Author’s Note: Since this article was published, development on the Lingua::Wordnet modules has been superseded by the Linguana project, an effort to create a dynamic, open source lexicon based on Wordnet. The author won the Damian Conway Award for Technical Excellence in 2001 for papers concerned with the project. More information is available from the Linguana web site at http://linguana.net.
My two-year-old son is just starting to grasp concepts like “A ‘car’ on the road is different from a ‘car’ on a railroad track.” Computer programs can begin to exhibit this understanding as well; this article discusses the use of the Perl module Lingua::Wordnet to answer questions like, “What are some different kinds of cars?” and “What articles of clothing are made from leather?”
Wordnet is a lexical database of the English language organized according to current psycholinguistic theories of human lexical memory. Developed at Princeton’s Cognitive Science department in the early 1990s, Wordnet was possibly the first undertaking to produce a machine-interpretable collection of English on a large scale. And like all really helpful and important software projects, it’s open source and just waiting to be used and extended.
The Wordnet package consists of several text database files, text indexes for those files, binaries for searching the files, and the source code for those binaries. A brief ...