UD for Nheengatu

Tokenization and Word Segmentation

• Words are generally delimited by whitespace or punctuation.

• According to the orthographic conventions adopted, a hyphen separates the syntactic words of most multiword tokens. For example, these are cases of incorporation of an auxiliary verb into the main verb (e.g., putari ‘to want’ in ayuíri-putari ‘I want to return’). However, the component words of a group of multiword tokens are not hyphen-separated, e.g., paraname, which derives from the cliticization of the postposition upé ‘on’ to paranã ‘river’.

Morphology

Tags

• The Nheengatu treebank presently employs 15 of the 17 universal POS tags, namely, ADJ, ADP, ADV, AUX, CCONJ, DET, INTJ, NOUN, NUM, PART, PRON, PROPN, PUNCT, SCONJ, and VERB. X and SYM have not come to use yet, since they didn’t occur in the sentences hitherto annotated.

• There is just one copula, ikú, which also functions as an auxiliary expressing progressive aspect.

• Nheengatu has a rich system of particles, which convey a variety of grammatical notions, e.g., negation, tense, and aspect.

• There are two classes of personal pronouns in Nheengatu, both tagged PRON. So-called first-class pronouns are free morphemes, while second-class pronouns are subject-agreement prefixes that attach to stative verbs, functioning as number-person agreement markers, or realizing the internal argument of nouns (genitive complement) or postpositions. According to the adopted orthographical conventions, these prefixes are written as independent tokens. Presently, these forms constitute independent words in our treebank. In the future, we may treat these prefixes as part of the verb, despite orthography.

• Determiners and (non-personal) pronouns seem to only differ in syntactic behavior, i.e., whether they modify a noun or function as the head of the NP.

• The base form, i.e., the non-inflected, infinitive form of certain verbs (e.g., kitika ‘to grate’) can function as ADJ (and possibly also as ADV or NOUN), but this was not attested in the sentences hitherto annotated.

Features

• Verbs inflect for the six logically possible combinations of the values 1, 2, and 3 for the person feature and the values singular and plural for the number feature. In conservative or historical varieties of Nheengatu, the number distinction is neutralized in Person 3 of the active series. In innovative varieties, the opposition between singular and plural seems to be emerging due to the grammaticalization of the person 3 plural pronoun as an inflectional prefix, see Avila (2021) and Cruz (2011, 2015).

• Verbs fall into one of two inflectional paradigms. Person-number of transitives and non-stative intransitives is marked with prefixes of the active series. Stative intransitives receive prefixes of the inactive series, also called second-class pronouns.

• Mood, tense, and aspect is expressed through auxiliaries or particles.

• Gender does not play any morpho-syntactic role in Nheengatu.

• Nouns and pronouns only inflect for number. There is no number agreement proper in the NP domain. The plural is marked with the suffix itá, which attaches either to the determiner or the noun (it is uncertain to us whether a double marking is possible). Nouns under the scope of a semantically plural quantifier (e.g., musapiri ‘three’) need not receive the plural suffix.

• The treebank employs the layered features Number[psor]=Sing, Person[psor]=3, and Person[grnd]. The last one encodes the person of the internal argument of a postpostion, i.e., the reference object, relatum or ground (also landmark in cognitive grammar), which is realized in Nheengatu in certain circumstances by the non-contiguity prefix s-. The first two layered features encode the number and person of the complement of a noun when it is realized by the non-adjacency relational prefix (e.g., suka ‘her/his/its house’).

• Some noun classes signal syntactic adjacency or non-adjacency between the head noun and its dependent employing relational prefixes. Relational prefixes also occur with subclasses of postpositions and inactive verbs.

Syntax

• The basic word order is SVO.

• Nheengatu has postpositions. Analogously, native subordinating conjunctions and relative pronouns are left adjacent to the verb, i.e., postverbal. By contrast, prepositions or conjunctions borrowed from Portuguese are preverbal. Some native conjunctions show a similar behavior.

• In the genitive construction, the possessor precedes the head (possessum) noun, which receives the relational adjacency prefix if it belongs to certain noun paradigms.

• There are two auxiliaries for the ingressive and the progressive aspect, namely sú and ikú, respectively. The former precedes, while the latter follows the main verb. Both auxiliaries inflect for person and number, as the main verb does.

• The base (non-finite) form of auxiliaries expressing modality, e.g., putari ‘to want’, left-adjoins the main verb, which inflects for person and number, e.g.ayuíri-putari ‘I want to return’.

• Individual-level predicates head copulaless sentences, while stage-level predicates require the copula. In both situations, an adjective or a postpositional phrase, inter alia, conveys the sentence predication.

• Subjects trigger agreement with the verb. Objects are either noun phrases (or pronouns) or postpositional phrases that realize a core argument. Ditransitive verbs govern a dative complement realized by a postpositional phrase.

• Subtype relations used so far: acl:relcl, acl:advcl, and nmod:poss.

Treebanks

There is one Nheengatu UD treebank:

* [UD_Nheengatu-CompLin](../treebanks/yrl_complin/index.html)

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