threads on patterns
kinman at HOTMAIL.COM
Fri Mar 2 10:11:50 CST 2001
I have said little on these particular threads, but I think Richard
Zander's point bears repeating (as I understand it). [Philosophical and
semantic discussions are interesting, but often go so far afield that
important points are lost in the shuffle].
Pattern recognition is a tricky business. Both humans and their
computers may see patterns that aren't really there (or magnify
insignificant ones), and this may seriously interfere in our search for
truly important patterns. Most of us might see a water stain on a wall and
wonder what might have caused it, but a few might discern a pattern in the
stain of the Virgin Mary crying and regard it as a miracle.
That's admittedly an extreme example, but if you push computer
algorithms beyond their limits, cladistic analyses will often come up with
patterns that aren't real or are unimportant. And if there are a number of
alternative trees that are somewhat equally parsimonious, there can be a
tendency to emphasize the wrong one (or in some cases they are all equally
wrong and worthless).
This is a vulnerability of cladistics, that overextrapolation can
result in serious errors. This problem is compounded when such results are
prematurely encoded into the formal nomenclature by phylogenetic
systematists. I am just as guilty of wanting to squeeze as much information
as I can out of the data, and I may overextrapolate---but I do oppose
finely-divided formal cladistic classifications, which often give formal
names to patterns that are unimportant, not there at all, or even totally
misleading (thus obscuring and slowing the discovery of the real patterns).
P.S. Don't get me wrong. I have nothing against Occam's razor and
parsimony in moderation. What makes me nervous are unwarranted attempts to
"maximize" parsimony (and oversimplify). And this probably happens more
often than we might think.
>From: Curtis Clark <jcclark at CSUPOMONA.EDU>
>Reply-To: Curtis Clark <jcclark at CSUPOMONA.EDU>
>To: TAXACOM at USOBI.ORG
>Subject: Re: parsimony/biology
>Date: Thu, 1 Mar 2001 18:13:31 -0800
>At 03:15 PM 2/28/01, Kirk Fitzhugh wrote:
>> My suggestion is that patterns don't
>>arise from homology or cladograms, but are explanations of patterns that
>>summarize in the form of applying the same names to properties. I agree
>>with you that pattern recognition is fundamental, otherwise there is no
>>impetus for explaining. This does not, however, mean that cladograms are
>>the patterns in need of explanation, but are the explanations.
>Okay, now I understand. Makes perfect sense.
>Curtis Clark http://www.csupomona.edu/~jcclark/
>Biological Sciences Department Voice: (909) 869-4062
>California State Polytechnic University FAX: (909) 869-4078
>Pomona CA 91768-4032 USA jcclark at csupomona.edu
Get your FREE download of MSN Explorer at http://explorer.msn.com
More information about the Taxacom