Re: Sentience

From: Dan Fabulich (daniel.fabulich@yale.edu)
Date: Tue Dec 19 2000 - 18:33:38 MST


Steve Nichols wrote:

> >I doubt that as well! But cite von Neumann, Turing and Church for
> >pointing out that a Turing machine can do everything a massively
> >parallel distributed system can do given enough time and tape. We
> >use massively parallel systems today because they are faster than
> >cheaper than doing these jobs serially, not because there are some
> >things that massively parallel systems can do but Turing machines
> >can't.
>
> The reason that neural computers are relevant in discussion of
> similarities to brains and that Turing machines are not is that
> in fact neural (massively distributed parallel) designs *come from*
> modelling and reverse engineering of brains! In no way are brains
> like Turing or Von Neumann architecture.

Brains are like Turing machines in that they are composed entirely of
simple physical parts that do simple physical things. These physical
parts move deterministically in lock-step. (Though, to qualify this
view, I'm a fan of many-worlds.)

Moreover, the physical world is causally closed. Nothing non-physical
interacts with the brain. Your tests establish a physical causal link
between the real pineal eye and a lack of intelligent behavior. As
the eye goes away, assuming you're right about this, intelligent
behavior begins to flourish.

But you did not measure consciousness; you cannot do so.

For the purposes of discussion with you, I'm an epiphenomenalist. To
the extent that we have conscious experiences at all, our experiences
do not cause anything to happen in the world at all. (That is,
everything physical would happen exactly the same whether or not we
had experiences; we'd act the same without them, etc.) Physical
things are correlated with (my) experiences, perhaps even cause our
experiences, and thus our experiences are fully physically determined.
But our experiences in no way affect the real world.

For all I know, the Turing machine DOES feel something when it
emulates me. I'll never know. I can ask it. It may pass the Turing
Test. But I'll still never know. At least at that point I'll know as
well as I know about you or John Clark or anyone else, which is good
enough for me.

> >Yes, there are. Try to think of a REALLY BIG NUMBER. Now think of a
> >number that long with all of its digits randomized. You can imagine
> >it as an opaque term (just like most of us can only imagine a number
> >as large as a trillion) but you can't actually hold it in your mind.
>
> But in inventing and articulating the above example you provide evidence
> to us that you HAVE thought of what you describe. Anyway, you pointed
> out earlier that true randomness doesn't exist, so a machine would have
> the same conceptual problem about that that we do.

No, that's thinking about it as an "opaque term." You can hold those
words in your mind, you can have some mental pictures, some notion of
what it would be like to actually imagine that, but you can't actually
have the whole thing in your mind.

I could go further into opaque terms, but it's a rather dull and
irrelevant philosophical sidetrack.

> >Alternately, try to imagine all of Shakespeare's Hamlet at once.
>
> By "at once" do you mean "instantaneously" or do you mean "a very short
> time, like 2 seconds." ? I can imagine both ... as a very blurry noise and
> flash of actors moving on stage.

I actually had the full text of Hamlet in mind when I said that.

> >Hey. That's sort of like the memory limitations on a computer. Hmm.
>
> I doubt a computer could "imagine" anything at all .......

That is why you fail.

> > Actually, you're right if you count games where the players keep
> > moving their pieces back and forth in a circle. They can do that an
> > arbitrary number of times. But I don't think that's quite what you meant.
>
> Absolutely, it doesn't matter that there is some repetition, I just used
> this to
> prove my point against your original claim about atoms.

This doesn't help your case. Sure, atoms can cycle through their
finite states an infinite number of times, just as a Turing machine
can, or a chess board. But that doesn't imply that there's anything
infinite state about them at all.

> Do you have problems accepting the evidence for phantoms limbs?
> These "non-physical components" have been widely accepted by science
> for hundreds of years, and certainly aren't disputed by those afflicted
> with phantom pain.

Phantom limbs aren't *real*. They're a handy term to explain real
phenomena, but they don't actually exist. It's like when I say things
like "Insects give me the creeps." I don't posit that there are
REALLY "creeps" which insects give me. I just mean that they make me
nervous. Similarly with phantom limbs: there's no *actual* limb
there, phantom or otherwise.

If you like, you can call the phantom limbs non-existent objects
interacting with the physical world. The car crashed on account of a
phantom driver at the wheel, ie no driver at all.

But that doesn't leave you in a good position to explain
consciousness. Just as *existing* physical phenomena provided us with
no good explanation of consciousness, *non-existing* objects do us no
better. Can you prove that it's the case that not-having a pineal eye
makes us *conscious* above and beyond making us intelligent?

> The key point of MVT, and how it DOES explain consciousness, is that
> the phantom limb (or phantom sense-organ) is a *felt* phenomenon.
> The difficulty in measuring such phenomena does not make them less
> real .... but just points to a shortfall in our methods of measuring.

No, it's deeper than that. This still offers us with no explanation
for *how* our consciousness happens. You tell me WHEN (when there's
no pineal eye) but not HOW.

> Only biological organisms can act independently from purely
> chemical stimuli ...so we are talking about a very small subset of
> "atomic structures" ... even more specifically, those animals with
> E-1 brains.

Brains *cannot* act independently of chemical stimuli. They are
purely physical objects, with no more independence from the world than
rocks have. Biological organisms are completely physically
determined.

> My personal (plenumist) view is that atoms are themselves doubtful
> as discrete entities, and I disagree with Heraclitus (?) claim that
> "only atoms and the void are real." I think that a Popperian third world
> of consciousness is independent of physical action, but is not "void."

If you like, we can be epiphenomenalists. My actual views on this are
so fully divorced from your own that I don't think it's even worth
bringing them up at this stage in the discussion.

> (S) Anyway, big blue is a CPU machine, not even a neural computer, so is
> (S) absolutely nothing like a brain. Are you saying the DNA is the
> (S) "program" or what other medium are you identifying?
>
> >The medium is your neurons. None of them can do anything
> >non-physical. So your brain can't do anything non-physical. So your
> >brain is entirely physical. Physical stuff happens in lock-step
> >(because time happens in very small quanta) so the brain is emulatable
> >with a sufficiently complex Turing machine.
>
> Sure, neurons in isolation can't do much, but taken as a matrix
> (maybe even including brain stem and both nervous systems)
> they are able to generate a phantom pineal eye,

How do you "generate" a non-existent entity?

> a complete theory of mind
> should explain how, for example, we can see objects although
> our lateral eyes are closed, or "hear" an internal voice even though
> our mouth is closed and we are not talking out loud.

No. We will find structural behavioral analogs, but we will never
figure out why we see anything at all. We'll be able to figure out
*when* we see something, but not how.

> There is even a question about non-physical quanta effects
> (action at a distance &c.) so even atomsaren't as "solid" and physical
> as you seem to claim ...

Again, I'm a many-worlds fan; I don't think action at a distance
happens.

> let alone if we speculate on the physical or
> non-physicalness at the very centre (the Singularity!) at the middle of
> a black hole).

Certainly true. New physics may lead me to reject my views about
what's physical and what's not. But consciousness will never be on
the list.

> Let me now reject your Turing machine nonsense yet again .....
> (1) Emulation is not the same as the thing being emulated, so even
> if a Turing machine could emulate the functions of a brain, we are
> no further forwards.

Yes, no further than any other physical theory of intelligence.

> (2) The Turing emulation would not "feel" anything, and could not
> dream (for example).

It would *emulate* dreaming. You can't show that this is not the same
thing, that it wouldn't feel like it was dreaming. How would you
know, if it described its dream to you as vividly as anyone else?

> (3) Your account does not get rid of the homunculus (in your deep blue
> case a whole team of homunculi ... human programmers tweaking
> the software). MVT is a fully self-servicing and evolutionary account.

Where's the homunculus? I'm saying that nature has evolved a Turing
machine. This is no more mysterious, I argue, than nature having
evolved any other machine, like a lever/pully system, a hydraulics
system, or dam building.

> I think there is a difference between look-up table "learning" and neural
> learning such as Reinforcement Learning (eg. in world champ. Backgammon
> software "NeuroGammon.") or Learning I.A.C (interactive activation &
> competition .. a distributed database alternative modelled on brains).

There are obvious differences; none of these differences are relevant,
however. (In particular, some of these solutions are more workable
approaches to AI than others.)

> Changing play style is trifling stuff .... and I should
> know, having designed or programmed three Chess-type games (Shogi,
> Chaturanga .. ancestor of common chess, which is a mere human-era
> Chaturanga variant!, and Enochian Chess, which includes an IAC neural
> patch).

Changing play style IS trifling. Doing it well isn't.

> >Improving your game of chess is learning. It's simple learning, but
> >it IS learning.
>
> Maybe, on the most simple definition, but nothing like human-scale.
> Deep Blue is an autistic moron!

Yes, but my point is that it's only a matter of scale.

> But it can't change its programming or deal with new situations,
> its microcosmical "chess-rule world" is an artificial human-made
> game-world to start with. It doesn't think like Kasparov, but uses
> brute-force algorithms suited to von Neumann processing.

It uses a different algorithm from the one Kasparov uses. But
Kasparov uses an algorithm too. (His is certainly better, but it's
hard to read off from the neurons, and ever trickier to program.)

> >Not all PEOPLE are good at that. But nonetheless, an adequately
> >complex Turing program can "deal with" circumstances as new as we can.
>
> Maybe not all are good at it, but even the stupidest human can do it.
> Deep Blue can't. Which Turing machine are you talking about here ...
> one of your (E-1 brain-powered) imaginary ones perhaps?

No, just an ordinary Turing machine built by evolution, like the one
you've got in your head.

> >The same thing happens to us. We don't have free will. Nothing is
> >within our remit.
>
> Plants, and simple E-2 (pineal eyed) animals have no free will.
> There behaviour is governed by the sunlight and seasonal shifts,
> plus chemical taxis &c. However, we E-1 animals have gained
> indepence from external governance of behaviour rom sunlight
> (via the old pineal eye) and *do* have self-volition or free will.

> The term "free will" would not ever have been invented if things were
> otherwise ..... the fact we can discuss it proves it almost.

<blink blink> Can you make that inference a little more explicit?

Certainly you don't mean the following:

(1) We came up with the term "free will."
(2) If we came up with a term, its referent must exist.
(3) Therefore free will exists.

The argument against free will follows directly from the causal
closure of physics. Even if you think that there are truly random
elements in the physical world, that doesn't mean that there's
anything going on in the world that's a *choice*.

Everything is either deterministic or random. Neither of these, nor
any composition of the two, constitute free will.

> >Neither can you. Your programming is more complex, but that's all.
>
> Yes I could ... like Captain Scott ... decide to override my body and
> voluntarily freeze to death!

Your decision to do so would be based on your *desires* to do so. You
can't change those effectively, as I'll discuss.

> >You act according to your desires and beliefs. You cannot change your
> >desires very much, and you can hardly change your desire to desire
> >different things at all. You cannot change your beliefs either.
>
> As an NLP and Hypnotherapy practitioner, of course I reject your
> claim as completely groundless.

Actually, I can accept an arbitrary degree of flexibility at a given
level of desire.

Suppose you can change any of your desires by hypnosis. Nonetheless,
some of your desires are determined externally: by your environment or
genetics.

You're not hypnotizing yourself right this instant. Presumably
because you don't desire to do so. So you have a desire, call it D,
the desire "not to change my own desires via hypnosis."

Naturally, you can't voluntarily hypnotize yourself when you have D.
If D is externally determined (by environment or genetics or whatever)
then YOU can't choose to change your desires via hypnosis at all. You
will simply *refuse* to change your own desires via hypnosis unless
and until the environment changes your mind about D.

Of course, you could change D via hypnosis, if you desired to. But D
includes a desire not to change D via hypnosis. So you can't
voluntarily change D via hypnosis, until your desires are changed by
the external world.

Now, if D was originally set by environment or genetics, and you just
never changed it, then it seems that you can't voluntarily use
hypnosis at all unless your environment changes your mind for you.
But if D has been set by *you*, it was set by you because you desired
to do so, that is, because you had some other desire, D', the desire
"to give myself desire D via hypnosis." But D' is in the same state
as D: it was externally set or set by you. And same with any D'', a
desire to set D'. Or D'''. And so on.

So if any of those D^(n)s are set externally, then your behavior is
externally determined: you just mechanically followed through with
your externally determined desires, even when you changed a few of
them, because even then you were under the control of externally
determined higher-level desires telling you what to do.

And, as it turns out, at least one of those D^(n)s IS externally
determined. You're born with a set of desires determined by genetics
and environment, which you later revise. You were built by something
which wasn't you. So your desires are externally determined.

So all of your behavior is externally determined: you simply follow
orders given by the environment and genetics.

> >You
> >can decide what to say, but you cannot decide, for example, to drop
> >your view about MVT and agree with me. You have to be convinced, and
> >you're not in control of when you get convinced.
>
> I could decide to drop MVT, and would if (1) you could refute it, or
> (2) you outline a more convincing alternative theory that solves more
> questions and I considered to be a better account. Go ahead,
> convince me!

Right. But you can't decide to drop MVT and those criteria. You
desire to keep them, and you can't help it. (Nor should you, in the
case of the criteria.)

> >You're not in control of when you're happy. You do things to make
> >yourself happy because you desire happiness because that's part of the
> >program. When you do things, you have only the slightest control over
> >whether they make you happy or not, and considerably less control over
> >your desire to feel happy. You're not in control of these things.
>
> What are you babbling about here? I don't recognise any of these
> statements in myself. Are you talking from your own, solipsistic viewpoint?

I'm not a solipsist.

> >And what suggests consciousness besides intelligence?
>
> Intelligence PLUS felt experience.

What suggests felt experience besides intelligence?

> I think the onus is on you to make a case why someone who claims
> to have consciousness is in fact a robot.

I concede. I think that ANYONE who plausibly claims to have
consciousness is conscious. Including robots.

> The subjective nature of the phenomena does maybe rule out
> "objective" tests, so indeed some new form of testing might be
> needed.

No form of testing is possible.

> Am more interested in developments from MVT, what I can do and
> predict from it, than in satisfying some sterile academic or
> philosophical debate.

Good! It IS a stupid debate. Why not just say that you've solved
animal intelligence and leave the sterile debate alone?

Realizing that a debate is fruitless is not to win the debate.

> S> The very fact that every aspect of Turing machines actions can be
> S> predicted, I would say, might even preclude them from consciousness.
>
> >No more than you are precluded from consciousness. Your atoms have no
> >free will. You are composed entirely of atoms. You have no more free
> >will than they do. An adequately informed and powerful computer could
> >calculate your every move in a controlled environment.
>
> Again (see above) you are looking at 'me' on the wrong level of
> description ..... why say I am composed of atoms rather than I am
> composed entirely of "cells" or "molecules" or "sub-atomic particles?"
> None of these levels is macro enough to explain consciousness, which
> requires the whole biomatrix, plus even phylogeny and embryonic history,
> to be properly understood. The whole is greater than the parts ... and the
> whole includes phantom body parts.

I can say the same thing back to you, and it would be just as silly.
You're looking at the Turing machine on the wrong level of
description. Why look at it as a punch card reader when you could
consider blocks of code, functional subsegments or bits on the tape?
None of these levels is macro enough to explain consiousness, which
requires the whole program, plus a Turing machine to operate them, to
be properly understood.

This explanation is empty.

> Ah, but you have accepted that at least you, even if nobody else)
> has *dreams* .... which are non-physical ...

<must resist blurting out my real opinion here>

> they are temporal events
> but cannot be detected by crummy human-era science. So just by
> accepting dreams, as you have, then a distinction is made between the
> conscious experience, and the PET or EEG activity of the brain, which
> is detectable to scientists (but is unknown to the dreamer ... you don't
> need
> to know which of your neurons are firing in order to have thoughts).

Sure thing. The brain activity causes the consciousness. But I don't
know how; can't know how. I can only notice THAT it does, and WHEN.
Not why. I can know WHICH brain activity results in which
experiences, but that's not knowing HOW that brain activity results in
experiences.

> Not even to "really big". A component can be scalar as
> opposed to digital over a very narrow spectrum. It is just that it is
> a continuum potential rather than just discrete digital switching
> that leads it to be called analog.

But the continuum isn't real, anymore then there's actually a smooth
curve on the letter O which appears on your computer screen. It's
jagged, pixellated. But it's so close to continuous, so near to
smooth, that you can and do over look it.

> S> No, phasic transients occur simply because a circuit is undergoing
> S> transformation from finite-state (lock step) to self-organising ....
> S> and they happen after the removal of an external clock (whether
> S> electronic or organic pineal eye). Aren't Turing machines always
> S> lock-stepped?
>
> >Why, yes. Along with ATOMS.
>
> There aren't any atoms in a phantom pineal eye, or in a thought ....
> your dreams cannot be observed precisely because they are NOT
> atomic!

Too true. But you just told me that you were using dreams to measure
consciousness. But you can't observe dreams. So you can't measure
consciousness.

> But MVT gives an account of why mammals dream, and furthermore
> throws new light on the melatonin hypothesis, and Menaker's work
> on REM and the evolutionary shift from pineal eye melatonin to
> retinal production. DREAMS are not "intelligence" and do not involve
> 'rational' cognition of processing of sensory information from external
> world ..... so MVT explains consciousness here rather than intelligence.

Nope, you're explaining intelligent behavior. In particular, there IS
dream behavior. Wake someone in the middle of REM and they'll tell
you about the dream that they were having, and what it felt like.
That's dream behavior, just like how smiling and laughing and telling
me how wonderful you're feeling is happiness behavior.

Don't you think that I'd ask any Turing machine trying to pass the
Turing Test what its last dream was like? Answering me just like a
human would answer me is part of passing the Turing Test. Reporting
on your dreams is part of what it is to exhibit intelligent behavior;
it's part of what it is to pass the Turing Test.

You've given me an explanation for why people go into REM and report
to be dreaming after the fact. This may be a great advance in the
theory of animal intelligence. But you have given no explanation of
how or why REM, brain activity or dream reports have anything to do
with actually dreaming. You could have and do all of these without
dreaming.

What I'm trying to convince you of is to settle for intelligence.
This philosophical discussion is geared for you to lose.
Consciousness is defined in such a way that you can't ever give a
scientific account of it, you can't EVER show how mere physical stuff
could be conscious. It's a waste of EVERYONE'S time.

Accept that you've got a theory of intelligence, of acting like you're
dreaming. That's more than enough.

> >The problem of other minds asks: "I know I'm conscious and
> >intelligent, and I know you act intelligently. But are you conscious,
> >or do you just act that way?" This is a problem which your theory
> >doesn't solve. Keep your pants on, a solid theory of animal
> >intelligence will still win you the Nobel, but it won't help you solve
> >the mind-body problem.
>
> No, see above. The "atomic" brain is necessary, but not sufficient
> for consciousness. The virtual sensor(gan) is need to complete the
> gestalt experience of being intelligent. It is a substitute for sense
> messages from the sun and environment that used, in early evolution,
> to govern behaviour .... but also includes the "felt" aspects of the organ.

Ahem. The "sensor" isn't there. How and why is it that something not
existing would cause me to have a feeling? How can you show that this
is the case? Can you give me any evidence that you have more on your
hands than a theory of intelligence?

> Sorry, I am a reductionist on these matters, and believe that human-era
> philosophers just perpetuate "problems" and do little to contribute to
> solutions. They are not qualified to make judgements on physiological or
> empirical issues ... academic philosophy seems to just be linguistic
> analysis.

Quite right!

> In fact, together with religion, they are an unnecessary Mystery
> Industry that actual obstructs new knowledge and solutions so as to
> safeguard their incomes and paltry reputations. MVT is a more
> complete and basic theory than any narrow linguistic
> (definititional) account of the mind-body problem.

I fully agree. So let them play in their little play pen. You have a
theory of intelligence. Let them keep "consciousness".

> It solves dualism ... since Leibnitz objection to Descartes is overcome.
> A non-physical part of the brain can interact with non-physico-spatial
> thoughts ... like can only interact with like. The phantom median eye
> is constructed out of the same neuronal type of information as the sensory
> contents, so can reintegrate or focus all the diverse types of mentation and
> give a pervasive locus of self. The phantom median eye is an illusion, or
> trick of nature ... sure ... but then what mental experience isn't illusory?

Again, why should I believe a word of this? Can you give evidence
that THIS is true?

In particular, suppose I advanced a theory related to MVT, MVT'.
According to this theory, the non-existence of a pineal eye causes all
sorts of interesting intelligent behavior. But MVT' makes no claims
as to what causes consciousness.

How can you show that MVT is superior to MVT'? What evidence do you
have that your EXTRA claims about solving dualism, fixing the
mind-body problem, etc. are true? Why should I accept MVT over MVT'?
What *scientific* evidence can you give for MVT over MVT'?

> But we have a difference of type here .... lie machines and Turing devices
> do not occur in nature. They are designed and built, and are not evolved,
> living beings.

An irrelevant distinction. How something came to be reflects not at
all on what it is. Turing-like machines COULD have evolved (and, I argue,
they have: in us) so what's the relevant difference?

> And besides, I deny totally that lie detectors are empathising ... they do
> not replicate the feelings of the subjects, just look for incidental tells
> like
> conductivity of sweat or whatever.

How do you know? Have you asked one?

> Do you agree with McGinn that the mind-body problem is insoluble?
> If so, then you can pack up and go away. If you think that it can be
> solved, then you must also allow that an evolutionary account is
> needed. I accuse academic philosophers of "THE ATHENA FALLACY (TM)"
> ... that they deal with mind as if it sprung fully formed as "human"
> from the split head of Zeus ... and ignore the intermediate stages
> and natural processes that led to the current state of things.

Don't be silly. I commit no Athena fallacy: I have no account
WHATSOEVER about how mind comes about, say nothing of one like you
describe.

Anyway, I COULD pack up and go away, but I won't. Indeed, I argue
that you can pack up and leave us alone. ;)

> MVT puts forwards a set of predicates that can be challenged, and
> conclusions that follow from these, that can also be challenged. Yet
> no philosophers can falsify MVT .. although I have been shoving it
> up their noses since 1980.

Ho ho ho. MVT consists of some philosophical claims plus MVT', the
core scientific position I established earlier. Sure, no philosopher
can refute MVT', but there's no reason to believe the philosophical
bits of MVT. I challenge those parts, and only those parts. MVT' is
fine and sound.

> As a qualified Philosopher (MA, Univ Leeds ... which I returned, &
> Bsc, Surrey) I despair at the waste of everyone's time and money
> spent on philosophy departments. The arrogance is bearable .. they
> think they can dictate to proper mathematicians on the "philosophy"
> or overview of maths, even though in practice they perhaps can't
> even add up! And they presume to tell medical doctors what to do
> regards medical ethics.

Yes. Philosophy of mind is a stupid discipline. But you can't beat
us on our own turf. Just drop the philosophy, stick to the science,
and let's get ON with it.

> >Why YES. I think we're finally on the same page!
>
> So YOU have the problem, not MVT, which comes to a high degree
> of certainty as having solved the mind-body problem.

Don't be silly. The philosophy part of MVT has no certainty
whatsoever, no scientific evidence whatsoever. MVT' is well grounded.
But MVT minus MVT' is founded only on your intuitions, which may
differ with my own.

> Let me put this to you. No scientific theorum can be proved 100%
> (cite Heisenberg's principle). But we have to judge between the contending
> theoretical models and make the best choice. As it happens, the only
> competitor to MVT is Jerison's Recency Theory .. which can be fairly
> easily dismissed. So if MVT is just 50.1% likely ... more likely than not,
> then we are OBLIGED to accept MVT until a fuller account arrives.

No. I can accept MVT' and reject the rest.

> >But what do YOU get besides intelligence? Besides intelligent
> >behaviour? Hell, what do you NEED more than than that, besides
> >philosophical drivel?
>
> MVT extends the Neuromatrix theories of self, and gateway theory of pain
> (used by most clinicians) to give an account of felt experience and
> self-referential states. It ties together an evolutionary story that
> includes
> mechanisms for transition to endothermy, the reasons for REM and
> mechanics of dreaming, PLUS it overcomes Leibnitz Law (to be fully
> identical the two things must be fully interchangeable) thus gets rid
> of Cartesian dualism.

I'm with you until you say you've explained dreaming, instead of dream
reports. You've explained consciousness reports, but not
consciousness. You've got MVT', but not MVT. The extra drivel is
built into MVT: it claims to have identified the source of
consciousness. Drop that part and you have a scientific theory.

> What philosophical drivel do you think is needed? Please make a case.

None, actually, which is why I reject MVT. You'll get quite a kick
when you find out what I *really* think about experiences.

> Do your Turing machines fall in love then?

I do, yes.

-Dan

      -unless you love someone-
    -nothing else makes any sense-
           e.e. cummings



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:32:28 MST