Taxacom: demystifying gender agreement ( was Re: Removals ofoffending scientific names)

[David Campbell]

>
>  An alternative solution would be to abandon the idea that any particular
> gender form of a name is the correct one. All gender variants of a name are
> the same name in effect, just with different gender forms. Hence, one could
> just use any gender variant of a name and anyone searching for the taxon
> would just have to search for the different variants (masculine, feminine
> and neuter), but would not need to know which of them is correct. We sort
> of already have to do that anyway for names with uncertain or disputed
> gender. We also have to do that with different combinations of the same
> name.
> Stephen
>

Regardless of whether this particular solution is deemed the most
practical, it raises the important point that awareness of alternate gender
forms is essential for distinguishing between versions of the same name
versus distinct names.  Even for variants that are not legal under the
current Code, recognizing grammatical variation is useful.  For example,
Galba viator is widely listed as Galba (or Lymnaea or various other genera)
viatrix.  The original description used viator, which as a noun does not
change under the current rules. Viator is a male traveller and viatrix is a
female traveller; the original description used a misspelled masculine
version of Lymnaea as the genus, which may have prompted the masculine
noun.  Probably shifting to viatrix reflected a feminine genus.  (The snail
itself is hermaphroditic and probably doesn't care what we call it.)  A
curious mixing of biological and grammatical gender is found in the case of
the deliberate change of the genus Viviparus to Vivipara, on the grounds
that one having live birth is feminine.  That change is also not allowed
under the current code, but does affect tracing the genus in literature.

Similar issues occur with cases where there are alternate spellings for
Greek or Latin terms.  I have seen a label written by someone who mixed up
sinensis and chinensis, for example.  The zoological Code has provisions
for common alternatives in transliteration (article 58).

Any effective AI approach will have to incorporate the recognition that two
similar names have a chance of being variants of the same name, as well as
the fact that two identical names might be homonyms.  It also must have
effective tools for submitting and verifying corrections.  As is, badly
done AI results are often treated as authoritative and get more support
than the basic work of generating and verifying the data.

-- 
Dr. David Campbell
Associate Professor, Geology
Department of Natural Sciences
110 S Main St, #7270
Gardner-Webb University
Boiling Springs NC 28017