Hacker News new | past | comments | ask | show | jobs | submit login

I believe the chart shown in the article titled, "Noam-enclature And The New Linguistics," mischaracterizes Chomsky's theory, by conflating the commonly used meaning of the word, "grammar," with grammar as it was defined in Chomsky's paper. In essence, this chart is a straw man argument.

https://www.scientificamerican.com/sciam/cache/file/A680FCC2...

The top part of the chart shows, "the brain's innate sentence diagraming machine," and then points toward a sentence structure broken down into designators, adjectives, nouns and verbs, which is what the common definition of the word, "grammar," means. This is not what Chomsky's theory stated. Chomsky's theory defined grammar as (I'm paraphrasing) a system for understanding language, involving a syntactic component (e.g. there should be some kind of order), a semantic component (e.g. there should be some kind of meaning) and a phonological component (e.g. there should be some kinds of sounds).

That's it! There was no inherent underlying English grammar structure hard-wired into babies' brains as the chart suggests. What is wired into a baby's brain, which assists with language, is the ability to order sensory inputs in a way that would result in language...that is what he called grammar. Basically what Chomsky did is, "psychologize," linguistics and I am not convinced by this article that this theory had been refuted. That would be like saying that Einstein refuted Newton.




> There was no inherent underlying English grammar structure hard-wired into babies' brains as the chart suggests.

There was in Chomsky's original theory. His original claims have been significantly mellowed out after the mounting weight of counter-evidence crashed down.


No, there wasn't. That's the poster's point. Universal grammar in Chomsky is lower-level than any concrete language, English most definitely included. https://en.wikipedia.org/wiki/Universal_grammar The first two sentences should suffice.


The characterization on that page is not good: virtually no one disagrees with "the ability to learn grammar is hard-wired into the brain". The substantive disagreement is whether grammatical categories and relations (or other knowledge that helps the child induce such) is hard-wired, or whether more general-purpose learning mechanisms are sufficient to abstract this structure from the environment.


>virtually no one disagrees with "the ability to learn grammar is hard-wired into the brain".

No, this is exactly where the field is split. Chomsky says that if a child socializes with people speaking a language with each other, the child will learn that language without teaching, punishment or reward. I.e. the human mind is not a blank slate buy has innate capabilities. Ideologically, this is an individualist take on linguistics.

The other side says that children have to be taught to speak a language properly and that language is difficult. Ideologically, this is the collectivist take on linguistics - children are blank slates to be filled in by society.

The people that have a problem with Chomsky's ideas are authoritarian collectivists. It's not about science.


You're making a straw man caricature out of the nativism detractors. Current researchers do not claim that first language is taught, that's why the field is called language acquisition regardless of whether one subscribes to nativism. The more nuanced non-nativist theories argue that grammar learning employs domain-general cognitive mechanisms, rather than language-specific innate grammar knowledge or principles. It's quite a stretch to connect any of this with political ideologies.


> The people that have a problem with Chomsky's ideas are authoritarian collectivists.

Wow, this is the worst argument for a scientific theory I have heard in a long time.


I might say that "universal grammar" is higher-level in the sense of more abstract than the grammar of any concrete language. Like Yacc.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: