Context: Aphrodite has beamed up Paris from the battlefield to his love nest. She then drags a protesting Helen through the halls of the castle and shoves her in.
Iliad 3.426-7:
ἔνθα κάθιζ ̓ Ἑλένη κούρη Διὸς αἰγιόχοιο
ὄσσε πάλιν κλίνασα, πόσιν δ ̓ ἠνίπαπε μύθῳ·
It seems clear to me that here ὄσσε πάλιν κλίνασα means that Helen is averting her eyes, with κλίνασα modifying Ἑλένη and taking neuter dual ὄσσε as its object. Cunliffe and Buckley seem to agree.
But the Perseus treebank has this lemmatized as ὄσσα, which I find in dictionaries only as a feminine word for “rumor.” Beekes assigns the words different etymologies, and he calls ὄσσε a dual inherited from pre-Greek. From the context, I can’t see how the meaning could be rumor. According to Cunliffe, the “rumor” word occurs only once in the entire Iliad.
(As an aside, it’s kind of interesting that this word occurs much more frequently in the Iliad than in the Odyssey. I don’t know if this kind of thing can be used as evidence that they were not written by the same author. I guess it’s also possible that the difference is due to the different content of the stories.)
So the possibilities that then occurred to me were that (a) Perseus wants to assign ὄσσε to a singular lemma, so they make a fictitious dictionary form ὄσσα (neuter), or (b) one of the people who constructed the Perseus treebank simply made a one-off mistake.
But actually, of the 56 times that Homer uses ὄσσε to mean “eyes,” 26 of those are lemmatized as ὄσσα, 30 as ὄσσε. So it seems like maybe either half the people who did the treebank for Homer did it differently from the other half, or maybe their software presented the wrong choice as a default, so that sleepy grad students who hadn’t had their second cup of coffee often didn’t realize that they needed to override it.
Does my interpretation make sense? Based on this, I would think that essentially all of these lemmatizations as ὄσσα should be patched for my purposes as a consumer of the treebank.