going rankless (was Article 71...)

[David Wright]

Barbara Ertter wrote:
>A rose can be a rose can be a rose without necessarily
>being a particular species of rose (I've certainly given up on most of
>the
>ones in my garden, anyhow!)  As is, I can still call it Rosa, but what
>uninomial would I refer to it by?


But in an unranked system the species in question would still be
diagnosable and identifiable as a member of some clade (e.g.,
"Antilinnaea" or "Antilinnaea sp.").


-----------------------------
Clades (evolutionary "lineages") exist in nature to be discovered, but
Linnaean categories (families, classes, genera, etc.) do not.  The
question of whether or not it is *possible* to convert to a purely
hierarchical, rankless, non-Linnaean taxonomy is distinct from the
question of whether or not it is *desirable* to do so.  Previously I
responded to a message suggesting that there are no real alternatives to
the Linnaean system.


Some of the impassioned defenses of the status quo seemed confused about
what is being proposed.  Here is an example of how moving to a rankless
system could happen:


Class Aves
  Order Passiformes
    Family Paridae
      Genus Parus
        Parus bicolor
        Parus major
        Parus rufescens
        Parus sclateri
        etc.

when stripped of ranks could look something like the scheme below, if
(e.g.) binomial species names were frozen into to functional uninomials on
1 January 2001:

Aves
  Passiformes
    Paridae
      Parus
        Parus-bicolor
        Parus-major
        Parus-rufescens
        Parus-sclateri
        etc.


It would not be absolutely necessary to join the genus and species names
with a space-filling character, but I think it might help to avoid
confusion of the old "genus" part of a frozen species-name with that of a
simple unranked clade.  Perhaps simply rigorously maintaining the
convention of italicizing species names would be sufficient to distinguish
them.  The exact typographical convention finally chosen is not crucial to
consider the possiblity (and not worth arguing about); I used the
connecting-character approach above to emphasize the uninomial nature of
species names.  Note that it is possible to mention the name of a more
inclusive taxon to provide a context for a species (it simply is not
necessary to do so to identify a species).  This unranked system would
preserve historical priority of names, etc.  But the names of species
would not change even if they were transferred out of Parus into a
different clade:


e.g., Paraparus, new clade
        Parus-rufescens
        Parus-sclateri
        etc.

Another possiblity would be to simply separate the specific/trivial part
of existing binomials from the genus part, but that would result in
massive numbers of homonyms to deal with.


Why bother changing?  Again, that is a distinct question.  Here I merely
wanted to point out that many of the faults attributed to a rankless
system over the past few days (e.g., necessity to have fully resolved
relationships for it to work, alternative must be icky numerical taxonomy,
etc.) simply do not exist.  The main benefit cited for ranks (which no one
tried to defend as natural) is that they provide mental crutches that help
us to remember which groups are more or less inclusive in a hierarchy.
Does anyone really believe ranks are *necessary*?  If not, it is worth
considering what is to be gained by simply dropping them.  The essay by
DeQueiroz and Gauthier (1994; Trends in ecology & evolution 9:27-31)
clearly sets out some of the problems with ranks and some of the benefits
of dropping them.  Ranks distort the way we view and interpret biological
diversity, despite glib assurances that "everyone knows they aren't real."


David Wright
dwright at u.washington.edu

PS -- nobody really took on the question of whether or not types are
necessary.  A typeless system wherein taxa are based purely on diagnosis
could work, but it would probably be more volatile than a system in which
taxa are based on types.