Re: [Robot-for-President] Re: Re: SUO: Re: Logic & ProgrammingLanguages
Hi Linda. If I am the "cognitive psychologist" below, I am flattered as
they are considered to be something of an elite in the profession. I
consider myself to be more of an eclectic or pragmatist as I will jump
from one school or model to another depending of what "works best" for the
circumstances. BTW, the first cognitive psych prof we got at U of A was
named...no kidding...Brainerd. When I got my first job in a clinical
setting, my boss was named...Nutter. What's in a name?
On Mon, 23 Jul 2001, L Misek-Falkoff wrote:
> Dear Seth:
>
> Does this have anything to do with differing value judgments about
> "instinct" and ":reason" ?
Once we have a simulation of human reasoning in a robot, does the
distinction between instinct and reason matter so much? We don't play
chess by instinct. We do so because we have learned to use a kind of
reasoning to play the game. But the chess-playing program installed in the
robot works just like instinct does it not? Present a real ant with a
complex situation and it will build a nest or herd aphids or whatever. It
does so by instinct. Present robo-ant with the same situation and it will
do the same. The program simulates the complex pattern of responses to the
complex pattern of stimuli. "Artificial instinct". The chess playing
program takes the complexity of chess playing which we humans handle by
learning and reasoning and for each stimulus pattern automatically
triggers the required response. Does it even matter to the result whether
"brute force" algorithms are used, as I was discussing with Jose Klingbeil
on BC Politics? Whether robo-arithmetician uses the algorithms of binary
arithmetic or uses "look-up tables" the result is the same. And the
machine process is....well, mechanical. It is after all a
machine. Machine-like and instinct-like seem very similar no matter how
sophisticated the machine.
Thanks for the question. It made me go back 35 years to a grad school
paper I wrote on instincts and apply some of that to the present-day world
of AI. That is one of the differences between me and some of the
robotics-AI people I argue with. I don't care so much about simulating
HOW the human brain or mind functions by using a machine. If the machine
comes up with WHAT the human comes up with, that's good enough for
me. Told you I was pragmatic!
> Recently I read a report reviewing what was described as Roger Schank's (as
> reviewed) arguments that some folks are discriminatory "fleshists," but I
> guess that concerns robots, not animals as per the often postulated
> distinction I have in mind.
Even a 100% mechanical humanoid like Honda's Asimo with some powerful AI
programming will find its share of "robot rights" advocates in society and
I don't think they should be considered on the whole as "nut cases". A
parallel development in society is biomechanics and cyborg technologies.
That will cause human society to wonder even more about robot rights and
robot sentience. No doubt we are on a path to a stage at which you lawyers
will be grappling with all kinds of robot rights issues. The TV series
"Total Recall 2070" is centred around this issue of "fleshism" ca. 2070.
Their latest model androids are actually machines with some flesh and
blood features, ie cyborgs. So there are more than two classes of
human-like entities on the planet in that fiction. Add a few more
categories due to in vitro gestation and artificial DNA in humans and
Robolawyers are going to be very, very busy. Speculations? Yes. But the
picture could get even more complicated than the few ideas I just
presented.
> Another candidate analogy, though analogies often don't work: neural memory
> and dna memory (suggesting CNS firing may link in with human concepts of
> reason or logic, DNA sequences to do with even more preconscious
> adaptation?)
I keep wondering about this little DNA-machine. We focus so much on
synaptic connections in the human brain as the sine qua non of human
intelligence. And some researchers will estimate so many hundreds of
billions of neurons and so many thousand synaptic connections per neuron
and so many firings per second...and they draw a parallel between that and
operations per second in a computer. Then we sometimes see statements as
to when computers will have as much processing power as the human
brain. But wait a minute. What about what goes on inside a cell, without a
synaptic discharge? Single celled organisms have no synaptic
discharge. But they do all kinds of things. They move, they seek food,
they ingest it, they reproduce, they avoid enemies...and sometimes they
learn. Isn't that little DNA-machine a biological computer too? It is the
one which builds what we have on our shoulders. Must have some kind of
smarts to do that.
"Preconscious adaption"...that's a very useful concept. It may be the
problem to overcome when it comes to developing a program for what they
call "natural language programming" or everyday language (and reading
ability). Strange isn't it, that we aren't fully conscious of such an
important human CREATION. I don't want to split hairs as I did with some
people previously on whether language/word usage is an invention,
innovation, creation or something homo sapiens kind of stumbled upon and
added to incrementally. Whatever it is, WE MADE IT UP. Now what I marvel
at is that we don't have conscious awareness of what the rules are for
word usage. If we did, we could have a conversing-reading machine pretty
fast could we not?
So we use 'preconscious adapation' for word usage. But that of course
doesn't mean language is preconscious to everyone. Remember the
philosopher Williams? I found him by pestering Philos-L with this
question. He said he was quite conscious of how words were to be used and
he could come up with the rules for grammar, reading etc. I couldn't
really figure him out and he was one of the most curmudgeonly people I
have ever corresponded with online. An eccentric genius with a gift worth
millions? Maybe. How much do you think IBM or Microsoft would pay for a
clear set of rules for how to converse and how to read which they could
turn over to their code writers? So there's an "intellectual property
rights" issue for you. It may well be that there are gifted people out
there who know these rules better than most of us, who can introspect and
turn the preconscious into the conscious. How do we get them to come
forward without having their gift given away, or stolen?
Best-FWP
> Don't know if these comparisons will kick in, but I did think about it and
> my instincts suggest they might. Forwarding to Neurosurgeon, and Cognitive
> Psychologist.
>
> :) L.
> ----------
> Dr. L. D. Misek-Falkoff, Speaker and Chronic Pain Chair
> The National Disability Party http://www.disabilityparty.com
> ----- Original Message -----
> From: "Seth Russell" <seth@robustai.net>
> To: "Chris Menzel" <cmenzel@philebus.tamu.edu>
> Cc: "Sergio Navega" <snavega@attglobal.net>;
> <standard-upper-ontology@IEEE.org>; <cg@cs.uah.edu>
> Sent: Sunday, July 22, 2001 10:59 PM
> Subject: CG: Re: SUO: Re: Logic & Programming Languages
>
>
> >
> > From: "Chris Menzel" <cmenzel@philebus.tamu.edu>
> >
> > > Seth wrote:
> > > > From: "Leo Obrst" <lobrst@mitre.org>
> > > >
> > > > > I tend to agree with John, as far as I can discern your original
> > issues.
> > > > > Much (most?) of what humans and other animals do are based on
> logical
> > > > > inferential processes, it seems to me (although this is disfavored
> in
> > > > > many literatures).
> > > >
> > > > Well me thinks you and John are playing fast and loose with this word
> > > > 'logical'. I don't think what happens when we flee from danger is
> > > > anything even remotely akin to an inferential process.
> > >
> > > Neither does Leo, I suspect. Fleeing is not reasoning. Leo is talking
> > > about the inferential process that leads to the fleeing.
> >
> > Sorry I also meant the process leading up to the fleeing
> >
> > >Action flows
> > > from both belief and desire. The belief component at least -- e.g., "a
> > > tiger is near", "tigers eat things like me", "if I run, the tiger won't
> > > see me", "tiger bad", etc -- requires reasoning.
> >
> > Yes, I would call that sequence of conscious events reasoning. But I am
> > simply saying that nothing like that sequence would ever happen in the
> > reality of this example, and if it did, it would have nothing to do with
> > triggering the actual fleeing.
> >
> > >You do some
> > > reasoning (very quickly) that leads to beliefs that, together with your
> > > desires (not to be tiger lunch, for example), lead to action.
> >
> > Your assumption seems to be (=> (and (belief a b) (desire b c)) (action
> > flee)). I am just saying there is no such implication. The action is not
> > triggered by such a structure, however coded. The action is triggered by
> a
> > different process having to do with the nature of triggering actions, and
> > not with the nature of beliefs and desires. Thinking yourself into
> action,
> > however reasonable your beliefs, never works. Just doing it, does. I
> > suspect that actions and their triggers, and beliefs and their reasons
> are
> > manifested by different parts of our brains; and have very little if
> > anything to do with each other. If I was studious I site actual
> neurology,
> > but that is not my field. I've cc Sergio Navega on this, perhaps he can
> > fill in the gap, or correct me if I am wrong.
> >
> > > > > Have these been compiled down to very quick responses? Yes. So one
> > > > > doesn't have to infer at run-time: jump from the set of premises to
> > > > > the conclusion (through the long intermediate chains, which have
> > > > > been compiled away) and run from the tiger.
> > > >
> > > > Again i doubt that a habitual (or conditioned) respons is anything
> > > > like being logical ... nor was the original response on which the
> > > > habit was formed logical.
> > >
> > > Granted, the inferential component no doubt becomes less prominent with
> > > regard to habitual behavior, but reasoning will arguably still be
> > > involved -- one needs at least to belief one is in the sort of situation
> > > where a certain habitual response can kick in.
> >
> > Well yes, one needs to recognize that one is in a situation before the
> > situation can trigger the behavior. But the recognition does not need to
> be
> > conscious. My thesis is that the conscious part of the recognition is
> > irrelevant to the actual action.
> >
> > > > I think logic is a game of rules for describing what happens after the
> > > > fact.
> > >
> > > Hm, where to start. Do you mean that we arrived at our logical "laws"
> > > simply by obvserving how people reason?
> >
> > No, logicians work out these 'laws' with pencil and paper.
> >
> > > But people reason in all sorts
> > > of ways, often bizarrely, as anyone can see by reading letters to the
> > > Editor of your local paper or listening to Rush Limbaugh. Why is it we
> > > only incorporate some of the things we observe and not others?
> >
> > I don't believe that the formal logic we practice today has stemmed from
> > such observations.
> >
> > >And,
> > > exactly who made up this game, and when?
> >
> > Well shucks it started with Aristotle, didn't it?
> > He made a big mistake too ...
> > http://robustai.net/ai/notnota.htm
> >
> > >And why is it so hard to
> > > change the rules?
> >
> > I think it has something to do with the way professors are tenured.
> >
> > Seth
> >
> >
> > ========================================================================
> > To post a message, send mail to cg@cs.uah.edu.
> > To unsubscribe, send mail to majordomo@cs.uah.edu with the command
> > 'unsubscribe cg' in the message body.
> > See http://www.virtual-earth.de/CG/cg-list/ for the mailing list archive
> > See http://www.cs.uah.edu/~delugach/CG for the Conceptual Graph Home Page
> > For help or administrative assistance, mail to owner-cg@cs.uah.edu
>
>
> *** The Era of Total Automation is Now ***
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/