[Corpora-List] ad-hoc generalization and meaning

[John F. Sowa]

I don't blame people for feeling overwhelmed with email, and I
try to limit my own comments to one message per day (except when
I get excited by something I feel is outrageous).

In this note, I'd just like to make a couple of remarks, starting
with a comment about the response Rob made to Paula:

PN> So the question, Rob, is what are you proposing?  Is it a new
> approach to linguistic investigation, or to NLP, or to ??

RF> Your comments conjure in me a rather odd picture of science where
> we assume everything which can be known, is already know, and it
> only remains to select what we want to do with that knowledge.

First, I don't think Paula intended that at all, but I do think it
is legitimate to ask anyone who is engaged in any aspect of science
or engineering what their goals might be.  In the long run, we would
all like to get a better understanding of the whole field, but in
the short run we have to focus on the things we can do best in the
time we have to devote to the study.

Right now, some metalevel discussion about where the field is going
is appropriate because many of the paradigms of the 20th century
are breaking down and many of the researchers who have been prominent
in those paradigms are questioning the foundations.

In an earlier note, I gave some pointers to comments by Hans Kamp
and Barbara Partee.  I also cited some books, whose contents are
not available on the web.  So I thought I would quote a few remarks
from them, which indicate how far they have diverged from the old
assumptions.

 From _Simpler Syntax_ by Culicover and Jackendoff (2005),
    "a far leaner syntax coupled with a somewhat richer syntax-semantics
    interface... that better facilitates the integration of linguistic
    theory with concerns of processing, acquisition, and biological
    evolution" (p. xiv). "an adequate theory of syntax should connect
    in a natural way to an account of how humans produce and understand
    sentences" (p. 546).

This is heresy when compared to Chomsky's old dogma, but there is much
more to language than just sentences.  The word 'sentence' should be
replaced by discourse, dialog, knowledge, human activity, and life.

 From _Dynamic syntax_ by Kempson, Meyer-Viol, and Gabbay (2001),

    "the time has come to shift from the static perspective of
    formalisms based on the metaphor of the familiar classical logics
    to more dynamic formalisms where the emphasis is on the process
    and the incremental development of structure" (p. x). "Language
    processing... involves manipulating incomplete objects at every
    stage except at the very end" (p. 3). "the DECLARATIVE part of
    the model, the decorated partial trees in which the results of
    the parsing process accumulate, and subsequently with the
    PROCEDURAL part of the model, the actions which transform one
    decorated partial tree into the next one" (p. 268).

This approach goes much farther than Culicover and Jackendoff by
not only simplifying the syntax, but by rejecting the idea that
there is a fixed set of grammar rules of any kind.  In that
regard, it is similar to variations of construction grammar.
But not the qualification "except at the very end" (p. 3).  The
last chapter of the book by Gabbay still has a very formal logic,
although the notation uses "decorated trees" instead of linear
formulas.

I am not opposed to formalization, where appropriate, but I would
like to see the logic made as dynamic as the syntax.  I don't
believe there is ever a "very end".  Instead, I would say that
there are just temporary resting points along a continuous range
of possible variations of grammar, logic, ontology, etc.

But I'm not saying that I adopt a relativistic view where "anything
goes".  I believe that the world imposes strong discipline by the
"school of hard knocks", and it punishes anyone who ignores the
limits imposed by the environment, human nature, and other people.

An important correction I would make to Rob's comments is that I
believe semantics is the study of how language relates to the world.
Studying a corpus is fine, but it is also necessary to relate strings
in the corpus to the world and to life.  Without some connection to
perception and action (even indirect), there is no semantics.

John Sowa




_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora