"Generative" serves them right

[Stephen M. Wechsler]

At 4:49 PM -0400 4/28/01, Carl Pollard wrote:
>  >
>Sergi, you say that Chomsky "reinvented" SLASH, and
>called it 'probe'.  Do you mean that SLASH and 'probe'
>are notational variants (laying aside who thought of
>it first)?
>>>
>
>Historical note: in linguistics, at least as far as I know, Gazdar
>deserves the credit (1981, Unbounded dependencies and coordinate
>structures, though I remember seeing it in drafts by Gazdar
>circulating in 1979), though he did not treat it as a feature. I think
>the first person to do that, in print at least, was John Bear, about
>1982. Ann Paulson and I independently implemented the same idea in an
>early HPSG system at Hewlett-Packard Labs, I think summer of 1983 (we
>found out about Bear's paper later).
>
>But if you go outside linguistics to logic, SLASH is basically the
>same idea as hypothetical reasoning in implicative logic, or
>introducing an indeterminate in lambda calculus, so it is an idea that
>goes back at least to the 30's and 40's

A footnote:  The slash idea was simultaneously and independently
developed by Gazdar and by the late C.L. (Lee) Baker, my former
colleague here.  See Baker's excellent textbook 'English Syntax' (MIT
Press).

Although Baker's work was in the transformational framework, he was
sympathetic towards GPSG-- and skeptical about minimalism.  His 1991
LI paper critiquing Pollack's influential split infl proposal is a
classic demonstration of how an analysis can appear to simplify but
carry steep hidden costs.

Steve