home edit page issue tracker

UD Guidelines

• Basic principles
  • Tokenization and word segmentation
  • Morphology
  • Syntax
  • Enhanced dependencies
  • CoNLL-U format and its extensions
  • Typos and other errors in underlying text
• Annotation guidelines
  • Nominals
  • Simple clauses
  • Complex clauses
  • Comparative constructions – working group materials
  • Other constructions
• Documentation of tags, features and relations
  • POS tags (single document)
  • Features (single document)
    • Layered features
    • Features in data (list of all features and values used in treebanks, including those that are not defined by the universal guidelines)
  • Syntactic relations (single document)
    • Relations in data (list of all relation subtypes that are used in treebanks)
  • Conversion from other tagsets to UD tags and features
  • MISC attributes
• Incubator for Construction-Oriented Documentation (it will be moved here when it is mature enough)

This is the online documentation of UD guidelines v2 (launched 2016-12-01 with subsequent revisions). For change history, see Guidelines Changes.

Language-specific Guidelines

Afro-Asiatic

Cushitic

• Beja

Egyptian

• Coptic

Semitic

• Akkadian
• Amharic
• Ancient Hebrew
• Arabic
• Assyrian
• Hebrew
• Maltese
• South Levantine Arabic

West Chadic

• Zaar

Arawakan

• Apurina

Arawan

• Madi

Austro-Asiatic

• Vietnamese

Austronesian

Central Philippine

• Cebuano
• Tagalog

Javanese

• Javanese

Malayo-Sumbawan

• Indonesian

Basque

• Basque

Bororoan

• Bororo

Chukotko-Kamchatkan

• Chukchi

Code switching

• Frisian Dutch
• Maghrebi Arabic French
• Turkish German

Creole

• Naija

Dravidian

• Malayalam
• Tamil
• Telugu

Eskimo-Aleut

• Yupik

Indo-European

Albanian

• Albanian
• Gheg

Anatolian

• Hittite

Armenian

• Armenian
• Western Armenian

Baltic

• Latvian
• Lithuanian

Celtic

• Breton
• Irish
• Manx
• Old Irish
• Scottish Gaelic
• Welsh

Germanic

• Afrikaans
• Danish
• Dutch
• English
• Faroese
• German
• Gothic
• Icelandic
• Low Saxon
• Norwegian
• Swedish
• Swiss German

Greek

• Ancient Greek
• Greek

Indic

• Bengali
• Bhojpuri
• Hindi
• Kangri
• Marathi
• Odia
• Prakrit
• Punjabi
• Sanskrit
• Sinhala
• Urdu

Iranian

• Balochi
• Khunsari
• Kurmanji
• Nayini
• Persian
• Soi

Italic

• Latin
• Umbrian

Romance

• Catalan
• French
• Galician
• Italian
• Ligurian
• Neapolitan
• Old French
• Portuguese
• Romanian
• Spanish

Slavic

• Belarusian
• Bulgarian
• Croatian
• Czech
• Old Church Slavonic
• Old East Slavic
• Polish
• Pomak
• Russian
• Serbian
• Slovak
• Slovenian
• Ukrainian
• Upper Sorbian

Japanese

• Japanese
• Old Japanese

Korean

• Korean

Macro-Je

• Xavante

Mande

• Bambara

Mayan

• Kiche

Mongolic

• Buryat

Niger-Congo

Defoid

• Yoruba

Northern Atlantic

• Wolof

Northwest Caucasian

• Abaza

Pama-Nyungan

• Warlpiri

Sign Language

• Swedish Sign Language

Sino-Tibetan

• Cantonese
• Chinese
• Classical Chinese

Tai-Kadai

• Thai

Tungusic

• Xibe

Tupian

Maweti-Guarani

• Guajajara
• Guarani
• Kaapor
• Mbya Guarani
• Nheengatu
• Teko
• Tupinamba

Munduruku

• Munduruku

Ramarama

• Karo

Tupari

• Akuntsu
• Makurap

Turkic

Northeastern

• Old Turkish
• Yakut

Northwestern

• Kazakh
• Kyrgyz
• Tatar

Southeastern

• Uyghur

Southwestern

• Turkish

Uralic

Finnic

• Estonian
• Finnish
• Karelian
• Livvi

Mordvin

• Erzya
• Moksha

Permic

• Komi Permyak
• Komi Zyrian

Sami

• North Sami
• Skolt Sami

Ugric

• Hungarian

Uto-Aztecan

• Western Sierra Puebla Nahuatl

© 2014–2022 Universal Dependencies contributors. Site powered by Annodoc and brat

.