to home

BYTE Letter on Voice-Writers
October, 1986

TONGUE-TYING VOICEWRITERS [headline by BYTE editor]

To the Editor:

In order to clear away hype and set standards, I would suggest that the following text be read in clear measured tones to all devices purporting to function as "voicewriters:"

"Cap Ten and my caftan," kept on crying the Captain, "Cap Ten!" Of numbered caps Cap Ten suited the Captain. He had numbered them as the numb bird landed. "Land dead ahead!" called out the Mate to cawing from the numb bird. Would its gnaw tickle? It cawed, but it was caught on the carpet. "Correction!" the Captain would carp pettily, "Your nautical law has no tickle. We've three types of wreck here: surface, deeper, core. Sir, fuss deeper, weave, hear the screw's deep purr. This ship's an old seesaw!" The Captain capped an epigram: "Old sea saw: Correction! Core wreck shun. Cap Ten and my caftan! I as a Copenhagen Captain coped on a capstan and kept on coping!"

If the device makes five or more errors in transcription (or about one per twenty words dictated), some careful examination of the voicewriting concept may be in order. While it may seem that this text has been contrived to place a fairly high strain on such machines, the type of difficulties it presents are not only normal but inevitable. Similar ones are found in everyday spoken English, where not only isolated homonyms but the ordinary internal phonetics of words and phrases can create obstacles to sound and word recognition. In some foreign languages, particularly Chinese and other members of the Sino-Tibetan family, this problem is many times more difficult than in English.

A few years ago there was a great deal of noise to the effect that computers would soon replace "analog" or human translators. This was before it was realized how difficult the problems of translating natural languages actually are, and now those companies involved in machine translation talk of working with rather than replacing the translator. Some of the same problems involved in machine translation are also present in transcribing spoken words, in some ways in even more acute forms. Hence it may be a bit premature to talk of replacing secretaries and stenographers with voice-writing devices.

Yours sincerely,
Aex Gross
Co-Director,
Cross-Cultural Research projects

COPYRIGHT STATEMENT:
This article is Copyright © 1986
by Alexander Gross and by
BYTE Magazine. It may be
reproduced for individuals and for
educational purposes only. It may
not be used for any commercial (i.e.,
money-making) purpose without
written permission from the author.

to linguistics menu
to top
to home