On 2/10/2012 13:54, Stephen P. King wrote:
On 2/9/2012 3:40 PM, acw wrote:
[SPK]
I do not see how this deals effectively with the concurrency problem!
:-( Using the Platonia idea is a cheat as it is explicitly unphysical.
But physics by itself does not explain consciousness either (as shown
by MGA). Maybe I just don't see what the concurrency problem is.
It has no constraints of thermodynamics, no limits on speeds of signals,
no explanation as to how an Ideal Form is defined, e.g. what is the
standard of its perfection, ect. It is no different from the Realm of
God in religious mythos, so what is it doing here in our rational
considerations? Forgive me but I was raised by parents that where
Fundamentalists "Believers", so please understand that I have an allergy
to ideas that remind me of the mental prison that I had to work so hard
to escape.
I'm not asking you to share all of Plato's beliefs here. It's merely a
minimal amount of "magic", not unlike the "magic" you have to accept
by positing a 3p world. The amount is basically this: arithmetical (or
computational) sentences have truth values independent of anything
physical and consciousness/qualia may be how some such arithmetical
truth feels from the inside. Without at least some axioms, one cannot
get anywhere, you can't reduce arithmetic to only logic and so on. Why
would Platonia have to have the same constraints as our physical
realms - it need only obey to constraints of logic and math, which
usually means stuff that is contained within the Church Turing Thesis
and its implications. Speed of signals? If some theory is
inconsistent, it's only there as part of the reasoning of some other
machine. Ideal Form? How do you define an integer or the axioms that
talk about arithmetic?
Popular religious mythos tend to be troublesome because they involve
*logically impossible* properties being attributed to Gods and other
beings - things which are inconsistent. It's not like one doesn't
assume some axioms in any theory - they are there in almost any
scientific theory. Yet, unlike popular religions, you're free to
evaluate your hypotheses and use evidence and meta-reasoning to decide
which one is more likely to be true and then try to use the results of
such theories to predict how stuff will behave or bet on various things.
Of course, it's not hard to get trapped in a bad epistemology, and I
can see why you'd be extra skeptical of bad theories, however nobody
is telling you to believe a theory is true or false, instead it asks
you to work out the consequences of each theory's axioms (as well as
using meta-reasoning skills to weed down overly complex theories, if
you prefer using Occam's) and then either choose to use or not use
that particular theory depending if the results match your
observations/expectations/standards/... (if expectations are broken,
one would either have to update beliefs or theories or both).
Hi ACW,
What ever the global structure that we use to relate our ideas and
provide explanations, it makes sense that we do not ignore problems that
are inconvenient. A big problem that I have with Platonia is that it
does not address the appearance of change that we finite semi-autonomous
beings observe. The problem of time is just a corollary to this. I would
prefer to toss out any postulates that require *any* "magic". "Magic" is
like Arsenic poison, every little bit doubles the harmful effects.
"Magic" is only used for things which have to either be axioms or which
just cannot be reduced further. Arithmetic cannot be reduced further.
What we have as subjective experience is not directly communicable, it
is very 'magical', yet our theories must explain it somehow. We may want
to have no axioms at all, but such theories are inconsistent as they can
prove anything at all.
Why
do we even need a notion of 3p except as a pedagogical tool? What we
need, at least, is a stratification scheme that allows us to represent
these differences, but we need to understand that in doing this we are
sneaking in the notion of a 3p that is equivalent to some kind of agent
whose only mission is to observe differences and that is a fallacy since
we are trying to explain observers in the first place.
Unless we have some way to handle a fundamental notion of change, there
is no way to deal with questions of change and time. Please notice how
many instances we are using verbs in our considerations of COMP ideas.
Where and how does the change implicit in the verb, as like "running the
UD", obtain? We cannot ignore this. I am highlighting the concurrency
problem b/c it shows how this problem cannot be ignored. The Platonic
Realm, especially the Arithmetic Realist one, is by definition fixed and
static, nothing changes in it at all! How do we get the appearance of
time from it? It is possible to show how, but the proponents of COMP
need to explain this, IMHO. It is incoherent at best to make statements
like "the UD is running on the walls of Platonia". How is that even a
meaningful claim?
Any objective model that features time can be made into a timeless one
by explaining how the state depends on the time and how they are
related. Time is a matter of relation between states. The continuity of
consciousness would be a matter of which OM you happen to be in and how
(or if) OMs are related to each other (for a relative measure, they have
to be, for an absolute one, no). This problem seems to be
similar/related to the problem of indexicals - which OM you happen to be
in, which observer you happen to be, what memories you have and so on.
Another problem is the problem of space as we see in the way that 1p
indeterminacy is defined in UDA. We read of a notion of "cutting and
pasting". Cut 'from" where and pasted "to" where? How is the difference
in "position" of say, Washington and Moscow, obtain in a Realm that has
nothing like "space"? Unless we have a substrate of some kind that
transformations can act upon and yet the transformations and the
substrate are not distinguished, there is no "cut and pasting" possible.
If you accept a digital substitution, then cutting and pasting is merely
a matter of instantiating some program twice, one in Washington and
another in Moscow. If you want it to be more fanciful, you can imagine
the reconstruction of the whole body, but I don't see how that would
change the outcome. The way I see it, you expect to be either in W or M
after such an experiment, and after performing it, there are now at
least 2 valid continuations (well, an infinity of them with the UD, but
let's ignore the "mostly the same" or "very unusual and very rare" ones
for now) which are the most probable continuations. It's undecidable
from your perspective before performing the experiment (let's call it
S0) if your next state will be Washington (S1W) or Moscow (S1M). You
could consider your history until now as being contained in S0, for
example in a more "proof theoretic" arithmetical way, it could be a
proof of the existence of your history and OMs until now. Now as the
next possible state is indeterminate (and undecidable), you'd have to
add the location as an axiom and thus you obtain 2 theories: S1M and
S1W. They would both have OMs associated with them, but they are now
different theories/histories.
The scope of the first part of the UDA is to show that if COMP (digital
subst.), you cannot depend on time and space or even implementation
substrate to tell what your next continuation will be like (one needs to
be able to show that most continuations should be local, otherwise you
have the white rabbit problem and COMP would be false as our experiences
are stable).
As a side-note I'll present a quick thought experiment similar to UDA
step 4, but instead of considering the indeterminacy as going forward in
time, I'll show that it goes backward in time as well (this becomes
irrelevant in UDA step 7, but might as well present it here):
Consider a machine with sufficient memory to run a SIM and offer an
interface with the world. Consider that in year 2050, such a SIM exists,
now he runs a special program (will explain later) on a different
machine, the program crashes/does nothing, the SIM now lives 100 years
until 2150, but accidentally his emulator crashes and he was careless
with backups and thus he's now 3p dead in that 2150. Suddenly the SIM
finds himself back in 2050 with his memories from 2150, what happened?
The program happened to run a truly random program that can be generated
from quantum noise (assuming MWI), given a 1bit QRNG: 1) if qrng() = 0,
stop and run program 2) if qrng() = 1, append qrng() to program, goto 1.
(the machine this runs on is the one originally considered)
This shows that the indeterminacy is strongly time independent (be it
backwards or forward), although that should already be obvious by UDA
step 7/8. Of course, the 2050 continuation was just one of an infinity
possible, but it is at least one possible one, thus it should also
happen, along with the rest (all the possibilities are as real, just not
all are as probable).
This might seem very formal and strange, so you could think of it
another way: with COMP, a mind/an observer is an abstract structure and
given some particular implementation of that observer there are many
possible implementations for the next state of the observer so that the
observer be consistent and work fine (at the low-level). The substrate
can very well vary below the substitution level (but such a level may
very well be quite low, and undergoing a substitution at a higher level
would entail some detail/memory loss and possibly even alter the
measure). The continuations are always random programs from the UD, and
there should be an infinite ensemble of them which happen to implement
the observer being in some particular state, the main requirement be
that they implement the next state of the observer correctly, such that
the observer remain consistent (relatively to the previous state). Also,
we cannot depend on the continuations to always lie at a later state
within the UD, a continuation can be at an earlier state as shown by my
previous thought experiment (unless of course you have a very strong
continuity assumption, in which case COMP is false, because you would
never survive the initial substitution, yet strangely your SIM version
would think he is you, which makes me wonder how this is any different
from someone undergoing an operation under anesthesia or just waking up
after sleeping - it should be noted that just a wipe of the short-term
memory is sufficient for such doubts).
What makes the difference between the substrate and the transformation?
Bruno et al seems to consider this in terms of Godel numbering schemes
but what distinguishes a Godel number of a program from a program?
The observer's abstract structure is what remains invariant, while all
the implementations change, quantum foam can be seen as that
"competition between machines which implement the observer". We cannot
know which program implements us or our lower substrates at any given
state and how much of the generalized brain does it include
(environment), this does not mean that we can't say "yes" to such a
digitalist doctor, merely that we can't be sure that we'll be correct
(of the level being correct, of continuity, of nature of qualia, etc),
so it's a bet. The observer itself is subject to change, consciousness
itself requires this change, learning and forgetting is just one example
of such a change. In a relative way, we might be able to think of a
continuous-OM which is represented by change between 2 states (such as
S0->S1M and S0->S1W), and of our histories as an infinite graph
representing such state changes (although finding even all continuations
from a given OM is uncomputable and a making a program which can decide
if any particular machine represents a continuation for some other one
is again impossible, although likely solvable for some specific cases,
just never *all* of them).
Onward!
Stephen
I'll try to answer some of the other posts a bit later. I'm still not
completely satisfied with my current answer as to what constitutes a
continuation as well as how OMs are selected (measure problem). I might
elaborate more on this in other posts.
--
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.
