[tdwg-content] tdwg-content Digest, Vol 20, Issue 17

[Richard Pyle]

Here's the thing:

What is fundamentally different between this statement:

> If someone wants to document that they have a jar full 
> of different individuals of several species taken from 
> a particular place and time then perhaps that should 
> be modeled as a separate kind of thing.

...and this statement:

> If someone wants to document that they have a jar full 
> of individuals from several different populations of 
> the same species taken from 
> a particular place and time then perhaps that should 
> be modeled as a separate kind of thing.

...and this statement:

> If someone wants to document that they have a jar full 
> of different individuals of the same species taken from 
> a particular place and time then perhaps that should 
> be modeled as a separate kind of thing.

My point is that DwC should not single out "species" as being a special kind
of Taxon, that is allowed to participate in certain kinds of relationships
that other ranks (or or even rankless) taxa are not allowed to participate
in.  There have been endless debates about the "specialness" on Taxacom, and
to my knowledge, none of those debates have ended in any sort of consensus
other than "agree to disagree".  Let's not "go there" in DwC.  Let's treat
relationships with taxa in a completely rank-agnostic way.

> Eventually one of the jars will be opened and 
> something will be identified to species.

Maybe the "somethings" will all be identified to the same species, and maybe
they'll all be identified to different species.  When we get to that stage,
if more than one taxon is represented within what was previosuly regarded as
a single Individual, then we generate new Individual instances that
correspond to the different taxon identifications, and tag them accordingly
with their own Identifications (maintaining, of course, the approproiate
"derived from" relatioships among the Individuals).

> They will want to relate that specimen back 
> to the jar it came from.

Yes, definitely!  We certainly want to have a mechanism to address Dean's
point about Individuals derived from Individuals, whether it be a
parent-child sort of derivation, or something more general, is a topic for
another thread.

> However, do we want these collection sets 
> to show up in a search of species occurrence records?

If we want to do a search of "species" occurrence records, then we add the
appropriate filter to the search that limits the results according to
dwc:taxonRank.  But we shouldn't constrain the underlying architecture to
prevent people from doing a search of genus occurrence records, or family
occurrence records.

> I also question the utility of analyzing 
> occurrence records at clades higher than species.

Really?  In the coral-reef fish world, there are some fasinating
biogeographic patterns where certain families only occur at continental
localities, and others that occcu only at oceanic insular localities.  When
I do queries of, say, larval fish holdings -- where species-level
identifications are ahrd to come by, but family-level identifications are
comparatively easy (e.g., a jar of 50 formalin-fixed larval specimens may
all be reliably be identified to the same family, but may be unidentifieable
ot species) -- I sure would like to know if the larvae of certain insular
families occur at contiental regions (and vice versa), because that would be
informative of the likely mechanisms behind the family-level geographic
distributions.

This is just an example I came up with off the top of my head just now.  I
seriously doubt that it is especially unique -- I'm sure there are other
kinds of biogeographic questions one would like to ask of our collective
data, that overalp with situations where identifications may be reliable
only to the higher-rank level.

> Species are assumed to be made up of population 
> of interbreeding individuals,
>
> But what is a genus vs, a subgenus, vs a tribe?

Again...these debates have happened ad-nauseum (almost literally) on
Taxacom, and I think it would be unwise to clutter DwC with presumptions
about the outcomes of these unresolved debates (or worse, repeat the deabtes
on this forum).  By its fundamental nature, DwC is accomodating to a broad
spectrum of user needs.  Why restrict its utility to a subset of needs --
especially when the subset can be accomodated by adding appropriate filter
criteria (e.g., in this case, via dwc:taxonRank)?

> If these higher clades were somewhat stable and 
> had some agreed on understanding then this might
> make sense.

In many groups they do.

> There is no clear reasoning behind why one clade 
> is a family in Mammals and a clade of similar 
> age is a genus in Beetles.

Perhaps not.  But within Mammals, or within Beetles, or within fishes, they
might have more useful meaning.

But that's not the point.  The real point is that non-trivial amounts of
data exist in our domain that prevents reliable assessments of aggregated
organisms to be cirumscribed within the same species-rank taxon concept, but
there are many reasons why we want to be able to assign reliable higher-rank
taxon identifications to these aggregates. As such, if an Identification is
a tuple of Individual and Taxon, then not only do we want to allow
Idetifications to link to higher-rank taxa, but we don't want to be forced
to restricting such links to be used only in cases where we have confidence
that the members of an instance of Individual reliably fall within a
species-rank taxon cocnept, even if they are not labelled as such.

I just re-read the preceeding paragraph, and it makes my head spin; so
apologies in advance to the dizzy among us....

Aloha,
Rich