home edit page issue tracker

UD for Northern Luri

Tokenization and Word Segmentation

• Words are delimited by whitespace. Punctuation marks (periods, question marks) are segmented as separate tokens.
• There are no multiword tokens in this treebank.
• There are no empty nodes used.
• Clitic pronouns and enclitic copulas are written attached to their host with = or - (e.g., =mo, =ēm, -é), but are tokenized as separate tokens in the treebank. Each such clitic receives its own line with the appropriate UPOS tag (PRON or AUX) and morphological features.

---

Instruction: The words are delimited based on whitespace and punctuation.

---

Morphology

Tags

• The following UPOS tags are used in this treebank: ADJ, ADP, ADV, AUX, CCONJ, DET, NOUN, NUM, PART, PROPN, PRON, PUNCT, SCONJ, VERB.
• The following UPOS tags are not used: INTJ, SYM, X.
• PART is used for clitics and particles that function as case markers or discourse particles. These include the negation/object marker né, enclitic definiteness/specificity markers (e.g., =ê, =ē, =é, -é), and the discourse particle Na. Many PART tokens carry ExtPos=ADP when functioning as postpositions or case markers.
• AUX covers the copula būdan (to be) and its enclitic present-tense forms (e.g., =ēm, =a), the progressive auxiliary dāštan, and the prospective/future auxiliary xāstan. The clitic copula forms (e.g., =a, =ēm) can appear as the sole root of a clause and are tagged AUX even in that position.
• DET is used for demonstratives (e.g., i, ū, ya) and quantifiers (e.g., gérd PronType=Tot, hama PronType=Tot, yé Definite=Ind). These are distinguished from PRON by their attributive (prenominal) function. Pronouns that stand alone as nominal heads are tagged PRON.
• PRON includes personal pronouns (free and enclitic possessive forms such as =mo, =m, =ēš), reflexive pronouns (xot), and indefinite pronouns (hiška). Possessive clitics carry Poss=Yes.
• Adjectives (ADJ) appear as both prenominal modifiers (amod) and as predicative complements (xcomp). Comparative adjectives carry Degree=Cmp; base forms carry Degree=Pos.
• There are no verbal forms tagged as ADV or NOUN. Verbal nouns and infinitival forms are tagged VERB with VerbForm=Inf.

---

Features

• NOUN: inflects for Number (Sing, Plur) and Definite (Def, Spec). Definiteness is typically marked by enclitic suffixes rather than free-standing articles.
• VERB: carries a rich set of inflectional features. The following are attested:
  • Aspect: Imp (imperfective), Perf (perfect), Prog (progressive), Prosp (prospective)
  • Mood: Ind (indicative), Imp (imperative), Sub (subjunctive), Opt (optative)
  • Tense: Pres, Past, Fut
  • Person: 1, 2, 3
  • Number: Sing, Plur
  • VerbForm: Fin, Inf
  • Voice: Act (active), Pass (passive)
  • Polarity: Pos, Neg (negation encoded on the verb itself)
• PRON: encodes PronType (Prs, Ind), Person (1–3), Number (Sing, Plur), Poss (Yes), Reflex (Yes), and Case (Acc) for reflexive forms.
• ADJ: attests Degree (Pos, Cmp).
• DET: encodes PronType (Dem, Tot, Art) and Definite (Ind) for the indefinite article yé.
• ADP: carries ExtPos=ADP on PART tokens that function as postpositions in a larger phrase.
• NUM: no features beyond POS are annotated in the current release.

---

Syntax

• The primary word order is SOV. The root of a sentence is almost always a finite VERB or a clitic copula (AUX).
• Subjects are tagged nsubj (for active clauses) and nsubj:pass (for passive clauses). Subjects are predominantly NOUN, PRON, or PROPN.
• Objects are tagged obj (direct object) and iobj (indirect object). The object marker né (PART) is used to mark specific/definite direct objects, attached via the case relation.
• Oblique arguments use obl, with the subtype obl:tmod for temporal modifiers.
• Copula constructions: non-verbal predication is handled via the copula būdan and its clitic forms, tagged AUX with the relation cop. Predicate nominals and predicate adjectives are the head of the clause, with the copula attached as a dependent. Clitic copulas (e.g., =ēm, =a) can also appear as the root of the clause directly.
• Light verb constructions: complex predicates consisting of a nominal or adjectival element plus a light verb are annotated using compound:lvc (e.g., némâz mēka ‘prays’). The nominal element is the compound:lvc dependent of the light verb.
• xcomp is used for predicative complements — both nominal (NOUN) and adjectival (ADJ) — in secondary predication constructions.
• Possessive clitics (e.g., =mo, =m, =ēš) attach to their nominal host and are annotated as nmod:poss.
• The following relation subtypes are used:
  • nsubj:pass — passive subject
  • nmod:poss — possessive nominal modifier
  • compound:lvc — light verb construction
  • acl:relcl — relative clause modifier
  • acl:lvc — adjectival light verb construction
  • obl:tmod — temporal oblique modifier

---

Treebanks

There are N Northern Luri UD treebanks:

• Northern Luri-A
• Northern Luri-B

---

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

.