From: ai-philosophy@yahoogroups.com [mailto:ai-
> philosophy@yahoogroups.com] On Behalf Of Ray Gardener
>
> I think that the solution is, instead of trying to neutralize the noise
> in a system, use it. Leverage the noise so that it becomes the ultimate
> foundation or driver of the system. In the noise, in the fundamental
> apparent randomness of reality, lies the qualities of truly living
> things.
>
> The human brain is a hierarchy of many minds, the smallest ones less
> capable than an insect's, the topmost one capable of grand abstractions,
> but all of them capable of experience and choice, the noise of each
> allowed to contribute and define the whole. It is a society of
> differring experiences, different opinions and different choices, and
> structured so that each higher mind experiences the choices of the ones
> below, and chooses from amongst them which is best. It is not a network
> of data, but a network of qualia, and it can only work if the initial
> inputs come from the very heart of reality -- the quantum noise. Dampen
> or ignore that, and the spirit is gone, the enterprise is lost.
>
> Ray
>
>

Very well said. In the minute fidelity of the noise lies some of the most
important information, yet of weak signal and computationally prohibitive to
extract directly it's influence ultimately could be an important
deterministic ingredient. Things are too clean w/o the noise, it doesn't
work that way, it's almost like noise is a complex system whose barrier
needs to be inclusively transgressed in order to get higher intelligence
leverage.

John

John G. Rose wrote:
>
> Very well said. In the minute fidelity of the noise lies some of the most
> important information, yet of weak signal and computationally prohibitive to
> extract directly it's influence ultimately could be an important
> deterministic ingredient. Things are too clean w/o the noise, it doesn't
> work that way, it's almost like noise is a complex system whose barrier
> needs to be inclusively transgressed in order to get higher intelligence
> leverage.

Something like that. I pictured noise as symptomatic of the very force
that animates the universe and gives it life, indeed defines life.

There was an article on Slashdot today referring to probabilistic
computing; how energy-efficient and fast it was. Given enough processing
elements, I believe one could both let the noise "speak for itself"
while simultaneously gaining useful reliability. A statistical averaging
can be used as an error detection/correction without adversely
offsetting the speed and energy efficiency.

Not to digress, but it's almost perfectly analogous to a dialogue I
imagine God had with the Devil concerning creation. God could create
life, but try as He might, it would either self-destruct inside a week
out of sheer boredom or putz around doing nothing of interest.

So finally the Devil slides up next to him and whispers, "I respect what
you're trying to do, old man... but you know in your heart, you can't
pull this off without me."

And God sighs so mightily that the heavens and earth tremble, and with
great sadness spreads his hands to let the Devil lend a hand. Satan
gleefully jumps down to Man and starts tinkering, introducing mortals to
apples and books and free will and so forth.

"You will be sinful," he says, teeth glistening. "and corrupt, and
craven, and decadent, and forever lusting for power. And paranoid, and
fearful, and stupid, and you will prey upon each other and cause untold
pain and suffering."

"But how can God allow it?" cries Man.

"Because," God answers from above, quietly, as if continuing to
acknowledge defeat. "it is the only way you will survive. Because to
exist without being interesting... it isn't enough."

"I'll start by giving you free will." says the Devil. "Then you can
choose yourself if you want all the rest. Fair enough?"

Man nods, and with his free will, thinks it through, and realizes the truth.

"Yes, I see now. You're right, there could be no other way."

And even as he destroyed himself, Man flourished, because he was no
longer bored. Great cataclysms and orgies of destruction waxed and
waned, and yet Man spread to every corner of the Earth, and eventually
the stars.

And God wept, and the Devil laughed, but one day he stopped laughing
long enough to say "You know, old man... I wasn't lying when I said that
I respected what you were trying to do. And I didn't really want to be
the bad guy, but I don't have the power of creation. I needed you as
much as you needed me. For what's it worth... if it's any consolation now."

And God watched as his creation -- perhaps, in all fairness, their
creation -- spread to one star after another, the galaxy slowly but
surely filling up.

"It's okay." He replied, even as His tears fell. He tried to focus on
all the good and wonderful things that were happening too, none of which
would ever have been without the Devil's help. And He knew that those
good things drove the Devil crazy, and as consolations went, it helped
keep him from destroying everything and trying yet again another
exercise in futility.

Ray

On Mon, Feb 9, 2009 at 9:39 PM, blroadies <blroadies@...> wrote:
> --- In analytic@yahoogroups.com, "Stuart W. Mirsky" <swmaerske@...> wrote:
>
>
>> The Searlean view seems to be that there is understanding and non-
>> understanding with a clear delineation between them.
>
> Tricky. Take the phylogentic scale. One sees a continuum from humans
> down to...you name it...beetles?
>
> But would you say that an amoeba has very minimum understandng or that
> it is inappropriate to apply the concept of "understanding" to it at all.
>
> My GPS gets me where I want to do. But is it appropriate to say it has
> any understanding of how it does this?

This doesn't seem to be quite the right way to approach the question.

The classical standard for "understanding" something, some sentence,
is being able to answer questions about it. To withstand a mini Turing
test about it.

If I want to test if you have understood something like general
relativity, I will ask you a problem about it, and expect you to
answer it right. Otherwise, I will think you have a superficial
understanding of it.

Searle is a vitalist, therefore he has double standards for  humans
and machines. He is a thinker that belongs to the better part of 19th
century, not 20th or 21st.

So, if we disregard the emotional drivel that is Searle, we can state
safely that: a formal logical language can be understood to the extent
that it can answer any questions stated in that language about a
particular logical sentence.

This is usually known as the completeness of the system in question.
Powerful enough systems are known to be incomplete, but that is not a
limitation just for computers. Humans are subject to *exactly* the
same limitations as computers. (And anyway, incompleteness is in a
particular sense, all those systems are computable in the limit, so
there is nothing to worry about even in theory)

That is, for instance in the case of answering questions about an FOL
sentence (given an interpretation domain) they are pretty much on the
same boat, and they understand the sentences to the same extent.

In the wider context, of course computers have a much harder job
because they scarcely know anything yet, and they work with quite
limited modes of reasoning (like some inefficient kinds of logical
inference). All of that is about to change, though.

In the case of a programming language, the analyzers can pretty much
answer many basic questions about the program: for instance, is this
variable used in a loop, is variable x compatible with variable y?
etc. and such information is used by compilers/interpreters. However,
of course a computer cannot answer all that can be answered by a
computer scientist because it has limited theoretical knowledge
encoded in it. For instance, a computer scientist like me can look at
a procedure and say that it has a running time complexity of O(n^2)
for input parameter n. A teenage C hacker might not be able to
"understand" this fact about the program, because he has limited
theoretical knowledge and thus the inference capacity is subdued, much
like the computer.

Therefore, one must understand that in the case of "open formal
systems" different agents attain different levels of  understanding,
properly understood as the extent of the set of questions that can be
answered by the agent about that very issue, or the problems that can
be solved by it. Thus, in the case of programming languages, a
computer has a basic understanding (i.e. it can answer what would
happen if this program ran with input "1"? etc. etc.) but not all
understanding that can be associated with the program. That, too, will
change, soon we will replace programmers with computers. The first
ones to lose their jobs will be low-level computer programmers. It was
a tedious job, anyway.

Best,

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

I am going to recap a thought experiment, that we discussed many times
on c.a.p and elsewhere. I was watching an episode of the Terminator
series, in which the ludicrously attractive gynoid Cameron converses
with John over this very issue (from my ever imperfect memory)

(John is driving and Cameron is waving her arm out of the window
trying to feel the breeze)
John: I don't think you can feel anything.
Cameron: I think you don't know how we work. I can feel the wind.

Well, it was a much better dialogue, but you get the idea. John has no
grounds to refute the hypothesis that she cannot experience. And
furthermore, questioning Cameron on whether she can really experience
something (i.e. not a zombie) is not going to prove anything as she
might be believing in a false thing. She can adequately explain data
processing but there is no simple way for her to compare their
phenomological worlds.

So, what do you think? Should John launch into a nonsensical diatribe
on property dualism or predicate dualism, after which the highly
intelligent robot would defeat the arguments? How can John be assured
that she does feel the wind in a way that is similar enough to his own
experience (i.e. fulfills conditions for being a subjective
experience)?

Cheers,

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

I think you're making two mistakes.

1. Just because a writer makes a robot say a particular
     line of dialogue, doesn't lend the idea credence.

2. Complexity and quantity make things more complex
     and bigger, but that's all. A quadrillion tinkertoys
     arranged to be a Turing machine is still just a
     lot of tinkertoys milling about. Before people
     build human-like robots, they'll simulate them
     in software, and a bunch of numbers being crunched
     doesn't feel anything, it is not anything more than
     a bunch of numbers changing value.

Derive material from observers, not observers from material.


Ray





Eray Ozkural wrote:
>
>
> I am going to recap a thought experiment, that we discussed many times
> on c.a.p and elsewhere. I was watching an episode of the Terminator
> series, in which the ludicrously attractive gynoid Cameron converses
> with John over this very issue (from my ever imperfect memory)
>
> (John is driving and Cameron is waving her arm out of the window
> trying to feel the breeze)
> John: I don't think you can feel anything.
> Cameron: I think you don't know how we work. I can feel the wind.
>
> Well, it was a much better dialogue, but you get the idea. John has no
> grounds to refute the hypothesis that she cannot experience. And
> furthermore, questioning Cameron on whether she can really experience
> something (i.e. not a zombie) is not going to prove anything as she
> might be believing in a false thing. She can adequately explain data
> processing but there is no simple way for her to compare their
> phenomological worlds.
>
> So, what do you think? Should John launch into a nonsensical diatribe
> on property dualism or predicate dualism, after which the highly
> intelligent robot would defeat the arguments? How can John be assured
> that she does feel the wind in a way that is similar enough to his own
> experience (i.e. fulfills conditions for being a subjective
> experience)?
>
> Cheers,
>
> --
> Eray Ozkural, PhD candidate. Comp. Sci. Dept., Bilkent University, Ankara
> Research Assistant, Erendiz Supercomputer Inc.
> http://groups.yahoo.com/group/ai-philosophy
> <http://groups.yahoo.com/group/ai-philosophy>
> http://myspace.com/arizanesil <http://myspace.com/arizanesil>
> http://myspace.com/malfunct <http://myspace.com/malfunct>
>
>

On Wed, Feb 11, 2009 at 3:08 AM, Ray Gardener <rayg@...> wrote:
> I think you're making two mistakes.
>
> 1. Just because a writer makes a robot say a particular
>    line of dialogue, doesn't lend the idea credence.
>
> 2. Complexity and quantity make things more complex
>    and bigger, but that's all. A quadrillion tinkertoys
>    arranged to be a Turing machine is still just a
>    lot of tinkertoys milling about. Before people
>    build human-like robots, they'll simulate them
>    in software, and a bunch of numbers being crunched
>    doesn't feel anything, it is not anything more than
>    a bunch of numbers changing value.
>
> Derive material from observers, not observers from material.

Ah, but my point is a little more than that. I am not assuming a
particular explanation of qualia, and I am simply expressing that it
may be impossible to tell the presence OR NON-PRESENCE of human-like
qualia from cognitive function, i.e. via a Turing Test.

So, it could turn out, since we are ignorant of that subject, that the
simulation of the robot possesses qualia as well, or just an AI that
runs on a computer. How do you know they don't? Do you have any
evidence for " a bunch of numbers being crunched doesn't feel
anything"?

How would John set out to prove that the robot doesn't "really"
experience anything, i.e. its report of its cognitive state is
misleading or is misinformed?

Best,

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

Eray said:
>
> How would John set out to prove that the robot doesn't "really"
> experience anything, i.e. its report of its cognitive state is
> misleading or is misinformed?
>

Bravo Eray!

As far as I can tell there is no such method possible. Currently in
physics there has been some controversy concerning the existence of
black hole-event horizons or if it's possible for singularities to be
naked. The common stance that singularities always results in "black
hole event horizons" (ie absolute cosmic censorship of direct
observational nature of singularities by event horizons) has been
successfully challenged on paper/in theory. But even if they succeed
in supporting this theory by finding and directly observing a naked
singularity thereby ruling out absolute cosmic censorship by black
holes, I'm afraid they will have a more difficult time (both in theory
and experimentally/observationally) ruling out the existence of the
absolute cosmic censorship of the nature of *someone else's* qualia
(be that someone else robot or human!).

But maybe I'm wrong (but what are the odds of that? really!)? ;o)

So maybe we should assume I am wrong and ask what would an argument
which would satisfy Eray's Terminator question look like? Certainly it
would have to be one which relies on the difference between robots and
humans at some particular level of organisation. The arguer must
establish how they know that "some-other person's" experience/qualia
is similar-enough to their-own first and only after this can they even
begin to rule out the Terminator as having anything like that...

JJG

--- In ai-philosophy@yahoogroups.com, Eray Ozkural <erayo@...> wrote:
>
> On Wed, Feb 11, 2009 at 3:08 AM, Ray Gardener <rayg@...> wrote:
> > I think you're making two mistakes.
> >
> > 1. Just because a writer makes a robot say a particular
> >    line of dialogue, doesn't lend the idea credence.
> >
> > 2. Complexity and quantity make things more complex
> >    and bigger, but that's all. A quadrillion tinkertoys
> >    arranged to be a Turing machine is still just a
> >    lot of tinkertoys milling about. Before people
> >    build human-like robots, they'll simulate them
> >    in software, and a bunch of numbers being crunched
> >    doesn't feel anything, it is not anything more than
> >    a bunch of numbers changing value.
> >
> > Derive material from observers, not observers from material.
>
> Ah, but my point is a little more than that. I am not assuming a
> particular explanation of qualia, and I am simply expressing that it
> may be impossible to tell the presence OR NON-PRESENCE of human-like
> qualia from cognitive function, i.e. via a Turing Test.
>
> So, it could turn out, since we are ignorant of that subject, that the
> simulation of the robot possesses qualia as well, or just an AI that
> runs on a computer. How do you know they don't? Do you have any
> evidence for " a bunch of numbers being crunched doesn't feel
> anything"?
>
> How would John set out to prove that the robot doesn't "really"
> experience anything, i.e. its report of its cognitive state is
> misleading or is misinformed?
>
> Best,
>
> --
> Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
> Research Assistant, Erendiz Supercomputer Inc.
> http://groups.yahoo.com/group/ai-philosophy
> http://myspace.com/arizanesil http://myspace.com/malfunct
>

This is a philosophy group, so just about anything is grist for
the mill, but as far as science is concerned, phenomena without
effects are the subject matter of the celebrated null hypothesis.
In other words, the phenomenological well of subjectivity is one
empty beaker.

--
Joe

On Wed, Feb 11, 2009 at 6:25 PM, Joe Legris <jalegris@...> wrote:
>
> This is a philosophy group, so just about anything is grist for
> the mill, but as far as science is concerned, phenomena without
> effects are the subject matter of the celebrated null hypothesis.
> In other words, the phenomenological well of subjectivity is one
> empty beaker.

Playing the devil's advocate, if Penrose were right, this would
probably mean that SE is not epiphenomenal, as I am hoping that if the
wave function is real, it accomplishes something. For instance it
might do quantum computation, achieving faster computation than
classical computation.

However, yes, phenomena with no effects... does not seem like a sound
concept, probably that's why epiphenomenal ascriptions to SE (like
property dualism, etc) are troubled, in general. As I indicated, we
expect such things like SE to be part of the big causal picture,
instead of being intangible entities.

Best,

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

--- In ai-philosophy@yahoogroups.com, Marvin Minsky <minsky@...>
wrote:
> However, our higher-level processes cannot detect all
> those intermediate perceptual steps and this lack of
> insight leads us to the belief that our sensations
> come to us in some way that is simple, direct, and
> immediate.

Those who know something about the brain wouldn't think
that these perceptions come without intensive computation.

> when you look at a color and sense that it is red, no
> intermediate steps seem to intervene and so, you can
> find nothing to say about it.


Just as SHRDLU could say nothing about its processes.

High level processes must involve short term memory so
they can be "thought about" whereas the processes that
tell us we are seeing yellow retain no trace or at least
not one accessible to the higher level processes we
call "being conscious".


> Gilbert Ryle 1949: A walker engaged in a heated dispute may
> be unconscious of the sensations in his blistered heel, and
> the reader of these words was, when he began this sentence,
> probably unconscious of the muscular and skin sensations in
> the back of his neck or his left knee. A person may also be
> unconscious or unaware that he is frowning, beating time to
> the music, or muttering.�
>
>
> Similarly, Joan might first notice a change in her gait, and
> only later notice that she's been favoring her injured knee.
> Indeed, her friends may be more aware than she is of how much
> that pain is affecting her.


That the content of consciousness is limited doesn't explain
anything special. Consciousness seems to be related to paying
attention to something. It is a selective process. You can
only act on one thing at a time. I can scratch that itch if
it comes to my attention.

It is only an itch if it comes to my attention.

Until then it nothing but a process without any qualia at all.



John

Marvin Minsky wrote:
>
> Agreed that a Turing Test may be too superficial.  However, this
> discussion has raised suggestions that it may be impossible to
> distinguish a person who 'feels' from a Zombie who doesn't -- if
> they both have the same behavior.

Well, it seems to me that it should come as no surprise that most (or
at the very least some) of us would strongly hold such opinions about
philosophical-zombies. After all, they were invented with exactly that
property of "impossibility to distinguish from a person who 'feels'"
intent in mind. But more than that, I'm not sure how anything within
the chapters you posted would help us to sort-out zombies?

JJG

Eray Ozkural wrote:
>
> Do you have any evidence for " a bunch of numbers being crunched
> doesn't feel anything"?

A failure to disprove such a supposition does not validate it. The
burden of proof is on you to show that numbers do feel. Otherwise we can
demand evidence for God's nonexistance, the nonexistance of celestial
teapots, unicorns and the like. Since the behavior of a computation can
be already explained by the processes acting upon it, saying that
numbers or machines feel is by far the more extraordinary claim. Demand
no disproofs, good sir! :)

If there are any feelings to be had, they are had outside the numbers,
by the observers who not only observe the numbers but in fact are the
ones, through their very observation, assigning the property of
numberness to them.

To be fair, I would conjecture that physical systems possess some kind
of observerness, since they are composed of matter/energy, but that's
not the same thing as saying that such systems feel after said
matter/energy has been configured to compute. In the machines we've
made, the information of the computations alone is being considered as
the whole of the intelligence, and to me, feelings cannot be formed out
of information but rather the other way around.

Success requires more than building devices that create information by
using the macroscopic chemical properties of matter. We must tap into
the very force that animates the universe, the very thing that causes it
to exist at all. And in that sense, the thing I find intriguing, is that
we seem to be on this path in a most inevitable fashion, as each
iteration of our computing machines utilizes ever smaller parts.

Ray

On Feb 12, 2009, at 1:18 AM, Ray Gardener wrote:

> Eray Ozkural wrote:
>>
>> Do you have any evidence for " a bunch of numbers being crunched
>> doesn't feel anything"?
>
> A failure to disprove such a supposition does not validate it. The
> burden of proof is on you to show that numbers do feel. Otherwise we
> can
> demand evidence for God's nonexistance, the nonexistance of celestial
> teapots, unicorns and the like. Since the behavior of a computation
> can
> be already explained by the processes acting upon it, saying that
> numbers or machines feel is by far the more extraordinary claim.
> Demand
> no disproofs, good sir! :)
>
> If there are any feelings to be had, they are had outside the numbers,
> by the observers who not only observe the numbers but in fact are the
> ones, through their very observation, assigning the property of
> numberness to them.
>
> To be fair, I would conjecture that physical systems possess some kind
> of observerness, since they are composed of matter/energy, but that's
> not the same thing as saying that such systems feel after said
> matter/energy has been configured to compute. In the machines we've
> made, the information of the computations alone is being considered as
> the whole of the intelligence, and to me, feelings cannot be formed
> out
> of information but rather the other way around.

Do you have any evidence for "feelings cannot be formed out
of [information-based processes]"?

It seems to me that you're making exactly the same kind of statement
Eray
made�and your  "for me" prefix doesn't provide any evidence.
Nor does it help to mention "matter/energy".

However,  it might help to mention "recursiveness".  Otherwise, it's
hard to see
how matter could matter.

Marvin Minsky wrote:
>
> Do you have any evidence for "feelings cannot be formed out
> of [information-based processes]"?

Five lines of thought:

1. I'm pretty confident that long before any life forms had any notions
about information, they had feelings. At the least, information is not a
requirement to have feelings.

2. If all the causality of an information system is attributable
to its workings, then there's nothing left over with which to feel.

3. I suspect that feelings are not a standalone item, but rather
tightly coupled with choice. Therefore a machine that could feel would
exhibit some physical behavior, however small, that would be
consistently different compared to a non-feeling machine.

4. If all the knowledge in a machine is represented by the
information in its memory, then there's nothing left over to represent
the knowledge of redness of red and other qualia.

5. The causality of claims to feeling can be traced to disprove feeling.
If a machine says that it feels pain, we can pause it, dump a stack
trace to see how this pain report came to be, and find that it breaks
down to nothing more than a number dance. The "pain" is just a label for
a value in some memory cell; the machine is saying what it was
programmed to say or what some genetic algorithm programmed it to say,
but either way, there was never a feeling of pain.

At best, it's a duet. The physical and the metaphysical require and play
off each other. So attempts to derive feelings from mechanism -- i.e.,
to have the mechanism be the causality of feelings -- will not work. The
height of a machine is the Universal Turing Machine, and any UTM can
emulate or host any other, thus mechanism only begets mechanism. The
arrangement of their parts doesn't ultimately matter, they're all
describing the same machine. It's a dead end, a UTM is as far as a
machine can go. To say the same thing another way, a UTM is the height
of what observers can imagine about physicality; it's the most complex
arrangement of parts we can think up.

One must take both meta and non-meta simultaneously and bring them
together. I favor this idea when thinking of biological evolution, which
let the natural dance of these two systems drive the form of "feeling
machines". It started right from the correct necessary basis -- atoms
and molecules -- and perhaps that basis extends much further back since
atoms evolved out of simpler things. And evolution leads to feeling
organisms precisely because a key selection pressure (fitness function)
is that organisms must feel. This fitness function is not defined by
some external agency (as in the case with programmers using genetic
algorithms to write software) but by the universe itself even as it is
evolving.

Even the fitness function evolves -- once evolution reached the
molecular or DNA level, traits such as reproducibility, strength, speed,
etc., got layered on top. These traits that are more mechanical in
nature couldn't get factored in until there was a more meaningful
foundation to build upon. I submit that in the early days, countless
combinations of DNA/RNA precursors were tried, but the only ones that
worked were those that took the universe's ability to feel into account.

It was like trying to produce a great work of art -- piece after piece
rejected because they never evoked any emotion. And then, finally, after
immeasurable patience, a molecular arrangement that touched the very
soul of reality and stirred its heart. A configuration that grouped the
simpler vibrations of atoms into a transcendant harmony. Here, the
fundamental force that animates the universe found purchase, could keep
being expressed, could operate in this new context. Feelings could be
had, choices could be made and noted. These new molecules were not just
matter, not just machines, but truly alive, feeling, the very passion of
the Big Bang flowing through them.

My other theory is that everything is alive and feeling, but we just
happen to do it on a much faster timescale relative to "non-living" things.

Ray

--- In ai-philosophy@yahoogroups.com, Michael Olea <oleaj@...> wrote:
>
>
> On Feb 11, 2009, at 11:34 AM, jgkjcasey wrote:
>
> > Those who know something about the brain wouldn't think
> > that these perceptions come without intensive computation.
>
> Hi, John.
>
> Good to see you are not on fire!
> (At least, I assume you aren't on fire.)
>
> -- M.
>

I will respond by email as this is off topic here.

John

--- In ai-philosophy@yahoogroups.com, "jgkjcasey" <jgkjcasey@...> wrote:
>
> --- In ai-philosophy@yahoogroups.com, Michael Olea <oleaj@> wrote:
> >
> >
> > On Feb 11, 2009, at 11:34 AM, jgkjcasey wrote:
> >
> > > Those who know something about the brain wouldn't think
> > > that these perceptions come without intensive computation.
> >
> > Hi, John.
> >
> > Good to see you are not on fire!
> > (At least, I assume you aren't on fire.)
> >
> > -- M.
> >
>
> I will respond by email as this is off topic here.
>

Wait, come back! It IS on-topic.

--
Joe

Eray Ozkural wrote:

> > Do you have any evidence for " a bunch of numbers being crunched
> > doesn't feel anything"?

Ray Gardener wrote:

> A failure to disprove such a supposition does not validate it. The
> burden of proof is on you to show that numbers do feel

I don't think so because Eray took no stance either way. But you did
when you said:

"a bunch of numbers being crunched doesn't feel anything"

Therefore the burden of proof of supporting such statements is yours.
The fact is that we can no more speak intelligently about what a bunch
of numbers feels than we can about what any other person on the face
of the planet feels, other than ourselves.

> Otherwise we can demand evidence for God's nonexistance, the
> nonexistance of celestial teapots, unicorns and the like.

Given any such ontological statements, such as your examples above,
for which there is zero evidence then yes, the safest stance to assume
is negative. But certainly there is a difference between Gods,
teapots, and unicorns for which it is true that there is zero evidence
and "feelings" for which there is zero doubt that I do have?

But, the question of how similar mine are to yours (or if you or the
number cruncher have them at all) is another thing altogether...

> Since the behavior of a computation can be already explained by the
> processes acting upon it, saying that numbers or machines feel is
> by far the more extraordinary claim. Demand no disproofs, good sir! > :)

No! It's no more or less an extraordinary claim than the claim that
humans feel.

JJG

On Thu, Feb 12, 2009 at 10:17 PM, John J. Gagne <john_j_gagne@...> wrote:
> Eray Ozkural wrote:
>
>> > Do you have any evidence for " a bunch of numbers being crunched
>> > doesn't feel anything"?
>
> Ray Gardener wrote:
>
>> A failure to disprove such a supposition does not validate it. The
>> burden of proof is on you to show that numbers do feel
>
> I don't think so because Eray took no stance either way. But you did
> when you said:
>
> "a bunch of numbers being crunched doesn't feel anything"
>
> Therefore the burden of proof of supporting such statements is yours.
> The fact is that we can no more speak intelligently about what a bunch
> of numbers feels than we can about what any other person on the face
> of the planet feels, other than ourselves.

Yes, you're right, I am simply stating that we don't know *if* that is
the case, I am not really asserting for or against it. To the best of
my knowledge, nobody on earth knows if that sentence is true.

And following Ockham's razor, doesn't help us much either, since our
knowledge of the basics of subjective experience is limited. When
prior knowledge is limited, pretty much anything goes...

At any rate, it would be a simpler theory if information processing
explained intelligence *and* subjective experience at the same time.
But equally simple explanations could be true: macro-scale quantum
effects of micro-tubules, DNA sex, the presence of a special hormone,
electrical currents of a special variety, quantum entanglement, black
holes, you name it. We simply don't know!

One reason I gave so much credit to information theoretic
understanding of subjective experience is because a) it seemed to
explain some interesting properties of SE b) it is just Newtonian
physics.

Best,

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

On Fri, Feb 13, 2009 at 5:33 AM, Pierre-Normand Houle
<houlepn@...> wrote:
> --- In analytic@yahoogroups.com, Eray Ozkural <erayo@...> wrote:
>>
>> On Thu, Feb 12, 2009 at 8:22 PM, Peter D Jones <peterdjones@...>
> wrote:
>>> --- In analytic@yahoogroups.com, Eray Ozkural <erayo@> wrote:
>>>
>>>> SE has all sorts of noun-ness and adjective-ness to it, and like
>>>> _anything_ it would be made up of parts so qualia is perfectly
>>>> clear to me as a concept, but I'd rather use something else as
>>>> that is a term tainted by vitalists like Searle.
>>>
>>> Searle doesn't use "qualia"
>>
>> Oh, he just uses Consciousness. Good for him!
>>
>> He is still an ignorant vitalist.
>
> But now maybe you should eschew the term 'consciousness' because it's
> tainted as well.

As Marvin would say, it's a suitcase word. And as previously discussed
in the parent thread, I'm indeed not using consciousness instead of
SE, the two are quite different things I believe.

I can imagine a biological neural network that constantly perceives a
world of blue. No relation to any truth. It only experiences a
sensation "out of the blue".

But I cannot imagine that hallucinating biological neural network to
have any consciousness whatsoever. Not in the ordinary sense. It
feels. It is not conscious: it knows nothing, it can think nothing but
a hallucination! It is not any more conscious than a simple reflex
agent (behaviorally *less* so) or a RAM chip that withholds a random
pattern indefinitely.

However, I do use the word C for what it *really* means: an array of
high-level cognitive functions that are essential for autonomous
intelligence.

Where possible, I will single out the cognitive function and focus on
it, instead of making some blanket statement about C.

Have fun with that!

PS: Also SWM's "conscious experience" may be quite confusing as well.
There are those things that we perceive, but that we aren't aware of
at a higher-level, and furthermore I might experience some basal state
of being when I'm asleep etc. etc. So I think that too is a term
loaded with presumptions and I will only use Subjective Experience.

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

Eray Ozkural wrote:
> > Ah, but my point is a little more than that. I am not assuming a
> > particular explanation of qualia, and I am simply expressing that
> > it may be impossible to tell the presence OR NON-PRESENCE of
> > human-like qualia from cognitive function, i.e. via a Turing Test.

Marvin Minsky wrote:

>
> Agreed that a Turing Test may be too superficial.
>

Barring any particular explanation of qualia, I can't think of any
better reason for granting possession of the property of qualia to
agents (all agents, regardless of composition) than Turing tests.
Certainly given that we currently have no particular explanation of
qualia (at all!), by what other method/test do we account for how we
do grant qualia to other humans? Question; is this "reason"
superficial? YES! I think it might be just that, but maybe not for the
reasons some might think! I'll do my best to explain�

If we have no particular explanation for qualia then by what means do
we attribute other humans having them too?

It seems to me before we can even begin to think about this question
we must establish what we mean by qualia. In what follows, I'm
discussing the quality of qualia differences and similarities between
humans (not between machines and humans)�

When someone refers to "red" they might be talking about a particular
range of the electromagnetic spectrum. If they did mean this, then any
system capable of sorting items/objects which emit (or reflect) that
range of the spectrum "sees/sorts-red".

On the other hand, if someone refers to the "qualia-red" then they
must mean that this "qualification" is not the same thing as that
particular range of the electromagnetic spectrum called (unqualified)
"red". We might even justify this by stating the fact that
"qualia-red" can be invoked/seen/known/or thought about even when my
eyes are closed and in the total absence of the frequency range of the
electromagnetic spectrum called "red".

Certainly anyone who might claim to understand this distinction
between "red" and "qualia-red" would be a candidate for granting
possession of the property of qualia. In other words, it seems to me
quite reasonable to assume that the claim of understand this
distinction is a necessary condition, even if though maybe not a
sufficient condition to conclude possession of qualia. Furthermore I
would submit that this act is nothing other than a Turing-like test we
use to conclude that other humans have qualia. But as I have already
stated this is an over generalization of a close enough condition
(though technically maybe insufficient condition to make such a
conclusion. Maybe�).

But why might this condition be insufficient? One very good reason is
that our definition of  "qualia" is still quite vague. How similar
must a particular "qualia" (say red) be to my own to qualify as
qualia? Lets simplify our choices to just two:

Choice #1. Qualia must be somewhat similar (recognizable) as something
I would call "red"
Choice #2. Qualia need not be at all similar to qualify as qualia

If we hold to C#1 then if it were somehow possible for me to know what
"qualia-red" is like for you and I was presented with an example of
such, I would tend to call it "red" without much training/getting use
to it.

If I hold to C#2 then the only requirement would be consistency. In
other words, no matter how different your "qualia-red" is to my
"qualia-red" it still qualifies as a "qualia-red" (or qualia-whatever)
as long as it's a consistent reference, even if it takes considerable
effort to learn and sort out those consistent elements.

Of the two choices above, it seems to me that C#2 is the fairest
standard and certainly the least arbitrary. Why? Well consider this,
by the C#1 definition it would be possible to for someone (a human
person) to pass a standard color-blindness test with flying colors
(matching your performance picture for picture) and still be
classified as "color-blind" just because their "qualia-color" is not
enough like my, your, or our "qualia-color". Doesn't that seem a bit
arbitrary to you? It does to me, anyway�

So, if I've done a good job so far, the reader should agree with me
that "red" is not the same thing as "qualia-red" and all "qualia-reds"
are equal as long as they are consistent references to "red". That
should have been the easy part. Now we move on to the hard part�

On Zombies

Some philosophers (apparently Chalmers is most notable) have
introduced the concept of zombies.

http://en.wikipedia.org/wiki/Philosophical_zombie

Quote:
A philosophical zombie, p-zombie or p-zed is a hypothetical being that
is indistinguishable from a normal human being except that it lacks
conscious experience, qualia, or sentience. When a zombie is poked
with a sharp object, for example, it does not feel any pain. However,
it behaves exactly as if it does feel pain (it may say "ouch" and
recoil from the stimulus), but it does not actually have the
experience of pain as a putative 'normal' person does.

The notion of a philosophical zombie is mainly used in arguments
(often called zombie arguments) in the philosophy of mind,
particularly arguments against forms of physicalism, such as
materialism and behaviorism.
End quote

Within this context, zombies can "see-red" without the property of
"qualia-red" (as I have defined that distinction above). Zombies would
also respond as if they understand this distinction when they simply
could not because zombies don't feel, don't experience, and absolutely
don't understand anything at all. But now we are face with a problem�
How could we ever hope to distinguish zombies for non-zombies, given
the choice of #2 as defining qualia above? The best you could hope for
is to say "I don't know but if it has consistent references I can't
find them". After all, what would it be like to "experience-nothing"?
It would be exactly like nothing at all. So if anyone says they
understand zombies (and it's TRUE!), you would certainly have
sufficient and necessary condition for being a zombie!

So, to conclude is the Turing test too superficial for sorting-out
zombies? Yes, maybe so.

John J. Gagne

1. Whenever my brain is working properly (in its main mode), I am
aware that I have a mind.
2. When something changes in my mind, something surely changes inside
my brain, or CNS to be more precise.
3. If my brain stops working properly (as in knocked out), the
experience goes away.
4. If my body freezes (ceases to work at normal temperature and
speed), then my mind freezes as well.
5. If my body then recovered without damage, my mind would continue to exist.
*6. Hence, the *physical operation* of my brain is what I call the
mind, i.e. it is a physical *process*
7. If I take neurological drugs, then the physical operation of my
brain and my mind are influenced.
8. I "perceive" something whenever I have ascertained a fact about
something through analysis of my sensory input, i.e. infer a physical
event.
9. I don't know if the red I see is the same as anyone else's
10. The way I see is different from the way I hear *qualitatively*
11. There is a geometry of my mind, so that there is a space, and
thoughts filling that space, and existing in geometric relation to one
another
12. I can experience many thoughts/feelings at once
13. If something in the current *operation* of my mind changed, then
my *experience* would change as well. (BUT THAT IS QUITE DISPUTABLE!)
14. If something in my brain changed that did not physically influence
the current operation of my brain, I would not feel anything
15. But then, for instance, it destroyed some part of my memory, it
might change my future mental states

There are many more but I will stop here. Obviously, this goes towards
a complete argument for computationalism, versions of which I've come
to defend over years.

Add to this list as you please.

--
Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara
Research Assistant, Erendiz Supercomputer Inc.
http://groups.yahoo.com/group/ai-philosophy
http://myspace.com/arizanesil http://myspace.com/malfunct

Consider this, wikipedia used the phrase:

"The notion of a philosophical zombie"

Certainly a "notion" implies some kind of understanding of zombies.

No?

But if it's true that Zombies don't feel, don't experience, and
absolutely don't understand anything at all then how could we possibly
understand what that would be like (we can't because it's not "like"
anything) and if we don't understand what that would be like then how
can we possibly have a "notion" of a zombies"?

The zombie liar says "I understand zombies" but has no ability to lie
and tells no truths...

Zombies are a rather silly and paradoxical "notion" (if they can be
called such a thing as a "notion" at all?) and certainly not worth
expending our efforts on.

But, maybe that just because "I don't understand zombies"?

JJG

PS I'm reading your post and I'll comment if I have something worth
saying. Thanks for your comments/interest as it tends to disturb me
what I think we're not exactly on the same page and they provide me
the feedback for locating where we branch and how far off we really
are from one another. It usually turns out to be far less than I
initially estimated (from my completely unbiased assessment of the
situation, that is).

;o)

Marvin Minsky wrote:

> Now, we could escape from such suffering, by rebuilding ourselves
> in the following way: Let's add, to your internal pain-machine, a
> copy of most of the rest of your brain. Then that copy could serve
> as your bodyguard, while the original one does your regular job
> (and philosophers start to argue about whether that copy now suffers
> instead). And now that we have imagined this scheme, we can see how
> to deal with that second question, of why did suffering evolve?
> It's because there was no good alternative:

Bingo, Bango, Bongo, and Irving, I think we have a winner!

Sure, any system capable of interpreting a situation which requires a
massive and immediate reassignment of critical resources, suspending
high level goals, would require some way of grabbing the attention of
those sub-process and taking whatever it needs. This would also
(ideally) have to be a rather abnormal situation, by definition. It
would be rather counter productive otherwise because these situations
usually have critical time-life. If the lions share of the resources
are tied up beyond this critical time-life dealing with a situation
which no longer exists, high level process will not be able to compute
ways to prevent these things form happing in the first place and
that's a very bad thing to pass on to our off spring...

JJG

On Feb 13, 2009, at 8:06 PM, John J. Gagne wrote:

> Consider this, wikipedia used the phrase:
>
> "The notion of a philosophical zombie"
>
> Certainly a "notion" implies some kind of understanding of zombies.
>
> But if it's true that Zombies don't feel, don't experience, and
> absolutely don't understand anything at all then how could we possibly
> understand what that would be like (we can't because it's not "like"
> anything) and if we don't understand what that would be like then how
> can we possibly have a "notion" of a zombies"?

Yes, I agree.  Of course, there is no problem with the idea of a
machine that always behaves in the same way as does  some particular
person�let's call him George.  For example, that machine could be a
big look-up table that has a response for each possible input sequence
of (say) 50 years of previous inputs. (But if we suppose, say. a new
input stimulus every second, that table would need to have n to the
1,500,000,000 entries (where 'n' is the number of bits in each
input).   The last time I saw an estimate, the then-current model of
our universe could only hold about 10 to the 80 bits.  (I don't
recollect the theoretical basis for this.)


>
> The zombie liar says "I understand zombies" but has no ability to lie
> and tells no truths...
>
> Zombies are a rather silly and paradoxical "notion" (if they can be
> called such a thing as a "notion" at all?) and certainly not worth
> expending our efforts on.

I agree with this, too.  I spent some time on it, more for political-
religious-philosophical reasons, than for technical or psychological
reasons.

On the other side, it seems to me that we *do* need technical theories
about the processes involved with "feelings" and so called 'subjective
experience.'

>
>
> But, maybe that just because "I don't understand zombies"?
>
> JJG
>
> PS I'm reading your post and I'll comment if I have something worth
> saying. Thanks for your comments/interest as it tends to disturb me
> what I think we're not exactly on the same page and they provide me
> the feedback for locating where we branch and how far off we really
> are from one another. It usually turns out to be far less than I
> initially estimated (from my completely unbiased assessment of the
> situation, that is).
>
> ;o)
>
>
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>

On Feb 13, 2009, at 9:41 PM, John J. Gagne wrote:

> Marvin Minsky wrote:
>
>> Now, we could escape from such suffering, by rebuilding ourselves
>> in the following way: Let's add, to your internal pain-machine, a
>> copy of most of the rest of your brain. Then that copy could serve
>> as your bodyguard, while the original one does your regular job
>> (and philosophers start to argue about whether that copy now suffers
>> instead). And now that we have imagined this scheme, we can see how
>> to deal with that second question, of why did suffering evolve?
>> It's because there was no good alternative:
>
> Bingo, Bango, Bongo, and Irving, I think we have a winner!
>
> Sure, any system capable of interpreting a situation which requires a
> massive and immediate reassignment of critical resources, suspending
> high level goals, would require some way of grabbing the attention of
> those sub-process and taking whatever it needs. This would also
> (ideally) have to be a rather abnormal situation, by definition. It
> would be rather counter productive otherwise because these situations
> usually have critical time-life. If the lions share of the resources
> are tied up beyond this critical time-life dealing with a situation
> which no longer exists, high level process will not be able to compute
> ways to prevent these things form happing in the first place and
> that's a very bad thing to pass on to our off spring...
>
> JJG

That's a nice way to look at the situation.  Interesting: religious
people find it hard
explain why a benevolent God would endow us with 'suffering.'  But
your arguments
suggests that this is useful, at least, in what you call those "rather
rather abnormal situations."

In fact, the next section (3.4) of my book suggests that this theory
does not explain why
some people are plagued with "chronic painful suffering " �which
persists for longer than does
the immediate "abnormal situation."   And the following section (3.4)
suggests that this may
simply be a 'bug' that evolution has not yet had time to remedy.
(Besides, aspirin and phenacetin
work well enough that Evolution might never get around to fixing that
bug!)

Marvin Minsky wrote:

> In fact, the next section (3.4) of my book suggests that this theory
> does not explain why
> some people are plagued with "chronic painful suffering " �which
> persists for longer than does
> the immediate "abnormal situation."   And the following section
> (3.4)
> suggests that this may
> simply be a 'bug' that evolution has not yet had time to remedy.
> (Besides, aspirin and phenacetin
> work well enough that Evolution might never get around to fixing
> that bug!)

Yes, but why deprive evolution of the credit for aspirin? By the
previous argument, it was she who provided the system capable of high
level reasoning and prediction for the express purpose of "fixing
bugs" (call it evolution 2.0)...

Of course, such a system also introduces a great many "bugs" in the
process and only time will tell if we can fix them faster than we tend
to introduce them (there seems to be an acceleration at work there).
If not she also has no particular problem starting over from scratch
with no hard feelings ;o) AI too will be a continuation of evolution
2.0 with even less in common with the products of evolution 1.0.. If
we do it right they will very good "bugs fixers" indeed.

JJG

Bob: I just had the oddest dream.

Alice: You don't say.

    Bob: But I do say. I was in a musuem looking at great
         works of art, when all of a sudden a burning bush
         appeared to my left.

Alice: _The_ burning bush?

    Bob: Yes, it was God himself, apparently visiting me.

Alice: So what did He say?

    Bob: He asked me if I'd seen anything interesting.

Alice: Did you?

    Bob: Well, yes, I had seen a few works I thought were
         very moving, but when I showed them to Him, He
         just yawned.

Alice: So this burning bush yawned... by opening its
         burning branches, I suppose.

    Bob: Well, it was definitely the sense I got of His
         reaction. I asked Him if He was okay, and He said,
         this is the problem with omniscience. It would
         take something really novel to impress me.

Alice: He's got you there. He's probably seen it all
         and then some.

    Bob: Yes. He offered to show me some of His art, and
         he gave me a careful peek, and just that small bit
         was enough to nearly erase my mind and leave me
         crying catatonically.

Alice: What was it you saw?

    Bob: You know those nice Pacific coast sunsets you're
         always raving about?

Alice: Uh-huh.

    Bob: Well, multiply the nicest one by a hundred and that
         would kind of come close.

Alice: Too bad I missed it.

    Bob: I felt sorry for Him. Apparently he's lost the
         ability to feel emotion, since nothing can impress
         Him anymore. Nothing can move Him... there's no way
         for art to evoke an emotional response in Him.

Alice: Really? How sad. Seems hard to believe.

    Bob: No, it's true. Oh, He can have feelings, but the
         way He explained it to me, is that they're like
         happening way down at some lower level. He knows
         that they're there, but they're distant things,
         mere data in a vast storehouse that can impinge upon
         His awareness but little else.

Alice: What are you saying? That God is a machine?

    Bob: Well, I guess it sounds crazy, but, yeah... in a
         certain way, He's... a machine. Sentience without
         feeling. He knows that Abel getting murdered
         was a downer, and that tidal waves drowning people
         isn't fun, but it's like they're just abstract
         datums. It's like He's... He's...

Alice: He's so above it all, you mean.

    Bob: Yes! He's above it all. He's so smart, that
         whatever He doesn't already know, He can simply
         predict accurately enough, so all the surprise
         is gone, and by the time anything good or bad
         happens He's already resigned Himself to it
         way ahead of time.

Alice: Poor deity. He must be bored too.

    Bob: Oh yes, very.

Alice: It is interesting, though.

    Bob: Yes. All that knowledge, all that intelligence...
         makes Him closer to a machine. Ironic. And if
         all this time we've been trying to make a strong AI
         by giving it smarts, maybe that's self-defeating.

Alice: How can He stand it? Or has even the feeling
         of boredom become abstract to Him?

    Bob: Perhaps. Before I woke up, He told me something,
         something important.

Alice: What?

    Bob: He said to me: Bob, you, Alice, and all the
         others, and everyone who has existed and who ever
         will exist, they exist for a reason.

Alice: Yes? Which is what?

    Bob: He said, I'm so bereft of feeling that I can
         become unaware of my own existence, stop feeling
         what it is like to be me, if you follow my meaning.

Alice: Really? Far out.

    Bob: Yes, yes. But there's more. He then said,
         but you exist, and with your limits, you can feel.
         So as I look out through your eyes, and live through
         your lives -- while simultaneously forgetting my own --
         I can feel again. Through you, I can be surprised
         again; take delight again; weep again.

Alice: Hmm. Was He talking about our mortal, limited lives,
         or about the physical realm in general?

    Bob: You know, I'm not sure. But if He was, it would
         explain why the physical realm and all of its
         measurables exists.


Ray