z47\doc\web\2000\12\evolsub.txt
From: Gregory Cochran [mailto:gmcochran#compuserve.com]
Sent: Wednesday, December 20, 2000 11:14 PM
Subject: [h-bd] Evolutionary flaws - or are they?


          When you read popular accounts of evolution, or sometimes even
more serious accounts,  you see  explanations of  how evolution sometimes
produces suboptimal designs.  The classic example is the vertebrate ( and
human) eye.  The neural connections are in _front_ of the light detectors -
clearly, it'd be better to have the connections in back.  Because of this,
we lose a little detection efficiency, and have a blind spot where the
wiring has to plunge back through the retina to the brain.
This doesn't have to be, and isn't the case for the eye of the octopus. 


        The reason is that as far as we understand it, evolution doesn't
look ahead and can get trapped in blind alleys, local optima.   
Unfortunately, this idea has been widely misapplied, mainly by saying that
any stupid-looking  phenomenon can be so explained.  Nope,.


         You see, although wandering into suboptimal solutions is certainly
possible, you can be sure to see slick, polished,  tested  and robust
implementations of those basically suboptimal designs.  Almost always,
given an environment that isn't too different fron the ancestral one, those
designs will work just as well as they can, which is usually  pretty damn
well.   And if they work well in anyone, you should expect to see them work
well in almost everyone. 


         I  have in mind three examples of such flawed analysis -
pregnancy, appendicitis, and lower back pain.


         Lots of times people have talked about the pain, difficulty, and
danger of human pregnancy, how  it's difficult to get that big baby brain
through a pelvis that still allows a women to run.    It _is_ a tight
squeeze and it's probably a suboptimal solution.,  But....  First, pain
doesn't count.   As long as it is brief, doesn't go on and on and get in
the way of  the work of the world,  pain just doesn't matter - not to
evolution, anyhow.    Second, it looks as if birth, in the ancestral
circumstances  that shaped human evolution, simply wasn't all that
dangerous. People  are thinking of childbirth around the turn of the
century or thereabouts, when  loopy medical intervention had greatly
increased a mother's risk in childbirth.   As it happened, MDs didn't know
nothin' bout birthing no  babies - but that didn't stop them.   So,
considered in the correct context, which is back during  hunter-gathering
days, rather than the lying-in hospitals of  Victorian Europe,   giving
birth involves work and pain, but is not spectacularly risky. 




         Next, appendicitis.  In today's world, the appendix is a nuisance.
 It  blows up  fairly often and can kill you, if untreated,  but doesn't
appear to do anything particularly important. Why then do we even have an
appendix? In _Why We Get Sick_ ,  the authors suggest that a skinnier
appendix is more likely to develp appendicitis than a larger one, and thus
evolution has somehow gotten itself caught in a Malayan thumb trap, unable
to rid itself of a useless organ.  Well, that's really stupid.    Just
gradually shortening the appendix would solve all the problems, and that's
hardly too complex  a hurdle for natural selection.  The real explanation
is much simpler - until recently, and even today in the  lands that time
forgot, like central Africa,   appendicitis is incredibly rare, almost
non-existent.   It's a new problem in the world. Something has changed in
ways that upset the colonic flora, or aided in the spread of a formerly
obscure pathogen.  Or, although I doubt it., maybe modern diets themselves 
cause  appendicitis in a more direct way.   The original question ( what's
the damn thing for?)  is very reasonable, but it should point you towards a
contemporary cause, properly understood.


     Last, lower back pain.  It's been around a long time, and according to
Dr. Harpending, even the Bushmen get it.  The usual explanation is that we
just haven't been  walking erect long  enough, that there are still a lot
of bugs in the system. So, five million years isn't long enough?   The hell
it isn't.   On the other hand, 'bugs in the system' is a pretty reasonable
explanation.  If some ubiquitous pathogen, say one of the herpes viruses,
caused  some kind of slow trouble in the sacroiliac, much would be
explained.   Natural selection can't trump  itself. That would make sense,
but saying that five million years just isn't long enough never will.




          gregory Cochran