Re: Evolution is NOT random

On Thu, 15 May 2008 11:09:14 -0600, dkomo <dkomo871@xxxxxxxxxxx>
wrote:

r norman wrote:

On Thu, 15 May 2008 10:24:10 +1000, j.wilkins1@xxxxxxxxx (John
Wilkins) wrote:


dkomo <dkomo871@xxxxxxxxxxx> wrote:


John Wilkins wrote:


dkomo <dkomo871@xxxxxxxxxxx> wrote:



John Wilkins wrote:



dkomo <dkomo871@xxxxxxxxxxx> wrote:




So why do people keep discussing it as though it were? An example of
this is the Texas Sharpshooter thread. Winning lotteries and drawing
targets around bullet holes paints a highly distorted picture of
evolution. I notice even Wilkins got into the act.


You mean this?

http://www.talkorigins.org/faqs/chance/chance.html


I was refering to your post in the Texas Sharpshooter thread. Perhaps
you might explain what winning lotteries has to do with the process of
evolution? Is this just an inappropiate metaphor?


The logic applied by creationists is a form of the Lottery paradox, here
referred to as the Sharpshooter Fallacy. These fallacies have been long
discussed in philosophy. The do help us clarify what the issues are.
They do not, nor did I ever suggest they did, tel us what is happening
in evolution, only that the argument by antievolutionists fails, for the
reason given.


I wasn't aware of your article. I'm in the process of reading it. I
think we agree on some things like that there is a difference between
objective randomness and unpredictability. Unpredictability is a result
of various deficiencies of observers. It seems to me that what is
referred to as randomness in evolution is actually unpredictability, as
illustrated by the responses in this thread.


I think that, whether or not physics is random at the quantum level, the
problem is not that we are not Laplacean Demons capable of making the
predictions on the totality of evidence, or not, but rather that
variation in evolution is not necessarily correlated with future
fitness. This is what people usually call "chance" in evolution - the
chance that a variation will turn out to be useful. It is this sense
that Monod, for example in his Chance and Necessity, was using the term.


You seem to be saying that we treat evolution as random because it looks
random *to us*. In other words, we are not Laplacean Demons, so we have
to compromise. I thought you were the one who was really big on
objective reality. Here, you're putting the observer very much into the
frame, and reality is seen through the observer's eyes.


No, you've misread me. As the Laplacean Demon is not at issue, it is
*not* about observers. Instead as I make clear in the essay, it is about
there being no a priori correlation between variation (mutations) and
future fitness or utility. In short, variation is "at random" relative
to fitness. This is the only relevant sense of "chance" (or "hazard" as
Monod called it) unique to evolution. All other senses of "chance" apply
equally to all domains and sciences, and so are not problematic for
evolution particularly.


Sorry about the empty post, I clicked the wrong button.

I have tried to keep out of this because I just get all ill-tempered
thinking about the inanity of the argument.

If so, then this reply should really get your blood pressure up there.

(Not yours, John.) If
you rolled back the tape of evolution you would never repeat anything.
You can't roll back and repeat anything where diffusion is involved,
for example. There are an incredible number of cellular events and
processes where Brownian motion plays a role in determining exactly
when and if something might happen, at least in determining it at the
molecular level. And all cell activity is nothing more than the
accumulation of events at the molecular level. And all organismal
activity is nothing more than the accumulation of events at the cell
level which is ... And all population activity is nothing more than
.... In other words, tiny blips of molecules can determine whether a
neuron fires, whether a molecule reacts, whether a mutation occurs.
You don't need Quantum indeterminacy, all you need is statistical
mechanical fluctuations which are not correlated with anything.

And the failure to predict Brownian motion except as a statistical
process is not simply due to our ignorance of the details. First, it
not possible in a finite time and with finite tools to represent
exactly the details of the setup. And then you do have uncertainty to
contend with. I mean the uncertainty that says not only can't you
_measure_ conjugate variables simultaneously, but particles do not
even possess exact values of the conjugate variables. At that level,
quantum fluctuations certainly can influence particle interaction
which influence Brownian movement which influences molecular activity
which influences cells which influence ...


I humbly disagree with your assertions. Brownian motion, diffusion and
thermal noise are still squarely in the classical realm, unless you
*explicitly* bring quantum fluctuations into the analysis. If I recall
correctly, for example, Einstein's treatment of Brownian motion did not
consider quantum effects. So as long as these phenomena are classical,
they are therefore deterministic and repeatable. End of story.

And, as I mentioned in my reply in sbe (which hasn't appeared yet), you
are repeatedly conflating our knowledge of reality with reality itself.
Our inability to track with exactitude the positions and velocities of
all the molecules involved in diffusion has nothing to do with the core
question of whether these flight paths are deterministic and repeatable.

Finally, I notice that the dread term "irreversible" has been bandied
about a few times in this thread. I wish to withdraw any references I
made in the past about time reversibility, rewinding and rolling back
because these are irrelevant to my central question about whether the
evolutionary process is deterministic and repeatable. If I don't
withdraw these references I fear I'll get stampeded by the insane
propensity of t.o. posters to go flipping off on irrelevant tangents and
spend absurd amounts of times arguing about minutia which have nothing
to do with the central question. Irreversibility, multi-worlds and
entropy are examples of these.

So, replace the tape of life with the DVD of life, and replace each
reference to "rewind and replay" with "reset, then run".

Now to you, John. What about genetic drift?

[Jumping in] What about it? In the classical realm genetic drift is
also repeatable.

Is that not a rather
significant factor in evolution? Is that not dependent on so many
contingencies, including ultimately Brownian motion of the molecules
and, of course, quantum effects that it another relevant sense of
"chance"?

Just to be clear, every time you mention quantum effects, I will agree
with your conclusions. But not otherwise.

OK, I'll bring quantum effects into it without excessive damage to my
blood pressure.

First, my impression is that there is an interpretation of quantum
mechanics that doesn't just say you can't _measure_ the position and
momentum but that a particle does not _have_ exact position and
momentum. The parallel, closely related by the agreement of the
underlying mathematics, is the exact time and frequency of a very
brief tone. An tone of extended duration may be of a specific
frequency but doesn't occur at a specific time; it is extended over an
interval. If you make it briefer and briefer, you lose the specific
frequency; the Fourier spectrum spreads out. The shorter the time
interval the wider the frequency band. You cannot have a tone at a
specific instant of a specific frequency. An event that occurs at a
specific time (a delta function or impulse) has an infinite frequency
bandwidth. The problem is not in the measurement; it is in the
concept itself. The wave function has the same properties; you
cannot confine it in space without altering it.

Okay, in that sense you cannot have Laplacian demon knowledge
necessary to compute the Newtonian deterministic universe, not because
we don't know the information but because the information does not
exist.

Another approach. In order to know the dynamics of, say, a confined
gas, you have to know the position and momentum not only of every
particle in the gas but also in the confining container because of
collisions with the wall. But the container is enclosed in its own
environment and the outside atmosphere impinges on the walls so in
order to compute a tiny volume of gas confined in a container, you
have to know the entire universe. You won't buy that because you
argue that the entire universe is still deterministic even though the
uncertainty principle I explained just above shows it is not.

Another approach. Here is were quantum mechanics is necessary to
understand classical statistical thermodynamics. First, particles in
a gas are not perfectly spherical billiard balls of specific mass and
diameter bouncing elastically off each other. They are squishy
electron orbitals around atoms interacting with each other in ways
demanding quantum mechanical calculations. I would imagine the
specific orbital state (energy state) of the outer electrons would
have some effect on the collisions and must be calculated. Even the
tiniest effect would get magnified by the chaotic nature of the system
until all predictability is lost.

Then there is radiation from external sources. If there happens to be
light (or any other form of radiation) shining on the gas, then photon
absorption is associated with absorption of the momentum of that
photon and the gas particle will be kicked off course. So to
calculate the deterministic statistical mechanics of a gas you have to
know the course of every photon (or other particle, cosmic ray or
whatever) that might possibly interact with the gas and also calculate
the details of that interaction. Since the interaction of a photon
with the electron orbitals of an atom is a quantum mechanical process,
I guess that firmly puts quantum mechanics into the realm of affecting
macroscopic classical dynamics.


.