home edit page issue tracker

UD for German

Tokenization and Word Segmentation

• In general, words are delimited by whitespace characters. Description of exceptions follows.
• According to typographical rules, many punctuation marks are attached to a neighboring word. We usually tokenize them as separate tokens (words) with the exception of abbreviations such as usw. “etc.” which are kept as one token with the period.
• German compounds are written as one word and we do not split them.
• There is one class of multi-word tokens: the contractions of prepositions and definite articles. Example: zum = zu + dem “to the”.

Morphology

Tags

• German uses all 17 universal POS categories, including particles (PART).
• The following words are particles in German: nicht “not”, and the infinitive marker zu “to”.
• The pronoun (PRON) vs. determiner (DET) distinction is based on word lists because the traditional grammar does not define determiners. In general, words that inflect for gender, to be able to agree with a modified noun, are tagged DET, even if they act independently in a given sentence; that includes possessives. There is more discussion on the page on DET.
• German auxiliary verbs (AUX) are:
  • sein for perfect tenses of some verbs (ich bin gekommen “I have come”) and as copula (er ist alt “he is old”)
  • haben for perfect tenses of the remaining verbs (ich habe gegessen “I have eaten”)
  • werden for future tense (ich werde reisen “I will travel”) and for the passive (es wurde gegessen “it was eaten”)
  • modal verbs dürfen “may”, können “can”, mögen “want”, müssen “must”, sollen “shall”, wollen “want”
  • The verbs sein, haben and werden can also occur as normal verbs (VERB), meaning “be, have, become”.
• There are four main (de)verbal forms, distinguished by the UPOS tag and the value of the VerbForm feature:
  • Infinitive Inf, tagged VERB or AUX.
  • Finite verb Fin, tagged VERB or AUX.
  • Participle Part, tagged VERB or AUX.
  • Verbal noun Vnoun, tagged NOUN, looks like a capitalized infinitive but has an article and may inflect.

Nominal Features

• Nominal words (NOUN, PROPN and PRON) have an inherent Gender feature with one of three values: Masc, Fem or Neut.
  • The following parts of speech inflect for Gender because they must agree with nouns: ADJ, DET, VERB, AUX. For verbs (including auxiliaries), only participles can inflect for gender. Finite verbs don’t.
• The two main values of the Number feature are Sing and Plur. The following parts of speech inflect for number: NOUN, PROPN, PRON, ADJ, DET, VERB, AUX (finite and participles).
• Case has 4 possible values: Nom, Gen, Dat, Acc. It occurs with the nominal words, i.e., NOUN, PROPN, PRON, ADJ, DET. However, case forms of nouns are extremely ambiguous and most of the time the case is distinguished only by the form of the article.
• Definite has 2 values: Ind, Def. It is used to distinguish the indefinite and definite articles (DET).

Degree and Polarity

• Degree applies to adjectives (ADJ) and adverbs (ADV) and has one of three possible values: Pos, Cmp, Sup.
• Polarity is used to mark the negative particle nicht, i.e., only the Neg value is used.

Verbal Features

• Finite verbs always have one of three values of Mood: Ind, Imp or Sub (called Konjunktiv in German).
• Indicative and subjunctive verbs always have one of two values of Tense: Past, Pres.
  • In the subjunctive mood, the tense feature is used to distinguish Konjunktiv I (Pres) and Konjunktiv II (Past).
  • Imperative forms do not have the Tense feature.
  • The Tense feature is also used to distinguish present and past participles (kommend “coming” vs. gekommen “come”).
• The features Aspect and Voice are not used in German because both the perfect aspect and the passive voice are expressed periphrastically.

Pronouns, Determiners, Quantifiers

• PronType is used with pronouns (PRON), determiners (DET) and adverbs (ADV).
• NumType is used with numerals (NUM), adjectives (ADJ) and determiners (DET).
• The Poss feature marks possessive personal determiners (e.g. mein “my”), possessive interrogative or relative determiners (e.g. wessen “whose”).
• The Reflex feature is always used together with PronType=Prs and it marks reflexive pronouns (mich/mir, dich/dir, sich, uns, euch). Note that their forms in the first and second person are ambiguous with irreflexive accusative forms, and the Reflex feature must be decided by context.
• Person is a lexical feature of personal pronouns (PRON) and has three values, 1, 2 and 3. With personal possessive determiners (DET), the feature actually encodes the person of the possessor. Person is not marked on other types of pronouns and on nouns, although they can be almost always interpreted as the 3rd person.
  • As a cross-reference to subject, person is also marked on finite verbs (VERB, AUX).
• The Polite feature distinguishes informal second-person pronouns (du, ihr, Polite=Infm) from the formal Sie (Polite=Form). The formal pronoun is phonologically equivalent in all its case forms to the third-person plural sie but it is distinguished in orthography by the capital letter S. We tag it as second person (because that is its meaning) and we do not tag its number (because it is used both for singular and plural addressees) despite the fact that it combines with third-person plural verbs. The parser must learn that Person=2|Polite=Form subject attaches to Number=Plur|Person=3 verbs, while Number=Sing|Person=2|Polite=Infm subject attaches to Number=Sing|Person=2 verbs.
• There are two layered features, Gender[psor] and Number[psor]. They appear with certain possessive determiners and encode the lexical gender/number of the possessor. The extra layer is needed to distinguish these lexical features from the inflectional gender and number that mark agreement with the modified (possessed) noun.

Other Features

• Besides the layered features listed above, there are several other language-specific features:
  • Typo
• The following universal features are not used in German: Animacy, Evident.

Syntax

This is an overview only. For more detailed discussion and examples, see the list of German relations.

Core Arguments, Oblique Arguments and Adjuncts

• Nominal subject (nsubj) is a noun phrase in the nominative case, without preposition.
  • A finite subordinate clause may serve as the subject and is labeled csubj.
  • If an infinitive verb is to serve as the subject, it becomes a verbal noun (it gets the neuter singular nominative article and it is capitalized in writing), thus it is labeled nsubj.
• Objects defined in the German grammar may be bare noun phrases in accusative, dative or genitive, or prepositional phrases in accusative, dative or genitive. For the purpose of UD the objects are divided to core objects, labeled obj, and oblique objects, labeled obl:arg.
  • Bare accusative objects are considered core.
  • Bare dative and genitive objects, and all prepositional objects are considered oblique.
    • This means that verbs of giving are not ditransitive predicates in German UD, as they have one oblique dative argument and only one core object.
    • In rare cases a verb may license two accusative objects and then the relation iobj will be used for one of them.
  • Accusative objects of some verbs alternate with finite clausal complements, which are labeled ccomp.
  • If a verb subcategorizes for the infinitive (e.g. phasal verbs or verbs of control), the infinitival complement is labeled xcomp.
• Adjuncts (or, following the German grammar, adverbial modifiers realized as noun phrases) are usually prepositional phrases, but they can be bare noun phrases as well. They are labeled obl:
  • Temporal modifiers realized as accusative noun phrases: ich arbeite jeden Tag “I work every day.”
  • All prepositional phrases that are not prepositional objects (i.e., their role and form is not defined lexically by the predicate) are adjuncts.
• Extra attention has to be paid to the reflexive pronoun sich. It can function as:
  • Core object (obj): er sah sich im Spiegel “he sighted himself in the mirror.”
  • Reciprocal core objects (obj): sie küssten sich “they were kissing each other.”
  • Inherently reflexive verb, cannot exist without the reflexive clitic, and the clitic cannot be substituted by an irreflexive pronoun or a noun phrase. In accord with the UD guidelines, we label the relation between the verb and the clitic as expl:pv, not compound. Example: wir müssen uns beeilen “we must hurry.”
• In passive clauses, the subject is labeled with nsubj:pass or csubj:pass, respectively.
  • The auxiliary verb in periphrastic passive is labeled aux:pass.
  • If the demoted agent is present, it has the form of a prepositional phrase and its relation is labeled obl:agent.

Non-verbal Clauses

• The copula verb sein (be) is used in equational, attributional, locative, possessive and benefactory nonverbal clauses.
  • Existential clauses use a different verb, geben (give) with an accusative object: es gibt Essen “there is food.”

Relations Overview

• The following relation subtypes are used in German:
  • nsubj:pass for nominal subjects of passive verbs
  • csubj:pass for clausal subjects of passive verbs
  • obl:agent for agents of passive verbs
  • obl:arg for prepositional objects
  • expl:pv for reflexive clitics of inherently reflexive verbs
  • aux:pass for passive auxiliaries
  • compound:prt for separable verb prefixes
  • det:poss for possessive determiners
  • nmod:poss for possessive modifier phrases
  • acl:relcl for adnominal relative clauses
  • advcl:relcl for relative clauses whose antecedent is a clause
• The following relation types are not used in German at all: clf, dislocated (dislocated could be used but it does not occur in the current data).

Treebanks

There are four German UD treebanks:

• German-GSD
• German-HDT
• German-PUD
• German-LIT

© 2014–2024 Universal Dependencies contributors. Site powered by Annodoc and brat

.