[Tcs-lc] name and concept <sigh>

[Richard Pyle]

> > The fundamental difference to me
> > between a Nominal TaxonConcept and a non-Nominal TaxonConcept
> (e.g. Original
> > TaxonConcept) is that the former is AccordingTo "nobody", and
> the latter is
> > AccordingTo "somebody".  The more complete version of the sample dataset
> > (with AccordingTos) is below.
>
> I interprete these two parts of paragraphs as your nominal
> TaxonConcept is name-literal in my terminology, or may be
> nomen nudum, rather than a taxon concept.  Does it sound
> right?

Close -- my Nominal TaxonConcept refers to what I have been calling a "Name
object". Depending on how you distinguish "name-literal" from "name-string",
it may be the same thing.  If all orthographic variant "name-strings" apply
to the same (one) "name-literal" (as seems to be explained on the LC Wiki),
then it may be that your "name-literal" and my "name-object" mean the same
thing, in which case the answer to your question above is "yes".

> > No -- to me, there is no such thing as "Nominal TaxonConcept1
> is congurent
> > to Nominal TaxonConcept2".  I believe the relationship types
> "is congruent
> > to", "is included in", "includes", "overlaps with" and
> "excludes" have no
> > real meaning for Nominal TaxonConcepts.
>
> I meant such as ambiregnal, equivalent and homotypic for nominal pair.
> I might misunderstand 'such a relationship'.

As I understand it, "is congruent to" has a very specific definition,
something like:

"TaxonConcept definition 1 circumscribes exactly the same set of organisms
as TaxonConcept definition 2"

Because a Nominal TaxonConcept (to me) has no specific circumscription, it
cannot participate in an "is congruent to" relationship.  Instead, there are
other kinds of defined relationships that express equivalency of names, such
as:

"is ambiregnal of"
"is basionym of"
"is replacement name for"

etc...

(all of these imply "is homotypic with")

None of these things have any real bearing on concept circumscriptions and
their relationships to other concept circumscriptions; rather, these are
examples of what I have been calling name-name relationships.

> > the difference between TC1/TC2 and TC3/TC4 is that the former two
> > represent only the name, and the latter two represent the taxon
> > circumscription as defined in the original definition of the concept
>
> I understand, so TC1/TC2 shouldn't have relationships to TC3/TC4.

Correct -- but they probably *should* somewhere have a pointer to the same
publication instance as is indicated in the "AccordingTo" of TC3/TC4.  But
this pointer means something different from what the pointer means in the
"AccordingTo" of TC3/TC4.  Within TC1/TC2, the pointer would mean "name was
established according to the relevant Code in"; whereas the pointer from
within the "AccordingTo" of TC3/TC4 would mean "concept circumscription as
defined in".

I think this is one of the things that people object to when separating out
name objects from circumscription objects.  If you already have an
"AccordingTo" pointer within the "Original" TaxonConcepts, why do you need
to repeat the pointer in the corresponding Nominal Concept.

For example:

Aus bus Jones SEC. Jones

could be thought of as having two separate pointers to the publication
authored by Jones: the first "Jones" pointer means the name-object "was
established according to the relevant Code in" the Jones publication; and
the second "Jones" pointer means the concept circumscription "as defined in"
the Jones publication.

As I understand the TCS approach, the "AccordingTo" is captured only once,
and the "was established according to the relevant Code in" part and the "as
defined in" part are captured via various "Relationships".

> I'm not confident that relationsihps with higher rank name literal
> are purely nominal in sense of name literal, although relationships
> between TC1/TC2 and TC11/TC12 are nomenclatural.  Do you extend
> similer relationships to TC21/TC22, i.e. families and higher?

No -- because no Code includes the Family name-unit as part of the "name"
object.

There would be a TC21/TC22 for family names, but there would be no direct
connection with TC11/TC12 via the Nominal concept, because the original
family placement of the name is not a nomenclatural property.

The only reason there is a nomenclatural connection between TC1/TC2 and
TC11/TC12 is that Linnean nomenclature is binomial, and thus the name-string
"Aus bus" means something different from the statement
"Concept-circumscription Aus includes Concept-circumscription bus".

In the zoological perspective, the "name" would be the terminal epithet
only, and all different genus combinations are concept-circumscription
issues, in the same way that genus-family connections are.  However, the
botanical perspective gives greater respect to the binomial nature of
Linnean names, and treats each combination as a different "name".  This
requires further discussion, but we first have to get past the point where
we decide whether or not to modularize name information within TCS at all,
and then how to do so.

> <TaxonConcepts>
> snip
> </TaxonConcepts>
>
> Now you include author part of name literal in NameSimple...

That was only so that I could show that non-Nominal-type TaxonConcepts have
an "AccordingTo", but Nominal-type TaxonConcepts do not.  In this way,
"AccordingTo" is restricted in meaning to "concept circumscription as
defined in", and does not included information about "name was established
according to the relevant Code in".

It would be better to have a complete dataset to work with.  Perhaps this
weekend I can assemble one.

Aloha,
Rich