Ontological semantics, formal ontology, and ambiguity

  • Authors:

  • Sergei Nirenburg;Victor Raskin

  • Affiliations:

  • New Mexico State University, Las Cruces, NM;New Mexico State University, Las Cruces, NM, and Purdue University

  • Venue:
  • Proceedings of the international conference on Formal Ontology in Information Systems - Volume 2001
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

Ontological semantics is a theory of meaning in natural languageand an approach to natural language processing (NLP) which uses anontology as the central resource for extracting and representingmeaning of natural language texts, reasoning about knowledgederived from texts as well as generating natural language textsbased on representations of their meaning. Ontological semanticsdirectly supports such applications as machine translation ofnatural languages, information extraction, text summarization,question answering, advice giving, collaborative work of networksof human and software agents, etc. Ontological semantics paysserious attention to its theoretical foundations by explicating itspremises; therefore, formal ontology and its relations withontological semantics are important. Besides a general briefdiscussion of these relations, the paper focuses on the importanttheoretical and practical issue of the distinction between ontologyand natural language. It is argued that this crucial distinctionlies not in the (inaccurately) presumed nonambiguity of the one andthe well-established ambiguity of the other but rather in theconstructed and overtly defined nature of ontological concepts andlabels on which no human background knowledge can operateunintentionally to introduce ambiguity, as opposed to pervasiveuncontrolled and uncontrollable ambiguity in natural language. Theemphasis on this distinction, we argue, will provide bettertheoretical support for the central tenets of formal ontology byfreeing it from the Wittgensteinian and Rortyan retreats from theanalytical paradigm; it also reinforces the methodology of NLP bymaintaining a productive demarcation between thelanguage-independent nature of ontology and language-specificnature of the lexicons, a demarcation that has paid off well inconsecutive implementations of ontological semantics and theirapplications in practical computer systems.