Re: Fw: Intro to natural language processing
On Friday 24 December 2004 02:14, John F. Sowa wrote:
> ...
> People speaking a foreign language can make themselves understood even
> with badly fractured syntax ... That supports Deacon's hypothesis
> that symbols are more important than syntax
I think that is a little hard on syntax. Is syntax unimportant just because it
is flexible?
Lexicon (symbolism in Peirce's sense of "convention"?) is one way of modelling
subjectivity. The observation of subjective syntax (collocation) explains the
popularity of Langacker's Cognitive Grammar and other lexically styled models
of language. But lexicon alone is a bit limiting. Collocation is not just
habit built on weak syntax.
Rather than marginalising syntax, wouldn't it be better to model its
flexibility?
While disdaining your (off-list) dispute with Jean-Luc ('I am very well aware
of the work on quantum logic. My only objection to it is the word "logic".'
-- Sheesh! That won't bake any bread!) I think I agree with Jean-Luc on the
substance. If we want to bake bread we have to do something different.
Contextualisation (if I can summarise the immediate appeal of QM in that way)
may be as good a rallying cry as any. I hesitate to talk about QM myself
because I think it will just alienate most readers on this list (I prefer to
limit myself to talking about the properties of ad-hoc generalizations, which
can be seen as QM by the way...), but whatever else can be said about the
relevance of QM to the system of language, the contrast between Classical
Mechanics and Quantum Mechanics does show systems can have materially
different properties. Find parallels and make logic and analogy the same if
you will, but do have a model where syntax depends on context (by ad-hoc
generalization over contrasting strings if you ask me :-)
-Rob Freeman