[tdwg-content] If you need something for referring to a population, then it is probably best to do it as a related class

Matt Jones jones at nceas.ucsb.edu
Wed May 4 18:30:06 CEST 2011


Earlier in the thread you commented that reasoning about taxonomies wasn't
your main goal, but that others might be interested in that facet.  I wanted
to point out that Dave Thau, who was extensively involved in the development
of TCS in his role on the SEEK project, did a lot of work on reasoning over
biological taxonomies in his dissertation work.  He published a number of
very useful papers on these experiments in taxonomic reasoning, which are
available here: http://wwwcsif.cs.ucdavis.edu/~thau/ -- I find the
Ecological Informatics article one of the more approachable of the set, but
they all say a lot about the power and pitfalls of reasoning over

Nice work on DSW, and thanks for the effort you've been investing here.


On Wed, May 4, 2011 at 7:49 AM, Nico Franz <nico.franz at upr.edu> wrote:

>  Hello all: I have a few inserted comments, hopefully with some clarifying
> effect:
> On 5/4/2011 10:54 AM, Peter DeVries wrote:
> Hi Steve,
>  I may have overloaded the term specimen to make the explanation easier to
> follow.
>  A specimen could be an individual or it could be part of an individual.
>  To some extent you need to think about how these models will be used.
>  If you subscribe to the model that a species is whatever a taxonomists
> says it is then it is difficult to make statements like.
>  X% of the world's species will be extinct by 2050.
>  If you mean a species as defined by the concept documented at this URI
> which is supported by these specimens, images, and DNA then you are on
> firmer ground.
>  Species in the natural world do a pretty good job recognizing those
> individuals that are appropriate mates. In other words members of their own
> species.
>  Are we modeling species or variations in human conceptualizations of
> species?
>     Assuming that there was only one individual organism identified there
>> is really one one species (or hybrid).
>>  Aaaaaaaaaaaaaah!  Please Nico C., don't take this one up!
>  I stick with this. Assuming you don't have a hybrid individual. That
> individual is one species. The fact that human may disagree on what species
> it is a human issue.
>  Again, Are we modeling species or variations in human conceptualizations
> of species?
> NMF: I believe there is a history of biodiversity database engineering that
> actually led some people, initially in Australia (Greg Whitbread?) and
> Europe (Berendsohn's group) to place more emphasis on modeling human
> conceptualizations of taxa. I understand this happened in the late 80s /
> early 90s. The motivation was, apparently, that no long-term consensus
> seemed available to model a single and sufficiently widely accepted taxonomy
> of, say, Central European vascular plant species or mosses. I believe that
> projects such as the Euro + Med PlantBase (
> http://www.emplantbase.org/home.html) are examples of the persistence
> (success?) of this 20+ year old practice. I think this is the kind of
> background that Steve is referring to.
>  Which of these is of primary importance to decision makers and
> non-taxonomist biologists?
>   Part of the problem with various publications relating to ontologies and
> taxonomy is that their species models entail a specific phylogenetic
> hypothesis.
>  In the real world taxa are not as clean as some would like to make them
> out to be.
>  Each individual is a unique combination of thousands of separate gene
> lineages which often do not follow clean monophyletic paths.
>  I would argue that most of those who work with species related data see
> them as useful typological constructs which in general follow the biological
> species model.
> NMF: The solution of how to properly model things such as species in RDF,
> OWL (etc.) should ideally accommodate all "species concepts" (BCS, ESC, PSC,
> GSC, there's dozens), i.e. it should be neutral towards this issue. The
> examples below are too simple, ignoring for example cases such as pro parte
> synonymy, or a concept ("species") that appears unique and monophyletic to
> one author but (vastly) polyphyletic to another. To do an accurate mapping
> in that latter case (which is not super-common but it's real), I think one
> would have to come down on saying we're modeling human conceptualizations.
> They are not just that, so really it's not a strict dichotomy, meaning that
> by modeling conceptualizations we do not have to say "the species only exist
> on our heads", or "well, then anything goes". I agree (sheepishly) that it
> does come down to what the input data look like, and what the expectations
> are towards just accommodating the user community versus pushing them in a
> particular direction.
> Some of this also on p. 47 of this PDF:
> https://journals.ku.edu/index.php/jbi/article/view/3927/3852
>  *Aedes triseriatus* owl:sameAs *Ochlerotatus triseriatus*
>  Others seem to see them as phylogenetic end nodes which entail a specific
> phylogenetic history.
>  *Aedes triseriatus* distinctFrom *Ochlerotatus triseriatus*
> *
> *
> If you are primarily interested in understanding issues of ecology,
> disease, diversity and conservation the former model is more appropriate
> than the later.
>  Respectfully,
>  - Pete
> Regards,
> Nico
> Nico M. Franz, Ph.D.
> Assistant Professor
> Director, UPRM Invertebrate Collection
> Department of Biology
> University of Puerto Rico
> Call Box 9000
> Mayagüez, PR 00681-9000
> Phone: (787) 832-4040, ext. 3005
> Fax: (787) 834-3673
> E-mail: nico.franz at upr.edu
> Laboratory website: http://academic.uprm.edu/~franz/
> UPRM-INVCOL: http://uprm-invcol-project.tumblr.com/
> _______________________________________________
> tdwg-content mailing list
> tdwg-content at lists.tdwg.org
> http://lists.tdwg.org/mailman/listinfo/tdwg-content
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.tdwg.org/pipermail/tdwg-content/attachments/20110504/c67f35cd/attachment.html 

More information about the tdwg-content mailing list