LCNameCreationTypeDiscussion

GregorHagedorn - Tue Dec 21 2004 - Version 1.18
Parent topic: LinneanCore
I believe two kind of nomenclatural relations among homotypic names exist: 1 Those detected at a later time (e.g. two names are based accidentially on the same type specimen or higher taxa on the same type name). This information is usually expressed as nomenclatural status or assessment (see LinneanCoreNomenclaturalStatus) 2 Those explicit in the creation of a name.

For the latter, I propose to record the information commonly expressed in "tax. nov., stat. nov., comb. nov., nom. nov." (where tax. nov. may be "gen. nov.", "sp. nov.", etc.) as an enumeration. In addition, this should also cover the cases of autonyms, sanctioning and validation. No conventional abbreviations exist for these cases. Sanctioning and autonyms are never published but implicitly created, validation technically creates a new species although part of the definition refers to older invalid publication attempts. Proposed enumeration:

Questionable further types:

Creation types may later turn out to be illegitimate (botany), or invalid (botany) = unavailable (zoology). In botany autonyms become invalid when it is detected that the name that implicitly created the autonym is invalid; there validity depends on this other name (currently no relation type for this present yet??).

The sanct code is probably the only exception (I am uncertain about whether auto may be invalid). However, if either proof exist that the name has been erroneously accepted in a usage context, or if the editors of the nomenclatural database believe this to be likely, it should be recorded in a nomenclatural database. This way both the usage history can be tracked, and the fact that the name has been recognized as illegitimate can be communicated. The nomenclatural status is recorded separately LinneanCoreNomenclaturalStatus).

Minor problem: in fungi, due to the existence of anamorphs, it is recommended to use "anamorph spec. nov." as a special form of "sp. nov." (compare ICBN, Recommendation 59A). -- I am not sure whether in the context of this enumeration, a separate category is needed or not. It may be sufficient to delegate this information to a verbatim or free-form text element.

For all creation types, except for the first, a basionym or replaced synonym relation with another nomenclatural object exists, through which the type specimen is defined (compare LCProtonymDiscussion). Only objects with creation type "tax" have their own nomenclatural type specimen or type name.

(Gregor, continued:) I assume these types name-types/creations types are exclusive, so a single relation pointer is enough. In LC 0.1.5 Rich called this pointer "Protonym" (it has a ref attribute pointing to the id of another record).

-- Main.GregorHagedorn - 02 Nov 2004

JMS: Creation type is type of events, not type of name. So I don't think that it fits scope of LC; perhaps can be covered by TCS. Keep It Simple and Stupid (KISS). -- 03 Nov 2004


Gregor: What am I after? The categorization should explain which role the nomenclatural base-name relationship - if present - would have. I think this means it should NOT be a property of the relation itself, because I may be able to say what the role would be, without being able to create the relationship (because the other LC-record that would need to be referred to does not exist, or current knowledge is insufficient to choose from multiple ambiguous ones.

What I propose here is largely based on current practices in nomenclature. I am thinking about capturing the "nom. nov." etc. information that accompanies the publication of a scientific name. At least in botany this is required: you must by some means - although not necessarily through fixed abbreviations - express your desire to create a name to do so effectively. Can someone clarify the zoology position?

I then propose to extend this also to autonyms and validation. Note also that Jerry's LinneanCore tries to capture IsAutonym and IsNovum and Combination information in separate elements - I propose to combine the exclusive statements using an enumeration.

(This information would be combined with the pointer information in some other field indicating a related name. Jerry Cooper proposed in an old exchange in 2002 the name "BasedOn" for this, which I second here.)

Information that is implicit in such a "kind-of-name-category" (if not "creation-type"...):

Questions:

  1. Are the creation or whatever "types" listed above really exclusive, i. e. can any name be assigned to one and only one category?

2. Name for this categorization: Does anybody has an name for a better categorization concept? * Is it NameType? Wolf-Henning Kusber notes that the use of "type" was confusing to him, since he thought of nomenclatural types. Should Mode or Kind-of be used? * I read from James's comments that the event-natural that seems to be implicit in "creation" causes confusion. We use the name "creation type" in the DiversityTaxonomy model but I welcome any better name! * Sally Hinchcliffe informs me that in IPNI the information about "comb. nov.", "nom. nov." is part of "record type" (only available in recent records, and IPNI treats stat. nov. as a special case of comb. nov.). * Paul Kirk writes (from email): "So, is not the term you are looking for a 'nomenclatural event' (NomenEvent)? The BasedOn (with or without the 'Is') would appear to mean that something has gone before and in a nomenclatural context for a sp.nov. this is not the case (hence the 'nov'). I can see the need for an 'BasedOn' in Index Fungorum. I have resisted changing the database structure for fear of making the web site fail so where I need extra links in a record I have added these to existing text strings with an appropriate separator..." -- 17. Dec. 2004 * Kanchi Gandhi writes (from email): "How about: "published as" or "status" (e.g., sp.nov., var.nov., comb. nov., stat. nov., nom. nov.). The term "status" may also include valid/invalid and legitimate/illegitimate." -- 20. Dec. 2004 * Gregor: my current shortlist: NameCategory? PublicationType?PublicationStatus? 3. I am worried about the possibility that in botany a nomen novum replaced synonym may be heterotypic (compare LCNomenNovumDiscussion, Jerry's comment #3 at the bottom). I do remember this case vaguely (I am not a nomenclature expert and have not yet encountered this). However, it seems to be the only case where the type-concept changes among all the nomenclatural changes above. 4. Validation in botany is discussed in http://www.bgbm.org/iapt/nomenclature/code/SaintLouis/0049Ch4Sec2a045.htm: "When the various conditions for valid publication are not simultaneously fulfilled, the date is that on which the last is fulfilled. However, the name must always be explicitly accepted in the place of its validation. A name published on or after 1 January 1973 for which the various conditions for valid publication are not simultaneously fulfilled is not validly published unless a full and direct reference (Art. 33.3) is given to the places where these requirements were previously fulfilled." -- How about Zoology?

Gregor Hagedorn -- 3/17. Nov. 2004 -- On Dec. 10 2004 Wolf-Henning Kusber, BGBM Berlin revised the proposal in the form of a Word document and improved it significantly. I used his annotations to update the page - many thanks! -- On Dec. 21, Franz Schuhwerk (München, Germany) reviewed the list, many thanks! -- Gregor Hagedorn, 21. Dec. 2004


Regarding the relation between type and name, Wolf-Henning discusses (by email in German, here my summary) that we need to be careful about distinguishing information asserted in the publication (e.g. about a designated type specimen) and information that becomes part of the nomenclatural type system. Only the types of botany.valid names are relevant for nomenclature.
1. Genus1 species1 Author1 1958, status: "nom. inval. (no latin Diagnosis)"
2. Genus1 species1 Author2 1960 (validating previous name by adding diagnosis and referring to earlier publication; status: bot.valid = zoo.available Name)

Wolf-Henning writes: Formally Name 1 is not a "scientific name" and the type designated has no meaning under ICBN. Name 2 may contain the same information and designates the type only be direct reference to the publication of Name 1 (German: "intendierter Typus"). The specimens in both database records may be identical, but formally the specimen is only the type of Name 2. One important conclusion is that because of the ambigouos status of Name 1 (source of type information as publication, but not scientific name with a type itself), the type information should probably be available directly in Name 2 (i.e. duplicated), rather than being available only through the reference to Name 1. Local database design can differ from data exchange, but the intended behavior in data exchange should be defined. -- During validation, other cases may occur: the original invalid Name 1 may miss the type, or have several types without identifying the holotype. In all these cases, the information in Name 2 is the only relevant type information. Also Wolf-Henning writes: "historical name often have no type, but are still valid. Not every name has a known type."

-- Gregor Hagedorn, 13. Dec. 2004


PS See also LinneanCoreTCSRelationshipTypes -- Gregor Hagedorn - 19. Nov. 2004