Machine Translation Conference
Roger M. Needham · ITNOW · 1961
The first International Conference of Machine Translation of Languages and Applied Language Analysis was held at the National Physical Laboratory from 5–8 September. Papers were presented from a large number of countries, and the Proceedings will give an excellent overall view of the state of this research. Many fewer people now say that MT is on the edge of solution than was the case a few years ago. There is a progression (which is sometimes denied by those who make it) from hopes of almost immediately usable results to a realisation that before MT can be done properly an immense amount has got to be found out about language and languages. As far as machinable operations are concerned, the the facts are just not there in the dictionaries and grammar books, which were made for use by human beings with a good deal of background knowledge. There is controversy in the field as to whether the missing facts are best to be found by more thorough dictionary making based, essentially upon people writing down what they know of how words are used, in some suitable standardised manner, or whether, alternatively, the solution will be found through the use of machines to examine, by scanning vast corpora of text, how the words are in fact used. The objection to the former course is based on possible non-objectivity, to the latter in doubtful practicability on a sufficiently large scale. The former approach was exemplified in the papers of M. Masterman, and of S. Ceccato and his collaborators; the latter is more popular in the USA, and was represented by the papers of M. Zarechnak, S. Chatman, K. Harper and S. Lamb.