With Greek, because we get the short definitions separately from the morphological parse, we are able to retrieve definitions for the lemmas identified by the treebank because we retrieve short definitions separately from the morphological parse. With Latin, we get our short definitions from Whitaker, in the same step as the morphological parse, and so if the treebank identifies a lemma that was missing from the Whitaker parser result, we don't get any definition, because we don't have a separate short definitions index for Latin. This is not good and we need to fix it.
Examples of words which cause a problem
nos, and mihi because Whitaker doesn't report the lemma ego as a possible parse
quaeque which when lemmatized as quisque
With Greek, because we get the short definitions separately from the morphological parse, we are able to retrieve definitions for the lemmas identified by the treebank because we retrieve short definitions separately from the morphological parse. With Latin, we get our short definitions from Whitaker, in the same step as the morphological parse, and so if the treebank identifies a lemma that was missing from the Whitaker parser result, we don't get any definition, because we don't have a separate short definitions index for Latin. This is not good and we need to fix it.
Examples of words which cause a problem
nos, and mihi because Whitaker doesn't report the lemma ego as a possible parse
quaeque which when lemmatized as quisque