Welcome to VerbNet!

VerbNet (VN) (Kipper-Schuler 2006) is the largest on-line verb lexicon currently available for English. It is a hierarchical domain-independent, broad-coverage verb lexicon with mappings to other lexical resources such as WordNet (Miller, 1990; Fellbaum, 1998), Xtag (XTAG Research Group, 2001), and FrameNet (Baker et al., 1998). VerbNet is organized into verb classes extending Levin (1993) classes through refinement and addition of subclasses to achieve syntactic and semantic coherence among members of a class. Each verb class in VN is completely described by thematic roles, selectional restrictions on the arguments, and frames consisting of a syntactic description and semantic predicates with a temporal function, in a manner similar to the event decomposition of Moens and Steedman (1988).

Acknowledgments

We gratefully acknowledge the support of the National Science Foundation Grant NSF-IIS-1116782, A Bayesian Approach to Dynamic Lexical Resources for Flexible Language Processing.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Check us out on GitHub!

You can find the most up-to-date information on, and recent changes to, VerbNet and its related projects on our GitHub page! Learn more...

What is VerbNet?

VerbNet (VN) (Kipper-Schuler 2006) is the largest on-line verb lexicon currently available for English. It is a hierarchical domain-independent, broad-coverage verb lexicon with mappings to other lexical resources such as WordNet (Miller, 1990; Fellbaum, 1998), Xtag (XTAG Research Group, 2001), and FrameNet (Baker et al., 1998). VerbNet is organized into verb classes extending Levin (1993) classes through refinement and addition of subclasses to achieve syntactic and semantic coherence among members of a class. Each verb class in VN is completely described by thematic roles, selectional restrictions on the arguments, and frames consisting of a syntactic description and semantic predicates with a temporal function, in a manner similar to the event decomposition of Moens and Steedman (1988).

(return to top)

The VerbNet Approach

VerbNet groups together verbs with identical sets of syntactic frames and semantic predicate structures, as in the example below for class CUT-21.1-1:

These classes may inherit frames from a parent class, resulting in a hierarchical structure:

The classes in VerbNet are based on an extension of Levin (1993) classes, with the understanding that the syntactic form of a verb and its arguments informs its semantics. Observe, for example, the differences and similarities between the following two classes, CUT-21.1 and BREAK-45.1:

(return to top)

How It All Works

Each VN class contains a set of syntactic descriptions, or syntactic frames, depicting the possible surface realizations of the argument structure for constructions such as transitive, intransitive, prepositional phrases, resultatives, and a large set of diathesis alternations. Semantic restrictions (such as animate, human, organization) are used to constrain the types of thematic roles allowed by the arguments, and further restrictions may be imposed to indicate the syntactic nature of the constituent likely to be associated with the thematic role. Syntactic frames may also be constrained in terms of which prepositions are allowed. Each frame is associated with explicit semantic information, expressed as a conjunction of boolean semantic predicates such as `motion,' `contact,' or `cause.' Each semantic predicate is associated with an event variable E that allows predicates to specify when in the event the predicate is true (start(E) for preparatory stage, during(E) for the culmination stage, and end(E) for the consequent stage). Figure 1. shows a complete entry for a frame in VerbNet class Hit-18.1.

Figure 1: Simplified VerbNet entry for Hit-18.1 class
Class Hit-18.1
Roles and Restrictions: Agent[+int_control] Patient[+concrete] Instrument[+concrete]
Members: bang, bash, hit, kick, ...
Frames:
Name Example Syntax Semantics
Basic Transitive Paula hit the ball Agent V Patient cause(Agent, E)manner(during(E), directedmotion, Agent) !contact(during(E), Agent, Patient) manner(end(E),forceful, Agent) contact(end(E), Agent, Patient)

VerbNet has recently been integrated with 57 new classes from Korhonen and Briscoe's (2004) (K&B) proposed extension to Levin's original classification (Kipper et al., 2006). This work has involved associating detailed syntactic-semantic descriptions to the K&B classes, as well as organizing them appropriately into the existing VN taxonomy. An additional set of 53 new classes from Korhonen and Ryant (2005) (K&R) have also been incorporated into VN. The outcome is a freely available resource which constitutes the most comprehensive and versatile Levin-style verb classification for English. After the two extensions VN has now also increased our coverage of PropBank tokens (Palmer et. al., 2005) from 78.45% to 90.86%, making feasible the creation of a substantial training corpus annotated with VN thematic role labels and class membership assignments, to be released in 2007. This will finally enable large-scale experimentation on the utility of syntax-based classes for improving the performance of syntactic parsers and semantic role labelers on new domains.

Integrating the two recent extensions to Levin classes into VerbNet was an important step in order to address a major limitation of Levin's verb classification, namely the fact that verbs taking ADJP, ADVP, predicative, control and sentential complements were not included or addressed in depth in that work. This limitation excludes many verbs that are highly frequent in language. A summary of how this integration affected VN and the result of the extended VN is shown in Table 1. The figures show that our work enriched and expanded VN considerably. The number of first-level classes grew significantly (from 191 to 274), there was also a significant increase in the number of verb senses and lemmas, along with the set of semantic predicates and the syntactic restrictions on sentential complements.

Table 1: Summary of the Lexicon's Extension
  Original VN Extended VN
First-level classes 191 274
Thematic roles 21 23
Semantic predicates 64 94
Syntactic restrictions (on sentential compl) 3 55
Number of verb senses 4656 5257
Number of lemmas 3445 3769
Table 2: Thematic roles and example classes that use them
Actor: used for some communication classes (e.g., Chitchat-37.6, Marry-36.2, Meet-36.2) when both arguments can be considered symmetrical (pseudo-agents).
Agent: generally a human or an animate subject. Used mostly as a volitional agent, but also used in VerbNet for internally controlled subjects such as forces and machines.
Asset: used for the Sum of Money Alternation, present in classes such as Build-26.1, Get-13.5.1, and Obtain-13.5.2 with `currency' as a selectional restriction.
Attribute: attribute of Patient/Theme refers to a quality of something that is being changed, as in (The price)att of oil soared. At the moment, we have only one class using this role Calibratable cos-45.6 to capture the Possessor Subject Possessor-Attribute Factoring Alternation. The selectional restriction `scalar' (defined as a quantity, such as mass, length, time, or temperature, which is completely specified by a number on an appropriate scale) ensures the nature of Attribute.
Beneficiary: the entity that benefits from some action. Used by such classes asBuild-26.1, Get-13.5.1, Performance-26.7, Preparing-26.3, and Steal-10.5. Generally introduced by the preposition `for', or double object variant in the benefactive alternation.
Cause: used mostly by classes involving Psychological Verbs and Verbs Involving the Body.
Location, Destination, Source: used for spatial locations.
Destination: end point of the motion, or direction towards which the motion is directed. Used with a `to' prepositional phrase by classes of change of location, such as Banish-10.2, and Verbs of Sending and Carrying. Also used as location direct objects in classes where the concept of destination is implicit (and location could not be Source), such as Butter-9.9, or Image impression-25.1.
Source: start point of the motion. Usually introduced by a source prepositional phrase (mostly headed by `from' or `out of'). It is also used as a direct object in such classes as Clear-10.3, Leave-51.2, and Wipe instr-10.4.2.
Location: underspecified destination, source, or place, in general introduced by a locative or path prepositional phrase.
Experiencer: used for a participant that is aware or experiencing something. In VerbNet it is used by classes involving Psychological Verbs, Verbs of Perception, Touch, and Verbs Involving the Body.
Extent: used only in the Calibratable-45.6 class, to specify the range or degree of change, as in The price of oil soared (10%)ext. This role may be added to other classes.
Instrument: used for objects (or forces) that come in contact with an object and cause some change in them. Generally introduced by a `with' prepositional phrase. Also used as a subject in the Instrument Subject Alternation and as a direct object in the Poke-19 class for the Through/With Alternation and in the Hit-18.1 class for the With/Against Alternation.
Material and Product: used in the Build and Grow classes to capture the key semantic components of the arguments. Used by classes from Verbs of Creation and Transformation that allow for the Material/Product Alternation.
Material: start point of transformation.
Product: end result of transformation.
Patient: used for participants that are undergoing a process or that have been affected in some way. Verbs that explicitly (or implicitly) express changes of state have Patient as their usual direct object. We also use Patient1 and Patient2 for some classes of Verbs of Combining and Attaching and Verbs of Separating and Disassembling, where there are two roles that undergo some change with no clear distinction between them.
Predicate: used for classes with a predicative complement.
Recipient: target of the transfer. Used by some classes of Verbs of Change of Possession, Verbs of Communication, and Verbs Involving the Body. The selection restrictions on this role always allow for animate and sometimes for organization recipients.
Stimulus: used by Verbs of Perception for events or objects that elicit some response from an xperiencer. This role usually imposes no restrictions.
Theme: used for participants in a location or undergoing a change of location. Also, Theme1 and Theme2 are used for a few classes where there seems to be no distinction between the arguments, such as Differ-23.4 and Exchange-13.6 classes.
Time: class-specific role, used in Begin-55.1 class to express time.
Topic: topic of communication verbs to handle theme/topic of the conversation or transfer of message. In some cases, like the verbs in the Say-37.7 class, it would seem better to have `Message' instead of `Topic', but we decided not to proliferate the number of roles.

Each verb argument is assigned one (usually unique) thematic role within the class. A few exceptions to this uniqueness are classes which contain verbs with symmetrical arguments, such as Chitchat-37.6 class, or the ContiguousLocation-47.8 class. These classes have indexed roles such as Actor1 and Actor2, as explained above.

(return to top)

Who Are We?

The VerbNet project is made possible by a highly talented team of researchers and students at the University of Colorado representing diverse domains of interest, expertise, and experience. Get to know them or get in touch by reading their introductions below!

Martha Palmer

Professor / Program Director

Martha Palmer is a Professor of Linguistics and Computer Science and an Institute of Cognitive Science Faculty Fellow. Her PhD is in Artificial Intelligence from the University of Edinburgh. She is an Association of Computational Linguistics (ACL) Fellow, and has won an Outstanding Graduate Advisor 2014 Award, a Boulder Faculty Assembly 2010 Research Award and was the Director of the 2011 Linguistics Institute in Boulder. Her research is focused on capturing elements of the meanings of words that can comprise automatic representations of complex sentences and documents. Supervised machine learning techniques rely on vast amounts of annotated training data so she and her students are engaged in providing data with word sense tags and semantic role labels for English, Chinese, Arabic, Hindi, and Urdu, funded by DARPA and NSF. They also train automatic sense taggers and semantic role labelers. A more recent focus is the application of these methods to biomedical journal articles and clinical notes, funded by NIH. She co-edits LiLT, Linguistic Issues in Language Technology, and has been a co-editor of the Journal of Natural Language Engineering and on the CLJ Editorial Board. She is a past President of ACL, and past Chair of SIGLEX and SIGHAN.

Susan Brown

Instructor / Research Associate

Susan Windisch Brown is the Associate Director of CLASIC, the professional M.S. in computational linguistics, and also serves as an instructor and research associate in the Computer Science and Linguistics Departments. She received her Ph.D. in Cognitive Science and in Linguistics from the University of Colorado in 2010. Her research is in natural language processing, especially computational semantics and ontology development. Other than a 2-year hiatus, she has devoted some of her time to expanding and revising VerbNet since 2007.

Benjamin Rohrs

Instructor / Research Associate

Dissertation: Vagueness and Propositional Content

Defended July 11, 2016 

AOS: Philosophy of Language, Metaphysics 

AOC: Philosophical Logic (classical first-order, non-classical first-order, modal), Epistemology 

Teaching Experience: Intro to Phil., Phil. and Society, Phil. Science, Feminist Phil., Phil. Race, Ancient, Modern 

Meredith Green

Professional Research Assistant

Meredith is an annotation project manager and adjudicator.

Kevin Stowe

Graduate Student / Research Assistant

Kevin is in his third year of the Linguistics Phd program at the University of Colorado, under the direction of professor Martha Palmer. His research is primarily in the capabilities of NLP systems for social media, as well as the automatic identification and understanding of metaphors. He is also curious about machine learning, logic, semantics, and the methods for extracting meaning from natural language accurately and automatically.

Ghazaleh Kazeminejad

Graduate Student / Research Assistant

Ghazaleh is a 3rd-year PhD student in the Department of Linguistics at University of Colorado Boulder, working under the supervision of professor Martha Palmer. She has worked on Persian Light Verb Constructions from a theoretical and computational perspective, creating lexical resources for the endangered language Arapaho, including creating a morphological parser and generator using finite-state automata techniques. She’s currently involved in a number of projects including bootstrapping a noun ontology corresponding to the Generative Lexicon theory based on the existing resources, improving and updating the VerbNet lexical semantic resource, and improving the Arapaho morphological parser. She’s also working on improving her skills in the machine learning and artificial intelligence fields to apply them to her academic research which is on communicating through gesture and extracting meaning from it.

Timothy O'Gorman

Graduate Student / Research Assistant

Specialization: Computational lexical semantics, construction grammar(s), distributional compositional semantics, information structure, Arabic syntax

Leo Kim

Student / Staff Programmer

Leo is a fourth-year undergraduate student in computer science and linguistics, and is enrolled in the concurrent MA program for linguistics.

Interests: natural language processing; models of semantic representation (computational, cognitive, or otherwise); computation for social good

How is VerbNet Used?

VerbNet is used in a variety of natural language processing applications and research projects, including:

Word Sense Disambiguation

VerbNet can be used to distinguish between multiple possible senses for a given verb.

http://verbs.colorado.edu/~mpalmer/papers/wsd.pdf
Palmer, Martha, Hoa Trang Dang, and Christiane Fellbaum. "Making fine-grained and coarse-grained sense distinctions, both manually and automatically." Natural Language Engineering 13.02 (2007): 137-163.

http://www.cs.huji.ac.il/~oabend/papers/VN.pdf
Abend, Omri, Roi Reichart, and Ari Rappoport. "A supervised algorithm for verb disambiguation into VerbNet classes." Proceedings of the 22nd International Conference on Computational Linguistics-Volume 1. Association for Computational Linguistics, 2008.

Figurative Language Detection

VerbNet can be used for metaphor and figurative language detection by looking for violations to selectional restrictions imposed on thematic roles.

For example, given the selectional restrictions to the thematic roles for verb DESTROY-44 as shown in the given image, we can say that any use of DESTROY-44 that violates the requirement for the patient to be a concrete object is a figurative use of the verb.

Literal :
The bomb destroyed the building [+concrete]
The rock damaged the fence [+conrete]

Figurative :
The speaker destroyed his argument [-concrete]
The politician damaged her reputation [-concrete]


https://pdfs.semanticscholar.org/29b1/65f82233f6c2e93d31dc274b57adc1da1d48.pdf
Wilks, Yorick, et al. "Automatic metaphor detection using large-scale lexical resources and conventional metaphor extraction." Meta4NLP 2013 (2013): 36.

Visual VerbNet

A computational resource using VerbNet and the MS COCO image dataset to construct a verb lexicon to generate descriptions of action in still images.

http://www.vision.caltech.edu/~mronchi/projects/Cocoa/

Caused-Motion Constructions (CMC)

A project extending VerbNet to appropriately integrate the semantics from (CMC) with verb class semantics. Consider the examples below:

She blinked the snow off of her eyelashes.

They hissed him off the stage.



http://www.aclweb.org/anthology/S15-1006
Hwang, Jena D., and Martha Palmer. "Identification of Caused Motion Constructions." Lexical and Computational Semantics (* SEM 2015) (2015): 51.

http://www.lrec-conf.org/proceedings/lrec2014/pdf/624_Paper.pdf
Hwang, Jena D., Annie Zaenen, and Martha Palmer. "Criteria for Identifying and Annotating Caused Motion Constructions in Corpus Data." LREC. 2014.

Communicating With Computers (CWC)

A DARPA program aimed at improving communication with computers through better representations of Elementary Composable Ideas. Maps VerbNet to Generative Lexicon and to Combinatory Categorial Grammars (CCG).

http://www.darpa.mil/program/communicating-with-computers

Involved in a project or paper that should be listed here? Let us know!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form

Try VerbNet for Yourself!

Anybody can use VerbNet! Visit the "Downloads" tab to work with the latest public release, or visit one of the following links for a more gentle introduction to working with VerbNet.

Unified Verb Index (UVI)

The Unified Verb Index is a system which merges links and web pages from four different natural language processing projects. The current release is based on VerbNet 3.2.

http://verbs.colorado.edu/verb-index/

GitHub

Visit us on GitHub to keep up with latest developments and newest changes.

https://github.com/leokim89/verbnet

Natural Language Toolkit (NLTK) Module

The Natural Language ToolKit (http://www.nltk.org) has a VerbNet Corpus Reader generously maintained by Ed Loper.

http://www.nltk.org/_modules/nltk/corpus/reader/verbnet.html

VerbNet Projects in Other Languages

The VerbNet project (an English language resource), has inspired similar initiatives in other languages. Check them out below!

French

https://verbenet.inria.fr/

Pradet, Quentin, Laurence Danlos, and Gaël De Chalendar. "Adapting VerbNet to French using existing resources." LREC'14-Ninth International Conference on Language Resources and Evaluation. 2014.

https://hal.inria.fr/hal-01084560/document

Urdu (Proposed)

Hautli-Janisz, Annette, Tracy Holloway King, and Gillian Ramchand. "Encoding event structure in Urdu/Hindi VerbNet." Proceedings of the 3rd Workshop on EVENTS at the NAACL-HLT. 2015.

http://www.aclweb.org/anthology/W15-0804

Involved in a project or paper that should be listed here? Let us know!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form

Related Resources and Links

Computational Linguistics EducAtion Research (CLEAR)

Martha Palmer is a Professor of Linguistics and Computer Science and an Institute of Cognitive Science Faculty Fellow. Her PhD is in Artificial Intelligence from the University of Edinburgh. She is an Association of Computational Linguistics (ACL) Fellow, and has won an Outstanding Graduate Advisor 2014 Award, a Boulder Faculty Assembly 2010 Research Award and was the Director of the 2011 Linguistics Institute in Boulder. Her research is focused on capturing elements of the meanings of words that can comprise automatic representations of complex sentences and documents. Supervised machine learning techniques rely on vast amounts of annotated training data so she and her students are engaged in providing data with word sense tags and semantic role labels for English, Chinese, Arabic, Hindi, and Urdu, funded by DARPA and NSF. They also train automatic sense taggers and semantic role labelers. A more recent focus is the application of these methods to biomedical journal articles and clinical notes, funded by NIH. She co-edits LiLT, Linguistic Issues in Language Technology, and has been a co-editor of the Journal of Natural Language Engineering and on the CLJ Editorial Board. She is a past President of ACL, and past Chair of SIGLEX and SIGHAN.

Unified Verb Index (3.2 - Stable)

The Unified Verb Index is a system which merges links and web pages from four different natural language processing projects: VerbNet, PropBank, FrameNet, and OntoNotes.

FrameNet

The FrameNet project is building a lexical database of English that is both human- and machine-readable, based on annotating examples of how words are used in actual texts. From the student's point of view, it is a dictionary of more than 10,000 word senses, most of them with annotated examples that show the meaning and usage. For the researcher in Natural Language Processing, the more than 170,000 manually annotated sentences provide a unique training dataset for semantic role labeling, used in applications such as information extraction, machine translation, event recognition, sentiment analysis, etc. For students and teachers of linguistics it serves as a valence dictionary, with uniquely detailed evidence for the combinatorial properties of a core set of the English vocabulary. The project has been in operation at the International Computer Science Institute in Berkeley since 1997, supported primarily by the National Science Foundation, and the data is freely available for download; it has been downloaded and used by researchers around the world for a wide variety of purposes.

PropBank

The original PropBank project, funded by ACE, created a corpus of text annotated with information about basic semantic propositions. Predicate-argument relations were added to the syntactic trees of the Penn Treebank. This resource is now available via LDC. This project was continued under NSF funding and DARPA GALE and BOLT. with the aim of creating Parallel PropBanks (the English-Chinese Treebank/PropBank) and also PropBanking other genres, such as Broadcast News, Broadcast Conversation, WebText and Discussion Fora, at the University of Colorado. PropBank is also being mapped to VerbNet and FrameNet as part of SemLink: Mapping together PropBank/VerbNet/FrameNet. PropBank's coverage is also being extended to provide support for AMR annotation, which makes heavy use of PropBank frame files. This is being funded by DARPA DEFT.

Downloads

You can help us to better understand the different ways that VerbNet is used around the world by completing the form below, in whole or in part. All fields are optional, but your participation is greatly appreciated by the VerbNet team!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form
VerbNet 3.3 (Latest Public Release)

VerbNet3.3 is now available. 3.3 includes new path_rel predicates, significant updates to semantics, and increased coverage. See the UVI for more.

VerbNet 3.2

In addition to the changes in syntactic frame names that characterized VerbNet 3.1, 3.2 also includes many updates to semantic role labels. These changes bring VerbNet semantic role labels more in line with the LIRICS labels. Documentation below. 

VerbNet 3.1

Version 2.3 reflected the integration of the first set of classes into VerbNet, Korhonen and Briscoe (2004), as did VerbNet 2.0 and VerbNet 2.1, as well as the second set of classes, Korhonen and Ryant (2005). 2.3 also included minor revisions resulting from the mapping to FrameNet. These are all incorporated in 3.1.

VerbNet 3.x Documentation

Documentation/ReadMe for VerbNet 3.0/3.1/3.2

VerbNet Java API

Implemented via the Inspector Application Framework, which allows researchers/developers to fire custom Java code as various VerbNet structures are encountered during the scanning of the XML files (which are available in the download above). Thanks to Derek Trumbo for this.

Unified Verb Index (UVI)

Unified Verb Index version 3.3

The Unified Verb Index is a system which merges links and web pages from four different natural language processing projects. The current release is based on VerbNet 3.3.