search

plazi / treatmentBank

Repository devoted to house keeping of treatmentBank
0 stars 0 forks source link

represent genus level synonymies at species level #46

Open myrmoteras opened 2 years ago

myrmoteras commented 2 years ago

Issue: When a genus level name is synonymized, all the included species level taxonomic names must change. In some cases, only the synonymy at genus level provided, but not the new combinations.

How can we annotate or represent this in TreatmentBank and then make sure these changes are also reflelected in the RDF we use in Synospecies?

@millerjeremya do you have an example UUID where is is happening?

millerjeremya commented 2 years ago

There is some discussion of the case of Dynamosaurus imperiosus case here: https://github.com/plazi/names_LOD/issues/113 Genus Dynamosaurus synonymized with genus Tyrannosaurus by Osborn 1906. The genera are monontypic and the clear implication is that the single species in each genus are synonyms. But when genera are synonymized, we should understand that to mean that all species in the junior genus are now in the senior genus. When we talked about it on Skype yesterday, I forgot that I had already attempted to encode the synonymy in a 1906 species treatment of T. rex. This should show up in Synospecies as a deprecation of Dynamosaurus imperiosus in 1906, but so far, this is not happening.

gsautter commented 2 years ago

Good question ... CoL and the GBIF backbone are dealing with similar issues, namely losing the former genus-species relationship when a genus is synonymized. I don't think there is a straightforward solution to this in the general case, especially considering the inherent danger of implying nomenclatorial acts down the taxonomic tree ... an error would have outright catastrophic consequences. Might be clear in the case @millerjeremya describes, but other cases might not be as clear, especially genera with many more species ... what if a genus gets synonymized with another ones, but the species are distributed to multiple new genera, for instance? There is a lot to sort out, and we need far more than a single clean-cut case to be able to decide upon a more general automatism ...

gsautter commented 2 years ago

Thinking about this for a little longer, I don't even think we should try to represent such inherent name changes in TB or export them to Synospecies ... what we model and export is the facts and nomenclature acts explicitly performed in treatments, and we should keep it this way ... the perfect place to make such inferences and implicit status changes is Synospecies proper, as part of its purpose is to compile the overall picture that arises our the treatment collection as a whole.