NATURAL-LANGUAGE PROCESSING BY COMPUTERS-1966.
Abstract
The eventual goal of language-processing research is the development of highly sophisticated question-answering systems. These would accept natural-language questions, search a multilingual library for pertinent factual text, translate and generate from that widely separated text essays that answered the question at any desired level of detail. Examples of existing question answerers, translations, essay writers, etc., are presented. These examples are evaluated as highly inventive linguistic gadgetry that is still far from practical usefulness. Linguistic, semantic, and psychological problems both of theory and practice must be solved before this gadgetry results in useful systems. Specific difficulties facing researchers in these areas are described. The paper concludes that the strongest existing machine capability with language is that of counting and indexing words of a text. This capability has already made the computer an indispensible partner in concordancing, information retrieval, linguistic analysis, and stylistic and context analysis. Although fascinating linguistic gadgetry is available on computers, the solid contribution so far lies in this well-developed capability of word counting and indexing. (Author) (For presentation at the Annual Meeting of the American Association for the Advancement of Science, Berkeley, Calif., Dec. 26-31, 1965)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1965
- Accession Number
- AD0625005
Entities
People
- R. F. Simmons
Organizations
- System Development Corporation