Abstract

The name of Yehoshua Bar Hillel is known to every student of machine translation. The late philosopher richly deserves this recognition: he had been a central figure in the early development of the field and contributed what should be considered the first set of sober assessments for MT at the time of the widespread gung-ho attitude to its prospects. The fame is probably due to the correctness of Bar Hillel’s forecasts of the field’s future stumbling blocks. Still, what is generally remembered of Bar Hillel’s contributions to MT is but a small fraction of his ideas and opinions on the subject. There is an unsettling tendency among my colleagues in MT, natural language processing, linguistics and AI tacitly to assume that their predecessors in the field, not having at their disposal either the latest machines or the latest theories, were somehow naive and “incomplete.” History of the subject starts for them with the dissertation work of their thesis advisor. For such people, Bar Hillel may only have the distinction of being the earliest widely quoted author in MT. In reality, reading Bar Hillel can be very instructive for today’s scholars, especially as he excelled in recognizing and assessing the intellectual evolution of entire fields (philosophy, logic, linguistics, MT) and pointing out lacunae in the applicability of their findings.

Citations