On 22 Dec 2011, at 15:40, smi...@zonnet.nl wrote:

Citeren Bruno Marchal <marc...@ulb.ac.be>:


On 21 Dec 2011, at 23:08, smi...@zonnet.nl wrote:

Citeren Bruno Marchal <marc...@ulb.ac.be>:


On 21 Dec 2011, at 14:06, David Nyman wrote:

On 21 December 2011 09:58, Russell Standish <li...@hpcoders.com.au > wrote:

Because Maudlin assumes a single universe physics,

Where? It assumes only the Turing emulabilty.

Its the only way to get inactive parts, and so force the absurdum. The
assumption is not explicit in Maudlin's work, but its there.

Russell, isn't it central to the multiverse view that distinct,
univocal observer experiences supervene on each branch?

I think so. But I think that Russell's point is logical. He tries to imagine a notion of physical reality which would make the MGA invalid. So he imagines a notion of multiple worlds physical supervenience. But this would contradict comp if the use of those other worlds is not Turing emulable. If it is, then we can emulate it in a single reality, and MGA can be reapplied.

Bruno

The observer cannot be located in a (fine grained) branch, as I explain here:

http://arxiv.org/abs/1009.4472

OK. But for the reasoning we don't have to localize an observer in a branch, we just need to conceive a program run by a single computer, even if that computer run other programs.

Bruno


I think you can get around this issue (I'm quoting what you wrote to Russell):

"The 323 principle *is* that later case. It says explicity that if a program does not need some register FOR A SPECIFIC COMPUTATION (supposed to be conscious, with sup-phys) then that specific computation will remain conscious when the register is removed.

I was correct that you were arguing against 323.

You have to explain how an inactive machinery has a physical activity in a computation."



Let's call the precise computational state of the computer a microstate. Then it is clear that the MGA applies to microstates, so you get into trouble when you assume that some specific evolution of the computer defined in terms of its microstates can represent some conscious entity.

OK.


Suppose that a computer is supposed to be conscious of shapes of objects. It takes as input millions of pixels and then it can recognize if there is a circle, triange or some other shape or nothing at all. Then, if it were to recognize a circle, then there are still an astronomically large number of different inputs that would have led to the same experience. The circle could have been mildly out of shape and the computer would still have recognized exatly the same thing. For the computer to recognize an "out of shape circle" as a different shape than a "perfect cirle", one would have to add an extra programming code for that.

OK.


Then if the computer is assumed to be conscious of recognizing a circle, then given that the exact pixel state can be any one of an astronomically large number of states, the exact quantum state of the environment-computer system should be some entangled state of the form: sum over inputs mapped to a circle of:

|input from environment>|computational state>

OK.



You can then associate this state which involves a huge number of microstates as representing the conscious experience of the computer recognizing a circle. The MGA then could perhaps allow you to remove inactive registers involved in the process of recognizing a triangle.

Yes.


But you will fail to remove register after register so as to make recognizing the circle a trivial movie where there are no counterfactuals. If you were to attempt to do this, at some point you would be removing terms from the above state.

Then the evolution of |input from environment>|computational state> is no more Turing emulable. If it is, I can emulate it in one boolean graph computation (like the UD will necessarily do in particular), and MGA will be applicable again. OK?


The interpretation of this is that to a superobserver who can be aware of any specific microstate the computer is in, that computer is not conscious. The superobserver is located in a much narrower branch.

The consciousness of an entity should not depend of the observer of that entity. Given that we can apply the MGA on the entire sum above, it seems to me that this leads to the idea that consciousness does not supervene on any particular physical implementation, but on the abstract relation which makes a computation a genuine, albeit immaterial, computation. Then we have to recover the material appearance from a mathematical theory of computation, number self- reference, etc.

Bruno




Saibal






Saibal


In which
case, isn't it correct to apply Maudlin's argument to each branch
separately? If so, to oppose the conclusion by appealing to all the branches simultaneously might seem like wanting to have your cake and
eat it too.

David


On Tue, Dec 20, 2011 at 09:06:10PM +0100, Bruno Marchal wrote:

On 20 Dec 2011, at 01:03, Russell Standish wrote:

Even though the parts may be distributed across multiple branches of
the MV, and have different counterfactual histories?

?

What is a branch of a W in a MW if you allow interaction between worlds?

Who said anything about interaction between the worlds? I assume by
interaction, you mean the usual interaction physicists means
(interference), or information being passed.

Comp, be it digital or quantum, makes classical computation non
interacting with parallel computation. Locally, if our brains were quantum computers this would be locally false, but not in a relevant way to contradict the MGA consequences, by the fact that if worlds
interfere that much still does not violate Church thesis, and
quantum computer are Turing emulable.

We're not discussing quantum computers here.


Are you arguing that comp does not entail the principle "323"?

I don't believe so.




Be careful of not including the conclusion in the definition of COMP.

...snip...


If the emulation is by means of dovetailing, then I think not. A
dovetailer is not conscious.

That is ambiguous. A dovetailer, like Robinson arithmetic (when
proving all its theorems) is not conscious, per se. But it
instantiates consciousness, indeed all possible machine's
consciousness.


Great. But this is more than mere terminological wrangling. To an
observer of the dovetailer, no conscious processes are visible.

No conscious processes are ever visible.




To do
that would require a means of determining whether a computation is conscious or not, something we don't have, and probably never will.

That's always the case. You judge by chatting with person or by
observing them and recognizing yourself.
That's how I became open to the idea that all löbian entities
(machine or not machine) are conscious.




It is another manifestation of there being no God's viewpoint in a
Multiverse.

I am not sure about that. Many would say that the very idea of
"multiverse" is an attempt of describing "God's viewpoint".
With mechanism we have universal dreams and universal dreamers,
sharing, or not, dreams and subroutines.


It cannot be a God's point of view. The Multiverse is too simple to
admit an observer...



I feel this invalidates applying Maudlin's argument to a
dovetailer.

Let me introduce a new definition. I define a closed generalized
brain (CGB) the portion of reality that you need to emulate a dream.

This may require the input of random numbers on the synapses. It seems to me that dreams are the result of filtering and amplifying random thermal noise with the brain. It is just a theory, of course, but it
would mean that the CGB is a Multiverse.

Many neurophysiologists would be that such a portion of reality is in the skull, and that the process is Turing emulable (and I think
it your position).

Sure, but the contents of the skull is an object that extends over
multiple branches of the Multiverse.

Comp implies that such CGB exists. That CGM can
be emulated by a turing machine, why would it matter the emulation
is done by dovetailing from the first person point of view?

Because in the 1st person POV, the "inert" parts are not inert. Only in the 3rd person dovetailed POV. And, I find it hard to think of the
dovetailer as conscious.


But put another way, perhaps it means that consciousness
cannot supervene on a physical implementation of a dovetailer.
Which
is probably what you're trying to get to.

I just reason from the assumption. Consciousness would supervene on the execution of a physical universal dovetailer. Why wouldn't it?

Because the dovetailer is an incredibly simple program. It hardly
seems conscious. If I ask it a question, it is mute, so the Turing
test hardly helps.

...snip...

It doesn't eliminate the supervenience of the consciousness on the simulated physics within the UD. It seems this is in accordance with
Brent's comments too.

Presumably you would argue that this is simulated matter, not
primitive matter. Sure.

You lost me here. The "primitive matter" in comp is not *a priori*
simulable, it appears below our sharable substitution level.


It may or may not be simulable a priori. Why would a materialist
assume that primitive matter is necessarily nonsimulable?



But what's to stop the primitive matter being
multiversal - whether it can be simulated or not is a little beside
the point.

On the contrary it is crucial. It makes the difference between
emulable in one reality (in our branch of the quantum multiverse in
case we imagine a concrete one), which is equivalent with Turing
emulable, or necessitating Non Turing emulable interactions or
interferences with parallel realities.

I don't expect there will be interference between the realities. Why
does supervenience over multiple branches entail there must be
interactions between realities?

The point is that if it is
Turing emulable, then the MGA applies.

I don't see this.

You have then to believe that
a physical inactive piece has a physical activity relevant in a
particular computation.

The physically "inactive" piece is only physically inactive in one branch. If the supervenience is across multiple branches, then the
absurdum is no longer.


...snip...

If a dream can supervene on a closed generalized brain Turing
emulation, then it has to supervene to its emulation in one of its
classical instantiation, either in a concrete quasi-classical
(normal) history (or, after MGA, in arithmetic, or in the UD*).

Why?

And
in that single reality emulation, MGA can be applied. If you give a
role to physically inactive, by making them active in some other
world you are forced to introduce a non Turing emulable *physical* component in matter playing a role in consciousness, where comp show
that we get it for free below our substitution level.


Why?

...snip...


The UD is not conscious, as a person, but once you add the
supervenience thesis, it instantiates consciousness at each moment where it executes a conscious program (say Russell Standish's one,
then Bruno's one, etc).

Why does this depend on supervenience?

I meant "physical supervenience". It is just introduced to get the
contradiction.
MGA is a reductio ad absurdum.

How does this work?


Because Maudlin assumes a single universe physics,

Where? It assumes only the Turing emulabilty.

Its the only way to get inactive parts, and so force the absurdum. The
assumption is not explicit in Maudlin's work, but its there.

...snip...


You might try to refute the 323 principle as clearly as possible by using a *physical* multiverse. I think you will see by yourself that
you have to endow some primitive Matter with some non Turing
emulable processes at some point.

I don't see the 323 principle as being relevant here - perhaps you can
explain more why its needed.


--

----------------------------------------------------------------------------
Prof Russell Standish                  Phone 0425 253119 (mobile)
Principal, High Performance Coders
Visiting Professor of Mathematics      hpco...@hpcoders.com.au
University of New South Wales          http://www.hpcoders.com.au
----------------------------------------------------------------------------

--
You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to everything- l...@googlegroups.com . To unsubscribe from this group, send email to everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


--
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 everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


http://iridia.ulb.ac.be/~marchal/



--
You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to everything- l...@googlegroups.com . To unsubscribe from this group, send email to everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


--
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 everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


http://iridia.ulb.ac.be/~marchal/



--
You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to everything- l...@googlegroups.com. To unsubscribe from this group, send email to everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


--
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 everything-list+unsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/everything-list?hl=en .


http://iridia.ulb.ac.be/~marchal/



--
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 
everything-list+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/everything-list?hl=en.

Reply via email to