Skip to content

2024 2 10

Jump to bottom
Furkan Akkurt edited this page Feb 28, 2024 · 1 revision

10 Feb Meeting

  • Participants:
    • Jonathan Washington
    • Bermet Chontaeva
    • Soudabeh Eslami
    • Chihiro Taguchi
  • Date: 10/2/2024 19:00 (UTC+3)
  • Place: Zoom

Agenda

Discussion

  • topics covered: Today we mainly covered issues found during the annotation of the Kyrgyz/Tatar translation.

Empty forms

Should we allow for empty syntactic words for all Turkic languages? Currently implemented in Kyrgyz. Kyrgyz:

sent_id = 9 (https://github.com/ud-turkic/udtw23/blob/main/Annotations/Kyrgyz.conllu)
…
8-9	эмес	_
8	эмес	эмес
9	_	э

sent_id = 13
…
3-5	бекен	_
3	б	бы
4	екен	экен
5	_	э
6	?	?

Multiple tokens with single surface token

Should we split the polar question morpheme (-mI/bI) as a separate word?

  • Kyrgyz:
    • коёсуңбу → коёсуң + бу
  • Tatar:
    • куясыңмы (currently annotated as one syntactic word)

Converb (-Ip) construction: advcl or root?

  • Kyrgyz:
    • Жубайым дачадагы балдар бөлмөсүнүн терезесин шаша-буша тазалап, ашкананыкын тазалабай коюптур.
    • Should the root be:
      • тазалап, given the principle to set the head in the first element in conjunctive phrases? (conj, parataxis, etc.)
      • тазалабай, given that тазалап is a non-finite verb form (VerbForm=Conv) and provides a background for the main event (тазалабай коюптур).
  • Currently, Tatar is following the second option (converbs are not the root)
    • cf. Washington et al. (2022) Non-finite verb forms in Turkic exhibit syncretism, not multifunctionality. Folia Linguistica 56(3): 693–742.
(14) A verbal adverb in /-GAnI/ in Kyrgyz
Досум			китепти		китепканага	тапшырганы	кезекте		турат	.
dosum			kitepti		kitepqanaʁa	tapʃərʁanə	kezekte		turat	.
dos-(I)m		kitep-NI	lkitepqana-GA	tapʃər-GAnI	kezek-TE	tur-É-t	.
friend-POSS.1SG		book-ACC	library-DAT	turn.in-VADV	line-LOC	stand-NPST-3
annotation 1

  • Kazakh:

    • annotation 2
    • “I almost ate the camel.”
    • The subject (1SG) and the object (the camel) are of the event expressed by the predicate жеп ("eating"); thus, they should have жеп as the head.
    • кете and жаздадым are auxiliaries.

  • cf. Tatar:

    • мин дөяне ашап китә яздым.

Kyrgyz negative past

Мен	барган  жокмун.
1SG	go-VN?	not-1SG
“I didn’t go”
  • forms like барган жокмун are primary negative form of бардым; (барбадым is also possible, but in much more limited contexts). Cf. Turkish forms gittim, gitmedim; барган жокмун would be equivalent to *gitmiş yokum.
  • The problem is, the negation and person agreement of the event “going” is expressed on жок
  • This does not happen in other Turkic languages like Tatar and Turkish
    • Tatar: Мин бармадым (1SG go-NEG-PST-1SG)
    • Turkish: Ben gitmedim
  • Tatar:
    • Барганым юк. (go-PST.PTCP-POSS.1SG not.exist): I have never been there.

Past perfect, even more complicated:

Мен барган эмес элем.
“I hadn’t gone.”

(This would be equivalent to Turkish *Ben gitmiş değil idim; this is negative of Мен барган элем, morphologically and semantically equivalent to Turkish Ben gitmiş idim.)

-AsI/-GI (bar/yoq)

  • Which is the head? The verb? or the existential morpheme?
  • -AsI/-GI: Washington et al. (2022) call it a “volitional non-finite verb form” and consider it to probably be a verbal noun
  • Examples:
    • Turkish: Gidesim yok.
    • Uzbek: Ichgim bor (drink-VN?-1SG exist)
    • possible analysis: VN (VerbForm=Gerund/VerbalNoun) is nsubj dependent of adjective head bar/yoq (in copula construction).
  • Examples of a slightly different strategy -AsI/-GI kil-/kel-
    • Tatar: Барасым килә (go-VN?-1SG come-PRS.3)
    • Kyrgyz: Баргым келет (go-VN?-1SG come-NPST-3)
    • possible analysis: VN (VerbForm=Gerund/VerbalNoun) is nsubj dependent of verb head.

Actions

  • Nothing decided on, no critical mass

Sidebar

Clone this wiki locally