Theories of Evolution

Tom Holroyd tomh at BAMBI.CCS.FAU.EDU
Mon Aug 29 15:12:08 EST 1994


>And by the way, a clock is not just a lot of pieces. It gives you the time.

Now you've done it.  Oh well, why not?

First, just so you all know where I'm coming from and so you won't label me
a vitalist or something :-), I'm a Ph.D. student at the Center for Complex
Systems and Brain Sciences here at FAU.  To save time, don't read the stuff
in [].

	[You could argue that the function of a clock is to wobble the
	escapement back and forth to create the ticking sound.  But I won't.
	:-)]

Let's examine the position that a clock is a collection of parts and nothing
more.  If we take a reductionistict stance, we can (literally) take the
clock apart, see how all the pieces interact, put it back together, and
create a perfect theory of clock function.  All *without* reference to
its supposed function of time telling.  And this would be perfectly
acceptable to any scientist.

Many biologists appear to take the same approach with living organisms,
especially molecular biologists.

There are those who would say, though, that a living organism is not like
a clock, that it is more than the sum of its parts.  What is meant by this?
	[The history of this line of thought is really old.  I've traced it
	back to Plato in the West, and I'm sure it's got roots in the Orient
	that go back even further.  The basic idea is that of an "undivided
	whole" to borrow the title of David Bohm's last book (q.v.).  Any
	attempt to reduce such a whole to a formal model (like the perfect
	theory of clock function), necessarily omits part of the whole; for
	example it may omit some interaction of a system with the
	environment.]

Formal systems are closed.  Real, natural systems are open.  For a clock,
it is possible to create a closed formal model of an abstract clock that
has no interactions with the environment; such a model of a living organism
is obviously absurd.  See, for example, Robert Rosen's work for more on
this point of view.

A clock's gears interact basically in only one way, at one level.  The
perfect theory of clock function chooses this level of description and
does everything there.

Living systems have highly complex interactions, at many levels and time
scales, with many other components of the environment and of themselves.
These interactions form a network of dependencies where one part of the
system depends in a non-linear way with other parts.
	[And when I say part I have to also say at what time scale and
	level of description I'm speaking on - bones are solid structures
	at one level but are constantly in flux on another.]
Removing any part affects the total system in often unpredictable ways.  It
is in this sense that living systems can be considered to be wholes.
The "whole" living system really needs to be the entire planet, the whole
solar system, etc., for each organism is part of a web of interactions with
other organisms, and isolating an organism from that web alters it, again
in often unpredictable ways.

So reductionism fails to be fully effective on living organisms, despite
the success with the clock.
	[Notice that equating a real physical clock with the formal model
	is like saying that there are no 'important' interactions between
	what is modeled and what is not.  So the model may or may not be
	successful, depending on the real importance of those interactions.]

So what's an honest reductionistic scientist going to do?  Well, you
can state your level of description, decide what time scales you are
interested in, and make a closed, formal model of that little bit.
Notice that you have to assume some underlying physical system which
has its own set of rules.  If you are really ambitious, you can even
try to make a model of the underlying system, at its level of
description, and show how the first model arises from it.  Statistical
mechanics is an example of how the macroscopic theory of thermodynamics can
be derived from the microscopic behavior of the components.  But it
only works at thermodynamic equilibrium, in a closed system, or at best
in the linear region near an equilibrium state (see Prigogine).

	[All this can be applied, by the way, to the brain.  In particular,
	it can be applied to the question of whether the brain is a
	computer.  A computer is like a clock, the brain is not.  If this
	sends up a red flag for you, send me private email and remember
	that I'm not a vitalist.]

So in summary, let us not forget that molecular evolution is a part of
this whole, this web of interactions.  The high-level behavior of animals
influences the course of molecular evolution, and all models live at
a chosen level of description, like statistical models of the likelyhood
of transposition/transition.  The probabilities are in flux themselves.

Tom Holroyd
Center for Complex Systems and Brain Sciences              The basis of
Florida Atlantic University, Boca Raton, FL 33431 USA      stability is
tomh at bambi.ccs.fau.edu                                     instability.




More information about the Mol-evol mailing list