On 02 Mar 2012, at 18:03, Craig Weinberg wrote:

On Mar 2, 4:43 am, Bruno Marchal <marc...@ulb.ac.be> wrote:
On 01 Mar 2012, at 22:32, Craig Weinberg wrote:

There is no such thing as evidence when it comes to
phenomenology. You don't need evidence to infer that a clock
know what time it is.

A clock has no self-referential ability.

How do you know?

By looking at the structure of the clock. It does not implement
reference. It is a finite automaton, much lower in complexity
than a
universal machine.

Knowing what time it is doesn't require self reference.

That's what I said, and it makes my point.

The difference between a clock knowing what time it is, Google
what you mean when you search for it, and an AI bot knowing how to
have a conversation with someone is a matter of degree. If comp
that certain kinds of processes have 1p experiences associated with
them it has to explain why that should be the case.

Because they have the ability to refer to themselves and understand
the difference between 1p, 3p, the mind-body problem, etc.
That some numbers have the ability to refer to themselves is proved
computer science textbook.
A clock lacks it. A computer has it.

"This sentence" refers to 'itself' too. I see no reason why any number
or computer would have any more of a 1p experience than that.

A sentence is not a program.

Okay, "WHILE  program > 0 DO program. Program = Program + 1. END

Does running that program (or one like it) create a 1p experience?

Very plausibly not. It lacks self-reference and universality.

By comp it
should be generated by the 1p experience of the logic of the gears
the clock.


If the Chinese Room is intelligent, then why not gears?

The chinese room is not intelligent.

I agree.

The person which supervene on the
some computation done by the chinese room might be intelligent.

Like a metaphysical 'person' that arises out of the computation ?

It is more like "prime numbers" arising from + and *. Or like a chess
player arising from some program, except that prime number and chess
player have (today) no universal self-referential abilities.

That sounds like what I said.

By comp logic, the clock could just be part of a
universal timekeeping machine - just a baby of course, so we
expect it to show any signs of being a universal machine yet,
but by
comp, we cannot assume that clocks can't know what time it is
because these primitive clocks don't know how to tell us that
know it yet.

Then the universal timekeeping would be conscious, not the baby
Level confusion.

A Swiss watch has a fairly complicated movement. How many watches
it take before they collectively have a chance at knowing what
time it
is? If all self referential machines arise from finite automation
though (by UDA inevitability?), the designation of any Level at
all is
arbitrary. How does comp conceive of self referential machines
evolving in the first place?

They exist arithmetically, in many relative way, that is to
numbers. Relative "Evolution" exists in higher level description of
those relation.
Evolution of species, presuppose arithmetic and even comp,
Genetics is already digital relatively to QM.

My question though was how many watches does it take to make an
intelligent watch?

Difficult question. One hundred might be enough, but a good engineers might be able to optimize it. I would not be so much astonished that
one clock is enough, to implement a very simple (and inefficacious)
universal system, but then you have to rearrange all the parts of

The misapprehensions of comp are even clearer to me imagining a
universal system in clockwork mechanisms. Electronic computers sort of
mesmerize us because electricity seems magical to us, but having a
warehouse full of brass gears manually clattering together and
assuming that there is a  conscious entity experiencing something
there is hard to seriously consider. It's like Leibniz' Windmill.

Or like Ned block chinese people computer. This is not convincing.

Why not? Because our brain can be broken down into components also and
we assume that we are the function of our brain?

We are relatively manifested by the function of our brain. "we" are not function.

If so, that objection
evaporates when we use a symmetrical form <> content model rather than
a cause >> effect model of brain-mind.

Form and content are not symmetrical.
The dependence of content to form requires at least universal machine.

is just helpful to understand that consciousness relies on logical
informational patterns that on matter. That problem is not a problem
for comp, but for theories without notion of first person. It breaks
down when you can apply a theory of knowledge, which is the case for
machine, thanks to incompleteness. Consciousness is in the "true"
fixed point of self-reference. It is not easy to explain this shortly
and it relies on Gödel and Tarski works. There will be opportunities
to come back on this.

All of that sounds still like the easy problem of consciousness.
Arithmetic can show *that* self reference exists but it does so by
drawing a circle around a hole where the self should be. It is a 3p
outside view looking in and finding only an abstract vector (pseudo
1p). This is indeed accurate from a 3p logical perspective, which is
why it is internally consistent and can be used to make sophisticated
puppets featuring trivial intelligence which can be elaborated to a
degree far exceeding human trivial intelligence, but still possessing
no feeling, understanding, or experience.

Gôdel 1931 stays in the 3p, and Gödel 1933 assess this by observing that Bp cannot be used for knowledge, but that's where you can apply Theaetus' theory of knowledge, by defining knowledge by Bp & p, factually. That truth restriction makes prossible to study meta- logically a non formalizable theory of knowledge associated to the correct machine. It behaves like a knower, and it mirrors well the Plotinian conception of everything.

Your non attribution of consciousness to the machine might comes from the fact that you believes that the machines is only handled by the 3p Bp, but it happens that the machine, and its universal self- transformation has self-referential correct fixed point, and who are you to judge if she meant them or not? If you define consciousness by the restriction of the Bp on the such true fixed point, the PA baby machine will already not be "satisfied" if you call her a zombie.

you were able to make a living zygote large enough to walk into, it
wouldn't be like that. Structures would emerge spontaneously out of
circulating fluid and molecules acting spontaneously and
simultaneously, not just in chain reaction.

It doesn't really make sense to me if comp were
true that there would be anything other than QM.


Why would there be any other 'levels'?

So you assume QM in your theory. I do not.

It doesn't have to be QM, it can be whatever you like - arithmetic
truth, Platonia, etc. Why have any other 'level'?

Nice. Let us chose first order arithmetical truth. The formula that we can write with "=", the logical symbols (with "A" for "for all", "E"" for "it exists", and x, y, z, ..., as variables), and the symbol "0", "+", "*" and "s".

Do you agree with the intended meaning of the axioms I use:

Ax ~(0 = s(x)) (For all number x the successor of x is different from zero). AxAy ~(x = y) -> ~(s(x) = s(y)) (different numbers have different successors)

Ax x + 0 = x
AxAy x + s(y) = s(x + y) ( meaning x + (y +1) = (x + y) +1) = laws of addition

Ax   x *0 = 0
AxAy x*s(y) = x*y + x    laws of multiplication

This defines a UD.

And in that theory, we can prove the existence of (tiedously) a machine which "believes" the axiom above together with the infinity of axioms (for all formula F translatable in the machine's language):

(F(0) & Ax(F(x) -> F(s(x))) -> AxF(x)

This defines the machines I will interview "in" the theory above.

Then the levels will grow, including many cycles, (strange) loops, self-reference, and relatively true self-reference, and "absolute" fixed points, ... well a whole "theology", I think.

No matter how complicated a
computer program is, it doesn't need to form some kind of non-
programmatic precipitate or accretion. What would be the point and how
would such a thing even be accomplished?


Deep Blue or Watson don't need to define some new 'level' of
interpretation which transcends programming or re-presents it in some

Because deep blue is programmed to play some toy game. Not the struggle of life game, in which you need self-referential control structure, short and long term memories, universality, Löbianity, etc.

Why go through the
formality of genetics or cells? What would possibly be the point? If
silicon makes just as good of a person as do living mammal cells,
not just make people out of quantum to begin with?

Nature does that, but it takes time. If you have a brain disease,
answer is like a doctor who would tell you, just wait life appears on
some planet and with some luck it will do your brain.
But my interest in comp is not in the practice, but in the conceptual
revolution it brings.

I think that comp has conceptual validity, and actually could help us understand consciousness in spite of it being exactly wrong about it.
Because of the disorientation problem, being wrong about it may in
fact be the only way to study it...as long as you know that it is only
showing you a shadow of mind, and not mind itself.

We don't know that.

I know it as much as I know anything.

- either you have the magical ability to distinguish humans from zombies, or - you must explain what in the brain is not Turing emulable, and how it interacts with the behavior.

A machine which can only add, cannot be universal.
A machine which can only multiply cannot be universal.
But a machine which can add and multiply is universal.

A calculator can add and multiply. Will it know what time it is if I
connect it to a clock?

Too much ambiguity, but a priori: yes. Actually it does not need a
clock. + and * can simulate the clock. Clock is a part of all
computers, explicitly or implicitly.

This is a good way to show the difference between the a-signifying,
generic 'sense' of time that you're talking about, versus the
anthropocentric, signifying sense. All of those old VCRs flashing
12:00 forever, even though there is a perfectly good clock on board
shows the extremely limited capacities of even a digital clock to tell
time. A microprocessor has only disconnected recursive enumeration.
There is no temporal context to it. If you set it to 7:00 or 13505:00 it makes no difference. Those symbols aren't grounded into anything at all, they are digital units representing nothing at all. No qualia, no
1p awareness.

So you assume a continuous time?

I assume no time other than memory of experience in the perpetual now.

That's duration. If you say it is perpetual, I think that you still assume some notion of time.

Knowing the time is a function of understanding. It only has relevance
in particular contexts, like 3D vision or olfactory sense.


You would be an alien, you might say that human people have no qualia,
given that they do not seems to be present in any cut section of a
human body.

They would be right that I have no alien qualia. I don't say that the
components of a clock don't have qualia - I think that they must, but
I suspect its much less (say one quintillionth) as significant as
ours. Because the qualia is so primitive, there is no 1p coherence to
the 'clock' assembly as a whole. There is zero increase in 1p
significance over and above the value of the parts. That is not to say
there is no increase in significance to us by virtue of possessing and
using the clock, of course there is great utility, joy, comfort,
learning, knowing, etc.

Why would the consciousness of the clock parts not add in the clock- reassembled computer, like you seems to say that human consciousness comes from the addition of the consciousness of its neurons?

The machine is a whole, its function belongs to none of its parts.
When the components are unrelated, the machine does not work. The
machine works well when its components are well assembled, be it
artificially, naturally, virtually or arithmetically (that does not
matter, and can't matter).

The machine isn't a whole though. Any number of parts can be
without irreversibly killing the machine.

Like us. There is no one construct in the human body which lasts for
more than seven years.

Not like us. If any major organ replacement fails for any reason, we
will die. A machine could sit in a machine shop for 100 years and be
perfectly viable if it gets fixed at that time.

Some seed can live thousand of years.

A thousand years means nothing to a seed. It doesn't begin living
until it germinates, that becomes year zero for it.

OK. But then why would not the machine live again when someone buys it and turns it on. Why would that not becomes the year zero for it (assuming its memory are virgin of experiences)?

You should not compare the crude
man made machine with natural nanotechnology having a very long
history. No one doubt that life is a very sophisticated technology.
Some frog can freeze completely, and after 4 month of seemingly death,
come back to their activities.

As far as I know, all living organisms arise from a single dividing
cell and no machines are built that way.

Is a ribosome alive?

This may be a much bigger
deal than it sounds if consciousness 'insists' through memory rather
than appears instantaneously as a function of objects in space.

Memory is a key, sure. But it is an information pattern, usually interpreted by some information handling.

I will not blame you to reintroduce the ghost in the machine. crudely said, computer science justifies the existence of the ghost (software) in the machine (hardware).

If we
started building machines this way, as nanotech seeds, I think we
would gain 1p sentience, but lose control of it.

Yes. But that's the point. Now when you accept an artificial brain, and if the level is chosen correctly, by definition, you will neither lose, nor gain more control on yourself than you already have, or have not.

Brains have much shorter material identity. Only bones change more
slowly, but are still replaced quasi completely in seven years,
according to biologists.

True, but they are replaced with tissues which are appropriately aged, not stem cells. The biographical narrative of the organism as a whole
is maintained.

Even if that is correct for current machines, comp is considering all

That's the theory.


All know theories in biology are known to be reducible to QM, which
Turing emulable. So your theory/opinion is that all known theories

They aren't false, they are only catastrophically incomplete.
biology nor QM has any opinion on a purpose for awareness or living
organisms to exist.

That does not entail that QM structures or biological structure
be aware, or bear local notion of persons.

If we were not ourselves aware, would anything that QM or biology
entails leas us to suspect that a such thing as awareness could be

Yes. Their ability to support universality and self-reference.

Why should universality and self-reference indicate awareness of any
kind. I have motion sensors on my garage lights. I could make them
universal by plugging them into a TV set instead of lights.


They are
self referential because whenever I go in the garage at night, they
turn on to greet me and to make their presence known.

This means perhaps that they are "Craig Weinberg" referential, not self-referential.

Turning emulation counts on computation being sufficient to
support life and awareness, but it's an arbitrary wish.

All theories collect evidences. Comp has many positive evidences. Non-
comp has only the absence of solution to a problem kind of evidence.

Non-comp has sense.

I completely agree.

It doesn't need evidence because the thought of
needing evidence is already non-comp symbol grounding.

I might agree with this, because comp implies something very akin to this, from the machine 1-povs.

A machine will
solve problems using whatever parameters or data it is given. It has
no capacity to doubt them unless programmed to act as if it were
doubting them.

It is here that you should study the math. Ideally self-referentially correct machines cannot miss the doubt. They becomes quickly modest, and know that whatever they learn, their ignorance will only get bigger. Forever.

Machines never want or need evidence. They extrapolate
recursively, forever.

That might be defended about the ideal virgin machine, dissociated from all other universal machine. But any reasonable computation will involve the many dependencies between machines, and this can play an important role, as Stephen suspects I think rightly, in the first person plural realities, which makes number dream consistent and coherent (multi-consistent with respect to universal numbers). In that case, the machine cries for evidences, when betting on such local universal neighbors.

But non-comp faces the same difficulties, except that it hides them
more easily, in special vague infinities.

It doesn't hide anything, any more than the our sense of humor hides.

That's only funny :)

It is theory which needs to justify it's relevance in terms of sense,
not the other way around.

You are right, but only from the first person point of view. It is a key point for the knower, perhaps the human right brain, and part of the limbic system. But doing science, the theory should no more refer to the sense of the one who does the theory, only to the sense "object of theory", if not you do pseudo-religion-science. You can do that. There is a public for that, but it is no more like searching the truth, but asserting personal opinions (in the best case, because usually the pseudo-things is just a selling strategy).

We aren't
seeing anything especially hopeful to back it up.

Study a book in computer science. Look at molecular biology, or
quantum mechanics.

Nice theory but no payoff in terms of breakthroughs in consciousness.

Read this list, study my work :)

You have to lower the comp level in the infinitely low, and
introduce special infinities, not 1p machine recoverable to make

No, you can just reject the entire presumption that computation by
itself has causal efficacy.

But it has causal efficacy, even with zombie, which can decide and
on the environment like us.

Only because there is a material body which can input and output to a
material environment.

Immaterial body can input and output to an immaterial environment.

Only in theory. I don't think it is the case in reality.

It is a consequence of the theory that this is the case in reality. We start from the innocuous local relative "yes doctor", and then the reasoning shows that we are already in an arithmetical matrix, and we can explain why, from inside, it looks like an analytical, physical, gigantic history.

A program, without a physical substrate, has no
causal efficacy (if it could, we wouldn't need computers).

Yes, but the point is that a physical substrate is a relative notion,
for relative (indexical) use.

We don't know that.

I was assuming comp. That is true as far as comp is true.

Its qualia of physicality to us is certainly a
relative notion, but with sense, it would be the case that the
relative notion is the actual concrete presentation of realism for us.
Its specular reflection - seeing 'through' the proximal surface to the
distal image. Sense means it is both relative and absolute.

Sense is absolute, in a sense, for machine. I don't think there is a problem here. Sense is absolute from the 1-pov, and relative from (and to) 3p local descriptions.

Computation to me is clearly an
epiphenomenon of experienced events, not the other way around.

Computation are well defined object in arithmetic. You cannot
standard notion to suit your point. Or you can conclude whatever you
want at the start.

Arithmetic is an experience too.

You confuse arithmetic and the experience of arithmetic.

No, it's just that I think all arithmetic is an experience of
something - whether it's a brain, a cell, or a semiconductor. Not
empty space.

It is easier to explain the sense of "cells", "brains", semiconductor" and the "experience of arithmetic" from arithmetic, than from brain and cells. you assume what we have to explain.

Comp makes this clear. And that is its virtue.

Your "non-comp theory" seems to assume both matter and mind, which is too much for me. (Sorry).

This is
another variation on the Chinese Room. The pig can walk around
30,000 feet and we can ask it questions about the view from up
but the pig has not, in fact learned to fly or become a bird.
has the plane, for that matter.

Your analogy is confusing. I would say that the pig in the plane
fly, but this is out of the topic.

It could be said that the pig is flying, but not that he has
to fly* (and especially not learned to fly like a bird - which
be the direct analogy for a computer simulating human

That why the flying analogy does not work. Consciousness concerns
something unprovable for everone concerned, except oneself.

No analogy can work any better because nothing else in the
universe is
unprovable for everyone except oneself except consciousness.


Nothing but consciousness is subjective. Nothing else besides
consciousness is unprovable to others but unnecessary to prove to

Good. that's a point for the machine's consciousness theory, which
relate consciousness and consistency. indeed only consistency and all
G* minus G propositions appears to the machine as unprovable to
others, but easily inferable to oneself.

Consistency is only a comment on an aspect of consciousness though,
just as a shadow of a tree has a basic tree shape. It doesn't define
the tree, it's a silhouette.

It is already something, and it is just Dt, looks at the many sensical combinations the machine will give sense too.

May I ask you a question? Is a human with an artificial heart
still a

Of course. A person with a wooden leg is still human as well. A
with a wooden head is not a person though.

OK. So the problem is circumscribe to the brain. Someone can have an
artificial body, but not an artificial brain.
Could someone survive with an artificial cerebral stem?

It depends how good the artificial brain stem was. The more of the
brain you try to replace, the more intolerant it will be, probably
exponentially so.

So, here you seem to agree that it is just a matter of complexity.

Not at all. If you are watering plants with vinegar, it is not the
complexity which makes it a poor substitute for water. Complexity is
important, but it's a red herring as far as a living organism having
parts replaced. It's more about organic authenticity and similarity on
all levels.


we abstract from this in the conceptual theory. Such a complexity is
irrelevant. We are not addressing any practical issue here.

Just as having four prosthetic limbs would be more
of a burden than just one, the more the ratio of living brain to
prosthetic brain tilts toward the prosthetic, the less person there is
left. It's not strictly linear, as neuroplasticity would allow the
person to scale down to what is left of the natural brain (as in cases
where people have an entire hemisphere removed), and even if the
prosthetics were good it is not clear that it would feel the same for
the person.

Theoretically, this will be true only if your lower the level in the
infinitely down.

Or if there is no level at all. At what level can water be substituted
with something that is not wet?

Easy. At the molecular level. One H20 is not wet, I think.
If you doubt this, you might go at the level of strings, if you want. A string is hardly wet!

If the person survived with an artificial brain stem, they
may never again feel that they were 'really' in their body again. If
the cortex were replaced, they may regress to infancy and never be
able to learn to use the new brain.

Why? You need infinities to asses such truth conceptually.

Because you are replacing part of a tree that knows it's a tree. It
remembers and has expectations. If someone suddenly replaced your home
with a structure that looked the same on the outside but was
cinderblocks and asphalt on the inside, you wouldn't be able to go on
living as usual.

It means someone has made a substitution at a wrong level, not that such a substitution level does not exist.



You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-list@googlegroups.com.
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to