On 05 Jul 2012, at 15:47, David Nyman wrote:
On 5 July 2012 08:25, Bruno Marchal <[email protected]> wrote:
The proof if by absurdo. Suppose there is an algorithm, or even just
a God capable of predicting the specific outcome among "1)" and
"2)". Suppose it is "1)", then the guy in Moscow refutes it, and
comp invites us to listen to him. If it is "2), then the guy in
washington refutes it, and comp invites us to listen to him.
Indeterminacy has not been assumed: it results from the trivial fact
that I am copied in the same state in two different place so that I
can't predict which differentation will occur from my first person
perspective.
This is where it may be more explicit - and hence clearer - to
express the logic in terms of Hoyle's heuristic, in my view. To
remind you, a unique first person locus is first assumed,
Is that different from the unique first person that we have to attach
to any universal machine?
all experiential content then being the consequence of a unique
stochastic multiplexing of the entire class of sentient moments,
from this singular perspective.
This seems like adding an ASSA on top of all RSSA. I fail to see the
simplification, but if it helps you why not. Do you think we can make
an experimental test showing a role to such an ASSA?
The "objective" substrate, on which sentience may be supposed to
supervene, is assumed to be deterministic, as is the experiential
multiplex, whose role is to make explicit a means of conceiving an
entire class as a serialisation. But this latter conceptualisation
allows us to bring an explicit indeterminism into the picture,
against what appears as an otherwise entirely deterministic
background.
But we can already justify the relative indeterminacy of the relative
first person perspective, from what is an entirely deterministic
background.
An explicit stochastic multiplexing of moments mandates that all
occasions of sentience must take their place in due course and in
due measure, relativised to whatever "personal histories" may be
recoverable from the internal logic of the deterministic substrate.
But then such stochastic process will interfere with the outcomes of
duplication, and transportation, at least to make sense. But then it
might be in conflict with computationalism.
There are several useful aspects of this way of thinking, I
believe. Firstly, it takes seriously the singularity of first-
personal location.
I try to assume the less, and it seems to me that for UDA this is
explained by the fact that machine have singular self-referential
memories. Then AUDA confirms this singular aspect with the use of the
classical theory of knowledge.
Secondly, it makes explicit a generalised first-personal
indeterminacy as a fundamental characteristic of experience.
I fail to see how this is needed, nor even if it makes thing simpler.
It is already hard for most to understand an indeterminacy whose
existence follows from deduction, so another one imposed on the top of
it might seem making the picture even more complex.
Thirdly, and importantly, it prises apart two distinct aspects of
temporal experience: "replaces" and "logically prior or next".
Stochastic multiplexing of moments mediates the former aspect: this
moment - the moment "as given" - replaces all others in experience.
It is only in the context of the moment as given that logico-
temporal ordering becomes relativised to a particular personal
history, as mediated by the deterministic substrate.. Thus Bruno's,
John's or David's occasions of experience become relativised to
their particular reference (or relevance) class through, in effect,
the selective memory of a more generalised stochastic "process".
You might have a correct intuition, and I can intuit something, but
all what I do, is to derive the consequence of surviving with
probability one in case of technically successful brain transplant.
What is the relation between the new stochastic process and the global
(on arithmetic) comp one.
Perhaps the stochastic process you describe is the same as the comp
one, applied to the virgin universal machine, more exactly to its
first person pov. "Total ammesy" would lead any one to that state.
But we don't have to assume it. Universal numbers exists, and their
domain of indeterminacy contains all possible experiences, except that
the probability of going amnesic and getting my personal memory back
might be a statistically rare event (I dunno, some experience reports
can be amazing on this point).
In this way, each of us can speak meaningfully of "my next expected
moment", or "my prior history" by exclusive reference to the memory
state defined within a given "spatio-temporal" location.
But machines/numbers can talk of their next expected moment without
invoking any stochastic process. Stochasticness appears because their
relative consciousness is related to the infinities of computations
existing in arithmetic.
It is curious at first to apply this heuristic to the UDA and
similar scenarios, though ultimately simplifying, in my view.
Firstly we must put aside any "fundamental" notion of past and
future except as a logical implication from the perspective of a
momentary localisation within a stochastic experiential multiplex.
It looks like the global first person indeterminacy of the universal
machine, in front of the (arithmetical or physical)) UD.
We must however continue to take personal history seriously - and
not only "our own" - because any other attitude would be bad faith.
From this perspective we see that John's description of the
objective situation after copying is perfectly reasonable and true
of the deterministic substrate, and indeed the class of all sentient
moments considered as a whole.
Locally. With that protocol. Yes. It is the 3-view on the 1-views.
It ceases, however, to make any sense at all in the context of the
present heuristic, in which first-personal experience is explicitly
recovered from a unique perspective ...
It seems to me that this happens, by definition, of the first person
experience, even when just approximated by the personal memory, or the
knower.
... by the mutual replacement of singular givens.
I am not sure if I understand. Is that not what is guarantied by comp:
belonging to (at least one) computation, then the singularity is
brought by the fact that the machine identify herself with its local
mental accessible self-knowledge, with personal diaries. And the
physics emerges from its ignorance on the computations below its local
substitution level, or local universal bearer.
This heuristic provides a view of the first-person as a singular
stream of consciousness in which all personal episodes emerge in due
course, in due measure and in due relation.
What I try to convey is that if computationalism is correct, the laws
of physics emerge from a statistics on (indexical view on) the
computations. The goal consists in making comp testable, and the meta-
goal is to illustrate that we can do theology following the usual
standard of rigor. (which amount to say we know about nothing but can
try making hypothesis and reasoning).
At first sight the status of your supplementary stochastic hypothesis
is unclear to me, with respect to the comp theory. I feel I miss
something to understand the motivation of your heuristic/supplementary
hypothesis. What you provide above seems to me already explained, at
least assuming the derivation of physics works fine before deciding
that comp is plausible. Here, Gödel, QM and the MW seems already to
conspire for the confirmation of the weirder aspect of comp: that our
consciousness is not related to a singular computation, but to
infinity of computations going through our state.
I am not sure I can make sense of an absolute probability on *all*
moments. In fact a "moment" is already an indexically first person
defined notion, so I fail to see what it could mean "in the absolute".
I am afraid you have to do some work to make this more precise,
especially in relation with the machines (relative numbers).
Keep in mind that I am a logician, and I like to assume the less, for
making the conclusions available in the greatest possible classes of
interpretations. Adding a supplementary assumption can help to provide
models (interpretations) in which we can make more sense. You might
need to justify the relative consistency of your hypothesis with
respect to comp, though.
I remember you are motivated by the question "why me?", but as I told
you before, I think that question is nonsensical, in the sense that we
can explain why all löbian machine will rightfully ask the same
question, in some genuine way. From inside eventually the question
becomes "who's me?", which comp explains admits both one very singular
unnameable answer + an infinity of different answers, which
(hopefully) can cohere in relatively sharable dreams (the
"multiverse(s)").
Try to make more precise what I, and the universal machine, might be
still missing, if we are. If you propose a different interpretation of
comp, you might also elaborate.
Not sure it can help John, but I wish being wrong on this.
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 [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en.
