Symbolic Systems Forum - Martin Kay, Linguistics

Martin Kay, Professor of Linguistics

"Machine Translation: A new frontier?"

ABSTRACT:

Serious research on machine translation began in the 1950's but was

brought almost to a stop towards the end of the 1960's by a critical

government report. It strongly advocated the view that practical

systems would require a great deal more work on fundamental issues in

language and computing, and it was here that the investment should be

made. This was probably the key event in bringing the field of

computational linguistics into being. Today, machine translation has

again become a concentrator of effort and a generator of excitement,

but not because of advances in computational linguistics. Quite the

contrary. Leaders of the field have come to the view, the belief,

nay, the realization, that an understanding of language and

linguistics is irrelevant, if not an impediment, to success in

machine translation research. This view is widespread, but not

universally held. As one who was there in the 1950's and is arguably

still there today, I will give you my two cents worth.