[SEEK-Taxon] RE: LinneanCore Group Work

[Richard Pyle]

> Question 1: I recently submitted a paper in which, as a result of a
genus-level
> analysis ("revision"), a species previously placed in its own genus now
must
> be "sunk" into another. So Hypoleschus atratus Fall should now be called
> (according to me) Phyllotrox atratus (Fall). I don't specifically publish
any
> reevaluation of identified (type) specimens, or species-specific
descriptions
> in my paper. So would the "Hypoleschus atratus Fall is now a synonym of
> Phyllotrox atratus (Fall)" statement qualify as a strictly nomenclatural
procedure?
> Please (Rich/Jessie/others) answer yes/no, and/or explain.

My answer: no -- not "strictly".  I see that as primarily a concept issue,
not a nomenclatural issue.  You did not create any of the names
"Hypoleschus", "atratus", nor "Phyllotrox", so there is no nomenclatural
actions here -- only concept actions. HOWEVER!!  -- and this is where my
Zoology brain is still struggling to think like a botanist -- if you *were*
the first person to create the combination "Phyllotrox atratus", then in the
botany world you did "create" a new name object: "Phyllotrox atratus (Fall)
Franz".  The botanical contributors are advocating that LC treat this as a
separate name-object from "Hypoleschus atratus Fall", and hence represent a
different LC instance for it (i.e., with a different GUID).  In other words,
the first creation of a new combination (by botanical rules) *would*
represent a nomenclatural act -- but in the example as you describe it, not
a *strictly* nomenclatural act.

> Question 2: Of course there are cases in taxonomy where by "just going
through
> the literature," one can see that things need to be renamed. In 1938
Hoffmann
> discovered that the name of his recently published new genus
Pseudoderelomus
> had already been used (for another taxon) by Champion in 1910. So Hoffmann
> published another very brief note (like 3 phrases total) in which he
renamed
> his new genus into Neoderelomus. There was certainly no mention of any
types,
> other constituents, diagnoses, etc. in that tiny note. So is the
> "Pseudoderelomus (sec. Hoffmann) is a synonym of Neoderelomus (replacement
name
> proposed by Hoffmann)" a strictly nomenclatural procedure? If yes, fine;
if no,
> please (Rich) provide an example where this does apply. How strictly
nomenclatural
> does one have to get?

I'm assuming that by "strictly nomenclatural procedure" -- you mean
something that would be represented by data contained in the LC structure
separately from being represented elsewhere in TCS.  There has been a bit of
discussion on what Zoologists would call a "replacement name", which is the
same is what botanists would call a "Nom. Nov.", on the LC Wiki.  These are
things that most people (including me) view as nomenclatural information,
that can exist without the existence of a concept to attach them to.

> Comments: I apologize if I'm not catching up to your level yet. As you
know
> I'm not a database specialist. So far as I can see, I agree with Jessie's
> view that there should be no instances of names used in taxonomic
> publications without a "sec." I do think, however, that Jessie's point
could
> be made more convincingly and will give it a shot.
>
>   If I understand the history of "concepts" in databases correctly, then
there
> has long been a conflict between two issues. One may describe those as (1)
> accurate labeling of separate pieces of information (so that the pieces
can be
> exchanged freely and reassembled into their original form); and (2) better
> expressing the dynamics of nomenclature and taxonomy than can be done
storing
> a single (however authoritative) view about how to classify nature.
>
>   The beauty of Berendsohn's "definition" of the term "concept" is that it
is so
> vague as to allow both goals to coexist, even if in reality they are
sometimes
> antagonistic (see Rich's slightly genial "To me, this is not a meaningful
> "taxonomic concept" -- by anyone's definition"). If you don't be believe
this,
> pease re-read his 1995 and 2003 papers - almost anything goes there.

If I'm not mistaken, Berendsohn uses the term "potential concept".  By using
the qualification "potential", I understand his unit to mean "here is an
instance where author(s) made reference to a taxonomic name -- there is
potentially a concept represented here, but not necessarily".  I think his
deliberate vagueness is brilliant as well, but I also think the word
"potential" is there for a reason.  I would have to go back and re-read his
papers to be sure, though -- I don't want to speak on Walter's behalf here.
And I still maintain that a "concept" that circumscribes only one specimen
(that explicitly excludes e.g. the parents and siblings of that one
specimen) is not a meaningful taxonomic concept by anyone's definition.

> The TCS - by experience - must adopt the same vague Berendsohnian
approach. It
> is not only a tool for managing different views about nature but also a
tool for
> managing "pieces of information" existing in the taxonomic literature and
in
> taxonomic databases. I am certain that this "almost-anything-goes"
approach
> violates many people's notions of what concepts should be.

This certainly doesn't violate my notion of what concepts should be.  In
fact, I think my perspective is much more liberal than yours or Jessie's,
because my scope of "potential concepts" extends effectively out to any
usage of a name, by anyone, ever.  But I don't think this is the issue we
are discussing here.

My understanding of what we are discussing is how to distinguish bits of
information that are relevant to nomenclature only (and by extension
relevant to concepts) from bits of information that are relevant to concepts
(but not strictly part of nomenclature).  Maybe I am exceptionally naive,
but from an information management, the distinction seems rather unambiguous
in all cases except that of the application of a name to the name's primary
type specimens (i.e., the one potential point of contention where I see name
data and concept data intersecting).  I see two universes: one is focused
around actual, physical organisms and abstract circumscription-sets of
organisms; and the other universe is focused around Codes of nomenclature
and rules that apply to the creation, establishment, and replacement of
scientific names (strings of text).  The only place where I see these two
universes colliding is the primary type specimen(s).  Coming at it from the
names perspective, I would slightly prefer to surrender this one point of
universe overlap over to the concept side, so we don't need two separate
information structures to reference specimen data.  If the other
nomenclaturalists agree (as Sally already has on the Wiki), then there may
be no point of contention at all.

>   Are the examples 1 & 2 above concepts in the information
management-related
> sense? Yes.

Agreed.  But a subset of concept information management is nomenclature
information management -- so the real issue is what bits fall within the
nomenclatural information structure (i.e., LC -- which I am presuming would
end up as the Names structure of TCS), and what bits fall outside of the
nomenclatural information structure (i.e., in other parts of TCS external to
LC).

> The TCS should eventually accommodate a Phyllotrox atratus (Fall) SEC.
FRANZ
> (2005, I hope...), and also a Neoderelomus (Hoffmann) SEC. HOFFMANN (1938,
> second tiny publication). These "strictly nomenclatural" procedures are,

I do not see these as "strictly nomenclatural" procedures!  I would break
them down as follows:

Nomenclatural information:
 1 - "Hypoleschus atratus Fall YEAR"
 2 - "Phyllotrox atratus (Fall YEAR) Franz 2005" (if zoologists adopt
botanical practice)
 3 - "Pseudoderelomus Champion 1910"
 4 - "Pseudoderelomus Hoffman 1938a"
 5 - "Neoderelomus Hoffman 1938b"
 6 - Linkage of 4 & 5 as homotypic synonyms

Concept information:
 1 - "Hypoleschus atratus Fall YEAR SEC. Fall YEAR"
 2 - "Phyllotrox atratus (Fall YEAR) Franz 2005 SEC. Franz 2005"
 3 - "Pseudoderelomus Champion 1910 SEC. Champion 1910"
 4 - "Pseudoderelomus Hoffman 1938a SEC. Hoffman 1938a"
 5 - "Pseudoderelomus Champion 1910 SEC. Hoffman 1938a"
 6 - "Neoderelomus Hoffman 1938b SEC. 1938b"
 7 - "Pseudoderelomus Hoffman 1938a SEC. Hoffman 1938b"
 8 - Mapping of 1-2 as Congruent? Overlapping? Including? Included in?
 9 - Mapping of 3-4 as "Excludes"
10 - Mapping of 3-5 as Congruent? Overlapping? Including? Included in?
11 - Mapping of 4-5 as "Excludes"
12 - Mapping of 4-6 as "Congruent"
13 - Mapping of 6-7 as "Congruent"

My point is, all the information applies to the concept objects, but not all
of the information applies to the name objects.

> after all, procedures enacted by expert speakers, published in expert
outlets,
> according to expert interpretations of how nature should be named and yes,
> how the rules of nomenclature interact (with some leeway on each side)
with
> taxonomy. I would like us to reach a consensus about the radical (not!)
idea
> that, even though taxonomy and nomenclature work partly independently,
> they are also in many ways interdependent!

Agreed!  And that is why I have always felt that LC should exist only as
embedded within a TCS wrapper.  But the point I am trying to get across is
that there is a block of nomenclatural information, that falls entirely
within the nomenclatural universe, that exists outside the context of a
concept. Stated another way: "Aus bus Smith 1949" is a fundamentally
different unit of information than "Aus bus Smith 1949 SEC. Smith 1949".  So
even though they appear to be the same, and appear to have many elements in
comment (e.g., "Smith 1949"), and even though I treat the latter as a proxy
for the former in my Taxonomer data model (as I *think* Jessie is suggesting
for TCS -- but not sure about this), they really are different sets of
information that (I think) need to be addressed in different parts of the
TCS schema.

> That is so because our adoption of Linnaean nomenclature is meant to carry
> and convey expert information content. Doesn't Hoffmann's 1938 name
replacement
> reflect a clear recognition of his genus being different from Champion's?

Yes -- both the name object is different and the concept object is
different.  Name object information would be stored with the LC substructure
of TCS, and concept information would be stored in other parts of TCS
outside of the LC substructure.

> Doesn't my renaming of Hypoleschus reflect insights of a lengthy cladistic
> analysis?

I would like to think so!

> With some leeway, nomenclature follows taxonomy, and sometimes later it
follows
> its own rules, but only after following taxonomic insights first. We
shouldn't
> ruffle over things like this.

I don't think this is what we are ruffling over.

>   Whether or not Franz (2005) and Hoffmann (1938) talk about one or more
specimens
> (Rich's angle)

Not my angle.

> is actually not the reason why the TCS treats "nomenclatural issues" as
concepts.
> The point of the latter is not to revolutionize what people consider a
"taxonomic
> definition," or to overly conflate nomenclature and taxonomy, but to
arrive at a
> future-proof solution to managing information stemming from many
taxonomy-related
> outlets.

Agreed.  And to some extent "past-proof" as well.

> That is the inherent strength of the "sec." annotation, more so than
telling us
> when a publication is "strictly nomenclatural" or "mostly taxonomic."

Publications are not "strictly nomenclatural" or "mostly taxonomic".  The
publication unit isn't the unit of discussion here.  What I understand the
issue to be is what bits of the information we are trying to manage in the
TCS/LC structure are name-object bits of information (stored within the LC
substructure), and which bits are concept-object bits of information (stored
outside of the LC substructure, within TCS). It seems clear to me that the
following two statements contain different kinds of information:

 1. "Smith 1949 included sufficient and appropriate information to establish
the name 'Aus bus' in accordance with the IC_N Code of nomenclature."

 2. "Smith 1949 applied the name 'Aus bus' to a taxonomic concept that [is
congruent with|includes|is included in|overlaps with|excludes] the taxonomic
concept that Jones 1990 applied the name 'Aus bus' to."

The first statement is in the realm of LC, and the second statement is
within TCS, but outside the realm of LC.

>   If we can come to a consensus on the partially independent/partially
interdependent
> statement above, then we might eventually see that "strictly nomenclatural
procedures"
> would do more damage to the concept approach if treated as in the LC than
their
> analogues would do to traditional nomenclature if treated as in the TCS.

I guess I really don't understand what you mean by "strictly nomenclatural
procedures".  Did I use that expression?  I know I have spoken of "strictly
nomenclatural information" -- by which I mean there is no concept
information implied (e.g., that the description of the Code-governed name
'Aus bus' appeared on p.XX of Smith 1949).  That is not to say that such
information is irrelevant to TCS -- but that it is strictly nomenclatural
information within TCS, and therefore belongs in the TCS "Name" element
(i.e., within the realm of LC).

Aloha,
Rich

Richard L. Pyle, PhD
Natural Sciences Database Coordinator, Bishop Museum
1525 Bernice St., Honolulu, HI 96817
Ph: (808)848-4115, Fax: (808)847-8252
email: deepreef at bishopmuseum.org
http://www.bishopmuseum.org/bishop/HBS/pylerichard.html