Hi Nico,
At a fairly high level, it does seems to me that if certain kinds of speaker intentions facilitate particular semantic integration services (or not so much), then finding a structured way to records these intentions may be worth exploring.
DEFINITELY agree!
For instance, could I possibly be the author of a biodiversity data paper where all TNUs are explicitly not "mine"?
By the definition of a TNU, no. The implication of a TNU is that the associated Reference explicitly asserts taxonomic opinion concerning spelling, rank, classification, and validity/synonymy, as captured in the core properties of a TNU. In cases where a Reference does not make an explicit assertion about one or more core properties (e.g., "I recognize this as a distinct species, but I don't know what genus it belongs to."; or "Some authors treat this as a valid species, and other treat it as a junior heterotypic synonym of another species, and I have no opinion on the matter."), the TNU simply records the ambiguous property as "Unspecified". However, the TNU itself, as anchored to Reference X, does not directly defer to a treatment by Reference Y. Such statements would be captured via a "Usage Citation" mechanism that I described in my previous message. In other words, even if Reference X ("my" reference) simply states, "With regard to Aus bus L., I follow the treatment of Smith", then we'd have two TNUs: one for Reference X, and one for Smith, both of which would otherwise have effectively identical TNU properties. Then we would need to leverage the "Usage Citation” mechanism (TCS: relationshipAssertion), to effectively state "[Aus bus L. sec. Reference X] is congruent to [Aus bus L. sec. Smith]". So, even if the taxon concept is not "mine" per se, the TNU most definitely is.
I am unaware of examples where someone asserts "Aus is a valid genus, and Xus is a valid genus (separate from Aus), and the species bus simultaneously belongs to both genera"; or "Aus bus L. is both a valid species and a junior heterotypic synonym of Aus dus". The latter is sometimes incorrectly assumed in cases where there are "In Part" synonymies. However, it's a Taxon NAME Usage, not a Taxon CONCEPT Usage; and as such, a NAME is anchored to a name-bearing type, so can only belong to one species taxon or the other. Edge cases of types that are of hybrid origins, and syntype series with multiple taxa represented, are rectified through mechanisms prescribed by the various Codes.
Or: how do I record a case where I say "our higher-level classification follows X", but then X's family-level concepts show up as paraphyletic on my phylogeny. Apparently I only followed X so much (at which point I 'said': "screw X, this is better").
This is why TNUs are clean and straightforward (almost always unambiguous), whereas relationshipAssertions are a bit messier (i.e., how MUCH do I actually follow of Smith's treatment of Aus bus? That it's a valid species? What classification to place it? What rank to treat it as? The spelling of the name? The taxon concept circumscription?). We have our five set-theory relationship types (congruent, includes, included in, overlaps, excludes), but those are somewhat limited when we have a long tradition of conflating hierarchical classification with taxon circumscriptions under the umbrella term of "taxon concept".
Aloha, Rich
P.S. I did not see Greg's reply or files on the list -- was that off-list? I would, of course, love to see them!
Richard L. Pyle, PhD Database Coordinator for Natural Sciences | Associate Zoologist in Ichthyology | Dive Safety Officer Department of Natural Sciences, Bishop Museum, 1525 Bernice St., Honolulu, HI 96817 Ph: (808)848-4115, Fax: (808)847-8252 email: deepreef@bishopmuseum.org http://hbs.bishopmuseum.org/staff/pylerichard.html