Re: Once again...ID

Brian D Harper (
Fri, 01 May 1998 23:59:54 -0400

At 03:54 PM 4/28/98 -0600, Garry wrote:

>Hi Brian,
>I have appreciated your posts recently concerning detecting design. Thanks
>for the interesting thoughts.

Thanks, and sorry about my delay in responding.

>At 04:30 PM 4/28/1998 -0400, Brian D Harper wrote:
>>I'm not sure if anyone has mentioned this yet, but according
>>to my understanding evolution is not a random process in the
>>usual statistical sense of the word random. Here I don't have
>>in mind the usual statement that Natural Selection is not
>>random. I mean that the parts of evolution we usually refer
>>to as "random" are not necessarily random. IOW, in evolution
>>random usually just means that an event does not anticipate
>>the needs of an organism. One can have highly deterministic
>>processes that satisfy this definition of "randomness".
>Perhaps you are right about this. The randomness in evolutionary processes
>is not the statistical randomness of a stochastic process. But certainly
>"chance" or probability comes into play, for example, in whether a
>particular mutation better enables a particular organism to survive a new
>environmental pressure and reproduce offspring with the mutation. So
>'random' may be the wrong term to use here. Nevertheless, the point I
>tried to make about the counterfactuals of chance still applies, for I
>think that indeterministic laws or causation, in which several outcomes are
>possible with associated probabilities, gives rise to these troubling
>counterfactuals and hence to God's ability logically to be able to create
>an indeterministic process which has a guaranteed outcome (Glenn Morton's
>example of a Sierpinski gasket notwithstanding).

I suspect that one's views on this are influenced not only by
where they stand on the theological issue of predestination vs.
free will (as I discussed last time) but also on their view
of how evolution occurs. A predominant view considers organisms
as passive and plastic. This is a very "Newtonian" view. By
passive I mean that organisms have little to say about what
happens to them, they are pushed this way and that by external
forces. Essentially, they are machines. It is interesting to
me that ultra-Darwinians such as Dawkins and Dennett describe
organisms in very much the same way as do ID'ers. Machine and
engineering metaphors abound. By plastic I mean that organisms
can be molded by external forces into practically any shape.
For some reason this reminds me of an essay written by Dawkins
where he tries to explain why organisms don't have wheels.
Basically, they don't have wheels because they're selfish.
Wheels require roads but roads are also beneficial to
competitors. Anyway, if one combines this view of evolution with
chance then its easy to become concerned about how anything might
be guaranteed to occur.

The self-organizational (SO) view of evolution paints a very
different picture. For one thing, organisms are not machines.
Also, the very essence of SO is anti-Newtonian. But I really
don't want to get into all that. The important point here is
plasticity. The example that I gave regarding spiral phyllotaxy
is an example of a SO process with developmental constraints.
The divergence angle cannot be just any old value. Actually
the model predicts, if I remember right, 13 different solutions.
If this view is correct, then one would expect the probability
distribution for the different forms to roughly coincide
with the frequency at which these forms are found in nature,
which happens to be the case. So, it is unreasonable (IMHO) to
say that these forms are accidents of history. This example is
actually very similar to Glenn's except that there is more
than one possibility.

Now I would like to consider another example somewhat similar
to the gasket but, IMHO, much more exciting. The problem
is called Langton's ant. The ant lives in rectangular grid
of squares that are either white or black. The theory of
everything for this world is very simple, there are only
two laws: (1) if the ant lands on a white square it changes it
to black and turns left, (2) if the ant lands on a black square
it changes it to white and turns right. That's it.

Now, one would reasonably expect such rules to just yield
random chaotic behavior and indeed this is what is observed
for many moves of the ant. If one waits long enough (in some
cases this may be hundreds of thousands of seemingly random
right and left turns) one gets to see something really
amazing. The ant will eventually build a bridge and use it
to climb off the screen! Descriptions can't do justice to
this, so fortunately I was able to locate three java applets
for Langton's ant on the web:




The first two are very similar except that (1) is very slow.
It overcomes this disadvantage a little by having the feature
that you can stop the run at any point and see both your
current location and where you just came from.

The third adds a new twist. About a dozen ants start at
random locations. For awhile they do their own thing in
isolation but eventually their worlds overlap yielding
interesting interference. The problem with this is that
the ants are very small and one cannot really see the
intricate details of the various types of bridges that
are built. Also, the simulation seems to stop at some
predetermined number of steps, sometimes before any ant
builds a bridge. If this happens just run it again.
Sometimes you get to see several ants build a bridge
all at once.

The only stochastic element in this problem is the initial
conditions. Once these are set the ant builds the same
bridge every time. With different initial conditions,
different bridges are built. One way of changing the
initial conditions is to randomly set the initial colors
of the squares to white or black. While a specific bridge
is never guaranteed, some type of bridge is guaranteed.

One also has uncertainty and unpredictability with this
model. When I say some sort of bridge is guaranteed I
base this on past experience. No one has found a case yet
where a bridge wasn't built. Yet, even though the problem
is very simple, no one has been able to prove that a bridge
will be built. Given this it should be obvious that one
also cannot predict what type of bridge will be built or
how many steps it will take.


>It has long been recognized that foreknowledge, even infallible
>foreknowledge, does not necessitate anything. Aristotle, in his famous
>discussion of the "sea fight tomorrow," seems to have concluded that there
>is a truth-value gap for future contingent propositions. Much later,
>William Ockham proposed what is now called the theory of soft facts: God's
>past knowledge of my future actions is a soft fact about the past. Suppose
>God believes that in two months I will freely do A. If I am truly free with
>respect to A, then I have it in my power to do not-A, and so bring it about
>that God believed in the past that I would do not-A.

I don't get this at all. This seems to say that I can, by
exercising my free will, change God's mind.


>>Now for a really speculative thought. Is it possible that
>>the creation of an intelligent agent with true freedom
>>requires some sort of evolutionary process with some
>>amount of chance and indeterminacy?

>Are you saying that no free agents could be produced except by an
>indeterminate process? I don't see how that argument would go.

Unfortunately, neither do I ;-). Its just kind of a gut
feeling I have at this point. Perhaps its better to say
that this would be one way of doing it. The idea is to
cut the causal connection between Creator and creation.
Not that God cannot intervene in the lives of men, but
rather that he can choose not to by cutting the cords,
so to speak.

Closely tied to this is my conviction that God acts in
such a way that his actions are never revealed too
clearly, giving men freedom to chose not to believe.

Brian Harper
Associate Professor
Applied Mechanics
The Ohio State University

"It is not certain that all is uncertain,
to the glory of skepticism." -- Pascal