search

IAHLT / UD_Hebrew

Hebrew Universal Dependencies Treebank
Other
2 stars 2 forks source link

secondary predication - noun vs. predicative - esp. after PAUL verbs #66

Open IsraelLand opened 2 years ago

IsraelLand commented 2 years ago

@amir-zeldes

העלאה טעונה אישור של ועדת המחירים vs. מצבם של החולים מוגדר קשה

While it seems the second is a secondary predication xcomp (though "mugdar [ke]kashe" :npmod is possible), I'm not sure if the first one is similar.

it seems ההעלאה טעונה את האישור של ועדת המחירים Is a fairly regular phrasing, and I wonder if the et marker is misleading. While "regel asuya etz" is obl:npmod, there's no - *הרגל עשויה את העץ

So either -

  1. obj, which the et points at, but may be misleading, and in any case problematic with a passive paul.
  2. obl:npmod - but said differences still apply...
  3. xcomp as secondary predication similar to "mugdar kashe" - what would indeed be the test?

Can the test between those be, if the connected element is predicative (hence clausal?), go for xcomp, otherwise obl:npmod?

Thank you

amir-zeldes commented 2 years ago

I think "mugdar kashe" is xcomp - but the one you're less sure about is "te'una" right?

Looking at it as a passive seems a bit implausible to me synchronically... Couldn't we just say it's a lexicalized ADJ at this point? Adjectives rarely take objects, but they can, and this was confirmed by Dan and Chris at some point for cases like "it's worth it/obj", which take ACC marking in a variety of languages. So I guess my vote would be for ADJ + obj for "te'una".

IsraelLand commented 2 years ago

Great! Did not expect that ;) Case=Acc as well, for adj?

So if I get it right -

xcomp - מוגדר קשה obl:npmod - השולחן עשוי עץ הרצפה עשויה אריחים obj (+adj) - ההעלאה טעונה אישור ?ההעלאה דרושה אישור?

Is the predicative test useful in determining between options (leaving the verb vs. adj issue for a minute), meaning if the connected element is predicative, xcomp, otherwise obl:npmod\obj? Or is this more complicated than that (i.e. "asuy etz" and "taun ishur" are both nouns but still get different treatments)?

Thanks

amir-zeldes commented 2 years ago

Those look right to me. Secondary predication indicates xcomp, but I'm not sure if there are no exceptions to this. "asuy" sounds to me like the depenent is just oblique, and indeed it is not right to say that the table is "etz", it's just made of that substance. Also note that verbs meaning 'become' typically get analyzed as xcomp, so you get things like "hamayim naasu kerax/xcomp"