CLS International Newsletter

by David Ellis

Spring 1999

Fast forward to the year 2010. You’re back in Hong Kong and you don’t have a visa. Is that a problem? Not when it’s a virtual Hong Kong, a three-dimensional photorealistic replica of the real thing. You’ve arranged to meet the Cantonese ‘Gang of Four’; – Messrs. Tse, Li, Wu and Ho – at sunset on the CLS balcony at Lyemun realistically simulated from records and plans archived on the Internet. You are all wearing skin tight data suits made of synthetic cloth and eyeglass-like HMDs [Head-Mounted Displays that give you the illusion of being inside the 3-D scene]. But you don’t see these paraphernalia. You see yourself and the others wearing whatever apparel you and they told the computer [which is integrated into your data suits] to dress you in.

I picture Mr. Tse in full Imperial regalia, flowing gold silk gown with the CLS dragon embroidered in red; Mr. Li in the sober robes of the Confucian scholar; Mr. Wu, calligraphy brushes in hand, as an artist; and Mr. Ho in snappy business suit and bright tie. You, needless to say, are in baggy shorts, vomit-streaked T-shirt and down-at-heel flipflops. But that’s irrelevant. What’s relevant is that the technology to do all I’ve just described is already here, though it’s not yet very well developed. But wait, as they say in the commercials; there’s more! As you all shake hands [feeling their individual strength and warmth through the haptic sensors -microscopic transducers woven into your data suits at a density of thousands per square centimetre], you hear their greetings enunciated in perfect English. Even more amazing, they hear your greetings in perfect Cantonese, except for Mr. Tse, who, having adopted the Mandarin befitting his virtual station in life, hears you speaking perfect 國 語.

That technology is here, too, I’ve tried it. It doesn’t work very well yet but it’s only a matter of time before it does, and the year 2010 is a very safe bet for it all to come together.

The technology is currently in two separate parts. First, there is the technology of automatic speech recognition [ASR]. Second, there is the technology of machine translation [MT]. Eventually, these technologies will be merged into one. [There is a third technology, voice synthesis, that has already been incorporated into ASR programs.]

ASR is software that lets you dictate through a microphone into the computer, which recognises what you say and types the words out on screen. ASR does not understand your words, yet , but that doesn’t matter for our purposes. The voice synthesis technology in ASR programs enables the computer to read aloud, through its loudspeakers, the words on the screen.

There are currently three inexpensive and competing ASR programs: Dragon Software’s ‘NaturallySpeaking’, Lernout & Hauspie’s ‘VoiceXpress’ and IBM’s ‘ViaVoice’. I’ve reviewed the first two of these products on the Web [see the links at the end of this essay for more information].

MT software takes text that’s already in the computer and translates it into another language. One such program available for ‘Windows’ machines is Systran, which can handle several European languages plus Chinese and Japanese. I’ve reviewed that, too. A version of Systran software now running on the AltaVista search engine will give you near-instant and free translations of Web pages and documents in French, German, Spanish and Portuguese, with more languages planned. It does not take much vision to see that ASR and MT can, and will, be merged to give us the equivalent of Douglas Adams’ [Hitchhiker’s Guide to the Galaxy] ‘Babelfish’ – an alien fish which, placed in the ear, translates between the ‘wearer’ and beings of all races in the universe. Given the exponential rate of increase in the power of computers, and the concurrent exponential decrease in their size, it will not be long before programs are woven into ‘wearable computers’. The world-renowned Media Lab at the Massachusetts Institute of Technology held a wearable computer fashion show in October 1997 that included a tunic which translated the wearer’s speech into a foreign language.

Those of you who have managed to keep your Chinese characters alive may be forgiven for wondering if, in the face of such technology, it was worth the effort. Of course it was, and it will always remain so.

Despite my own loss of the characters, the experience of having had even a smattering for a short time opened a priceless window onto the Chinese mind and soul that no interpreter, human or machine, could emulate or substitute for. I am deeply honoured to have been afforded a view through that window through the teaching of Mr. Tse and his colleagues.