COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	43 l.-3 Is there always a deeper explanation?
C00012 00003
C00015 ENDMK
C⊗;
43 l.-3 Is there always a deeper explanation?

Can a purpose ever be proved or does it depend on circumscription?

142 Why can't a simulation "hurry"?  If it can't "hurry", perhaps
it can ""hurry"".

148 She points out that Dreyfus has 3 not 2 challenges.

154 "the human neurophysiologist may never be able to offer more
than the general principles of human brain function".  Well maybe,
but surely pessimism about this is premature.

169 Identifying motives and emotions - McD didn't.

176 promissory notes

181 McD's list of instincts.  Not very plausible.

212 her diagram of compound emotions doesn't take into account more
complex structures, e.g. servomechanisms

268 Are consciousness and goal-seeking contingently or necessarily
connected?  Besides the ways of investigating the question discussed
by Boden, it can also be investigated by constructing systems that
exhibit one feature and not the other.  It seems to me that either
is possible without the other.  The simple goal-seekers have very
little if any consciousness.  In particular, they never manipulate
any idea of self.  On the other hand, it is possible to make a purely
passive reasoner, that generates consequences of its previous beliefs
and its observations and which does not pursue any goals - either in
the external world or among its own thoughts.

From this point of view, it is hard to understand Wittgenstein's contention
that a private language is impossible.  It is quite easy to construct
a computer program that reasons in some language and never receives
or emits expressions in it.

274 Boden decides for a necessary relation between consciousness and
goal-seeking.

280 she says that consciousness needn't be ascribed to machines

287,320 Boden and Miller, Galanter and Pribram are mistaken in supposing that
"so long as people are behaving %2some%1 plan or other must be
executed".  Executing a plan is a complex form of behavior, unless the
notion of plan is taken quite differently from either the common sense
or the AI usage, and many simpler behaviors exist.

288,320 The question of whether there is something like psychic energy in
humans should remain open.  Certainly, there is no need for anything
of the kind in computer programs.  Their operation is entirely logical,
and the amount of electrical energy used by most computers is almost
independent of what they are doing.  The computer may use just as
much energy in a wait state as in active computation.  Suppose we
ask whether the amount of mental activity of a human can be attributed
to some kind of energy that can be used up independently of muscular
energy and which can be renewed by processes other than eating and
breathing, e.g. by being inspired by some event.  In my opinion, the
probability that something like that will be found is not high, but
it can't be excluded.  McDougall may have been impressed by the
difference between mentally active and mentally sluggish people, and
so am I.

291 The separation of conation and cognition is somewhat dubious.  It
is true that without at least one %2ought%1 statement, no other can
be derived, but given just one, almost every further derived statement
can contain an ought, i.e. be a conative statement.  This is again a
contingent matter.

306 I see no reason to be discouraged about eventually reaching a
detailed understanding of how ideas are represented in the human
nervous system.  It took a long time to crack DNA, but progress in
that area is now very fast.  It may also take a long time to crack
the nervous system, but cracking molecular biology will help directly
and in suggesting the kind of effort required.

308 Boden doesn't seem to have noticed the problem of modeling
partial information and the problems thereof, c.f. Moore's discussion
of disjunctions.

317 Consider the hemiplegic who ignores half his body.  Since his verbal
concept of his body is unitary, it must be that washing and dressing
etc. have been programmed (in order to save memory) to have been
triggered by proprioceptive signals from the parts of the body.  This
seems to be the most plausible explanation of how loss of these signals
can cause him to dress only half the body.
Another person might use different cues to keep track of where he is
in the program, and so might dress and wash ok even with the hemiplegia.
Introspecting, I think I might dress ok, because it's governed by sequence
but might wash only half - that being triggered by signals from the
unwashed part.

324 "But the same may be sadid about the physical and the information-processing
aspects of computers".

328 I agree that one can view man as a machine without becoming demoralized,
but this doesn't argue one way or the other about the truth of the proposition.
A scientist should try to find out the truth irregardless of whether
the truth will make people feel good or behave well.

331 The discussion of morality and character is likely to be useful,
but we are a long way from being able to simulate or even axiomatize the
growth of character.  The discussion of freedom can be made more precise.

336 The purposes of all present programs are unconscious.
The idea that information about purposes is present "at a higher level of the
mind" is persistent but probably wrong and maybe even paradoxical, since
one can think about "my present thoughts".

McDougall and Boden are correct in rejecting hedonism as an explanation of
behavior.

What was his purpose in doing X?  He did X in order to achieve Y.

Is there not a goal of having goals?

	The idea of sketching a simulation of a person based on
McDougall's scheme is a good one.  It's complexity is closer to
reality than that of any behaviorist theorist.  If Boden had actually
attempted to carry out the simulation, she would have learned the
inadequacy of some of her and McDougall's basic schemes.  (I say this
on the basis of contemplating cognitive simulations that have been
done).  Namely, the relations between concepts are still too
associationist.

	Consider the diagram on p. 213 offering alternate structures for
the sentiment of reverence.  If these were expressed in the
language of logic they would be simple propositional compositions.