Docs on SUD Guidelines

-t-il

Mon, 01 Jan 0001 00:00:00 +0000

How to annotate a-t-il ? #

We choose this annotation for -t-il :

Thus, we consider -t-il as a PRON with the lemma lui. We consider -t-il as an allomorph of il. -t is not a syntactic element, it is a morphophonological rule that introduces the sound /t/.

ADJ

Mon, 01 Jan 0001 00:00:00 +0000

Adjective #

Universal #

Definition from de UD website

Adjectives are words that typically modify nouns and specify their properties or attributes:

The oldest French bridge

They may also function as predicates, as in:

The car is green.

Some words that could be seen as adjectives (and are tagged as such in other annotation schemes) have a different tag in UD: See DET for determiners and NUM for (cardinal) numbers.

ADJ is also used for “proper adjectives” such as European (“proper” as in proper nouns, i.e., words that are derived from names but are adjectives rather than nouns).

aff

Mon, 01 Jan 0001 00:00:00 +0000

aff #

Universal #

French #

TODO

Overview #

Specific Pattern #

Aspect

Mon, 01 Jan 0001 00:00:00 +0000

`Aspect` #

Aspect page on UD website
Tables built on UD treebanks:
- Treebanks using Aspect
- Values used with the Aspect feature

Beja #

In Beja, the three values Perf, Imp and aor are used for Aspect.

pattern { X [Aspect] }
  X.Aspect 
  X.upos

Case

Mon, 01 Jan 0001 00:00:00 +0000

`Case` #

Case page on UD website
Tables built on UD treebanks:
- Treebanks using Case
- Values used with the Case feature

CCONJ

Mon, 01 Jan 0001 00:00:00 +0000

Coordination conjunction #

Universal #

Definition from de UD website A coordinating conjunction is a word that links words or larger constituents without syntactically subordinating one to the other and expresses a semantic relationship between them.

For subordinating conjunctions, see SCONJ.

Examples

comp

Mon, 01 Jan 0001 00:00:00 +0000

`comp` #

Universal #

The comp relation (without extension) is used when one has difficulty deciding between comp:obj and comp:obl.

French #

pattern { e: X -[1=comp, !2]-> Y }
  e.label

Overview #

In French, the comp label is frequently used to annotate reflexive pronouns and other pronominal clitics which contribute to the formation of pronominal verbs when it is difficult to determine the role of the pronoun. In constructions such as Il s’en sort (en: He’s doing well) the pronoun se no longer provides the semantic value of an argument of the verb. However, it fits so well into the typical argument structure that it is hard to recognize that it cannot be de-pronominalized. For this reason, it is annotated with the comp relation.

comp:aux

Mon, 01 Jan 0001 00:00:00 +0000

`comp:aux` #

Universal #

The relation comp:aux is used for the argument of auxiliaries. It can be used with the following deep features:

pattern { GOV-[comp:aux]->DEP }

English

comp:cleft

Mon, 01 Jan 0001 00:00:00 +0000

`comp:cleft` #

Universal #

The comp:cleft relation is used in cleft sentences for the dependency from the head of the sentence to the head of the complement clause.

English

comp:obj

Mon, 01 Jan 0001 00:00:00 +0000

`comp:obj` #

Universal #

The relation comp:obj is used for direct object complements, including direct complements of an adposition or a subordinating conjunction.

comp:obj can have the following deep features:

pattern { GOV-[comp:obj]->DEP }

Direct object complements #

English

comp:obl

Mon, 01 Jan 0001 00:00:00 +0000

comp:obl #

Universal #

The comp:obl relation is used for oblique arguments of verbs, adjectives, adverbs, nouns or pronouns, regardless of their form. Unlike UD, SUD does not make a distinction between indirect objects and oblique arguments. SUD instead makes the distinction between arguments and modifiers.

The relation comp:obl can have these deep features:

pattern { GOV-[comp:obl]->DEP }

English

comp:pred

Mon, 01 Jan 0001 00:00:00 +0000

comp:pred #

Universal #

The comp:pred relation is used for predicative arguments introduced by a copula or a verb.

pattern { X -[comp:pred]-> Y }

predicative arguments of copula #

pattern { 
	X -[comp:pred]-> Y; 
	X [upos=AUX] 
}

English: comp:pred introduced by the copula to be

Comparative

Mon, 01 Jan 0001 00:00:00 +0000

Comparative constructions #

In comparative constructions where each term is introduced with a grammatical pair like more… than, the relation between the two grammatical words is labeled comp:obj, with the first word being the head of the second. Consider the example:

French

conj:appos

Mon, 01 Jan 0001 00:00:00 +0000

conj:appos #

Universal #

This relation is used for appositions. In an apposition, conjuncts have the same referent, but two clearly different denotations. The conjunct in apposition forms a predication in the background (in Mary, my best friend, the second conjunct my best friend is a predication on the first conjunct Mary, it is equivalent to Mary is my best friend).

pattern { X -[conj:appos]-> Y }

Be careful: conj:appos is only used when the two conjuncts form two separate phrases. When the apposition is a modification, the relation mod@appos must be used, like my friend Mary or le boulevard Saint-Michel.

conj:coord

Mon, 01 Jan 0001 00:00:00 +0000

`conj:coord` #

For more general information about coordination, see the coordination construction page.

The conj:coord relation is used to link elements connected with a coordinating conjunction such as and and or. The conjunction itself is linked to the second element with a cc relation. It is analogous to conj relation in UD.

pattern { X -[conj:coord]-> Y }

French

conj:dicto

Mon, 01 Jan 0001 00:00:00 +0000

conj:dicto #

Universal #

This relation is used to indicate difluencies when a speaker corrects his speech and to analyse reformulations. The relation conj:coord is used to link two different objects, two referants (ex: Mary and John are two different referents). The relation conj:dicto on the other hand is used to link two denotations of the same referent (ex: the desert in Kenya, the Kenya desert is denoting the same referant).

Coordination

Mon, 01 Jan 0001 00:00:00 +0000

Coordination #

In SUD, the dependency relation between conjucts or a coordination.

pattern { X -[conj:coord]-> Y }

Chained conjuncts #

In SUD, when a coordination concerns more than two conjuncts, each conjunct is attached to the head of the previous one in a chain. This helps to reduce cumbersome long-distance relations. We also believe it to be a more accurate representation of the cognitive process of coordination.

CorrectForm

Mon, 01 Jan 0001 00:00:00 +0000

`CorrectForm` #

Universal #

The CorrectForm feature is used in the annotation of typographical errors; it gives the expected corrected form. It is always used with the features Typo=Yes

Definite

Mon, 01 Jan 0001 00:00:00 +0000

Definite #

Universal #

Definition from de UD website

Definiteness is typically a feature of nouns, adjectives and articles. Its value distinguishes whether we are talking about something known and concrete, or something general or unknown. It can be marked on definite and indefinite articles, or directly on nouns, adjectives etc. In Arabic, definiteness is also called the “state”.

Values #

Indefinite, Specific indefinite, Definite, Construct state, Complex.

Degree

Mon, 01 Jan 0001 00:00:00 +0000

Degree #

Universal #

Definition from de UD website Degree of comparison is typically an inflectional feature of some adjectives and adverbs. A different flavor of degree is diminutives and augmentatives, which often apply to nouns but are not restricted to them.

Values #

Positiv, Equativ, Comparative, Superlative, Abolute, Diminutive, Augmentative.

DET

Mon, 01 Jan 0001 00:00:00 +0000

Determinant #

Universal #

See UD DET page for a definition.

Haitian Creole #

Overview #

Determiners are words that modify nouns or noun phrases and express the reference of the noun phrase in context. In Haitian Creole the determiners are placed always after a noun. However, the “yon” determiner has an exceptional behaviour and is placed before a noun.

Specific Pattern #

Yon determiner #

The yon determiner is an indefinite singular article. It has an exceptional behaviour compared to other determiners in Haitian Creole and it is always placed before a noun.

Disfluency

Mon, 01 Jan 0001 00:00:00 +0000

Disfluency #

In oral language, speakers form sentences while speaking, which means that sometimes they might search for words while speaking.

They may repeat a word until they find the right one, or they may try with one word and then decide to change it.

Repetitions and reformulations #

To analyse repetition or reformulation, we use the relation conj:dicto.

French

dislocated

Mon, 01 Jan 0001 00:00:00 +0000

`dislocated` #

Universal #

The dislocated relation is used to analyze detached elements that do not enter in the core grammatical relations of a sentence.

pattern { X -[dislocated]-> Y }

French

Foreign

Mon, 01 Jan 0001 00:00:00 +0000

Foreign #

Universal #

Definition from de UD website Boolean feature. Is this a foreign word? Not a loan word and not a foreign name but a genuinely foreign word appearing inside native text, e.g. inside direct speech, titles of books etc. This feature would apply either to the X part of speech (unanalyzable token), or to other parts of speech if we know and are willing to annotate the class to which the word belongs in its original language.

Gender

Mon, 01 Jan 0001 00:00:00 +0000

`Gender` #

Gender page on UD website
Tables built on UD treebanks:

French #

In French, the two values Masc and Fem are used for Gender.

pattern { X [Gender] }
  X.Gender 
  X.upos

Idioms and Titles

Mon, 01 Jan 0001 00:00:00 +0000

Idioms and Titles #

Universal #

SUD offers several features which allow annotators to mark idiomatic expressions and titles while still preserving the internal syntactic structure between their various components. We have decided to distinguish these two categories from Multi-Word Expressions (MWEs), which represent a broader category which also includes named entities.

For our purposes, “title” refer to any title given to a film, book, painting, or other work of art, such as Planet of the Apes, Dark Side of the Moon, American Gothic, or Super Mario Bros. However, this excludes other named entities like events, holidays or locations, such as The Gulf War, Good Friday, or The Eiffel Tower.

INTJ

Mon, 01 Jan 0001 00:00:00 +0000

Interjection #

Universal #

See UD INTJ page for a definition.

French

Mood

Mon, 01 Jan 0001 00:00:00 +0000

Mood #

Universal #

Definition from de UD website

Mood is a feature that expresses modality and subclassifies finite verb forms.

Value #

Ind, Imp, Cdn, Pot, Jus, Sub, Prp, Opt, Des, Nec, Qot, Int, Irr, Adm.

NOUN

Mon, 01 Jan 0001 00:00:00 +0000

Noun #

Universal #

Definition from de UD website

Nouns are a part of speech typically denoting a person, place, thing, animal or idea.

The NOUN tag is intended for common nouns only. See PROPN for proper nouns and PRON for pronouns.

Note that some verb forms such as gerunds and infinitives may share properties and usage of nouns and verbs. Depending on language and context, they may be classified as either VERB or NOUN.

NUM

Mon, 01 Jan 0001 00:00:00 +0000

Numeral #

Universal #

Definition from de UD website

A numeral is a word, functioning most typically as a determiner, adjective or pronoun, that expresses a number and a relation to the number, such as quantity, sequence, frequency or fraction.

Note that cardinal numerals are covered by NUM whether they are used as determiners or not (as in Windows Seven) and whether they are expressed as words (four), digits (4) or Roman numerals (IV). Other words functioning as determiners (including quantifiers such as many and few) are tagged DET.

Number

Mon, 01 Jan 0001 00:00:00 +0000

`Number` #

Number page on UD website
Tables built on UD treebanks:

French #

In French, the two values Sing and Plur are used for Number.

pattern { X [Number] }
  X.Number 
  X.upos

Numeral

Mon, 01 Jan 0001 00:00:00 +0000

Numeral #

Written in Letter #

Numbers composed of more than one word, such as five hundred or six thousand are primarily chained together with the flat relation. If the number contains the coordinating conjunction and, such as in one hundred and one, the integer directly preceding the coordinating conjunction is connected to one directly following it with a conj:coord relation.

pattern { X1 [upos=NUM]; X2 [upos=NUM]; X1 -[flat]-> X2 }

Naija

NumType

Mon, 01 Jan 0001 00:00:00 +0000

NumType #

Universal #

Definition from de UD website

Some languages (especially Slavic) have a complex system of numerals. For example, in the school grammar of Czech, the main part of speech is “numeral”, it includes almost everything where counting is involved and there are various subtypes. It also includes interrogative, relative, indefinite and demonstrative words referring to numbers (words like kolik / how many, tolik / so many, několik / some, a few), so at the same time we may have a non-empty value of PronType. (In English, these words are called quantifiers and they are considered a subgroup of determiners.)

parataxis

Mon, 01 Jan 0001 00:00:00 +0000

`parataxis` #

Universal #

The parataxis relation is used to analyse two elements that are placed side by side with no explicit marker of coordination, subordination, or argument relation with the head word.

pattern { X -[parataxis]-> Y }

Incisive proposal #

French

parataxis:insert

Mon, 01 Jan 0001 00:00:00 +0000

`parataxis:insert` #

Universal #

The relation parataxis:insert is used for inserted clause. Contrary to a parenthetical clause, an inserted clause could not form an independent sentence

pattern { X -[parataxis:insert]-> Y }

French

parataxis:parenth

Mon, 01 Jan 0001 00:00:00 +0000

`parataxis:parenth` #

Universal #

The parataxis:parenth relation is used for parenthetical clauses, in the sense that the clause is backgrounded (as if it were inside parentheses). A parenthetical clause could form an independent sentence.

pattern { X -[parataxis:parenth]-> Y }

French

PART

Mon, 01 Jan 0001 00:00:00 +0000

Paticle #

Universal #

Definition from de UD website

Particles are function words that must be associated with another word or phrase to impart meaning and that do not satisfy definitions of other universal parts of speech (e.g. adpositions, coordinating conjunctions, subordinating conjunctions or auxiliary verbs). Particles may encode grammatical categories such as negation, mood, tense etc. Particles are normally not inflected, although exceptions may occur.

Note that the PART tag does not cover so-called verbal particles in Germanic languages, as in give in or end up. These are adpositions or adverbs by origin and are tagged accordingly ADP or ADV. Separable verb prefixes in German are treated analogically.

Person

Mon, 01 Jan 0001 00:00:00 +0000

Person #

Universal #

Definition from de UD website Person is typically feature of personal and possessive pronouns / determiners, and of verbs. On verbs it is in fact an agreement feature that marks the person of the verb’s subject (some languages, e.g. Basque, can also mark person of objects). Person marked on verbs makes it unnecessary to always add a personal pronoun as subject and thus subjects are sometimes dropped (pro-drop languages).

Polarity

Mon, 01 Jan 0001 00:00:00 +0000

Polarity #

Definition from de UD website Polarity is typically a feature of verbs, adjectives, sometimes also adverbs and nouns in languages that negate using bound morphemes. In languages that negate using a function word, Polarity is used to mark that function word, unless it is a pro-form already marked with PronType=Neg (see below).

Positive polarity (affirmativeness) is rarely, if at all, encoded using overt morphology. The feature value Polarity=Pos is usually used to signal that a lemma has negative forms but this particular form is not negative. Using the feature in such cases is somewhat optional for words that can be negated but rarely are.

Poss

Mon, 01 Jan 0001 00:00:00 +0000

Poss #

Universal #

Definition from de UD website Boolean feature of pronouns, determiners or adjectives. It tells whether the word is possessive.

While many tagsets would have “possessive” as one of the various pronoun types, this feature is intentionally separate from PronType, as it is orthogonal to pronominal types. Several of the pronominal types can be optionally possessive, and adjectives can too.

Values #

Yes.

Predicate nominal construction

Mon, 01 Jan 0001 00:00:00 +0000

Predicate nominal construction #

Zero copula strategy #

French Predicate Nominal Construction #

Overview :

Pattern :

Example :

English Predicate Nominal Construction #

Overview :

Pattern :

Example :

Verbal copula strategy #

Verbal strategy #

PRON

Mon, 01 Jan 0001 00:00:00 +0000

Pronoun #

Universal #

Definition from de UD website Pronouns are words that substitute for nouns or noun phrases, whose meaning is recoverable from the linguistic or extralinguistic context.

Pronouns under this definition function like nouns. Note that some languages traditionally extend the term pronoun to words that substitute for adjectives. Such words are not tagged PRON under our universal scheme. They are tagged as determiners in order to annotate the same thing the same way across languages.

PROPN

Mon, 01 Jan 0001 00:00:00 +0000

Proper Noun #

Universal #

Definition from de UD website

A proper noun is a noun (or nominal content word) that is the name (or part of the name) of a specific individual, place, or object.

Note that PROPN is only used for the subclass of nouns that are used as names and that often exhibit special syntactic properties (such as occurring without an article in the singular in English). When other phrases or sentences are used as names, the component words retain their original tags. For example, in Cat on a Hot Tin Roof, Cat is NOUN, on is ADP, a is DET, etc.

PUNCT

Mon, 01 Jan 0001 00:00:00 +0000

Punctuation #

Universal #

Definition from de UD website

Punctuation marks are non-alphabetical characters and character groups used in many languages to delimit linguistic units in printed text.

Punctuation is not taken to include logograms such as $, %, and §, which are instead tagged as SYM. (Hint: if it corresponds to a word that you pronounce, such as dollar or percent, it is SYM and not PUNCT.)

Spoken corpora contain symbols representing pauses, laughter and other sounds; we treat them as punctuation, too. In these cases it is even not required that all characters of the token are non-alphabetical. One can represent a pause using a special character such as #, or using some more descriptive coding such as [:pause].

Reflex

Mon, 01 Jan 0001 00:00:00 +0000

Reflex #

Universal #

Definition from de UD website Boolean feature, typically of pronouns or determiners. It tells whether the word is reflexive, i.e. refers to the subject of its clause.

While many tagsets would have “reflexive” as one of the various pronoun types, this feature is intentionally separate from PronType. When used with pronouns and determiners, it should be combined with PronType=Prs, regardless whether they really distinguish the Person feature (in some languages they do, in others they do not).

Reported

Mon, 01 Jan 0001 00:00:00 +0000

Reported #

Universal #

The feature Reported=Yes is used to annotate reported speech in the spoken corpora. The features is on the head of the reported speech.

See also the page about reported speech.

Reported speech

Mon, 01 Jan 0001 00:00:00 +0000

Reported Speech in Spoken Corpora #

Universal #

Reported speech has a feature Reported=Yes on its head. It is generally the comp:obj of a speech verb, such as dire ’to say'.

pattern { X -[comp:obj]-> Y; Y[Reported=Yes] }

French

SCONJ

Mon, 01 Jan 0001 00:00:00 +0000

Subordinating conjunction #

Universal #

Definition from de UD website

A subordinating conjunction is a conjunction that links constructions by making one of them a constituent of the other. The subordinating conjunction typically marks the incorporated constituent which has the status of a (subordinate) clause.

We follow Loos et al. 2003 in recognizing these three subclasses as subordinating conjunctions:

Complementizers, like [en] that or whether Non-ADV markers that introduce an adverbial clause, like [en] because, since, before, or once (when introducing a clause, not a nominal) Non-pronominal relativizers, like [he] še. Words in this category simply introduce a relative clause (and normally don’t inflect). This excludes words that have a nominal function within the relative clause; relative and resumptive pronouns (e.g., English relative that and which) are analyzed as PRON. For coordinating conjunctions, see CCONJ.

subj

Mon, 01 Jan 0001 00:00:00 +0000

The `subj` relation #

Universal #

The subj relation is used for all subjects, regardless of their form, nominal or clausal.

English

English

suffixes -ci and -là

Mon, 01 Jan 0001 00:00:00 +0000

Annotation of -ci and -là #

In French, we consider -ci and -là as ADV.

pattern { N [form="-ci"|"-là"] }

See #12

SYM

Mon, 01 Jan 0001 00:00:00 +0000

Symbol #

Universal #

Definition from de UD website

A symbol is a word-like entity that differs from ordinary words by form, function, or both.

Many symbols are or contain special non-alphanumeric characters, similarly to punctuation. What makes them different from punctuation is that they can be substituted by normal words. This involves all currency symbols, e.g. $ 75 is identical to seventy-five dollars.

Mathematical operators form another group of symbols.

Tense

Mon, 01 Jan 0001 00:00:00 +0000

Tense #

Universal #

See UD Tense page for a description.

French #

Haitian Creole #

Usage of Tense in Haitian Creole UD treebank 2.15

Typo

Mon, 01 Jan 0001 00:00:00 +0000

`Typo` #

Universal #

The feature Typo=Yes is used to indicate that a token has a typographic error. The corresponding tokern should have the feature CorrectForm with the expected form of the token.

French

VERB

Mon, 01 Jan 0001 00:00:00 +0000

Verb #

Universal #

Definition from de UD website

A verb is a member of the syntactic class of words that typically signal events and actions, can constitute a minimal predicate in a clause, and govern the number and types of other constituents which may occur in the clause. Verbs are often associated with grammatical categories like tense, mood, aspect and voice, which can either be expressed inflectionally or using auxilliary verbs or particles.

VerbForm

Mon, 01 Jan 0001 00:00:00 +0000

VerbForm #

Universal #

Definition from de UD website

Even though the name of the feature seems to suggest that it is used exclusively with verbs, it is not the case. Some verb forms in some languages actually form a gray zone between verbs and other parts of speech (nouns, adjectives and adverbs). For instance, participles may be either classified as verbs or as adjectives, depending on language and context. In both cases VerbForm=Part may be used to separate them from other verb forms or other types of adjectives.

Voice

Mon, 01 Jan 0001 00:00:00 +0000

Voice #

Universal #

Definition from de UD website

Voice is typically a feature of verbs. It may also occur with other parts of speech (nouns, adjectives, adverbs), depending on whether borderline word forms such as gerunds and participles are classified as verbs or as the other category.

For Indo-European speakers, voice means mainly the active-passive distinction. In other languages, other shades of verb meaning are categorized as voice.

X

Mon, 01 Jan 0001 00:00:00 +0000

Other #

Universal #

Definition from de UD website

The tag X is used for words that for some reason cannot be assigned a real part-of-speech category. It should be used very restrictively.

A special usage of X is for cases of code-switching where it is not possible (or meaningful) to analyze the intervening language grammatically (and where the dependency relation flat:foreign is typically used in the syntactic analysis). This usage does not extend to ordinary loan words which should be assigned a normal part-of-speech. For example, in he put on a large sombrero, sombrero is an ordinary NOUN.

mod

Mon, 01 Jan 0001 00:00:00 +0000

`mod` #

Universal #

The mod relation is used for modifiers of verbs, nouns, adjectives, adverbs, auxiliaries, adpositions and conjunctions.

English

Subject

Mon, 01 Jan 0001 00:00:00 +0000

Subject #

Universal #

The Subject feature is specific to SUD. It is used to annotate the control of infinitive subject. In version 2.15, it is annotated in the Naija corpus and in the four French corpora.

See discussion #29 on GitHub.

⚠️ In previous versions of SUD, control were annotated with the deep edge feature @x.

The @x feature is now deprecated and must not be used in new annotations.

@agent

Mon, 01 Jan 0001 00:00:00 +0000

`@agent` #

Universal #

The @agent feature is used for arguments that are deep subjects of their governor. A deep subject is an argument that is subject in the base construction but has been demoted in redistribution of argument position. Typical cases are passive, causative, and impersonal constructions.

passive (English)

@caus

Mon, 01 Jan 0001 00:00:00 +0000

`@caus`: Causative #

Universal #

The @caus feature is used for causative constructions:

the promoted subject is analyzed subj@caus(while the demoted subject, if present, should have a @agent deep subrelation;
the relation between the causative auxiliary (when the causative construction uses one) and the verb is comp:aux@caus.

Causative construction (French)

@emb

Mon, 01 Jan 0001 00:00:00 +0000

`@emb`: Embedded Coordination #

The deep feature @emb is used only with the relation conj:coord to mark embedded coordination.

See here for more details.

@expl

Mon, 01 Jan 0001 00:00:00 +0000

`@expl`: expletive #

Universal #

The @expl feature is used for arguments of predicates without semantic content (impersonal subjects for instance). In UD, such elements ara analyzed with expl, without indicating their surface syntactic function. The feature @explallows us to compibe the surface syntactic relation with these more semantic information.

Impersonal subject (English)

@foreign

Mon, 01 Jan 0001 00:00:00 +0000

`@foreign` #

Universal #

The @foreign feature is used with the flat relation for annotation of foreign language phrases. Note that foreign tokens are tagged upos=X and Foreign=Yes. An ExtPos feature can indicate the role of the ofreign phrase.

English

@lvc

Mon, 01 Jan 0001 00:00:00 +0000

`@lvc`: Light Verb Construction #

Universal #

A light verb construction (LVC) is a type of verbal structure in which a verb is coupled with another element – typically a noun phrase – which provides the primary semantic value. Common examples of LVCs in English include take a walk, give a kiss, or have a drink. In SUD, LVCs are marked with the deep feature @lvc.

Note that in case of a LVC, the arguments of the LVC attached to the predicative noun are compand not udep.

@name

Mon, 01 Jan 0001 00:00:00 +0000

`@name` #

Universal #

The deep feature @name is used in combination with the flat relation, moslty in proper nouns. See table for the use of flat@name in SUD 2.15.

Finnish

@pass

Mon, 01 Jan 0001 00:00:00 +0000

@pass #

Universal #

The @pass deep feature is used for the arguments of passive auxiliaries.

The relation between the auxiliary and the verb is marked comp:aux@pass and the subject of the auxiliary bears the subj@pass relation.

English

@relcl

Mon, 01 Jan 0001 00:00:00 +0000

`@relcl`: Relative clause #

Universal #

The @relcl feature is used for the relation between the head of a relative clause and its antecedent. See table for the use of @relcl in SUD 2.16.

English

@tense

Mon, 01 Jan 0001 00:00:00 +0000

@tense #

Universal #

The @tense feature is used for the argument of tense auxiliaries.

English

à part

Mon, 01 Jan 0001 00:00:00 +0000

The bigram of lemmas à[ADP] + part[NOUN] can be used in several idiomatic expressions:

as an adverbal idiom (cf. Wiktionary)
as an adjectival idiom (cf. Wiktionary)
as a prepositional idiom (cf. Wiktionary)
in the larger preposional idiom mis à part (cf. Wiktionary)
in the larger adjectival idiom à part entière (cf. Wiktionary)
in the larger adverbial idiom à parts égales (cf. Wiktionary)

Adverbal idiom #

à part can be an adverbal idiom, meaning separately

ADP

Mon, 01 Jan 0001 00:00:00 +0000

Adposition #

Universal #

Definition from de UD website

Adposition is a cover term for prepositions and postpositions. Adpositions belong to a closed set of items that occur before (preposition) or after (postposition) a complement composed of a noun phrase, noun, pronoun, or clause that functions as a noun phrase, and that form a single structure with the complement to express its grammatical and semantic relation to another unit within a clause.

ADV

Mon, 01 Jan 0001 00:00:00 +0000

Adverb #

Universal #

Definition from UD website

Adverbs are words that typically modify verbs for such categories as time, place, direction or manner. They may also modify adjectives and other adverbs, as in very briefly or arguably wrong.

There is a closed subclass of pronominal adverbs that refer to circumstances in context, rather than naming them directly; similarly to pronouns, these can be categorized as interrogative, relative, demonstrative etc. Pronominal adverbs also get the ADV part-of-speech tag but they are differentiated by additional features.

AUX

Mon, 01 Jan 0001 00:00:00 +0000

Auxiliaire #

Universal #

Definition from UD website

An aux (auxiliary) of a clause is a function word associated with a verbal predicate that expresses categories such as tense, mood, aspect, voice or evidentiality. It is often a verb (which may have non-auxiliary uses as well) but many languages have nonverbal TAME markers and these are also treated as instances of aux. Auxiliares used to construct the passive voice are now also labeled aux.

Cleft constructions

Mon, 01 Jan 0001 00:00:00 +0000

Cleft sentences and questions #

Cleft sentences are an extremely common construction in Naija, making the comp:cleft relation a particularly important for the annotation of this language. The basic cleft construction in Naija includes the phrase na im (it’s him), followed by a verb phrase, though a number of variants exist. The following provides several such examples.

The comp:cleft relation is also used in questions containing interrogative words such as who or where. In such cases, the wh-word is annotated as the root, and is connected to the verb via a comp:cleft relation.

Compounds and phrasal verbs

Mon, 01 Jan 0001 00:00:00 +0000

Compounds and phrasal verbs #

Our annotation of Naija makes frequent use of the compound relation. In our annotation system, this relation is systematically applied to relationships between two nouns in which one of them acts as a form of modifier. In this sense, compound functions much like the mod relation, except that it links two nouns together rather than a noun to an adjective.

The compound relation is also used in some relations between nouns and adjectives, such as dry cleaner, which are considered fixed expressions whose meaning cannot be directly understood from its constituent parts

discourse

Mon, 01 Jan 0001 00:00:00 +0000

`discourse` #

Universal #

The discourse relation is used to link discourse markers. These markers are not clearly linked to the structure of the sentence, except in an expressive way.

In some cases the discourse marker can be another proposition.

Ex: He comes on wednesday, I think. In this sentence I think is a discourse marker because it lacks an object (the object of the verb think is the rest of the sentence) and because they are fixed (I think doesn’t commute with he thinks).

flat

Mon, 01 Jan 0001 00:00:00 +0000

`flat` #

Universal #

The flat relation plays a similar role to the compound one, and its exact usage also varies on a language-by-language basis. However, it is most frequently used to connect the various elements of proper names to one another, including titles and honorifics.

Some examples in SUD_English-EWT:

pattern { X -[flat]-> Y }

English

French auxiliaries

Mon, 01 Jan 0001 00:00:00 +0000

comp:aux #

In French, only three verbs are considered auxiliaries: être, avoir and faire.

The comp:aux relation can be used with one of the three sub-relations @tense, @pass or @caus, depending on whether the auxiliary is expressing a tense, a passive or a causative construction. In French, the relation comp:aux@tense is used with both verbs être and avoir and is the most common sub-relation. The relation comp:aux@pass is only used with the verb être, while comp:aux@caus is only used with the verb faire.

Il y a

Mon, 01 Jan 0001 00:00:00 +0000

Il y a #

There are three possible annotations for il y a in French:

Il y a as an adposition
Il y a as an expletive construction
Il y a as a locative pronoun

`Il y a` as an adposition #

Sometimes, il y a is used to express a temporal argument.

pattern {
  N1 [form="il"]; N2 [lemma="y"]; N3 [lemma="avoir"]; N1 << N2; N2 << N3;
  N3-[comp]-> N2; 
  N3-[subj]-> N1
}

`Il y a` as an expletive construction #

Sometimes, il y a is used to express an expletive construction.

Modifiers of predicative constructions

Mon, 01 Jan 0001 00:00:00 +0000

Modifiers of predicative constructions #

Modifiers of predicative constructions in the form ((VERB | AUX )+ ADJ) are attached to the verb if they are sentence modifiers (temporal, locative, judgment … modifiers), otherwise they arre attached to the adjective.

You can find more information about this decisions here

Multi-word placenames and organizations

Mon, 01 Jan 0001 00:00:00 +0000

Multi-word placenames and organizations #

In Naija, multi-word placenames and organizations are currently annotated with a simple flat relation, though their constituents retain their typical parts of speech.

Naija Copulas

Mon, 01 Jan 0001 00:00:00 +0000

dey, be, na and the zero copula #

dey #

The term dey in Naija performs two primary roles. The first is that of a copula. In these instances, dey is annotated as a verb and is connected to its complement with a comp:pred relation, as in the examples below.

This word is also used as an auxiliary verb which marks the imperfective aspect. In these cases, dey is annotated as an auxiliary and is connected to the following verb with a comp:aux relation.

Numbers and dates

Mon, 01 Jan 0001 00:00:00 +0000

Numbers and dates #

If the number contains a decimal, the point is marked as a noun and is integrated into the number with a simple flat relation.

If numerals are listed a sequence, such as in telephone numbers, the constituents are chained together with the conj:coord relation.

Pronominal verbs

Mon, 01 Jan 0001 00:00:00 +0000

Pronominal verbs #

Reflexive marker (se, me, te, nous, vous) are analysed as a reflexive pronoun Reflex=Yes with one of the lemmas moi, toi, soi, nous, vous)

Four relations are considered for the reflexive marker se: comp:obj, comp:obl, comp@expl, and comp@pass.

pattern { 
  X [Reflex=Yes, lemma = "moi"|"toi"|"soi"|"nous"|"vous"];
  e: Y -> X;
}
  e.label

The semantic distinction between reflexive meaning (je me rase) and reciprocal meaning (ils s’aiment) is not marked.

Reported speech

Mon, 01 Jan 0001 00:00:00 +0000

Reported speech #

Reported speech has a feature Reported=Yes on its head. It is generally the comp:objof a speech verb, such as dire ’to say'.

en mode: Reported speech can be introduced by the idomatic preposition en mode.

être là Sometimes, en mode is absent and there is a direct relation between the idiom être là and the reported speech, which we decide to label comp:obj.

Titles and honorifics

Mon, 01 Jan 0001 00:00:00 +0000

Titles and honorifics #

Honorifics such as Mister or President are connected to the names they precede with a simple flat relation.

However, this is not the case when a title is connected to a determiner or otherwise modified in some way. In these cases, a mod:appos relation is used.

udep

Mon, 01 Jan 0001 00:00:00 +0000

`udep` relation #

Universal #

The udep relation is used for complements when one does not wish to or is unable to distinguish between arguments and modifiers. This relationship can notably be used in cases where the complement’s relationship with its governor is ambiguous. When there is little debate about the nature of the relationship, the more specific label is preferred.

pattern { X -[udep]-> Y }

The governor and the dependent of a udep relation can have any POS. The udep relation is frequently given to the complements of nouns, since it is generally more difficult to distinguish between arguments and modifiers for nouns than it is for verbs.

VERB pas de NOUN

Mon, 01 Jan 0001 00:00:00 +0000

VERB + pas de + NOUN #

In the construction VERB + pas de + (NOUN | PRON), we consider de as a preposition introducing NOUN|PRON and depending on VERB through comp:obj.

Examples:

Je ne veux pas de ça ici. (I don’t want that here)
```
  veux -[comp:obj]-> de. de -[comp:obj]-> ça
```

Je ne mange pas de fruits. (I don’t eat fruit.)

  mange -[comp:obj]-> de. de -[comp:obj]-> fruits

Mange de ce gâteau ! (Eat some of this cake!)

vocative

Mon, 01 Jan 0001 00:00:00 +0000

`vocative` #

Universal #

The vocative relation is used when a speaker directly addresses a listener in an utterance.

pattern { X -[vocative]-> Y }

French

compound

Mon, 01 Jan 0001 00:00:00 +0000

`compound` #

Universal #

SUD shares with with UD its compound relation, though the exact usage of this relationship is determined on a language-by-language basis. SUD annotations for some languages, such as French, do not use the compound relation at all.

English

@scrap

Mon, 01 Jan 0001 00:00:00 +0000

`@scrap` • Deprecated #

The @scrap feature was previously used for annotating incomplete utterances or constructions.

It has been replaced by the MISC node feature Scrap.

⚠️ parataxis:obj

Mon, 01 Jan 0001 00:00:00 +0000

`parataxis:obj` #

⚠️ deprecated

Universal #

The depency relation parataxis:obj has been deleted. Direct discourse is now attached its governor by the relation comp:obj with the features Reported=Yes. See Reported Speech

ctxt

Mon, 01 Jan 0001 00:00:00 +0000

`ctxt` #

Universal #

In French, adjectives agree with nouns in gender and number. Gender on adjectives functions as an agreement morpheme, marking their relationship with the noun.

Although French adjectives generally agree in gender, many share the same form for masculine and feminine. This raises questions about whether such adjectives should be assigned a Gender feature. This phenomenon, already frequent in written French, is even more prevalent in spoken French.

ExtPos

Mon, 01 Jan 0001 00:00:00 +0000

`ExtPos` #

The ExtPos feature was introduced in SUD and it is now used in many UD treebanks as well.

The ExtPos feature is used to facilitate the annotation of idioms, titles, and other multi-word units which behave like a certain part of speech, even though none of their constituents necessarily carry that part of speech. This feature allows the annotator to preserve the internal syntactic structure between the components of these units. See Idioms and titles for examples.

Scrap

Mon, 01 Jan 0001 00:00:00 +0000

`Scrap` #

In spoken data, sometimes speakers utter a half-finished construction. In that case, it may happen that a word cannot be attached to its head, because the speaker decided not to utter it.

pattern { GOV -> DEP; DEP[Scrap=Yes] }

Universal #

French

pronouns

Mon, 01 Jan 0001 00:00:00 +0000

Pronouns in French #

Universal #

French pronouns encode grammatical features such as person, number, case, and sometimes gender. However, several forms are ambiguous or exhibit inconsistent annotation in current treebanks. This guide clarifies how to annotate pronouns more consistently, particularly regarding French.

Lemma normalization #

Pronoun lemmas should be unified across paradigms to reflect core referential identity rather than surface distinctions.

Subject pronouns

In many current annotations, the singular pronouns “il” and “elle” are assigned the lemma lui, while the plural pronouns “ils” and “elles” have a different lemma, eux. This distinction is unnecessary and inconsistent because these pronouns share core referential properties and syntactic distribution. Therefore, we propose unifying both singular and plural third-person pronouns under the single lemma lui, rather than maintaining separate lemmas like eux.

Sauf, soit, dont as CCONJ

Mon, 01 Jan 0001 00:00:00 +0000

Sauf, soit, dont as CCONJ #

sauf, soit, dont can be used as markers in lists. They can link NP, as well as PP:

tous les couches, sauf la première ‘all layers, except the first one’
partout, sauf en France ’everywhere, except in France' This property brings these markers closer to coordinating conjunctions than to adpositions. We decided to analyze them as CCONJs.

We distinguish here between:

Coordination (conj:coord)
Apposition (conj:appos)

“sauf” #

Function: Subtractive connector, introduces an exception to a preceding phrase.
POS: CCONJ
Deprel: Prefer conj:appos over conj:coord when the second conjunct is a proper subset or exception.

@fixed • Deprecated

Mon, 01 Jan 0001 00:00:00 +0000

`@fixed` • Deprecated #

Universal #

The @fixed feature was previously used for multi-word expression.

It has been replaced by nodes features: see here

@x • Deprecated

Mon, 01 Jan 0001 00:00:00 +0000

@x • Deprecated #

Universal #

NOT USED ANYMORE. REPLACE BY FEATURES Subject

The @x deep feature was used to annotate relationships with verbs that do not have an explicitly marked subject, but whose semantic subject can nevertheless be found in the sentence. This relationship therefore applies essentially to infinitive verbs which relate to a subject introduced earlier in the sentence. For example, in the sentence he wants to eat, he has a clear semantic relationship with to eat.

Mon, 01 Jan 0001 00:00:00 +0000

Shared #

Universal #

Left dependents on coordination are attached to the head of the coordination. Since the head of a coordination is the head of the leftmost conjunct, it is not possible to distinguish these dependents from the own dependents of the leftmost conjunct.

Right dependents on a coordination are attached to the head of the rightmost conjunct, but they can be confused with the own dependents of this conjunct.

Mon, 01 Jan 0001 00:00:00 +0000

The Beja treebank contains 684 words if we count both stems and clitics; 39% of words are clitics. The
treebank contains 244 affixes for 418 stems, or a proportion of 58%. 59% of them are suffixes and 41%
prefixes. 88% of the affixes are on verbs, 7% on nouns, and 5% on auxiliaries. All affixes on nouns are
suffixes. Not all inflectional morphemes are affixes: 44% of the stems are in fact inflected forms,
containing an inseparable inflectional morpheme, which increases the proportion of inflectional
morphemes to 102% (102 inflectional morphemes for 100 stems).
Beja is a head-final language: only 11% of the dependencies between two stems have the governor
before the dependent in the SUD version of the treebank. Among the 31 dependencies concerned, 11
are for modifiers, 6 for discourse markers, 4 for dislocated objects, 2 for objects in canonical position,
and 2 for determiners. Clitics occur on both sides: 47% are proclitics and 53% are enclitics. Clitics are
mainly on verbs (56%) and nouns (38%). Clitics on nouns are determiners (70%), possessives (15%),
postpositions (11%), and coordinating conjunctions (3%). Clitics on verbs are subordinating
conjunctions (35%), object pronouns (25%), coordinating conjunctions (14%), an optative particle (2%),
and, on nominalized forms, determiners (15%) and copulas (8%).
Beja adpositions are postpositions, either as independent words (Figure 10) or as enclitics
(Figure 11)

Mon, 01 Jan 0001 00:00:00 +0000

In Beja, verbal and nominal coordination are expressed with different enclitic morphemes. The texts
contain 20 tokens of the verbal coordinating conjunction =t (and its allomorphs =it and =ajt) (Figure
13). For half of the tokens, the conjunctions occur at the end of a prosodic unit, be it a major or a minor
prosodic break, or a sentence. For this reason, we attach the coordinating conjunctions to the first
conjunct and we consider that the second conjunct is the head of the coordination. See Kanayama et al.
(2018) for a similar analysis in two other head-final languages, Japanese and Korean. As the conj relation
is forbidden from right to left in UD, we introduced a dep:conj relation

Mon, 01 Jan 0001 00:00:00 +0000

`ADV` + de + `NOUN` #

In order to analyse these expression :

(un) peu de + NOUN
beaucoup de + NOUN
plein de + NOUN
nombre de + NOUN

We use the features Idiom and ExtPos to consider the initial ADV as a PRON in theses expressions. So we have the following annotation :

pattern { N1 [upos=ADV]; N2 [form=“de”]; N3 [upos=NOUN] ;N1 < N2 ; N2<N3}

un peu de #

beaucoup de #

plein de #

nombre de #

Mon, 01 Jan 0001 00:00:00 +0000

AUX or VERB? #

It is sometimes not trivial to chose between the part of speech AUX and VERB for être.

AUX #

être is an AUX:

when it has a predicative complement (it is considered as a copula)

when it is a tense auxiliary
when it is a passive auxiliary

VERB #

être is a VERB:

Mon, 01 Jan 0001 00:00:00 +0000

être situé #

For VERB that are in relation of comp:pred with the AUX être, we chose this analysis :

B: est -[comp:pred]-> située:VERB

The same relation comp:pred with a verbal dependant can be find with several other verbs:

pattern { X[lemma="être"]; X -[comp:pred]-> Y; Y[upos=VERB] }
  Y.lemma

Mon, 01 Jan 0001 00:00:00 +0000

puis · Issue #15 · surfacesyntacticud/guidelines (github.com)

Mon, 01 Jan 0001 00:00:00 +0000

un peu #

We analyse peu from un peu as a NOUN with an ExtPos=ADV because it can modify an ADJ, ADV, VERB and an SCONJ. We also add the features Idiom.

You can find the conversation about this annotation here

Mon, 01 Jan 0001 00:00:00 +0000

Interjections and other discourse markers #

BOUGER cette section dans les INTJ !!!

Pure interjections (such as ah, hein, ouais, euh, etc.) are analysed as INTJs. Discourse markers coming from other POS (such as enfin, chouette, disons, etc.), as well as idioms (such as en fait, tu sais, etc.), keep their original POS but have an additional ExtPos = INTJ feature. Except 4 of them which are frequent and are analysed as pure INTJs: bon, ben, quoi, and tiens.

Mon, 01 Jan 0001 00:00:00 +0000

Cleft constructions #

As shown in the page about the comp:cleft relation, the cleft constructions are analysed with a comp:cleft relation going from the head of the main sentence to the head of the complement.

With qui and que direct object, the cleft clause is analysed as a relative clause. in particular, qui and que are analysed as relative pronouns (PRON).

Example #

In the other cases, the cleft clause is analysed as a complement clause, with que as a SCONJ.

Mon, 01 Jan 0001 00:00:00 +0000

Complement of light verb construction #

The complement of a VERB -[comp:obj@lvc]-> OBJ construction is attached to OBJ if it can be considered independently of VERB, otherwise it is attached to VERB.

Examples :

j’ai peur de mourir (I’m afraid to die) - la peur de mourir me hante (the fear of dying haunts me) peur -[comp:obl]-> de

Je me rends compte de mon erreur (I realize my mistake ) rends -[comp:obl]-> de

See issue #6

Mon, 01 Jan 0001 00:00:00 +0000

Interrogative #

Qu’est-ce que: Interrogatives with PRON est-ce qui/que are analysed as particular cases of cleft clauses.

Est-ce que ? #

Example #

Qu’est-ce que ? #

Example #

Mon, 01 Jan 0001 00:00:00 +0000

`mod` or `discourse` ? #

In Spoken, we much more use the relation discourse, because there is much more discourse marker in oral productions, but also because we become aware that some units can be discourse markers rather than modifiers.
Nevertheless, even in Spoken, we often hesitate between discourse and mod and some annotations are not very consistent. For instance, en fait is analyzed in both ways for similar constructions:

Mon, 01 Jan 0001 00:00:00 +0000

mod/m #

TODO

Mon, 01 Jan 0001 00:00:00 +0000

ADP #

We follow the same rules than Chinese Mandarin UD for ADP but with the following two differences:

the addition of some instances of 和 which can act as a preposition in some constructions.
These adpositions are the head of the adpositional phrase and govern (as it is in Stanford Chinese). This way, the hierarchical information between the preposition and postposition is not lost as it is in UD (and criticized by the author of Chinese UD guidelines themselves in \cite{ud_hk_2016})
Locative

Mon, 01 Jan 0001 00:00:00 +0000

AUX in mandarin #

We follow the same convention than Mandarin UD for AUX except that

the three aspect markers 过着了 are annotated as particle (PART) instead of auxiliaries.
肯 / kěn “be willing to”, 敢 / gǎn “dare” are NOT annotated as auxiliaries (as they can be modified by the aspect marker 过) and we don’t think they have a specific syntactic function in the sentence.

Mon, 01 Jan 0001 00:00:00 +0000

Aspect in Mandarin #

Mon, 01 Jan 0001 00:00:00 +0000

Dative Shift #

According to Liu (2066), mandarin has 3 possible dative construction, and not any verbs can be used in all of them. ( https://en.wikipedia.org/wiki/Dative_shift#Mandarin)

Remarks #

At the moment, we have a group of words which we annotes as ADP that looks like these dative construcition (给 / 不如 / 让 /(比？)). It seems that it could also be seen as a dative verb construction with a VERB and two complements.

Mon, 01 Jan 0001 00:00:00 +0000

Directional Complement in Mandarin Chinese (comp:dir) #

A “directional complement” (known as “directional verb compound” in UD) consists of a series of at least two verbs where the second verb is one of the directional or deitic motion verbs (or a combination of the two) in the below exhaustive list :

VERB + A : Directional motion #

Eight verbs expressing directional motion:

上 (shàng) - “我爬上山” (wǒ pá shàng shān) - “I climb up the mountain”
下 (xià) - “他跑下楼” (tā pǎo xià lóu) - “He runs down the stairs”
出 (chū) - “我走出房间” (wǒ zǒu chū fángjiān) - “I walk out of the room”
进 (jìn) - “我们走进电影院” (wǒmen zǒu jìn diànyǐngyuàn) - “We walk into the movie theater”
回 (huí) - “他跑回家” (tā pǎo huí jiā) - “He runs back home”
过 (guò) - “他跳过河” (tā tiào guò hé) - “He jumps over the river”
开 (kāi) - “她拉开门” (tā lā kāi mén) - “She pulls open the door”
起 (qǐ) - “他拿起书” (tā ná qǐ shū) - “He picks up the book”

The main verb is the head of it’s directional complement : VERB<-[comp:dir]- A

Mon, 01 Jan 0001 00:00:00 +0000

Filler words in Mandarin Chinese #

In Mandarin Chinese, there are several filler words (also known as “filler sounds” or “hesitation sounds”) that are used, much like “um”, “uh”, “like”, “you know”, etc. in English. Here are some of the most common ones:

那个 (nèi ge / nà ge) - Similar to “uh”, “um”, or “like” in English.
这个 (zhè ge) - Used similarly to 那个 (nèi ge / nà ge). It can also mean “this one” in other contexts.
嗯 (ń / ň / ǹg) - Used to show agreement or acknowledgment, similar to “uh-huh” in English.
就是 (jiù shì) - Similar to “it’s just like”, “you know”, or “so” in English.
对 (duì) - Means “correct”, but it can also be used as a filler word similar to “right” in English.
然后 (ránhòu) - Similar to “and then” or “so then” in English.

As in other languages, filler words in Mandarin should be used sparingly in formal speech or writing but can make informal speech sound more natural.

Mon, 01 Jan 0001 00:00:00 +0000

Interrogative Clauses #

Direct Interrogative Clauses #

yes/no questions #

Using 吗 #

By using 吗 (ma) at the end of the sentence, we can turn any sentence in a yes/no question. This form is straightforward and neutral in tone, making it suitable for a wide range of contexts without implying urgency.

Using the positive/negative/positive pattern VERB/ADJ + 不/没 + VERB/ADJ #

For indicating a choice between doing and not doing an action (or being and not being in the case of some adjectives), the positive/negative/positive structure can be used. It emphasizes the dichotomy, almost as if saying “or not” in English. In some contexts, especially when spoken with particular intonations, this form can sound a bit more pressing or urgent, as if seeking an immediate answer or confirmation.

Mon, 01 Jan 0001 00:00:00 +0000

Measure words in Mandarin Chinese #

“Measure words”, commonly referred to as “classifiers” in linguistics, are utilized in conjunction with numbers to denote the quantity of a noun or, on occasion, an action. These terms, measure words, essentially classify or categorize the noun or action under consideration for counting.

How to annotate #

The head of the measure head will be the noun it qualify.
The dependant will be the quantity and/or the determinant

Mon, 01 Jan 0001 00:00:00 +0000

Negation in Chinese #

There are several ways to express negation in Mandarin Chinese depending on the context. Two ways are syntactic while the other are morphological.

Syntactical negation #

1) 不 (bù) #

This is the most common way to express general negation. It’s typically used before a verb or an adjective. For example, “我们不喝酒。” (Wǒmen bù hējiǔ) means “We don’t drink alcohol.”.

Mon, 01 Jan 0001 00:00:00 +0000

New Sentence #

In case a sentence is made of two seperated sentence, we encode this as a parataxis:new relation between the root of the second sentence to the root of the first sentence. The final punctuation of each sentence is attached to its corresponding root.

Mon, 01 Jan 0001 00:00:00 +0000

Passive construction in Mandarin Chinese #

“Passive construction”, also known as “passive voice”, is a grammatical structure that shifts the focus from the agent of the action to the recipient. In passive voice, the subject of the sentence becomes the entity affected by the action, while the doer is often omitted or indicated by the passive marker (typically auxiliary verbs) in Madarin Chinese.

Here are the most commun four auxiliary verbs used as the passive voice indicator :

Mon, 01 Jan 0001 00:00:00 +0000

Predicative construction #

是 #

他是老师

算 #

好像 #

看起来 #

Mon, 01 Jan 0001 00:00:00 +0000

Resultative Complement (comp:res) #

This relationship applies to verb-verb and verb-adjective combinations in which the second verb or adjective encompasses what are termed “resultative complements” and “phase complements” in the study of Chinese linguistics.

Like directional complements, the affirmative potential 得 de and negative potential 不 bù can intervene between the compound. However, unlike directional complements, this is the only situation in which the resultative and phase resultative complements can be separated.

Mon, 01 Jan 0001 00:00:00 +0000

Sentence Final Particle (discourse:sp) #

Like in Chinese UD, we use the discourse:sp relation to annotate sentence final particle.

Citation of Leung 2016 :

Similar to classifiers, sentence-final particles are an areal feature across many East and Southeast Asian languages. They occur at the end of a sentence or utterance (8), and may have a wide variety of nonreferential meanings and functions that modify the entire sentence, including modality, speech register, “speaker/hearer orientation”, and other discourse and pragmatic uses. In Mandarin, these particles include 嗎 ma (interrogative), 了 le (new situation), 吧 ba (command/suggestion), 的 de (certainty), among others. They are integrated into the sentence they attach to as part of its prosodic contour with no pause in between.

Mon, 01 Jan 0001 00:00:00 +0000

Serial Verbs Construction in Mandarin Chinese (compound:svc) #

See the universal explanation of UD compound:svc

In Mandarin Chinese, a Serial Verb Construction (SVC) is a sequence of verbs or verb phrases that share the same subject without any explicit conjunction or marker of subordination between them. This grammatical structure is used to express various complex actions, states, and events. SVCs are very common in Mandarin and provide a concise way to express complex ideas.

Mon, 01 Jan 0001 00:00:00 +0000

Subordinating conjunction in Mandarin Chinese #

Subordinating conjunction is a conjunction that links two clauses within a sentence, creating a hierarchy where one clause relies on the other. It introduces the dependent clause, which cannot exist independently as a sentence and instead depends on the main clause for meaning. In Madarin Chinese, the subordinating conjunction can mark both the main clause and the subordinate clause. Subordinating conjunction indicates various relationships between the clauses, here are the most commun relations :

Mon, 01 Jan 0001 00:00:00 +0000

`comp` deprel #

pattern { e:X -[1=comp]-> Y }
  e.label

`comp:obj@R`: Recipient #

`comp:obj@T`: Theme #

`comp:obj@G`: Goal #

`comp:obj@L`: Locative #

@Bernard: un seul comp:obj verbal avec le dépendant à gauche:

Mon, 01 Jan 0001 00:00:00 +0000

`compound` deprel #

The base relation compound is used without extension or in one of the three relations compound:prt, compound:svc and compound:redup.

pattern { e:X -[1=compound]-> Y }
  e.label

Il n’y a que 5 cas ( https://universal.grew.fr/?custom=68e2ba235e8b1) entre 2 AUX, je ne sais pas si c’est important de garder l’exemple…

============ Old guidelines ============ #

compound:svc #

The compound:svc relation defines a serial verb construction that is normally connecting two (different) verbs. In Zaar, this relation can also connect two auxiliaries that are repeated.

Mon, 01 Jan 0001 00:00:00 +0000

============ Old guidelines ============ #

J’avais finalement une question concernant le point (1) ordre Verbe-Complément différent pour le progressif % aux autres TAM; ordre qui se retrouve dans les syntagmes nominaux ayant un nom verbal comme tête. Je pensais avoir compris sur le coup, mais en écrivant cette requête

pattern { GOV -[comp:aux]-> DEP ; GOV[Aspect=“Prog”]; DEP[VerbForm=“Vnoun”]}

et en comparant avec les autres auxiliaires non progressifs je n’ai finalement pas réussi à voir de différence. J’avais pourtant noté quelque chose comme suit : le progressif suivi d’un nom verbal peut prendre un complément avant ou après le verbe, tandis qu’il est toujours après le verbe en temps normal. J’ai l’impression d’avoir loupé quelque chose.

Docs on SUD Guidelines

-t-il

How to annotate a-t-il ? #

ADJ

Adjective #

Universal #

aff

aff #

Universal #

French #

Overview #

Specific Pattern #

Aspect

Aspect #

Beja #

Case

Case #

CCONJ

Coordination conjunction #

Universal #

comp

comp #

Universal #

French #

Overview #

comp:aux

comp:aux #

Universal #

comp:cleft

comp:cleft #

Universal #

comp:obj

comp:obj #

Universal #

Direct object complements #

comp:obl

comp:obl #

Universal #

comp:pred

comp:pred #

Universal #

predicative arguments of copula #

Comparative

Comparative constructions #

conj:appos

conj:appos #

Universal #

conj:coord

conj:coord #

conj:dicto

conj:dicto #

Universal #

Coordination

Coordination #

Chained conjuncts #

CorrectForm

CorrectForm #

Universal #

Definite

Definite #

Universal #

Values #

Degree

Degree #

Universal #

Values #

DET

Determinant #

Universal #

Haitian Creole #

Overview #

Specific Pattern #

Yon determiner #

Disfluency

Disfluency #

Repetitions and reformulations #

dislocated

dislocated #

Universal #

Foreign

`Aspect` #

`Case` #

`comp` #

`comp:aux` #

`comp:cleft` #

`comp:obj` #

`conj:coord` #

`CorrectForm` #

`dislocated` #

`Gender` #

`Number` #

`parataxis` #

`parataxis:insert` #

`parataxis:parenth` #