Apertium has moved from SourceForge to GitHub.
If you have any questions, please come and talk to us on #apertium on irc.freenode.net or contact the GitHub migration team.

Annotation guidelines for Kazakh

From Apertium
Jump to: navigation, search


[edit] Verbal noun or noun

[edit] Nominal compounds

When choosing between <attr> and <nom> in noun1-noun2 compounds, the choice basically depends on if noun2 is marked for possession. If it is marked for possession then you should chose <nom>, if not, then choose <attr>.

  • <attr>:
    • көрші елдер
  • <nom>:
    • əлем чемпионаты

However, there are cases when noun2 is marked for possession but noun1 is not its possessor, e.g. "жазба әдебиеті" in a phrase "қазақ жазба әдебиеті".

қазақ       жазба          әдебиеті
Kazakh.NOM  written.ATTR   literature.3SG

If we blindly applied the above rule for "жазба әдебиеті", then we would tag "жазба" as <nom>, but actually "әдебиеті" is possessed by "қазақ", not by "жазба". Moreover, it is possible to drop "i" in "жазба әдебиеті", thus "жазба" is <attr>.

[edit] Specific words

[edit] "-DA"

The word "-DA" can be a conjunction or a postadverb:

  • cnjcoo [joins two or more noun/verb phrases; it's conjoining two parallel things in the same phrase, as opposed to saying that it's adding one thing to something from before]
    • Үстелде қалам да, қарындаш та, дәптер де жатыр.
    • Абай әуелі ауылдағы Ғабитхан молдадан сауатын ашады да, 10 жасқа толған соң 3 жыл Семейдегі Ахмет Риза медресесінде оқиды.
  • postadverb [means 'also', 'even', or used for emphasis]
    • Мен де барамын.
    • Аузы қисық болса да байдын баласы сөйлесін.

[edit] "бұл", "мынау", "осы", "мына", "анау", "ана", "сол"

The word "бұл" (along with "мынау", "осы", "мына", "анау", "ана", "сол") can be either a determiner, modifying a noun phrase, or a pronoun, replacing a noun phrase.

  • det.dem
    • Мынау үй жаңа.
  • prn
    • Мынау — терезе емес.

The way to tell is whether it's part of the following noun phrase (det dem) or separate from it (prn)

[edit] Verbs in dictionary form

A verb in a dictionary form (gerund) can sometimes be a noun.

  • ger
    • Кітап оқу адамдарды ақылдырақ етеді.
  • n
    • Оқу басталды.

With gerunds you often get embedded arguments, like "кітап" in the example above.

If there is an adjective or determiner preceding: noun.

Some cases are ambiguous: Балалардың оқуы жақсы болды. Here, it's unclear whether бала is the subject of a verbal noun or the possessor of a noun. In cases like this, the context can sometimes provide some intuition; e.g., in "Балалардың оқуы басталды.", оқу seems more like a noun and less like a gerund. In such ambiguous cases, if you have an intuition, go with that. If not, go with verbal noun (the reason being that we would need more things in our lexicon, and also that these "noun" are derived from verbs anyway).

[edit] Perfect participle or Verbal adverb

Perfect participle (prc_perf) if there is an auxiliary following, otherwise verbal adverb (gna_perf)

  • prc_perf
    • Ол кәзір ұйықтап жатыр.
  • gna_perf
    • Мектепті бітіріп, университетке түстім.
    • Орталық Азиядан арий тайпалары келіп қоныстанды.

There is also other evidence that келіп in the example above is a verbal adverb: In particular, you can put arguments/adjuncts of қоныстанды between келіп and қоныстанды, like "Орталық Азиядан арий тайпалары келіп бұл аймақта қоныстанды." In this case the two verbs function as separate predicates, which participle+auxiliary constructions don't do.

[edit] Futher reading

Personal tools