>At 01:18 AM 7/11/00 -0500, Chris wrote:
>
>>The following little essay was written for another list, but it is
>>relevant to the topic of this list, so I reproduce it here. It was
>>originally written in response to the old creationist canard that complex
>>organization cannot be obtained from random processes. The idea is that
>>this would be like the parts of a BMW miraculously assembling themselves.
>>Of course, though the evolutionary process involves a kind of randomness,
>>it is not even *remotely* like such a miraculous occurrence, as should be
>>clear from this essay, if not from ten seconds thought.
>>------------------------
Brian
>Thanks Chris, a very interesting essay. Just a few random :-[0] thoughts.
>
>Chris
>Why is it that processes of replication-with-variation nature that
>generate complex organization from randomness work?
>
>One reason is that randomness *is* complex organization.
Brian
>I think this was stated better later on. Anyway, by most definitions I've
>seen randomness represents the upper end of the complexity scale. But
>also, according to most definitions I know of, random would not be
>organized. This really follows as a matter of the most usual definitions
>employed.
Chris
Well, we might say that something that is random is organized so complexly
that it no longer means much to say that it is organized, because the
principle of organization is itself too complex to be usable. The most
organized structure is the simplest as well, so I agree that, if we had a
scale like that of probability, from zero to one, a purely random sequence
would be rated as zero organization.
In between zero and one are those things that we normally call complex
organization: Things that are complex but still simple enough that the
organizational principle is shorter than the specification of the value of
every individual item in the sequence would be. At an organizational level
of one, we'd have things like a string of zeros. Complexity would be
inversely proportional to organization, at least for sequences of the same
length
>Chris
>>But, more importantly, these processes work by *accumulating* small
>>changes in organization over a number of generations of replication and
>>near-replication. Selection is not necessary. In fact, the process works
>>*better* if there is no selection at all, and every copy is saved for
>>further replication.
Brian
Here I disagree. Selection is I believe critical for the generation of
organization. A random element generates complexity, selection generates
organization. In my view anyway. But I believe my comment here is in the
spirit of what was written later.
No, because selection is *removing* of variations. It does not effect the
variations left behind (at least not directly and certainly not in my
example). The "tree" of branchings has fewer branches because of selection
than it would otherwise.
Selection *may*, of course, function in such a way as to select out strings
that are less organized, leaving a population of the more-organized
strings. But this is not *creating* such strings; it is merely allowing
them to continue as members of the population of "surviving" strings.
Thus, in my example, some of the resulting strings would be merely complex
(but not organized). Others would be extremely organized but not complex.
And some, would be in between, with a fair degree of complexity, and yet
also a fair degree of organization.
Now, in the real world, of course, even in the case of computer models, we
eventually reach a point where *some* kind of selection occurs, because we
don't have infinite computer memory/or disk space. In *this* case,
selection by environmental criteria of organization will yield more
organized strings than will selection *merely* by randomly removing strings
in order to free up memory or disk space. This is because, once some
strings have been excluded because they are too "random", the remaining
ones will produce "offspring" that are now more organized than the
population would have been otherwise. If we *again* select out the
least-organized strings, we will continue to drive the population toward
organization that would not have occurred had each successive generation
been culled by random selection instead.
>Chris
>>To see how such an unintelligent process can (and does) work, imagine a
>>machine that copies strings of information, with a generally high degree
>>of accuracy.
Brian
>That a random process will generate complexity is, I believe, relatively
>obvious given the usual definitions. one severe problem is confusing
>random (technical meaning) with "lack of purpose" etc.
>
>The example is, nevertheless, a good one. There was a paper published in
>journ. theor. biol. awhile back that essentially does the same thing
>except with DNA and point mutations.
>
>OBTW, why is this process called unintelligent?
Chris
To emphasize that no "smarts" are involved in either the generation of
variations or in the selection process, and that, therefore, no *design* is
involved (or needed). That is, in my example, *both* variation and
selection are as mindlessly random as possible (hence the use of the cosmic
ray detector to eliminate any human intelligence from the selection
process, on the assumption that the cosmic rays could not *know* which
strings they might "select" for removal from the population). I suppose
that Jones, et al, in order to defend their turf, will *now* insist that
randomness is God-like intelligence, but that's a risk I had to take.
>[skipped details of cosmic ray detector]
>
>
>Chris
>>In fact, if the input was *not* random, the process *possibly* would not
>>work as well. Consider what would happen if every input was simply an
>>instruction to set the existing bit to zero in making any copies. This
>>would ensure perfect uniformity and simplicity in the entire population,
>>even after billions of generations. Or, if the only change made was the
>>adding of a zero bit to the string, so that all strings would be strings
>>of zeroes.
Brian
>I've thought a lot about this over the years and yes, I agree. It is
>fairly common to see some bemoaning the fact that mutations are random wrt
>to the needs of an organism. "Wouldn't it be nice if mutations occurred in
>such a way as to benefit the organism? But we live in a cruel,
>purposeless, world. When we grow up we learn that things aren't always the
>way we want." etc. followed by the theme from hee haw "gloom despair and
>agony, oh me". I've seen this crap in a well known biology text.
>Embarrassing. But, as you say, a random process seems to be the best
>choice. By way of analogy, consider the engineer that uses a random search
>combined with selection to produce complex designs.
Chris
Eventually, the variation process may get smarter than pure randomness. For
example, some "errors" in DNA may be corrected, while others are allowed to
stand because the DNA/organism/mitochondria/reproductive-mechanism "knows"
that some things are critical to the organism, while correcting other
things may be too expensive or even harmful to the survival of the
corrective mechanism.
The reproduction mechanism rules. That's why genomes that reproduce by
recombining DNA between individuals remain so "popular": This method may
lose some genes (or alleles) but the increase in variability may promote
the survival of the reproductive mechanism's genes. Obviously, any gene
that significantly "fights" the reproductive mechanism will also be
"fighting" its own chances of survival. Thus, while allowing the
reproductive mechanism such strong control may not be ideal for other genes
(because they then run a risk of being lost), it is *still* their *best*
strategy; everything else is even worse.
Thus, sexually-reproducing genomes allow a certain degree of randomness but
not so much as to severely damage the reproductive mechanism itself.
Further, complex genomes seem to include a kind of dispersed "library" of
not-currently-used genetic material that is randomly activated (but not so
often as to disable too many organisms that carry them). Thus, even if a
gene or allele is no longer needed, it may not be gotten rid of, but merely
"turned off." This allows it to continue being reproduced. Being reproduced
or otherwise perpetuated is the whole "purpose" of genetic material, but
being active in the organism is of no interest to it. Thus, *its* "selfish"
interest is satisfied even though such material may lie dormant in the
genome for millions of years (aside from an occasional random activation).
The big advantage of the "library" is that its contents are *not* random.
They have been found to work in the past, so "trying" them from time to
time can be far more effective in dealing with environmentally recurring
problems than having to reconstruct them from scratch. Finch-beaks are
probably like this (of course, at one time, the various genes for Finch
beaks must have been generated by a less-"intelligent" process -- but, once
developed, they could be saved to be re-activated (over several
generations) throughout the population when the environment made them
useful again). This could be several orders of magnitude faster than such
genes were originally created. Hence the survival value to the genome that
carries them.
Thanks for your comments, Brian.
--Chris
This archive was generated by hypermail 2b29 : Sun Jul 16 2000 - 18:33:03 EDT