Hypothesis: I am a Transducer (Formerly "Virtual Grounding")
lang...@memstvx1.memst.edu
> COMPUTATIONALISM = HOMUNCULARISM
> (Or, I AM A TRANSDUCER)
>
> Many readers still do not seem to have understood my point about
> transduction, so I will try yet another intuition pump: Although it is
> an over-simplification, consider my hypothesis to be that you ARE a
> transducer. If that hypothesis is correct, then there may be many
> different ways to implement you -- namely, all the different ways of
> implementing a transducer with your capabilities (TTT) -- but among
> those ways is definitely NOT one in which instead of a transducer there
> is a computer simulation of a transducer (a "virtual" transducer). I
See message immediately preceeding this one.
> repeat, my hypthesis is that you ARE a transducer. If that is true,
> anything that is not really a transducer is not really you.
So, conversely, anything that is really a transducer has the possibility
of being me?
Or, on another tack, if something in the environment plays a role in the
system of transduction (e.g., a phone, a CRT, sunglasses), do I suddenly
have an extended 'self' due to the transducting system as a whole?
>
> I might add that there is some real homuncular thinking involved in the
> persistent misunderstanding of my hypothesis. People keep reverting to
> the rival computational hypothesis (which I have tried to show is
> false, Harnad 1989, 1990, 1991, 1992) in which you are a computational
> core, with the transducers simply carrying information TO it ("you"),
> as our senses do to "us."
>
Assuming I am a transducer (which I will not), and the one-to-one and
many-to-one dilemma is ignored for the moment (which it should not be),
What about allowing for a one-to-many mapping? Or is this theory simply
a jazzed up behaviourist approach?
As I interpret it, your "I am transducer" stance incorporates the brain
as part of the transduction system. If this is the case, and assuming a
one-to-one mapping, as you seem to desire, does this imply that each
stimulus received from the sensory system will elicit one response from the
various effectors? Or, must this be broken down, treating each _neuron_ as
a transducer in itself. And, if this _is_ the case, doesn't the number of
transducers in the human system under this hypothesis increase to the point
of being irrelevant, in the sense that you are simply giving a name to each
element in the total information processing system?
How does this shed new light on anything?
Confused,
--
Mark C. Langston "What concerns me is not the
Psychology Department way things are, but rather
Memphis State University the way people think things
LANG...@MEMSTVX1.MEMST.EDU are." -Epictetus
"...a brighter tomorrow?!? How about a better TODAY?" -me
Daryl McCullough
> COMPUTATIONALISM = HOMUNCULARISM
> (Or, I AM A TRANSDUCER)
>
>Many readers still do not seem to have understood my point about
>transduction, so I will try yet another intuition pump: Although it is
>an over-simplification, consider my hypothesis to be that you ARE a
>transducer. If that hypothesis is correct, then there may be many
>different ways to implement you -- namely, all the different ways of
>implementing a transducer with your capabilities (TTT) -- but among
>those ways is definitely NOT one in which instead of a transducer there
>is a computer simulation of a transducer (a "virtual" transducer). I
>repeat, my hypthesis is that you ARE a transducer. If that is true,
>anything that is not really a transducer is not really you.
I don't have any problem with your hypothesis so far; however, your
later statements don't seem consistent with that hypothesis:
>Because we have more than one sense modality (and because most of our
>brain just consists of analog extensions of the sensory and motor
>projections), we know that we are still "us" if we are blind, or deaf
>or paralyzed.
If I am understanding correctly what you mean by "transducer", a
person who becomes blind, deaf or paralyzed person is *not* the same
transducer that he was before. How are we then still "us" after such
an accident? It seems to me that in order to maintain personal
identity, one must posit some kind of "core" which remains constant.
The other problem that I have with your hypothesis is why you think it
is important for the transducers to be analog. That seems to be the
core difference between your position and that of computationalists,
not the issue of "homunculi". I pretty much agree with your hypothesis
that humans are essentially transducers, but I don't see what is
special about analog processing.
Daryl McCullough
ORA Corp.
Ithaca, NY
Harry Erwin
evidence that > 50% of the brain is involved directly in sensory
processing. I have a very preliminary inkling that although the data do
not seem to support Pribram's holographic memory model, the related
concept of "adaptive beam forming" may be productive. The difference
between the two concepts has to do with the holographic memory model
actually generating an internal scene from the sensory data array, which
the frontal and motor areas observe passively, while the "adaptive beam
forming" model instead allows the frontal and motor cortexes to actively
interact with the sensory data array to specify the point of view and
sensory modalities involved.
--
Harry Erwin
Internet: er...@trwacs.fp.trw.com
Thomas Clarke
> COMPUTATIONALISM = HOMUNCULARISM
> (Or, I AM A TRANSDUCER)
>
> an over-simplification, consider my hypothesis to be that you ARE a
> transducer. If that hypothesis is correct, then there may be many
> different ways to implement you -- namely, all the different ways of
> implementing a transducer with your capabilities (TTT) -- but among
> those ways is definitely NOT one in which instead of a transducer there
> is a computer simulation of a transducer (a "virtual" transducer).
[Warning: I am about to bring up quantum mechanics and chaos again
for those who find these topics uninteresting.]
I like to look at the problem of AI/consciousness from a physical point
of view. I lean toward agreement with SH's conclusion that bare
computation cannot implement consciousness. Unfortunately, this
raises the ugly question of where the consciousness resides.
From the standpoint of classical phsyics, one can take the viewpoint of
Laplace that given initial conditions and perfect calculation, the
physical future can be perfectly predicted or simulated. We have
learned that perfect calculation is captured by Turing machine
computation (Church-Turing thesis), so a modern day Laplacian would
ascribe perfect simulative capabilities to a suitably programmed
Turing machine.
Now if mind is physical, then from this viewpoint it can be
Laplace-Turing simulated. Since the L-T simlation is completely
accurate, then everything about the mind, including conciousness,
will be automatically captured by the L-T simulation. The L-T simulation
of the mind may have to expand to include physical transduction,
perhaps even the entire universe, but given Laplace's classical view
of physics, the Church-Turing thesis, and the physicality of
mind, the conclusion that consciousness can result from a computation
seems inevitable.
SH (and others) argue plausibly that computation alone cannot give
consiousness. If physicality is not abandoned, then either the C-T
thesis is false or Laplace's view is wrong. C-T is only a hypothesis,
but nothing constructive is available that can exceed Turing machine
computation in power. On the other hand, we know Laplace was wrong.
Laplace was wrong in two ways: one the robustness of physical
calculation has been overthrown by chaos theory, and determinism
has been overthrown by quantum mechanics.
Chaos is attractive in that it a purely "classical" phenomena
resulting only from the difficulties of non-linear calculation.
The problem is that Laplace is still right, it is only that the
initial conditions must be know to accuracy beyond that achievable
by any conceivable physical measurement. Any error in the initial
conditions results ultimately in enormous error in the simulation.
But this is a problem only for mere mortals, not for Philosophical
Principals.
Quantum Mechanics is a horse of a different color. It sets
fundamental limits on what is simulable. It is not possible
in principle to predict certain things. Attempts to make quantum
mechanics classical and simulable by postulating an underlying reality
lead to predictions that contradict experiment (e.g. Bell's inequality).
Worse (for physics) the standard attempts to interpret the math of
QM lead to says things like "the way function collapses upon
observation." Observation? by who? Other interpretations lead
to equally mysterious things like:
many universes existing in parallel, each containing a different
version of every possible observation, or
correlations between observations extending through all of space
acting at infinite speed, but having no locally observable
consequences.
Big difficulties for the Laplace-Turing simulation to capture all
quantum effects. Some interpretations of quantum pheonomea require
a conscious observer (?), others require simlation of every possible
world history, and yet others (equivalently) require taking into account
all possilbe events in the entire universe.
Big opportunities for non-simulable, but physical phenomena to
bring consciousness into the world in a way that cannot be simulated.
For example, the unobservable correlations traveling at infinite
speed sound a lot like a consciousness residing in the brain but
having no observable behavioral effects.
At any rate, to end this already too long post, there seems to be
ample room in the strange discoveries of 20th century physics for
conscious phenomena that Steven Harnad postulates cannot be simulated
via a Turing machine.
--
Thomas Clarke
Institute for Simulation and Training, University of Central FL
12424 Research Parkway, Suite 300, Orlando, FL 32826
(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
Neil Rickert
>SH (and others) argue plausibly that computation alone cannot give
>consiousness. If physicality is not abandoned, then either the C-T
>thesis is false or Laplace's view is wrong. C-T is only a hypothesis,
What has C-T to do with anything. C-T is neither true nor false. It
is not the type of thing that can ever be true or false. It is merely
a cultural belief withi mathematics. People can stop believing it
if they like. But you can't talk about it being false.
--
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
Neil W. Rickert, Computer Science <ric...@cs.niu.edu>
Northern Illinois Univ.
DeKalb, IL 60115 +1-815-753-6940
Thomas Clarke
> In article <1992Jun1.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
Clarke) writes:
>
> >SH (and others) argue plausibly that computation alone cannot give
> >consiousness. If physicality is not abandoned, then either the C-T
> >thesis is false or Laplace's view is wrong. C-T is only a hypothesis,
>
> What has C-T to do with anything. C-T is neither true nor false. It
> is not the type of thing that can ever be true or false. It is merely
> a cultural belief withi mathematics. People can stop believing it
> if they like. But you can't talk about it being false.
>
way to compute. If the formalisms of the science of physics can compute
anything physical and C-T is true, then a Turing machine properly programmed
could predict (simulate) anything, including the behavior of
intelligent critters.
C-T could be falsified by the discovery of another means for effective,
that is precisely specifiable, scientifically reproducible calculation.
Something like Douglas Adam's computational bistro would qualify :-)
More seriously David Deutsch has speculated that there may exist quantum
mechanical computers that transcend Turing computation.
A comment on your other posting:
>nonsense, that there is nothing special in the analog representation of
>information, and that digital representation is often preferable due to
This is precisely the point I am trying to get at: there may well
be something special about raw, analog, undigitized information.
To make my case more succinctly, according to QM, one can predict the
probability of given experimental outcomes, but not the experimental
outcome itself. Thus a Turing machine can only calculate the probabilities
of quantum events, not the events themselves. The only way to determine
the outcome of events is to look at the physical (e.g. analog) outcomes
themselves. Nor can heuristics such as randomly throwing non-quantum
dice be used to determine outcomes since nastinesses such as
Bell's inequalities will cause the simulation to depart from reality.
The only way to simulate quantum events is to carry out the full
complex-valued, superimposed wave state calcualtion. Even then
the only result is probabilities of events, not predictions of
the actual events.
QM is not relevant to your digitally interfaced room, but may well
play a role in the observer within the room in a way consistent with
Harnad's transducer idea. Were the digital data fed directly into a
Turing machine-equivalent, the result might not be conscious perception
in much the same way as a Turing machine cannot predict quantum events.
Stephen P Spackman
observable. The fact that a simulation of system X may not yield the
same results as system X itself is wholely uninteresting, since if
there IS randomness out there, then "two runs" of the "real" system X
(whatever tht means) would not be obliged to return the same result
as EACH OTHER.
Your simulation is successful whenever you predict ANY of the possible
outcomes of a physical experiment - since that's the best that reality
itself can offer.
(Another way of phrasing this is that if you randomised algorithm
fails when a non-random source is substituted for its RNG, then you
never had an algorithm in the first place.)
(And ANOTHER way of phrasing it is that (given totality) a
strengthened postcondition is interconvertible with a weakened
precondition, and vice-versa. Nondeterminism and (enough) hidden state
amount to the same thing. Given how LITTLE of the state of the "real
world" will ever be communicated to you, equivalence up to
(personal!) observability is an incredibly weak criterion - but is
also the strongest you can have).
----------------------------------------------------------------------
stephen p spackman Center for Information and Language Studies
ste...@estragon.uchicago.edu University of Chicago
----------------------------------------------------------------------
Believe in Strong AI? I don't even believe in Strong I!
Nick Cassimatis
>Big difficulties for the Laplace-Turing simulation to capture all
>quantum effects. Some interpretations of quantum pheonomea require
>a conscious observer (?), others require simlation of every possible
>world history, and yet others (equivalently) require taking into account
>all possilbe events in the entire universe.
I've always been bothered by arguments using quantum mechanics. One
reason is all of those books claiming to explain Experience,
Consciousness, Art and Everything else using Quantum Mechanics. But a
alot fruit brains don't falsify a theory. The more principled reason:
a scientific theory is not set in stone. There is every reason to
believe that it will be revised, extended or even thrown out. So
using such theories (and dubious interpretations of the math therin)
puts those arguments on very shaky ground. Science is certainly a
marvel, but not final.
In particular, there are a few facets of arguments using QM that have
always puzzled me. As I have yet to take a course in QM (though I may
in the Fall) perhaps someone can enlighten me: it seems that peaple
like to revel in Heisenberg's Uncertainty Principle and see it as the
downfall of Laplace. But I see it this way. That Principle is only a
statement of our experimental impotence. It suggests that it is
impossible to know the state of the universe at a particular moment.
Even if this is true, so what? Laplace's statement is based on a
counterfactual: "If I knew the state of the universe...."
Heisenberg's principle only suggests that we can't know the state of
the universe, it says nothing about what could be done *if* we did
know the state of the universe.
The other thing that bothers me is the probablistic character of QM.
That our theory of certain phenomena is only probablistic is no reason
to elevate our ignormance to a property of the universe. Even if
Heisenberg's principle is true and thus prevents us from getting a
deterministic theory, this would seem not imply that the universe is
random, but only that our understanding of it is constrained.
Though I emphatically do not mean to suggest that this is the
motivation of Thomas' sober comments, Chaos and QM seem to be all to
ready excuses for people to wax mystical (Dawkings' phrase.)
Randall Holmes
A QM machine can't predict quantum events, either!
>
>--
>Thomas Clarke
>Institute for Simulation and Training, University of Central FL
>12424 Research Parkway, Suite 300, Orlando, FL 32826
>(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
I'm quite certain that the numerical probabilities in the state
calculation are Turing computable to any desired degree of accuracy
(when our theory is adequate to set up the calculations at all), and
beyond this we cannot go; no dice-rolling, even with the benefit of
contact with real QM phenomena, will give any more information than
the bare probabilities, if QM is valid. Nowhere do you exhibit any
evidence for a super-Turing capability coming out of QM; it is
certainly unreasonable to fault any machine for being unable to
predict what cannot be predicted, and it is not evidence for the claim
that there is a better kind of machine: _no_ "machine" can predict QM
effects.
--
The opinions expressed | --Sincerely,
above are not the "official" | M. Randall Holmes
opinions of any person | Math. Dept., Boise State Univ.
or institution. | hol...@opal.idbsu.edu
Mark Rosenfelder
>I've always been bothered by arguments using quantum mechanics. One
>reason is all of those books claiming to explain Experience,
>Consciousness, Art and Everything else using Quantum Mechanics. But a
>alot fruit brains don't falsify a theory. The more principled reason:
>a scientific theory is not set in stone. There is every reason to
>believe that it will be revised, extended or even thrown out.
Don't get your hopes up. After 70 years there are no known holes in QM--
it's always agreed with experiment. Nor is there any guarantee
that if it *is* replaced, you'll like the replacement any better. :)
>In particular, there are a few facets of arguments using QM that have
>always puzzled me. As I have yet to take a course in QM (though I may
>in the Fall) perhaps someone can enlighten me: it seems that peaple
>like to revel in Heisenberg's Uncertainty Principle and see it as the
>downfall of Laplace. But I see it this way. That Principle is only a
>statement of our experimental impotence. It suggests that it is
>impossible to know the state of the universe at a particular moment.
>Even if this is true, so what? Laplace's statement is based on a
>counterfactual: "If I knew the state of the universe...."
>Heisenberg's principle only suggests that we can't know the state of
>the universe, it says nothing about what could be done *if* we did
>know the state of the universe.
It sounds like you're thinking there is a "real" state of the universe--
or, to put it into more concrete terms, an electron does have a definite
position and momentum-- only we can't find out what they are. Heisenberg
actually goes further than this: there is no "real" state of the universe;
you can't find out the electron's definite position and momentum because
it doesn't have one.
Common sense rebels at this; but common sense needs to spend more time
thinking about waves. Any wave can be built up out of superimposed simple
waveforms-- e.g. sine waves or impulse waves. A particular wave might
be built up out of just 3 sine waves, in which case it will take a whole
bunch of impulse waves to synthesize it. Or if it only takes a few impulse
waves, the analysis into sine waves will be very complex. You can't have
a wave that's extremely simple in both sine-wave terms and impulse-wave
terms, because these are opposite types of waveforms.
This isn't just an analogy-- it's the reason behind the "uncertainty
principle." Position and momentum are different aspects of the quantum
wavefunction of a particle; you can't get both at the same time for the
same reason you can't analyze a wave into simple superimpositions of
opposite waveforms.
Now, your idea that the universe does have "real states" even if we can't
(yet?) observe them is not new; such "hidden variable" theories can be
built which are compatible with quantum facts. However, it's been
pretty well established that such theories can't explain the facts without
recourse to faster-than-light signalling.
For further reading, I'd recommend Richard Feynman's _QED_, Nick Herbert's
_Quantum Reality_, and J.C. Polkinghorne's _The Quantum World_.
Nick Cassimatis
>Don't get your hopes up. After 70 years there are no known holes in QM--
>it's always agreed with experiment. Nor is there any guarantee
>that if it *is* replaced, you'll like the replacement any better. :)
70 years is not that long of a time. And to the best of my knowledge,
QM has evolved since it's invention. That it and the rest of science
keeps on changing seems to be good enough reason to not let a priori
arguments based on *any* scientific theory move us too far on issues
of such importance. Though you're right that replacements to QM might
be even more flaky. My point was only that we can never be too
certain about this flakiness.
>This isn't just an analogy-- it's the reason behind the "uncertainty
>principle." Position and momentum are different aspects of the quantum
>wavefunction of a particle; you can't get both at the same time for the
>same reason you can't analyze a wave into simple superimpositions of
>opposite waveforms.
Again, this seems to use language internal to the theory of QM.
Hence, it is a valid argument only insofar as QM is a valid theory.
This isn't enough for an a priori argument.
Thanks for the comments and references,
-Nick
PS Of course, I don't see any of the possible outcomes such
discussions would changing the way I think about AI or MOI (my own
intelligence).
Patrick L Faith
>>Even if this is true, so what? Laplace's statement is based on a
>>counterfactual: "If I knew the state of the universe...."
>>Heisenberg's principle only suggests that we can't know the state of
>>the universe, it says nothing about what could be done *if* we did
>>know the state of the universe.
Good insight !!! Took me awhile to completely understand the ramifications
of assuming an initial condition, a extremely non-trivial problem.
>Now, your idea that the universe does have "real states" even if we can't
>(yet?) observe them is not new; such "hidden variable" theories can be
>built which are compatible with quantum facts. However, it's been
>pretty well established that such theories can't explain the facts without
>recourse to faster-than-light signalling.
Agreed, and the fact that it always points to finer granularity of space/time
should tell us something.
>For further reading, I'd recommend Richard Feynman's _QED_, Nick Herbert's
>_Quantum Reality_, and J.C. Polkinghorne's _The Quantum World_.
Not going to find what you're looking for in these books, try looking
for Einsteins later notes (especially discrete space/time), or general
relativistic pregeometry ( is a graduate level concept but has a lot
of good computer science).
I was suprised to find this thread in the ai loop, but pleasently because
it belongs in here. There is a lot of work going on in connectionist
oriented geometry (i.e. n-dimensional geometries) and other so such
weird things that require ai techniques to fully pull them off - for example
if every event had a connection (dimension) with every other event, then
computationally you would have to throw out connections and make new
connections to model (with error) any complex event with a computer, the
ability to chunk and generalize being the ai techniques I think that are
most useful working on such problems.
PLai
Stephen P Spackman
hidden-variable model of reality WITH ftl signalling. After all,
claiming that the topology/locality properties of space are other than
approximately euclidean is no weirder than saying that the kinetic
properties of particles are other than approximately newtonian.
In fact, htis is a VERY familiar notion. As a computer scientist, I am
entirely familiar with data structures that are superficially a lot
like strings, exhibiting what are for the most part regular grammars
and so forth - but which have POINTERS in them. Pointers are not
locally analysable, and are "faster than light" - they refer to
locations that are _exponentially_ distant from themselves in a
universe where scanning operations are "basically" O(n) or slower.
So: there's a lot of hidden state that is indetectable and thus not a
fit topic for contemplation; and space isn't as flat as you think
(having a topology, in fact, that is intimately connected with its
history). Sounds reasonable enough to me.
(Of course, this isn't the only model available, but it is adequate
and meshes well with the fact that there is no good mathematical
definition of randomness formulated as other than a guarantee of
relative ignorance!).
Mark Rosenfelder
>Waves aren't the whole picture either. Life would be simpler if it
>were!
Yes, I know, but *you* try summarizing QM in three paragraphs. :)
What I was trying to point out was that the linked indeterminacy of quantum
mechanics is not an arbitrary restriction (as suggested by the term
"uncertainty _principle_") but falls naturally out of the wave nature
of quantum particles. Describing the U.P. as arbitrary makes QM sound
much stranger than it really is.
Mark Rosenfelder
>70 years is not that long of a time. And to the best of my knowledge,
>QM has evolved since it's invention.
I don't know what you mean by "evolved" here, but I don't think the basic
nature of the theory has changed much. I don't think anybody's corrected
the Dirac equation, for instance. The changes have mainly been applying QM
to new things and finding better ways to calculate with it.
>That it and the rest of science
>keeps on changing seems to be good enough reason to not let a priori
>arguments based on *any* scientific theory move us too far on issues
>of such importance. Though you're right that replacements to QM might
>be even more flaky. My point was only that we can never be too
>certain about this flakiness.
Certain, no. Philosophical conclusions based on QM could be wrong.
But I like my philosophy to be consistent with the known facts,
and for now that means facing up to QM. (I don't like it when
sf stories posit faster-than-light travel, either...)
>PS Of course, I don't see any of the possible outcomes such
>discussions would changing the way I think about AI or MOI (my own
>intelligence).
I'm with you here. Attempts to link QM with intelligence have so far been,
to my mind, pretty wispy.
Thomas Clarke
> There is, of course, nothing even remotely disagreeable about using a
> hidden-variable model of reality WITH ftl signalling.
A very disconcerting idea if you've had a lot of physics courses, though.
... Analogy of quantum mechanical hidden variable theories with
pointers in computer science ....
> So: there's a lot of hidden state that is indetectable and thus not a
> fit topic for contemplation; and space isn't as flat as you think
> (having a topology, in fact, that is intimately connected with its
> history). Sounds reasonable enough to me.
Something like this could be true, I suppose. But the computational
nature of your physical theory points out the danger of saying that
physics (e.g. analog transduction) has nothing to do with intelligence.
If physics is computational and our Turing computations are a poor
approximation thereof, it is not unreasonable the physical computations
implemented in brains can exceed Turing computations by, for example,
exhibiting consciousness.
Thomas Clarke
Waves aren't the whole picture either. Life would be simpler if it
were! Energy insists on being transferred only in discrete quanta of
size Planck's constant times frequency. I once came across an analogy:
It's as if when ocean waves hit the beach, instead of washing up onto the
sand, all of sudden frogs were to leap out, the number of frogs being
determined by the size of the wave. Catch a bunch of frogs and they turn
back into waves in your bait bucket.
Thomas Clarke
> That our theory of certain phenomena is only probablistic is no reason
> to elevate our ignormance to a property of the universe. Even if
> Heisenberg's principle is true and thus prevents us from getting a
> deterministic theory, this would seem not imply that the universe is
> random, but only that our understanding of it is constrained.
The QM literature shows that people have been trying to squirm out
of the conclusion that the ignorance is a property of the universe
since QM's inception. The jury is still out, but the simplest, if most
puzzling interpretations of QM have an ignorant universe.
> Though I emphatically do not mean to suggest that this is the
> motivation of Thomas' sober comments, Chaos and QM seem to be all to
> ready excuses for people to wax mystical (Dawkings' phrase.)
I thought discussions of mind were inherently mystical :-)
Thomas Clarke
> I'm quite certain that the numerical probabilities in the state
> calculation are Turing computable to any desired degree of accuracy
> (when our theory is adequate to set up the calculations at all), and
> beyond this we cannot go; no dice-rolling, even with the benefit of
> contact with real QM phenomena, will give any more information than
> the bare probabilities, if QM is valid. Nowhere do you exhibit any
> evidence for a super-Turing capability coming out of QM; it is
> certainly unreasonable to fault any machine for being unable to
> predict what cannot be predicted, and it is not evidence for the claim
> that there is a better kind of machine: _no_ "machine" can predict QM
> effects.
While not (Turing) machine can predict QM effects, QM effects happen so
the universe "predicts" them. Perhaps a subset of the universe can
be boxed into a convenient black box machine that can "predict" these
effects in a computationaly useful way. (Is the brain such a black box?)
For details I can only refer you to
Deutsch, David (1985). Quantum thoery, the Curch-Turing principle and
the universal quantum computer. _Proc Royal Soc of London, A400,_
97-117.
I could not begin to do justice to Deutsch's detailed discussion.
By invoking quantum mechanics I was trying to show the fallacy I see
in the argument that everyone is seeming to make implicitly:
1. Mind is physical.
2. Computers can simulate all relevant physics relevant to mind.
Therefore computers can simulate mind.
Quantum mechanics is an example of physics that cannot be fully
simulated.
To me the question of whether QM is relevent is open.
Wayne Throop
>Don't get your hopes up. After scores of years there are no known holes
>in Kpelerian dynamics... it's always agreed with planetary observations.
>Nor is there any guarantee that if it *is* replaced, you'll like the
>replacement any better. :)
70 years is not that long of a time. And to the best of my knowledge,
KD has evolved since it's invention. That it and the rest of science
keeps on changing seems to be good enough reason to not let a priori
arguments based on *any* scientific theory move us too far on issues
of such importance. Though you're right that replacements to KD might
be even more flaky. My point was only that we can never be too
certain about this flakiness.
>This isn't just an analogy-- it's the reason behind the deviations from
>circular orbits. The eliptical shape of the orbit is why it tends to
>look so much like a circle with some corrections, but there isn't really
>any circle involved at all.
Again, this seems to use language internal to the theory of KD.
Hence, it is a valid argument only insofar as KD is a valid theory.
This isn't enough for an a priori argument.
And on another parchment:
>Now, your idea that the orbits do have "real circles" even if we can't
>(yet?) observe them is not new; such "epicycle" theories can be
>built which are compatible with observed orbits. However, it's been
>pretty well established that such theories can't explain the facts without
>recourse to potentially infinite regress.
There is, of course, nothing even remotely disagreeable about using an
infinite epicycle model of orbits. After all, claiming that a finite
formulation is more "real" is nothing more than a computational
prejudice.
Wayne Throop ...!mcnc!aurgate!throop
Randall Holmes
>(Randall Holmes) writes:
>> I'm quite certain that the numerical probabilities in the state
>> calculation are Turing computable to any desired degree of accuracy
>> (when our theory is adequate to set up the calculations at all), and
>> beyond this we cannot go; no dice-rolling, even with the benefit of
>> contact with real QM phenomena, will give any more information than
>> the bare probabilities, if QM is valid. Nowhere do you exhibit any
>> evidence for a super-Turing capability coming out of QM; it is
>> certainly unreasonable to fault any machine for being unable to
>> predict what cannot be predicted, and it is not evidence for the claim
>> that there is a better kind of machine: _no_ "machine" can predict QM
>> effects.
>
>While not (Turing) machine can predict QM effects, QM effects happen so
>the universe "predicts" them. Perhaps a subset of the universe can
>be boxed into a convenient black box machine that can "predict" these
>effects in a computationaly useful way. (Is the brain such a black
box?)
The universe does not predict the QM effects; it simply exhibits them.
It would only predict them if it were actually deterministic on a
deeper level; i.e., if QM were only an approximation to the real
situation!
>
>For details I can only refer you to
>Deutsch, David (1985). Quantum thoery, the Curch-Turing principle and
>the universal quantum computer. _Proc Royal Soc of London, A400,_
>97-117.
>
>I could not begin to do justice to Deutsch's detailed discussion.
>
>By invoking quantum mechanics I was trying to show the fallacy I see
>in the argument that everyone is seeming to make implicitly:
>
>1. Mind is physical.
>2. Computers can simulate all relevant physics relevant to mind.
>Therefore computers can simulate mind.
If premises 1 and 2 are valid, we can indeed simulate mind. The
simulated mind will behave unpredictably at certain points, just like
real minds. (As I have indicated elsewhere, I think that "mind" is
one of the fixed points of the function (X |-> simulated X)).
>
>Quantum mechanics is an example of physics that cannot be fully
>simulated.
It certainly can be simulated. You simply have to roll dice at the
right point (and not stupidly -- i.e., not locally). Also, you have
to do really tough mathematics to set up your simulation correctly. I
believe that this is actually done -- for very small systems (atoms!).
>To me the question of whether QM is relevent is open.
I suspect that QM effects may be relevant, but mostly in the form of
providing random numbers when these are useful; sometimes an
intelligence needs to make a decision when there are no grounds for
the decision (nothing to do with "free will"; our behaviour in
situations where "free will" is supposed to be involved had better be
determined!)
>
>--
>Thomas Clarke
>Institute for Simulation and Training, University of Central FL
>12424 Research Parkway, Suite 300, Orlando, FL 32826
>(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
Randall Holmes
>(yet?) observe them is not new; such "hidden variable" theories can be
>built which are compatible with quantum facts. However, it's been
>pretty well established that such theories can't explain the facts without
>recourse to faster-than-light signalling.
If you think that the waves are real physical phenomena, then FTL
signalling is involved. I think it is quite reasonable to suppose
that the universe has "real states"; the particle side of the
particle-wave duality is real; the wave side encodes restrictions on
what information we can have about the real situation. There is no
reason why we cannot know that there are correlations between the
facts about two events which have a space-like separation, while we do
not know the actual facts; once we acquire further information about
one of these events, we immediately gain information about the other,
but there is no superluminal communication involved (the correlations
between the two facts derive from their common relationship with a
third event which has time-like separation from each of them!). The
illusion of superluminal communication comes in when you suppose (and
I believe this is built into the axioms of the arguments against
hidden variables) that the waves are physical phenomena (and so behave
"locally"). Heisenberg's step from "we cannot know the position and
momentum at the same time" to "the particle does not have a position
and a momentum at the same time" was motivated by philosophical
considerations (and mathematical convenience); the physics does not
dictate this. Also, I believe that it has been argued that we _can_
determine position and momentum of a particle simultaneously in
retrospect (we can't know both when they would be any use to us...).
>
>For further reading, I'd recommend Richard Feynman's _QED_, Nick Herbert's
>_Quantum Reality_, and J.C. Polkinghorne's _The Quantum World_.
Stephen P Spackman
with intelligence - in fact, physics has a LOT to do with
intelligence. My contention is, rather, that computation can, and
should, be presumed adequate to intelligence until proven otherwise,
because it seems increasingly clear that intelligence is a very
mundane phenomenon, determined by boring macrophysics (in the form of
social and linguistic interactions), most of whose "interesting"
properties are subjective anyway.
The clever thing about people is that they get away with being so
stupid, but that's to do with evolution, not quantum mechanics or
divine intervention.
The whole discussion seems to me like _Chariots of the Gods_ - why go
to all thi trouble to construct an exotic solution to a problem that
can clearly be solved in fifty conventional ways, the hard part being
just pinning down which one obtains in practise (and which other one
gives the most efficient engineered implementation, if that's even
useful at all).
Thomas Clarke
> In article <1992Jun2.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
Clarke) writes:
> >While not (Turing) machine can predict QM effects, QM effects happen so
> >the universe "predicts" them. Perhaps a subset of the universe can
> >be boxed into a convenient black box machine that can "predict" these
> >effects in a computationaly useful way. (Is the brain such a black
> box?)
>
> The universe does not predict the QM effects; it simply exhibits them.
> It would only predict them if it were actually deterministic on a
> deeper level; i.e., if QM were only an approximation to the real
> situation!
Hence the quotation marks around predicts! Neither may my brain
calculate (or simulate) intelligence, it may just exhibit it.
> >Quantum mechanics is an example of physics that cannot be fully
> >simulated.
>
> It certainly can be simulated. You simply have to roll dice at the
> right point (and not stupidly -- i.e., not locally). Also, you have
> to do really tough mathematics to set up your simulation correctly. I
> believe that this is actually done -- for very small systems (atoms!).
I suspect our respective undestandings of QM differ too much to
really reach an agreement now. My intutition is that a
correct QM simulation would have to simulate every possiblity in an
Everett style many world's simulation. Shades of the humongous LUT.
With regard to atoms - I refer you to
T. Erber and S. Putterman, 1985, "Randomness in quantum mechanics - nature's
ultimate cryptogram?", Nature, 318, 41-43.
The time series of radiative transititions of a single atom is shown to
be ultimately random in the sense that no algorithm can generate such
a time series, much as in Chaitin's algoritmic information theory. A single
atom is thus not simulatable. You can add "dice" I suppose, but they
must be dice correlated with dice everywhere else so that the simulation
becomes less of a model and more just a restructuring of the actual
physical process.
> >To me the question of whether QM is relevent is open.
>
> I suspect that QM effects may be relevant, but mostly in the form of
> providing random numbers when these are useful; sometimes an
> intelligence needs to make a decision when there are no grounds for
> the decision (nothing to do with "free will"; our behaviour in
> situations where "free will" is supposed to be involved had better be
> determined!)
But consider the following correspondences:
QM correlations <--?--> qualia
In QM correlations have no effects, but are necessary to understand the
theory. In the same way qualia are necessary to understand the mind.
hidden variables <--?--> homomculus
Both hidden variables and homonculi are attractive explanations that
must be avoided, but often crop up in unexpected ways.
These parallels between AI and QM are interesting on a metascientific level;
both disciplines have interpretation problems that most practitioners
ignore. But to me the analogies are so close that one is tempted to
apply Occam's razor and argue for a deeper connection.
Philip Santas
>reason is all of those books claiming to explain Experience,
>Consciousness, Art and Everything else using Quantum Mechanics. But a
>alot fruit brains don't falsify a theory. The more principled reason:
>a scientific theory is not set in stone. There is every reason to
>believe that it will be revised, extended or even thrown out. So
>using such theories (and dubious interpretations of the math therin)
>puts those arguments on very shaky ground. Science is certainly a
>marvel, but not final.
Every successfull model is consistent with the facts observed till the point
of its introduction and must be able to predict more. If there are more
facts to be included, then a more general model has to be introduced
(relativity vs newtonian mechanics).
Notice that religions have already introduced the most abstract models :-)
>like to revel in Heisenberg's Uncertainty Principle and see it as the
>downfall of Laplace. But I see it this way. That Principle is only a
>statement of our experimental impotence. It suggests that it is
>impossible to know the state of the universe at a particular moment.
>Even if this is true, so what? Laplace's statement is based on a
>counterfactual: "If I knew the state of the universe...."
then we could do things which cannot be verified because,
>Heisenberg's principle only suggests that we can't know the state of
>the universe, it says nothing about what could be done *if* we did
>know the state of the universe.
if we cannot know the state of the universe, then we cannot apply
Laplace's statement. Why is this not a counterfactual?
>The other thing that bothers me is the probablistic character of QM.
But most of the models of the macroworld are probabilistic.
Gas laws are based on statistical assumtions. But this does not seem
to prevent pilots from flying airplanes :-)
>That our theory of certain phenomena is only probablistic is no reason
>to elevate our ignormance to a property of the universe. Even if
>Heisenberg's principle is true and thus prevents us from getting a
>deterministic theory, this would seem not imply that the universe is
>random, but only that our understanding of it is constrained.
Heisenberg's principle does NOT imply randomness. There can be a reason,
which WE cannot observe with physical means. Where do you see the randomness?
Philip Santas
--------------------------------------------------------------------------------
email: san...@inf.ethz.ch Philip Santas
Mail: Dept. Informatik Department of Computer Science
ETH-Zentrum Swiss Federal Institute of Technology
CH-8092 Zurich Zurich, Switzerland
Switzerland
Phone: +41-1-2547391
Philip Santas
In article <1992Jun2.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas Clarke) writes:
>ste...@estragon.uchicago.edu (Stephen P Spackman) writes:
>> There is, of course, nothing even remotely disagreeable about using a
>> hidden-variable model of reality WITH ftl signalling.
>
>A very disconcerting idea if you've had a lot of physics courses, though.
>
>... Analogy of quantum mechanical hidden variable theories with
>pointers in computer science ....
Nothing strange if you consider the analogies drawn in this group about
psychology and quantum mechanics :-)
>> So: there's a lot of hidden state that is indetectable and thus not a
>> fit topic for contemplation; and space isn't as flat as you think
>> (having a topology, in fact, that is intimately connected with its
>> history). Sounds reasonable enough to me.
>
>Something like this could be true, I suppose. But the computational
>nature of your physical theory points out the danger of saying that
>physics (e.g. analog transduction) has nothing to do with intelligence.
>If physics is computational and our Turing computations are a poor
>approximation thereof, it is not unreasonable the physical computations
>implemented in brains can exceed Turing computations by, for example,
>exhibiting consciousness.
which is computable by itself?
Then where is the danger?
Nick Cassimatis
>
>In article <1992Jun1.2...@news.media.mit.edu> n...@media.mit.edu (Nick Cassimatis) writes:
>>
>>The other thing that bothers me is the probablistic character of QM.
>
>But most of the models of the macroworld are probabilistic.
>Gas laws are based on statistical assumtions. But this does not seem
>to prevent pilots from flying airplanes :-)
While I haven't seen the gas laws formally worked out, from what I've
heard, I get the impression that the "probablilistic assumption" is
not about the laws that govern those particles. But the point is
moot. I never meant to imply that probabalism bothers me, only that
bad arguments for probabalism do. As interesting they may be, the
results of physics never really moved me either way. Whatever they
turn out to be, it wouldn't change the way I hear a symphony or the
way I think about psychology and a whole bunch of more trivial things.
>Heisenberg's principle does NOT imply randomness. There can be a reason,
>which WE cannot observe with physical means. Where do you see the randomness?
Again, since I haven't read a *real* quantum mechanics bood, I
wouldn't know. The passage you quote was my reaction to someone who
said the H's principle *is* randomness.
On the relevence of QM to AI: there are many arguments that maintain
that we cannot achieve AI on a computer because a computer can't model
quantum effects. Even if it can't so what? There are quantum effects
in water molecules, but we have had some success in modelling fluids.
The neuron is such a big thing that I don's see why quantum effects
would be relevant. Even if it were relevent on the neuronal level,
who says we have to go all the way down to that level to achieve AI?
I may be wrong, but I'm pretty sure we got to the point of building
transistors smaller than a neuron without taking quantum effects into
account.
-Nick
Gordon Joly
> From: n...@media.mit.edu (Nick Cassimatis)
> [...]
>
> Again, since I haven't read a *real* quantum mechanics bood, I
> wouldn't know. The passage you quote was my reaction to someone who
> said the H's principle *is* randomness.
No. It is very precise.
And choatic systems are completely deterministic: it is the initial
conditions that are the "problem". Heisenberg affects measurements as
well.
> On the relevence of QM to AI: there are many arguments that maintain
> that we cannot achieve AI on a computer because a computer can't model
> quantum effects. Even if it can't so what? There are quantum effects
> in water molecules, but we have had some success in modelling fluids.
> The neuron is such a big thing that I don's see why quantum effects
> would be relevant.
The golf ball travelling at 30 mph has a wavelength, via
E= h v (de Broglie's equation)
then why ingore (possible) quantum effects at the neuronal level?
> Even if it were relevent on the neuronal level,
> who says we have to go all the way down to that level to achieve AI?
Roger Penrose.
> I may be wrong, but I'm pretty sure we got to the point of building
> transistors smaller than a neuron without taking quantum effects into
> account.
>
> -Nick
Ouch. The Zenner diode, which use quantum tunnellling, is just one
example of QM. And CCD cameras(?)
Gordon.
Mark William Hopkins
>it's always agreed with experiment. Nor is there any guarantee
>that if it *is* replaced, you'll like the replacement any better. :)
There's big, obvious, gaping holes in QM!
The biggest of them all is that it can't explain gravity.
The second is that it can't explain the dimensional or topological
structure of spacetime.
The third is that it can't explain why this dimensionality is 3+1.
The fourth is that it can't explain the origin or significance of the Planck
units.
The fifth is that it can't explain why or how the continuum breaks down at
Planck scales.
The sixth is that it can't explain the origin of the Universe or its fate!
Need I go on? The list is endless.
Mark Rosenfelder
>There's big, obvious, gaping holes in QM!
>
>The biggest of them all is that it can't explain gravity.
>The second is that it can't explain the dimensional or topological
>structure of spacetime.
>The third is that it can't explain why this dimensionality is 3+1.
>The fourth is that it can't explain the origin or significance of the Planck
>units.
>The fifth is that it can't explain why or how the continuum breaks down at
>Planck scales.
>The sixth is that it can't explain the origin of the Universe or its fate!
>
>Need I go on? The list is endless.
I'll add some to your list: It can't explain why electrons have the mass
they do; it can't explain the three-body problem; it can't explain
population genetics; and it can't explain Cubism.
What I meant by a hole was a case where QM theory predicts one thing and
experimental evidence contradicts it; as for instance the Michelson-Morley
experiment, the photoelectric effect, the black-body radiation curve,
and the problem of the electron's orbit contradicted the predictions
of classical physics.
Mark Rosenfelder
>signalling is involved. I think it is quite reasonable to suppose
>that the universe has "real states"; the particle side of the
>particle-wave duality is real; the wave side encodes restrictions on
>what information we can have about the real situation. There is no
>reason why we cannot know that there are correlations between the
>facts about two events which have a space-like separation, while we do
>not know the actual facts; once we acquire further information about
>one of these events, we immediately gain information about the other,
>but there is no superluminal communication involved (the correlations
>between the two facts derive from their common relationship with a
>third event which has time-like separation from each of them!). The
>illusion of superluminal communication comes in when you suppose (and
>I believe this is built into the axioms of the arguments against
>hidden variables) that the waves are physical phenomena (and so behave
>"locally").
I'm not sure I understand you, but I suspect you're wrong. It's not just
a matter of observing a correlation across a huge distance. It's that
you and another observer do something (twist a calcite crystal) and
immediately, before any signal could move from one to the other, the
correlation rate changes, in a way that can't be explained by local
properties of the particles. There's something non-local going on here.
I don't see what difference it makes if you call it "phyical" or not.
Philip Santas
> san...@inf.ethz.ch (Philip Santas) writes:
>>
>>But most of the models of the macroworld are probabilistic.
>>Gas laws are based on statistical assumtions. But this does not seem
>>to prevent pilots from flying airplanes :-)
>
>While I haven't seen the gas laws formally worked out, from what I've
>heard, I get the impression that the "probablilistic assumption" is
>not about the laws that govern those particles. But the point is
All these particles together are modelled by means of statistics.
Boyle's law is based on statistical assumptions for ideal gas.
The reason for using statistical models in such cases is the plethora of
molecular interactions and the uniformity of the molecules.
>moot. I never meant to imply that probabalism bothers me, only that
>bad arguments for probabalism do. As interesting they may be, the
>results of physics never really moved me either way. Whatever they
>turn out to be, it wouldn't change the way I hear a symphony or the
>way I think about psychology and a whole bunch of more trivial things.
There is a relationship between arts and physics or mathematics.
Impressionists were influnced by the QM and Cubists by Relativity.
Xenakis is influenced from Markovian Proceses.
Ancient Greek statues and temples obey to certain eucledian abstract rules.
The list can be very long.
What I want to say, is that you personally may not be influenced by science
while listening to a symphony, but some composers certainly did
during the production stage.
>On the relevence of QM to AI: there are many arguments that maintain
>that we cannot achieve AI on a computer because a computer can't model
>quantum effects. Even if it can't so what?
This is indeed an unfortunate argument.
>There are quantum effects
>in water molecules, but we have had some success in modelling fluids.
>The neuron is such a big thing that I don's see why quantum effects
>would be relevant. Even if it were relevent on the neuronal level,
>who says we have to go all the way down to that level to achieve AI?
I have to add to your questioning, that even if it was relevent then
the intelligence AMONG HUMANS should be qualitatively distinct!
But again statistical interactions put things back in order.
But these models work for computers too.
>I may be wrong, but I'm pretty sure we got to the point of building
>transistors smaller than a neuron without taking quantum effects into
>account.
Here you are wrong. Certain light diodes are based exactly on this phenomenon.
But as you said, it is irrelevant to the point of AI.
Randall Holmes
>In article <1992Jun2.1...@guinness.idbsu.edu> hol...@opal.idbsu.edu
>(Randall Holmes) writes:
>> In article <1992Jun2.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
>Clarke) writes:
>> >While not (Turing) machine can predict QM effects, QM effects happen so
>> >the universe "predicts" them. Perhaps a subset of the universe can
>> >be boxed into a convenient black box machine that can "predict" these
>> >effects in a computationaly useful way. (Is the brain such a black
>> box?)
>>
>> The universe does not predict the QM effects; it simply exhibits them.
>> It would only predict them if it were actually deterministic on a
>> deeper level; i.e., if QM were only an approximation to the real
>> situation!
>
>Hence the quotation marks around predicts! Neither may my brain
>calculate (or simulate) intelligence, it may just exhibit it.
Good point. (Not that I agree! -- I still think that thought is
"calculation", but this is an assumption which has to be made
explicit).
>
>> >Quantum mechanics is an example of physics that cannot be fully
>> >simulated.
>>
>> It certainly can be simulated. You simply have to roll dice at the
>> right point (and not stupidly -- i.e., not locally). Also, you have
>> to do really tough mathematics to set up your simulation correctly. I
>> believe that this is actually done -- for very small systems (atoms!).
>
>I suspect our respective undestandings of QM differ too much to
>really reach an agreement now. My intutition is that a
>correct QM simulation would have to simulate every possiblity in an
>Everett style many world's simulation. Shades of the humongous LUT.
I agree absolutely; nobody said it was _feasible_! That's why I
brought up the difficulty of simulating even atoms. But simulation of
the classical universe would be just as hard.
[...]
>But consider the following correspondences:
>
>QM correlations <--?--> qualia
>
>In QM correlations have no effects, but are necessary to understand the
>theory. In the same way qualia are necessary to understand the mind.
I'm not sure that I understand this analogy. I also doubt that one
needs to understand qualia to understand the mind.
>
>hidden variables <--?--> homomculus
>
>Both hidden variables and homonculi are attractive explanations that
>must be avoided, but often crop up in unexpected ways.
I think that "hidden variables" is the correct explanation of the
physical reality (which is not to say that I am disputing QM
predictions). I think that particles are real (they leave little
spots on emulsions; how real can you get?) but that the "waves" are
not. The wave side of the duality is (I think) a mathematical
representation of what information we are allowed to have about the
real particles (restrictions imposed by the quantum of action); and we
actually do "observe" the effects of waves via statistical
distribution of observations of _particles_ (waves are arguably never
directly observed!) However, the behaviour of the waves is so elegant
that we "reify" them. What is necessary to make progress is to figure
out why information about particles (the actual reality) is restricted
in this way. This is hard work; we will make no progress in this
direction as long as we continue to take the easy course of treating
the waves as an independent physical reality (an attractive
explanation which probably should have been avoided). Heisenberg's
original argument treated particles as real and indicated why we could
not determine their position and momentum at the same time; I suspect
that it would be instructive to continue deducing the effects of the
quantum of action in a world of real particles (maybe massless
particles would still be most conveniently treated as waves, but maybe
not) -- the same (maybe not _exactly_ the same) consequences would
follow, but we would understand them better (and we would have to do a
lot more work to derive them -- this is the cost). I think that there
are _reasons_ for the behaviour of QM waves which can be _understood_
in terms of a world of real particles with a quantum of action; at
present the behaviour of the waves is simply postulated. If the exact
QM formalism arose from this explanation (I think that there is an
excellent chance that this is what would happen), we would have the
curious consequence that the "hidden variables" would be unobservable:
first, we do not know that this would be the outcome until we try;
second, a conceptually accessible (not necessarily simple) explanation
with unobservable factors could be preferable to an opaque one
without; third, I believe I have seen an argument that it is possible
to determine the position and momentum of a particle simultaneously
_in retrospect_; information is not necessarily invisible from all
standpoints.
>
>These parallels between AI and QM are interesting on a metascientific level;
>both disciplines have interpretation problems that most practitioners
>ignore. But to me the analogies are so close that one is tempted to
>apply Occam's razor and argue for a deeper connection.
I think there is a connection; both fields suffer from the tendency of
people to ascribe occult powers to themselves.
>--
>Thomas Clarke
>Institute for Simulation and Training, University of Central FL
>12424 Research Parkway, Suite 300, Orlando, FL 32826
>(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
Randall Holmes
It's local properties of the _waves_ that are involved. Such an
experiment is carried out by setting up a state in two widely
separated locations which depends on an unobserved factor in a single
earlier event. One then makes observations at the widely separated
points, and, lo, they agree with one another. These results make
perfect sense (require no explanation at all, in fact) on a hidden
variables interpretation, i.e., on the interpretation that there was a
real underlying value to the unobserved factor in the earlier event
which we did not in fact observe (we couldn't observe it and do the
experiment, in fact). The "non-locality" has to do (on my
interpretation) with the fact that getting extra information about
event A may immediately give me extra information about event B even
if A and B have space-like separation (suppose that event A and event
B are individuals listening to television broadcast years earlier from
Earth, at the same distance from the Solar System in opposite
directions in the conventional frame of reference of the Solar System,
and the information that I have acquired is a copy of the TV schedule;
I then know that if A watched "I Love Lucy" at a certain time, so did
B). The extra factor in the QM experiments is that the information
involved is information which we are not "allowed" to have when the
experiment is first performed (there is no contemporary TV schedule to
be had); but the actual situation is one in which it is perfectly easy
to understand why the same thing will happen at the two widely
separated points, for essentially the reason described. Suppose that
I do not have access to contemporary TV schedules; I can still be
certain that when I find out what was being watched at event A, I will
discover that the same program was watched at event B.
Thomas Clarke
> In article <1992Jun2.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
Clarke) writes:
> >In article <1992Jun2.1...@guinness.idbsu.edu> hol...@opal.idbsu.edu
> >(Randall Holmes) writes:
> >> In article <1992Jun2.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
> >Clarke) writes:
> >> >While not (Turing) machine can predict QM effects, QM effects happen so
> >> >the universe "predicts" them. Perhaps a subset of the universe can
> >> >be boxed into a convenient black box machine that can "predict" these
> >> >effects in a computationaly useful way. (Is the brain such a black
> >> box?)
> >>
> >> The universe does not predict the QM effects; it simply exhibits them.
> >> It would only predict them if it were actually deterministic on a
> >> deeper level; i.e., if QM were only an approximation to the real
> >> situation!
> >
> >Hence the quotation marks around predicts! Neither may my brain
> >calculate (or simulate) intelligence, it may just exhibit it.
>
> Good point. (Not that I agree! -- I still think that thought is
> "calculation", but this is an assumption which has to be made
> explicit).
>
distinction I perceive on odd days.
> I think that "hidden variables" is the correct explanation of the
> physical reality (which is not to say that I am disputing QM
> predictions). I think that particles are real (they leave little
> spots on emulsions; how real can you get?) but that the "waves" are
> not. The wave side of the duality is (I think) a mathematical
> representation of what information we are allowed to have about the
> real particles (restrictions imposed by the quantum of action); and we
> actually do "observe" the effects of waves via statistical
> distribution of observations of _particles_ (waves are arguably never
> directly observed!)
We're getting into sci.physics territory, but I think you get into trouble
with a pure particle picture also. See Bernard d'Espagnat "The Quantum
Theory and Reality", Sci. Am. (1983? amazing - I tore the article out of
my copy and it has no date on it - e-mail me and I'll try to find the
citation). d'Espagnat uses a wave-free discussion to show how reasoning
about independently existing particles with definite properties (spin
in this case) leads to Bell's inequalities which contradict quantum
mechanics. The experiments have been done to test the inequalities
and quantum mechanics wins!
Intense debate about what this all means continues in philosophical circles.
Physicists just yawn and say "of course."
Thomas Clarke
> In article <1992Jun2.1...@guinness.idbsu.edu> hol...@opal.idbsu.edu
> (Randall Holmes) writes:
> >If you think that the waves are real physical phenomena, then FTL
> >signalling is involved. I think it is quite reasonable to suppose
> >that the universe has "real states"; the particle side of the
> >particle-wave duality is real; the wave side encodes restrictions on
> >what information we can have about the real situation.
> I'm not sure I understand you, but I suspect you're wrong. It's not just
> a matter of observing a correlation across a huge distance. It's that
> you and another observer do something (twist a calcite crystal) and
> immediately, before any signal could move from one to the other, the
> correlation rate changes, in a way that can't be explained by local
> properties of the particles. There's something non-local going on here.
> I don't see what difference it makes if you call it "phyical" or not.
My two cents. The particle picture and the wave picture are mathematically
equivalent. I think it was Dirac who dicovered that Heisevberg's and
Schrodinger's views could be reconciled in the arena of Hilbert space;
von Neuman dotted the mathematical i's and crossed the t's.
Heisenberg has an infinitely long vector of amplitudes which represent the
strength of various particle states (square magnitude for probability).
Mathematically Heisenberg's vectors are in l2, the space of square summable
infinite sequences with dot product equal to sum of products.
Schrodinger has wave functions whose amplitude represents state amplitude
at a given point in space (square for probability density). Mathematically
Schrodinger's wave functions are in L2, the space of square integrable
functions with dot product equal to the integral of the product.
In functional analysis it is proved that L2 is isomorphic ot l2 (at least
for the "nice" cases involved in quantum mechanics). Whatever happens
in l2 has a corresponding result in L2. (Think of the Fourier transform
of a periodic continuous signal which results in a infinite series of
Fourier coefficients).
Thus it seems to me the mathematics says that if waves give you heartburn,
then particles will give you indigestion.
Gordon Joly
< > I may be wrong, but I'm pretty sure we got to the point of building
< > transistors smaller than a neuron without taking quantum effects into
< > account.
< >
<
< Ouch. The Zenner diode, which use quantum tunnellling, is just one
< example of QM. And CCD cameras(?)
FYI, there is an article in the Juine Scientific American on
"single-electron" devices.
Gordon.
Mark William Hopkins
William Hopkins) writes:
>There's big, obvious, gaping holes in QM!
>
>The biggest of them all is that it can't explain gravity.
>The second is that it can't explain the dimensional or topological
>structure of spacetime.
>The third is that it can't explain why this dimensionality is 3+1.
>The fourth is that it can't explain the origin or significance of the Planck
>units.
>The fifth is that it can't explain why or how the continuum breaks down at
>Planck scales.
>The sixth is that it can't explain the origin of the Universe or its fate!
>
>Need I go on? The list is endless.
In article <1992Jun03.2...@spss.com> mark...@spss.com (Mark Rosenfelder) writes:
>I'll add some to your list: It can't explain why electrons have the mass
>they do; it can't explain the three-body problem; it can't explain
>population genetics; and it can't explain Cubism.
I see the logic clearly:
<List of highly relevant and topical Quantum Theoretic issues>
>"here lemme add some more"
<Semi-relevant issue, also false>
<Irrelevamnt issues>
>"See? Your list is irrelevant. Therefore Quantum Theory is complete."
Gee, what an ingenious ploy. "My theory can't be wrong, because your
exceptions don't count"...
It can't explain gravity: it gives all the WRONG predictions concerning
anything that has to do with gravity.
Let ME add one to the list:
It can't explain the strong force. It can't even give any predictions about
it. It's not renormalizeable.
The theory has gaping holes in it and is woefully incomplete.
Daryl McCullough
>I think that "hidden variables" is the correct explanation of the
>physical reality (which is not to say that I am disputing QM
>predictions). I think that particles are real (they leave little
>spots on emulsions; how real can you get?) but that the "waves" are
>not. The wave side of the duality is (I think) a mathematical
>representation of what information we are allowed to have about the
>real particles (restrictions imposed by the quantum of action); and we
>actually do "observe" the effects of waves via statistical
>distribution of observations of _particles_ (waves are arguably never
>directly observed!) However, the behaviour of the waves is so elegant
>that we "reify" them.
Hi, Randall! I don't know if you remember me, but we were in a set
theory course together at Cornell. We discussed your research into
NFU quite a bit.
Anyway, while I have enjoyed your comments about AI, I have to
disagree with your discussion of quantum mechanics. I don't think that
there is a consistent interpretation of a wave function as a
measurement of our information about the location of a particle. The
main difficulty with such an interpretation is that there is no role
for the phase of the wave function in such an interpretation, and the
phase is crucial for such things as interference effects.
I tend to believe that *particles* are the reification of terms in an
expansion of the wave function. In quantum field theory, the reality
of particles is even more in question. The fundamental quantity seems
to be the field (the analog of the wave function). Particles are seen
to be excitations in the fields, not entities in their own right.
>What is necessary to make progress is to figure out why information
>about particles (the actual reality) is restricted in this way.
>This is hard work; we will make no progress in this direction as
>long as we continue to take the easy course of treating
>the waves as an independent physical reality (an attractive
>explanation which probably should have been avoided).
Once again, I disagree strongly. There is no reason (that I know of)
to doubt that waves are part of physical reality. The so-called
"collapse of the wave function" tends to make people think of the
parallel with ordinary probability: When we flip a coin in the dark,
we can describe the situation as "Heads, with 50% probability, and
tails, with 50% probability." However, when we turn on the light,
the coin seems to "collapse" into a define state, either heads or
tails. Obviously, in this case, there really is no collapse; the
coin was either heads or tails before we turned on the light, we
just didn't know which.
Discussion of quantum mechanics, without the mathematics of
interference, might lead people to think that the situation is
comparable to the case with flipping a coin. The particle really has a
position, we just don't know what it is. When we observe the particle,
the collapse of the wave function is simply a change in our knowledge.
What's the big deal?
However, the mathematics does not support this interpretation, because
of interference effects between alternatives. There is a testable
difference between the following two situations:
1. The particle either has spin-up, or spin-down, we just don't know
which.
2. The particle is in a superposition of a spin-up state and a spin-down
state.
Mark Rosenfelder
>It's local properties of the _waves_ that are involved. Such an
>experiment is carried out by setting up a state in two widely
>separated locations which depends on an unobserved factor in a single
>earlier event. One then makes observations at the widely separated
>points, and, lo, they agree with one another. These results make
>perfect sense (require no explanation at all, in fact) on a hidden
>variables interpretation, i.e., on the interpretation that there was a
>real underlying value to the unobserved factor in the earlier event
>which we did not in fact observe (we couldn't observe it and do the
>experiment, in fact). The "non-locality" has to do (on my
>interpretation) with the fact that getting extra information about
>event A may immediately give me extra information about event B even
>if A and B have space-like separation (suppose that event A and event
>B are individuals listening to television broadcast years earlier from
>Earth, at the same distance from the Solar System in opposite
>directions in the conventional frame of reference of the Solar System,
>and the information that I have acquired is a copy of the TV schedule;
>I then know that if A watched "I Love Lucy" at a certain time, so did
Perhaps we are not talking about the same thing. Let me attempt to
describe the problem in more detail. I hope physicists reading this
will gently correct any errors I make.
An apparatus is set up which emits a stream of spin-correlated photon pairs
in opposite directions. Observers at points A and B are using calcite
crystals to test the polarization of each pair of photons. The following
results are noted:
Angle between orientation Errors in their
of A's and B's calcite spin observations
0 0 in 4
30 1 in 4
60 3 in 4
90 4 in 4
Now, this seems to be what you are talking about: we observe a correlation
between A's and B's observations, but we have no reason to suppose any
superluminal (or even subluminal) connection between the photons once they
leave the apparatus.
A local hidden-variable theory could explain the observations by supposing
that each photon, as it leaves the apparatus, "knows" how it is going to
react to each possible setting of the observer's calcite (perhaps by
having an explicit angle of polarization, rather than the binary spin
value assumed by QM).
A photon leaves the apparatus, and A twists his calcite 30 degrees left.
No problem: the photon knows how to respond to any angle. Before either
photon is measured, but after they've left the apparatus, however, B turns
*her* calcite 30 degrees right. If the locality assumption is true, that
can make no difference in A's observations. But it does: the observations
are now correlated at the 3-errors-in-4 rate.
To know how to respond to A's observation, the photon needs to react not
only to the orientation of A's calcite, but to that of B's too, and its
knowledge is not limited by the speed of light.
Hidden variables are not ruled out, but solely local ones are.
Mark Rosenfelder
>I see the logic clearly:
><List of highly relevant and topical Quantum Theoretic issues>
>>"here lemme add some more"
><Semi-relevant issue, also false>
><Irrelevamnt issues>
>>"See? Your list is irrelevant. Therefore Quantum Theory is complete."
>
>Gee, what an ingenious ploy. "My theory can't be wrong, because your
>exceptions don't count"...
>The theory has gaping holes in it and is woefully incomplete.
Gee, what an ingenious ploy. Redefine your opponent's terms, fantasize
about what he's saying, and establish your point by repetition.
I stated very clearly, in two separate posts, exactly what I meant by
"holes" (namely, contradictions with experiment). If you want to use the
same word to mean something else (namely, things not covered by a theory),
go ahead, but don't pretend that you're addressing what I'm saying.
Someone who reads English would also have noted that I made no claims about
QM being "right" or "complete".
A discussion consisting of endless correction of your misinterpretations of
my words is infinitely dull; let's stop right here.
Randall Holmes
No, they aren't. The particle has a whole spectrum of reactions to
the settings of the instruments, which are probabilistically
correlated with each other. They are decided at the source.
Certainly a single angle of polarization won't work, but independent
yes-or-no answers for each angle which are correlated with one another
with probabilities depending on angles in the usual way work fine. No
one said that the hidden variables were simple!
Randall Holmes
I know all of this very well. Nonetheless... I will try to respond
at more length later.
Daryl McCullough
[about quantum nonlocality]
>It's local properties of the _waves_ that are involved. Such an
>experiment is carried out by setting up a state in two widely
>separated locations which depends on an unobserved factor in a single
>earlier event. One then makes observations at the widely separated
>points, and, lo, they agree with one another. These results make
>perfect sense (require no explanation at all, in fact) on a hidden
>variables interpretation, i.e., on the interpretation that there was a
>real underlying value to the unobserved factor in the earlier event
>which we did not in fact observe (we couldn't observe it and do the
>experiment, in fact).
Randall, I think you are behind the times on this. Einstein *thought*
that a hidden-variables interpretation would explain the seeming
nonlocality of quantum mechanics, but John Bell in fact showed just
the opposite: the nonlocality *cannot* be explained by hidden
variables theories (Bell's Theorem).
>The "non-locality" has to do (on my interpretation) with the fact
>that getting extra information about event A may immediately give
>me extra information about event B even if A and B have space-like
>separation
That explanation has been pretty much ruled out. There is no way
to reproduce the statistical predictions of quantum mechanics by
such a hidden variables theory.
Daryl McCullough
ORA Corp.
Ithaca, NY
Nick Cassimatis
>In article <1992Jun3.0...@news.media.mit.edu> n...@media.mit.edu (Nick Cassimatis) writes:
>> san...@inf.ethz.ch (Philip Santas) writes:
>
>There is a relationship between arts and physics or mathematics.
>Impressionists were influnced by the QM and Cubists by Relativity.
>Xenakis is influenced from Markovian Proceses.
>Ancient Greek statues and temples obey to certain eucledian abstract rules.
>The list can be very long.
Though the cubists and the impressionists said they were influenced by
those areas of physics, I really doubt it. First of all, how many of
them could understand physics? (The first Impressionist exhebition
was in 1862 -- way before QM.) Second, I think their talking about
physics is on a par with those people who call the 20th centruy an age
of anxiety and then say that the results of physics were part of the
cause of this. My opinion on this is that those results were the
cause of no real anxiety (except to scientists!!) but that instead,
they were just something that came out of people's mouths and pens
when they reveled in their anxitety and alienation. QM and relativity
(along with consciousness and a few others) seem to be these big
bogeys that are thrown into a discussion whenever someone want's to
interject some despair or pessimism or mysticicism.
>>On the relevence of QM to AI: there are many arguments that maintain
>>that we cannot achieve AI on a computer because a computer can't model
>>quantum effects. Even if it can't so what?
>
>This is indeed an unfortunate argument.
It's nice to be in agreement with someone on this!! I really can't
believe how wide-spread discussions of Quantum Mechanics are. It
seems they're everywhere. There was (maybe still is) a big discussion
on sci.phil.meta on the relation between consciousness and quantum
mechanics. DID ANYONE EVER WONDER WHY YOU RARELY SEE DISCUSSIONS ON
THE RELEVENCE OF QM TO MOTOR CONTROL OR LANGUAGE OR ANY OTHER ASPECT
OF COGNITION? Why focus only on consciousness? This phenomena
convinces me even more that that QM is just a wild card that makes
people glassy eyed enough to accept mysticism.
Much talk about Quantum Mechanics a sign of intellectual decadence.
>>I may be wrong, but I'm pretty sure we got to the point of building
>>transistors smaller than a neuron without taking quantum effects into
>>account.
>
>Here you are wrong. Certain light diodes are based exactly on this phenomenon.
I meant to suggest not that we got to the present state of the art
without QM, but to the state of a transistor being smaller than a
neuron -- which happened some years ago -- without QM.
>But as you said, it is irrelevant to the point of AI.
But as it is irrelevant, I won't mention it again.
-Nick
Stephen P Spackman
observable, there's no problem with it being "superluminal". After
all, the problem with FTL "information" transmission is an intutive
one, but the intuitions it offends are those honed by questions of
_observability_, which is precisely what is _not_ at issue here. You
may be seeing global effects, but locality isn't "leaking".
Or perhaps I'm being dense.
$0.02
Thomas Clarke
> DID ANYONE EVER WONDER WHY YOU RARELY SEE DISCUSSIONS ON
> THE RELEVENCE OF QM TO MOTOR CONTROL OR LANGUAGE OR ANY OTHER ASPECT
> OF COGNITION? Why focus only on consciousness? This phenomena
> convinces me even more that that QM is just a wild card that makes
> people glassy eyed enough to accept mysticism.
>
> Much talk about Quantum Mechanics a sign of intellectual decadence.
>
Much talk about QM and the mind _is hokum_, but by totally rejecting the
possible relevance of quantum effects to mind, the baby may be going out
with the bathwater. Quantum mechanics is ill-understood and mysterious
and inasmuch as QM is a "theory of everything", the universe is ill-
understood and mysterious.
I think my original point is still valid: Harnad's thesis that grounding
in the physical world is important to mind may find a physical explanation
in QM phenomena. Harnad's argument therefore cannot be refuted by
arguments based on simulation of classical physics.
The central problem of AI - Searle's anyway - is that a machine behaving
intelligently may not be conscious - have qualia etc. etc. Even Searle
would agree that it is possible to build a zombie - use a humongous LUT
if all else fails.
Achieving consciousness may require something else, that something else
might well be quantum. Thus it might be the case that you can build
a zombie using deterministic computational devices, but that to achieve
consciousness you'll have to wire in a few Josephson junctions.
{Consciousness might arise through grounding via correlations with the
rest of the quantum world. The quantum correlations are of course
not directly observable, but then neither is consciousness}
DID ANYONE EVER WONDER WHY ALL THE DISCUSSIONS OF THE PHILOSOPHICAL
PROBLEMS OF QUANTUM MECHANICS ALWAYS SOMEHOW INVOLVE A
CONSCIOUS OBSERVER?
Drew McDermott
>
>DID ANYONE EVER WONDER WHY ALL THE DISCUSSIONS OF THE PHILOSOPHICAL
>PROBLEMS OF QUANTUM MECHANICS ALWAYS SOMEHOW INVOLVE A
>CONSCIOUS OBSERVER?
The many-worlds interpretation dispenses with conscious observers,
which is good, because if there is ever to be an explanation of
consciousness as a property of matter, our fundamental theory of matter
cannot presuppose consciousness.
-- Drew McDermott
Mark Shanks
>Cassimatis) writes:
>> DID ANYONE EVER WONDER WHY YOU RARELY SEE DISCUSSIONS ON
>> THE RELEVENCE OF QM TO MOTOR CONTROL OR LANGUAGE OR ANY OTHER ASPECT
>> OF COGNITION? Why focus only on consciousness? This phenomena
>> convinces me even more that that QM is just a wild card that makes
>> people glassy eyed enough to accept mysticism.
I don`t see the connection between QM and mysticism, "A spiritual discipline
aiming at direct union or communion with God or with ultimate reality through
trancelike contemplation or deep meditation." Are there aspects of quantum
theory that suggest to you something less than (or maybe "other than" is
better) an attempt at objective scientific investigation? Why is QM a "wild
card"? Is QM an ideology?
>> Much talk about Quantum Mechanics is a sign of intellectual decadence.
>>
>As the "originator" of the current QM threads I must comment.
>
>Much talk about QM and the mind _is hokum_, but by totally rejecting the
>possible relevance of quantum effects to mind, the baby may be going out
>with the bathwater. Quantum mechanics is ill-understood and mysterious
>and inasmuch as QM is a "theory of everything", the universe is ill-
>understood and mysterious.
As has been argued in previous postings, QM has not explained everything,
and possibly has fundamental inherent flaws. But to date, it's hypotheses
have consistently been verified through experimentation. It seems that
many of the elements of the theory have an irritating effect on people
who seem to thing that QM has a hidden agenda to warp "Western rationality"
and inflict some form of mysticism on the scientific community.
>
>I think my original point is still valid: Harnad's thesis that grounding
>in the physical world is important to mind may find a physical explanation
>in QM phenomena. Harnad's argument therefore cannot be refuted by
>arguments based on simulation of classical physics.
>
>The central problem of AI - Searle's anyway - is that a machine behaving
>intelligently may not be conscious - have qualia etc. etc. Even Searle
>would agree that it is possible to build a zombie - use a humongous LUT
>if all else fails.
>
>Achieving consciousness may require something else, that something else
>might well be quantum.
Agreed. And it might be something else. But QM is a valid concern, and
IMHO, is NOT a wild card for glassy-eyed decadent mystics.
------------------------------------------------------------------------------
Mark Shanks
Principal Engineer, 777 Displays
------------------------------------------------------------------------------
Thomas Clarke
Many worlds does eliminate the requirement that a conscious
observer determines the moment of observation. It does so,
however, by postulating that all possible outcomes of all
possible experimental observations occur and continue to
evolve in parallel. Conscious observers then must have
the peculiar ability to sense only one possible observational
track. How this observational track is selected is not
specified. It always struck me that selecting a single
observational track for a conscious observer in Many Worlds
was exactly equivalent to the conscious observer "collapsing
the wavefunction" in more conventional approaches. No matter
which way you slice it, consciousness seems to still have a
peculiar relation to the physical world in quantum mechanics.
Is it a priori obvious that consciousness is not a part of
the material world? Indeed, physicists seem to be driven to this
conclusion as the only way to theoretically account for observation.
Randall Holmes
>hol...@opal.idbsu.edu (Randall Holmes) writes:
>
>[about quantum nonlocality]
>
>>It's local properties of the _waves_ that are involved. Such an
>>experiment is carried out by setting up a state in two widely
>>separated locations which depends on an unobserved factor in a single
>>earlier event. One then makes observations at the widely separated
>>points, and, lo, they agree with one another. These results make
>>perfect sense (require no explanation at all, in fact) on a hidden
>>variables interpretation, i.e., on the interpretation that there was a
>>real underlying value to the unobserved factor in the earlier event
>>which we did not in fact observe (we couldn't observe it and do the
>>experiment, in fact).
>
>Randall, I think you are behind the times on this. Einstein *thought*
>that a hidden-variables interpretation would explain the seeming
>nonlocality of quantum mechanics, but John Bell in fact showed just
>the opposite: the nonlocality *cannot* be explained by hidden
>variables theories (Bell's Theorem).
It just isn't so. I produced a hidden variables explanation of the
behaviour of Mermin's device after about an hour of reflection (the
reflection was required to recover my earlier thinking on the subject,
which took a lot more than an hour!). Bell's Theorem says that a
certain class of hidden-variables theories cannot explain the Mermin
results; these theories are deterministic, among other things
(determinism is not needed; a suitably simplistic notion of what the
hidden variables look like will get one into trouble by itself).. I
am not advocating a deterministic position, nor do I think that the
hidden state is very simple (it is not anything like a macroscopic
spin). Suppose that a device emits two electrons with opposite spin
in opposite directions. Whenever the spin of an electron along a
given axis is measured, it will be found to have its axis of rotation
pointing in that direction or in the opposite direction. Set up two
measuring devices to catch the two electrons emerging from the device,
which are set up to measure spin along an axis which can be adjusted
(independently for the two devices). The events of the two devices
capturing the two electrons emitted in a given trial have space-like
separation. When a given pair of electrons is emitted, if the two
devices are at the same setting, they will measure the electrons as
having opposite spin. Now suppose that we adjust one of the
detectors; there is suddenly a fall-off in correlation of the two
measurements, depending on the angle of adjustment. Now suppose
detector A is at a slight angle from detector B, so we are set up to
expect a certain failure of correlation. We send off a pair of
electrons. While they are in flight, we twiddle detector B back into
line with detector A -- notice that no information can be transferred
from detector B to detector A in time to affect detector A in any way
-- suddenly, we have a perfect correlation between the two spins.
This is supposed to establish superluminal communication between the
two detectors. It establishes no such thing. Here is an alternate
hypothesis. The state of a particle, when it is emitted, consists of
a definite yes-no answer for each angle (not a hidden single axis of
polarization -- Bell's argument does kill this). Each of these
answers is diametrically opposed as between the two electrons. The
answers vis-a-vis two angles for a given one of the two electrons are
correlated probabilistically in the appropriate degree determined by
the angle. This kind of hidden state, decided at the source, not by
random events with space-like separation at the detectors, will
exhibit the exact behaviour seen in the Mermin experiment and does not
involve non-locality. What is unclear is what the physical meaning of
the state is, but that is already unclear in QM as it stands.
>
>>The "non-locality" has to do (on my interpretation) with the fact
>>that getting extra information about event A may immediately give
>>me extra information about event B even if A and B have space-like
>>separation
>
>That explanation has been pretty much ruled out. There is no way
>to reproduce the statistical predictions of quantum mechanics by
>such a hidden variables theory.
I just did it. The point is that the rather stringent conditions on
Bell's Theorem are not satisfied by this hidden state hypothesis.
Read the conditions. What is true is that the hidden states have to
be very complicated and the resulting theory will not be
deterministic. A more succinct answer is that this is a hidden
variables theory, but it is not "such" a hidden variables theory.
The question of cost arises; is this too complicated an expedient for
avoiding the non-locality in the QM formalism revealed by Bell's
Theorem? I think the answer is that it is necessary at any cost to
avoid the confusion of physical objects (particles) with mathematical
constructs (probability waves) which are not physical objects [of
course, I maintain that they are real objects, but they are not
physical objects]; as long as the probability waves are treated as if
they were an independent reality, we are evading the difficult problem
of "explaining" the reasons why the probability waves work to describe
the behaviour of particles with information restrictions imposed by
the quantum of action. As regards unobservable factors, there is
nothing to choose between the theories; hidden states of which only a
part can be observed in a single operation versus superluminal
communication which cannot be used to transmit information make an
unappealing choice either way. "I was thinking of a plan to dye one's
whiskers green..."
>
>Daryl McCullough
>ORA Corp.
>Ithaca, NY
Randall Holmes
>In article <1992Jun4.2...@oracorp.com> da...@oracorp.com (Daryl McCullough) writes:
>>hol...@opal.idbsu.edu (Randall Holmes) writes:
>>
>>[about quantum nonlocality]
>>
>>>It's local properties of the _waves_ that are involved. Such an
>>Randall, I think you are behind the times on this. Einstein *thought*
>>that a hidden-variables interpretation would explain the seeming
>It just isn't so. I produced a hidden variables explanation of the
>behaviour of Mermin's device after about an hour of reflection (the
>reflection was required to recover my earlier thinking on the subject,
>which took a lot more than an hour!). Bell's Theorem says that a
>>
>>>The "non-locality" has to do (on my interpretation) with the fact
>>>that getting extra information about event A may immediately give
>>>me extra information about event B even if A and B have space-like
>>>separation
>>
>>That explanation has been pretty much ruled out. There is no way
>>to reproduce the statistical predictions of quantum mechanics by
>>such a hidden variables theory.
>
>I just did it. The point is that the rather stringent conditions on
>Bell's Theorem are not satisfied by this hidden state hypothesis.
>
>
>
>>
>>Daryl McCullough
>>ORA Corp.
>>Ithaca, NY
>
>
>--
>The opinions expressed | --Sincerely,
>above are not the "official" | M. Randall Holmes
>opinions of any person | Math. Dept., Boise State Univ.
>or institution. | hol...@opal.idbsu.edu
Fragments included to indicate what material I am talking about.
Please, let's not discuss hidden variables further here. The original
point I wanted to make is that QM has nothing to do with the nature of
consciousness; this is adequately established by the many-worlds
interpretation. I don't think that the discussion of hidden variables
is appropriate in comp.ai.philosophy. My apologies for getting
carried away with it! I'll be happy to discuss it via e-mail.
By the way, Daryl, I do remember who you are; I recognized your name
and employer on your postings some time ago. Hi!
Richard Ottolini
discretization and probability. People who use QM to explain the metaphysics
of the universe and the mind are overinterpreting the framework. Often it
seems to be people using one thing they don't understand to explain something
else they don't understand.
Thomas Clarke
> QM is a mathematical framework for explaining physical observations employing
> discretization and probability.
Not so. Discretization arises from the mathematics of operators in
Hilbert space. Probability arises from quantum measurement theory
which specifies how to turn a vector into Hilbert space into
probabilities of observation.
> People who use QM to explain the metaphysics
> of the universe and the mind are overinterpreting the framework. Often it
> seems to be people using one thing they don't understand to explain something
> else they don't understand.
Actually, I'm impressed by the level of knowledge of quantum mechanics
expressed in this newsgroup.
Philip Santas
In article <1992Jun5.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas Clarke) writes:
>In article <1992Jun5.0...@news.media.mit.edu> n...@media.mit.edu (Nick
>Cassimatis) writes:
>> DID ANYONE EVER WONDER WHY YOU RARELY SEE DISCUSSIONS ON
>> THE RELEVENCE OF QM TO MOTOR CONTROL OR LANGUAGE OR ANY OTHER ASPECT
>> OF COGNITION? Why focus only on consciousness? This phenomena
>> convinces me even more that that QM is just a wild card that makes
>> people glassy eyed enough to accept mysticism.
What is not well understood leads ALWAYS to mysticism.
Since the majority of people have no idea about QM it is obvious
that there many such missconceptions.
>> Much talk about Quantum Mechanics a sign of intellectual decadence.
^^^^^^^^^
I would say that the much blaming about QM is a sign of impotence to produce
something positive.
>Much talk about QM and the mind _is hokum_, but by totally rejecting the
>possible relevance of quantum effects to mind, the baby may be going out
>with the bathwater. Quantum mechanics is ill-understood and mysterious
>and inasmuch as QM is a "theory of everything", the universe is ill-
>understood and mysterious.
QM is NOT theory of everything, in the sence in which humans need explanation of them.
To explain the weather phenomena you do not go down into the subatomic level,
you do not arrive even to molecular level, but you remain at the macroworld
and with statistical methods you try to find laws being statistically valid.
You cannot examine evrything by QM, as a doctor does not treat your deseases
by relativity theories, but he fights against viruses with elements of the same
scale. If universe is ill-understood this is NOT a problem of medicine,
mechanical or computer engineering, nor a problem of AI.
>I think my original point is still valid: Harnad's thesis that grounding
>in the physical world is important to mind may find a physical explanation
>in QM phenomena. Harnad's argument therefore cannot be refuted by
>arguments based on simulation of classical physics.
But even simulation of QM does not reinforce this argument either.
There is absolutely no relevance between these two fields.
Harnad has to prove that his statement is true, but the people of the
opposite field do not need to prove anything.
>The central problem of AI - Searle's anyway - is that a machine behaving
>intelligently may not be conscious - have qualia etc. etc. Even Searle
>would agree that it is possible to build a zombie - use a humongous LUT
>if all else fails.
Without a definition of conciousness one can prove ANYTHING about it.
If you define conciousness as an entirely human property, then obviously
machines do not fit into this definition. If you define it as a non human
property, then humans do not fit into it. So what is conciousness after all?
BTW who said that AI research is for producing conciousness?
>Achieving consciousness may require something else, that something else
>might well be quantum.
Of course! But you cannot know what is required if you don't know what you want
to produce. Can one tell what is required for achieving ojhflg1pi :-)
>DID ANYONE EVER WONDER WHY ALL THE DISCUSSIONS OF THE PHILOSOPHICAL
>PROBLEMS OF QUANTUM MECHANICS ALWAYS SOMEHOW INVOLVE A
>CONSCIOUS OBSERVER?
I thought that they involve photons.
Nick Cassimatis
>In article <1992Jun5.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas Clarke) writes:
>>In article <1992Jun5.0...@news.media.mit.edu> n...@media.mit.edu (Nick
>>Cassimatis) writes:
>>> DID ANYONE EVER WONDER WHY YOU RARELY SEE DISCUSSIONS ON
>>> THE RELEVENCE OF QM TO MOTOR CONTROL OR LANGUAGE OR ANY OTHER ASPECT
>>> OF COGNITION? Why focus only on consciousness? This phenomena
>>> convinces me even more that that QM is just a wild card that makes
>>> people glassy eyed enough to accept mysticism.
>
>I don`t see the connection between QM and mysticism, "A spiritual discipline
>aiming at direct union or communion with God or with ultimate reality through
>trancelike contemplation or deep meditation." Are there aspects of quantum
>theory that suggest to you something less than (or maybe "other than" is
>better) an attempt at objective scientific investigation? Why is QM a "wild
>card"? Is QM an ideology?
I have not quarrel with QM itself. What bothers me is that it is
Philip Santas
In article <1992Jun5.0...@news.media.mit.edu> n...@media.mit.edu (Nick Cassimatis) writes:
>san...@inf.ethz.ch (Philip Santas) writes:
>>
>>There is a relationship between arts and physics or mathematics.
>>Impressionists were influnced by the QM and Cubists by Relativity.
>>Xenakis is influenced from Markovian Proceses.
>>Ancient Greek statues and temples obey to certain eucledian abstract rules.
>>The list can be very long.
>
>Though the cubists and the impressionists said they were influenced by
>those areas of physics, I really doubt it. First of all, how many of
>them could understand physics?
Van Gogh's style was definitely influenced by the theory of quanta.
The temples in anccient Athens were built following mathematical models.
The same is with computer music and Xenakis. Enjoying them of course
does not require such knowledge: driving a car does not need theory
thermodynamics, but motor designers have this very deep into their mind.
>Second, I think their talking about
>physics is on a par with those people who call the 20th centruy an age
>of anxiety and then say that the results of physics were part of the
>cause of this.
Humans had fear of many things durings during the centuries.
But poor journalism and superficial knowledge of reality are responsible
for the fact you refer to.
>QM and relativity
>(along with consciousness and a few others) seem to be these big
>bogeys that are thrown into a discussion whenever someone want's to
>interject some despair or pessimism or mysticicism.
Or to write some papers and improve his academic position :-)
Randall Holmes
[...]
>
>Many worlds does eliminate the requirement that a conscious
>observer determines the moment of observation. It does so,
>however, by postulating that all possible outcomes of all
>possible experimental observations occur and continue to
>evolve in parallel. Conscious observers then must have
>the peculiar ability to sense only one possible observational
>track. How this observational track is selected is not
>specified.
The whole point is that it is not _selected_!
It always struck me that selecting a single
>observational track for a conscious observer in Many Worlds
>was exactly equivalent to the conscious observer "collapsing
>the wavefunction" in more conventional approaches.
There is no selection; it is simply that the different versions of the
observer in different worlds (all of whom exist) have no way to
communicate with one another, so are not aware of one another
(presumably interference phenomena remain possible, but are unlikely
due to the macroscopic nature of the observer). Each one wonders why
he was "selected", along with all of the others.
[...]
>Thomas Clarke
>Institute for Simulation and Training, University of Central FL
>12424 Research Parkway, Suite 300, Orlando, FL 32826
>(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
--
David Chalmers
>>The many-worlds interpretation dispenses with conscious observers,
>>which is good, because if there is ever to be an explanation of
>>consciousness as a property of matter, our fundamental theory of matter
>>cannot presuppose consciousness.
>Many worlds does eliminate the requirement that a conscious
>observer determines the moment of observation. It does so,
>however, by postulating that all possible outcomes of all
>possible experimental observations occur and continue to
>evolve in parallel. Conscious observers then must have
>the peculiar ability to sense only one possible observational
>track. How this observational track is selected is not
>specified. It always struck me that selecting a single
>observational track for a conscious observer in Many Worlds
>was exactly equivalent to the conscious observer "collapsing
>the wavefunction" in more conventional approaches. No matter
>which way you slice it, consciousness seems to still have a
>peculiar relation to the physical world in quantum mechanics.
There doesn't need to be any "selection" of a "single observational
track". All of these "observational tracks" will have conscious
observers in them. To be sure, to each of those observers, it will
seem as if a single track has been slected, but that's just an
illusion of perspective. One has to get away from a simple
deterministic view of personal identity here: you might think
that there's a fact of the matter about which of several "tracks"
*I* will be in in the future -- as opposed to those other beings
that aren't "me". But there need be no such facts. All of those
future beings bear exactly the same relation to me now.
What does need to be explained in the Everett interpretation is why,
given that the world is a giant superposition, there are observers
that are conscious of seemingly non-superposed states. But that's
not too much of a problem, with a decent theory of consciousness
in hand: such observers are only determined by a limited part of
the wavefunction, namely by information present in one of its
eigenstates; they simply don't have access to information elsewhere
in the wavefunction.
I don't agree with Drew McDermott that the Everett interpretation
"dispenses" with conscious observers; they're still around, and
understanding the role of observership helps us understand the
theory (in particular, it helps us understand why the world doesn't
*seem* superposed). It's just that on this interpretation,
consciousness doesn't play any fundamental role in the physical
laws (which is probably what McDermott meant). Which makes for
a much more satisfying physical theory, if you ask me.
--
Dave Chalmers (da...@cogsci.indiana.edu)
Center for Research on Concepts and Cognition, Indiana University.
"It is not the least charm of a theory that it is refutable."
Neil Rickert
>
>But even simulation of QM does not reinforce this argument either.
>There is absolutely no relevance between these two fields.
>Harnad has to prove that his statement is true, but the people of the
>opposite field do not need to prove anything.
This discussion sure has wandered far.
To be fair to Harnad, I think I should point out that he said nothing about
QM in his postings. I suspect he would agree that QM is probably not
relevant.
Harnad's assertions about grounding don't depend on QM. They are based
on his experience as a psychologist. He sees interactions between the body
and the world which to him seem critical to human intelligence, and he is
convinced the equivalent cannot be achieved digitally. At least some of
those who disagree with Harnad's assertions (and I include myself here) do
so because we believe he underestimates the capabilities of a computer.
Perhaps if we knew more psych and Harnad knew more CS, we would discover
that our disagreements are quite minor.
David Chalmers
>Here is an alternate
>hypothesis. The state of a particle, when it is emitted, consists of
>a definite yes-no answer for each angle (not a hidden single axis of
>polarization -- Bell's argument does kill this). Each of these
>answers is diametrically opposed as between the two electrons. The
>answers vis-a-vis two angles for a given one of the two electrons are
>correlated probabilistically in the appropriate degree determined by
>the angle. This kind of hidden state, decided at the source, not by
>random events with space-like separation at the detectors, will
>exhibit the exact behaviour seen in the Mermin experiment and does not
>involve non-locality.
This simply doesn't work. The hypothesis that the state of each
particle carries a definite yes-no answer for each angle (or even
a definite probability for each angle) is inconsistent with the
experimental results -- unless, of course, we allow action at
a distance. Try reading Mermin's paper again.
Philip Santas
In article <1992Jun4.1...@oracorp.com> da...@oracorp.com (Daryl McCullough) writes:
>hol...@opal.idbsu.edu (Randall Holmes) writes:
>
>>about particles (the actual reality) is restricted in this way.
>>This is hard work; we will make no progress in this direction as
>>long as we continue to take the easy course of treating
>>the waves as an independent physical reality (an attractive
>>explanation which probably should have been avoided).
>
>Once again, I disagree strongly. There is no reason (that I know of)
>to doubt that waves are part of physical reality. The so-called
>"collapse of the wave function" tends to make people think of the
>parallel with ordinary probability: When we flip a coin in the dark,
>we can describe the situation as "Heads, with 50% probability, and
>tails, with 50% probability." However, when we turn on the light,
>the coin seems to "collapse" into a define state, either heads or
>tails. Obviously, in this case, there really is no collapse; the
>coin was either heads or tails before we turned on the light, we
>just didn't know which.
This argument is a bogus. Light seems to be the reason why one cannot say
much about an electron :-)
Harry Erwin
I've given some physicists the willies with this interpretation, and yet
others have told me it's not far off: Consider an electron. In our universe,
you can't have just "an" electron, because it is surrounded by a cloud of
virtual particles, and identifying the real electron from the virtual ones
nearby is impossible. Hence, the wave function actually describes a cloud
of particles. The value of the wave function integrated over a small
region is the expected excess of electrons in that cloud in that region
when you sum _everything_ in the cloud. Hence there is no _real_ electron,
but rather a cloud of virtual particles that has an excess averaging one
electron.
Since the sum we are dealing with is a conditional sum over an infinite
set, we can get other results besides an excess of one electron. If you're
in an accelerated frame, you will see particles that you wouldn't see in a
frame at rest... So the nature of the reference frame may well define your
rules for doing the conditional sum...
Not being a physicist, I'll leave it at that.
Cheers,
--
Harry Erwin
Internet: er...@trwacs.fp.trw.com
Nick Cassimatis
>>intelligently may not be conscious - have qualia etc. etc.
This statement came as a real shocker. To me, someone who plans to do
AI work some day and who thinks about it alot, the central problems of
AI are getting a machine to use and understand language, building a
robot that can learn to use tools to build things, and so on.... I'm
willing to bet quite a bit that these problems won't be solved by
people who expend most of their "mental energies" to solve Searle's
puzzle.
>> Even Searle
>>would agree that it is possible to build a zombie - use a humongous LUT
>>if all else fails.
Assume we have built a "zombie" -- let's say it's something like Data
from Star Trek: TNG, only lets say that it is a bit weaker, that it
lies on occasion, that it goes to great length tos win the affection
of someone of the opposite sex, that it commits many of the sins one
reads about in the Old Testament and so on. Now let's say someone
like Searle comes along and convinces us that while quasi-Data is in
all other ways indistinguishable from humans, he is not conscious, he
doesn't have qualia, and that his (positronic?) "brain" states are not
grounded; all because [insert whatever reasons you wish.] Now let's
assume that we are convinced by this quasi-Searle. So what? How is
Mr./Ms. quasi-Searle's statement that interesting? Should the fact
that we don't use a certain of words in relation to quasi-Data, change
our behavior to it? Let's assume that a latter day Dennett come along
and proves irrefutably that we can use the same "mentalistic" idioms
towards quasi-Data as we would to humans. So now we make a minor
adjustment to our linguistic conventions -- but the fact remains that
the machine we se before us is still understood in the same way as it
was before we changed our linguitic habits: our conception of the
mechanism that is behind this marvel hasn't changed, the jokes we tell
to him don't change, the way we listent to his violin playing doesn't
change and so on. So if the arguments about Qualia and all that had
no appreciable affect either way, why did we get so worked up in the
first place?
Stephen P Spackman
doing anything clever come from? How am I supposed to aduce that *I*
have consciousness in this special sense?
Since I'm unable to confuse myself enough to not understand how I
could be a fully deterministic machine, why should I assume that
something bizarre is going on in me at all? Why should I not be a
Data?
Randall Holmes
>In article <1992Jun5.1...@guinness.idbsu.edu> hol...@opal.idbsu.edu (Randall Holmes) writes:
>
>>Here is an alternate
>>hypothesis. The state of a particle, when it is emitted, consists of
>>a definite yes-no answer for each angle (not a hidden single axis of
>>polarization -- Bell's argument does kill this). Each of these
>>answers is diametrically opposed as between the two electrons. The
>>answers vis-a-vis two angles for a given one of the two electrons are
>>correlated probabilistically in the appropriate degree determined by
>>the angle. This kind of hidden state, decided at the source, not by
>>random events with space-like separation at the detectors, will
>>exhibit the exact behaviour seen in the Mermin experiment and does not
>>involve non-locality.
>
>This simply doesn't work. The hypothesis that the state of each
>particle carries a definite yes-no answer for each angle (or even
>a definite probability for each angle) is inconsistent with the
>experimental results -- unless, of course, we allow action at
>a distance. Try reading Mermin's paper again.
I think it works. The point is that the relationship between the
various yes-or-no answers is not what you expect in a hidden-variables
theory. (This is not a hidden variables theory of the type refuted by
the experiments being discussed). But let's stay off this topic on
this list.
>
>--
>Dave Chalmers (da...@cogsci.indiana.edu)
>Center for Research on Concepts and Cognition, Indiana University.
>"It is not the least charm of a theory that it is refutable."
The Technicolour Throw-up
> Think of questions to the computer : "Computer, correlate positronic
> emission anomalies with unusual Romulan activity in sector 6." which
> require intelligence. Then compare Data's questions, "Jordy. Why do
> you laugh?" which show consciousness. It is interesting that the
> writers have made Data the only one of his kind;
But he isn't the only one of his "kind". Moriarty was conscious.
> the secret of the conscious (?) robot died with his creator.
Except that the Enterprise's computer systems was able to make a conscious
mind following a simple verbal request from Geordi. Sheesh. I mean why
bother employing zillions of federation scientists when a ship's computer
can solve your major research problems. (Can you say continuity error?)
--
Just my two rubber ningis worth.
Name: Michael Chisnall (chis...@cosc.canterbury.ac.nz)
I'm not a .signature virus and nor do I play one on tv.
Wayne Throop
>> cla...@acme.ucf.edu (Thomas Clarke):
>> the secret of the conscious (?) robot died with his creator.
> Except that the Enterprise's computer systems was able to make a conscious
> mind following a simple verbal request from Geordi. Sheesh. I mean why
> bother employing zillions of federation scientists when a ship's computer
> can solve your major research problems. (Can you say continuity error?)
There is another point I noticed. First, it seems that the basic
premise of the show's treatment of AIs is very much akin to Harnad's.
For example, the mind of a dying man, transplanted into Data's
sensorimotor system, produced a living, thinking, TTT-passing entity.
The same programs and data, when transfered to the ship's computer,
resulted only in "squiggles and squoggles". They even showed some
squiggles and squoggles on a display panel, to reinforce the
non-living-ness of the information in the computer.
And then there's the Holodeck, where simulated entities are very wooden
and unlifelike, since after all they are "computer generated".
The problem is, they keep slipping up, and examples abound: Moriarity,
the designer of the enterprise's engine systems, the seductress used to
manipulate Riker (and Picard by accident).
( Interesting that two of these counterexamples involve Geordi simply
asking the computer to simulate a conscious being... maybe Geordi has
some psychic talent he's unaware of, involving the ability to
"ground" the symbols used by computer simulations. After all, when
Riker makes a very similar request of the computer to try to
recreate his seductress, the result is still wooden and unlifelike. )
But how can you expect consistency from a show who's writers think
it is plausible that Data can do all of the things he can do, but
(of all stupid things) can't bring himself to use contractions. Ghak.
( Note that by poking fun at the ST:TNG, I don't mean to
ridicule the notion of grounding, or that there might be a
distinction between a thing and a simulation of a thing. )
Wayne Throop ...!mcnc!aurgate!throop
Drew McDermott
In article <1992Jun8....@cs.ucf.edu>, cla...@acme.ucf.edu (Thomas Clarke) writes:
|> In article <1992Jun5.2...@bronze.ucs.indiana.edu>
|> chal...@bronze.ucs.indiana.edu (David Chalmers) writes:
|> >
|> > track". All of these "observational tracks" will have conscious
|> > observers in them. To be sure, to each of those observers, it will
|> > seem as if a single track has been slected, but that's just an
|> > illusion of perspective.
|> > What does need to be explained in the Everett interpretation is why,
|> > given that the world is a giant superposition, there are observers
|> > that are conscious of seemingly non-superposed states. But that's
|> > not too much of a problem, with a decent theory of consciousness
|> > in hand: such observers are only determined by a limited part of
|> > the wavefunction, namely by information present in one of its
|> > eigenstates; they simply don't have access to information elsewhere
|> > in the wavefunction.
|> >
|> "choosing" which state to observe?
|>
Only if you picture minds as being outside the universe and looking in.
If minds are just physical systems, then the problem of why a mind in a
given branch of the universe is conscious of just that branch is exactly
the same as why a billiard ball collides only with billiard balls in its
branch of the universe.
-- Drew McDermott
Jason Snell
as it doesn't degenerate into a goofy trekker flame war...
thr...@aurs01.UUCP (Wayne Throop) writes...
>> chis...@cosc.canterbury.ac.nz (The Technicolour Throw-up)
>>> cla...@acme.ucf.edu (Thomas Clarke):
>>> the secret of the conscious (?) robot died with his creator.
>> Except that the Enterprise's computer systems was able to make a conscious
>> mind following a simple verbal request from Geordi. Sheesh. I mean why
>> bother employing zillions of federation scientists when a ship's computer
>> can solve your major research problems. (Can you say continuity error?)
I think the idea behind this episode ("Booby Trap") is that Geordi had the
computer create a personality framework through which it could present him the
data, rather than just doing it on screen. Rather that create an "Enterprise
Computer Person" template from scratch, it used the actual engineer's
personality profile.
My view of this scene was not that Geordi was talking with a "living person,"
but that he was talking with the intelligence of the enterprise computer
itself, filtered through the superficial personality traits of a random human
female, Leah Brahms.
So Geordi wasn't given the answers by the computer-person-- he was just using
a more, uh, interactive form of computer assistance.
>And then there's the Holodeck, where simulated entities are very wooden
>and unlifelike, since after all they are "computer generated".
>
>The problem is, they keep slipping up, and examples abound: Moriarity,
>the designer of the enterprise's engine systems, the seductress used to
>manipulate Riker (and Picard by accident).
Moriarty is a stupid slip-up. If it's so easy to create holodeck
personalities, just by asking...
I've countered the enterprise engine designer argument above; as for the
seductress, that was a specially-programmed computer enhancement, designed by
the Binars (binary computer people -- there's another thread. arrgh.) to
corral Riker and, with luck, Picard. It wasn't stumbled upon -- it was a
specially-designed program.
Moriarty is the only character that was a real slip-up.
>But how can you expect consistency from a show who's writers think
>it is plausible that Data can do all of the things he can do, but
>(of all stupid things) can't bring himself to use contractions. Ghak.
No kidding. And don't you think that if a super genius like Dr. Soong could
make such an amazing android, he could also give him correct skin tone? Or
didn't the good doctor get a lot of sun?
--
Jason Snell / jsn...@ucsd.edu / Ashamed owner of a Communication BA!
InterText - A Net Fiction Magazine - ASCII or PostScript - Mail me for info!
"I thought for a second that my monkey had rabies --
It turned out he had just gotten into the Cool Whip." -- Dave Letterman
Nick Cassimatis
>In article <1992Jun7....@news.media.mit.edu> n...@media.mit.edu (Nick
>> I'm
>> willing to bet quite a bit that these problems won't be solved by
>> people who expend most of their "mental energies" to solve Searle's
>> puzzle.
>limits are. We already know time should not be wasted on the halting
>problem.
Yes, but the halting problem is well defined (it's part of math --
which means that it is well-defined by definition!) The rest of my
post was an attempt to show that the putative limits of understanding
and so forth are not well defined.
>Think of questions to the computer : "Computer, correlate positronic
>emission anomalies with unusual Romulan activity in sector 6." which
>require intelligence. Then compare Data's questions, "Jordy. Why do
>you laugh?" which show consciousness. It is interesting that the
>writers have made Data the only one of his kind; the secret of
>the conscious (?) robot died with his creator.
What is so special about a why-question that it requires consciousness?
There is a debate among some biologists concerning whether a virus is
alive or not. My Bio class in high school spent about 30 minutes one
time debating this. Though I wasn't particularly interested in the
begining, I became so when it the following question dawned on me:
"What would it matter either way?" Assumet that you call it alive.
Then so what? Assume that you call it not-alive. Agin, so what?
Assume that we agree now that it is not alive, but then 30 years later
we come up with a formal definition of life that clearly includes the
virus. So what? Will this change the methods we use to combat the
virus, the way we sneeze when infected, the chances of us curing the
common cold .... Not one bit! So the only effect the debate could
possibly have is a change in linguistic convention. Notice that no
self-respecting biologist would suggest that rats have a
life-substance that a virus doesn't -- the way we think about
organisms is sophisticated enough not to need vitalistic terms.
Now think about the debate over whether computers can *really* be
intelligent or *really* understand. How is it different from the
question of wether a virus is alive or not?
Thomas Clarke
Observer O1 sees events {e(t<T)}U{e1(t>=T)}
Observer O2 sees events {e(t<T)}U{e2(t>=T)}
where e1(T) is different from e2(T) (subsequenct events may differ
also).
A physical argment from continuity would say that since
O1=O2 for t<T then O1=O2 at t=T. So as Einstein might have
said "God rolls the dice and lets O1 see e1 and O2 see e2."
I fail to see how this differs in essentials from a single
observer watching "God roll the dice" to determine
the outcome of an observation (that is a
"wavefucntion collapse").
A static view from outside the universe in space and time
would of course reveal a complicated interweaved web of observers
according to many worlds. The discovery of Bell-type inequalities
shows that according to the conventional interpretation, the web
of observers is replaced by a web of correlations between
observational choices. I suspect the two webs are isomorophic and
equally unobservable.
--
Thomas Clarke
> In article <1992Jun8....@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
Clarke) writes:
> >In article <1992Jun7....@news.media.mit.edu> n...@media.mit.edu (Nick
> >Cassimatis) writes:
> >> I'm
> >> willing to bet quite a bit that these problems won't be solved by
> >> people who expend most of their "mental energies" to solve Searle's
> >> puzzle.
> >It seems to me important to establish, if possible, what the fundamental
> >limits are. We already know time should not be wasted on the halting
> >problem.
>
> Yes, but the halting problem is well defined (it's part of math --
> which means that it is well-defined by definition!) The rest of my
> post was an attempt to show that the putative limits of understanding
> and so forth are not well defined.
intelligence. No?
> >Think of questions to the computer : "Computer, correlate positronic
> >emission anomalies with unusual Romulan activity in sector 6." which
> >require intelligence. Then compare Data's questions, "Jordy. Why do
> >you laugh?" which show consciousness. It is interesting that the
> >writers have made Data the only one of his kind; the secret of
> >the conscious (?) robot died with his creator.
>
> What is so special about a why-question that it requires consciousness?
Excerpt from the official TT(T) scoring guide:)
Reiteration of previous statement: 1 point
Change of subject: 2 points
...
Knowlege of larger world 25 points
...
Introspective why question: 100 points
...
> There is a debate among some biologists concerning whether a virus is
> alive or not. My Bio class in high school spent about 30 minutes one
> time debating this. Though I wasn't particularly interested in the
> begining, I became so when it the following question dawned on me:
> "What would it matter either way?" <deleted> Notice that no
> self-respecting biologist would suggest that rats have a
> life-substance that a virus doesn't -- the way we think about
> organisms is sophisticated enough not to need vitalistic terms.
>
> Now think about the debate over whether computers can *really* be
> intelligent or *really* understand. How is it different from the
> question of wether a virus is alive or not?
I think biochemists can reproduce a virus from basic chemicals right
now. Make up a strand of DNA/RNA with right sequence, make up some
proteins for the coat (this might actually be harder), put all in
test tube and shake. Voila! Virus particles.
No way can they yet make a functioning cell. I'll bet that they'll
never make one by puting chemicals in a test tube and shaking!
This is one difference between viruses and "living" cells. I see
a similar difference between intelligence and conciousness.
Functions of life (reproductive structure) are "easy",
life itself is hard. In the same way conscious functions (chess
playing, specific disease diagnosis) are proving doable, but
consciousness, thinking itself, is much harder.
I'll bet the techniques used in computer chess and expert systems
will not play a roll in achieving artificial conciousness.
Drew McDermott
In article <1992Jun10.1...@cs.ucf.edu>, cla...@acme.ucf.edu (Thomas Clarke) writes:
|> In article <1992Jun9.2...@cs.yale.edu> mcdermo...@CS.YALE.EDU (Drew
|> McDermott) writes:
|> >
|> > In article <1992Jun8....@cs.ucf.edu>, cla...@acme.ucf.edu (Thomas
|> Clarke) writes:
|> > |> >
|> > |> Isn't being conscious of non-superimposed states isomorphic to
|> > |> "choosing" which state to observe?
|> > |>
|> >
|> > Only if you picture minds as being outside the universe and looking in.
|> > If minds are just physical systems, then the problem of why a mind in a
|> > given branch of the universe is conscious of just that branch is exactly
|> > the same as why a billiard ball collides only with billiard balls in its
|> > branch of the universe.
|>
|> Observer O2 sees events {e(t<T)}U{e2(t>=T)}
|> where e1(T) is different from e2(T) (subsequenct events may differ
|> also).
|> A physical argment from continuity would say that since
|> O1=O2 for t<T then O1=O2 at t=T. So as Einstein might have
|> said "God rolls the dice and lets O1 see e1 and O2 see e2."
|>
|> I fail to see how this differs in essentials from a single
|> observer watching "God roll the dice" to determine
|> the outcome of an observation (that is a
You still have the wrong picture, in which O1 and O2 are floating outside
the universe, and God is assigning them different branches to observe.
The right picture is that there is a physical system O that observes its
universe, and at time T it splits into O1 and O2. O1 and O2 are *defined*
in terms of which branch they're in, so there is nothing arbitrary about
which branch each observes. (Of course, the reality is that the universe
is splitting into uncountably many copies constantly, but these binary
pictures are simpler to visualize.)
-- Drew McDermott
Thomas Clarke
somewhat arbitrary and ad hoc? Why does O split at time T?
(at every time or at any time for that matter)
Many worlds is a consistent interpretation of quantum mechanics
in that it produces exactly the same observations and predictions
as the standard model (otherwise it would probably be wrong since it
would violate some experiment), and because of that I still maintain
many worlds is isomorphic to the standard interpretation.
Randall Holmes
>in that it produces exactly the same observations and predictions
>as the standard model (otherwise it would probably be wrong since it
>would violate some experiment), and because of that I still maintain
>many worlds is isomorphic to the standard interpretation.
>--
>Thomas Clarke
>Institute for Simulation and Training, University of Central FL
>12424 Research Parkway, Suite 300, Orlando, FL 32826
>(407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
Its physical predictions are the same as those of the usual
interpretation; thus the characteristic features of the usual
interpretation (collapsing wave functions, etc.) are _not part of the
physics_.
Some comments on the way that "many-worlds" looks to me:
a. The universe in the "many-worlds" interpretation actually has a
single state, which never collapses; we are not really in a "garden of
forking paths" with massive proliferation of universes. The
particular state which we observe locally is not the whole thing.
b. Only certain states of an observer are "suitable" as subjective
states (identified with certain physical eigenstates); superpositions
of these "suitable" subjective states are interpreted as "parallel
experiences of versions of the observer in different worlds (again,
identified with physical phenomena)". This is no more problematic
than the idea that a spin 1/2 particle has either "spin up" or "spin
down" as eigenstates, and other states are a superposition of these
basic states, and the interpretation of this situation with particles
is handled the same way in "many-worlds" (except that the particle
situation is much simpler). [there _is_ a problem for some members of
the audience, which is that I am flatly identifying subjective states
with concrete physical states, but this is a philosophical position
"orthogonal" to questions about QM]
c. A "suitable" state of an observer can evolve through interactions
into a superposition of "suitable" states; this corresponds to a
bifurcation in the "garden of forking paths" metaphor. But we are
_not_ in the garden of forking paths; don't lose sight of the overall
state of the universe. Further bifurcations will _not_ proliferate
endlessly; they will start to "interfere" with one another.
_Locally_, as long as we are talking about relatively few decisions,
the number of separate future worlds for an observer seems to increase
exponentially; interference within the overall world state prevents
this from actually happening.
Randall Holmes
On reflection, I suppose that the number of branches _could_
proliferate indefinitely, with the probabilities of individual
branches getting very small, but all of the branches would still be
embedded in the original world state, which is in this case perhaps
getting more "chaotic" all the time. But all these branches are still
embedded in the common world state; there is not the exponential
expansion of needed "resources" which appears to arise if we think of
the world (the particular one we suppose that we are actually
"subjectively" observing) as reduplicating itself continually.
bay...@force.ssd.lmsc.lockheed.com
>thr...@aurs01.UUCP (Wayne Throop) writes...
[lots of stuff deleted]
>>But how can you expect consistency from a show who's writers think
>>it is plausible that Data can do all of the things he can do, but
>>(of all stupid things) can't bring himself to use contractions. Ghak.
It gets even better. While re-watching a 1st or 2nd season episode
I caught Data using a contraction, either immediately after or WHILE DENYING
THAT HE COULD. It is, of course, meant to be an enderring little reminder
of our mechanical friend's limitations. Actually, I thought the point was
better (& more plausably) made when Dr Crusher was teaching Data to dance.
He could do a complicated precision tapdance routine with ease, but
couldn't figure out the more loosely structured slow dance.
Thomas Clarke
(Randall Holmes) writes:
> In article <1992Jun10....@guinness.idbsu.edu> hol...@opal.idbsu.edu
(Randall Holmes) writes:
> >In article <1992Jun10.1...@cs.ucf.edu> cla...@acme.ucf.edu (Thomas
Clarke) writes:
> >[...]
> >>
> >>Many worlds is a consistent interpretation of quantum mechanics
> >>in that it produces exactly the same observations and predictions
> >>as the standard model (otherwise it would probably be wrong since it
> >>would violate some experiment), and because of that I still maintain
> >>many worlds is isomorphic to the standard interpretation.
> >Its physical predictions are the same as those of the usual
> >interpretation; thus the characteristic features of the usual
> >interpretation (collapsing wave functions, etc.) are _not part of the
> >physics_.
> >
> On reflection, I suppose that the number of branches _could_
> proliferate indefinitely, with the probabilities of individual
> branches getting very small, but all of the branches would still be
> embedded in the original world state, which is in this case perhaps
> getting more "chaotic" all the time.
I think enough debate about interpretation of QM for now. We're
not going to settle a debate that has been raging for sixty years
this month in ai.philosophy.
Very tempting to try for a last word, but I think I'll mull over
my ideas for a try a more coherent and expository posting later
(maybe even a real paper :-).
Robert W Murphree
>What is so special about a why-question that it requires consciousness?
>There is a debate among some biologists concerning whether a virus is
>alive or not. My Bio class in high school spent about 30 minutes one
>time debating this. Though I wasn't particularly interested in the
>begining, I became so when it the following question dawned on me:
>"What would it matter either way?" Assumet that you call it alive.
>Then so what? Assume that you call it not-alive. Agin, so what?
>Assume that we agree now that it is not alive, but then 30 years later
>we come up with a formal definition of life that clearly includes the
>virus. So what? Will this change the methods we use to combat the
>virus, the way we sneeze when infected, the chances of us curing the
>common cold .... Not one bit! So the only effect the debate could
>possibly have is a change in linguistic convention. Notice that no
>self-respecting biologist would suggest that rats have a
>life-substance that a virus doesn't -- the way we think about
>organisms is sophisticated enough not to need vitalistic terms.
I've been reading a book by lynn marguluis the biologist called
five kingdoms. In it she adresses the question of whether viruses
are alive or not. Apparently viruses are considered "part" of
the organism they infect. The tobacco mosaic virus is much more
closely related to the tabacco plant than to animal viruses.
The polio virus is much more closely related to people than to
other types of virus's. So If your going to be a good reductionist
you need to start with bacteria, not viruses, according to this book.
While I wouldn't want to treat a bacteria as a sentient being, I think
it safe to say that it is not, at least to present science, an organism
without choices and probably not presently modelable as a deterministic
system. Now most biologist know this and are still staunch reductionists.
But the reduce everything to a virus=chemical argument doesn't work for
the reasons I"ve advanced.
>Now think about the debate over whether computers can *really* be
>intelligent or *really* understand. How is it different from the
>question of wether a virus is alive or not?
If the virus-other species system is sufficiently complicated to defy
simulation by symbol manipulating devices, which maybe it is, then maybe
understanding is also something which requires something more than
symbol manipulation-being as such or whatever.
Daryl McCullough
>Observer O1 sees events {e(t<T)}U{e1(t>=T)}
>Observer O2 sees events {e(t<T)}U{e2(t>=T)}
>where e1(T) is different from e2(T) (subsequenct events may differ
>also).
>A physical argment from continuity would say that since
>O1=O2 for t<T then O1=O2 at t=T. So as Einstein might have
>said "God rolls the dice and lets O1 see e1 and O2 see e2."
>
>I fail to see how this differs in essentials from a single
>observer watching "God roll the dice" to determine
>the outcome of an observation (that is a
>"wavefucntion collapse").
In practice, there is no testable difference between the Many-Worlds
view and the collapse view. However, in the details there is a big
difference. With the standard notion that the wave function is
collapsed by observation, the collapse is objective, and everyone is
affected by the collapse (in that certain outcomes that were once
possible become no longer possible). In the Many-Worlds view, the only
collapsing that happens is in the subjective world of a single
observer. Therefore, the wave function can be "collapsed" relative to
some observers, and "uncollapsed" relative to others.
While MWI and the collapse interpretation make roughly the same
predictions, they are not, strictly speaking identical. The MWI
predicts that there could be interference effects between different
states of a human brain, while the collapse hypothesis says that such
effects are impossible. This is not really a testable difference,
since interference between different states of macroscopic objects is
in practice impossible to detect.
Daryl McCullough
ORA Corp.
Ithaca, NY
Drew McDermott
|>
|> Don't you find the split at time T at least a little strange,
|> somewhat arbitrary and ad hoc? Why does O split at time T?
|> (at every time or at any time for that matter)
As Everett explains, in the many-worlds interpretation observation is
just a special case of correlation, in which two systems formerly
described by separate wave functions interact in such a way that only
one wave function will suffice to describe them. (Or, to be more precise
(I hope) that their joint wave function is no longer describable as the
Cartesian product of the wave functions for the two original systems.)
E.g., put an electron and a proton into a box (separately). Eventually
you will find a hydrogen atom in the box. The two particles have
"observed" each other. The location of the hydrogen atom is itself given
by a wave function, but the positions of the electron and proton are no longer
independent. Their wave functions have "collapsed" in a sense, although
each particle could still be anywhere.
|> Many worlds is a consistent interpretation of quantum mechanics
|> in that it produces exactly the same observations and predictions
|> as the standard model (otherwise it would probably be wrong since it
|> would violate some experiment), and because of that I still maintain
|> many worlds is isomorphic to the standard interpretation.
I question whether there is a coherent standard model. As I understand it,
the standard model asks us to take the notions of "macroscopic object"
and "measuring instrument" as given (or even take the notion of "conscious
observer" as given, God help us), and states the laws governing the
behavior of elementary particles in those terms. Such a theory cannot
in principle explain how a universe containing no macroscopic objects
or conscious observers came to contain such things. The many-worlds
interpretation works just fine.
If you want to put this in terms of predictions, how about this: The
standard interpretation predicts that if the universe had begun as a
gigantic wave function (which presumably it did), it never would have
collapsed, because it never would have encountered a measuring device.
This prediction has been refuted by observation.
Actually, I'm being sophistical. The standard interpretation predicts
exactly what the many-worlds interpretation predicts: that this gigantic
uncollapsed wave function will come to consist of an unimaginable number of
projections of alternative worlds containing different ways the primordial
universe could have evolved. What's striking about the many-world
interpretation is that it simply takes QM at face value. It seems as
if the hard part for opponents is to explain why it isn't obviously true.
-- Drew McDermott
Thomas Clarke
>
> In article <1992Jun10.1...@cs.ucf.edu>, cla...@acme.ucf.edu (Thomas
> Clarke) writes [in connection with the many-worlds interpretation of QM]:
> |>
> |> Don't you find the split at time T at least a little strange,
> |> somewhat arbitrary and ad hoc? Why does O split at time T?
> |> (at every time or at any time for that matter)
>
> As Everett explains, in the many-worlds interpretation observation is
> just a special case of correlation, in which two systems formerly
> described by separate wave functions interact in such a way that only
> one wave function will suffice to describe them. (Or, to be more precise
> (I hope) that their joint wave function is no longer describable as the
> Cartesian product of the wave functions for the two original systems.)
> E.g., put an electron and a proton into a box (separately). Eventually
> you will find a hydrogen atom in the box. The two particles have
> "observed" each other. The location of the hydrogen atom is itself given
> by a wave function, but the positions of the electron and proton are no
longer
> independent. Their wave functions have "collapsed" in a sense, although
> each particle could still be anywhere.
You get a photon or two to carry off the energy/momentum in creating
the hydrogen atom. Not sure what that has to do with philosophy of
many worlds, butdoes complicate the picture somewhat.
> |> Many worlds is a consistent interpretation of quantum mechanics
> |> in that it produces exactly the same observations and predictions
> |> as the standard model (otherwise it would probably be wrong since it
> |> would violate some experiment), and because of that I still maintain
> |> many worlds is isomorphic to the standard interpretation.
>
> I question whether there is a coherent standard model. As I understand it,
> the standard model asks us to take the notions of "macroscopic object"
> and "measuring instrument" as given (or even take the notion of "conscious
> observer" as given, God help us), and states the laws governing the
> behavior of elementary particles in those terms. Such a theory cannot
> in principle explain how a universe containing no macroscopic objects
> or conscious observers came to contain such things. The many-worlds
> interpretation works just fine.
Probably the closed thing to a standard model is the wavicle
picture most people formulate :-)
> If you want to put this in terms of predictions, how about this: The
> standard interpretation predicts that if the universe had begun as a
> gigantic wave function (which presumably it did), it never would have
> collapsed, because it never would have encountered a measuring device.
> This prediction has been refuted by observation.
Has it been refuted? Wheeler, that old radical, says something like that
we live in an observer created universe. Don't tell the fundamentalists,
but you've just proved the existence of God :-)
> Actually, I'm being sophistical. The standard interpretation predicts
> exactly what the many-worlds interpretation predicts: that this gigantic
> uncollapsed wave function will come to consist of an unimaginable number of
> projections of alternative worlds containing different ways the primordial
> universe could have evolved. What's striking about the many-world
> interpretation is that it simply takes QM at face value. It seems as
> if the hard part for opponents is to explain why it isn't obviously true.
>
> -- Drew McDermott
I'm not opposed to many worlds, in fact I rather like it. I just think
it doesn't get rid of the observer problem. It just gives another
way of looking at what is essentially the same problem. How do
you form a coherent consciousness out of the superposition of
every possible quantum event?
Doug Dawson
>thr...@aurs01.UUCP (Wayne Throop) writes...
[lots of stuff deleted]
>But how can you expect consistency from a show who's writers think
>it is plausible that Data can do all of the things he can do, but
>(of all stupid things) can't bring himself to use contractions. Ghak.
Well, this particulary inconsistency was explained, sortof. Data is the
second android produced by Dr. Soong...the earlier model, Lore, was in
fact able to use contractions and in general emulate humanity. He was
so human in fact, that the other colonists requested that Dr. Soong make
him less human. Dr. Soong responded by producing Data, complete with
limitations to prevent him from acting too human.
Geeze, I haven't even watched the show for a year, but I still remember
all this stuff from 1st or 2nd season...
Nick Cassimatis
>> >> I'm
>> >> willing to bet quite a bit that these problems won't be solved by
>> >> people who expend most of their "mental energies" to solve Searle's
>> >> puzzle.
>> >It seems to me important to establish, if possible, what the fundamental
>> >limits are. We already know time should not be wasted on the halting
>> >problem.
>>
>> Yes, but the halting problem is well defined (it's part of math --
>> which means that it is well-defined by definition!) The rest of my
>> post was an attempt to show that the putative limits of understanding
>> and so forth are not well defined.
>
>Let's define them, if possible! A good job for collective net
>intelligence. No?
Yes, I think that this would be a wonderful task. But most of the
discussion here doesn't even bother to do so. (I've even seen "I
won't play the definition game.") There is a criticism of AI that
goes something like the following: "Define Intelligence. I don't have
a definition. So how can you work on something that is not
well-defined?" The response to this is that we are trying to get
comuters to use language, solve problems and all sorts of other things
that are not hard to define. Now if there were a defintion of
intelligence that included a procedure for solving the halting problem
(or any other unsolvable problem.) Then we would have set a genuine
limit on a research program called artifical intelligence. But this
isn't really that interesting considering nobody was working on the
halting problem in the first place (since we new it was unsolvable
before we got our definition for intelligence.)
Now lets turn this on Artifical Consciousness. Since we don't really
have a good definition of consciousness, we can only hope do build
machines that exhibiti what we would say required consciousness for a
human. If we did settle on a definition, then we could see if there
were any mathematically imposed limits. But, I doubt we could get a
good definiton of consciousness untill we had more precise terms about
cognition. Beforethen, all we can do is mimick "conscious behavior".
Since I have yet to have been given a convincing and precise defintion
of understanding, intelligence, consciousness, etc., I (like the good
Math major that I am) won't let Searle's (or anyone elses) proofs
alter my activites.
How can people let a priori arguments from un- or ill- defined terms
changed their behavior?
>> Now think about the debate over whether computers can *really* be
>> intelligent or *really* understand. How is it different from the
>> question of wether a virus is alive or not?
>
>I think biochemists can reproduce a virus from basic chemicals right
>now. Make up a strand of DNA/RNA with right sequence, make up some
>proteins for the coat (this might actually be harder), put all in
>test tube and shake. Voila! Virus particles.
>
>No way can they yet make a functioning cell. I'll bet that they'll
>never make one by puting chemicals in a test tube and shaking!
It probably won't be so simple, but why do you think that they'll
never be able to make up a cell? Is there any principled reason to
think that this won't be possible some day? Even if we couldn't
reproduce a cell molecule by molucule, why couldn't we build something
that did the same sorts of things that cells do?
>This is one difference between viruses and "living" cells. I see
>a similar difference between intelligence and conciousness.
The difference is only one of complexity.
>Functions of life (reproductive structure) are "easy",
>life itself is hard. In the same way conscious functions (chess
>playing, specific disease diagnosis) are proving doable, but
>consciousness, thinking itself, is much harder.
>
>I'll bet the techniques used in computer chess and expert systems
>will not play a roll in achieving artificial conciousness.
Even if this is so, so what? To say that AI or AC won't be achieved
by the resent tools alone (which I'm pretty certain of) is not to say
that the whole research program is bancrupt. Galileo didn't have
calculus, but this didn't mean that he was wrong for aiming at a
mathematical model of mechanis (assuming that he was.)
-Nick
Bill Skaggs
cla...@acme.ucf.edu (Thomas Clarke) writes:
>> [ . . . ] The rest of my
>> post was an attempt to show that the putative limits of understanding
>> and so forth are not well defined.
>
>Let's define them, if possible! A good job for collective net
>intelligence. No?
Yes, I think that this would be a wonderful task. But most of the
discussion here doesn't even bother to do so. (I've even seen "I
won't play the definition game.") There is a criticism of AI that
goes something like the following: "Define Intelligence. I don't have
a definition. So how can you work on something that is not
well-defined?" The response to this is that we are trying to get
comuters to use language, solve problems and all sorts of other things
that are not hard to define.
Definitions are a bit like laws. There are two kinds of law,
prescriptive (like the "law against committing murder") and
descriptive (like the "law of gravity"). Sophisticated people like us
understand the difference and do not confuse the two.
There are also prescriptive and descriptive *definitions*. A
prescriptive definition is a set of instructions on how a word is to
be used. Hard scientists, especially mathematicians, like to begin
with prescriptive definitions of their terms, because then the meaning
of what they say is perfectly clear.
Descriptive definitions are descriptions of how words are used in
practice. Dictionaries (like Webster's) consist of descriptive
definitions.
When we are thinking about something like "intelligence", we cannot
begin with a prescriptive definition, because we are trying to
understand the meaning of the word as it is generally used by people.
To prescribe a new meaning is simply to duck the problem. This is
what Jeff Dalton meant when he said "I won't play the definition
game". It is quite reasonable, though --- and even necessary -- to
look for a descriptive definition in a situation like this; otherwise
we can never be sure we're all talking about the same thing.
To sum up, I believe the argument here is between people who are
thinking about two different kinds of definition.
-- Bill
Nick Cassimatis
>
>When we are thinking about something like "intelligence", we cannot
>begin with a prescriptive definition, because we are trying to
>understand the meaning of the word as it is generally used by people.
>To prescribe a new meaning is simply to duck the problem. This is
>what Jeff Dalton meant when he said "I won't play the definition
>game". It is quite reasonable, though --- and even necessary -- to
>look for a descriptive definition in a situation like this; otherwise
>we can never be sure we're all talking about the same thing.
>
>To sum up, I believe the argument here is between people who are
>thinking about two different kinds of definition.
>
> -- Bill
What do you [plural] consider a satisfactory "descriptive" definition?
Searle (to my knowledge) hasn't given one to consciousness and
intelligence, Dennett rerfuses to give one with an "if and only if" in
it, and the discussion here hasn't done so either. The possiblility
that we could clearly explicate our notions of intelligence and
consciousness as they are now is clearly in question. I don't think
that those notions as they stand are coherent.
So if you [plural, as always] want to establish some "limits" on AI or
"Artifical Consciousness" or Artifical Life for that matter, give us a
definition and call it whatever kind of definition you want. Until
then, researchers in these fields should just get to work making
machines that can speak and plan and so forth.
The Chinese room doesn't even attempt to show that machine translation
is possible (in fact, it assumes that it is.) It only tries to show
that "understanding" is impossible. There would have been arguments
against robotics in the early 19th century that would have said, even
if you can make machines that look like humans, that talk like them,
that behave like them, and soi forth, you can't make a machine that
has "life". Vitalism such as that is out of fashion now. But it is
not really that different from the "intellectualism" of Searle and all
them. If you use some wishy-washy notion such as "life" or
"intelligence", and then show that artificial life and/or intelligence
are impossible, then you haven't really done anything but proliferate
confusion and bastardize the language.
Brad Vender
is unimportant. In order to perceive the state of a particle,
a device is required to either interact with or act upon the particle
to produce a response (e.g. particle detectors). The consciousness of
the scientist observing the results plays no part in it (unless someone
wishes to put consciousness in the measuring device :) ).
My other objection to the inclusion of this thread here is
If an object's state is not observable, it is of no importance.
(This I learned in High School science classes)
Obviously since we are physical objects, this instance of us represents
a collapsed wave function of some sort (or perhaps we're just inside the
wave function at some state or other).
In any event, the interactions required to observe the particle's state
are what collapse the wave function of the particle, not the intelligence
of the observer.
--ven...@plains.nodak.edu (who may or may not be human)
Kristoffer Eriksson
> But all these branches are still
>embedded in the common world state; there is not the exponential
>expansion of needed "resources" which appears to arise if we think of
It seems to me, that your view of the many-worlds interpretation is
fairly close to an interpretation where there is just one world, and
where the wave functions never actualy collaps - they just appear that
way on measurements. Which is an interpretation that I wouldn't find
unbelievable (but I'm certainly no expert), but perhaps would not
call "the many-worlds interpretation".
--
Kristoffer Eriksson, Peridot Konsult AB, Hagagatan 6, S-703 40 Oerebro, Sweden
Phone: +46 19-13 03 60 ! e-mail: s...@pkmab.se
Fax: +46 19-11 51 03 ! or ...!{uunet,mcsun}!mail.swip.net!kullmar!pkmab!ske
Randall Holmes
>In article <1992Jun10....@guinness.idbsu.edu> hol...@opal.idbsu.edu (Randall Holmes) writes:
>> But all these branches are still
>>embedded in the common world state; there is not the exponential
>>expansion of needed "resources" which appears to arise if we think of
>
>It seems to me, that your view of the many-worlds interpretation is
>fairly close to an interpretation where there is just one world, and
>where the wave functions never actualy collaps - they just appear that
>way on measurements. Which is an interpretation that I wouldn't find
>unbelievable (but I'm certainly no expert), but perhaps would not
>call "the many-worlds interpretation".
As far as I can tell, that is exactly what the many-worlds
interpretation is, mathematically. The local effect on observers
inspires the unfortunate idea that there is a wild proliferation of
worlds going on.
>
>--
>Kristoffer Eriksson, Peridot Konsult AB, Hagagatan 6, S-703 40 Oerebro, Sweden
>Phone: +46 19-13 03 60 ! e-mail: s...@pkmab.se
>Fax: +46 19-11 51 03 ! or ...!{uunet,mcsun}!mail.swip.net!kullmar!pkmab!ske
Stanley Friesen
|n...@media.mit.edu (Nick Cassimatis) writes:
|If the virus-other species system is sufficiently complicated to defy
|simulation by symbol manipulating devices, which maybe it is, then maybe
|understanding is also something which requires something more than
|symbol manipulation-being as such or whatever.
It is certainly beyond the capabilities of currently available computational
devices, but is this an intrinsic limitation, or just a current engineering
limit? The answer is not clear. (There is *also* the fact that we do not
know enough about the virus-cell system to fully specify the parameters of
the simulation).
--
sar...@teradata.com (formerly tdatirv!sarima)
or
Stanley...@ElSegundoCA.ncr.com
Robert W Murphree
>|simulation by symbol manipulating devices, which maybe it is, then maybe
>|understanding is also something which requires something more than
>|symbol manipulation-being as such or whatever.
>It is certainly beyond the capabilities of currently available computational
>devices, but is this an intrinsic limitation, or just a current engineering
>limit? The answer is not clear. (There is *also* the fact that we do not
>know enough about the virus-cell system to fully specify the parameters of
>the simulation).
When I think the the defintion of "state" in control theory (EE control
theory). I think of an ideal system which, to the degree to which you
specify the parameters of the system ( R's and L'c and C's , masses and
distances, or whatever) you can specify the future state of the sytem,
given the initial conditions to any desired accuracy (see Karl Popper's
definition of experimental determinimism). I just cannot think of how
one would do this with living things, even a bacteria.
If the "state"
is living versus dead, it seems no matter how carefully you specify the
situation, you can almost always wriggle out of any given prediction of
this state. Physicians make their money trying to predict and effect
this living or dead state and as an ex-medical technician, let me tell
you its hard to do.
I suppose the current work to predict the tertiaru
teriary structure of proteins from primary sequence structure is an
example of how this sort of attempt to find a "state" is going. I hope
a molecular biologist will read this and comment. My reading of the
work in the problem of predicting tertiary protein structure is that
the process of how proteins fold is so complicated that describing them
is more like geology (trying to find a few generalizations from vast
amount of knowledge-a little pattern recognition)than physics or
computer science. I know very little of the current science in this
area so watch out.
Nick Cassimatis
Nick Cassimatis
KNOW WHAT HAPPEND -- HERE IS WHAT I MEANT TO SEND.
In article <1992Jun20....@constellation.ecn.uoknor.edu> rwmu...@uokmax.ecn.uoknor.edu (Robert W Murphree) writes:
>>|If the virus-other species system is sufficiently complicated to defy
>>|simulation by symbol manipulating devices, which maybe it is, then maybe
>>|understanding is also something which requires something more than
>>|symbol manipulation-being as such or whatever.
>
>>It is certainly beyond the capabilities of currently available computational
>>devices, but is this an intrinsic limitation, or just a current engineering
>>limit? The answer is not clear. (There is *also* the fact that we do not
>>know enough about the virus-cell system to fully specify the parameters of
>>the simulation).
>
If the answer is not clear and there are constant examples from the
history of science of seemingly impossible things being done, then why
think that it is not possible? Now I'm not sure what is meant by
symbol manipulation, but if the manipulation is powerful enough to
simulate a Turing Machine, then there a whole lot of computable
functions that have not been dreamed of.
Whether we can ever have computers that use language and organize
their behavior on the basis of scene recognition and plan their motion
and so on -- this is an empirical question. Searle doesn't seem to
doubt the impossibility of this, but that IF we got computers to
behave thus, they would lack "understanding" or "intelligence" or
whatever. Now boggling over what this could mean and if it is true is
a waste of time -- because it doesn't matter either way. The thing
would work the way it does and a verbal hernia isn't going to change
anything.
>definition of experimental determinimism). I just cannot think of how
>one would do this with living things, even a bacteria.
With all due respect, this does not mean much. The history of science
is riddled with "I just cannot think...." statements that someone has
thought out. NOw I'm not a bundle of ideas on this subject on myself,
but I am trying and I see no reason to doubt significant progress in
my lifetime.
Here's another little thought experiment: imagine that you are borned
into a society in which the computers have run themselves long before
anyone can remember, there is nothing in the culture to make one think
otherwise. Now imagine looking at the marvel of some computer
animated version of Swan Lake. It would be as natural to think "I
can't think of how you would model that computer. What it does is so
beautiful and complex." Since the computer is something as given in
nature as a rock or a tree OR A CELL in this society, this would be no
different from looking at a cell and boggling at the task of
describing its state. OF course, the computer can be modelled easily.
WHy not the cell, or the brain?
Another caveat on the whole modelling quantum effects contreversy:
since one mu st take into account quantum effects to design a
transister today, one would thi nk that one would need to understand
these to model the comptuer. BUT WE DON't. The computer can
completely be modelled as a turing machine. Of course, thiswould
explain what happens when you hit it with gamma rays, but that's not
the s ort of understanding we want. Why should it be any different
with the brain --even if we must take into account quantum effects to
understand the design of th e neuron?
-Nick
Thomas Clarke
Cassimatis) writes:
> Another caveat on the whole modelling quantum effects contreversy:
> since one mu st take into account quantum effects to design a
> transister today, one would thi nk that one would need to understand
> these to model the comptuer. BUT WE DON't. The computer can
> completely be modelled as a turing machine. Of course, thiswould
> explain what happens when you hit it with gamma rays, but that's not
> the s ort of understanding we want. Why should it be any different
> with the brain --even if we must take into account quantum effects to
> understand the design of th e neuron?
>
> -Nick
To my way of thinking, computers are not modelled as Turing machines,
computers are designed and built to BE Turing machines. If any
non-deterministic effects sneak in, through race conditions leading
to metastable states in flip-flops, or whatever, this is considered
a design defect and is rapidly fixed.
Neurons are not nearly so well understood as transistors. While
transistors in computers are deliberately operated in regimes where
quantum effects are unimportant on the scale of the computer as a
whole, the modes in which real neurons operate in real brains are not
so well understood. It is of course entirely possible that evolution
produced brains which operate entirely in deterministic modes,
readily modelled by Turing machines, but I have my doubts.
In fact see the July '92 Scientific American for an article about G
proteins which communicate between and change the states of enzymes
and receptors bound to cellular membranes. It seems entirely likely
that a universal Turing membrane could be constructed from these
components. The components being molecular in size, quantum
effects would be hard to avoid.
Kind of boggles the mind, the brain potentially being an array of
10^10 (quantum) Turing machines :-)
Harry Erwin
variants of tubulin and speculate they can build a Turing Machine using
microtubule/MT-associated protein networks. They're motivated in this by
the evidence for learning in Paramecia, which seems to be associated with
the microtubule cytoskeleton.
So your speculation may not be that far away from the truth.
--
Harry Erwin
Internet: er...@trwacs.fp.trw.com
El Gordo
>> Kind of boggles the mind, the brain potentially being an array of
>> 10^10 (quantum) Turing machines :-)
>> --
>> Thomas Clarke
>> Institute for Simulation and Training, University of Central FL
>> 12424 Research Parkway, Suite 300, Orlando, FL 32826
>> (407)658-5030, FAX: (407)658-5059, cla...@acme.ucf.edu
So much for the brain: how goes it with the mind?
____
Gordon Joly +44 71 387 7050 ext 3703
Internet: G.J...@cs.ucl.ac.uk UUCP: ...!{uunet,uknet}!ucl-cs!G.Joly
Computer Science, University College London, Gower Street, LONDON WC1E 6BT
Matthew P Wiener
>Judy Dayhoff and Stuart Hameroff have been looking at the configurational
>variants of tubulin and speculate they can build a Turing Machine using
>microtubule/MT-associated protein networks.
Last Dec/Jan, PNAS had articles on using enzyme oscillations to encode
neural nets, Turing machines, etc.
--
-Matthew P Wiener (wee...@sagi.wistar.upenn.edu)
Thomas Clarke
> >> [...]
> >> Kind of boggles the mind, the brain potentially being an array of
> >> 10^10 (quantum) Turing machines :-)
> >> --
>
> ____
>
> Gordon Joly +44 71 387 7050 ext 3703
> Internet: G.J...@cs.ucl.ac.uk UUCP: ...!{uunet,uknet}!ucl-cs!G.Joly
> Computer Science, University College London, Gower Street, LONDON WC1E 6BT
It convulses the brain, to compute that the mind might be an
unobservable quantum correlation.