Fantastic breakdown of how Chomsky's formalism collapsed under scrutiny. The Peters-Ritchie proof showing the theory was Turing-complete is brutal, basically revealing it explained everyting and nothing at the same time. What's interesting is how this parallels his political work, stripping away messy historical context to build grand unified theories. The Piraha counterexample is particularly damning since recursion was supposed to be the last remaining universal. It's kinda wild how someone can dominate a field for decades while the empirical foundation keeps eroding beneaththem.
Exactly. The one field where he could have been useful decided to abandon formal grammars as too cumbersome and moved to statistical approxations, which most modern nlp and llm algos.are based on
In the early 1980s, there was a contest at CMU to provide a context for "Colorless green ideas dream furiously." which would give the sentence a meaning. The winner was something like "After admiring the glorious pictures in the catalogs, one plants one's bulbs with hope in the autumn, and all through the bleak winter these colorless green ideas dream furiously." That final clause is in the middle voice, but the meaning is obvious in context.
Showed this piece to a friend with a PhD in linguistics — their response: “ This is largely a stupid critique. A theory of syntax cannot be faulted for being divorced from how language is used socially unless there is evidence that no aspect of syntax is independent from how language is used socially. The actual evidence goes the other way: Syntax is chock full of properties for which no one has been able to provide any sort of social explanation. For example, how come the pronoun "her" can have the same meaning as "Sarah" in 1 but not in 2?
1. Sarah wants Jane to love her.
2. Sarah wants her to love Jane.
The answer seems to be that a principle of Universal Grammar forbids any pronoun from co-referring with the nearest noun phrase that precedes it (I am simplifying this slightly for the sake of comprehensibility). To fault this appeal to a theory of Universal Grammar for not being based in the social use of language without offering a successful social-use theory in its place is merely to spout dogma. This is largely a stupid critique. A theory of syntax cannot be faulted for being divorced from how language is used socially unless there is evidence that no aspect of syntax is independent from how language is used socially. The actual evidence goes the other way: Syntax is chock full of properties for which no one has been able to provide any sort of social explanation. For example, how come the pronoun "her" can have the same meaning as "Sarah" in 1 but not in 2?
1. Sarah wants Jane to love her.
2. Sarah wants her to love Jane.
Will Esha next trash the binomial theorem for being divorced from the communicative processes that gave rise to the algebraic principles that it explains?"
No. because binomial theorem does not posit to model communication. It is a shorthand about how to multiply (a+b)^n, which you can perfectly do manually without the binomial theorem.
Not my field, but it seems that Chomsky did create a series of frameworks that provided a new and useful way of looking at linguistics?
Oversimplified and wrong, perhaps, but logical and well-organized, amenable to falsification and useful for others? Very roughly akin to Einstein's work, that wrongly didn't acknowledge quantum effects but whose relativity work provided a coherent base for others to work from?
Sadly no. jHe is constantly goal post shifting... so he is unfalsfiable. also burden is lm him to prove the existence of UG or merge "genes.". The burden is not on others to prove the lack of it
Chomsky's idea of a "deep structure" as part of language rather apart from semantics never made sense. All human languages are about expressing the same semantics. Children will either adopt the ambient language or develop their own, and that language will be fully expressive.
One current theory of human language breaks it into two subsytems, a syntactic one like the one used by birds to generate songs and a more semantic one predicate structures as used in primate calls and bee waggle dances. The former combined with statistical analysis has given us a lot of modern AI while the latter remains inscrutable though now and then brain science offers hints.
Einstein didn't reject quantum mechanics so much as feel that it led to "spooky" phenomena. This is true. Some QM results seem spooky, but so do a lot of ideas in general relativity. If you follow through on some of the ideas about light cones, you quickly get into spooky territory.
Fantastic breakdown of how Chomsky's formalism collapsed under scrutiny. The Peters-Ritchie proof showing the theory was Turing-complete is brutal, basically revealing it explained everyting and nothing at the same time. What's interesting is how this parallels his political work, stripping away messy historical context to build grand unified theories. The Piraha counterexample is particularly damning since recursion was supposed to be the last remaining universal. It's kinda wild how someone can dominate a field for decades while the empirical foundation keeps eroding beneaththem.
And my Theory is linguistics of secondary importance.. his primary function is that of an acceptable dissident
Exactly. The one field where he could have been useful decided to abandon formal grammars as too cumbersome and moved to statistical approxations, which most modern nlp and llm algos.are based on
In the early 1980s, there was a contest at CMU to provide a context for "Colorless green ideas dream furiously." which would give the sentence a meaning. The winner was something like "After admiring the glorious pictures in the catalogs, one plants one's bulbs with hope in the autumn, and all through the bleak winter these colorless green ideas dream furiously." That final clause is in the middle voice, but the meaning is obvious in context.
Thank You Esha
Showed this piece to a friend with a PhD in linguistics — their response: “ This is largely a stupid critique. A theory of syntax cannot be faulted for being divorced from how language is used socially unless there is evidence that no aspect of syntax is independent from how language is used socially. The actual evidence goes the other way: Syntax is chock full of properties for which no one has been able to provide any sort of social explanation. For example, how come the pronoun "her" can have the same meaning as "Sarah" in 1 but not in 2?
1. Sarah wants Jane to love her.
2. Sarah wants her to love Jane.
The answer seems to be that a principle of Universal Grammar forbids any pronoun from co-referring with the nearest noun phrase that precedes it (I am simplifying this slightly for the sake of comprehensibility). To fault this appeal to a theory of Universal Grammar for not being based in the social use of language without offering a successful social-use theory in its place is merely to spout dogma. This is largely a stupid critique. A theory of syntax cannot be faulted for being divorced from how language is used socially unless there is evidence that no aspect of syntax is independent from how language is used socially. The actual evidence goes the other way: Syntax is chock full of properties for which no one has been able to provide any sort of social explanation. For example, how come the pronoun "her" can have the same meaning as "Sarah" in 1 but not in 2?
1. Sarah wants Jane to love her.
2. Sarah wants her to love Jane.
Will Esha next trash the binomial theorem for being divorced from the communicative processes that gave rise to the algebraic principles that it explains?"
No. because binomial theorem does not posit to model communication. It is a shorthand about how to multiply (a+b)^n, which you can perfectly do manually without the binomial theorem.
Not my field, but it seems that Chomsky did create a series of frameworks that provided a new and useful way of looking at linguistics?
Oversimplified and wrong, perhaps, but logical and well-organized, amenable to falsification and useful for others? Very roughly akin to Einstein's work, that wrongly didn't acknowledge quantum effects but whose relativity work provided a coherent base for others to work from?
Sadly no. jHe is constantly goal post shifting... so he is unfalsfiable. also burden is lm him to prove the existence of UG or merge "genes.". The burden is not on others to prove the lack of it
Chomsky's idea of a "deep structure" as part of language rather apart from semantics never made sense. All human languages are about expressing the same semantics. Children will either adopt the ambient language or develop their own, and that language will be fully expressive.
One current theory of human language breaks it into two subsytems, a syntactic one like the one used by birds to generate songs and a more semantic one predicate structures as used in primate calls and bee waggle dances. The former combined with statistical analysis has given us a lot of modern AI while the latter remains inscrutable though now and then brain science offers hints.
Einstein didn't reject quantum mechanics so much as feel that it led to "spooky" phenomena. This is true. Some QM results seem spooky, but so do a lot of ideas in general relativity. If you follow through on some of the ideas about light cones, you quickly get into spooky territory.