vu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) (03/12/88)
In article <45134@sun.uucp> randolph@sun.UUCP (Randolph Fritz) writes: >Cats sometimes move their eyes rapidly while sleeping. The motions are similar >to the rapid eye motions (REM) of a human dreamer; cats probably dream. >Apparently cats can operate on representations not physically present . . . >don't know about gorillas, though. I believe that it is quite clear that cats and other animals dream. But I wouldn't want to call dreaming a kind of thinking. The difference is that dreams are uncontrolled. In other words, I don't think you can say that the dreamer is operating on the dream images (representations). . .more like the dreamer's brain (as a part of the dreamer), or a *part* of the dreamer's brain (i.e. the "dream center," "libmic system" (BTW, what is the limbic system?)) is operating on them. Uh oh, I see an argument about personal identity coming on. . . >Conversely, many humans engaging in those activities called "thinking" move >through an imagined environment. Yes, this is my contention. >Human activity and animal activity is, in >this way, quite similar. But I would deny that this follows, since I don't think dreaming is thinking. >Is there a qualitative difference? I believe there is, which is the ability to control the generation and manipulation of representations. Symbolic activity is evidence of this ability, and to the extent that we can teach animals to talk (speak, sign) perhaps we can teach them to think (?). Please note that my theory leaves the point of animal and machine intelligence as empirical points, offering no theoretical solution. O----------------------------------------------------------------------> | Cliff Joslyn, Professional Cybernetician | Systems Science Department, SUNY Binghamton, New York, but my opinions | vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . .
randolph%cognito@Sun.COM (Randolph Fritz) (03/12/88)
In response to an article of mine, Cliff Joslyn
(vu0112@bingvaxu.cc.binghamton.edu) writes:
I wouldn't want to call dreaming a kind of thinking. The difference is
that dreams are uncontrolled.
and later,
[the qualitative difference between human and non-human animal
intelligence] is the ability to control the generation and manipulation of
representations.
In many senses thinking is not controlled. Can you stop yourself from
visualising a color when you hear "red" in the appropriate context? Change
the way you feel when you greet your lover? Change the meanings you use for
the symbols of C, if you know the language? [To step briefly to an unrelated
discussion, this is part of making an effective advertisement -- choosing
symbols to which a large group of people respond uncontrollably in some
desired way.]
Humans can plan strategies which result in changes in their own responses
(learning). Yet we do not control that which make us want to learn nor
that which makes us choose particular subjects. I wonder . . . is the
distinctive quality of human intelligence simply that we want to learn? Or is
that, too, something we share with other animals?
__Randolph Fritz
randolph@sun.com
sun!randolphvu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) (03/13/88)
In article <45232@sun.uucp> randolph%cognito@Sun.COM (Randolph Fritz) writes: >In response to an article of mine, Cliff Joslyn >(vu0112@bingvaxu.cc.binghamton.edu) writes: > > I wouldn't want to call dreaming a kind of thinking. The difference is > that dreams are uncontrolled. > >and later, > > [the qualitative difference between human and non-human animal > intelligence] is the ability to control the generation and manipulation of > representations. > >In many senses thinking is not controlled. Can you stop yourself from >visualising a color when you hear "red" in the appropriate context? Change >the way you feel when you greet your lover? Change the meanings you use for >the symbols of C, if you know the language? These are all examples of "unconscious thought", i.e., these things just "happen" to you, these representations "appear" in your mind, sometimes we know why, sometimes not. *IS THIS THINKING?* If so, then what is the difference between thinking and complex reflex? Again, *IS DREAMING THINKING?* I think you'd be hard-pressed to say so. >Humans can plan strategies which result in changes in their own responses >(learning). Yet we do not control that which make us want to learn nor >that which makes us choose particular subjects. I wonder . . . is the >distinctive quality of human intelligence simply that we want to learn? Or is >that, too, something we share with other animals? It has been often said (no, I can't quote!) that the distinctive aspect of human mentality, perhaps a definition of intelligence, is our ability to meta-learn, that is, learn how to learn, change our learning strategy, etc. Presumably learning *per se* is an instinct in both humans and non-humans (of course, *what is learned* is not instinctive). The theory says that in humans, we can *control* learning, just as we can control (some of) our other instincts (e.g. fasting, biofeedback, football :-)). If this is so, then I would say that *imagination* is the mechanism by which we control learning, in the ways I've outlined before, e.g. generation and testing of representations, etc. O----------------------------------------------------------------------> | Cliff Joslyn, Professional Cybernetician | Systems Science Department, SUNY Binghamton, New York, but my opinions | vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . .
randolph%cognito@Sun.COM (Randolph Fritz) (03/13/88)
In response to an article of mine, Cliff Joslyn
(vu0112@bingvaxu.cc.binghamton.edu) writes:
These are all examples of "unconscious thought", i.e., these things just
"happen" to you, these representations "appear" in your mind, sometimes we
know why, sometimes not. *IS THIS THINKING?* If so, then what is the
difference between thinking and complex reflex?
I've the strong impression that "thought" for you has at least the following
two properties:
1. Conscious control
2. Uniquely human behavior
You seem to presuppose that I use the same two criteria; my arguments must
sound like so much noise to you. If you choose to use these criteria, fine.
I wouldn't want to control :-) your thinking.
Yet these two criteria leave out so much behavior that they seem to me
limiting. Nor is it well understood behavior. Clinical psychology is
largely the long slow study of methods of changing unconscious behavior; an
artificial intelligence researcher would be delighted to be able to produce
the behaviors you disparage as "unconscious".
__Randolph Fritz
randolph@sun.com
sun!randolphvu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) (03/14/88)
In article <45257@sun.uucp> randolph%cognito@Sun.COM (Randolph Fritz) writes: >I've the strong impression that "thought" for you has at least the following >two properties: > > 1. Conscious control > 2. Uniquely human behavior Well, maybe, especially point 2 (I said we might be able to train animals to think). >You seem to presuppose that I use the same two criteria; my arguments must >sound like so much noise to you. If you choose to use these criteria, fine. >I wouldn't want to control :-) your thinking. No, of course not noise! I thought we were engaged in a debate about the definition of thought? How can we engage in debate if I think you have the same assumptions I do? If argument is thought control, so be it, all the better. It's a good think that if I lose an argument my thought is changed to reflect those of the one who won the argument. This is the way we learn things. >Yet these two criteria leave out so much behavior that they seem to me >limiting. Nor is it well understood behavior. Clinical psychology is >largely the long slow study of methods of changing unconscious behavior; an >artificial intelligence researcher would be delighted to be able to produce >the behaviors you disparage as "unconscious". Hey, you're getting me upset! I don't understand the above, perhaps you could give some examples? Behavior is not thought. I disparage nothing. I seek truth. Please answer my question: what is the difference between thought and complex reflex? >H_Randolph Fritz >randolph@sun.com >sun!randolph O----------------------------------------------------------------------> | Cliff Joslyn, Professional Cybernetician | Systems Science Department, SUNY Binghamton, New York, but my opinions | vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . .
hadj@sbcs.sunysb.edu (Michael Hadjimichael) (03/15/88)
In article <927@bingvaxu.cc.binghamton.edu>, vu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) writes: > In article <45134@sun.uucp> randolph@sun.UUCP (Randolph Fritz) writes: > >Cats sometimes move their eyes rapidly while sleeping. The motions are similar > >to the rapid eye motions (REM) of a human dreamer; cats probably dream. > >Apparently cats can operate on representations not physically present . . . > > I believe that it is quite clear that cats and other animals dream. But > I wouldn't want to call dreaming a kind of thinking. The difference is > that dreams are uncontrolled. In other words, I don't think you can say > that the dreamer is operating on the dream images (representations). . > .more like the dreamer's brain (as a part of the dreamer), or a *part* > of the dreamer's brain (i.e. the "dream center," "libmic system" (BTW, > what is the limbic system?)) is operating on them. In reference to dreaming cats... In experiments at the University of Pennsylvania, researchers removed parts of the pontine area in cats (this is the area which inhibits the body from acting on messages from the brain during REM sleep) and then allowed them to enter REM sleep. As a result, the cats acted out their dreams, acting as if they were awake. The only way to tell that they were asleep was by noting that they were not responsive to external stimuli. I realize that this doesn't show that the cats are thinking, but it does show that interpretable signals are being sent out by the brain. We can't really say what causes these signals to be generated, but we can see that the cat is indeed acting on them (or trying to act on them, in the case of a normal sleeping cat). A reference to the experiment mentioned above may be found in SCIENCE 80 Magazine, Vol.1, No.4, May/June 1980, pp36-43. -mike hadjimichael. -- / "We are creatures of love..." - Talking Heads \ \ / / hadj@sbcs.sunysb.edu {philabs, allegra}!sbcs!hadj hadj@suny-sb \ \ departmentofcomputersciencesunystonybrookstonybrooknyoneonesevenninefour /
randolph%cognito@Sun.COM (Randolph Fritz) (03/27/88)
Cliff Joslyn (vu0112@bingvaxu.cc.binghamton.edu) writes -
Please answer my question: what is the difference between thought and
complex reflex?
I would say that complex reflex is a kind of thought.
Also, as part of a comment that Cliff defines thought as "consciously
controlled human activity", I wrote:
Yet these two criteria leave out so much behavior that they seem to me
limiting. Nor is it well understood behavior. Clinical psychology is
largely the long slow study of methods of changing unconscious behavior; an
artificial intelligence researcher would be delighted to be able to produce
the behaviors you disparage as "unconscious".
Cliff writes;
Hey, you're getting me upset! I don't understand the above, perhaps you
could give some examples? Behavior is not thought. I disparage nothing.
I seek truth.
Hmm. I guess that was a bit terse. Ok, here's a longer explanation.
"Clinical psychology is largely the long slow study of methods of changing
unconscious behavior." A client who voluntarily visits a psychologist does so
be cause he has mental behaviors ("thoughts") which he can't consciously
control. If the client was aware and in control of his behavior, he could
have solved his problem without assistance! Thus, a large interest of
clinical psychologists is the changing of unconscious, uncontrollable
mental behavior ("thought").
"an artificial intelligence researcher would be delighted to be able to
produce the behaviors you disparage as 'unconscious'". If I could build a
machine that could, from the sensation of reflected light from a page,
construct a representation of the words on the page (the "meaning" of the text
on the page), I'd have achieved something that would awe an AI researcher. It
would turn the world upside down! And yet, in a skilled human reader, reading
is unconscious mental behavior ("thought").
The transformation of sensation to representation, the operation on
representation, the transformation of representation to action (I'm tempted to
write sensation, but that would cause more confusion), all of these are
unconscious. And, so far as I know, humans are better at them than any other
creature; we have more options in our behavior, we can create a broader range
of representations, we do more things with them. Not thought? This is most
of what minds do!
__Randolph Fritz (sun!randolph; randolph@sun.com)vu0112@bingvaxu.cc.binghamton.edu (vu0112) (03/27/88)
In article <47166@sun.uucp> randolph%cognito@Sun.COM (Randolph Fritz) writes: >Cliff Joslyn (vu0112@bingvaxu.cc.binghamton.edu) writes - > > Please answer my question: what is the difference between thought and > complex reflex? > >I would say that complex reflex is a kind of thought. I must admit that after reading your posting I have become more sympathetic to your position. In particular, the idea that *uncontrolled*, but still spontaneously generated (as opposed to perceived) mental representations are thoughts seems unaviodable. So that when I close my ears and hear voices, or close my eyes and see pictures, it must be said that I am thinking. Yet *when* I control those voices and pictures, that is a different kind of thought. Any suggestions as to a name? So while I'll admit that thoughts need not be controlled generation of mental representations (although this is typical), on the other hand I must insist that complex reflexes are not thoughts. Do you want to argue that *all* mental activity is a kind of thought, including perception? If so, then you must say that *all* neural creaturres are also *thinking* creatures, something which is not really common sensical. Indeed, if you allow complex reflexes to be thoughts, then to the extent that algorithms are complex reflexes (this seems unproblematic to me now), existing machines think! Perhaps we can draw the line between thoughts and non-thought mental phenomena based on thoughts being internally generated, as opposed to received from the environment. However, on this view we quickly run into problems w/hallucinations and other forms of "induced" mental presentation. In particular, when the neurosurgeon's probe or a stray neurotransmitter causes a mental presentation, is that a thought? The problem here is a viable definition of "environment:" is my endocrine system outside or inside? >"Clinical psychology is largely the long slow study of methods of changing >unconscious behavior." Thank you, I think I understand now. On the above reading, schizophrenic hallucinations and such may or may not count as thoughts, depending on the brain science involved. I have little idea what an unconscious thought might be. I'm inclined not to believe in them. What I will grant is that some (most? all?) thoughts are caused by unconscious processes (viz. my voices and images), but I couldn't possibly say that the thoughts themselves were unconscious. >If I could build a >machine that could, from the sensation of reflected light from a page, >construct a representation of the words on the page (the "meaning" of the text >on the page), I'd have achieved something that would awe an AI researcher. It >would turn the world upside down! And yet, in a skilled human reader, reading >is unconscious mental behavior ("thought"). I think you'd agree with me that the problem above is *very* difficult to define. OCRs construct representations of words on pages, but that's not what you mean, is it? What you mean is that the OCR must *understand* what the words mean. We all know the difficulties involved in finding criteria for "understanding." On the other hand, I suspect that if I was a highly skilled accountant or editor, that there could be many tasks working with numbers and words which I could accomplish with very sophisticated complex reflexes (algorithms), and thus without thought. Such activities may be normally and commonly accomplished *with* thought, but perhaps not necessarily so. And of course, automated accounting and editting is proceeding apace, yet no one would call a spell-checker a thinking machine, or understands anything about what it is doing. >The transformation of sensation to representation, the operation on >representation, the transformation of representation to action (I'm tempted to >write sensation, but that would cause more confusion), all of these are >unconscious. And, so far as I know, humans are better at them than any other >creature; we have more options in our behavior, we can create a broader range >of representations, we do more things with them. Not thought? This is most >of what minds do! I have no idea what an unconscious thought could be. Perhaps minds do more than think? Perhaps they do all kinds of other things, things like the kinds of things that computers do, like complex algorithms? People *do* think, in fact we're so good at it, and it demmands so much of us that sometimes we forget that brains and minds do other things, like keep us alive, as they have kept countless non-thinking creatures alive for billions of years. >__Randolph Fritz (sun!randolph; randolph@sun.com) O----------------------------------------------------------------------> | Cliff Joslyn, Professional Cybernetician | Systems Science Department, SUNY Binghamton, New York | vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . .