Deos: additional morphology entry from treebank data #13
Description
in the case of deos, we have two definition for deos and the game is set for the less common word with missing declension. I have compared the result with Whitaker and the second
definition (the less common one) does not exist. in the inflection table,
all the matching ending for deos (2nd decl nom and acc and 4th decl acc)
are highlighted. Could it be that declension information for deos is coming
from matching the endings???? two things here:
- the less common word is selected
- based on the partial information on this word, the declension is chosen by the matching
Bridget said that it's related to alpheios-project/components#249 but a slightly different problem.
The disambiguation is failing to find the right entry from the morphological output for deos I think because of the capitalization. It should probably be doing a case-insensitive match, although I am not positive about this, because in fact capitalization can mean a different lemma, as in the case of a proper-name. (I.e God and god are quite legitimately different lemmas). (There is also an interesting parser behavior here with the variation of o vs u in the lemma, because deos gives you only the Deus, Dei lemma, whereas deus gives you both).
it's creating the entry from the treebank data, which specifies the lemma as 'deus' rather than 'Deus'. In this case, I think actually the treebank data is wrong, the lemma should be Deus.
There are other words similar to "deos": me (ego), fugiendo (it does look for fugio not
fugiendo. the matching in the inflection table is wrong), Partheniis (it is a noun but in the morphology pop-up with diagram is an adj)