[tcs-lc] Names as Objects

[Gregor Hagedorn]

Jessie wrote:
> > we want them to be
> > concepts even if in the original publication of that name there
> > is no concept described as such there will be reference to the
> > original concept (in your terms the original name embedded in the
> > original concept) that this is a correction for.

Rich wrote:
> That's where we disagree, I think.  If I read you correctly, you suggest
> that a name without an elaborated concept should default to the "Original"
> concept.  I say it should default to the corresponding "Nominal" concept.
> That's why I think the "Nominal" concept is such a powerful tool -- it says "we
> don't know what concept was asserted, but we know that it likely falls within
> the set of possible concepts for which this name has been assigned". That's
> exactly the sort of "concept" we want name-only data to default to.

I would like to support Rich here. The original concept at the time of 
publication of the nomenclatural name may indeed be meaningful, and it should 
be possible to make an assertion that something has been identified according 
to the original concept. However, if this is not know, "defaulting" to it would 
be highly confusing.

This may be a social question depending on the group you work with. Probably in 
most vascular plants, the original concept has reasonable circumscription. 
However, in groups like fungi, bacteria, viruses, microscopic insects, 
probabaly many molluscs, for many species older than 50 or 100 years the 
original concept is so far removed from current knowledge as to be meaningless.

I also want to support Rich in the stressing the importance of distinguishing 
between name-string and name-object identity. I think separating name objects 
(= type concepts) and circumscription concepts is often vital. I am in danger 
of repeating myself, but I argued repeatedly that the name string, even where a 
unique id for a name, is a bad basis for implementing name-equality 
comparisons, because of misspellings when used and because of misspelling and 
grammatical errors in original publication that have to be corrected upon 
citation, at least in botany. My own experience from matching two lists of 
names, one small and one large authoritative is that only 30% of names from the 
small list could be found in the authoritative one based on string comparison, 
whereas almost all names could be mapped manually.

My largest concern about NOT being able to cross-reference name-object and 
circumscription object is that much of the biological reality I encounter in 
our diagnostic work of plant pathogens is not expressible:

Publication1 publishes a name "Ustilago violacea" with a historic character 
circumscription. Publication2 20 years ago studied morphology and produced 
keys/diagnostic descriptions for "Ustilago violacea" defining a circumscription 
concept. Publication3 studies chemical or molecular characters and concludes it 
needs to be recombined into Microbotryum. However, in my group it is highly 
unlikely that Publication3 either explicitly bases its character 
circumscription on any specific publication, nor gives a character 
circumscription itself.

If I identify this pathogen, I would call it "Microbotryum violaceum 
(nomenclatural recombination Publication 3) sensu Publication 2". Note I may be 
the first that uses the names from Publication3 in the circumscription sense of 
Publication 2, a fact that rightfully is not even noticed in science. 

Conclusion: biologists freely recombine knowledge on nomenclatural on type-
concepts with existing character circumscriptions concepts. This can be easily 
modeled of the two concept types are different object types, but not so easily 
in a single object type. 

Alternatively the nomenclatural type can be represented by a TCS type 
"nominal". However, it must be elaborated who shall create the additional 
relationship asserptions between the four other TCS concept types and the 
nominal TCS record. I am afraid that none of the current data providers is 
interested in doing this work. With a separation of name and concept object 
types, these relations would already exist and allow actions such as above.

Gregor 
----------------------------------------------------------
Gregor Hagedorn (G.Hagedorn at bba.de)
Institute for Plant Virology, Microbiology, and Biosafety
Federal Research Center for Agriculture and Forestry (BBA)
Königin-Luise-Str. 19           Tel: +49-30-8304-2220
14195 Berlin, Germany           Fax: +49-30-8304-2203