4 ms·

Very true, but the problem is in paying these models to the actual study of linguistics and human language. For example, you can approach computation using some

by ribit 6y ago

Very true, but the problem is in paying these models to the actual study of linguistics and human language. For example, you can approach computation using something like a Turing Machine, but this won't help you understanding how to understand a modern CPU architecture or how to write well-optimized code. Chomsky's work is very interesting and relevant in the domain of formal languages, but as far as human languages goes, it always struck me as a self-indulgent intellectual exercise with little practical relevance.

foldr 6y ago
> it always struck me as a self-indulgent intellectual exercise with little practical relevance. It largely is. But you could say that of most work in pure mathematics or physics, or of pure research in any number of other disciplines. Generative grammar led to an enormous leap in our understanding of the grammatical structure of human languages. For example, we now know with a high degree of confidence that all natural languages have mildly context-sensitive string languages. There wasn’t even the glimmer of that result in the 1950s. It’s also worth noting that any rule-based NLP system makes use of a wealth of syntactic, semantic and phonological knowledge gained from Chomsky’s research program and its outshoots.
- ribit 6y ago
  This is very true. There is no denying that Chomsky and colleagues have contributed significantly to our understanding of the language. My comment is more in regards to the current day state of the field, which in my experience is overly driven by dogma. Science should evolve and improve it's methods, and sadly, the modern "generative grammarian" tends to live in the shadow of the glorious 80-ties.
  - foldr 6y ago
    Interesting take. I would say that the field is actually less dogmatic than it's ever been. Chomsky's personal influence is waning, and as the prestige of pure generative linguistics fades, younger researchers are having to broaden out their research programs to stay competitive.
    - ribit 6y ago
      Maybe you are right. What I am reporting here are the impressions from my early PhD around 10 years ago when I was still actively participating in conferences and workshops on generative grammar. Most contributions I have seen were along the lines of "I am trying to apply the minimalist program to this endangered language X and it doesn't work directly, but if I make the system more powerful, I can fit my examples in!" — and I just didn't see this approach leading us anywhere... I hope that you are right and that the field is evolving to everyone's benefit. We need different smart people with different opinions and believes to drive the science forward.
ukj 6y ago
It all depends on whether you understand what it means "to understand". There are two schools of thought: School 1: User's knowledge (this idea goes all the way back to Plato). You know how to use existing CPUs and existing programming language. This is akin to understanding the existing limits of existing technology. School 2: Maker's knowledge. You know how to invent CPUs and programming languages from first principles. The latter school inevitably leads to Wadler's law: semantics. It leads to asking "What is it that I want to express and why?". This mode of thought pre-supposes pragmatic/goal-driven intent.

by ribit 6y ago

foldr 6y ago

> it always struck me as a self-indulgent intellectual exercise with little practical relevance. It largely is. But you could say that of most work in pure mathematics or physics, or of pure research in any number of other disciplines. Generative grammar led to an enormous leap in our understanding of the grammatical structure of human languages. For example, we now know with a high degree of confidence that all natural languages have mildly context-sensitive string languages. There wasn’t even the glimmer of that result in the 1950s. It’s also worth noting that any rule-based NLP system makes use of a wealth of syntactic, semantic and phonological knowledge gained from Chomsky’s research program and its outshoots.

ribit 6y ago
This is very true. There is no denying that Chomsky and colleagues have contributed significantly to our understanding of the language. My comment is more in regards to the current day state of the field, which in my experience is overly driven by dogma. Science should evolve and improve it's methods, and sadly, the modern "generative grammarian" tends to live in the shadow of the glorious 80-ties.
- foldr 6y ago
  Interesting take. I would say that the field is actually less dogmatic than it's ever been. Chomsky's personal influence is waning, and as the prestige of pure generative linguistics fades, younger researchers are having to broaden out their research programs to stay competitive.
  - ribit 6y ago
    Maybe you are right. What I am reporting here are the impressions from my early PhD around 10 years ago when I was still actively participating in conferences and workshops on generative grammar. Most contributions I have seen were along the lines of "I am trying to apply the minimalist program to this endangered language X and it doesn't work directly, but if I make the system more powerful, I can fit my examples in!" — and I just didn't see this approach leading us anywhere... I hope that you are right and that the field is evolving to everyone's benefit. We need different smart people with different opinions and believes to drive the science forward.

ukj 6y ago

It all depends on whether you understand what it means "to understand". There are two schools of thought: School 1: User's knowledge (this idea goes all the way back to Plato). You know how to use existing CPUs and existing programming language. This is akin to understanding the existing limits of existing technology. School 2: Maker's knowledge. You know how to invent CPUs and programming languages from first principles. The latter school inevitably leads to Wadler's law: semantics. It leads to asking "What is it that I want to express and why?". This mode of thought pre-supposes pragmatic/goal-driven intent.