search

IAHLT / UD_Hebrew

Hebrew Universal Dependencies Treebank
Other
2 stars 2 forks source link

"let us"-"bou" Hebrew construct #38

Closed IsraelLand closed 2 years ago

IsraelLand commented 2 years ago

Hi @amir-zeldes

We've stumbled upon these cases, bou+Verb - "בואו נעשה את הניסיון", "בואו נביא את הבשורה"

Quite common, but maybe not so much in the old Haaretz HTB literary style... It seems that the options for deprels are -

  1. xcomp- this is the the English GUM "let's" usage afaic and should reflect some sort of predication. Question is, if it fits the Hebrew usage
  2. advcl
  3. aux- some sort of an aux structure, like "rotze lalechet" reflecting the grammaticality of "bou/boi/bo+conj. verb" structure, though I'm not sure it's quite the same, it's tough seperating semanticity from function in this case
  4. discourse- kind of a discourse marker like "shma, tire"
  5. or just go the (easy?) depor parataxisroute

I'm undecided between 1 and 3, as I think that "bou" definitely serves a grammatical structure, which involves some 'sub-predication' perhaps Thank you!

Hilla-Merhav commented 2 years ago

@amir-zeldes I don't think it's xcomp because I don’t think it's a core argument, syntactically (in English it is xcomp – let's is a contraction of let us ---> let's do it – let us do it - but in Hebrew it's different). We don't analyze יצא לעשות as xcomp, but as advcl, so I think when we analyze בואו לעשות – (bou, laasot) should be advcl also. image

As for בואו נעשה I am not sure whether it's advcl or parataxis, or maybe it's better נעשה would be the head somehow.

IsraelLand commented 2 years ago

I see. Personally I do think it's a core argument, or a head in itself, as בואו נעשה as a verbal phrase is completely different than נצא לעקוב, where in the latter you actually go out &/then track, whereas with the former the doing is key, i.e. no one is actually coming, just doing (basically, Modern Hebrew grammaticalization)

amir-zeldes commented 2 years ago

This looks to me like what is usually referred to as a serial verb construction, and is also attested in English in things like "to go do something". Some languages use a special type compound:svc for these, see the guidelines here:

https://universaldependencies.org/u/dep/compound-svc.html

Relevant discussions:

My reading of the situation based on the Hebrew speakers who chimed in, and the comparison with English, is that the consensus favors conj, and in any case not advcl or xcomp, because you can insert 've'. I think the languages that have opted for compound:svc are the ones in which this is extremely productive and regular, verging on an auxiliary as part of a paradigm, which forms something like a new "cohortative verb" complex, or "desiderative" (often with "want"), "benefactive" (often with "give") etc., whereas in Hebrew I think it's still more or less a transparent coordination with pragmatic constructional flavor.

IsraelLand commented 2 years ago

Huh, didn't even think of conj., and that it can be used with a coordinator, so I guess conj. makes sense (Though still no one is actually coming, only doing, but it holds true for the English as well) The compound nature of the phrase isn't entirely shown imo by conj either, but I agree it's still much more transparent in Hebrew, still. Thanks.

Hilla-Merhav commented 2 years ago

@amir-zeldes Thanks! Seems reasonable. So conj for (bou, naase), and what about (bou, laasot)? We can't insert 've', and the example from the guidelines seems comparable to me, so I think advcl is right for infinitive after בא, בואו: image

הוא בא לטורקיה לעקוב אחרי נבחרת הנוער הוא בא לטורקיה (כדי) לעקוב אחרי נבחרת הנוער בואו לעשות את הניסיון בואו (כדי) לעשות את הניסיון

amir-zeldes commented 2 years ago

Yes, this seems normal: if "kedey" is possible, it's a purpose advcl, otherwise it's an xcomp infinitive (like "nasu la'asot/xcomp")