Taxacom: Can GPT-4 replace us

[Kipling W Will]

Prompted by Doug Yanega's recent email suggesting technology might be part
of the cure for the binomial gender agreement issue we nearly all face at
some point, I submitted 5 queries to chat GPT-4 to see how the LLM AI would
do. I submitted the question in this form:

"The carabid beetle species Loxandrus tropicus Allen, 1972 has been moved
from Loxandrus to the genus Oxycrepis. How should the new binomial
combination be formed to have the correct gender agreement?"

swapping out the taxon names each time. These are all real names and real
changes I made in my 2020 reclassification* of the group.

It performed about as good as a freshman who has little or no Latin
training or someone that learned Latin from old Warner Bros. cartoons. So
it seems I am still useful, for the moment. The group has >400 valid
species and a large number of them moved from a masculine to feminine
genus. I probably did not get 100% correct since it can sometimes be a
challenge to figure out which epithets are variable (adjectival) or stable
(nouns). I suspect with relatively modest training and a little human
guidance a task-specific LLM AI could handle the vast majority of cases,
leaving the hard ones for subscribers of Taxacom fight over solve.

enjoy,

Kip


 Copied text responses from GPT-4 with a comment by me in brackets at the
end:

1.

To form the correct binomial combination with gender agreement for the
carabid beetle species formerly known as Loxandrus tropicus Allen, 1972,
now placed in the genus Oxycrepis, you would follow the rules of gender
agreement in binomial nomenclature.

The genus name "Oxycrepis" is neuter in gender, so the species epithet
should also be in agreement with the neuter gender. In this case,
"tropicus" needs to be modified to agree with the neuter gender of the new
genus.

To achieve gender agreement, the modified species epithet would be
"tropicum" instead of "tropicus." Therefore, the new binomial combination
for the carabid beetle species in the genus Oxycrepis would be "Oxycrepis
tropicum" (Allen, 1972).

*[This is not correct. Oxycrepis was originally and always treated as
feminine.]*

2.

To form the correct binomial combination with gender agreement for the
carabid beetle species Loxandrus gagatinus Laporte, 1867, which has been
moved to the genus Zeodera, you would follow the rules of gender agreement
in binomial nomenclature.

The genus name "Zeodera" is feminine in gender, so the species epithet
should also be in agreement with the feminine gender. In this case,
"gagatinus" needs to be modified to agree with the feminine gender of the
new genus.

To achieve gender agreement, the modified species epithet would be
"gagatina" instead of "gagatinus." Therefore, the new binomial combination
for the carabid beetle species in the genus Zeodera would be "Zeodera
gagatina" (Laporte, 1867).

*[This is correct]*

3.

To form the correct binomial combination with gender agreement for the
carabid beetle species Feronista monteithi Baehr, 2007, which has been
moved to the genus Cerabilia, you would follow the rules of gender
agreement in binomial nomenclature.

The genus name "Cerabilia" is feminine in gender, so the species epithet
should also be in agreement with the feminine gender. In this case,
"monteithi" needs to be modified to agree with the feminine gender of the
new genus.

To achieve gender agreement, the modified species epithet would be
"monteithiae" instead of "monteithi." Therefore, the new binomial
combination for the carabid beetle species in the genus Cerabilia would be
"Cerabilia monteithiae" (Baehr, 2007).

*[Not correct for several, obvious reasons.]*

 4.

To form the correct binomial combination with gender agreement for the
carabid beetle species Loxandrus ater Tschitscherine, 1901, which has been
moved to the genus Oxycrepis, you would follow the rules of gender
agreement in binomial nomenclature.

The genus name "Oxycrepis" is feminine in gender, so the species epithet
should also be in agreement with the feminine gender. In this case, "ater"
needs to be modified to agree with the feminine gender of the new genus.

To achieve gender agreement, the modified species epithet would be "atra"
instead of "ater." Therefore, the new binomial combination for the carabid
beetle species in the genus Oxycrepis would be "Oxycrepis atra"
(Tschitscherine, 1901).

*[Correct. But, see #1 where chat gpt said Oxycrepis was neuter.]*

 5.

To form the correct binomial combination with gender agreement for the
carabid beetle species Loxandrus discolor Allen 1972, which has been moved
to the genus Oxycrepis, you would follow the rules of gender agreement in
binomial nomenclature.

The genus name "Oxycrepis" is feminine in gender, so the species epithet
should also be in agreement with the feminine gender. In this case,
"discolor" needs to be modified to agree with the feminine gender of the
new genus.

To achieve gender agreement, the modified species epithet would be
"discoloris" instead of "discolor." Therefore, the new binomial combination
for the carabid beetle species in the genus Oxycrepis would be "Oxycrepis
discoloris" (Allen 1972).

*[Nope. Shouldn’t change. Note it followed my formatting of no comma
between author and year in this case.]*


***Despite my reclassification and making all the nomenclatural changes, a
majority of web resources still are using the old, wrong names. Will, K. W.
(2020). Phylogeny and classification of the genus-group taxa of Loxandrina
(Coleoptera, Carabidae, Abacetini). Deutsche Entomologische Zeitschrift,
67(2), 151-182. doi:10.3897/dez.67.55985