Page One

Will technology deliver us from the Tower of Babel?

Walter Truett Anderson Pacific News Service
Friday August 10, 2001

SAN FRANCISCO – As the information and communications technology revolution rolls along, we have become accustomed to people talking to their computers, asking for information on electronic data bases, having conversations with robot voices on the telephone. Speech-recognition systems are becoming increasingly sophisticated, and a plethora of new smart gadgets for homes, offices, even cars, are coming on the market. And more are on the way. Despite the current hard times in the high-tech industries, billions are being invested in the development of future speech-recognition devices – many of which most of us could probably do 

without, and some of which will really make the operation of many kinds of machines easier and more efficient. 

And while we get used to people talking to their computers, other new technologies are moving in the direction of making it easier for people to talk to one another – even when they don't speak the same language. 

This is a line of technological development that obviously has great social implications as well as profit potential, and people have been working at it for some time. Various software programs for translating written text into other languages have been on the market for years, although most of them – as disappointed owners have discovered – leave a lot to be desired. 

Translating the spoken word is an even steeper challenge. As anybody knows who has ever read a literal transcript of a conversation, people rarely speak in clear and complete sentences. They repeat phrases, skip over things, change direction, throw words in the general direction of a meaning and hope the listener will understand. And the listener – if he or she speaks the same language with approximately the same accent, and is paying attention – usually does. But it's asking a bit much of a machine, even a smart one, to turn a spoken language into clear sentences and translate that smoothly into another language. The best way devised so far to get simultaneous translation is to hire a human being who is fluent in both languages – and even that is far from perfect. 

But serious research into simultaneous electronic translation has been underway for a decade or so, under the leadership of an international network of universities and private laboratories – the Consortium for Speech Translation Advanced Research (C-STAR). With member organizations in the United States, Europe and Asia, C-STAR is dedicated to developing a technology that will enable linguistically challenged tourists to travel with “wearable translators” that could ask questions and get answers from the natives, or make it possible for an international group 

to hold an electronic meeting in several different languages – sort of a multilingual conference call. 

The first translating machines could handle only carefully-worded statements within a limited vocabulary and subject area. But the technology is improving steadily, to the point that the current prototype systems can handle spontaneous speech in vocabularies of 100,000 words, translating into a number of different languages. Not long ago the Consortium put on a demonstration video conference linking six countries, in which participants not only heard their statements successfully translated into other languages, but were also able to watch computer-animated projections of the faces of speakers in other countries, lips moving in synchronization with the translated language. 

Impressive progress, but the technology still has a long way to go, and is hardly likely ever to be able to translate rapid colloquial speech without difficulty. 

Nevertheless, even a moderately reliable speech-translation system would have a multitude of uses beyond those envisioned so far by the consortium researchers – for example, although the wearable translators are an obvious plus for tourists, future-minded law enforcement agencies are already talking about how useful they might be to peace officers in multicultural communities. And the technology also suggests the science-fictionish but inspiring notion that people might someday be able to hold global town meetings on a grand scale, connected as never before. 


Pacific News Service commentator Walter Truett Anderson, ( is the author of the forthcoming book, All Connected Now: Life in the First Global Civilization.