Wait! What? Chomsky is wrong about something?

Noam Chomsky

I am a Chomsky enthusiast, admittedly. I have read several of his articles and books, and have viewed countless videos. I am no expert in Linguistics, but would sit through a video of his lectures on it.

So, when Tom Wolfe wrote in Harper’s magazine, an article entitled The Origins of Speech, I learned about the fallability of one of the “most important intellectuals alive”. Not that this is tragic. After all, who cares if an intellectual makes a mistake or not? It would be beyond naive to think that Chomsky’s pronouncements on any topic are flawless, and I am sure that Chomsky would subscribe to that admission, although much has been made on the part of others about his legendary rhetorical skills, which had served him well in live public debates and his countless speeches.


Tom Wolfe mentioned Chomsky’s insistence that Linguistics be seen as a science. As a science, it is unique, as Wolfe states, in that it requires zero field study. Linguistics is the only “science” that does this. Rather than be seen as, say, another Linus Pauling or Frederick Banting, it would be better to see him as an ancient Greek philosopher like Democritus. Democritus asked the question about the essential nature of matter and came up with “atoms” without all the bother of needing to observe samples under a scanning tunnelling microscope, for which he would have had to wait 2500 years for such a microscope to be invented.

toddler-blocksSimilarly, Chomsky asked the question about the essential nature of language acquisition as a philosopher. He asked the question of how children first acquire language from truly degenerate and imperfect samples found in the environment around him. The child is a toddler, perhaps age 2, so there is no way for the child to look up words somewhere on Google; no way to get perfect grammar samples; no way to know the parts of speech. The child’s cognitive skills are not up to a sufficient level, nor his motor skills, nor his reasoning skills. In addition, toddlers acquire language in very adverse situations, where the parent’s education and vocabulary are not great, and the growing environment is not the best. So, Chomsky reasoned, acquisition of language must be innate to our species. We chat up a storm just as surely as beavers build dams and birds build nests. Therefore, language acquisition must be tied to neurological processes more than it is tied to processes outside of the language learner. So, maybe there is a “language centre” in the brain, and maybe a universal grammar … ?

That’s a lot of reasoning, and a true scientist would demand “real-world” evidence, such as those obtained from field study, but as Wolfe points out, Chomsky himself downplayed that. That still doesn’t fault Chomsky, since that could still be for others to do, and I am sure a great deal of this has been done over the past 60 years that Chomskyan Linguistics has been around, if only by detractors of Chomsky’s theories and would try to knock it down somehow.

Peter Norvig

There has, and any linguistics based on samples and statistics have been treated as a separate paradigm of thought. While these would be dismissed by Chomsky and his followers, it did give us Google Translate. Google Translate, invented by Google research director Peter Norvig, is a unique beast. Google Translate knows nothing about any language, relying only on statistical occurrences of actual use of specific words and phrases in documents on the Internet. It uses what linguists call “corpus” linguistics: the study of language as it is actually used and observed to be used by people in the real world. And it does it with a significant degree of success. I have used it extensively, and can attest to its usefulness. Corpus linguistics would be dismissed by Chomsky as one dismisses a paradigm that is outside of one’s own.

Chomsky and those invested in his paradigm would have not been able to create Google Translate from universal grammar (UG) theory. UG is touted to exist, but such a UG is not fully realized, at least not to the point where a translating engine could auto-detect your language and translate it to another one for you the way Google Translate does. Corpus Linguistics appears to do the real work manifest in our techhnologies and education systems, while UG is that shiny new theory, 60 years old but largely still on the drawing board, attractive but too new to be applied to existing technologies right now. UG is like the wave mechanical model in physics, where you are better off with high school physics most of the time. Einstein may have been right and Newton may have been wrong, but you can still build bridges and skyscrapers with Newtonian physics.

Wolfe does not go into any of this stuff about paradigmatic differences, by the way. While there is effectively a schism between modern linguists, Wolfe chose to attack Chomsky in ways that border on ad hominem. Wolfe has been criticized for misinterpreting actions of Chomsky to fill out a particular narrative. He was painted, for example, as a kind of defensive person, unresponsive to criticism. This is laughable, since he is famous for responding to email from just about anybody. I had even corresponded with him in the past, and to this day I have never met him in person. His correspondence with me had always been thoughtful and helpful, and far from defensive or unresponsive.