On 15 Dec 2016, at 22:02, Brent Meeker wrote:
On 12/15/2016 7:47 AM, Bruno Marchal wrote:
On 14 Dec 2016, at 23:49, Russell Standish wrote:
On Wed, Dec 14, 2016 at 05:23:16PM +0100, Bruno Marchal wrote:
On 14 Dec 2016, at 02:12, Russell Standish wrote:
I don't see why you would say physicalism needs to be assumed to
explain the predictive power of physics.
To predict (exactly and in principle) something physical you have
only one way: to compute the relative FPI on UD*. (Which is
obviously highly non-computable, as even "W" or "M" is already not
computable in the simple self-duplication, with "W" and "M"
refering
to the experiences of finding oneself in Washington and finding
oneself in Moscow respectively (step 3 of the Sane04 slide).
Predictions are never exact in 100% detailed, so running a
dovetailer is not
necessary.
It is run in arithmetic, and that is what we have to take into
account to explain that we need physicalism to get physical
prediction when assuming mechanism.
Probabilistic predictions are just fine too.
OK. But they are based on some theory. Lottery assumes balls and a
bit of mechanics for example. The problem is that if we assume
mechanism, we cannot rely a priori on the physical laws. We are
under the global FPI.
So for any
class of system (presumably containing our world to be of interest),
?
I was just answering your question above.
I think you digress and talk about prediction when assuming some
world, which is not available when we assume mechanism.
there will be some properties that remain constant, or will change
in
predictable (ie mechanistic or computable) ways. Mostly we have
just a
model (physicist's model, not logician's) to work with -
OK. That is: a theory.
which of
course brings to light the problem of induction that the model
needn't be faithful to the system being modelled.
We are searching the fundamental theory. Not doing prediction, but
explaining how physical 3p prediction can be assessed by a digital
machine, which necessarily belongs to infinities of computations
(with infinities of inputs, oracles, etc.).
To get a special physicalness or a physical universe, capable of
selecting some special computations on all computations which go
through the actual state of the guy doing the physical test, you
need to invoke some non-computable element, different from the
statistics on all computations, (which, as I just said, is not
computable).
None of this is required to get predictive power.
I was just explaining that with mechanism, physicalist physics does
not make sense. See your question above.
Models needn't have
any ontological status - the vast majority of physical model are
_known_ not to have ontological status.
Now, the probability distribution might be computable, or the logic
of the "certain events" might be axiomatisable, and indeed, is (by
S4Grz1; Z1* and X1*).
Particularly when the
whole induction process is explained quite neatly with the
Solomonoff-Levin universal prior and Bayes theorem over a
multiversal
set of events that naturally arises in the context of
computationalism.
Unless the universal prior is based on the assumption of a unique
physical reality, that makes my point.
?
If you derive the multiversal
set of events from computationalism, physicalism needs to add
something which has no role at all, from the computationalist
perspective, and yet has to have some role to not contradict the
"yes" doctor, or it has to bring some strange actions from some
object having no interaction with a machine (like in the movie
graph
or Olympia).
We seemed to have diverged from predictive power of physics to
physicalism? Why?
?
Look at your question. I quote it "I don't see why you would say
physicalism needs to be assumed to
explain the predictive power of physics.".
The whole problem is that, when we assume mechanism, physics can
have a predicting power only if the measure on the relative first
person experience of the machine realized in UD* (alias the sigma_1
complete part of arithmetic, alias all computations) is given by
the probability on those arithmetical experience.
... rest snipped as it is along the same digression ...
I don't think so.
The question you asked was (I quote):
I don't see why you would say physicalism needs to be assumed to
explain the predictive power of physics.
Let me try to explain again.
How do a physicist make a prediction about his future first person
experience?
To fix the things, why am I pretty sure I will fell like seeing an
eclipse when predicted by Newton's law.
The usual materialist/physicalist answer is roughly like this.
There is the assumption of a physical reality(*) and that it
contains or realized objects obeying laws. We assume that our first
person experience is related or attached or realized by our brain.
Then we assume that during the evolution of the object of that
reality, our first person experience remains connected to the brain
of the observer interacting with those objects, so that by using
the laws of evolution of the objects he can predict its personal
feeling he will get when interacting with the objects.
We basically do that all the time. OK? And I am pretty sure it is
the good things to do FAPP.
Now, but this is a digression, in quantum physics, this picture
needs already to be quite nuanced, indeed, pretty much like
Mechanism will show it has to be.
Indeed. With mechanism, we know (in the sense of true and
justifiable) the existence of all computations,
I don't think we know that. I think you are equivocating on
"existence".
First logic is used to avoid any equivocation on "existence", and
avoid any form of ontological commitment, as we do in science (even in
metaphysics if we do it with the scientific method). I use "exists" in
the sense of classical first order logic. No metaphysical baggages.
The basic existence is then the existence of the interpretation of the
basic term, which in the theory are 0, s(0), s(s(0)), ...
The non basic existence, which are phenomenological are given by the
modalities of self-reference, and take shape like []Ex[]P(x) or many
others.
and no universal machine can distinguished from its first person
pov which universal machine(s) run it. On the contrary, the
universal machine can deduce, from the assumption that she is a
machine at some substitution level, that below that level: *all*
universal machines participate in emulating the computations going
through their state.
In this case, when we *assume* Mechanism to be precise, evoking a
physical reality, or any *special* reality (Matter, God, etc.),
becomes a theological or metaphysical or philosophical error. It
consists in adding a selection principle which assures that the
first person is more "real" in some computation(s) than all its
dreaming counterparts in the Turing complete part of arithmetic. So
the only way to be in a stable physical reality, is that the
physical reality, or the observable, becomes some statistical sum
of all computations going through our state.
This might apply to all representational theory of mind, be them
mechanist or not, as this phenomena extends in the transfinite.
Now, in an non-mechanist and non-representational theory of mind, I
think you can escape that problem (but then you might decide to
ignore the origin of the physical laws) by adding some special
actual infinities in the brain and in the observed objects and in
their relation to assure some one-one absolute correspondence, like
Searle did somehow, but that seems premature to me. Why not, and it
is a merit of machine's theology to invite to possible precise
mathematical non mechanist theory of consciousness.
But you see that with Mechanism, physicalism needs to make a
physical universe into a sort of God capable of making some
computations (of your brain), or some subset of computations (of
your brain), more "real" than all computations (of your brain).
Those computations exists independently of all observers in the
sense that we agree that x+2=4 admits a solution independently of
all observers.
But that's not the sense in which our experience exists.
You are right. It is a phenomenological existence. It emerges from
special modal relations that numbers can have with other numbers. But
computations are just numbers having some properties with respect to
(universal) numbers.
Bruno
Brent
(Then the logic of the "measure one" is given by []p & <>t (& p),
with p sigma_1, and up to now, it works, as we get (3) quantum
logic(s).)
Bruno
(*) 'which make it primary, by definition. (primary relatively to
the theory T mean assumed by the theory T). With computationalism,
you can take as primary any terms of a Turing-complete theory, be
it lambda calculus, combinators, or natural numbers. Of course, if
you choose the combinators, the natural numbers are no more
primary, but you still get all computable number relations, and
vice-versa, so, by abuse, as they are Turing equivalent and define
the same computation space (universal deployment) I take them as
being as much primary ...
--
----------------------------------------------------------------------------
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders
Visiting Senior Research Fellow hpco...@hpcoders.com.au
Economics, Kingston University http://www.hpcoders.com.au
----------------------------------------------------------------------------
--
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 https://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 https://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 https://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
