--- title: "VERB: verb — definition and key papers — Elamur Research" description: "VERB is the Universal Dependencies part-of-speech tag for content words that typically name events, actions, processes or states." canonical: "https://research.elamur.ai/free/en/concepts/upos-verb/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/concepts/upos-verb/) [home](https://md.research.elamur.ai/free/en/) / [concepts](https://md.research.elamur.ai/free/en/concepts/) / VERB: verb # 🏷️ VERB: verb Universal Dependencies · 7 key papers · [all concepts](https://md.research.elamur.ai/free/en/concepts/) Part of: [🗨️ Natural language processing](https://md.research.elamur.ai/free/en/concepts/natural-language-processing/) · [🌐 Universal Dependencies](https://md.research.elamur.ai/free/en/concepts/universal-dependencies/) VERB is the Universal Dependencies part-of-speech tag for content words that typically name events, actions, processes or states. Such words head clauses as predicates and inflect for categories such as tense, aspect, mood, person and voice. Auxiliaries and copulas are tagged separately. Also: verb, UPOS VERB, V ## What it is In [Universal Dependencies](https://md.research.elamur.ai/free/en/concepts/universal-dependencies/) (UD), VERB is one of the universal part-of-speech tags assigned during [part-of-speech tagging](https://md.research.elamur.ai/free/en/concepts/part-of-speech-tagging/). A word gets VERB when it is a lexical verb: it carries the meaning of an event or state and usually acts as the predicate of a clause. In *The dog barked*, the word *barked* is a VERB. UD's theory rests on three kinds of units. Nominals denote entities, clauses denote events, and modifiers specify them ([Marneffe et al. 2021](https://md.research.elamur.ai/free/en/papers/marneffe-2021-universal-dependencies/)). The verb is the typical core of a clause. Because UD makes content words the heads of the tree, the lexical verb governs its subject, its objects and the function words around it ([Marneffe et al. 2021](https://md.research.elamur.ai/free/en/papers/marneffe-2021-universal-dependencies/)). The main borderline is with AUX. Tense, aspect and modal auxiliaries are tagged AUX, and so are copulas. In *has eaten*, only *eaten* is a VERB. Other boundaries: - Participles used as plain modifiers (*a broken window*) are often tagged ADJ. - Verbal nouns and nominalized gerunds (*the singing was loud*) are tagged NOUN. - Light-verb and phrasal-verb combinations keep VERB on the verb. The particle or noun gets its own tag (see [idioms & multiword expressions](https://md.research.elamur.ai/free/en/topics/multiword-expressions/)). The tag is older than UD. Petrov et al. 2012 defined a compact set of twelve coarse categories, and VERB was one of them ([Petrov et al. 2012](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/)). They built the categories pragmatically from existing corpus guidelines. One of their contentious decisions put Korean stative verbs under adjectives ([Petrov et al. 2012](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/)). That set had no separate auxiliary tag. UD v1 extended it to 17 tags, adding AUX and PROPN, which narrowed VERB to lexical verbs ([Nivre et al. 2016](https://md.research.elamur.ai/free/en/papers/nivre-2016-universal-dependencies-v1-multilingual-treebank/)). The list stays closed at 17 in v2 ([Nivre et al. 2020](https://md.research.elamur.ai/free/en/papers/nivre-2020-universal-dependencies-v2-evergrowing-multilingual/)). ## Examples | Language | Sentence | VERB | Not VERB | | --- | --- | --- | --- | | English | *She has written letters.* | written | has (AUX) | | French | *Il veut partir.* | veut, partir | — | | German | *Er ruft sie an.* | ruft | an (ADP or PART, by treebank) | | Spanish | *Está lloviendo.* | lloviendo | está (AUX) | Typical features (FEATS) on a VERB token: - VerbForm: Fin, Inf, Part, Ger, Conv - Mood, Tense, Aspect and Voice - Person and Number for agreement - Polarity - Evident, in languages that mark evidentiality Several of these names changed in v2. Negative became Polarity, and the value Trans became Conv ([Nivre et al. 2020](https://md.research.elamur.ai/free/en/papers/nivre-2020-universal-dependencies-v2-evergrowing-multilingual/)). The feature inventory itself descends from Interset, a universal feature interlingua used to convert between tagsets ([Zeman 2008](https://md.research.elamur.ai/free/en/papers/zeman-2008-reusable-tagset-conversion-tagset-drivers/); [Nivre et al. 2016](https://md.research.elamur.ai/free/en/papers/nivre-2016-universal-dependencies-v1-multilingual-treebank/)). A verb's [lemma](https://md.research.elamur.ai/free/en/concepts/lemma/) is its citation form, such as the infinitive *write* for *written*. ## Across languages [Morphology](https://md.research.elamur.ai/free/en/concepts/morphology/) decides how much a VERB carries. In Turkish or Finnish, a single verb form may encode tense, mood, person, negation and voice together. In Mandarin, the verb is largely uninflected, and aspect is often marked by separate particles. Russian verbs mark Aspect lexically, with perfective and imperfective pairs. The line against ADJ also moves between languages. In Korean and Japanese, predicative adjective-like words inflect much like verbs. The coarse tagset forced a choice here ([Petrov et al. 2012](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/)). Copula behaviour varies too. Some languages omit the copula, and some use verbs that UD treats as AUX only in a copular role. UD always tags the syntactic word, not the morpheme. Fused forms such as French *du* or Spanish *dámelo* are split before tagging, while the original surface token is kept ([Nivre et al. 2016](https://md.research.elamur.ai/free/en/papers/nivre-2016-universal-dependencies-v1-multilingual-treebank/); [Marneffe et al. 2021](https://md.research.elamur.ai/free/en/papers/marneffe-2021-universal-dependencies/)). ## In treebanks and research In UD [treebanks](https://md.research.elamur.ai/free/en/topics/treebanks/) of English and French, roughly a tenth of all tokens are VERB. By far their most common relation is root, since the main predicate heads the sentence. In both languages acl, advcl, xcomp and conj follow. English verbs attach more often as xcomp, and French verbs more often as acl and advcl. As heads, verbs take the core arguments nsubj, obj and iobj, as well as obl, advmod, aux and mark. This makes them the anchor of predicate–argument structure, which matters for [dependency parsing](https://md.research.elamur.ai/free/en/concepts/dependency-parsing/) and [semantic role labeling](https://md.research.elamur.ai/free/en/concepts/semantic-role-labeling/) ([Marneffe et al. 2021](https://md.research.elamur.ai/free/en/papers/marneffe-2021-universal-dependencies/)). Coarse tags like VERB proved useful early. Delexicalized parsers trained on the twelve universal tags transferred directly from English to other languages and beat the unsupervised methods of the time ([Petrov et al. 2012](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/)). Tagging the VERB/AUX boundary consistently remains hard, so one proposed check compares UPOS trigram distributions between treebanks of the same language ([Aggarwal & Zeman 2020](https://md.research.elamur.ai/free/en/papers/aggarwal-2020-estimating-pos-annotation-consistency-different/)): ``` θpos(A, B) = KLcpos3(A, B) + KLcpos3(B, A) ``` Here A and B are two treebanks. KLcpos3 is an asymmetric Kullback–Leibler divergence over UPOS trigram frequencies. Adding both directions makes the measure symmetric ([Aggarwal & Zeman 2020](https://md.research.elamur.ai/free/en/papers/aggarwal-2020-estimating-pos-annotation-consistency-different/)). Broader consistency audits of UD corpora address the same problem ([Marneffe et al. 2017](https://md.research.elamur.ai/free/en/papers/marneffe-2017-assessing-annotation-consistency-universal-dependencies/)). How often the tag VERB occurs: its share of all tokens Counted by code over the Universal Dependencies 2.18 treebanks of English and French (every token once, subtypes with their base relation). **The numbers** | Treebanks | Share | | --- | --- | | English treebanks | 11.43% of tokens (128,753 tokens) | | French treebanks | 10.61% of tokens (115,700 tokens) | The relations VERB words take most often The five most frequent in each language's treebanks, by tokens. **The numbers** | Pair | Tokens | | --- | --- | | English: root | 60,308 | | English: acl | 14,552 | | English: xcomp | 12,763 | | English: advcl | 12,718 | | English: conj | 11,842 | | French: root | 47,471 | | French: acl | 21,801 | | French: advcl | 13,905 | | French: xcomp | 11,099 | | French: conj | 10,197 | ## Key papers The works this page cites, the most cited first; citations as counted by OpenAlex plus the papers of this library that cite the work, read 2026-10-24. - [Universal Dependencies v1: A Multilingual Treebank Collection](https://md.research.elamur.ai/free/en/papers/nivre-2016-universal-dependencies-v1-multilingual-treebank/) — Joakim Nivre, Marie-Catherine de Marneffe, Filip Ginter et al. · 2016 · 1,741 citations - [A Universal Part-of-Speech Tagset](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/) — Slav Petrov, Dipanjan Das, Ryan McDonald · 2012 · 1,180 citations - [Universal Dependencies](https://md.research.elamur.ai/free/en/papers/marneffe-2021-universal-dependencies/) — Marie-Catherine de Marneffe, Christopher D. Manning, Joakim Nivre et al. · 2021 · 480 citations - [Universal Dependencies v2: An Evergrowing Multilingual Treebank Collection](https://md.research.elamur.ai/free/en/papers/nivre-2020-universal-dependencies-v2-evergrowing-multilingual/) — Joakim Nivre, Marie-Catherine de Marneffe, Filip Ginter et al. · 2020 · 325 citations - [Reusable Tagset Conversion Using Tagset Drivers](https://md.research.elamur.ai/free/en/papers/zeman-2008-reusable-tagset-conversion-tagset-drivers/) — Daniel Zeman · 2008 · 181 citations - [Assessing the Annotation Consistency of the Universal Dependencies Corpora](https://md.research.elamur.ai/free/en/papers/marneffe-2017-assessing-annotation-consistency-universal-dependencies/) — Marie-Catherine de Marneffe, Matias Grioni, Jenna Kanerva et al. · 2017 · 30 citations - [Estimating POS Annotation Consistency of Different Treebanks in a Language](https://md.research.elamur.ai/free/en/papers/aggarwal-2020-estimating-pos-annotation-consistency-different/) — Akshay Aggarwal, Daniel Zeman · 2020 · 7 citations ## Related concepts - [Universal Dependencies](https://md.research.elamur.ai/free/en/concepts/universal-dependencies/) — Universal Dependencies is a framework and an open collection of treebanks that annotates the grammar of many languages with one shared… - [Part-of-speech tagging](https://md.research.elamur.ai/free/en/concepts/part-of-speech-tagging/) — Part-of-speech tagging is the task of assigning each token in a text a grammatical category, such as noun, verb, adjective or determiner… - [Morphology](https://md.research.elamur.ai/free/en/concepts/morphology/) — Morphology is the branch of linguistics that studies the internal structure of words. It covers how words are built from meaningful units… ## More papers Topics: [🔤 Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/) · [🌐 Universal Dependencies](https://md.research.elamur.ai/free/en/topics/universal-dependencies/) Shelves: [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) - [← previous concept Universal Dependencies](https://md.research.elamur.ai/free/en/concepts/universal-dependencies/) - [next concept → vocative: vocative](https://md.research.elamur.ai/free/en/concepts/ud-vocative/) - [↑ back to concepts](https://md.research.elamur.ai/free/en/concepts/)