home edit page issue tracker

UD for Tagalog

Tokenization and Word Segmentation

• Words are generally delimited by whitespace. Exceptions:
  • compound words separated by a hyphen are tokenized independently (e.g. bukod + - + tangi) if each word can be taken as a separate semantic expression forming a whole
  • contractions are tokenized independently whenever possible (e.g. iba + ‘t)
  • occurrences of the linker -ng are annotated as multi-word tokens, (e.g. surface form: maraming, underlying words: marami + -ng)
• Words with spaces do not regularly occur in Tagalog.

Morphology

Tags

• Tagalog uses all the 17 universal POS tags.
  • The only word that is currently tagged as auxiliary (AUX) is the negative imperative (prohibitive) marker huwag. There are no copulas.
• Several word types are classified as particles (PART). These fall under two categories:
  • grammatical particles (the inversion marker ay; and relative or linking markers na and -ng)
  • lexical particles (the question marker ba, the negation marker hindi, fillers such as e, o, and a, and the honorific po)
• Quantifiers (including the pluralizer mga and indefinite quantifiers such as marami “many” and ilan “some”) are classified as determiners (DET). Possessive pronouns are classified under pronouns (PRON) since several identical word forms are also used to indicate the objective case.
• Like other Philippine-type languages, Tagalog employs nominal markers that have characteristics of both determiners (‘articles’) and prepositions (Himmelmann 2005, pages 145–147). Their functions include case marking, semantic role marking, topic and definiteness marking. There is no standard terminology in use for these words in the literature. Some authors classify them as prepositions (e.g., Schachter and Shopen 2007, page 35), some as articles or determiners (e.g., Dryer 2007, pages 94–95 and 121–122), and many authors avoid either of the terms and use the term ‘markers’ instead (e.g., Andrews 2007, page 203). Within the UD framework, we have to choose either the ADP tag or the DET tag and the corresponding dependency relations; the adposition analysis has been adopted. The markers include:
  • the topic nominative marker ang
  • the non-topic genitive marker ng
  • oblique markers such as dative sa, mula sa, para sa, sa pamamamgitan ng, etc.

Nominal Features

• Nouns (NOUN) and proper nouns (PROPN) do not have Number because plural, if overtly expressed, is signaled by the function word mga or by reduplication. In the former case, the pluralizel mga bears the feature Number=Plur. In case of reduplication, there are two tokens connected via compound:redup and neither bears a Number feature.
• However, the Number feature is marked for pronouns (PRON), with three possible values: Sing, Dual, Plur (the dual is used only with one pronoun that refers to the speaker and the listener).
• Some nouns have an inherent Gender with one of two values: Masc, Fem. This involves mostly person-denoting loanwords from Spanish and personal first names. Other nouns do not have the feature.
• Case has 3 primary values: Nom, Gen, Dat. The case names are not without controversy, as some authors would argue for an ergative analysis of the Philippine languages (while others would either use the labels adopted here, or use Acc instead of Gen, or argue that the Philippine type is neither ergative-absolutive nor nominative-accusative).
  • The nominative case corresponds to the subject of a clause regardless of voice (like in many European languages).
  • The genitive case is used for non-subject core participants in a clause, and for modifiers of other nominals.
  • The dative case is oblique and may encompass various locative and directional meanings.
  • In full noun phrases, Case is a lexical feature of the case-marking word or phrase, i.e., ADP. For case markers of obliques, more fine-grained values may be specified, e.g. Loc for locative sa, Ben for benefactive para sa, and Ins for instrumental sa pamamagitan ng. Note that there are also other prepositions which are not tagged with any Case feature, although they are attached to a noun via the case relation.
  • For pronouns (PRON), Case is their inflectional feature.

Degree and Polarity

• Degree applies to adjectives (ADJ) and has one of two possible values: Pos, Sup. Note that the comparative is not formed morphologically.
• Polarity has two values, Pos and Neg, and applies marginally to some existential verbs (VERB), as well as to the negation particle hindi and to the response interjections oo and hindi.

Verbal Features

• Verbs inflect for Mood, Aspect (rather than Tense) and Voice.
• The current data contain only verbs in the indicative mood (Ind).
• Aspect is perfective (Perf), imperfective (Imp), prospective (Prosp) and habitual (Hab).
• The Austronesian Voice system is quite different from the active-passive opposition in Indo-European languages. The morphological voice of the verb “focuses” on one particular argument, which is annotated as subject. However, the voices are symmetric in the sense that focusing on the more patient-like argument (“passive” voice) does not transform a transitive clause into intransitive: the more agent-like argument, if present, is still a core argument. Moreover, it is possible to focus on arguments with other semantic roles as well. (Note that the “focus” of the verb on a particular semantic role should not be confused with pragmatic focus as in topic-focus articulation. The focused argument here is actually always the topic, not the focus of the utterance.) The following Voice features are defined in the universal guidelines and apply to multiple Philippine languages. Not all verbs have forms for all the voices, though.
  • To reduce proliferation of feature values, the actor-focus voice uses the same label as the active voice in Indo-European languages, that is, Voice=Act.
  • Analogously, the undergoer-focus voice is conflated with Indo-European passive (Voice=Pass), although its grammatical behavior is different, as explained above.
  • The locative-focused voice (Voice=Lfoc) marks the location as the topic of the sentence.
  • The beneficiary-focused voice (Voice=Bfoc) marks the beneficiary as the topic of the sentence.

Pronouns, Determiners, Quantifiers

• PronType is used with pronouns (PRON), determiners (DET) and adverbs (ADV).
• NumType is used with cardinal numerals (NUM).
• The Reflex feature marks the reflexive pronoun sarili. It is always used together with PronType=Prs, although it does not bear the Person feature.
• Person is a lexical feature of personal pronouns (PRON) and has three values, 1, 2 and 3. Person is not marked on other types of pronouns and on nouns, although they can be almost always interpreted as the 3rd person.
• Clusivity is used with 1st person plural (and dual) pronouns, and distinguishes inclusive tayo “we” from exclusive kami “we”.
• Deixis is a lexical feature of demonstratives (PRON, ADV) and has three values, Prox, Med, Remt.

Other Features

• PartType distinguishes various Tagalog-specific particles (besides the negative particle hindi, which is tagged with Polarity=Neg):
  • Int … question particle ano (tag question), ba (yes-no question), kaya (speculative yes-no question). The particle ano is actually a conversion from the pronoun ano “what”, but as a tag question marker it no longer is a pronoun and should be tagged PART rather than PRON.
  • Des … particle sana “hopefully”
  • Nfh … non-first-hand particle daw “they say”
• Link=Yes is currently used in the TRG treebank to signal the linker suffix. In the Ugnayan treebank, linkers are treated as separate syntactic words and the surface word is treated as a multi-word token.

Syntax

Core Arguments and Adjuncts

• Tagalog uses the Austronesian voice system, a typologically unusual class of morphosyntactic alignment. In Tagalog, two core arguments (the topic argument and non-topic argument) are marked by prepositions, with their semantic roles determined by voice affixes on the verb. For the purposes of UD annotation, the following guidelines are used:
  • The topic/nominative argument (marked by the preposition ang) is attached to the predicate using the nsubj relation.
  • The non-topic/genitive argument (marked by the preposition ng) is attached to the predicate using the obj relation.
  • If there are two ng-arguments, the more agent-like argument is attached as obj and the more patient-like argument is attached as iobj.
  • Other nominals are attached to the predicate using the obl relation.

Non-verbal Predicates

• Non-verbal clauses (such as noun, adverb, or adjective phrases) may be used as predicates in Tagalog. In these instances, the highest node of the clause is labeled as root, and all other nodes typically connected to the root verb are instead connected to this node. There is no copula.

Relations Overview

• The following relation subtypes are used in Tagalog:
  • nsubj:pass for nominal subjects of verbs in patient-focused voice
  • nsubj:lfoc for nominal subjects of verbs in location-focused voice
  • nsubj:bfoc for nominal subjects of verbs in beneficiary-focused voice
  • csubj:pass for clausal subjects of verbs in patient-focused voice
  • csubj:lfoc for clausal subjects of verbs in location-focused voice
  • csubj:bfoc for clausal subjects of verbs in beneficiary-focused voice
  • obj:agent for agents (actors) of verbs in a non-actor-focused voice
  • iobj:patient for patients (undergoers) of verbs in a voice where neither agent nor patient is the subject
  • compound:redup for reduplicated nouns and adjectives (reduplication intensifies their meaning)
  • nmod:poss for possessive relation between nominals
  • acl:relcl for relative adnominal clauses
• The following relation types are not used in Tagalog at all: clf, expl, dislocated

Treebanks

There are 2 Tagalog UD treebanks:

• Tagalog-TRG
• Tagalog-Ugnayan

References

• Avery D. Andrews (2007). The major functions of the noun phrase. In Timothy Shopen (ed.): Language Typology and Syntactic Description vol. I: Clause Structure. Cambridge University Press, Cambridge, UK.
• Matthew S. Dryer (2007). Word order. In Timothy Shopen (ed.): Language Typology and Syntactic Description vol. I: Clause Structure. Cambridge University Press, Cambridge, UK.
• Nikolaus P. Himmelmann (2005). Typological characteristics. In Alexander Adelaar and Nikolaus P. Himmelmann (eds.): The Austronesian Languages of Asia and Madagascar. Routledge, London and New York.
• Paul Schachter and Timothy Shopen (2007). Parts-of-speech systems. In Timothy Shopen (ed.): Language Typology and Syntactic Description vol. I: Clause Structure. Cambridge University Press, Cambridge, UK.

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

.