Following from TDWG2004, and building on the LinnaeanCore proposal of Jerry Cooper, Richard Pyle assisted by many others (Sally Hinchcliffe, Gregor Hagedorn etc.) are developing the slightly renamed LinneanCore as a schema to exchange taxonomic name strings.
They are maintaining an active WIKI site to discuss these proposals @
http://wiki.cs.umb.edu/twiki/bin/view/UBIF/LinneanCore
The Napier TCS team has considered how the proposals being developed as the LinneanCore relate to the TCS.
Specifically:
1. Whether TCS could include/use LC as its 'Name' representation
or
2. Whether TCS could represent LC 'Name objects' adequately as TCS concepts
This is presented as LCAndTCSCompared
The Napier TCS Team has put together an imaginary Timeline example of Taxonomic and Nomenclatural changes recorded for the genus Aus. This is shown in a powerpoint presentation TimeLine.
We have created a TCS instance document to represent the concepts created and described in this timeline as a demonstration of how Taxonomic Concepts are represented: XML.
To view the XML you can use a free version of XMLSpy or a simpler XMLViewer, or you can open them in InternetExplorer (you might need to enable Pop-up and ActiveContent).
In order to explain more clearly how Names are handled by the TCS this example records a number of concept definitions that are published as taxonomic revisions, and one example of a purely nomenclatural observation. We believe that there is considerable overlap between a Nomenclaturist view of the Taxon history and a Taxonomist's: Taxonomists are interested in the published definitions of Original and Revision Concepts; whilst a Nomenclaturist is not interested in the definitions of Concepts but needs to refer to the original publications where names were first introduced. A Nomenclaturist is therefore interested in the same Original Concepts - but only the Name elements of these (i.e. a limitted view on the Original Concepts). In addition we need to represent purely nomenclatural observations such as Aus bea should be known as Aus beus; and we have modelled this as the creation of another original concept - without any defintional information other than a relationship to the original concept for the name that the recorded observation is 'correcting'. We think that it should be possible to represent all Nomenclatural observations using this model, but we are interested in testing this with further examples (see below).
Each first-use/definition of a name is recorded as an Original Taxon Concept
Each first use of a Name also creates a Nominal Concept which only exists to allow markup and resolution of biological data on the basis of undefined names.
We think that nomenclaturists should be able to represent all the information that they require by including LinneanCore as a subset of the TCS; and that nomenclatural relations should be modelled as relationships between taxonomic concepts. We suggest that that the information that Nomenclaturists wish to track can be represented as a subset of the TCS, and believe that this can be tested with real examples. We foresee difficulties if it is possible to represent nomenclatural relationships in two ways - as attributes of a Linnean Core object and as relationships between concepts with Names.
(Excerpt from posting on same date to the tcs-lc@ecoinformatics.org email list.)
TimeLine Powerpoint fileI have a couple of questions/comments about the PPT file, which might help spark/focus additional discussion. My references to slide numbers refer to actual slide numbers when viewed in Design mode -- not to animation steps (often several per slide).
Probably trivial, but do the little symbols represent only those particular specimens examined by each "in" author -- or do they represent the full scope of individual organisms (living, recently dead, and yet-to-be-born) on the planet that each author would have included within each concept circumscription? I'm assuming the latter, because it is very unlikely that each author would have access to the same particular set of specimens. But then again, from the "Demo_v2.xml" file, it looks like the symbols represent actual vouchered specimens. Like I said, probably trivial, but I just want to make sure one way or the other.
Need to be really careful with the words "valid" and "invalid" (side-box for Pyle 1990). Those words mean different things to different people and different nomenclatural codes. In this case "bea" and "cea" would have been treated by Pyle as incorrect spellings, not as "invalid" names; and "beus" and "ceus" would considered "correct" spellings of the same name (at least from a zoologist point of view).
Question to Botanists: In this example, would you consider "Aus bea" and "Aus beus" as "different names", or "different spellings of the same names"?
Maybe yes maybe no. On the grounds they are different strings we should model them as different names/concepts. A data source can then point one of these names/concepts to the other one with the appropriate relationship flag. Another data source may have a different opinion and assert a different relationship. If we say 'they are the same' we are expressing an opinion that may be wrong. The exchange format should allow for different opinions to be expressed. People have opinions about nomenclature as well as about taxa. When presented with two strings coming from two sources we should start by assuming they refer to two different things then model their relationships. RogerHyam 2005-03-04
Whenever a new Code-compliant original description is published, two concept records are simultaneously created: one that refers to the concept circumscription intended by the Author ("Original" concept), and the other is an "empty" concept automatically created to represent the name by itself, without any specific implied concept circumscription ("Nominal" concept). In the PPT file, you provide an implied "link" between each pair of "Concept" records via the "ID" attribute (e.g., "cl1" and "cl1n"; "cl2" and "cl2n"). I'm 99% sure that you don't intend the actual ID value to have embedded within it some sort of self-evident link between an Original concept and its corresponding Nominal concept -- but could you please erase my 1% uncertainty with an explicit confirmation of this?
The Codes (at least the ICZN code) would not consider the gender-corrected "Aus beus" as a "replacement name" to "Aus bea". Indeed, I don't think any zoologist would think of these as different "names" -- but rather as different spellings of the same "name". I don't believe the Pyle 1990 act of noting the corrected spelling would represent either an "Original" concept, or a "Nominal" concept. More of a "usage" instance. Hence, I am either uneasy about: a) creating two different concept instances for each spelling correction; or b) the way that "Original" and "Nominal" concept types are defined. Perhaps it would be appropriate to create only a new "Nominal" concept for each of the two misspelled names (not sure I would agree with this), but to me, what distinguishes an "Original" concept from a "Nominal" concept is that the former has a specific implied concept circumscription, whereas the latter is purely nomenclatural. Pyle 1990 never made any assertions about concept circumscriptions for either name, and therefore should not have an "Original" concept associated with either (by my understanding, anyway).
On the other hand, if Franz had discovered that Xus Pargiter 2003 was preoccupied by Xus Smith 1950 (thus Pargiter's "Xus" being a junior homonym of Smith's "Xus"), then published the replacement name "Zus", it would be appropriate to create a new "Original" concept and a new "Nominal" concept. But this is a special case, and there are botanical/zoological differences here (with respect to basionym treatment).
You illustrate "Xus beus" as having a new "Original" concept and new "Nominal" concept. This the botanical perspective, but not the zoological perspective. There's been a general assumption that the Zoological perspective would change to accommodate the botanical perspective, but I think it depends ultimately on how the names data is ultimately modeled within TCS (i.e., as attributes of Concepts, or as objects to themselves). Needs more discussion.
I had been under the impression that "Original" concept was restricted to Basionyms, true "replacement names", and perhaps new combinations (if the botanists win) -- but not alternate spellings. However, I get the impression that here a new "Original" concept is created whenever there is a novel "NameString" (specific string of characters). Also, I thought they were created only in cases where an actual concept (circumscription of specimens) was asserted -- not for purely nomenclatural acts (without any concept implications).
Because Pyle never made any statements about the shape/size/position of the concept circumscriptions associated with the two names he was correcting, I don't see why there should be any "SEC Pyle" instances.
Also, I think there is an error in the box for TCrev19 -- shouldn't that be "Aus cea Fry 1989", instead of "Aus cea Archer 1965"? Also, what's with the "BFry"? Is that a botanical author code?
XML Demo File v2 file:Why don't any of the Species-rank names have Relationships? Are there no reciprocal "is type of" or "is child of" relationships?
Does <Relationship type="has type"> mean "Has type species."? If so, I understand why cl1 has this relationship, but why is it repeated for ca1, cf1, ct1, etc. Is this only because these are "Revision" concepts, and in each case the author re-affirmed the type species of the genus? Or will all concept instances of Aus L. 1758 contain a copy of this connection with the type species? This is an example of something that is a property of the name "Aus", not of every single Concept instances of "Aus".
For TaxonConcept cp5 (original concept of "Xus beus (Archer 1965) Pargiter 2003"), there are "Relationships" with ct2 and cf3. Does their representation here (instead of as a RelationshipAssertion)? Is this because relationships stated by Pargiter 2003 himself about his own concepts belong within his TaxonConcept instance, and RelationshipAssertions are for "third-party" relationship assertions? (Is this effectively the definition distinction between Relationships and RelationshipAssertions? I.e., between intra- and inter-publication relationship statements?)
Intuitively, it makes sense to me that if you're going to go to the trouble of automatically creating a Nominal Concept for every Original Concept, then the Nomenclatural information should be embedded within the "Nominal" version, and that should be the unit of exchange for nomenclators. It seems intuitive to me that a "Nominal" concept would be a container for "all the name stuff". I think an "Original Concept" should be used for a true concept -- the concept that was asserted by the Author(s) of the original description of the name.
Feel free to scrutinize and comment on how we have captured our imaginary example...then...
We would love to model real examples of Nomenclatural and Taxonomic history/relationships, and to include further examples of nomenclatural changes. If anybody could provide these it would be great - obviously it would help if they were conceptualized somewhat as we can't quite follow abbreviated taxonomic latin! cartoon diagrams or similar would be good.