home edit page issue tracker

UD for Huave

The variety documented here is the Huave of San Mateo del Mar. The first treebank consists of Huave sentences (and their Spanish translations) from the SIL-produced Diccionario Huave de San Mateo del Mar, compiled by Glenn Albert Stairs Kreger and Emily Florence Scharfe de Stairs. The text is written in the newer orthography typical of younger speakers (k, kw, g; see the note on spelling below). The original SIL spelling of the source text is kept in the text[orig] comment.

Tokenisation and Word Segmentation

• Words are delimited by whitespace. Punctuation marks, including the inverted question and exclamation marks (¿ ¡), are separate tokens and are marked SpaceAfter=No when no space follows.
• Several short function words are written together with the following/preceding word. These are split into multiword tokens (the surface form is kept on the range line, and the parts are numbered separately). These include:
  • The preposition tiül + a noun, written as one word with the preposition reduced to ti-: tikambaj = ti (lemma tiül, ADP) + kambaj (NOUN) “in the town”; also tikorrül, timanchiük, timüx, tixiül, titiiüd, tindek.
  • The locative copula al + tiül: altiül = al (AUX) + tiül (ADP), “there is / has”; also altikambaj = al + ti + kambaj.
  • The aspectual proclitic la (punctual, “already”) + the next word: landoj = la + ndoj, lango = la + ngo.
  • Final clitics -e and -a, tagged PART: ngome = ngom + e, ngana = ngan + a, leawa = leaw + a.

• The annotation files also carry Error=TokenLevel in MISC on many of these tokens.

• Spanish loanwords are tokenised like native words. See the note on spelling below.

Spelling. The texts use the orthography of the recent Huave materials (k and kw instead of c, qu and cu; g instead of gu before e and i). Unadapted Spanish words may keep Spanish spelling (e.g., pomada and calle).

Morphology

Tags

• Unused tag: SYM. All other universal tags occur. X and INTJ are relatively rare.
• NOUN, PROPN. Possessed nouns are tagged NOUN and lemmatised as the bare root (xekit → kit, xemüm → müm). Deverbal nouns such as nendok “fisherman” are NOUN with VerbForm=Vnoun. Personal names are PROPN (with Gender).
• VERB. Verbs inflect by prefixes and suffixes for person, number, tense, aspect and mood. The lemma is the bare root without those (ajlüy → jlüy, akwiür → kwiür).
• AUX. The locative copula al (“be located, there is”; Herrera 2016 glosses it COP.LOC) is AUX, as is the aspectual proclitic la.
• PART. Interrogative clitics -e and -a (ngome, ngana), with PronType=Int, and the evidential chük (Evident=Nfh, one token in these data; it is very frequent in the story collections).
• DET and PRON. The definite article a (variants aaga, aag) is DET and precedes the noun. Herrera 2016 analyses it as the base of the demonstratives (pronoun, adnominal, locative adverb and manner adverb are all built on it plus a deictic element). When aag stands alone it is PRON. Personal pronouns (nej 3rd person, xik 1st person singular, ik 2nd person singular, xikon / xikona 1st person plural exclusive) are PRON with PronType=Prs. Possessive pronouns (saneay “mine”, okweaj) are PRON.
• NUM. The numeral no- “one” with a classifier suffix (nop, noik) is NUM with NumType=Clf, and depends on the noun as nummod. Herrera 2016 (ch. 6) argues that it also works as an indefinite article, but it seems not to have been tagged as DET.
• ADP. tiül “in, to, with” and similar words are ADP. They precede their noun and depend on it as case.
• SCONJ and CCONJ. pero / pere is CCONJ. Subordinators are SCONJ (wüx “when”, kos “because”, sitiül “if”).
• ADJ and ADV. Words that modify nouns are ADJ and precede them (nine nench “the small child”). Manner and degree words are ADV.

Features

• Nouns. Possession is marked by a prefix on the noun. Person and number of the possessor are Person[psor] and Number[psor] (xe- and xi- for first person singular; ow- and ap- for third person). The prefix mi- occurs on nouns whose possessor is a separate noun or pronoun (mikwal nej “his/her child”); it carries no feature. Plural is rarely marked on nouns in the data (3 tokens). Number (mostly Sing, also Plur and Dual) is used on verbs and pronouns.
• Verbs. Subject person and number: Person, Number (also Number=Dual, and Clusivity=Ex|In for first person plural and dual). Tense=Pres|Past, Aspect=Prog (the progressive te(a)-) and Aspect=Punt (la “already”), Mood=Sub and Mood=Irr. VerbForm=Fin|Vnoun. Voice=Caus for causatives (3 tokens). Subcat=Tran|Intr marks transitivity.
• Pronouns. PronType=Prs|Dem|Int, Person, Number, Clusivity.
• Numerals. NumType=Clf for classifier numerals.
• Definiteness. The feature Definite is not used. Indefinite and free-choice pronouns (the al- series: aleaik, aljane; the nej series: jangantanej, kwajantanej) are marked PronType=Ind, and negative pronouns (the ni…ind series: nikwajind) PronType=Neg (Herrera 2016, ch. 6).

Syntax

• Core arguments. nsubj, obj and iobj are used. Word order is relatively free (Herrera 2016: ch. 1). In the annotated clauses that have both a subject and an object, SVO is the most common order, VOS the next. Objects follow the verb in most clauses; subjects are about equally often before and after it.
• Within the noun phrase. Determiners, numerals, adjectives, prepositions and subordinators precede their head (det, nummod, amod, case, mark). Possessors and noun modifiers follow it (nmod, 101 of 105). A pronominal possessor after a possessed noun (mikwal nej) is nmod.
• Predicates. Most roots are verbs. Nonverbal predicates (nouns, adjectives, pronouns, adverbs) are roots with no copula, as in Aag agüy saneay “this is mine”. The locative copula al is cop when it takes a nominal complement (altiül “have, there is”).
• Subordinate and complement clauses. Clauses introduced by wüx and kos are advcl with a mark dependent. The modal ndom “can” and the negator ngo are aux of the following verb when they are tagged AUX (21 aux relations, mostly ndom, ngo and ombas); that verb is then often an xcomp of the main verb. Other verbs also take an xcomp. Clausal arguments are ccomp and csubj. A clause modifying a noun is acl.
• Parataxis. Juxtaposed clauses are parataxis, including clauses with a subjunctive verb (Mood=Sub) after a comma.
• Other relations. dislocated (a topicalised noun before the clause), compound (mostly noun + noun, e.g. vida + müm; also reduplicated numerals: nop nop “one by one”), fixed, flat, goeswith and discourse are rare.
• Subtype relations. nmod:poss is the only subtype used.

Treebanks

There is 1 Huave UD treebank:

• Huave-Mesotree

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

.