Investigating the Possibility of a Microprocessor-Based Machine Translation System

This paper describes an on-going research project being carried out by staff and students at the Centre for Computational Linguistics to examine the feasibility of Machine Translation (MT) in a microprocessor environment. These incorporates as far as possible features of large-scale MT systems that have proved desirable or effective: it is multilingual, algorithms and data are strictly separated, and the system is highly modular. Problems of terminological polysemy and syntactic complexity are reduced via the notions of controlled vocabulary and restricted syntax. Given these constraints, it seems feasible to achieve translation via an 'interlingua', avoiding any language-pair oriented 'transfer' stage. The paper concentrates on a description of the separates modules in the translation process as they are currently envisaged, and details some of the problems specific to the microprocessor-based approach to MT have so far come to light. This article is from 'Proceedings of the Conference on Applied Natural Language Processing Held at Santa Monica, California on 1-3 February 1983,' AD-A129 824.