On 16 Mar 2011, at 22:40, Russell Standish wrote:
On Wed, Mar 16, 2011 at 06:03:12PM +0100, Bruno Marchal wrote:
On 16 Mar 2011, at 09:53, Digital Physics wrote:
You are right, the "theory of nothing" is an old hat, published in
the 1990s. Hutter's new contribution is the observer localization:
how many bits are necessary to identify the observer's location in
the multiverse? Random locations cost many bits. Non-random ones
are much more likely and therefore more predictive in the Bayesian
framework. This "allows to distinguish meaningful from
predictively meaningless theories".
Marcus Hutter, like Schmidhuber, seems to be not aware of the mind
body problem, notably as formulated in the computationalist theory
of mind. His self-localization used implicitly the mind-brain
identity thesis, which cannot be sustained in any computationalist
framework. I already give you the references on sane04.
Please don't hesitate to ask questions if you find something unclear
in it.
You will understand that the notion of self-localization of "myself"
in a universe or a multiverse does not make sense, except in a
precise *relative* way which needs the use of the classical theory
of knowledge of Plato (Theaetetus), and its translation into
arithmetic (something which can be done by using a technic due to
Gödel) to be defined precisely. This makes computationalism
predictive and falsifiable. The key idea consists in distinguishing
the logics of the first person views (and the first person plural
views) from the third person views.
Bruno
Yes, he is a little biased by the Kolmogorov complexity result that
random strings are the most complex. The problem is in treating all
random bits as significant, when usually they aren't. I argue (On
complexity & emergence, 2001) that the relevant complexity measure
takes into account the significance of the information to the
observer, which really means that random strings are amongst the
simplest.
Of course this touches on your point that one cannot localise an
observer to a particular bitstring, but rather to all consistent
strings (that have the same meaning to the observer).
I would say this is my main significant departure with Schmidhuber
and Hutter.
You are right. This is old stuff already refuted in this list a long
time ago by different people.
Then, as you say, we are obliged to restrict the measure on the
relative consistent extensions of the strings/theories/bodies, and
this forces us to take the self-referential logics into account,
together with their POV variants. This lead to the G* theory, and the
G/G* splitting, and their corresponding POV variants splittings, so
that we get a theory of qualia, with quanta as a particular sharable
case, making the whole theory testable, and rather well confirmed up
to now. Indeed most quantum weirdness (indeterminacy, non locality,
non cloning, etc.) are easy consequences of the comp assumption (and
even of much weaker similar assumptions).
The problem of Schmidhuber, Hutter and also Tegmark is that they
continue to hide or trivialize the person's consciousness, if not the
person per se, under the rug. They do use a more Platonic framework
than most materialists, which makes their work being a sort of
progress compared to purely physicalist approaches, but, like most
physicists, they are still blinded by the Aristotelian frames which
they continue to take for granted.
They missed the first person indeterminacy, which makes their position
hard to maintain. Not only they miss the qualia, but they have an
inconsistent (with comp) theory of quanta.
Such Aristotelian materialism/naturalism has been probably the root of
the apparent unsolvability of the mind body problem. Nagel says that
some 'revolution' is needed to solve the mind-body problem.
Computationalism shows that such a revolution is only a coming back to
Plato and Plotinus' type of conception of reality. This announces a
coming back of theology, soon or later, in the scientific curriculum.
A good thing, I think, because the lack of modesty and rigor in
theology is responsible for a lot of unnecessary suffering on this
planet. Authoritative arguments, conscious one or not, are bad in
*all* fields, and even worst as much as the fields are concerned with
fundamental questions.
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.
