RE: mo' FIN

From: Marchal <marchal.domain.name.hidden>
Date: Mon Sep 3 07:26:09 2001

Jacques Mallah wrote:


> Actually I am still waiting to see the full UDA argument! I don't
>think you ever posted more than bits and pieces of it, without the precise
>definitions that I requested, and you referred people to papers written in
>French. But I'll check ...


The full UDA appears at http://www.escribe.com/science/theory/m1726.html

A better more recent version is in the conversation with Joel Dobrzelewski:
see the main links at http://www.escribe.com/science/theory/m3044.html
(note also that some people on the list seems to arrive independently to
similar conclusions).

(The older version in the list was called P-omega experience.


> I did explain my view, check the archive. But don't say that I believe
>in what you call "comp". I never claimed to believe it, mainly because I
>have yet to see a clear definition of it. I believe in computationalism,
>meaning that computations (effectively) give rise to consciousness. The
>word "survive" appears nowhere in any definition of computationalism I
> might give.


You *did* accept my way of presenting "comp" (saying yes to the brain
surgeon proposing an artificial digital brain/body made at *some* level of
description). You *do* give some meaning to survive, if only to prove us
that comp immortality (QTI) entails that we should find ourself older than
the expected age common for our species.


> More precisely, I don't see any 1st person view, other than as a
>description of what an observer-moment experiences.


That's not too bad, and it is enough for the UDA. (the translation
of the UDA in the language of a sound machine need a little more
rigorous definition, possible with G and G*, we will probably
come to that later).


>But the set of all such
>experiences is what the objective (which you might call 3rd person, but I
>don't) view describes. The objective truth has all the information, which
>we should guess as best we can.


Both the thought experiences and the incompleteness phenomena can
be used for explaining why the "objective truth" cannot have all the
information. This is important and can be made completely precise.
(something different but related happens in quantum self-reference,
cf Albert's work).


>You say "me" (and thus, the 1st person concept) can't be defined. Maybe
>you have some idea of what you mean by it, but if you can't define it,
>there's no way you could ever convey that information to another person.
>that case, you might as well stop posting.


First babies cannot define milk but can convey information about
milk to another person. It is naive to believe we must define
all the terms of our talk for using them. Second:

"me" can be *both* interpreted from a 1 and 3 person pov.
(pov = point of view).

"me" can be precisely defined from a 3 pov: through a bet on a level of
substitution.

The fact that "me" cannot be defined from a first person pov does not
entail that "me" cannot be made mathematically precise in a theory
of the 1 person.
I agree this is a subtil point (clearly ununderstood by people
like Penrose or Lucas).
We will see that the logic S4Grz is able to formalise the notion
of informal and unformalisable proof!!! The trick is made simple once
you realise the gap between truth and provability (captured by the
gap between G and G* in the modal setting).
My problem: logic is not very well known, but then I explain all
details when people asks me so.


>I have little doubt that you pull some questionable tricks
>in reaching that conclusion, but here we get into the technical
>part that you have never
>fully explained on this list, I believe.


Almost fully. The missing part are the non original part of my
thesis: mainly Godel, Lob, Solovay theorems. I have begin recently
a path toward such an explanation (diagonalisation 1:
http://www.escribe.com/science/theory/m3079.html).
I intend to go slowly but surely ...


>[...] You have never given a precise definition, but you
>always have the word "survive" in there, so maybe when you translate the
>implications of "comp" into precise mathematics you in effect are
>effectively assuming that conclusion from the start.


At some time I will explain that indeed there is no
possible precise definition of "survive". Still the word has a
folk meaning precise *enough* to understand that comp entails the
reversal psycho/physics.
If you don't understand the word "survive" you should remain
silent about any mortality/immortality question.
To believe in mortality = to believe there is an experience x such
that I don't survive x. To believe in immortality
is to believe that for all experience I survive x.
If you don't understand the word "survive" you should be agnostic
about FIN (your term).

Now I agree with you: it can be argued that the immortality is build
in in comp. No problem with that. You should show then comp being
false or inconsistent.


> It's not meant to be comical. Scientists always doubt that
>they really have the complete right answer, but one the other hand,
>it is much more certain to be able to identify a wrong answer.
> Newtonian mechanics is
>certainly wrong, but our QM may not be the whole story.
>(Yet I have little
>doubt that QM is basically right, for that matter.)
>FIN is certainly wrong.


I could understand that "FIN" (comp immortality or Quantum
immortality) is vague, weird, counterintuitive, etc.

But you are the one making the precise proposition "FIN is
wrong". You are the one who has to show a precise contradiction
from it.
You cannot at this stage just refer to imprecisions, for in that
case you would have told us that FIN is imprecise.

You really seems to want FIN wrong. I can understand. Unlike
George Levy, I am not sure I wish FIN to be true, I don't
know. The real question is harder with comp than with QM,
and also there is some sort of possible backtracking if you
take the amnesia effect into consideration which makes the
whole things still more difficult, with our current knowledge
of comp/QM. What happens clearly,
through the UDA for instance, is the fact that we loss
any form of certainty about mortality. That gives us just one
more doubt forcing us to be a little more humble on those
sort of questions(°).


> Where do I say I definitely haven't solved it?
> [...]

I have nowhere pretend I definitely haven't solved
the prime twins problem in number theory. This does not
entail I have solved it. I just say I have not found a solution
of your implementation problem in your web pages, nor in the list.


> Why? I will, for purposes of this argument, grant one concession
>however. To compare FIN vs. non-FIN, it is not acceptable to make the two
>categories of possible observations be something like A="I am younger than
>N" and B="I am older than that", where N = some arbritrary age that may
>(perhaps secretly) be based on my own such as twice my age. That would
>obviously be cheating to reject the FIN. I do not cheat.
> On the other hand if a more objective standard is used, like
>A = "I am younger than the 'thermodynamically unlikely' age for my
>species", or even better like A ="I am young enough to calculate (and
>store) my own age (perhaps normalized so that the typical lifetime of my
> species is set to
>1)", and B= the opposite, then Bayesian reasoning is perfectly valid and
>appropriate in this context.
> If the FIN were (cough) true, then people using this reasoning would
>usually (100% of the time in fact) be older than that, so it would almost
>always (100% in fact) work, except for a set of effective probability 0.
> If
>the FIN is not true, it would usually work, but not quite 100%. So I am
>erring on the side of letting very old people believe the FIN. Luckily I
>have other arguments against it.

Where?

The preceding move need a non physical definition of thermodynamics or even
species, and an explanation how mind is "attached" to "physical" machine.
For doing that you need to solve your implementation problem.
Anyway, such ideas does not survive (!) through UDA. Please show me a flaw
in it. I suggest you follow the "Joel" version which, although a little
less rigorous, is much simpler and direct than the older one (in the
archive). (see links above).


Bruno

(°) I just see Saibal Mitra use the "amnesia" effect. That is ok,
but use of amnesia effect can make "immortality" trivial in the sense
that each of us can be any other person, like if there was only one
person.
I recall that the "number of 1-person" is still an open problem,
and probably needs refined definitions of a lot of terms we are using.
But "comp entails FIN" works with the actual definition, and it is
unpedagogical to define a term more precisely than it is needed
in a course of a reasoning.
Received on Mon Sep 03 2001 - 07:26:09 PDT

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:07 PST