From: Crosby_M (CrosbyM@po1.cpi.bls.gov)
Date: Tue Sep 24 1996 - 12:07:00 MDT
On Tuesday, September 24, 1996 11:36AM, Anders Sandberg wrote:
>On Mon, 23 Sep 1996, Crosby_M wrote:
>> The implications of these latest theories of evolution & physics take us
>> far beyond the simplistic entropy and heat death scenarios of
>> traditional physics (note that Rosen is explicitly saying that complex
>> adaptive systems cannot be modeled as Markov processes, which is a basic
>> assumption in Frank Tipler's "Physics of Immortality", for example)
>What evidence is there that Rosen's models describe reality better than
>the usual ones?
Uh Oh - I was afraid that I'd bitten off a little too much with that
post!
I warned that I had not studied Rosen. I believe his point is that
mechanical models of physics do not allow systems to act in an
anticipatory fashion, something that some believe to be the case in
living organisms, while others might argue that it is simply a matter of
mechanical natural selection. As for what Rosen means by "much else
besides" organic phenomena, I haven't a clue at this point.
That's a poor response so I guess I'll have to get and read Rosen's
latest book:"Life Itself: A Comprehensive Inquiry into the Nature,
Origin, and Fabrication of Life". Here's one review of it:
"Book News, Inc., 02/01/92: In a sophisticated (but clear and
understandable) exposition, Rosen (physiology and biophysics, Dalhousie
U.) rejects both the necessity and the sufficiency of the mechanistic
model/metaphor for biological life, and proposes an alternative, drawn
equally from experience in biology, physics, and mathematics. Annotation
copyright Book News, Inc. Portland, Or. "
>And where does Tipler rely on modelling complex systems as Markov
>processes? His main arguments about the Eternal Return and the Bekenstein
>Bound applies to just about any system; even if the system does it best
>to be unpredictable or organized it *has to* repeat itself in a finite
>time if it is finite and non-growing. Markov processes are awfully general.
You are probably correct that Tipler does not rely on Markov processes
in his arguments. The following is my *very* _naive_ interpretation of
this from a fragment of some notes I took after reading "The Physics of
Immortality". I think I was using Markov processes as a proxy for the
even more general idea of determinism.
10/31/94. I submit that most real-world processes do not evolve as
Markov chains because state Xi+1 is NOT dependent ONLY on the state the
system is in at state Xi. Random inputs from the environment during the
transition to state Xi+1 can cause the next state to be quite
discontinuous from state Xi. For example, a machine normally wears down
in a manner consistent with a Markov process; but, accidents can happen
at any point and thus break the chain. Of course, in this case, the
probability of an accident could be incorporated into the calculation of
the transition probabilities. However, in the case of living systems
that are also actively seeking new inputs as well as encountering them
by chance, the assumption of stationary transition probabilities seems
absurd.
We are not Markov chains because we have memory AND we can plan ahead.
In other words, in addition to accidents and active seeking of
alternatives, we also are able to apply past states (experience) to
current situations. The teleology of being nostalgic and goal-directed
violates the first-order Markovian property of finite-state systems.
What this means is that, yes, we may go along for a while, stuck in a
rut so to speak, advancing step by step according to stationary
transition probabilities; but, we are also able to suddenly recall a
previous, perhaps long-ago state, that somehow seems relevant to our
current situation, and we make a 'quantum leap', an attitude adjustment.
In short, Tipler's basic assumption that humans are finite-state
machines and, therefore, can be simulated, by brute force in the
all-powerful future, is irrelevant. Even if A SINGLE entity can be IN
only a finite number of states, this does not mean that there are ONLY a
finite number of TOTAL states that an entity can choose from.
While acknowledging free will and the randomness of experience, Tipler
seems to imply that, in the long run, none of these idiosyncracies
matter because we will settle into a 'pattern'. It is the common
patterns, not the individual peculiarities, which are 'immortal', which
will be resurrected by the Omega Point. All flaws will be removed in
the 'heaven' of the Omega Point, except perhaps as isolated, static
perversions available only for pious reflection and academic reference,
but not allowed to exist in practice.
>> Some recent posts have claimed that the universe is compressible because
>> it contains so much space, ignoring that this 'space' is far from empty,
>> being filled with sundry species of cosmic rays and 'vital' dust
>> (archives of the universe?)
>What is the evidence that vacuum is not a state of near maximal entropy?
>(One vacuum with virtual particles looks like any other vacuum -> *very*
>high entropy)
Vacuum: 1. emptiness of space, 2. a space absolutely devoid of matter
(or voided of matter to the highest degree possible).
I think you are correct that vacuum is "in a state of near maximal
entropy." I am saying that space cannot be considered a mere vacuum, it
must be seen as a communications medium, even a cyberspace, for
electromagnetic energy (not to mention gravity). As far as cosmic rays
& 'vital' dust being the 'archives of the universe', that was pretty
much wild speculation on my part. However, these cosmic rays and dust
do contain reams of information about the past cosmos. Perhaps space
could be compressed, if that wouldn't impact its ability to transmit
information or allow material processes to occur. One of the things I
criticized about the Strong Nanotech posts (as Dan Clemmensen wrote on
9/22/96 : "With MNT, all I need to produce an item is the design for the
item, and matter, and energy.") was that they assumed no need for
'geospace' or other infrastructure.
Also, Taoists and others are fond of telling us to study the spaces
between things (the relationships) rather than just the things
themselves.
Mark Crosby
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:35:45 MST