A BRIEF HISTORY OF MACHINE TRANSLATION I: The early years
The first phase of machine translation began in 1945, the year the dream was born and took shape. That was the year American writer Murry Leinster popularised the idea of a mechanical translator that could help humans communicate with aliens in his science fictions tale, First Contact.
In 1949, American researcher Warren Weaver prepared the first list of proposals – Translation Memorandum – for translation using a computer.
Then in New York in 1954, scientists from IBM and the University of Georgetown demonstrated the world’s first machine translation system, which was able to translate two-and-a-half lines of Russian text into English every second.
This first phase of the dream was followed by disappointment; when experts presented conclusions that were less promising and much closer to reality. In 1960, the Israeli linguist Bar-Hillel concluded machine translation was still a long way off because the equipment was incapable of understanding the context in which a word was being employed. Then in 1966, the US government asked a group of American scientists to prepare a report, which determined that the technology was too expensive and unable to achieve the quality of a human translator. This was to have a profound impact on the sector.
Despite this disappointment, Systran was founded in 1968. Systran is one of the world’s oldest machine translation systems and was for decades used by both the European Commission and the United States Department of Defense.