From: Josh Bembenek (jbembe@hotmail.com)
Date: Tue Nov 12 2002 - 12:31:47 EST
Dr. Campbell-
"As a rule, this method will tell us whether the random probability is so
low as to make another explanation preferable, rather than being able to
truly rule out random events. The only exception is if we can truly assign
a zero probability to something."
If we take this reasoning, we should accept nicely Dawkins explanation of
miracle-like events in "The Blind Watchmaker," where even if we saw a stone
statue waving its hand at us, we should not suspect miracle. Rather we
should conclude that an extremely low probability event occured where the
random thermal movement of all atoms in the arm of the statue move
simultaneously in a given direction to produce the effect. Dembski's method
leads us to the best inference, given that we allow supernatural
explanations to be considered among the possibilities, and I think it is a
valuable tool.
"The complexity must be pre-specified, an issue that Dembski has noted but
not solved to my satisfaction.
hwkurgtrawhgtuszhrgturehbgjsdfhgbhfgbrbrewbfwbfrshgbrer is a complex
sequence of letters, and could be made more complex by my randomly banging
on the keys for longer. The probability that I would generate that exact
sequence of letters is very low. However, the probability that I would
generate some sequence of letters was 1, assuming that the computer works.
Only by having a narrow goal, not defined with reference to the results, can
the random interpretation be adequately tested. Without a good idea of the
relevant probabilities, we cannot tell if we have a decent test case, much
less the result of the test."
This reflects very closely to a discussion we were having previously, which
was abandoned based upon the difference of how we interpret protein
function. It also brings up a line of thinking that I need to do some
investigations into, as someone on the list-serve graciously sent me some
primary-literature references concerning this exact topic. Regardless, I
think the essence of this problem is: Just how specified is biological
activity and function among protein sequences? The critical distinction
between the analogy of your computer sequence of letters and the formation
of life and the generation of biological complexity and information is the
fact that although a computer can generate any sequence of letters, and this
can be considered specified in any kind of way, biological sequences must be
specified in a biologically relevant way: sequences must confer functional
advantages that lead to evolutionarily selective adaptations. So the
ultimate and crucial question is exactly how narrow is this goal? My
opinion, not greatly supported by exhaustive research, but rather my
inclinations developed from structural biology and biochemistry states that
this goal is in fact quite narrow and that we can expect specified
complexity as assigned to biological function to be quite narrow.
This is the essence of the problem as applied to biological information:
Consider three inter-related "landscapes" of all possible identities:
1. The Sequence Landscape of genetic information (>is greater than)
which generates the
2. Protein Landscape of amino acids, who fold and form global structures
(>is also greater than)
which generates the
3. Function Landscape of possible biological activities.
Now, landscape one is greater encompassing in relationship, by nature, than
either Landscape 2 or 3 when you require the outcome to be relevant
biologically active function that confers selective advantage, which is all
that we are interested in terms of the evolution or (not possible)-evolution
of biological information. Function landscapes are narrow in comparison to
sequence landscapes, not all sequences can lead to biologically active
functions. (For example, some percentage of landscape one forms insoluble
proteins which cannot perform any function.) This also requires an
environment consideration: protein sequences that could not perform a role
at normal physiological conditions may be super-active at
non-physiologically relevant conditions such as extreme salt, heat, pH, etc.
(for halobacter, extremes will be different than mammalian tissue culture
but you get the idea.) The question is, out of the sequence landscape, what
percentage of sequences are there which generate proteins that form folded,
global structures and that ultimately lead to biologically active functions
within the function landscape and confer selective advantage. The
requirements of function and selective advantage, mandated by evolutionary
developed processes, places strict requirements on the output, that greatly
narrows (imo) the sequence landscape available for the access of biological
information. In fact, we currently know that all of life samples only an
infinitesimally small amount of the available sequence landscape. Whether
these sequences are only islands of function that stand above the sea of
modern-day requisite biological function verses the lower threshold function
of pre-biotic or ancient precursor biotic system function is a
largely-unaddressed and completely-open-to-interpretation question, imo.
Josh
_________________________________________________________________
STOP MORE SPAM with the new MSN 8 and get 2 months FREE*
http://join.msn.com/?page=features/junkmail
This archive was generated by hypermail 2.1.4 : Tue Nov 12 2002 - 14:08:01 EST