On 26 May 2015, at 02:54, Bruce Kellett wrote:
Bruno Marchal wrote:
On 25 May 2015, at 08:58, Bruce Kellett wrote:
Part of my problem is that the UD does not execute any actual
program sequentially: after each step in a program it executes the
next step of the next program and so on, until it reaches the
first step of some program, at which point it loops back to the
start.
The UD does execute sequentially *each* specific program, but the
UD adds delays, due to its dovetailing duties, which we already
know (step 2) that it does not change the first person experience
of the entiuty supported by that execution.
So if the conscious moment is evinced by a logical sequence of
steps by the dovetailer, it does not correspond to any particular
program, but a rather arbitrary assortment of steps from many
programs.
?
Each execution of the programs is well individuated in the UD*. You
can descrbied them by sequences
phi_i^k(j) k = 0, 1, 2, ... (with i and j fixed).
But each step of the dovetailer is just a single application of the
axioms of the Turing machine in question: one of the
Kxy gives x,
Sxyz gives xz(yz),
for example. These single steps are all that there is in the
dovetailer. But such steps lack a context -- they make no sense on
their own. You could simply claim that the two basic steps are all
that is needed -- consciousness self-assembles by taking as many of
these in whatever order is needed.
?
The context will be given by the combinators. To dovetail universally
with the combinators, you need to generate them all: K, S, KK, KS, SK,
SS, KKK, K(KK), KKS, K(KS), ...
If comp is true, the combinators running your current brain states
will be executed, and probably with some rich context in most of them
(if not, and can prove it, comp is refuted).
If the next step of program phi_i is some 10^50 or so dovetailer
steps away, the only thing that could possibly link these is the
program phi_i itself -- the actual execution of the steps is
entirely secondary. In which case, one would say that consciousness
resides in the program phi_i itself - execution on the dovetailer is
not required. I do not think you would want to go down this path, so
you need something to give each step a context, something to link
the separate steps that are required for consciousness.
Consciousness is associated to the execution, not to the programs.
That would not make sense, even if the existence of the program
entails the existence of its execution in arithmetic. The relative
probabilities depends on the execution and the mathematical structure
which exists on the set of continuations (structured by the first,
third, ... points of view).
The teleportation arguments of Steps 1-7 are insufficient for this,
since in that argument you are teleporting complete conscious
entities, not just single steps of the underlying program.
?
Of course, given that all programs are executed, this sequence of
steps does correspond to some program, somewhere, but not
necessarily any of the ones partially executed for generating that
conscious moment.
Yes. So what?
I was trying to give you a way out of the problems that I have
raised above. If you don't see that the sequential steps of the
actual dovetailer program give the required connectivity, then what
does?
?
I do see that the sequential steps of the *many* computations give the
required statistical connectivity.
You did, some time ago, claim that the dovetailer steps gave an
effective time parameter for the system.
An infinity of them.
But even that requires a contextual link between the steps --
The UD brought them all.
something that would be given by the underlying stepping -- which is
not the stepping of each individual program phi_i.
It depends at which level you describe the happenings. The FPI makes
your subjective future statistically defined on all the UD*, by the
first person non awareness of the underlying stepping of the UD itself.
Bruno
Bruce
--
You received this message because you are subscribed to the Google
Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
http://iridia.ulb.ac.be/~marchal/
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
