COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "memo.pub[let,jmc]" source
C00021 00003	.cb CHALLENGE TO A DEBATE
C00024 ENDMK
C⊗;
.require "memo.pub[let,jmc]" source;
.cb REPLY TO WEIZENBAUM'S REPLY and CHALLENGE TO A DEBATE

	This is a response to Weizenbaum's response to my draft review.
Weizenbaum's response is weizen.ans[pub,jmc].  The draft of the
review will be further modified, and this may make some of his
points puzzling to the new reader.  First I will respond to his
points as he makes them, and then I will raise further questions
on the issues.

	My sentence "It also owes much in rhetorical style and
political presupposition to the new left" apparently means different
things to different people and will have to be omitted unless
I can find a way of putting the point more clearly and still
concisely; it is not central.  What struck me was the
chapter title "Against the imperialism of instrumental reason"
the phrase (p. 37) "imprisons poor people in inner cities...",
and following the phrase about "science ...converted into poison"
on p. 13 by an association of "logicality" with American policy
in Vietnam which he assumes his readers must agree was "murderous".

	With some readers the above explanation may lose me credence,
because they will agree with Weizenbaum about Vietnam.  I request
such readers to make their own judgment about how relevant this is
to my main criticisms of Weizenbaum.  I put in the point, and refuse
to drop it, not in order to strengthen my case about the anti-science
tendency of the book, but specifically to defend the Defense
Department.  I think the country has suffered much and will
suffer more because of the estrangement of scientists from
defense.

	In the response Weizenbaum says "I would have thought, but
for this exhibit, that all participants in scholarly debates had
by now renounced argument by irrelevant political association".
However, later in the response, he defends referring to the
Department of "Defense" as merely pointing out that the emperor
wears no clothes.  However, the main way in which the Defense Department
is used in the book is in order to criticize ideas by association
with it.

	Nevertheless, whether the Defense Department is good or bad
is not an issue on which either of us has much new to say,
and if Weizenbaum would reformulate his views omitting attacks on it,
I would cheerfully omit defending it.

	I think I should reformulate my comment about the relation
of the book to academic politics.  However, the computer linguists
of my acquaintance are offended by the persistent refusal of the
Chomsky school to acknowledge their contributions (they switched
emphasis to semantics before Chomsky did) and to discuss their
contentions.  Morris Halle's report to the M.I.T. conference on
the centenary of the telephone confirmed my impression that
the Chomsky school is behaving arrogantly.  Weizenbaum's acceptance
of their arrogant dismissal of computer linguists as hackers
in a book in which he acknowledges the scientific value of much
of the computer linguists' work seems obtuse.
Moreover, whether Weizenbaum intends it specifically or not, he
is off-handedly confirming the view that the M.I.T. Linguistics
Department need not look beyond it own followers in making
appointments, and when a physicist as prominent at M.I.T. as Weisskopf
endorses the book, one can see that the book is
effective, and one can see the usefulness of the mutual
admiration society incorporating Weizenbaum and Chomsky.
(I refer to the quotations in an ad in current ⊗Scientific ⊗American). 

	I understand his point that because of a difference in
experience, computers can't understand humans.  I disagree and
cite the counter-example of the deaf-dumb-and-blind who overcome
their handicaps and develop world-views surprisingly similar
to those of people without those handicaps.

	With regard to questions of "power", I think that the
technology assessment movement is wrongly attempting to get
power of approval over scientific work and that Weizenbaum
is helping them.  When Colby's work is regarded as based
on an obscenity, the question of how to prevent such work
arises in the minds of many people and proposals have been
made, e.g. by the %2Science for the People%1 thugs.

	Now we have a disagreement about the word "cure".  I
prefer not to say %2Prefrontal lobotomy cures people but
destroys essential parts of their personalities%1, but rather
%2Prefrontal lobotomy doesn't cure people, because it destroys
essential parts of their personalities%1 assuming this
criticism of prefrontal lobotomy is correct, which I gather
is the present general opinion.

	A similar remark applies to the use of the words "rationality"
and "logicality".  When someone, e.g. Meadows or the Defense
Department, puts what I
consider wrong assumptions into a computer program and
comes up with wrong answers, I don't want to criticize rationality
or logicality, but merely the wrong assumptions (which
may be merely wrong data or worse yet built into
the program).

	I can't continue the argument about whether
a computer program could cure people "in a sufficiently encompassing
interpretation of that word", because there is no
candidate that I would support.  Moreover, I think
work on a computer psychotherapist is premature until
a better understanding of mental process is achieved.  However,
I take it that Weizenbaum considers the necessary alienness
of computers sufficiently proved that all work aimed at computer
psychiatry is "obscene" per se.

	Weizenbaum has caught me in a misstatement about science
being the main source of knowledge, and his examples of common
sense knowledge demonstrate it.  However, his quarrel with excessive
claims for science do not consist in saying that people wrongly
claim that science proves the paternity of their children.  Let
me revise my contention to assert that science is the main reliable
source of general knowledge including knowledge about human
behavior.  However, science is regrettably slow, and areas exist
in which present scientific theories may be worse than
unguided common sense.  Let me admit that I am not satisfied
with this formulation.

	Weizenbaum indeed has quotes that counter his criticism
that AI hasn't produced industrial results in 20 years.  As I
remark, many of his assertions are coupled with counter-assertions,
and his inconsistency is a major fault.

	My analysis of the import of "Those who know who and what
they are do not need to ask what they should do" is based on
my encounters with would-be psycho-priests, not on assuming that
Weizenbaum is one of them.  On rereading its context, I can see
that Weizenbaum doesn't intend to set himself up as a judge of
whether someone "knows who and what he is".  Unfortunately,
whenever a vague statement like that is ascribed great
importance by a social group, it acts as an attractant
for would-be priests.  At least, I have seen this very question
used in encounter groups to put people in a position in which
every answer is rejected as trivial or wrong.  My further opinion
is that "what should I do" and "what may I do" and "what do I
want to do" are more fruitful questions than "who and what
am I?".  At least I can understand these questions and
try to answer them for myself, while I can only give
trivial answers to the latter.
(There is a poignant article entitled "Trashing: When 'Sisters'
turn on you" in the April 1976 ⊗Ms.  The author complains
that women are unjustly attacked within the women's movement
for "what they are" rather than for anything they have done
and that it is impossible to defend oneself against such an
attack.)

	I don't claim to have proved in my review that Mumford,
Roszak, Ellul or the new left are wrong, but when Weizenbaum
says without supplying supporting argument "Men could instead
choose to have truly safe automobiles, decent television, decent
housing for everyone, or comfortable, safe, and widely distributed
mass transportation", he is making presumptions about technology,
economics, the definition of social justice, what constitutes
good taste, and possibly about what government action is legitimate
in order to achieve social goals.  I suspect Weizenbaum and I
could agree on what TV commercials are offensive.  We might find
it harder to agree on how TV should be financed or whether it
should be done in a uniform way.  I take it back.  He might not
agree with me in finding the nagging self-praise of the "non-commercial"
stations even more offensive than commercials for vaginal
deodorants.

	Lastly, if there were a plausible model man's beliefs,
I think that it would be mentioned in the admittedly few sources
I have read, but I admit a certain probability that someone has
found the truth in this area but it hasn't achieved prominence.
My own experience has been that if one of my ideas didn't achieve
acceptance in ten years or so, I could usually find something
wrong or at least grossly incomplete about it.

SOME QUESTIONS

	In the foregoing, I have dealt as best I can with all the
issues raised in Weizenbaum's reply to my review, but I think
he has not dealt with some important questions raised in the
review.  Indeed he hasn't dealt with them seriatim at all.
Here are some issues I would like him to address
specifically:

.item←0;

	#. Clarify the position with respect to research
on recombinant DNA expressed on p. 260.  Since the issues
have been much discussed in the scientific and popular press,
it would help to say how to apply the principle "to exempt
life itself from the madness of treating everything as
an object".  The biologists were apparently concerned only
with the possibility that something infectious would escape
the laboratory.

	#. Do the hackers Chomsky is justified in ignoring
inlude Winograd and Schank?  Compare the goals of the two
groups and say why one line of research is acceptable and
the other is "at best ... another example of the drunkard's
search". p.200

	#. Is the mini-theory about clocks correct?
Don't the hackers eat more spontaneously than people did before
clocks were invented?

	#. What is the the definition of obscenity, and how
can a poor AI graduate student be sure he won't be guilty of it?
Perhaps you can illustrate the difference by comparing the work
of Schank and Chomsky.

	#. Should present support of research in AI be continued?

	#. Explain the phrase "the dogmatic coupling of reason
to power" in your reply.  I simply don't understand it.

	#. Explain the import of the statement
%2"Scientists who continue to prattle on about 'knowledge for its own
sake'  in order to  exploit that  slogan for their  self-serving ends
have detached science  and knowledge from any  contact with the  real
world".%1  Don't scientists have some reason to fear what may come of
Senator Kennedy's reading it?
Who are the prattlers in the computer science field?

	#. Please name a compulsive computer programmer or two.  If
you don't want to do so, say how many employees, including graduate
students, of the AI Lab and the Laboratory for Computer Science are
compulsive computer programmers.


.skip to column 1

.cb CHALLENGE TO A DEBATE


	I would like to get Weizenbaum's and my positions on the more
important  issues raised  in his  book better  defined.   The present
procedure of exchanging polemics  has the disadvantage that it  tends
to concentrate  on the  areas in  which verbal  slips have  been made
rather   than  on  the  areas  in   which  the  positions  have  been
well-formulated.

	%3Therefore, I offer a challenge to a debate.%1

	If possible, I would like to  eliminate all issues concerning
whether  the Defense Department  is good  or bad or  the intellectual
source of Weizenbaum's or my views, or academic politics.  Here  is a
list  of questions  that  might be  discussed, but  I  will agree  to
modifications.

.item←0;

	#. Has science become "a slow-acting poison" in any sense of
the word?

	#. Is it impossible for a computer to understand humans?

	#. Is  there a concept  of obscenity that  can be  applied to
scientific  research?  Is it  immoral to couple an  animal brain to a
computer?  Is it immoral to try to make a computer program  that will
give psychiatric  treatment -  assuming the precautions  customary in
medicine  for  testing  efficacy  and  side-effects?    Is  the  ARPA
supported speech research immoral?

	#. Do compulsive computer programmers exist in a significant
sense?

	#. Is the AI approach to natural language of Winograd and
Schank a reasonable scientific enterprise?

	I will also be glad to defend AI  research it it is attacked,
but  I  gather  that  it can  be  stipulated  that  this research  is
proceeding reasonably if not optimally.

	An oral  debate at  M.I.T. at  some mutually convenient  time
would  suit me  fine, but  if this  is  not agreeable,  I am  open to
counter suggestion.