Pizza Evening Translation

Pizza Evening Translation
It was unusually quiet at the pizza evening today, although this does happen sometimes at this time of the year. There were just the five of us: Alex and Pauline, Anne-Marie and Patrick and myself. Conversation stagnated after a while until Alex started expounding on the virtues of his iPhone and, somewhat unconnected, the prospects for machine translation of natural languages. He ignited my long but now-dormant experience with computer technology and languages.

First things first. Alex had downloaded an edition of The Economist onto his iPhone and I had to admit that the reproduction was impressive. However, some problems were apparent. The amount of text visible at any one time was small and so you probably wouldn't want to read a long article on it. Secondly, illustrations in which detail is significant are a problem. Large tables, for instance, can be seen only part at a time or appear in almost invisible font size. The problem is absolute, for the moment at least. If the device has to fit comfortably into a pocket, the screen size has to be small; some kind of folding screen would be conceivable but would make the device very thick and cumbersome. The iPhone provides an impressive compromise but a compromise nonetheless. I've noticed the same with friends Steve and Jo's Kindle eBooks. They are fine for reading novels (or other text) and excellent as an alternative to packing several such books into a suitcase; but inadequate for many illustrations and very awkward if you need to check cross-references or footnotes. The real answer, at least to the illustration problem, would be some kind of holographic facility (not inconceivable in the future).

Machine translation of natural languages is another kettle of fish: Babel fish in fact. I've never seen a decent machine-generated translation and am aware of the many problems. Turing's test has yet to be passed by a long way. Turing's test, incidentally, was that a human being should be able to have a conversation with a machine without being aware that the other party was a machine. Alex said his boss was excited by the advances in machine generated translations for intelligence (defence) purposes. Admittedly, if a few key words were all that was important in the translation, a machine-generated one might do; anything requiring appreciation of subtle wording, tone, etc, would fail.

However, the discussion turned my mind to the army of interpreters in Brussels and the UN and what had happened in IT in the 1960s and 1970s. In IT, language incompatibility was already a problem by then. Broadly, if you had 4 machines and 4 languages, you needed 16 interpreters/compilers to bridge between them. The idea came about to produce an intermediate language, from and into which each language and machine code was translated, which meant you would need only half the number of interpreters/compilers. There were various attempts at such an intermediate language: UNCOL (Universal Computer Language) was one and BCL (Basic Compiler Language) another but they all failed for technical reasons of levels of machine language and operating systems that I won't go into here. The point is that the idea was basically a good one.

So taking that idea and thinking of the armies of interpreters, I thought: why not take English as a universal intermediate natural language? Excellence in English would be required of every interpreter. If, say, a Russian was speaking at a conference, the Russian interpreter would translate into English; every other interpreter would then translate from English into their own native tongue. The delay in simultaneous translation would be imperceptible. The problems of having enough interpreters to translate between, e.g. Russian, Norwegian, Swedish, Spanish, Dutch, etc, would disappear at a stroke and the armies of interpreters and their costs would be decimated. Technically, implementation of this idea would be simple and bring enormous cost savings. It couldn't fail for any of the technical reasons that the analogous IT equivalent had. Indeed, it is already practised in many international commercial situations where interpreters are not available: English is the lingua franca. Diplomatically.......?I wonder what the Académie Française would have to say about it? Incidentally, if machine-generated translations (and Babel fish) are ever to succeed, this is probably the route they will have to take.