search

UniversalDependencies / UD_Irish-IDT

Irish data
Other
6 stars 7 forks source link

Nouns v Adjective - some cases unclear #89

Closed tlynn747 closed 3 years ago

tlynn747 commented 3 years ago

Unclear whether NOUN our ADJ should be used in some case. Might not be so easy to automate this cleanup...

Inconsistencies arose in original source data (3000 pos-tagged corpus).

1207 tíortha forbartha forbartha forbartha ADJ Adj VerbForm=Part or 1426 cúrsaí forbartha forbartha forbairt NOUN Noun Case=Gen|Gender=Fem|Number=Sing

sent 1248 Arm Slánaithe

Cosanta - another example

michealjohnny commented 3 years ago

-- Just a comment, sorry I cant offer a fix -- This issue has caused a lot of Irish-language experts a lot of trouble - the terminology committee have to spend a lot of time on constructions like this. Some can be worked out manually with considerable effort, others can only really be disambiguated by the speaker/writer really. Why? In the broader sense, which POS is chosen can have implications on a séimhiú being used on the modifier or no where the head noun is feminine. Also: According to some/former Irish grammars/standards those forms that could take a séimhiú need not in every situation, as a means of differentiating between meanings when the modifier is a noun.

So you would be right to say automating the clean up will be problematic.

kscanne commented 3 years ago

I'm going through these now; there are about 1000 ambiguous cases. Most are clear in context so far... I'll post any tricky cases here if necessary.