Dependencies
Note: nmod, neg, and punct appear in two places.
|
|
|
|||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||
|
|
|
acl
: clausal modifier of noun
acl
stands for finite and non-finite clauses that modify a noun, in
contrast to the advcl relation which is used for adverbial clauses
that modify a predicate. The head of the acl
relation is the noun
that is modified, and the dependent is the head of the clause that
modifies the noun.
We use acl
for:
Verbal adjectives that modify nouns:
Тез жүгіретін спортшылар көрдім . \n Fast running athletes saw-I .
acl(спортшылар-3, жүгіретін-2)
advmod(жүгіретін-2, Тез-1)
obj(көрдім-4, спортшылар-3)
punct(көрдім-4, .-5)
This relation may be tagged with acl:relcl
in the future, as it is the way
in which Turkic languages do relative clauses.
Gerunds in genitive modifying a noun:
Некенің тоқтатылуы жаңадан некеге тұру құқығын туғызады . \n Marriage-of terminating new-from marriage-to ? right is-born .
acl(құқығын-6, тұру-5)
Fronted relative clauses with empty copula:
Шашы ұзын қызды көрдім . \n Hair long girl saw-I .
acl(қызды-3, ұзын-2)
obj(көрдім-4, қызды-3)
subj(ұзын-2, Шашы-1)
punct(көрдім-4, .-5)
Conditional phrases with ‘болса’:
Құлдық пен құл саудасына , қандай түрде болса да , тыйым салынады . \n Slavery and slave trade , whichever form-in may-be also , prohibition is-placed-on .
acl(түрде-7, болса-8)
advmod(болса-8, да-9)
Secondary predication:
advcl
: adverbial clause modifier
Adverbial clause modifiers (advcl
) are subordinate clauses that
are not complements.
Note that unless there is a separate subject for the “subordinate” clause, the subject will be the same as for the main clause, but is not directly connected.
Айгүл санап біткеннен кейін айналасына қарады . \n Aygül counting finishing-from after around looks .
advcl(қарады-6, санап-2)
aux(санап-2, біткеннен-3)
case(біткеннен-3, кейін-4)
subj(қарады-6, Айгүл-1)
obj(қарады-6, айналасына-5)
advmod
: adverb modifier
The dependency type advmod
is used for adverb modifiers of verbs,
nominals and adverbs alike.
Governed by a verb:
Бақыткүл мүлдем өзгерді . \n Baqıtkul completely changed .
advmod(өзгерді-3, мүлдем-2)
nsubj(өзгерді-3, Бақыткүл-1)
punct(өзгерді-3, .-4)
Governed by an adjective:
Бірақ кеше өте суық еді ! \n But yesterday very cold was-it !
advmod(суық-4, өте-3)
cop(суық-4, еді-5)
Note that adjectives used adverbially are also given this relation.
amod
: adjectival modifier
Nouns may take adjectival modifiers, which are marked with the
dependency type amod
.
Азамат үлкен бір ағаш көрді . \n Azamat big one tree sees .
amod(ағаш-4, үлкен-2)
subj(көрді-5, Азамат-1)
obj(көрді-5, ағаш-4)
Ordinal numerals:
The amod
relation is also used for ordinal numerals. Note that these may not be explicitly marked in the morphology.
1968 жылдан бастап Ширазда театр фестивалы өткізіліп тұрды . \n 1968 year-from starting Shiraz-in theatre festival take-place does .
amod(жылдан-2, 1968-1)
nmod(бастап-3, жылдан-2)
Қызметін 1933 жылы ғана бастаған . \n His-service 1993rd year only he-began .
amod(жылы-3, 1993-2)
Locative attributives:
Substantives in the attributive form of the locative (-DAGI) are also marked with amod
. (Pending clarification)
Олар әрдайым үлкен үйдің алдындағы бақшада бірге ойнайды . \n They always big house-of front-in garden-in together play .
amod(бақшада-6, алдындағы-5)
nmod(ойнайды-8, бақшада-6)
subj(ойнайды-8, Олар-1)
appos
: apposition
An appositional modifier of a noun is a nominal immediately following the first noun that serves to define or modify that noun. It includes examples in parentheses, as well as defining abbreviations in one of these structures.
Олар екеуі де көңілді үйге қарай келе жатыр . \n They both also happily house-to towards coming go .
appos(Олар-1, екеуі-2)
nsubj(келе-7, Олар-1)
aux
: auxiliary
In Kazakh, only a closed list of verbs can act as auxiliaries. These are words take part in participle constructions.
The full list of auxiliaries is as follows:
- ал- “do, make”
- баста- “start”
- бер- “”
- біт- “”
- бітір- “”
- бол- “be”
- ет- “”
- жазда- “”
- жат- “”
- жүр- “”
- қал- “”
- кел- “come”
- кет- “”
- көр- “see”
- отыр- “”
- сал- “”
- тұр- “stand”
- шық- “”
Note: Some of these are ambiguous with lexical verbs.
Examples
Ол терезеден Азамат пен Айгүлдің ойнағанына қарап тұр . \n She window-through Azamat and Aygül's playing watching stands .
nsubj(қарап-7, Ол-1)
nsubj(ойнағанына-6, Айгүлдің-5)
aux(қарап-7, тұр-8)
punct(қарап-7, .-9)
nmod(қарап-7, терезеден-2)
cc(Айгүлдің-5, пен-4)
conj(Азамат-3, пен-4)
ccomp(қарап-7, ойнағанына-6)
auxpass
: passive auxiliary
This document is a placeholder for the language-specific documentation
for auxpass
.
case
: case marking
The dependency type case
is used for the postposition in
postpositional phrases. The head of
a postpositional phrase is the nominal, not the postposition, so as to
analyse postpositional phrases similarly to nominal modifiers without a
postposition. (Such nominal modifiers are frequent in Kazakh, as cases
are often used for the same purpose as postpositions.) To the same end,
the type case
is used in combination with the type nmod
, which is
also used for nominal modifiers when no adposition is present (see
nmod).
Азамат ағашқа қарай жүгірді . \n Azamat tree-to towards ran .
nsubj(жүгірді-4, Азамат-1)
case(ағашқа-2, қарай-3)
punct(жүгірді-4, .-5)
nmod(жүгірді-4, ағашқа-2)
Note that case is not used with auxiliary nouns (sometimes called “postpositions”) in the form of N¹.gen N².poss.case, for those nmod should be used (following treatment in English of prepositional constructions like “in front of”).
Айгүл бақшаның ішінде ақырын басып жүр . \n Aygül garden-of inside-in quietly walking is .
nsubj(басып-5, Айгүл-1)
nmod:poss(ішінде-3, бақшаның-2)
nmod(басып-5, ішінде-3)
advmod(басып-5, ақырын-4)
aux(басып-5, жүр-6)
punct(басып-5, .-7)
cc
: coordinating conjunction
Coordinating conjunctions are marked as dependents of the last
coordinated element, and the dependency type used is cc
.
Олар Финляндия , Швеция және Эстонияидан кейін Ресейге барды . \n They Finland , Sweden and Estonia-from after Russia-to went .
case(Эстонияидан-6, кейін-7)
conj(Эстонияидан-6, Швеция-4)
cc(Эстонияидан-6, және-5)
conj(Эстонияидан-6, Финляндия-2)
punct(Эстонияидан-6, ,-3)
nmod(барды-9, Эстонияидан-6)
subj(барды-9, Олар-1)
nmod(барды-9, Ресейге-8)
Note: Coordination directionality is under discussion.
ccomp
: clausal complement
Clausal complement is an object like clausal dependent. The governor is most commonly, although not always, the main verb or predicative of the main clause, and the dependent is the main verb or predicative of the dependent clause. The clausal complement can also modify a word other than a verb, most often a noun or pronoun. Most commonly clausal complements are verbal nouns in accusative or dative.
Ол терезеден Азамат пен Айгүлдің ойнағанына қарап тұр . \n She window-through Azamat and Aygül's playing watching stands .
nsubj(қарап-7, Ол-1)
nsubj(ойнағанына-6, Айгүлдің-5)
aux(қарап-7, тұр-8)
punct(қарап-7, .-9)
nmod(қарап-7, терезеден-2)
cc(Айгүлдің-5, пен-4)
conj(Азамат-3, пен-4)
ccomp(қарап-7, ойнағанына-6)
Another example:
Сайттың маңызды екенін түсінбей жатыр . \n Site importance being not-understanding is .
ccomp(түсінбей-4, маңызды-2)
cop(маңызды-2, екенін-3)
nmod:poss(маңызды-2, Сайттың-1)
aux(түсінбей-4, жатыр-5)
punct(түсінбей-4, .-6)
Note that if the complement is completely controlled by the matrix verb, that
is it does not permit another subject or object, then the relation should be
xcomp
.
We also use ccomp
for the complement of reported speech clauses with де- (e.g.
деп, деген, …)
« Төрге шық , тамақ іш » , - демепті . \n « Tör-to go-back , food drink » , - said-not-they .
ccomp(демепті-10, іш-6)
obj(іш-6, тамақ-5)
conj(іш-6, шық-3)
nmod(шық-3, Төрге-2)
punct(демепті-10, .-11)
compound
: compound
compound
is one of the three UD relations UD for compounding,
together with mwe and name.
Nouns should modify appropriate noun in the compound in order to respect the branching structure.
Most uses of attr will be tagged with compound.
Қазақстан көрші елдерге үлгі бола алады . \n Kazakhstan neighbour countries-to example be can .
compound(елдерге-3, көрші-2)
nsubj(бола-5, Қазақстан-1)
aux(бола-5, алады-6)
punct(бола-5, .-7)
Nouns in the izafet construction (e.g. possessive on the final noun) should not get the compound relation,
they should instead be labelled nmod:poss
, e.g.
Дүниежүзілік экономикалық дағдарыс Иран экономикасын тұралатып тастады . \n Global economic depression Iran-of economy declining threw .
nmod:poss(экономикасын-5, Иран-4)
obj(тастады-7, экономикасын-5)
Numerical expressions consisting of multiple tokens are annotated
using the compound
dependency type. The last word of the numerical
expression is the governor, and the number dependencies are chained.
Мұнай экспортынан жыл сайын 16,3 миллиард АҚШ доллары түседі . \n Oil exports-from year per 16,3 milliard USA dollar falls
nmod(доллары-8, АҚШ-7)
nummod(доллары-8, миллиард-6)
nsubj(түседі-9, доллары-8)
compound(миллиард-6, 16,3-5)
punct(түседі-9, .-10)
conj
: coordinated element
A conjunct is the relation between two elements connected by a coordinating
conjunction, such және, мен, немесе, etc. We treat conjunctions
asymmetrically: The head of the relation is the last conjunct and all the
other conjuncts depend on it via the conj
relation.
Олар Финляндия , Швеция және Эстонияидан кейін Ресейге барды . \n They Finland , Sweden and Estonia-from after Russia-to went .
case(Эстонияидан-6, кейін-7)
conj(Эстонияидан-6, Швеция-4)
cc(Эстонияидан-6, және-5)
conj(Эстонияидан-6, Финляндия-2)
punct(Эстонияидан-6, ,-3)
nmod(барды-9, Эстонияидан-6)
subj(барды-9, Олар-1)
nmod(барды-9, Ресейге-8)
Warning: If two sentences are joined with a comma and there is no relation between them, the relation should be parataxis.
Diffs
Note: Coordination directionality is under discussion.
cop
: copula
Warning: DRAFT
A copula verb is a linking verb that joins a subject with a predicate.
Copular clauses receive a special treatment. The
predicative acts as the head word of the clause, and the copular verb
depends on it using a cop
(copula) dependency. The cop
relation
is only applied when the verb is used to link a subject to its predicate
(although the subject may be left out).
The full list of copula verbs is as follows:
- бол- “be, become”
- е- “be” (defective)
If the morphological analyser used outputs an е- copula in aorist third person then this is attached as a leaf node.
Бұл ойын <ø . \n It game is .
cop(ойын-2, <ø-3)
nsubj(ойын-2, Бұл-1)
punct(ойын-2, .-4)
If the copula is not third person it has overt person marking:
Мен студент <пін . \n I student am .
cop(студент-2, <пін-3)
nsubj(студент-2, Мен-1)
punct(студент-2, .-4)
In the past it surfaces as еді-:
Бірақ кеше өте суық еді ! \n But yesterday very cold was-it !
advmod(суық-4, өте-3)
cop(суық-4, еді-5)
Use of “бол” without a predicate:
Эрте заманда Эрназар деген киши болуптур . \n Early time-in Ernazar called person was .
Use of “бол” without subject or predicate:
Шылым шегуге болмайды . \n Cigarette smoking is-allowed-not .
Warning: DRAFT
csubj
: clausal subject
A clausal subject (csubj
) is a clause that acts as the subject of
another clause.
Мүсін жасау , жартастарға рельеф қашау елеулі орын алды . \n Sculpture making , cliffs-to reliefs carving important place took .
csubj(алды-9, қашау-6)
dobj(алды-9, орын-8)
dep
: unspecified dependency
This document is a placeholder for the language-specific documentation
for dep
.
det
: determiner
A determiner is the relation between the head of an NP and its determiner.
Тойға көп кісі жиналыпты . \n Feast-to a_lot people gathered .
subj(жиналыпты-4, кісі-3)
nmod(жиналыпты-4, Тойға-1)
det(кісі-3, көп-2)
punct(жиналыпты-4, .-5)
Қайсы кітапты жақсы_көресің ? \n Which book prefer-you ?
det(кітапты-2, Қайсы-1)
punct(жақсы_көресің-3, ?-4)
obj(жақсы_көресің-3, кітапты-2)
Another example:
Барлық шығыстарды бұл одақ көтереді . \n All expenses this union lifts .
det(шығыстарды-2, Барлық-1)
discourse
: discourse element
Discourse elements are interjections, exclamations or emoticons. They are attached to the main verb or predicative of the
sentence with the discourse
dependency type.
Жоқ , Айгүлдің күшігі жоқ . \n No , Aygül's dog not-existing-is .
discourse(жоқ-5, Жоқ-1)
nsubj(жоқ-5, күшігі-4)
And another example:
Жоқ , хат жазып берейін . \n No , letter writing let-me .
discourse(жазып-4, Жоқ-1)
The discourse
label is also used for modal words, and the question word (ма):
Ал Айгүлдің күшігі бар ма ? \n And Aygül's dog existing-is does-it ?
discourse(бар-4, ма-5)
nsubj(бар-4, күшігі-3)
dislocated
: dislocated elements
This document is a placeholder for the language-specific documentation
for dislocated
.
dobj
: direct object
The dependency type dobj
is used for (nominal) direct objects of the
verb.
Direct objects are typically in “nominative” (unmarked, indefinite accusative) or “accusative” (marked, definite accusative), but may be in other cases (e.g. dative).
Ол барлық жерді қарап жүр . \n She all place-in looking is .
dobj(қарап-4, жерді-3)
aux(қарап-4, жүр-5)
Another example:
Бұл көрсеткіш 75 пайызды құрайды . \n This indicator 75 percent makes-up-for .
dobj(құрайды-5, пайызды-4)
nummod(пайызды-4, 75-4)
A dative direct object:
Айгүл терезеде тұрған анасына қарайды . \n Aygül window-in standing mother-her watches .
dobj(қарайды-5, анасына-4)
acl(анасына-4, тұрған-3)
nmod(тұрған-3, терезеде-2)
nsubj(қарайды-5, Айгүл-1)
punct(қарайды-5, .-6)
expl
: expletive
This document is a placeholder for the language-specific documentation
for expl
.
foreign
: foreign words
This document is a placeholder for the language-specific documentation
for foreign
.
goeswith
: goes with
This document is a placeholder for the language-specific documentation
for goeswith
.
iobj
: indirect object
The indirect object of a verb is any nominal phrase that is a core argument of the verb but is not its subject or direct object. The prototypical example is the recipient of ditransitive verbs of exchange:
19 ғасырдың 2 жартысынан Иран музыкасына еуропалық музыка мәдениеті әсер етті . \n 19 century's 2nd part-from Iran music-to European music culture impression made .
iobj(етті-11, музыкасына-6)
obj(етті-11, әсер-10)
Note: Discussion of what constitutes a core argument in Kazakh is ongoing. The iobj
relation will
probably be discarded in favour of subcategorising nmod
.
list
: list
This document is a placeholder for the language-specific documentation
for list
.
mark
: marker
A marker (mark
) is the subordinating conjunction in a
non-complement subordinate clause.
The subordinating conjunctions in Kazakh are:
- егер “if”
- өйткені “because”
Егер көп жаңбыр жауса , түсімнің бәрі өледі . \n If much rain rain-CND , crop's all die-will
mark(жауса-4, Егер-1)
det(жаңбыр-3, көп-2)
subj(жауса-4, жаңбыр-3)
The dependency relation mark
is also used to identify the
complementiser appearing in most clausal complements, where the head
of the dependency is the main verb of the subordinate clause.
mwe
: multi-word expression
This document is a placeholder for the language-specific documentation
for mwe
.
name
: name
Proper names constituted of more than one word are annotated using the dependency type name
. The last (rightmost) word is the head, and the other words are direct dependents of the head. The last word is chosen to be the head because in Kazakh the last word carries the inflectional information of the whole structure.
The name
dependency relation is used in cases where the multi-word name does not have an obvious internal syntactic structure, as is the case with for instance names of people (**).
If a name has an obvious internal structure, as is often the case in names of books and movies for instance, this structure is marked instead.
Соғыстан кейін Федор Федорович Қарағандыда тұрды . \n War-from after Fëdor Fëdorovič Qarağandı-in stood .
name(Федорович-4, Федор-3)
nsubj(тұрды-6, Федорович-4)
Diffs
Contrary to the general UD definition of u-dep/name, in UD Kazakh the last word of the multi-word name expression is considered the head.
neg
: negation modifier
The neg
relation is not used in Kazakh.
nmod
: nominal modifier
Nominal modifiers are inflected nominals which modify most commonly a verb or a noun phrase. They can occur alone or together with an postposition in an postpositional phrase. Both cases are analysed similarly, as semantically nominal modifiers and postpositional phrases are similar.
nmod
is by far the most used relation in UD Kazakh.
Postpositional phrases:
Азамат ағашқа қарай жүгірді . \n Azamat tree-to towards ran .
nsubj(жүгірді-4, Азамат-1)
case(ағашқа-2, қарай-3)
punct(жүгірді-4, .-5)
nmod(жүгірді-4, ағашқа-2)
Izafet construction:
Genitive nouns in the izafet construction get a special relation, nmod:poss
, be they indefinite:
Дүниежүзілік экономикалық дағдарыс Иран экономикасын тұралатып тастады . \n Global economic depression Iran-of economy declining threw .
nmod:poss(экономикасын-5, Иран-4)
obj(тастады-7, экономикасын-5)
or definite:
Object of comparison:
Nouns in ablative case which depend on adjectives as object of comparison get nmod:comp
:
nmod:comp
: object of comparison
We use the nmod:comp
relation for the comparator in comparison constructions like “X is bigger
than Y”, in Turkic, the “than Y” is in the ablative case and this depends on the adjective X.
nmod:poss
: nmod:poss
This document is a placeholder for the language-specific documentation
for nmod:poss
.
nsubj
: nominal subject
The dependency type nsubj
marks nominal subjects a clause.
Азамат ағашқа қарай жүгірді . \n Azamat tree-to towards ran .
nsubj(жүгірді-4, Азамат-1)
case(ағашқа-2, қарай-3)
punct(жүгірді-4, .-5)
nmod(жүгірді-4, ағашқа-2)
Another example:
Мұнда үлкен шара өтеді . \n Here big event takes-place .
nsubj(өтеді-4, шара-3)
amod(шара-3, үлкен-2)
advmod(өтеді-4, Мұнда-1)
punct(өтеді-4, .-5)
Note that the governor may not always be a verb, in copula predication, non-verbal predicates are possible:
Сары жылы түс . \n Yellow warm colour .
nsubj(түс-3, Сары-1)
amod(түс-3, жылы-2)
punct(түс-3, .-4)
nummod
: numeric modifier
Numeric modifiers of a noun or NP, when they function as quantifiers
are marked with the nummod
dependency type.
Ordinal numbers should get the amod
type.
Бразилия өз жерінде чемпионатты екі рет өткізген бесінші ел болды . \n Brazil own land-in championship two times XX fifth country was .
nummod(рет-6, екі-5)
nmod(ел-9, рет-6)
amod(ел-9, бесінші-8)
Another example:
Баба-атаға да үш күн түнеді . \n Baba-ata-to also three days slept-over .
nummod(күн-4, үш-3)
parataxis
: parataxis
When two sentences share no relation but are written together in a single
sentence (delimited by comma, dash, semicolon or other punctuation) then we use the relation parataxis.
As with explicit coordinations, the last element is the governor.
Also parenthetical and interjected clauses can receive the parataxis
dependency.
Азамат пен Айгүл ойнағанды жақсы көреді , олар әрдайым бақшада бірге ойнайды . \n Azamat and Aygül playing XX see , they always garden-in together play .
parataxis(ойнайды-12, көреді-6)
punct
: punctuation
The dependency type punct
is used to mark punctuation. The dependent is the punctuation
symbol, and the governor is the element which the punctuation symbol delimits. For instance,
with coordination, the last coordinated element is the head of all punct
dependencies
in the coordination.
Coordination and parataxis:
The punctuation attaches to the governor.
Олар Финляндия , Швеция және Эстонияидан кейін Ресейге барды . \n They Finland , Sweden and Estonia-from after Russia-to went .
case(Эстонияидан-6, кейін-7)
conj(Эстонияидан-6, Швеция-4)
cc(Эстонияидан-6, және-5)
conj(Эстонияидан-6, Финляндия-2)
punct(Эстонияидан-6, ,-3)
nmod(барды-9, Эстонияидан-6)
subj(барды-9, Олар-1)
nmod(барды-9, Ресейге-8)
Adverbial clauses:
The punctuation attaches to the head of the adverbial phrase.
remnant
: remnant in ellipsis
The remnant
relation is used to provide a treatment of ellipsis (in
the case of gapping and stripping, where a predicational or verbal
head gets elided). In particular, remnant
aims to provide analyses
that do not postulate empty nodes.
Ашылу матчы Сан-Паулуда , ал финалы Рио-де-Жанейрода орын алды . \n Opening match São-Paulo-in , and final Rio-de-Janeiro-in place took .
remnant(Рио-де-Жанейрода-7, Сан-Паулуда-3)
remnant(финалы-6 ,матчы-2)
nmod(алды-9, Рио-де-Жанейрода-7)
nsubj(алды-9, финалы-6)
nmod(матчы-2, Ашылу-1)
reparandum
: overridden disfluency
This document is a placeholder for the language-specific documentation
for reparandum
.
root
: root
The root
grammatical relation points to the root of the sentence. A
fake node ROOT
is used as the governor.
ROOT Азамат ағашқа қарай жүгірді . \n ROOT Azamat tree-to towards ran .
root(ROOT-1, жүгірді-5)
nsubj(жүгірді-5, Азамат-2)
case(ағашқа-3, қарай-4)
punct(жүгірді-5, .-6)
nmod(жүгірді-5, ағашқа-3)
vocative
: vocative modifier
The dependency type vocative
is used for vocatives, that is,
expressions where someone is being addressed. The governor of the
dependency is the main predicate of the clause where the addressing
occurs.
Жоқ , Айгүл , айта алмаймын . \n No , Aygül , say can-not-I .
vocative(айта-5, Айгүл-3)
discourse(айта-5, Жоқ-1)
aux(айта-5, алмаймын-6)
xcomp
: open clausal complement
The dependency type xcomp
is reserved for clausal complements which have an external
subject, that is, whose subject is shared with the complemented verb (a phenomenon also
known as subject control). Note that the subject of the complementing clause must be
the subject of the complemented verb, not any other sentence element.
Ол Азаматты табуға әрекет етіп жүр . \n She Azamat finding activity doing is .
xcomp(етіп-5, табуға-3)
nsubj(етіп-5, Ол-1)
dobj(табуға-3, Азаматты-2)
Note: The existence of xcomp
is uncertain in Kazakh, the vast majority of clausal
complements are not controlled and should get ccomp
instead.