[Taxacom] Semantic Web: What is a species?

Peter DeVries pete.devries at gmail.com
Sun Jan 25 16:09:55 CST 2009


Thanks to everyone for posting their comments. ;-)


I think that this discussion is illustrative of two issues.


Different people often have different ideas of what the same term means and
in discussions these

subtile differences are not always apparent.


When matching or linking two terms or concepts via the Semantic Web, one
needs to be aware that

they may not actually have the same semantic meaning.


If one of my collaborators chooses to tag their specimens with my species
concept URI, they

are saying that we share a common species concept.


What I have been thinking about ,and what prompted the original post, was
that I needed a more formal

definition of what this species concept is.


David Baum has a paper that will be coming out in Systematic Biology that
discusses the issue

of species concepts, and has some valuable insights.


This issue is very complicated and I was not trying to oversimplify it. My
goal was to foster thinking

about the best way to implement these concepts on the Semantic Web.


I had some additional thoughts that I would like to put out there.


Animals seem to do a pretty good job of "recognizing" conspecifics through a
number of different

mechanisms. For them, species act as if they are real.


What if species were mainly "real" things with fuzzy borders but our ability
as humans to correctly

recognize them and and their instances (specimens) is limited. We can only
hypothesize that

something is a species, but overtime, our hypotheses get closer to
approximating the real species.


This is a take on Plato's Allegory of the cave
http://en.wikipedia.org/wiki/Allegory_of_the_cave


And may be best illustrated by an example.


We all know what a chair is, right? Now try to write a formal description of
a chair that correctly

accounts for all the known instances of chairs.


Thanks to everyone for posting their thoughts, it has been very helpful.


- Pete



More information about the Taxacom mailing list