Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 170
AIList Digest Monday, 6 Jul 1987 Volume 5 : Issue 170
Today's Topics:
Theory - Symbol Grounding Metadiscussion
----------------------------------------------------------------------
Date: 3 Jul 87 01:02:48 GMT
From: mnetor!utzoo!utgpu!water!watmath!watcgl!ksbooth@seismo.css.gov
Subject: Re: The symbol grounding problem - please start your own
newsgroup
Hooray for David Harwood.
------------------------------
Date: 5 Jul 87 05:39:38 GMT
From: mind!harnad@princeton.edu (Stevan Harnad)
Subject: Re: The symbol grounding problem - please start your own
newsgroup
In Article 186 of comp.cog-eng, ksbooth@watcgl.waterloo.edu (Kelly Booth)
of U. of Waterloo, Ontario writes:
> Hooray for David Harwood.
David Harwood has made two very rude requests that I stop the symbol grounding
discussion, which I ignored. But perhaps it's time to take a poll. Please send
me e-mail indicating whether or not you find the discussion useful and worth
continuing. I promise to post and abide by the results.
--
Stevan Harnad (609) - 921 7771
{bellcore, psuvax1, seismo, rutgers, packard} !princeton!mind!harnad
harnad%mind@princeton.csnet harnad@mind.Princeton.EDU
------------------------------
Date: 5 Jul 87 05:05:53 GMT
From: mind!harnad@princeton.edu (Stevan Harnad)
Subject: Re: The symbol grounding problem: 3 routes to grounding
needed?
In Article 181 of comp.cog-eng berleant@ut-sally.UUCP (Dan Berleant)
of U. Texas CS Dept., Austin, Texas writes:
> may not be much difference between a classical view augmented to...
> *arbitrary* boolean expressions of features...and a probabilistic view
I agree that such a probabilistic representation is possible. Now the question
is, will it work, is it economical (and is it right)? Note, though, that even
graded (probabilistic) individual features must yield an all-or-none feature
SET. So even this would not be evidence of graded membership. (I don't think
you'd disagree.)
> need to...explain...typicality and reaction time results...interpreted
> as supporting probabilistic and exemplar-based category representations
Yes, but it seems only appropriate that we should account for the
categorization performance capacity itself before we worry about its
fine tuning. (Experimental psychology has a long history of bypassing
the difficult but real problems underlying our behavioral capacities
and fixating instead on fine-tuning.)
> may [be] 2 representations for categories: a 'core' of defining features
> and a heuristic categorizer... 2 pathways [grounding] categories
You may be right. It's an empirical question whether the heuristic component
will be necessary to generate successful performance. If it is, it is still not
obvious that the need for it would be directly related to the grounding problem.
> [Re:] Anders Weinstein [on] the semantic meaning of...thunder/...`angry
> gods nearby'...: The terms in the definition presumably are grounded
> via the 2 routes discussed above... [now] Consider a sentence with 2
> variables, e.g. FISH SWIM... Obviously, many bindings would satisfy
> the sentence. [But]...by adding many more true sentences, the possible
> bindings of the variables become much more constrained.
I accepted this argument the first time you made it. I think it's
right; I've made similar degrees-of-freedom arguments against Quine myself,
and I've cross-referenced your point in my response to Weinstein. I
don't believe, though, that this reduction of the degrees of freedom
of the interpretation (even to zero) is sufficient to ground a symbol
system. Even if there's only one way to interpret an entire language,
the decryption must be performed; and it's not enough that the mapping
should be into a natural language (that's still a symbol/symbol
relation, leaving the entire edifice hanging by a skyhook of derived
rather than intrinsic meaning). The mapping must be into the world.
But, in any case, you seem to rescind your degrees-of-freedom
argument immediately after you make it:
> On the other hand... Maybe a Martian [or] your neighbor... could
> figure out [an alternative] way to do it consistently... but as long
> as you both agree on the truthfulness of all the sentences you are
> mutually aware of, there is no way to tell! Shades of the Turing test...
This is standard Quinean indeterminacy again! So you don't believe
your degrees-of-freedom argument! Well I do. And it's partly because
of degrees-of-freedom and convergence considerations that I am so
sanguine about the TTT. (I called this the "convergence" argument in
"Minds, Machines and Searle": There may be many arbitrary ways to
successfully model a toy performance, but as you move toward the TTT,
the degrees of freedom shrink.)
> would this method of 'grounding' the semantics of categories be
> sufficient to do the job? Only in theory? Potentially in practice? ...
I think it would not (although it may simplify the task of grounding
somewhat). Even if only one interpretation is possible, it must be
intrinsic, not derivative.
> Are you assuming a representation of episodes (more generally,
> exemplars) that is iconic rather than symbolic?
Yes, I am assuming that episodic representations would be iconic. This is
related to the distinction in the human memory literature concering
"episodic" vs. "semantic" memory. The former involves qualitative
recall for when something happened (e.g., Kennedy's assassination) and
the particulars of the experience; the latter involves only the
*product* of past learning (e.g., knowing how to ride a bicycle, do
calculus or speak English). It's much harder to imagine how the former
could be symbolic (although, of course, there are "constructive" memory
theories such as Bartlett's that suggest that what we remember as an
episode may be based on reconstruction and logical inference...).
> *no* category representation method can generate category boundaries
> when there is significant interconfusability among categories!
I would be very interested to know your basis for this assertion
(particularly as "significant interconfusability" is not exactly a
quantitative predicate). If I had said "complete indeterminacy," or even
"radical underdetermination" (say, features that would require
exponential search to find), I could understand why you would say this
-- but significant interconfusability... Can you remember first
looking at cellular structures under a microscope? Have you seen Inuit snow
taxonomies? Have you ever tried serious mushroom-picking? Or chicken
sexing? Or tumor identification? Art classification? Or, to pick some
more abstract examples: paleolinguistic taxonomy? ideological
typologizing? or problems at the creative frontiers of pure mathematics?
--
Stevan Harnad (609) - 921 7771
{bellcore, psuvax1, seismo, rutgers, packard} !princeton!mind!harnad
harnad%mind@princeton.csnet harnad@mind.Princeton.EDU
------------------------------
Date: 5 Jul 87 18:34:37 GMT
From: bloom-beacon!bolasov@husc6.harvard.edu (Benjamin I Olasov)
Subject: Re: The symbol grounding problem - please start your own
newsgroup
I personally don't feel that it's Harwood's place to make a
recommendation such as the one he made (rude or otherwise). If the
discussion is germaine to the stated purpose(s) of the newsgroup
(which it is), and is carried on in an intellectually responsible
manner (which it certainly has been), why should it not be allowed to
continue?
Isn't the solution for those who don't find the topic interesting to
simply not read the messages bearing that topic on the subject line?
After all, any number of discussions can be carried on concurrently.
------------------------------
Date: 5 Jul 87 17:31:15 GMT
From: harwood@cvl.umd.edu (David Harwood)
Subject: Re: The symbol grounding problem - please start your own
newsgroup
In article <977@mind.UUCP> harnad@mind.UUCP (Stevan Harnad) writes:
>
>David Harwood has made two very rude requests that I stop the symbol grounding
>discussion, which I ignored. But perhaps it's time to take a poll. Please send
>me e-mail indicating whether or not you find the discussion useful and worth
>continuing. I promise to post and abide by the results.
As I have told others, I don't really want you to quit posting
altogether to this or other newsgroups. And I would be glad for you to
form your own group for your "dialogues," such as they are. But I have
to complain about your insufferable postings on two grounds: (i) they
have nearly nothing to do with computer science, nevertheless preoccupy
comp.ai with your various and sundry self-referential, just vaguely
intelligible musings; (ii) your postings, in my opinion, are the heighth,
width, and breadth of unresponsive, presumptuous, and condescending
twaddle. Worse than anything which I've read which was contributed as
an original article to BBS, for example. Of course, as my colleagues
advise, BBS does not publish my research - and is unlikely to in the
near distant future. Such are the wages of public sin.)
Yes, my two replies to you were sarcastic (more than "very rude,"
I think; I never recieved any serious complaint about either, perhaps
because others knew what I meant, even if they did not quite agree with
me.)
Let me give you back an illustration of how you talk. You just
a moment ago replied to D.S. who question what psychological evidence you
have that perceptual categorization is usually "all-or-none." He seemed to
question your expertese as a perceptual psychologist. (I might add that
you have tried to impress us with generally slighting remarks about
psychologists as well as computer scientists, but this may be a "policy
of controversy" (perhaps used to secure competitive funding - who knows;-).
Anyway, your one line reply did not answer the question, but was
more of a silly riposte, something like, "Check the concrete nouns in
your dictionary." He asks you something, and you ignore this. Or, taking
you seriously, you tell him to go supply his own evidence for your claims.
(I suppose that if he were your research assistant, that you would sagely
explain that a "concrete" noun is one admitting "all-or-none" categorization.)
I have no prejudice concerning your views - to be sure, I rarely
can make sense of them. But I wish you would simply take your own advise,
"Check the concrete nouns of your dictionary," and use them sometimes to
good effect in your postings. Define your abstractions. Cite evidence for
your speculations. Do not cite your own damn article like a parrot. If you
prefer, post the damn thing, which has got to be more intelligible than
your recent stuff, and we will be done with this particular "symbol grounding
problem."
Then I will look forward to your new occasional postings, even
in this newsgroup.
David Harwood
------------------------------
Date: 5 Jul 87 21:48:28 GMT
From: harwood@cvl.umd.edu (David Harwood)
Subject: Re: The symbol grounding problem - please start your own
newsgroup
Letter sent by email to Stevan Harnad (with postscript added)
re his postings to comp.ai about "the symbol grounding problem."
I don't want you to quit posting altogether - I would just like
you to realize that you are hogging comp.ai with what seems, to me at least,
to be mostly pompous and unintelligible postings, that have very little to
do with computer science. I heard from a student colleague, who is not opposed
to a "cognitive science" viewpoint (if this means anything to you), that the
first thing you did to explain your views at a recent colloquim was make
reference to your net discussions.
My oh my, either you are an modest comedian, or these dialogues of
yours - why - if even they be blarney and posing of feathers - why they be
verily verily immortal.
You have made your views, whatever these are, resoundingly reknown
- by, I suppose, half or more of the recent volume of comp.ai. I simply wish
you'd pipe down for awhile, especially about your "symbol grounding problem."
I will be especially verily verily glad to see you post the source
code which implements your theoretical improvements; this should keep us off
the streets for awhile; and I will try to be first to applaud your success.
David Harwood
Computer Vision Laboratory
Center for Automation Research
University of Maryland
My views are simply my own. Please note all typos and mistakes, as I prepare
to publish an edition (with permission which is surely forthcoming) of
_Recent Contributions to the Dialogue de Problem Profundo Symbo-Grundo:
New Foundations and New Vocations in Computer Science_.
[This postscript added to my letter emailed S.H.]
------------------------------
Date: 6 Jul 87 02:19:01 GMT
From: bloom-beacon!bolasov@husc6.harvard.edu (Benjamin I Olasov)
Subject: Re: The symbol grounding problem - please start your own
newsgroup
In article <2328@cvl.umd.edu> harwood@cvl.UUCP (David Harwood) writes:
> I don't want you to quit posting altogether - I would just like
>you to realize that you are hogging comp.ai with what seems, to me at least,
>to be mostly pompous and unintelligible postings, that have very little to
>do with computer science.
^^^^^^^^ ^^^^^^^
This point should not need to be made, but this newsgroup doesn't deal
exclusively with computer science issues per se. Many important
contributions to AI, after all, have come from outside the field of CS,
as conventionally understood- much of Marvin Minsky's research for example,
is not restricted to CS, and yet has significant implications for AI.
Some of the most challenging and interesting problems of AI are philosophical
in nature. I frankly don't see why this fact should disturb anyone.
Perhaps if more of us pursued our theoretical models with comparable rigor
to that with which Mr. Harnad pursues his, the balance of topics represented
on comp.ai might shift .....
------------------------------
End of AIList Digest
********************
