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). Complete lists of the thematic roles, selectional and syntactic restrictions, predicates,
and frame types are available on the Unified
Verb Index Reference Page.
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
|Roles and Restrictions: Agent[+int_control] Patient[+concrete] Instrument[+concrete]|
|Members: bang, bash, hit, kick, ...|
||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|
|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
||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).
||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.
||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 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.
||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.
||used mostly by classes involving Psychological Verbs and Verbs Involving the Body.
|Location, Destination, Source:
||used for spatial locations.
||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.
||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.
||underspecified destination, source, or place, in general introduced by a locative or path prepositional phrase.
||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.
||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.
||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.
||start point of transformation.
||end result of transformation.
||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.
|| used for classes with a predicative complement.
||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.
||used by Verbs of Perception for events or objects that elicit some response from an xperiencer. This role usually imposes
||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.
||class-specific role, used in Begin-55.1 class to express time.
||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.
Figure 2: Selectional Restrictions associated with thematic roles