Controlled English for Knowledge-Based MT: Experience with the KANT System

In this paper, we describe the design and deployment of KANT Controlled English (KCE) for knowledge-based machine translation in the KANT system. KCE combines three kinds of constraints: constraints on the lexicon; constraints on the complexity of sentences; and the use of generalized markup language. We describe how each of these types of language control are utilized in the implementation of a typical KANT application. The principles described are not specific to knowledge-based MT, and can be applied in the design of controlled languages for any kind of MT application. 1 Introduction In recent years, more emphasis has been given to the usefulness of controlled source languages in MT (cf. Adriaens and Schreurs, 1992 and the references cited there). In this paper, we focus on the use of controlled input languages in the KANT translation system (Mitamura, et al., 1991). Controlled English is used to improve the clarity of expression in the source text and to improve the quality of the..