[Taxacom] Reproducibility of descriptive data

[Edwards, James]

Stephen -

Your argument doesn't hold water. Let's for the moment accept that you are correct that there are "hard" species boundaries in nature and all we have to do is discover them (which I don't believe is true, but am willing to accept for argument's sake). The problem is that we do not *know* what these boundaries are. We have to infer them.

In your scenario, Smith and Jones have published different descriptions (concepts) of what is included in the species. We don't know which one of them is "correct" (as you admit in your message below). So whenever a *good* taxonomist (i.e. one who knows the literature and therefore knows that there are differences of opinion about the species concepts) makes an identification (be it in a paper or on an herbarium sheet or on a specimen tag or in a database entry), it would help all future workers if that individual indicates which concept s/he had in mind when making the identification. That is all that Jim Croft (and many others) have been arguing for. It doesn't depend on whether the data are in a database or elsewhere; it is just good scientific practice, similar to describing whose protocol you used when doing an experiment.

If, in your scenario, Brown is making identifications based solely on comparison to a holotype (rather than using the literature and examining the additional specimens that Smith and Jones include in their species circumscriptions), then Brown is not a very good taxonomist. As many previous posters have said, holotypes are exemplars and name bearers; they are not species descriptions. In the most trivial case, for example, a single holotype cannot encompass sexual dimorphism.

Finally, it is immaterial whether someone else later has a different opinion of where the species boundaries lie. We can't expect anyone to be able to divine future events. What is important is that *at the time the person made the identification*, s/he was making it based upon a particular published species concept and identifies that concept along with the identification. If the identifier does not include which species concept was used, then future workers will have great difficulty in understanding what s/he meant.

            --  Jim

Dr. James L. Edwards
Executive Director
Encyclopedia of Life
National Museum of Natural History
Smithsonian Institution
P.O. Box 37012, MRC 106
Washington, DC 20013-7012

Phone:  1 202 633 8730
Mobile:  1 571 230 4098
Fax:      1 202 633 8742
Email:   edwardsjl at si.edu
-----Original Message-----
From: Stephen Thorpe [mailto:s.thorpe at auckland.ac.nz]
Sent: Saturday, September 12, 2009 10:14 PM
To: Edwards, James; Jim Croft
Cc: TAXACOM; Mike Dallwitz
Subject: RE: [Taxacom] Reproducibility of descriptive data

Hi again Jim :)

As a first step to answering you on this, let me restate what I originally said, as you seem to be unsure how to interpret my words on some points:

taxonomist Smith considers B to be a ("subjective") synonym of A (and publishes the synonymy)
taxonomist Jones disagrees and considers them to be distinct (and publishes the "non-synonymy" - let us please avoid the issue of who (Smith or Jones) published first or last!)
identifier Brown identifies a specimen as A
Jim (Croft) wants to include Brown's identification in his database, but exclaims "by crikey, mate! Did you mean that they are A sensu Smith, or A sensu Jones?"

The points to ponder about this example are as follows:

(1) is Jim asking a well-defined, sensible, meaningful question; and

(2) if so, does the answer matter? Does it help us in any non-trivial way?

One thing is clear: if the specimen identified by Brown as A is a "typical A", then it doesn't matter which sense it was identified as (Smith or Jones). But if it is a "typical B", then the two senses give different answers (Smith = yes, Jones =no)

Another thing also seems clear: if species boundaries are purely subjective, like I say genera are, and as Richard Pyle seems to be suggesting also for species, then there is no objective fact of the matter regarding whose concept is "the correct one" (Smith's or Jones'). If that is so, then an identification would simply be correct or not relative to a species concept, and not depending on any facts in the world.

Let us first consider point (1) above: my concern over this is that the basis for Brown's identification could be one or more of any number of things, possibly unrelated in any obvious way to Smith or Jones' published "concepts". Brown might not even be aware of Smith or Jones. Brown might have just looked at the holotype of A, compared it to his specimen, said (to himself) "well it looks pretty darn close to me!" and identified it as A on that basis. If the specimen had of been a "typical B", Brown may have scratched his head and said "I can't decide!" But , supposing that the specimen happened to be a "fairly typical A", this didn't happen and Brown felt confident enough to make a decision. I suggest that a scenario more or less like this is typical of how identifications are actually done in practice. But where does it leave Jim's question? I am unsure ...

Regarding point (2): again it is not entirely clear to me at this point, but I think that IF species boundaries are purely subjective, AND Jim's question above is meaningful and well-defined, THEN the answer might be useful and helpful. But, I don't of course think that species boundaries are purely subjective! If I am right, then at least one thing is clear, namely that either Smith or Jones is correct, and the other the opposite. So when Jim asks "by crikey, mate! Did you mean that they are A sensu Smith, or A sensu Jones?", the correctness of the identification might hang on the answer to that (if the specimen is a "typical B", anyway). Typically, we won't know who is correct - Smith or Jones. What Jim really wants to know is "is Brown's identification correct or not?" If Jim knew that the specimen was a "typical B", then this would help to know if the identification was correct, given a decision about the synonymy being correct or not. So, the idea seems to be to keep track of this information until such time as a decision is made, and then we can retrospectively re-evaluate all past identifications without having to track down and revisit all the actual specimens. BUT, to my mind, this all collapses to needing to have a "complete" description of the specimen identified in order for the identification of it to be meaningful! It has nothing to do with Smith or Jones, but rather has everything to do with whether the specimen has characters consistent or not with whichever "concept" of the species is eventually decided upon (if such a decision is ever made!). But this is absurd, and is precisely one of the main reasons why we should keep safe specimens that have been identified - so we can go back to them if "concepts" change, and re-evaluate them!

Think of it this way. Brown identifies the specimen. Then, a year LATER, Steve comes along and lumps C into "A sensu Smith", to get "A sensu Steve"! What use is Jim's question now? We need to retroactively know if Brown identified the specimen as A sensu Steve, A sensu Smith, or A sensu Jones! THAT was what I meant when I said 'It also doesn't make much sense to me to give Smith and Jones' concepts any special status compared with the huge number of other POSSIBLE concepts (defined by synonymy) for the same name, that nobody has proposed yet!'

I think that what this all boils down to is this:
(Species level) taxonomy isn't an exact science. But this doesn't mean (contra Richard) that it is purely subjective! It is just an "inexact science"! People can cope with this by just getting on with it, but I'm not sure about computers ...

Cheers,

Stephen

________________________________________
From: Edwards, James [EDWARDSJL at si.edu]
Sent: Sunday, 13 September 2009 9:27 a.m.
To: Stephen Thorpe; Jim Croft
Cc: TAXACOM; Mike Dallwitz
Subject: RE: [Taxacom] Reproducibility of descriptive data

Stephen Thorpe said:

It also doesn't make much sense to me to give Smith and Jones' concepts any special status compared with the huge number of other POSSIBLE concepts (defined by synonymy) for the same name, that nobody has proposed yet!

My response:

In this scenario, Smith and Jones have presumably published their taxon concepts. Otherwise we have no way to know what their concepts are.

If you are saying that unpublished concepts don't have special status, then I fully agree with you, because unpublished concepts are, for all intents and purposes, of no consequence. Since they are not published, we have no way of intuiting them.

But published concepts, of course, do have special status, by virtue of having been published. It is ludicrous to say they are no different from all other possible concepts that anyone might think up in the future. We have to deal with species determinations here and now, based upon what people have published in the literature.

            --  Jim

Dr. James L. Edwards
Executive Director
Encyclopedia of Life
National Museum of Natural History
Smithsonian Institution
P.O. Box 37012, MRC 106
Washington, DC 20013-7012

Phone:  1 202 633 8730
Mobile:  1 571 230 4098
Fax:      1 202 633 8742
Email:   edwardsjl at si.edu

-----Original Message-----
From: taxacom-bounces at mailman.nhm.ku.edu [mailto:taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Stephen Thorpe
Sent: Friday, September 11, 2009 12:55 AM
To: Jim Croft
Cc: TAXACOM; Mike Dallwitz
Subject: Re: [Taxacom] Reproducibility of descriptive data

Just to throw a spanner in the works (!), there is something else that Jim might "really mean" by species concepts. If so, then such concepts are defined by inclusions (of specimens), but not quite in the way that Jim seems to think. Consider this:
taxonomist Smith considers B to be a ("subjective") synonym of A
taxonomist Jones disagrees and considers them to be distinct
identifier Brown identifies some specimens as A
Jim wants to include Brown's identification in his database, but exclaims "by crikey, mate! Did you mean that they are A sensu Smith, or A sensu Jones?" It might not matter if the specimens are "typical A", and so fit both concepts equally well, but what if they are "typical B"???
So, there appears, at least on first glance, to be a notion of species concept defined by inclusion of synonyms (i.e., type specimens). So, now the questions become: (1) is that notion well-defined; and (2) is it important?

I think that, although it is called "subjective synonymy", this is misleading for species. The distinction is really one of what you might call "taxonomic synonymy" vs. "nomenclatural synonymy", not "subjective" vs. "objective"! On my "natural boundary" view of species, Smith is (objectively) either right or wrong, and Jones the opposite.

Note that Brown might have misidentified the specimens altogether, so they may be neither A sensu Smith, nor A sensu Jones! On my view, either A sensu Smith or A sensu Jones is a misidentification anyway, so isn't what we really want to know "is Brown's identification correct?", and not whether it is correct sensu Smith or sensu Jones?

If we don't know which sense of A (Smith or Jones) is the correct one, which is the typical situation, then it seems to make sense to keep track of which sense specimens have been identified. But is this possible in practice? If possible, is it even worthwhile? Remember that Brown could have misidentified the specimens for any number of reasons, not just by choosing to follow Smith over Jones or vice versa. Perhaps, Brown just saw a specimen, identical to his specimens, which had a spurious label on it saying "Holotype of A"! Maybe we could track down that specimen and see if it fits A sensu Smith, or A sensu Jones, or A sensu neither ... Doesn't sound very practical to me! To my mind, it seems to collapse to the absurdity of claiming that an identification/name instance has no meaning unless you can verify that it isn't a misidentification!

I think that whenever we encounter a name instance, we must be aware that it may be a misidentification, but we can't hope to track the reasons why it might be a misidentification (= the basis for the identification), because that basis may be a complex and vague mixture of factors, and not just simply a matter of having followed either Smith or Jones. It also doesn't make much sense to me to give Smith and Jones' concepts any special status compared with the huge number of other POSSIBLE concepts (defined by synonymy) for the same name, that nobody has proposed yet!

These are just my preliminary thoughts on this matter...

Stephen

________________________________________
From: Jim Croft [jim.croft at gmail.com]
Sent: Friday, 11 September 2009 9:15 a.m.
To: Stephen Thorpe
Cc: Mike Dallwitz; TAXACOM
Subject: Re: [Taxacom] Reproducibility of descriptive data

Descriptions do not 'define' a concept (at any level). They... um...
describe it.

If you want a definition, you could use the list of those thing you
might include in that concept. Species in genera, specimens in
species, etc.

Having settled on a concept by inclusion, you can then go about
describing it, listing the characters/attributes that, in your mind,
set the boundaries.  It is conceivable that a taxonomist could account
for all relevant specimens, species, etc.  This is, after all, why we
do revisions.  Any character/attribute list is arbitrarily selected
and can never be complete.

Your 'country' analogy is spurious.  Like the US, Australia is an
historical federation, defined by inclusion of various bits of land,
including disjunct offshore islands and territories.  That the bulk of
Australia happens to coincide with recognizable continent is an
irrelevant artefact.  Notions of artificial or real in this context
have no meaning.

jim


On Thu, Sep 10, 2009 at 9:26 AM, Stephen Thorpe<s.thorpe at auckland.ac.nz> wrote:
> Jim, put another way:
>
> Genera are defined by descriptions (= genus boundary circumscriptions)
>
> Species are defined by nominating an individual as the type. Species descriptions (=species boundary circumscriptions) do not define the species. They can be incorrect descriptions of the true boundaries (unlike generic descriptions)
>
> Analogy: Australia can be defined by sticking a flag in the ground and saying "I hereby define Australia to be all the land in all directions from this flag to the sea". So, Australia is like a species (it has natural boundaries). U.S.A. is like a genus (it has artificial boundaries).
>
> Your "taxon concepts" are a mixture of two very different things: (1) generic descriptions; and (2) species descriptions.
>
> ________________________________________
> From: Jim Croft [jim.croft at gmail.com]
> Sent: Thursday, 10 September 2009 9:46 a.m.
> To: Stephen Thorpe
> Cc: Mike Dallwitz; TAXACOM
> Subject: Re: [Taxacom] Reproducibility of descriptive data
>
> Don't buy this.  At all.  And I do not think the codes do either.  Nor
> many/most taxonomists.   The type does not define the species (which
> are in nearly every case variable).  It is an exemplar (not always
> 'typical' in the English sense) which anchors the name.  The extreme
> example of this are species that have multiple synonymic types.  In a
> type-defined species, concepts of lumping and splitting have no
> meaning - yet we all do it.
>
> Have a chat to Pete deVries.  He would argue that a species knows what
> a species is and does not care what we call it or think it is.  Humans
> develop a concept of what we think it is, sometimes (maybe even often)
> a reasonably good approximation of what a species knows it is.  And we
> give a name to this human concept a name.
>
> There are three things: a species entity, a species concept and
> species name.  The first is defined by biology and evolution, the
> second by humans, and the third is defined by the code and selected by
> humans.
>
> The problem we have, and why taxacom exists at all, is someone utters
> the third, a listener assumes the first, without considering the
> second, of which there are often several alternatives.
>
> jim
>
> On Thu, Sep 10, 2009 at 7:03 AM, Stephen Thorpe<s.thorpe at auckland.ac.nz> wrote:
>> [Mike Dallwitz wrote] _Whatever_ we want to say about a taxon (e.g. what its boundaries, distribution, abundance, or uses are), we need to define the
>> taxon that we want to talk about. And the only way to do that is to describe it in a reproducible way, so that people can identify individuals as belonging or not belonging to the taxon
>>
>> [reply] Species are defined by their name-bearing types (holotypes or lectotypes or neotypes or syntypes). A description of a species is a circumscription of its boundaries, according to the describer. So, we don't describe a taxon in order to define it so that we can then talk about boundaries. Rather, by describing it, we ARE talking about its boundaries, but the species is defined by its type.
>>
>>>describe it in a reproducible way, so that people can identify individuals as belonging or not belonging to the taxon
>> No, describing it in a reproducible way only allows people to identify individuals as being within or else outside the boundaries of the species as circumscribed in the description. These boundaries could be wrong, so the description is certainly not a DEFINITION of the species (definitions are true by definition and cannot be wrong!)
>>
>> What you say applies more to genera and other "subjective" taxa, but not to species, which are objectively defined once a type is designated...
>>
>> Stephen
>>
> --
> _________________
> Jim Croft ~ jim.croft at gmail.com ~ +61-2-62509499 ~
> http://www.google.com/profiles/jim.croft
> ... in pursuit of the meaning of leaf ...
> ... 'All is leaf' ('Alles ist Blatt') - Goethe



--
_________________
Jim Croft ~ jim.croft at gmail.com ~ +61-2-62509499 ~
http://www.google.com/profiles/jim.croft
... in pursuit of the meaning of leaf ...
... 'All is leaf' ('Alles ist Blatt') - Goethe
_______________________________________________

Taxacom Mailing List
Taxacom at mailman.nhm.ku.edu
http://mailman.nhm.ku.edu/mailman/listinfo/taxacom

The Taxacom archive going back to 1992 may be searched with either of these methods:

(1) http://taxacom.markmail.org

Or (2) a Google search specified as:  site:mailman.nhm.ku.edu/pipermail/taxacom  your search terms here