RT list: Grice´s Sticky Wicket

Someone on _this_ list should get enough funds to go to the H. P. Grice
Papers Collection, deposited in the UC/Berkeley, Calif. and find,

   _how_ used, as he apparently did,
    Â´relevance´
    in his "Conversation" lectures given
    at Oxford -- privately, as University Lecturer --
    in 1966,
    and which thus pre-date by a year the MS
    most often cited, the William James Memorial
    Lectures as typed by D. S. M. W.

(The ref. is to Chapman, op. cit., ch. on ´conversation´).

But on this post, rather, which I have titled apres Grice´s problem
with computers -- indeed the computers´ problem with Grice -- "this
d*mn machine can´t spell ´sticky wicket´" he would complain, cited by
Chapman.

Vis a vis D. Sperber´s recent ref. to this essay by White, PhD
UCalifornia Berkeley, now Emeritus, on the relevance of librarians when
giving hits to consulters.

I tried to access the link D. Sperber provides, but it´s almost $40
dollars to pull it, if you can believe that! No way, I would think. I
don´t get charged half as much by my boy as he pulls leaves from my
Swimming-Pool Library! But in any case, I found White´s powerpoint
presentation as he delivered it in Brazil. Fun, and full of exclamation
marks, which are good with some!

But whatever the merits of "evidence" (¨new evidence´ in White´s words)
 from "information science" -- a branch of ...? for ¨RT", I would like
to share with this list, with a view to positive constructive feedback,
my qualms, which I share with Grice, on computer-models of _anything_.

Mind, I _love_ computer models. Anything citing Grice I collect, and I
thus have a huge literature (by the lean Swimming Pool Library
standards, granted) on this. Power´s book, GUS model, Reichman on
implicature, Reiter, which I reviewed on the public domain, Thomason,
Rosenschein, and many many many others.

So in any case I would think we should consider:

computational models of paleo-Gricean issues.
computational models of Gricean issues.
computational models of neo-Gricean issues, or post-Gricean issues. (I
never know which one I have to avoid! just joking).

To me, the problems are two: a good and a bad.

The good is: I learned a lot from computer models of, say, metaphor.
How do we provide an algorithm, never heuristic, for the proof, in
formal terms, by which we can derive, say, that ¨No man is an island¨,
is, _pace_ Grice, a metaphor?

The bad is: Haugeland. He became the anti-computationalist at Berkeley
apres Grice, and he had the cheek -- just joking -- of including in one
collection of his essays, a joint paper with Grice, on Hume and
personal identity. The issues with computers is an old Berkeleian one.
Searle, for one, detests them, and would not even grant that a computer
can ´naturally mean´, "print!".

But of course, there are computers and there are computers. In
abstract, they are abstracts, and nobody should panic at them. The
issue with Gricean theories is the complexity of the
intention-embedding, and the defeasibility, non-monotonic, even
context-sensitiveness of much of the issues he discusses. Makes you
wonder: rough enough to deal with conversation as it happens in Oxford
-- the last ch. of my PhD was on Oxonian conversation, since I wanted
to refute anti-Gricean arguments against the universality of
implicature, by arguing -- even if just Oxonian, what´s wrong with
that? -- to even care to consider what new evidence for our
philosophical insights can be gained by a science of ... info?

In true vein, White´s essay is pretty complicated, from what I could
gather from the powerpoint presentation. The ¨Logic of Scientific
Revolutions¨ is just one of his examples, of at least 5 of them. But
they all seemed to simplify the original Gricean insight on what counts
as ´relevant´ on various levels. If a lister can access it and share
with the list, that would or should be fun.

Cheers,

J. L. Speranza
   for the Grice Circle
Received on Tue Jan 12 21:45:33 2010