# Re: Non-truths that do not transform

From: Iain Strachan <igd.strachan@gmail.com>
Date: Tue Apr 26 2005 - 08:29:17 EDT

Vernon:

You wrote:

> I trust you would agree that anything created by divine fiat ex nihilo (as
> strongly implied by the words of Genesis 1) must be fully functional from
> day one, and hence must inevitably assume an 'appearance of age'. Surely, to
> deny this is to question the capabilities of the Creator. So, if we believe
> in a God for whom nothing is impossible (Gen.18:14), how can we Christians
> deny the possibility of something being younger than it appears?

Vernon,
Certainly I believe in a God for whom nothing is impossible. But your
argument cuts both ways. As you know I happen to find Behe's
arguments about "irreducible complexity" to be very strong ones - and
one way of interpreting these arguments is that evolution is
impossible and by implication impossible for God. And this is where
I run into problems ... because as you say nothing is impossible for
God.

As someone who has worked in earnest with Genetic Algorithms, I see
the problem of irreducible complexity all the time. It crops up
trivially if you have a numeric parameter that is encoded via standard
binary encoding. If you are solving a problem where fitness surface
is smoothly varying, then you run into big problems if you want to go
from e.g. 511 to 512 because the binary coding of the two numbers
requires ten bits to change simultaneously ( 0111111111 ->
1000000000). This is never going to work with an evolutionary
algorithm, and is well known in the GA community, and it is called the
"Hamming Cliff" (a cliff in Hamming space). However, that doesn't
mean to say you can't design a GA that works well. You simply choose
a different encoding scheme. The one chosen nearly all the time is
the "Gray code", for which all the changes from one integer to the
next one are a single bit change. This is done by a walk over the
unit hypercube. A 2-bit gray scale would be:

0 -> 00
1 -> 01
2 -> 11
3 -> 10

As you can see, the relative positions of 2 and 3 are swapped over
from a binary code.

If you do this, then your GA will work on simple problems. And it is
absolutely clear that you had to input a "design feature" _up front_
to get the evolution to work.

Gray coding is rarely enough even for comparatively simple problems -
the mathematical representation of the parameters has to be designed
so that the variables are decoupled as well.

But the point here is that I can in principle make an evolutionary
simulation work _IF_ I design it correctly at the outset.

Now I can't see easily at all how you design the evolutionary
algorithm of life to make bacterial flagella, the ATP synthase
molecule etc. But if I say that means it's impossible, I'm somehow
limiting God's intelligence, whose thoughts and ways are far higher
than ours. Just because with my feeble intellect I can't see how to
do it doesn't mean that God couldn't see how to do it.

So I MUST conclude, by the same argument as yours that God could
indeed have designed the laws of the universe so that evolution could
take place and lead to people with the capacity to be aware of Him and
worship him. It is surely the most fantastic feat of design ever, and
one to marvel at.

I realise you wil see this as a significant change in my viewpoint,
and hope you will not think me a "devil worshipper" as a result (as
Michael seems to think you do).

I still believe the numerics of Gen 1:1 are highly significant, but
they need to be separated from the "apologetics" angle in order to get
people to take a serious look at them without feeling that their own
beliefs are thereby being rubbished.

In Christ,
Iain
Received on Tue Apr 26 08:30:58 2005

This archive was generated by hypermail 2.1.8 : Tue Apr 26 2005 - 08:31:02 EDT