Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!ucsd!nosc!humu!uhccux!lee
From: lee@uhccux.uhcc.hawaii.edu (Greg Lee)
Newsgroups: comp.lang.misc
Subject: Re: Syntax vs Semantics (silliness!)
Message-ID: <3706@uhccux.uhcc.hawaii.edu>
Date: 10 Apr 89 13:43:13 GMT
References: <7708@batcomputer.tn.cornell.edu>
Organization: University of Hawaii
Lines: 30

From article <7708@batcomputer.tn.cornell.edu>, by garry@batcomputer.tn.cornell.edu (Garry Wiegand as guest):
" ...
" I'm not even real sure of the *informal* English distinction between 
""syntax" and "semantics". 
" 
" How do you linguistics people use these words?

We are no less confused than anyone.  Semantics concerns reference, but
in common usage in linguistics and logic, cross-reference doesn't make a
thing semantic.  In the typedef affair, I take it we're talking about a
sort of cross-referencing.

For instance, in predicate logic, in the most commonly used notation,
arguments are cross-referenced by re-using variable names within the
scope of a single quantifier for that name.  Yet one can still give a
logical syntax for predicate logic.  Perhaps that is justified in logic,
because variables can be regarded as notational conveniences.  In a
variant of ordinary logical notation, combinatory logic, there need be
no variables.  [Ref.  W.V.O Quine, "Variables explained away", in Proc.
Nat. Acad. of Sciences, late '40's.]

A kind of cross-referencing in natural language similar to typedef's is
grammatical agreement, which is usually taken to be syntactic in nature.
It can be argued that agreement in natural language does not prevent it
from being context-free, but the argument depends on the fact that the
number of agreement categories (singular/plural, masculine/feminine,
Bantu noun class, ...) is finite (and small).  This is not the case
with typedef's, so I would say that C is not context-free.

			Greg, lee@uhccux.uhcc.hawaii.edu