Linguistic knowledge for language technology: what for?
This came from a recent article in CNET news, talking about machine translation:
In the past few years, however, researchers have switched to using statistical analysis to get the job done.
"It doesn't go through a deep understanding of the meaning of a sentence. It maps one word to another," Waibel said. "Increases in computer speed and power and databases have made this a winning approach...We essentially gave up trying to do the full semantics of this thing.
I hope no linguist still believes that linguistic knowledge can ever be the main force is language technology. It won't happen in useful applications. Disappointing? Maybe. Should linguists go deeper into statistical approaches? Probably not. Unless we all sell out!
The article stresses some weaknesses of the statistical approach, especially the lack of databases. If one wishes a system translating directly between some 100 languages or so, not a tiny fraction of data exists for it.
Another curious quote from the same Waibel:
"I was born German and spent my childhood in Spain and I speak German, English, Spanish, French and Latin," he said. "My wife is Japanese so I am sort of culturally messed up."
How can anyone claim to speak Latin? One can even have a deep knowledge of Latin, but speak it? With whom? Why? As an exercise or for showing off, I guess.