[_glass]

I think it is important to realize, that we need to understand language on our own terms. The logic of LLMs is not unlike alien technology to us. That being said, the minimalist program of Chomsky lead to nowhere, because just like programming, it found edge case after edge case, reducing it further and further, until there was no program anymore that resembled a real theory. But it is wrong to assume that the big progress in linguistics is in vain, the same reason Prolog, Theorem provers, type theory, category theory is in vain, when we have LLMs that can produce everything in C++. We can use the technology of linguistics to ground our knowledge, and in some dark corner of the LLM it might already have integrated this. I think the original divide between the sciences and the humanities might be deeper and more fundamental than we think. We need linguistic as a discipline of the humanities, and maybe huge swaths of Computer Science is just that.

[js8]

I agree with you. I think the fundamental problem is we don't have a good unified theory of fuzzy reasoning. We have a lot of different formal approaches but they all have flaws.

Now LLMs made a big breakthrough that they showed we can do decent fuzzy reasoning in practice. But at the cost of nobody understanding the underlying process formally.

If we had a good unified (formal) theory of fuzzy reasoning, we could build models that reason better (or at least more predictably). But we won't get a better theory by scaling the existing models, I think Chomsky is right about that.

We lack the goal, not the means. If I am asking LLM a question, what answer do I want? A playfully creative one? A strictly logical one? A pleasingly sycophantic one? A harshly critical one? An out of the box devil's advocate one? A beautiful one? A practical one? We have no clue how to express these modes in logical reasoning.

[nz]

By way of analogy, the result of the theorem prover is usually actionable (i.e. we can replace one kind of expression with its proven equivalent for some end like optimizing code-size or code-run-time), but mathematicians _still_ endeavor to translate the unwieldy and verbose machine-generated proofs into concise human-readable proofs, because those readable proofs are useful to our understanding of mathematics even long after the "productive action" has been taken.

In a way, this collaboration between the machine and the human is better than what came before, because now productive actions can be taken sooner, and mathematicians do not have to doubt whether they are searching for a proof that exists.

[suddenlybananas]

>That being said, the minimalist program of Chomsky lead to nowhere, because just like programming, it found edge case after edge case, reducing it further and further, until there was no program anymore that resembled a real theory

As someone who has worked in linguistics, I don't really see what you're talking about. Minimalism is not full of exceptions (please elaborate on a specific example if you have one). Minimalism was created to make the old theory, Government and Binding, simpler.

[_glass]

Yes, and the project can be criticised by reducing until there's no value anymore. Well known instances of this process:

- Predicate Fronting in Free Relatives: In sentences like “What John saw was a surprise,” labeling the fronted predicate is not without problems, Merge doesn’t yield a clear head.

- Optional Verb Movement in Persian: Yes-no questions where verbs can optionally move (e.g., “Did you go?” vs. “You went?”) messes up feature-checking’s binary mode.

- Non-Matching Free Relatives with Pied-Piping: Structures like “In whichever city you live, you’ll find culture” mess up standard labeling, needs extra stipulations.

- Some Subjects in Finnish: Nominative vs. non-nominative subjects (e.g., “Minua kylmä” [me-ACC cold]) complicate that Minimalist case assignment.

[vrighter]

but we don't have llms that can "produce everything in c++".

We have LLMs that can get some boilerplate right if you use it in a greenfield project, and will repeatedly mess up your code once it grows enough for you to actually need assistance grokking it.