant cutting-edge research and technologies
in naturallanguage processin~ and second, it would permit
a single tool to be used for all languages, including less commonly
taught ones, for which limited funding for resource development is
a bottleneck.