Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 177
AIList Digest Monday, 13 Jul 1987 Volume 5 : Issue 177
Today's Topics:
Theory - Symbol Grounding Poll: Yea's
----------------------------------------------------------------------
Date: 9 Jul 87 03:32:45 GMT
From: mind!harnad@princeton.edu (Stevan Harnad)
Subject: Results of Symbol Grounding Poll (1st of 3 parts)
In the poll on whether the symbol grounding discussion was useful and
worth continuing there were 24 yea's and 37 nays (with some ambiguous
ones I have tried to classify non-self-servingly), so the nays have it.
As promised, I am posting the results (yea's in part 2 and nays in
part 3) and I will abide by the decision. Perhaps I may be allowed a few
parting reflections:
(1) It is not entirely clear what the motivation of the nays was:
ecological/economic considerations about overuse of the airways or
reluctance to perform the dozen or so keystrokes per week (or to
put in the software filter) that would flush unwanted topic headers.
(2) There were distinct signs of the default option "I can't follow it,
therefore it makes no sense" running through some of the nays (and indeed
some of the discussion itself). This may be a liability of polling as a
method of advancing human inquiry.
(3) Along with several thoughtful replies, there was unfortunately also some
ad hominem abusiveness, both in the poll and in the discussion. This is the
ugly side of electronic networks: unmoderated noise from the tail end of the
gaussian distribution. It will certainly be a serious obstacle to making the
Net the reliable and respectable medium of scholarly communication that I
and (I trust) others are hoping it will evolve into. It may turn out that
moderated groups, despite the bottle-necking they add -- a slight step
backward from the unique potential of electronic nets -- will have
to be the direction this evolution takes.
(4) I continue to be extremely enthusiastic about and committed to
developing the remarkable potential of electronic networks for scholarly
communication and the evolution of ideas. I take the present votes to
indicate that the current Usenet Newsgroups may not be the place to attempt
to start this.
(5) Starting a special-interest Newsgroup every time a topic catches
on does not seem like the optimal solution. It is also unclear whether
even majority lack of interest should prevail over minority interest
when all that seems to be at issue is a keystroke. (Not only is there
software to screen out unwanted topics, but to filter multiple postings
as well. I have been posting to both comp.ai and comp.cog-eng because they
each have a relevant nonoverlapping sub-readership. I subscribe to both; my own
version of "rn" only displays multiple postings once. Secondary
digests like the ailist are another matter, but everyone knows that
half or more of it duplicates comp.ai anyway. The general ecology and economy
of the airwaves, on the other hand, should perhaps be deliberated at a higher
level, by whoever actually pays the piper.)
(6) The current majority status of engineers, computer scientists and
programmers on the Net also seems to be a constraint on the development of
its broader scholarly potential. Although these two disciplines developed the
technology and were the first to use it widely, it's now rather as if
Guttenberg and a legion of linotype operators were largely determining not
just the form but the content of the printed page. The other academic
disciplines need *much* greater representation in the intellectual Newsgroups
(such as those devoted to biology, language, philosophy, music, etc.)
if the Net's scholarly contribution is ever to become serious and lasting; right
now these Newsgroups seem only to be outlets for the intellectual hobbies of the
two predominant disciplines. This may just be a quirk of initial conditions
and a matter of time. I wlll certainly do my best to get the other disciplines
involved in this unique and powerful new medium.
[N.B.: I am of course in no way deprecating the great value or contribution
to knowledge of the two disciplines I mentioned; I just believe that their
incidental monopoly over the electronic networks should be benignly dissolved
as soon as possible by the entry of the other disciplines that have a hand in
the written word, scholarly communication and the advancement of knowledge.
The interdisciplinary field of cognitive science happens to be a microcosm of
this larger problem of temporary disciplinary imbalance on the Net,
and the subfield of artificial intelligence -- though of course legitimately
skewed toward computer science -- seems to be showing some of its effects too,
especially on foundational topics like the symbol grounding problem.]
--
Stevan Harnad (609) - 921 7771
{bellcore, psuvax1, seismo, rutgers, packard} !princeton!mind!harnad
harnad%mind@princeton.csnet harnad@mind.Princeton.EDU
------------------------------
Date: 10 Jul 87 11:32:00 EST
From: "Robert Breaux" <breaux@ntsc-74.arpa>
Reply-to: "Robert Breaux" <breaux@ntsc-74.arpa>
Subject: SYMBOL GROUNDING DIES DOWN
It occurs to me that the flare up then dying of symbol grounding in
the ai list is an evolution not possible until recently. I believe
it is good. In the "old days" prior to electronic bulletin boards,
this argument would have raged for years, camps divided, universities
would have created "schools of thought", and perhaps books written
which would not have stood the "test of time" as a classic issue.
Now, we can have "face to face", so to speak, discussions early on,
resolve the issues which are not "classic" or seminal, and get on
with it.
It's GREAT, wouldn't you say?
------------------------------
Date: 9 Jul 87 03:41:27 GMT
From: mind!harnad@princeton.edu (Stevan Harnad)
Subject: Results of Symbol Grounding Poll: Yea's (2nd of 3 parts)
[These are the 24 yea's in response to the poll on whether or not to
continue the symbol grounding discussion on comp.ai/comp.cog-eng. I have
removed names and addresses because I had not asked for permission to repost
them. If you wish to communicate with anyone, specify by number (*and* whether
"yea" or "nay") and I will forward it to the author.]
-------------------------
1. I am finding the symbol grounding discussion very interesting and would
like it to continue. More generally, the community is better served by having
too much information flow than too little. I hope the discussion will continue
even if most respondents to your poll disagree.
----------------------
2. 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 germane 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.
---------------------
3. I vote to continue on symbol grounding. And by all means, keep going
with your good and interesting work.
-------------------
4. I don't read this discussion anymore. I couldn't find the beginning,
and never felt that I really understood what the problem was.
However, I have absolutely no objection to the discussion continuing.
I presume that the discussants get value out of it.
----------------------
5. Although I only peruse most of the symbol grounding discussion I think
it is well placed in comp.ai and I vote to see it continue. Personally, I do
not see why intelligent use of the NET needs to be defended but apparently
there is always an 'offended' party.
---------------------
6. [re. ailist] I initially found some of the symbol grounding discussion
interesting, but at the moment it is getting in my way, interfering with my
work of reviewing what is already too much material in AIList. Perhaps a
general solution to "what belongs on AIList" is to put lengthy, continuing
discussions which are of a temporary nature in separate issues, each clearly
titled so it can be deleted by the recepient at the title level without danger
of deleting other AIList topics.
[Ken Laws, Ailist's moderator, then replied that he was sorting already]
Thanks for the reply. Indeed, you are sorting the material already.
Thanks for the reference to the mail scanning program. It, or an enhancement of
the one I am using, could fill the bill nicely. Perhaps a one-character
appendage to the digest name to indicate the issue pertains exclusively to a
continuing lengthy discussion? Then, if desired, a smart mailer could
automatically omit or delete them. Just a thought.
------------------------
7. I would, with the following reservation, vote against splitting
off this discussion. It is tangential to some important aspects
of AI and discussions of this sort tend to emphasize areas which
need further scientific exploration.
My reservation, which I have until now contained, is that your
contributions do tend to be lengthy, wordy, vague, and full of
(sigh) ungrounded symbols. At times they also appear to lack
respect for the views of other contributors. If you're looking
for a soapbox, please find one that doesn't appear in my
mailbox. If you have a point to make, and can do so precisely,
concisely, and with an open mind towards the responses you receive
and respect for their contributors, please contribute to the AIList.
This is offered in the spirit of constructive criticism, and I hope
you can accept it as such.
----------------
8. I think symbol grounding discussion are *very* critical to
the AIList and count me as pro-discussion on the AIList.
--------------------
9. Ha! I subscribe to quite a few bulletin boards. The symbol grounding
problem is the only discussion topic for which I religiously archive all notes.
It's far, FAR more important than 99.9% of the drivel you see on the net.
What are your critics suggesting? Free up more slots for dumb jokes and
sophomoric opinions about the nature of intelligence? I say, "Right on! Keep
the symbol grounding discussion going."
If you want to be magnanimous, you might request that the discussion
be confined to one bulletin board. It seems to inhabit ai, cog-eng, and language
boards, at least, now. If you decide to start your own board, however, please
let me know.
---------------------
10. Please continue! Critics who care would notice that (in the ailist
version at least) these discussions are usually in a posting on their own, and
are thus easily discarded by those uninterested.
---------------------
11. Mark one with thumbs up.
-------------------
12. As per our phone conversation this morning... continue the dialogue.
-------------------------
13. Please continue the very enlightening discussion on symbol grounding in
its present arena. And thanks very much for the effort you put into explaining
quite carefully what you propose.
----------------------------
14. I consider the recent discussion on the symbol grounding problem to be
very interesting and relevant. Please continue.
--------------------------
15. What I am doing is responding to your poll request. Please continue
the discussion of the symbol grounding problem. I have not had time to
contribute, but I find the contributions, especially yours, quite valuable.
(Your contributions are good, but I also value "bad" contributions, since they
are often clear examples of the bad philosophy and epistemology which people
inflict on themselves and others.) My vote: continue posting.
-------------------
16. Despite the complaints from McCarthy and Minsky, there does seem to be
some benefit of the Symbol Grounding discussion for we lurkers. Sometimes I
almost think I understand what the issue is.
However, I do find it distracting that essentially the same material is
arriving by both comp.ai and comp.cog-eng newsgroups. I don't want to
unsubscribe to either, but I'd like to have to see the material only once.
Is it possible to move this discussion to just comp.cog-eng, since it
seems to be the (weak?) AI community that finds much of this correspondence
tiresome?
I think if you simply announce your intention to operate on one group,
and then make all your submissions there (while monitoring both, of course),
the news stream will become a bit easier to cope with for many of us.
[See earlier material on filtering multiple postings.]
---------------------------
17. I followed your early discussion in symbol grounding but now skip
over it. Maybe its gone on too long? But * as long as Ken Laws [ailist]
separates it into its own volumes * (as he has been doing) I can skip it and
others can follow it as they wish. If he decides this is too much work for him,
I would suggest moving it to a different forum.
--------------------------
18. I find the discussion of symbol grounding useful and worth
contuinuing. I vote to continue.
-----------------------------------
19. You get my vote for continuing the discussion.
---------------------------
20. Simple response. I don't participate, but I enjoy the discussion.
I'm a novice in this area, and seeing exchanges like this help educate.
-----------------------
21. Yes I find it useful and worth continuing.
[Mild ad hominem remarks about a prior rude poster deleted]
-----------------------------
22. My response to your request for a vote: I am emphatically *FOR*
keeping discussions such as the symbol grounding discussion *ON* Ailist Digest.
Though I don't always read all of them (I'm amazed at your energy and
ability to sustain these discussions on "paper") as a philosopher I find
discussions such as yours the the most important part of the digest. If
people think that AI is just computer science, let them start another list.
Laws obviously thinks that these discussions are part of AI and he's right.
I think that your policy of initially ignoring the rude remarks made
against you was a good one. It is unfortunate that some people lose their
manners when they go electronic.
----------------------
23. I vote that you continue the symbol grounding discussion and related
topics in the present forum. I've found these articles to be far more
enlightening, useful, and relevant than the typical requests and responses
for the latest references on KB techniques or expert systems marketing. Not
to say that such articles are inappropriate, but that this forum is for all
AI-related discussion. Please continue to ignore Booth and Harwood.
-------------------------
24. A difficult question. The discussion HAS been going on at considerable
length, but it evolves, and maintains a certain interest. Many people
(including me) seem not to work from the same foundation as you, and
therefore you need many words to get across what often sounds like
reiterations. But if you used fewer words, perhaps we might misunderstand
worse than we do.
Personally, I think you skirt some important points about
categorization, which may be in your book: that it is probably required
only for communication (perhaps for a conversation within a single brain,
as Gordon Pask would insist); that it usually depends on the existence
of a catastrophe function (anywhere near the border of a category,
the data may lead unequivocally to more than one result depending
on historic and local context); that symbols need not be grounded
in real-world phenomena, but in agreed categories constrained by context
(people DO communicate about religion and politics, in which fields
there is unlikley to be any real-world grounding of the symbols).
There are probably other issues. As for continuing the discussion,
I would say yes if the contributions could be kept under 75 lines,
no otherwise. Or else act as a moderator and submit weekly digests
of the arguments people send you privately.
------------------------
--
Stevan Harnad (609) - 921 7771
{bellcore, psuvax1, seismo, rutgers, packard} !princeton!mind!harnad
harnad%mind@princeton.csnet harnad@mind.Princeton.EDU
------------------------------
End of AIList Digest
********************
