search

IAHLT / UD_Hebrew

Hebrew Universal Dependencies Treebank
Other
2 stars 2 forks source link

Naming construction revisions #3

Open amir-zeldes opened 3 years ago

amir-zeldes commented 3 years ago

Naming constructions with a secondary predication should have xcomp to the name

And similarly נקרא etc.

Note that just introducing a name is not necessarily xcomp and can be appos, e.g.:

השם דני

Hilla-Merhav commented 3 years ago

@amir-zeldes Is this rule relevant also for the construction מסוג? For example – צהבת נגיפית מסוג הפטיטיס תרופה אנטיביוטית מסוג בטא לקטם Is it – xcomp(soog, Hepatitis) like ילד בשם דני is xcomp(shem, Danny)?

amir-zeldes commented 3 years ago

My gut feeling is no, I'd do it literally as appos, not sure why I feel that way. In English I'm pretty sure "type" wouldn't take xcomp. Maybe for literal double object siveg without ke- I would do it with xcomp:

hu siveg oto "mesukan"

xcomp(siveg, mesukan)

strasss commented 3 years ago

@amir-zeldes @Hilla-Merhav The problem with using appos for סם מסוג קוקאין and the like, I think, is that it doesn't account for the fact that סוג is indefinite.

So, similarly to how xcomp and appos are used with 'beshem' and 'bashem' respectively, I agree with Hilla that the different deprels can help us make the distinction between צהבת מסוג הפטיטיס סי and צהבת מהסוג הקטלני הפטיטיס סי.

amir-zeldes commented 3 years ago

OK, you've convinced me - I can't find a way how this is syntactically distinguishable from בשם.