From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!mips!darwin.sura.net!cs.ucf.edu!barros!gomez Mon May 25 14:07:34 EDT 1992
Article 5886 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!mips!darwin.sura.net!cs.ucf.edu!barros!gomez
>From: gomez@barros.cs.ucf.edu (Fernando Gomez)
Subject: Re: Can Abstract Categories Be Grounded?
Message-ID: <1992May25.081017.27092@cs.ucf.edu>
Keywords: horse, stripes, zebra, goodness, truth, beauty
Sender: news@cs.ucf.edu (News system)
Organization: University of Central Florida, Orlando
References: <1992May24.172516.15226@Princeton.EDU>
Date: Mon, 25 May 1992 08:10:17 GMT
Lines: 48

(This is a reply to Harnad's 6522 article.)

Unanswered Objections:

The reason why concepts like "bachelor," "employee," "divorcee," etc.
present a more serious objection to Harnad's grounding project than concepts
like "good," "truth," etc. is that the former are not abstract concepts.
They are physical entities. Bachelors are spatial, eat, live, ... You
do not break any ontological category by saying "Bachelors eat apples,"
but, yes, by saying "Truth eats apples." But, in spite of being physical
entities, class membership is not determined by empirical stimuli.
That is the interesting aspect of "bachelor"!

The second objection in my posting and due to Quine has nothing to do
with Wittgenstein's statement that concepts are not bounded
categories defined by necessary and sufficient conditions.
The problem is completely a different one. 
Basically, it attacks the notion that the relation of symbols to
the physical world is one-way street. According to Quine,
it is rather a multi-way street in two directions. Symbols *ground*
even the most overt empirical experiences. (Cf. Word and Object, and
Two Dogmas of Empiricism. )

My third objection is based on the fact that when one observes
the evolution of modern language, the trend of this evolution
has been to become increasingly abstract. It is amazing the
number of scientific concepts that have found their way in
language. In many senses, modern language looks very much like a
popular science. As result of this,
language has become more and more distant from the physical world,
becoming an immense fabric of concepts whose relation to
empirical stimuli is rather insignificant. (Symbols take off
leaving the physical world far behind.)

A final comment to Harnad's statement "A roboticist is just
modeling performance capacity, not ontology." I wonder how a robot
can ground symbols without an ontology. As Kant said, referring to
Locke and Hume's, even for experience to take place
one needs a set of *a priori* concepts that are the conditions of possibility
of experience. Work by Keil and Carey on the role of ontology in
children is highly relevant, and certainly not introspection.
Ontological matters matter.

Fernando Gomez






