[Taxacom] Are species real? Doesn't matter.
deepreef at bishopmuseum.org
Thu May 31 13:43:52 CDT 2007
Thanks for commenting. Actually, my recent post was more about seeking
recent literature on species concepts and application of binomials and
trinomials. I only mentioned the "species are/aren't real" issue
tangentially. The topic has been discussed in much greater detail on this
list in years past.
The comparison with physicists and the Platonist/Cartesian perspectives is
useful, I think, and I'd not seen that comparison made before. Maybe I
should also sort through the literature surrounding the Physicists' debates
on this issue as well.
> I think we have the nominalism/realism divide here in
> taxonomy, with alpha taxonomists worrying and arguing over
> their interpretations, preferring to always cite alternative
> concepts, while databasers (e.g.
> EoL) deal with what they prefer to consider real things, one of each.
> Oh, sure, we are all scientific realists of a sort, but
> differ in what we worry about.
Hmmmm...I wonder where I fit in. I am certainly at the "alpha" end of the
taxonomist spectrum; but I am also a hard-core databaser (including
involvement with EoL). The databaser in me does indeed want to deal with
"real" things, but the alpha taxonomist in me knows that those real things
cannot be "species". In fact, I am increasingly convinced that they cannot
be "names" either -- not because they aren't "real" (after all, we have
Codes to sort them out), but because there are too many different "flavors"
of how different biologists think of a "name". My current belief, as I have
articulated on another recent thread on this list, is that the closest thing
we have to a "real" entity of practical database management value for
taxonomy is the "usage instance". That is, a massive index tracking how text
stings have been used as names for organisms within the context of various
forms of documentation (e.g., publications -- see
access to such a "BIG Index", the important information concerning "name"
and "species" can be algorithmically derived (regardless of whether the
algorithm is processed in a human brain -- as all biologists have always
done when assessing taxa -- or in a computer).
While I was writing the above, Doug Yanega responded:
> It matters to more than just practicing taxonomists and
> "databasers", and I think you are downplaying how vastly
> asymmetrical the dichotomy is here; there are far, FAR more
> people in the world whose need for taxonomic information
> requires it to be as consistent, objective, and unambiguous
> as possible.
I agree (except for the "objective" bit; consistent and unambiguous, for
sure) -- and I see this need being easily fulfilled through a "layer"
interface between the "BIG Index" and the end consumer. ITIS, Species2000,
FishBase, and others currently perform this function.
> For one of the simplest such observations, and
> one close at hand, is the fact that every specimen in every
> museum in the world that has been identified to species is
> filed under a single name.
[Note: I wrote the following 3 paragraphs when I thought by "filed" you
meant "data managed". I now realize you meant physically stored, but since
I wrote the stuff below already, and it does have some relevance, I'll leave
Ummm...mostly, yes. But I think that is changing, or will change. Very
early on in my career as a technician in a museum collection, I started
wrestling with cases where specimens could either be labeled as one species
(following one expert), or another species (following another expert).
Indeed, it was this very issue that first led me into the whole
names/concepts thing many years ago. I am in the process of completely
overhauling the management of specimen collections data at our Museum, and
will be solving both the species concept issue and the specimen
identification issue through the same silver bullet (as you might guess: the
usage instance). Instead of maintaining a single list of taxonomic names as
our "taxonomic authority", we will instead maintain a potentially much
larger index of usage instances. To serve the needs of the many data
consumers who need a consistent/unambiguous list of species, our institution
will play the role of "meta authority" and select one specific usage
instance for each Code-governed "name" that we feel "got it right". Through
this collection of "Accepted Usages", we can very easily derive not only a
taxonomic authority/checklist in the traditional sense, but also accommodate
the translation of existing specimen identifications into "accepted"
taxonomy (sensu our Museum) on the fly. What this means is that we will
maintain specimen identifications as they are, in the form of a usage
instance ("Smith identified this specimen to be Aus bus L."), which can be
instantaneously translated into "Xus bus (L.)" through our Museum's
taxonomic "meta authority" service.
In many cases, specimens have multiple identifications through history.
Some of these are "real" determination events, where an expert examined the
specimen and declared it to belong to a particular species; and in other
cases, nobody every examined the specimen, but the identification was
updated merely to reflect changing prevailing taxonomy. My ambition is to
track all of the "real" determination events for each specimen as
functionally equal, and then select one of them (usually the most recent) as
the "current identification" -- in much the same way that we select one
usage instance of a taxon name as the "current" usage sensu us. I would like
the end the practice of creating artificial "re-identifications" simply for
the sake of maintaining current taxonomy, and fulfill that need through the
dynamic taxonomic meta-authority service.
The point is, specimen data can be found (searched) or rendered (specimen
labels, web pages, etc.) in a way that serves the needs for a consistent and
unambiguous single name for each specimen, even though the actual database
management scheme does not treat the specimens as having "identified to
species ... filed under a single name".
> I have yet to see an institution
> that maintains specimens in multiple places in their
> collection if there are multiple alternative opinions as to
> what name should be applied.
Where a specimen is physically stored is a completely separate issue from
how the information is managed. All that matters in terms of physical
storage is that the specimen can be located when needed. Traditionally,
this has been accomplished by arranging the specimens taxonomically -- and
with good reason. Obviously, a specimen cannot be on two separate shelves
at the same time (which I now gather is your original point). But a
well-designed database system will obviate the need to informatically brand
each specimen under a single name in order to make it easy to access the
specimen. Keeping track of where the physical specimen is located on a
shelf becomes less of an example for a need for consistent taxonomy, than a
need for creative and careful data management.
> It is also true that there are
> probably no two institutions in existence who have the same
> names applied to ALL of the taxa that they have in common;
> that is why a "rosetta stone" taxonomic resource is essential
> to the future of collections-based research - so multiple
> collections can compare their holdings and accurately assess
> where they overlap and where they differ, without being
> thrown off by idiosyncrasies introduced by alternative
> If you wish to draw analogies, then why not use quantum
> physics? It's fine for some people to think about an electron
> as a cloud of energy whose location and definition in
> space/time is probabilistic, but for most people, they need
> to treat it as a particle.
In the time it took me to write this WAY-too-long post, Richard Zander
already replied with:
> Sure specimens have different names in collections. Maybe they
> are under one name as a filing aid, but for difficult groups
> each may have annotation labels with all kinds of names.
...which very nicely distills the point it took me pages and pages to
More information about the Taxacom