On 12 Jun 2012, at 22:48, David Nyman wrote:
On 12 June 2012 17:36, Bruno Marchal <marc...@ulb.ac.be> wrote:
Yes, but the expression "from the current state of any universal
machine" (different sense of universal, of course) already *assumes*
the restriction of universal attention to a particular state of a
particular machine.
But is that not the result of the fact that each machine has only
access to its own configuration?
That's too quick for me. To say that "each machine has only access
to its own configuration", is still merely to generalise; to go from
this to some particular machine requires one instance to be
discriminated from the whole class. So what, you may retort, "your"
states just "discriminate themselves as you". The problem to my
mind, with looking at things in this way, is that for there to be a
universal knower, each state must primarily belong to "you qua that
knower" (which is what makes it universal) and only secondarily to
"you qua some local specification". If this be so, it is circular
to invoke those secondary characteristics, which become definite
only after discrimination, to justify the discrimination in the
first place.
ISTM that the two of us must actually be thinking of something
rather different when we conceive a universal person or knower. For
you, IIUC, this idea is consistent with many different states of
consciousness obtaining "all together"; consequently the viewpoint
of this species of universal person can never be reducible to any
particular single perspective. I'm unsatisfied with this (as
presumably was Hoyle) because it leaves me with no way of justifying
"why am I David" that isn't circular.
I think it is justified by the working of a computer and the computer
science definition of self. It explains self-reference in a non
circular way.
The question "why am I David" is twofold:
- One aspect is "trivial" and admit the same explanation as "why am I
in W and not in M" in the WM-duplication. Here comp can explain that
there is no answer possible to that question (first person
indeterminacy).
- The other aspect is less trivial and concerns the very ability to
refer to oneself. Here comp, or just computer science can entirely
explain what happens. I often explain quickly (perhaps too much
quickly) that such reference result from the double diagonalization
trick on duplication. Cf: if Dx produces xx, then DD produces DD. We
can come back on that, and will, on the FOAR list, most probably.
I can of course say that I'm David because the given state (here,
now) happens to be one of David's states of mind, but the problem in
this view is that this is completely consistent, mutatis mutandis,
with Bruno's saying exactly the same.
But that is why we can both exist, consistently. If that would not be
the case, it seems to me that one of us would be a zombie for some
moment.
By contrast, Hoyle's heuristic allows me to say I'm David because a
state of David happens momentarily to be the unique perspective of
the whole.
I don't see why the beam heuristic would explain that, nor why the
absence of the beam would prevent that. I just don't see.
The beam seems to reintroduce an external structure to the 3p reality,
which is hard to justify by itself.
If such a beam is conceivable, why could we not conceive a lot of
beams, perhaps as much beams as they are subjects, so that we can
disallow zombie?
What is the nature of the beam? How does it work?
Is the existence of the beam compatible with comp? How would the
doctor justify that the beam will target my digital brain state? Is
the action of the beam Turing emulable?
In a sense, the DU itself acts like a beam, given that it go through
each possible subjective state *one* at a time, but then the UD-time
itself is non material, and the states are linked only by their
relation with (all) the universal numbers linking them, and that makes
the explanation by comp complete, with respect to that qustion. So the
call for such a beam seems to me very similar to the argument that a
concrete physical universal dovetailer is needed for consciousness to
exist, which makes the explanation of physics by comp circular, and
then cannot work because it relies on a physical supervenience which
happens to be incompatible with comp (step 8). Here too the beam seems
to introduce a difficulty in both the physical supervenience and the
computational one.
As Schrödinger puts it, not a piece of the whole, but in a certain
sense the whole; Hoyle's heuristic makes explicit that "certain
sense".
I agree with Schroedinger. Given that "time is an illusion", I have no
problem with a universal observer looking through both our eyes, and
even being our common deep self simultaneously.
I suspect that the difference between us is that it is not your
intention to justify the feeling of change directly from your
mathematical treatment,
I am not sure. The logic of self-reference ([]p, in G) leads to the
first person ([[]p &p), S4Grz), which can be used to explain why the
subject feels that things are time-asymmetrical, why knowledge
evolves, etc). Adding a beam would makes all machines states, when not
under the beam, into sort of zombie states.
but rather to demonstrate the existence of an eternal structure from
which that experience could be recovered extra-mathematically.
OK, but no more extramathematical than the notion of arithmetical
truth is for arithmetical machine. That "extramathematics" is the same
as the one use implicitly each time we grasp the notion of some
infinite set, like with the statement, let N be {0, 1, 2, ...}. It is
minimal and unavoidable. And withoit doubt mysterious, but that is why
we have to assume arithmetic explicitly.
You often refer to the inside view of numbers in this rather
inexplicit manner (forgive me if I have inadvertently missed your
making the details explicit elsewhere).
May be you are just missing the logic of self-reference. The inside
view of numbers is easily conceivable once you assume comp, and accept
to attribute consciousness to statical computations as they are
defined by the arithmetical relation. That is why we need so few math
in the UDA. But for those people who feel uneasy with cognitive
science and thought experience, it is the AUDA which makes the whole
picture arithmetically consistent, by providing the explanation of how
machines are able to refer to themselves, in a third person way first
(G, G*), and in a first person way (S4Grz). The feeling associate with
them are given by other modalities (like Bp & p & Dt), which are
unavoidable for any sufficiently rich (cognitively, Löbian) machine,
by the incompleteness phenomenon. Without this, I would have abandoned
comp since long.
Hoyle however seemed to be directly concerned with rationalising
this feeling by associating it with a unique dynamic process
operating over the system as whole.
If only I could figure out what it does, and how it does it.
That's the difference, I think, and it may be irreconcilable.
I am not sure what you really mean here.
Is it irreconcilable with the comp hypothesis? I tend to think so,
because I don't see the need to add in a program some call to the beam
(subroutine, oracle?) to make the universal person conscious through
it. The implementable reference to oneself seems to play that trick.
Or the UD and arithmetical truth is "beam" enough, so to speak.
Is it irreconcilable with the UD Argument? I tend to think so, because
it introduces a non mechanical element, and a curious, very weird,
form of supervenience.
(If the beam is mechanical, then it is just a question of substitution
level, I guess).
I am aware you are using it only as an heuristic, but as I said, I
don't see what it solves. It looks like it would only reintroduce a
"real" time for consciousness in the picture, and that it reintroduces
mystery, to account for what should be, with comp, only an illusion,
like the fact that "her, now and me" would have a sort of absolute
status, which I think has only a relative existence, even if we feel
the contrary "here and now".
Bruno
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.
