Re: Devi (Re: Adams-Webber (Re: Lefebvre))

anima@devi.demon.co.uk
Sun, 12 May 1996 22:31:08 +0000

Dear Jack,

Thank you for the copy of your summary of Lefebvre, and its thoughtful motive,
>Perhaps the following might save you some time (or save you sometime).

On the contrary, alas! I'll have to expend some time if I'm to take
Lefevbre with the seriousness that his ideas, as instanced by your summary,
deserve. Fortunately, it all sounds so interesting that that'll be no great
burden.
(And it'll only "save me sometime" if I get down to him properly: so here's
my summer vacation fun job, I guess!)

There's certainly plenty to get my head round. Firstly, the material that
you sent isn't self-evidently an answer to the apparent paradox I raised:
there'll be some additional inferences involved.

Another way of saying this is that there is sufficient loose thinking in my
own understanding of the issue at present, for me not to be able to
appreciate what's being said: particularly when I suspect that some of the
authors you quote are themselves "loose", not in the sense of being
imprecise, but rather in offering an argument which is orientated to rather
different issues than my own. (Strictly speaking and in fairness to both
parties, _the mapping between their concerns and my own_ is loose.)

*****
For example, you quote Stephen Hawking on the necessary unidirectionality
of memory in a universe whose entropy is always increasing (Hawking, 1988
p. 155 rather than p. 147, by the way: same edition by Bantam, must be a
different printing): "one can show that this increase in disorder is always
greater than the increase in the order of the memory itself".

I've always understood that the statement that "entropy is always
increasing" is true only when a system boundary is closed to energy: one
can always reverse entropy by inputting energy, as Boltzmann first argued,
and as Hawking concedes when he gives the example of painting a house.
However, "that requires expenditure of effort or energy and so decreases
the amount of ordered energy available" Hawking (1988: p.108); and you
provide the same caveat when you say
>any model of the organization of information in memory, either human or
>>computer, must entail a net increase in entropy _when we take into account
>>the amount of energy dissipated in its implementation_", my emphasis; and so
>the argument is that on the larger scale, entropy can only increase.

Now, it strikes me that this argument is applicable or not, depending on
where one chooses to place the system boundary.

Hawking's concern is with cosmology and so he's quite right to place the
boundary "around the Whole Works", as it were!
(BTW: I don't believe this is quite that same use of the word "boundary" as
his later discussion of possible reversals of entropy in a cyclic,
expanding- contracting universe pp 156-161, where the term relates to the
consequences of the initial starting position, as it were, of the "Whole
Works".)

Our concern in the present discussion is with the local system which models
the human being, or the computer, processing information into and out of
memory; the "boundary" then, is that which distinguishes the set of
elements in this system from everything else: its environment.

And I don't _think_ it necessarily follows that the argument should be the
same for the local system with which we're concerned as it might be for the
"Whole Works" system of the entire universe. Imagine, for example, an
information-processing system in which a human being is making choices: we
can pump energy into that system to ensure entropic reversal (growth,
development, choice, increased information, indeed increased wisdom if we
get lucky) as much as we like, without being concerned that in some other,
wider, "Whole-Works" system there would be a net loss of information: we're
mainly concerned about a level of discourse within the single life-span of
the indivdual we're modelling, and we don't _have_ to see the local system
of this model as included within the wider "whole-Works" universe. That's a
choice of scope that we make to suit our current purposes.

We can mix and match. Ross-Ashby had an interesting concept of the
"Intelligence Amplifier". If I remember rightly, this was a system capable
of choosing among alternatives (and thereby one which _increased
information_, i.e. reversed entropy) which could do so much more
effectively when interconnected in just the right way with a system of
increasing entropy, "hitching a ride" on the entropic increase, as it were.
The fact that the whole contraption dissipated some total energy store in
order to do so was neither here nor there.

In other words, the two kinds of model (Hawking's and our more localised
ones) don't necessarily map onto each other- we're talking different
explanatory structures here- and I'm not sure that explanations in our own
model are illuminated by explanations at the "whole-universe" level.
*****

Or are they?

There's a fascinating publication by Coveney and Highfield (1991) in which
the authors' point of departure is the apparent paradox that for so many
natural systems, the equations which model them could equally be run
backwards and forwards in time and yet time's arrow only goes one way. How
come? They cover similar ground to Hawking and yourself; towards the end,
they say:
"Have we therefore made a universal link between the thermodynamic arrow of
time and the reversible equations which govern the microscopic world? Not
quite; after all, we did not consider all possible kinds of motion..." (the
descriptions) ..."we have so far encountered depict only the extremes of
behaviour, contrasting 'simple' and 'complex' dynamical systems", (Coveney
& Highfield 1991: p. 281): and yet it has to be said that their many
examples of local (natural biological) systems dependent on time's arrow
are very suggestive of the universality of their argument.

And there's certainly something very absolute about the very concepts of
order, logical necessity, and of the "pressure towards consistency" without
which it's impossible to model peoples' thinking and action, whether in
terms of construct systems or anything else! (It just seems such a great
leap to go from cosmologies to psychologies...)

And so I look forward to tackling Lefebvre, who, you once told me in an
earlier mailing, is all about that!

Kindest regards,

Devi Jankowicz

PS I'm suspicious of the "anthropic principle" too, BTW: sounds very
_circular_ to me!

Coveney P. & Highfield R. (1991) _The Arrow of Time_ London: Flamingo
(Harper-Collins).

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%