Newsgroups: comp.ai.nat-lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!satisfied.elf.com!news.mathworks.com!uunet!harlequin.com!epcot!usenet
From: norvig@comet.menlo.harlequin.com (Peter Norvig)
Subject: Re: Context Sensitivity
In-Reply-To: billf@osi.ncsl.nist.gov's message of 6 Feb 1995 18:24:24 GMT
Message-ID: <NORVIG.95Feb7162032@comet.menlo.harlequin.com>
Lines: 24
Sender: usenet@harlequin.com (Usenet Maintainer)
Nntp-Posting-Host: comet.menlo
Organization: Harlequin Inc, Cambridge, MA
References: <447@mango.ukc.ac.uk>
	<GRIFFITH.95Jan30101222@filippo.sfs.nphil.uni-tuebingen.de>
	<3gmcck$nkm@csnews.cs.Colorado.EDU> <QOBI.95Feb1105826@qew.cs>
	<NORVIG.95Feb2165128@comet.menlo.harlequin.com>
	<3h5pgo$i2j@dove.nist.gov>
Date: Wed, 8 Feb 1995 00:20:32 GMT


I showed a purported counter-argument to the claim that the "respectively"
construction proves that English is not context-free:

> Contestants 1 through 3 drove Melanie, Norman, and Oliver to Xanadu, York,
> and Zanzibar respectively.
> 

Then billf wrote:

>  What difference does it make if it's syntactic or semantic?  It's still
> true that somewhere in the language facility there's context sensitivity.

The difference is between making a precise, measurable claim (that the
English language is or is not context free in the sense defined by
formal language theory) versus making a vague, unfalsifiable claim
(that "somewhere in the language facility there's context
sensitivity", without defining what "context sensitivity" means in
this sense).
-- 
Peter Norvig                          
Harlequin Inc.                        Email: norvig@harlequin.com
1010 El Camino Real, Suite 310        Phone: 415-833-4022
Menlo Park CA 94025                   Fax:   415-833-4111
