Chomsky and “walks like a duck” empiricism

Replicated Typo and @jasonbaldridge inter alia have been discussing a recent forum where Chomsky and Minsky and Labov (among others) met to discuss the past and future of artificial intelligence. Replicated Typo points out the most interesting contribution from Chomsky:

Chomsky derided researchers in machine learning who use purely statistical methods to produce behavior that mimics something in the world, but who don’t try to understand the meaning of that behavior. Chomsky compared such researchers to scientists who might study the dance made by a bee returning to the hive, and who could produce a statistically based simulation of such a dance without attempting to understand why the bee behaved that way. “That’s a notion of [scientific] success that’s very novel. I don’t know of anything like it in the history of science,” said Chomsky.

More evidence, as I suggested before, that Chomsky — and his acolytes — really don’t mean the same thing that empiricists do when talk about “what we mean when we say we know something”. Having a good predictive model — even if it’s statistical, or perhaps especially if it’s statistical — is a form of knowing something, and it’s a form that’s younger than Noam himself.

Empiricists adopt a “walks like a duck” model for demonstrating their understanding. I trace it to Shannon and Turing (pace Babbage and Lovelace).  The Turing Test itself is a form of “walks like a duck” empiricism (“if it looks and acts like an intelligence….”), and statistical models that behave the same as a natural phenomenon demonstrate an understanding of that phenomenon. Rejecting an explanation because it is messy is a form of willful blindness — the real world sometimes is messy, and an explanation must actually match observations: simplicity is not the only criterion.


Posted

in

,

by

Tags:

Comments

8 responses to “Chomsky and “walks like a duck” empiricism”

  1. Zoltán Varjú Avatar

    Speaking about scientific knowledge, you have to face the problem of simple description and explanatory models. The “mechanistic balkanization” (such a great term – used by Winston in the TR article) is the other side of the coin, empiricism can turn into a baroque rhetoric – just like rationalism 😀

    Empirical science is not only driven by empiria, but the researchers’ commitments and preconceptions – you can’t exclude those factors. If it walks like a duck and quacks like a duck, it is a duck. But before making the decision, you know what is a quack….

    1. Jeremy Avatar
      Jeremy

      Having categories in which to observe? Absolutely useful. Even the notion of a “word” is a category with some widespread utility.

      But the categories are fuzzy, even for widely-shared abstractions like word: How many words are in the phrase “Slytherin’s oft-cited Utility Model”? (I get between three and six [inclusive], depending on how I’m counting!)

      So yes, the assumptions and the units of observation are absolutely valuable. Nobody — not even die-hard empiricists — is saying that understanding your commitments and preconceptions is a waste of time. Please don’t straw-man.

      1. Zoltán Varjú Avatar

        In the empiricism-rationalism debate, these things were central, and later a similar debate started in the philosophy of science. I agree with you in more questions than you think, but assuming a kind of blank paper is a theoretical consequence of strict empiricism. If you leave room for some sort of top-down mechanisms, you have to accept that one can reason about these mechanisms.

        Let me put my thoughts into other words. There is some sort of difference between qualitative and quantitative reasoning. What Chomsky says is nothing more than highlighting this difference. But this kind of theory is not natural science! Chomsky defines his approach as Cartesian linguistics, and Cartesianism lost the battle with Newtonians. What Chomsky asks is not a scientific question, but a deep philosophical question.

        1. Jeremy Avatar
          Jeremy

          Z, which one do you think Chomsky identifies with “quantitative” and which with “qualitative”? I think I understand your analogy but I’m not sure it holds.

          Cartesian (in the philosophy-of-mind sense) is what I’m seeing in Chomskyan linguistics — a sense that the “i-language” (read “true self” or “soul” in Original Cartesianism) exists somehow separately from the “e-language” (read “material body” in OC). But Dennett and Dawkins and many others have really put the last nails in the coffin of Cartesian dualism — minds (and language faculties, I would say) are emergent from the material and social substrata, not hidden within them (or separate from them).

          1. Asad Sayeed Avatar

            Did they? Fodor might have something to say about that.

  2. […] p.m. crowd (a.m. in Europe): New Trochaism: Chomsky and "walks like a duck" empiricism https://trochee.net/2011/05/walks-like-a-duck/ […]

  3. Asad Sayeed Avatar

    Look. Here is the conflict in a nutshell as I see it:

    We observe structure in the universe, name it, and attempt to explain its provenance. Sure, fine, it emerges or whatever from some other physical substrate, which has structure that emerges from some other physical substrate, and it’s turtles all the way down until we get to fermions and bosons and then maybe it goes down still further.

    The question is, at each emergent level, there is an independent abstract reality to each structural level that merits description and explanation insofar as it is not possible to directly extend the properties of the substrate to the emergent structure—and it is not always possible, if it is ever (see economics for a case in point).

    Now, there is one side of the debate who has decided that they want to abjure “dualism” or whatever they want to call it for varied grab-bag of reasons and substitute, in place of rationalistic explanation, statistical models or something, and call that “empirical” explanation, and therefore better.

    To me they are hiding the underlying “rationalizations” that they are making in order to give themselves angelic scientific halos in comparison to the mean old Chomskyans. In the rush to explain the day-to-day and epoch-to-epoch variations of language—linguistic performance—and out of a mistaken and very naive interpretation of the concept of scientific evidence, there is a very real risk that they will successfully throw the baby out with the bathwater.

  4. […] who got fired up about Chomsky’s difficult comments regarding empiricism, including myself, will be gratified to see that Peter Norvig, patron saint of data-driven […]

Leave a Reply

Your email address will not be published. Required fields are marked *