History

Machine translation technology has become an integral part of modern life. Just 50 years ago, in order to read a foreign article or translate a letter received from abroad, people spent a long and meticulous time flipping through dictionaries or turning to professional translators. Today the best way to understand the “secret” meaning of a foreign language is to use an online translator or mobile app.

Behind the accessibility, simplicity and convenience of computer-based translators is a tremendous amount of work done by scientists, mathematicians, engineers and linguists. This article explores how machine translation technology was born and developed from the 17th century to the present day.

andrew-neel-QLqNalPe0RA-unsplash
Background: 17th century
For thousands of years, people have been trying to break down the language barrier, to create a common language for all people, to find a way to learn all languages, or to create a technology that allows people to understand foreign languages without wasting time and effort in learning them. The first ideas date back to the 17th century, when René Descartes and Gottfried Leibniz independently proposed a universal language as a new basis for logical thinking, and for eliminating the mutual misunderstanding that occurs due to illogical languages.
 
In 1668, John Wilkins, a British clergyman, published a treatise, “Experience on the true symbolism and philosophical language”, in which he presented his approach to creating a universal language for scientists and philosophers, which would replace Latin. However, his proposal did not meet with approval among linguists.
 
Much later, in the 19th and 20th century, scientists returned to developing a single international language, which led to the creation of Esperanto. However, the first attempts at machine translation were still decades away.

The 1930s: Trojan's system - the first step towards machine translation

In the 1930s, scientists started talking about developing machine translation technology. Georges Arzrouni, a French scientist of Armenian origin, created a bilingual automatic dictionary on punched tape, while the Soviet engineer Pyotr Trojansky invented “a machine for selecting and typing words when translating from one language to another”.

ranurte-R65mxDXDi4o-unsplash

Trojansky’s invention was a table with a slanted surface and a camera combined with a typewriter. The typewriter keys would encode morphological and grammatical information and its ribbon would feed simultaneously with the photographic film. A movable plate of printed words, called the glossary field, was attached to the surface of the device. The words on the glossary field were accompanied by translations into at least three languages and arranged like letters on a keyboard: the most frequently used ones were closer to the centre of the field. The ideas of Peter the Trojan remained unknown even to scholars for a long time. They were rediscovered only in the 1950s.

yves-junge-kEv-Ls1t9Zg-unsplash
lee-campbell-6njoEbtarec-unsplash

1949: The first computer-based MT developments

The concept of machine translation was not formulated until the late 1940s, by cryptographer Warren Weaver, director of the Rockefeller Foundation’s Natural Sciences Division. In a letter to his colleague Norbert Wiener, he proposed that the task of translating from one language to another should be seen as a new application of decoding technology. Imagine that the original text is in your own language, but encoded with special characters, and all you have to do is crack that code to make sense of the information in the text.

Weaver’s letter caused a great resonance and in 1949 the cryptographer published a memorandum in which he justified the possibility of implementing machine translation technology based on decoding. This document became a milestone in the development of machine translation. In it, the scientist described the concept of interlingva, in which the process of information transfer is carried out in two stages:
 
– the original sentence is translated into a simplified version of English (the interlingual language);
 
– the result is converted into a sentence in the target language.

1950s and 1960s: the Georgetown experiment and the ALPAC report

An important event in the development of MP was the Georgetown Experiment which took place on 7 January 1954 in New York, at the headquarters of IBM Corporation. As part of this experiment, the IBM 701 computer translated 60 sentences from Russian into English for the first time in the world.

Shopping Basket