Over the years I have seen predictions come and go. Some go quietly when it is clear they are wrong, while others hang on, particularly if the predictor has a vested interest in having people believe the prediction, at least for a short time. Let’s take a look at a few.
More than a decade ago, MT reared its hopeful head, promising to make quick work of the annoyingly expensive process of translating one natural language to another. And it would do it at a fraction of the cost then incurred for what some thought was a necessary, but too-expensive evil.
At the time, most translators scoffed at the notion of an MT takeover, but more than a few translators were somewhat alarmed. A non-translator friend of mine asked me if I wouldn’t shortly be out of work, as computers were taking over.
Well, that prediction was a quite premature, but it appears that the introduction of AI is combining with not only a long-standing distaste for spending money for translations but also the realization that the good-enough paradigm is largely valid. The result has presented a very real existential crisis for many translators. The prediction was quite premature back then, but it has turned out to be correct.
Switching from the sublime to the ridiculous, we have the recent promise of flying cars, like something revived from Popular Science articles in the 1950s. This time, however, there are working models. But the catch is that flying cars are aircraft, with inevitable regulatory and safety requirements that will sink this dream before flying car startups go into production.
The chances are microscopically small of us seeing the skies over highways (or over anywhere, realistically) blackened by flying cars, piloted (or unpiloted), carrying passengers to their destinations without having to deal with traffic congestion down below. I think it is a safe bet that, for all intents and purposes, the flying car hype doesn’t amount much to more than showcasing of technology. A proof of concept does not mean proof of acceptance by a society and its regulators.
Turning back to languages, I ran across a book about writing systems today (Andrew Robinson, The Story of Writing, Thames & Hudson 1995) that I bought more than a decade ago at the British Museum. My motive was to learn about various writing systems that have evolved throughout the ages, and I was interested to see some of the things the author had to say about the Japanese writing system.
He effectively and correctly disabused readers of the commonly heard mistaken notion that kanji are ideograms that represent ideas and more properly characterized them as logograms, which represent words.
Then I encountered a section in which the author posits that kanji are on their way out.
It looks likely that the need for computerization must one day lead to the abandonment of kanji in electronic data processing, if not in other areas of Japanese life.
Written, as amazing as it seems, in 1995, this was clearly off the mark. I guess the author was a bit prescience-challenged. It is a shame that the author fell into this trap, particularly only a few pages after he correctly explained the large number of homophones in Japanese as the reason the Japanese are not able to easily abandon kanji. What, I wonder, would an AI system do with texts written only in kana phonetics? Something quite comical (or tragic), I would imagine.
Predications will continue, including those that are premature, wrong, and spot-on. Which they are only time will tell.