home edit page issue tracker

UD for Tigrinya

Tigrinya (ትግርኛ) is an Ethio-Semitic language spoken in Eritrea and northern Ethiopia, written in the Ge’ez script (called Ethiopic in Unicode). In the examples below, Tigrinya words are followed by a transliteration in italics and an English translation.

Tokenization and Word Segmentation

• Words are delimited by whitespace. Punctuation marks (፡ ። ፣ ፤ ፧ !), which are often attached to a neighbouring word, are separate tokens. Questions may end with ? as well as with the Ge’ez question mark ፧. Abbreviations written with a slash, such as ወ/ሮ / w/ro for ወይዘሮ / wəyzəro “Mrs.”, are single tokens.
• There are no words with spaces.
• Multiword tokens are frequent. The following elements are written together with their host but are segmented as separate syntactic words:
  • prepositions written as prefixes (ን- n-, ብ- b-, and ናይ / nay in ናይቲ / nayti), e.g. ነቲ / nəti = ን + እቲ “OBJ the”;
  • possessive suffixes on nouns, adjectives and verbal nouns, e.g. ኣቦይ / ʔaboy = ኣቦ + -ይ “my father”;
  • subordinating prefixes of verbs (ዝ- z-, ክ- k-, እንተ- ʔntə-, ከም kəm, ስለ slə, እንክ- ʔnk-);
  • the negation affixes ኣይ-…-ን ʔay-…-n and ዘይ- zəy-;
  • the coordinating suffix -ን -n, and enclitics such as -ስ -s “as for” and ‘ውን ‘wn “also”.
• Other verb morphology (subject and object agreement, aspect, mood and voice) is not segmented.
• Exceptions: the negative copula ኣይኮነን / ʔaykonən “is not” is one token; in the negative existential (የለን / yələn “there is not”) only the suffix -ን is segmented, since the prefix is fused with the stem; case-marked personal pronouns (ንዓኣ / nʕaʔa “her”, ከምኡ / kəmʔu “like him”, ናብኡ / nabʔu “to him”) and possessive pronouns (ናታ / nata “hers”) are not segmented; lexicalised adverbs such as ብጣዕሚ / bt’aʕmi “very” are not segmented.

Morphology

Tags

• All UPOS tags are used except SYM and X.
• PART is used for discourse and question particles (እንዶ / ʔndo, ዶ / do, -ስ / -s) and for the particles of light verb compounds (ጽን / s’n, ደስ / dəs, ድንን / dnn, …).
• The negation affixes ኣይ-, -ን and ዘይ- are tagged ADV with Polarity=Neg.
• AUX is used for the copula እዩ / ʔyu (including the negative ኣይኮነን /ʔaykonǝn), for ነበረ / nəbərə as past copula and past auxiliary, for ኣሎ / ʔalo when it follows another verb and agrees with its subject (progressive, perfect), and for forms of ኮነ / konə when they follow a verb in subordinate clauses. Elsewhere, ኣሎ is VERB (existence, possession, obligation), and ኮነ is VERB in the sense “become”.
• Determiners (the article እቲ / ʔti and demonstratives such as እዚ / ʔzi) are DET before a noun, adjective or relative clause, and PRON when used alone. Possessive suffixes are DET; separate possessive words are DET before a noun and PRON when used alone.
• Verbal nouns (ምኻድ / mxad “going”) are NOUN with VerbForm=Vnoun. Converbs are VERB with VerbForm=Conv. Ordinal numerals are ADJ with NumType=Ord.

Features

• Nouns: Gender, Number; VerbForm=Vnoun on verbal nouns; Abbr=Yes on abbreviations.
• Verbs: Gender, Number, Person (subject agreement); Gender[obj], Number[obj], Person[obj] (object suffixes); Gender[app], Number[app], Person[app] (applicative suffixes with -l-); Aspect=Imp,Perf; Mood=Imp,Ind,Jus; VerbForm=Conv; Voice=Pass,Cau (Voice=Pass marks the ተ- / tə- stem, whether the clause is passive, reflexive, reciprocal or middle); Polarity=Neg on forms with a fused negative prefix.
• Copula and auxiliaries: Tense=Pres (non-past) on እዩ / ʔyu and ኣሎ / ʔalo, Tense=Past on ነበረ / nǝbǝrǝ; Polarity=Neg on ኣይኮነን / ʔaykonǝn.
• Pronouns and determiners: PronType=Art,Dem,Prs,Int; Poss=Yes; Reflex=Yes; Person, Gender, Number; Case=Acc,Equ,All on case-marked personal pronouns.
• Adjectives: Gender, Number; NumType=Ord on ordinals.
• Passive and causative verbs are lemmatised by their derived stem (ተፈሊጡ / təfəlit’u → ተፈልጠ / tǝfǝlt’ǝ “be known”).

Syntax

• Subjects (nsubj) agree with the subject marking of the verb and are not case-marked. In the obligation construction with ኣሎ / ʔalo, the obliged person is the subject (nominative) but is indexed by an object suffix on ኣሎ.
• Objects (obj) are cross-referenced by an object suffix on the verb and take the objective prefix ን- / n- when definite. iobj is used for arguments marked with ን- and cross-referenced by an object or applicative suffix (recipients, causees, benefactives, malefactives, instruments). Arguments with other prepositions are obl.
• Experiencers of experience verbs and possessors of ኣሎ / ʔalo, cross-referenced by the object suffix, are obj:exp and obj:poss.
• Copula: the present copula is እዩ / ʔyu (negative ኣይኮነን / ʔaykonǝn), the past copula is ነበረ / nǝbǝrǝ; the copula is attached to the nonverbal predicate as cop. When both the subject and the predicate are nominals or headless relative clauses, the constituent preceding the copula is the predicate. The existential verb ኣሎ / ʔalo is not a copula. In clefts, the clefted element is the predicate and the relative clause is csubj:cleft.
• Converbs and subordinate verbs form adverbial clauses (advcl); relative clauses with ዝ- / z- are acl; headless relative clauses are annotated according to their function (csubj, ccomp).
• Universal subtypes used: nsubj:pass, csubj:pass, nsubj:outer, csubj:outer, det:poss, compound:lvc, compound:redup.
• Language-specific subtypes: xcomp-pred, ti-dep/csubj-cleft, ti-dep/obj-exp, ti-dep/obj-poss, ti-dep/advmod-aff, ti-dep/case-pre.

Treebanks

There is 1 Tigrinya UD treebank:

• Tigrinya-TiTB

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

.