non-Linnean taxonomy

Robin Panza panzar at CLP2.CLPGH.ORG
Mon Mar 20 18:06:25 CST 1995

Except for Dr. Reveal's much-appreciated (by me, at least) history of
nomenclature and biological naming, the majority of contributions to this
thread have been, overtly or implicitly, about removing the classificatory
aspect to the current naming system used in biology.  Unfortunately, I have yet
to read any alternative proposal for conveying currently-believed or proposed
information about relatedness or where, in the overall evolutionary array, this
particular item fits.  All this arguing over the terminology to apply to the
endpoints on the evolutionary tree does not address how to communicate
information on the path to that address.

I agree wholeheartedly that the Linnean (kind of an odd name, since it bears
so little resemblance to the system he actually devised) system of
hierarchical classification terminology has serious flaws, but where
are the _reasonable_ alternatives that will work any better?

Certainly, when everything is known about the branching patterns (including
reticulations and partial reticulations) and divergences of all
life forms, we can create the Ur-tree, showing all the nodes (and antinodes)
leading to each taxon.  Then we can name the endpoints Fred, George, and Tara,
and project the pertinent segments of the Ur-tree to show a colleage (or
government official, or lay-reader, or whomever) the place in life's scheme
occupied by George.  Alternatively, we can give them all numbers indicating
which path (0="right", 1="left" at each node) leads from the root of all
life to the taxon in question.  There are certain problems with such an
attitude that I've not heard addressed in this thread.

First and most obvious is that we are not likely to ever produce the Ur-tree
and, even if "we" (not in the lifetime of anyone on this thread, at least) ever
did, what do we do in the meantime?  We don't even know all the endpoints on
the tree (undescribed, undiscovered taxa), or all the branchings, much less the
order of the branching.  Many, probably most, major taxonomic groupings (by
whatever rank-name you wish to use) have not even been addressed
phylogenetically.  In those taxonomic groups that have been so addressed, "The
Answer" has not been anything but final, and has been argued and altered
by many who used a slightly different dataset.  We don't have an
uncontrovertible tree for _any_ taxon of which I am aware.  In other words,
even the segments of interest to one person keep changing and, until all is
known, the "taxonomy" (by which I mean the path) of each endpoint will be in
_every bit_ as much flux as it is under our current system.  The instability is
due to a lack of knowledge, not the fault of the classification scheme.

Aside from the unknowability and instability mentioned above, such a system is
unwieldy in the extreme.  Without rank names, however inadequate the current
ones are, one must mention all the nodes in the path to the taxon in question
(or at least all the nodes currently known or believed to be along the path,
given current knowledge) in order to direct attention to that taxon.  "Now,
students, we are going to discuss mating behavior in 010011101010001001110101."
Yeah, sure.  That sequence of digits _might_ be long enough to get one to
"Phylum Cnidaria", but probably not.  Yes, current rank names are inexact and
inconsistent across taxa, but at least they simplify (by approximation) things
to a level that can be grasped by human minds.

In addition, we do not have the technology to deal with the complexities of
life.  The Great God Paup cannot distinguish between "real" polychotomies (and
Its disciples therefore maintain that such do not occur) and collapsed nodes
due to insufficient data.  It cannot even _address_ the problem of reticulate
evolution.  Taxa _have_ combined genomes to produce self-perpetuating hybrid
lines, and have "shared" genes in less grand scale.  Aside from denial, how
does one code such an antinode?  A minus sign?  And what does one do with
asexual clonal lines that slowly diverge from each other via mutation?

The above paragraph assumes one accepts a semi-typological species
concept--that life forms can be grouped into taxonomic units of shared genetic
descent reasonably independent of other such units.  There are those who do not
accept even this degree of typology (hence, the discussion about the
abolishment of types).  Do we need to use the pedigree of each _individual_
organism currently alive today?  My, what a tree that would make!

The Linnean system of hierarchical classification is far from ideal.  Its
current applications are inconsistent (the fault of the user, not the tool).
The labels used above the species level are oversimplifications and, at best,
approximations.  However, I have yet to hear of an alternative that will be any
improvement.  The only alternatives I have heard are:  let's do away with
classification altogether (contrary to human nature, and it would do away with
the jobs of most of the people subscribing to taxacom);  let's use a scheme
that will accurately portray all branches along the path to each endpoint
(which has the flaws stated at such length above); or let's feed all our data
into the black box known as Paup and accept Its divine word (which is nothing
but an approximation, called a concensus tree, and will be altered as soon as
someone comes up with more/new/different data or introduces a new endpoint to
be included).  Does _anybody_ have any idea what actually goes on inside that
black box?  Do we have any really good reason to blindly believe its Answer?

Until someone comes up with a better alternative, why should we do away with
the current system?   Better an approximation, based on weaknesses we know and
understand, than a system that is accurate-but-too-big-to-grasp or blind
acceptance of a black box whose weaknesses are unknown.

Robin Panza
Section of Birds, an admittedly paraphyletic group
Carnegie Museum of Natural History

Return-Path: <Postmaster at>
Received: from by (MX V4.1 VAX) with SMTP; Mon, 20
          Mar 1995 12:10:07 EST
Received: from by NET2.WAU.NL (PMDF V4.3-10 #6821) id
          <01HOD9IIB14000084W at NET2.WAU.NL>; Mon, 20 Mar 1995 18:09:00 +0000
Received: by; Mon, 20 Mar 95 18:04:38 +0100
Date: Mon, 20 Mar 1995 18:04:35 +0100 (CET)
Subject: Undeliverable Message
To: panzar at CLP2.CLPGH.ORG
Message-ID: <Rbf8+XKPPja at>
X-Envelope-to: panzar at CLP2.CLPGH.ORG
Content-Transfer-Encoding: 7BIT
X-Priority: 3 (Normal)

More information about the Taxacom mailing list