Learning a Second Language

We present a system, called IMMIGRANT, which learns rules about the grammar of a second language from instructions. We explore the implications of this task on the representation of linguistic knowledge in a natural language understanding system. We conclude that the internal representation of linguistic knowledge used in IM- MIGRANT, which is unification-based, is more amenable to language learning from instructions than other representation schemes.

Authors

• Steven L. Lyffnen Carol E. Moon

Keywords

No keywords are indexed for this paper.