On 19 Aug 2014, at 17:59, meekerdb wrote:
On 8/19/2014 3:10 AM, David Nyman wrote:
On 19 August 2014 01:53, meekerdb <meeke...@verizon.net> wrote:
So the idea is that comp necessarily entails epistemological logics
(the "dreams of the machines")
Except that it seems to be an epistemology very different from ones
we usually practice. What's the last time you learned a fact about
the world by proving it from Peano's axioms?
Well, you frequently counsel me against arguing on the basis of
personal incredulity. So I could respond by asking you when was the
last time you learned a fact about the world by deducing it from
the molecular structure of your brain.
But that's not my theory of epistemology. I infer the existence of
my brain and molecular structures via a very long chain of
inferences and hypotheses starting with my perceptions.
OK. that is a definition of []p, if you bet on comp, so that your
representational believes are hardwired and
evolves obeying some induction axioms. Unfortunately, you can be
hallucinated (G* proves <>[]f), so to get knowledge, we just add "God
knows it", which we locally emulate by the sentence p.
Bruno defines []p & p as knowledge, but that doesn't show any way of
getting knowledge except []p,
Not at all. []p alone is only belief. Bye incompleteness, the meta-
definition []p & p changes the game completely. the knower is no more
even a machine from its points of view. Like Alice wandering in some
forest, we lose our name, there. We lose any 3p description, and that
is why the saying "yes" to the doctor will ask for some act of faith.
i.e. proving p from axioms (and p happening to be true, i.e. the
axioms are true).
We never know that. I agree we can be pretty sure for the axioms of
first order arithmetic.
So the epistemology is either mathematical proof or it's left to the
hoped-for "statistical mechanics" analysis of the UD.
Only the doxastic (belief) is mathematical, and semi-computable
(computable when true/converging).
The first person defined by 1p, is for us, beyond all theories and
realms. But a rich Löbian machine can still study the theology of
simpler machine, and she can know that it applies to her if God will
(if she is correct).
Then the UD gives the restriction on the sigma_1 sentences, in the
arithmetical interpretation of the modal logics, and that is used for
the FPI.
[]p & <>t gives the proba, not knowledge, that is why []p & <>p & p,
will still add a nuance (the feeler).
I guess we will come back on this.
Bruno
Brent
Given that we are committed to explaining the complex in terms of
something simpler, then some sort of structure, defined molecularly
or otherwise, must surely be implicated in what it means to learn a
fact, even though we can't yet say precisely what it is.
I guess I take the logics that Bruno investigates in AUDA to be at
something analogous to the "molecular" level vis-a-vis any
explanation of cognition or perception that would strike us as
intuitively familiar. So just as an understanding of the dynamics
of molecular bonding has turned out to be crucial to an
appreciation of the possibilities of large-scale structure, the
hope (or project) is that we can derive something of analogous
relevance, to the structure of human-like cognition and perception,
from a rigorous study of particular classes of more basic logical
relations.
that are *prior* to physics in the sense that only certain sub-
classes will be characterised by the statistical dominance of
physically-lawlike relations over their range of reference.
It's pretty much like Bertrand Russell's neutral monism. There are
events or states that classified one way constitute experiences or
thoughts of individuals, and classified another way, some of them
constitute objective physical events.
That's not a bad way of putting it as a general position. My point
though was that if we want to start from a very general notion of
computation that doesn't presuppose physics, we must seek to
justify the differentiation of a sub-class of lawlike physical
realities from a much larger totality. According to comp, this
differentiation is rooted in the statistical dominance of certain
classes of internal "belief" or reference that are
deducible from a quasi-ubiquitous form of self-referential "machine
psychology". I guess it is only to be expected that a fundamental
concept of this sort would strike us as being at some remove from
any putative elaboration at the human scale. The devil, as ever,
will be found in the detail.
David
On 8/18/2014 4:23 PM, David Nyman wrote:
On 18 August 2014 23:27, Stathis Papaioannou <stath...@gmail.com>
wrote:
I'm not entirely clear on Bruno's argument on this last point. The
way
I see it, if a brain is simulated by a computer program, what is
being
simulated is the physics; and if comp is true, that means that
simulating the physics will also reproduce the brain's
consciousness.
I'm not sure about computations instantiating consciousness without
instantiating physics, and I'm not sure how instantiating the
appearance of physics is different to instantiating (virtual)
physics.
I've always understood him to be saying, in the first place, that
the dovetailer necessarily generates certain classes of self-
referential computations. Very generally, such computations are
then regarded as emulating self-referred (i.e. first-personal or
indexical) logics that in turn are amenable to treatment as
"beliefs" in realities or appearances. So the idea is that comp
necessarily entails epistemological logics (the "dreams of the
machines")
Except that it seems to be an epistemology very different from ones
we usually practice. What's the last time you
learned a fact about the world by proving it from Peano's axioms?
that are *prior* to physics in the sense that only certain sub-
classes will be characterised by the statistical dominance of
physically-lawlike relations over their range of reference.
It's pretty much like Bertrand Russell's neutral monism. There are
events or states that classified one way constitute experiences or
thoughts of individuals, and classified another way, some of them
constitute objective physical events.
Brent
I've always assumed that it's this logical priority of "machine
psychology" over the subsequent appearance of lawlike physical
relations that constitutes the postulated "reversal".
David
--
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.