UD for Romanian 
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 always tokenize them as separate tokens (words); that holds even for hyphenated words pronounced joining two syllables in a single one, such as ne-am, v-ați.
Morphology Tags This is an overview only. For more detailed discussion and examples, see the list of Romanian POS tags and Romanian features. • Romanian uses all 17 universal POS categories.
• The pronoun (PRON) vs. determiner (DET) distinction is based on the syntactic position; PRON is independent, DET is subordinated to a NOUN or PRONOUN.
• CONTEMPORARY ROMANIAN (RRT) has three auxiliary verbs (AUX), fi (“to be”), avea (“to have”) vrea (“to want”). In Nonstandard Romanian, the auxiliaries putea (“can”), trebui (“must”), ști (“to know”) were annotated. The auxiliary vrea (“to want”) is used in the old language in many forms than in the contemporary, ex.: participle.
o The copula with non-verbal predicates was annotated as AUX.
o Periphrastic future tense (vrea + infinitive of the main verb; vrea + fi+ participle of the main verb).
o Periphrastic past tense (avea + participle of the main verb).
o Periphrastic conditional (conditional form of avea + future of the main verb; conditional form of avea+ fi + participle of the main verb).
o Periphrastic passive (any form of fi, including periphrastic forms, + participle of the main verb). o • There are four main (de)verbal forms, distinguished by the UPOS tag and the value of the VerbForm feature:
o Infinitive Inf, tagged VERB or AUX.
o Finite verb Fin, tagged VERB or AUX.
o Participle Part, tagged VERB or AUX (in the nonstandard treebank) or ADJ (in the RRT treebank).
o Gerundive verb Ger, tagged VERB or AUX.
Nominal Features
• Nominal words (NOUN, PROPN and PRON) have an inherent Gender feature with one of two values: Masc or Fem. The neuter is in Romanian classified as masc singular and feminine plural.
o The following parts of speech inflect for Gender because they must agree with nouns: ADJ, DET, NUM, VERB, AUX. For verbs (including auxiliaries), only participles have gender.
• 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, participles), marginally NUM.
• Case has 5 possible values: Nom, Gen, Dat, Acc, Voc. It occurs with the nominal words, i.e., NOUN, PROPN, PRON, ADJ, DET, NUM. It can occur with participles but only with those tagged as ADJ. To nouns there is no distinction of form, they are annotated as nom.,-acc. and gen.,-dat. Distinct forms for acc., dat. have only some pronouns.
o The Case feature also occurs with prepositions (ADP). Here it is a lexical feature. Prepositions do not inflect for case but they subcategorize for the case of their noun phrase.
Degree and Polarity
• Degree applies to adjectives (ADJ) and adverbs (ADV) and has (rarely, in RRT) one of three possible values: Pos, Cmp, Sup.
• Polarity has two values, Pos and Neg, and applies primarily to verbs (VERB, AUX), adjectives (ADJ) and adverbs (ADV) that can be negated using the paricle “nu”, the ADV “nici” (or the morpheme “ne-” in nonstandard treebank).
o The Polarity feature is not used with pronouns and determiners, although there is a subset of negative pronouns and determiners. The PronType=Neg feature is used there instead.
Verbal Features o • Finite verbs always have one of three values of Mood: Ind, Imp or Subj.
• Verbs in the indicative mood have one of four values of Tense: Past simple, Pres, Imperfect, Pluperfect.
• The future are distinguished analytically.
• Imperative forms do not have the Tense feature
• Past and present conditionals are distinguished analytically; also the past subjunctive and the perfect future).
Pronouns, Determiners, Quantifiers
• PronType is used with pronouns (PRON), determiners (DET) and adverbs (ADV).
• NumType is used with numerals (NUM), The Poss feature marks possessive personal determiners.
• The Reflex feature is marked by the language specific features: expl:pv, expl:poss, expl:impers, and expl:pass.
• 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.
o As a cross-reference to subject, person is also marked on finite verbs (VERB, AUX).
• There are two layered features, Gender[psor] and Number[psor]. They appear with certain possessive adjectives and 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:
• NumType
• Definite
• Polite
• Variant long distinguishes long forms of infinitives and imperative negative in nonstandard.
• Variant short distinguishes the words written with hyphen, whish loss some characters in RRT.
Syntax
Core Arguments, Oblique Arguments and Adjuncts
• Nominal subject (nsubj) is a noun phrase in the nominative case, without preposition.
o An infinitive verb may serve as the subject and is labeled as clausal subject, csubj.
o A finite subordinate clause may serve as the subject and is labeled csubj.
• For the purpose of UD the objects are divided to core objects, labeled obj or iobj, and oblique objects, labeled obl.
o All prepositional objects are considered oblique.
o Accusative objects of some verbs alternate with finite clausal complements, which are labeled ccomp.
o If a verbal construction expand the iobj, it is labeled xcomp.
o If a verb subcategorizes for two core objects, one of them accusative (or ccomp) and the other non-accusative, then the non-accusative object is labeled iobj. Core nominal objects in other situations are labeled just obj.
• Adjuncts are usually prepositional phrases, but they can be bare noun phrases as well. They are labeled obl.
• For the prepositional objects which cannot be omitted we used the label nmod:pmod.
• This label is also used for the direct object which in Romanian language is preceded by the preposition “pe” to show that it names a person.
o Temporal modifiers are annotated as: nmod:tmod; advmod:tmod or advcl:tcl.
• Extra attention has to be paid to clitic forms of pronouns (accusative) and (dative). They can function as:
o Core objects with weak form can be (obj or iobj): or expl if the obj. iobj is named also by a noun or another pronoun.
o The reflexive can be expl:pv, expl:impers, expl:poss, expl:pass.
• In passive clauses (both reflexive and periphrastic passive), the subject is labeled with nsubj:pass or csubj:pass, respectively.
o The auxiliary verb in periphrastic passive is labeled aux:pass.
o If the demoted agent is present, it has the form of a bare instrumental phrase and its relation is labeled nmod:agent.
Non-verbal Clauses
• The copula verb fi (be) is used in equational, attributional, possessive and benefactory nonverbal clauses. Purely existential clauses use fi as well but it is treated as the head of the clause and tagged VERB.
Relations Overview
• The following relation subtypes are used in Romanian:
o nsubj:pass for nominal subjects of passive verbs
o csubj:pass for clausal subjects of passive verbs
o nmod:agent for agents of passive verbs
o nmod:pmod for prepositional objects which cannot be omitted as determiners of some verbs.
o expl:impers for impersonal value of the reflexive
o expl:pv for reflexive clitics of inherently reflexive verbs
o expl:pass for reflexive clitics with passive value
o expl:poss for reflexive clitics with possessive value
o aux:pass for passive auxiliaries
o nmod:tmod for prepositional nouns with temporal value
o advcl:tcl for adverbial clauses with temporal value
o advmod:tmod for adverbs with temporal value
• The following main types are also used alone: expl if a nsubj, obj or iobj is repeated by a pronoun.
Treebanks There are two UD-Romanian treebanks:
UD_Romanian-RRT
UD_Romanian-Nonstandard