Stephen Paul King (stephenk1@home.com)
Mon, 29 Nov 1999 01:56:29 -0500
Dear Prof. Matsuno and Friends,
Koichiro Matsuno wrote:
>
> Dear Matti, Hitoshi, Robert, Stephen & All
>
> At 21:35 on 28 Nov 99, Matti Pitkanen <matpitka@pcu.helsinki.fi> wrote:
>
> >One must simply give up the idea about physics
> >as evolution for geometry of single time=constant snapshot.
> > In quantum jump between quantum histories picture (effectively
> >quantum jumps between superpositions of macroscopically equivalent
> >spacetimes) grandfather paradox disappears.
>
> This is an important point, no doubt about it. The remaining problem must
> be how to implement this picture in a nittygritty manner down to the earth.
Perhaps the fist step is to look for concrete experiments that could
falsify this picture... I believe that the grandfather paradox is
unsound since it tacitly implies the ability to violate the second law,
e.g. it would be possible to create a Maxwell Daemon in the form of a
computer program that, given the information representing the exact
future states of the hot and cold "sources" of an adiabatic engine and
controlling the "valve" connecting the two, that it could generate more
work that would be possible otherwise; it would be at least 100%
efficient! So this implies a strong but subtle connection between the
thermodynamics and computational pictures of q-jumps between
superpositions of "macroscopically equivalent" spacetimes.
Question to Matti: Could you elaborate on what the phrase,
"macroscopically equivalent", means?
> >Just the keyworkds: classical nondeterminism
> >of Kaehler action; cognitive and material spacetime sheets,
> >localization in zero modes; psychological time as center of mass
> >time coordinate for cognitive spacetime sheet and as zero mode:
> >average increase of psychological time by 10^4 Planck times per
> >quantum jump.
>
> May I add one more to the list? That is thermodynamics taking note of the
> presence of heat sinks. Heat sinks are and have to be concrete in what they
> have absorbed so far.
Umm, but would you not agree that this thought of "heat sinks have to
be concrete in what they have absorbed so far" as implying a temporal
duration? I see wish as saying "Heat content E == Time period". Would it
be correct to consider a recording as a heat sink? There is something
about what Bennett discovered in studying the Maxwell daemon; it is that
heat is only necessarily generated when a bit is erased.
If the "agency of making distinctions in progress" is a material body
capable of acting as a heat sink, could we see the convergence of the
"present progressive" to a "present perfect"? This looks to me like the
situation of a (thermo)dynamical system going into equilibrium, where
all subsystems are completely synchronized with each other and no
clocking can take place any more. Time has become frozen for such.
This looks like Matti's picture of a q-jump!
> In order to make any jump, the concerned party must make a decision or a
> choice. That is quite concrete. Quantum mechanics is generous in allowing
> whatever concretization by the name of measurement, but it is still not
> specific enough about how such concretization could proceed in reality.
We need to look very carefully at the concept of "choice"!!!! :-) I
will tentatively suggest that one of the key necessary signatures of
"choice" is, metaphorically" the ability to change one's mind! Umm, heat
is only generated when we change the bits in a memory! Is this not
"changing one's mind"?
Within the computer science arena, the term "branching time" and/or
"lazy binding" is used to address this "Choice" situation. The dynamical
information structures that we call distributed computational networks,
(such as that described by Ben's work!
http://intelligenesis.net/webmindoverview.html) are set up so that their
state space orbits are not a priori definable! See:
http://www.cs.cmu.edu/~rwh/introsml/techniques/perseph.htm
http://Theory.Stanford.EDU/people/rvg/branching/
http://liawww.epfl.ch/lia/aipc/T_B/sld013.htm
The key notion is that when perfect ab initio information is not
available or a priori definable, systems can "guess" their histories
and/or "only make decisions when the exact (up to an epsilon!)
information becomes available". Contrast this notion with the classical
LaPlacean notion of all of the Universe is knowable by an Observer that
has access to the "exact" initial conditions of such and, tacitly, that
It is capable of computing without heat dissipation the equations that
represent the evolution of the Universe!
> A supreme example of concretization is ticktocking of a clock. (A cesium clock
> the physicist has invented is fine in itself. However, if the clock is left
> unattended by the physicist who is supposed to do the job of exciting the
> cesium atom to its excited state all through the time, no ticktocking.) The
> underlying issue is operation and synchronization in the set of nested
> interacting local clocks in which each clock is part of the others while
> constantly reading and moving the others.
I have been trying to think of this idea, but focusing on the
information content of the clocks. This is the subject of periodic
gossipings. If we could equate the particular information content of
each clock with the "class" (forgive me Lance!) of possible observations
of such, and the synchronization of the clocks with the idea that the
sequence of disjoint pairwise gossipings between observer ("I tell you
all I know so far, and you tell me what you know so far!"), eventually
leads to a situation where all of the observer's clocks have an
equivalent information content class, then I am communicating well my
crazy idea! :-)
One thing that must be noted is that if the number of clocks N allowed
to interact is infinite, the period (number of steps in the sequence
needed to reach information "equilibrium") is infinite. If N is finite
and the N observers can be represented by the vertices of a complete
graph, it takes \upperbound (Log_2 N + Re_2 N) steps to reach
equilibrium. This also assumes no pathologies like dissipation of the
information, as happens in real examples of gossiping, unitarity of
clocks/observers and that the group of clocks/observers is closed!
Mathematical question: what is the rate at which the number of steps
required to reach equilibrium increases as N -> \infinity? What happens
to this toy model when we use allow for dissipation and alteration of
the clocks/observer that are involved in the gossiping?
> Actually, any pair of interacting
> bodies can serve as a clock to a third party since the clock is something
> displaying a relative movement toward somebody else nearby. The act of
> reading one clock in the form of an interaction constantly generates an
> impetus for moving other clocks in the neighborhood. In this picture, time
> is taken to be an attribute of the act of that somebody reading the clock.
So you seem to define a clock as something given by an "agency of
making distinctions in progress" compared to another? So this is like
the relative motion of the "hands" against the "face" in an analogue
watch? I agree completely that the very act of reading a clock is the
"cause" of the disturbance in the other clocks "in the neighborhood"
(e.g. are relatively synchronized to each other and the "clocking"
observer). This is like saying that it is the act of observing a system
that selects what "time" it "at". Umm, if time is canonically conjugate
to energy... :-)
key: observation <=> perturbation
> Time conceived in this way is locally relational and not absolute in the
> Newtonian sense.
Yes! But, does this give us a way of creating the "illusion" of
absolute time for finite neighborhoods ("nebula") of interacting clocks?
I think so! :-)
> However, if we fail in coming up with time that we can share among us, it
> would be just a waste of time to muse on time. One of the loopholes I can
> think of for escaping this stalemate might be to wishfully conceive of a
> most encompassing clock out of the nebula of the nested local clocks, such
> that it can be referred to by any others inside but it does not react upon
> itself accordingly. The most encompassing clock, if conceivable, can be
> referred to by any of the local clocks as a common reference. The
> encompassing clock is a material system that can act upon or can be read out
> by others but does not react upon itself in return. Of course, this must be
> a very funny system.
Umm, perhaps we are not seeing the "backaction" in the ways that we are
used to! What properties would it have to have? Umm, this looks like a
Machian effect if we consider the way that large groups of entrained
oscillators tend to "keep each other in line"! Perhaps this explanation
of mass is hiding here!
> Nonetheless, there seems to be at least one clue for giving this funny
> story a legitimate treatment. Think about a heat sink conceived within
> thermodynamics. The heat sink is something that can act upon others but does
> not react upon itself accordingly. Heat sinks are heat sinks whatever they
> may absorb. If one can associate the most encompassing clock with a heat
> sink, there might be a possibility of imagining a common time to be referred
> to and read by each of the nested local clocks in the nebula. This has been
> a scenario of how to come up with a common time shared by all of the local
> participants without recourse to Newtonian absolute time. If one wants to
> talk about time in quantum mechanics without relying upon Newtonian time,
> thermodynamics equipped with heat sinks must constitute its supporting
> infra-structure. Of course, the fate of this funny scenario must be upon
> what can we expect under the name of heat sinks in thermodynamics.
Here I see a heat sink as a material body or system that is capable of
or the potential for "tending toward equilibrium". Once it is "at"
equilibrium" (present perfect?) it would be unable to be a heat sink any
more. It is interesting that the notion of open "Hubble expansion" of a
space-time acts like a perpetually cold heat sink, except at the Big
Bong singularity! :-)
> Cheers,
> Koichiro
Onward to the Unknown!
Stephen
This archive was generated by hypermail 2.0b3 on Wed Dec 01 1999 - 01:15:41 JST