Deverbal Nouns in Knowledge Representation

Abstract

Deverbal nouns pose serious challenges for knowledgerepresentation systems. We present a method of canonicalizing deverbal noun representations, relying on a rich lexicon of verb subcategorization frames, the WordNet database, a large finite-state network for derivational morphology, and a series of heuristics for mapping deverbal arguments onto the arguments of corresponding verbs.

OpenWordNet-PT: Taking Stock

“…They are useful for linguistic research as well as for information extraction. Since one team member was already conducting research on nominals in English [16] and another one was writing a doctoral thesis on them in Portuguese [28], it made sense to investigate how nominalizations could and should be incorporated into wordnet-like resources.…”

Mapping Verbal Argument Preferences to Deverbals

“…In the current system implementation, we attempt to map deverbal nouns onto corresponding verbs using a small set of heuristics similar to that described in [7]. We distinguish between event nouns like destruction, agentive nouns like destroyer, and patient-like nouns like employee.…”

Mapping Verbal Argument Preferences to Deverbal Nouns

“…For example, the noun destruction refers to the action described by the verb destroy, and destroyer may refer to the agent of that event. Deverbal nouns are very common in English texts: by one count, about half of all sentences in written text contain at least one deverbal noun [7]. Thus, a computational system that aims to match multiple ways of expressing the same underlying events (such as question answering or search) must be able to deal with deverbal nouns.…”