NIST logo

Publication Citation: A Restricted English for Constructing Ontologies (RECON)

NIST Authors in Bold

Author(s): Edward J. Barkmeyer; Andreas K. Mattas;
Title: A Restricted English for Constructing Ontologies (RECON)
Published: July 10, 2012
Abstract: Verifying correctness of stated facts, rules, and term definitions for an industrial domain of work requires the contributions of experts in the domain. Conversely, use of automated reasoning technologies to assist in making industry decisions and validating industry information demands that facts, rules and the definitions of terms be stated in a formal logic language. Such languages are very difficult for engineers and operations experts to understand, and generally difficult to use. A restricted English is a language that looks like English, but is carefully restricted in grammar, so that it has an unambiguous equivalent in a formal logic language. This paper presents the grammar for RECON -- a restricted English language whose purpose is to capture industry concepts, facts and rules in such a way as to permit conversion to an extended firstorder logic language (IKL). The vocabulary of the language -- the nouns, verbs and adjectives -- is defined by the industry experts. The grammar regulates the forms of definitions and sentences that use those terms. The RECON grammar supports the use of complex noun phrases involving individuals, groups, classifiers and properties, quantifying expressions and ad hoc classifiers that are qualified by adjectives and properties. It supports verb phrases involving verbs and prepositional phrases, and it supports simple, compound and conditional sentence structures. RECON English can be read by experts in the industry domain, and can be written more easily by analysts who are capturing the knowledge of those experts. It facilitates the transfer of human knowledge to computational forms.
Citation: NIST Interagency/Internal Report (NISTIR) - 7868
Pages: 57 pp.
Keywords: ontology; controlled natural language; grammar
Research Areas: Data and Informatics
PDF version: PDF Document Click here to retrieve PDF version of paper (872KB)