Re: Noam Chomsky born (7/12/1928)

 Ar an deichiú lá de mí na Nollaig, scríobh Jeff Barnett:

 > [...] I interacted with several MIT folks who worked with and around Chomsky
 > in the 1960s and 1970s though I never met him. I formed a few impressions
 > that were favorable. The first was that he provided several theories about
 > the origins of language and the development of its structure in enough
 > detail that one could criticize the ideas and, in so doing, learn things
 > whether or not his original speculations were completely correct or not.

And then he was consistently unwilling to accept data that falsified his
theories.

 > The second was his work in formal language theory: the identification of
 > relative powers of representations. (In computer science, the hierarchy from
 > finite state to touring machines.) This work led to computer tools that
 > allowed various grammars to describe languages and either attempt
 > recognition or generation. I believe that these tools brought a new wrinkle
 > into linguistics: You got a theory, show me your grammar and I'll quickly
 > generate a few thousand examples and see if they all ring true.

Almost never relevant to conventional languages (because the grammars were
always incomplete), and a dead-end for computer languages. Neither are used
today in anger. Machine translation and large language models use statistical
methods. Compilers that aim to give helpful feedback with compile errors (that
are not toys) need heavy hand-parsing of the input.

 > All in all, the above brought a degree of science to linguistics that hadn't
 > been there before because practitioners were now asked to provide at least
 > partially testable theories. This supposedly deprecated the artistic flare
 > that linguists brought to the field but I don't think that was the case.
 > Rather, it eliminated a substantial amount of bull shit that was foist by
 > reputation rather than testing.

As I say, a dead end, his own brand of bullshit.

But he made a good living from the military-industrial complex while loudly
decrying the military-industrial complex. Maybe his underlying motivation was
to waste US taxpayer money.

 > Not being a linguist myself, I may hold beliefs that are counter to your more
 > informed opinions. In part these opinions came from seeing the "MAC hackers"
 > coding up some professors' linguistics publications and generating raucous
 > nonsense with the code. Of course you'd hear horse laughs all through the AI
 > lab.

--
‘As I sat looking up at the Guinness ad, I could never figure out /
How your man stayed up on the surfboard after fourteen pints of stout’
(C. Moore)