[Taxacom] Reproducibility of descriptive data

[Mike Dallwitz]

Stephen Thorpe wrote:

> Robust! Another way of saying that species have a real, natural 
> boundaries! Absolutely, yes! :)

The principles I'm writing about have nothing to do with taxonomic levels, 
as I tried to make clear by writing "a class ... (such as a subspecies, 
species, or genus)". Indeed, they apply to classes of non-living things 
too, e.g. books.

By 'robust description' or class definition, I mean one such that it will 
usually be fairly obvious whether or not a new specimen belongs to the 
class. If a new specimen matches the definition fairly well, we can assign 
it to the class (and, if we can be bothered, adjust the definition as 
necessary), without agonizing over it too much.

(Note that this is not, by any means, the detailed discussion that I 
threatened to give!)

Of course, if the class has a 'natural' boundary, with a moat or wall 
separating it from other classes (for whatever reason), making a robust 
description will tend to be easier. Genera can have moats at their 
boundaries, too.

>> Which comes first, the concept or the inclusions (exemplars)?
> 
> Actually, you mean which defines the concept, the description or 
> the inclusions (exemplars)?

No, Jim and I were both talking (I think!) about the mental process of 
forming classes - which everyone does, not just taxonomists. The process 
of making definitions for communicating the class concept to others comes 
later, if it's done at all.

-- 
Mike Dallwitz
Contact information: http://delta-intkey.com/contact/dallwitz.htm
DELTA home page: http://delta-intkey.com