Ontology

As Tom Gruber puts it, "An ontology is a specification of a conceptualization."^[*] This means that an ontology is a way of specifying what the concepts behind a set of terms mean. Ontologies are the evolution of other approaches to organizing meaning, specifically glossary, vocabulary, and taxonomy:

• Glossary—A glossary is a collection of terms used in a domain, with their definitions. It is not meant to be comprehensive or precise, but to inform a human reader of enough of the meaning of a term to comprehend the sentence.

• Vocabulary—A vocabulary, or sometimes a "controlled vocabulary," is an exhaustive list of all the terms for a given domain or subdomain. Often a vocabulary has definitions associated with it, but this is not essential.

• Taxonomy—A taxonomy is an organized set of terms. The organization is almost always a hierarchy. In a rigorous taxonomy the meaning of the hierarchic relationship is the same throughout. For example, the biologic classification of species into genus, family, order, and phylum is a rigorous taxonomy in that each hierarchic relationship means "is a type of." Many other taxonomies are far less consistent, with the result that the taxonomy becomes more of a grouping mechanism that is intended to help people find similar terms.

Before saying any more about ontologies, let's briefly discuss dictionaries, thesauri, and encyclopedias.

• Dictionary—The purpose of a dictionary is to define terms. Defining a term should concentrate on making sure the reader knows the general type of thing that the term refers to and then distinguishing the term from any other closely related terms. Dictionaries occasionally do this, but they devote a surprising amount of space to "encyclopedia-style" definitions, as well as circular definitions (terms defined in terms of other terms, which eventually are defined in terms of the first term^[*]).

• Thesaurus—A thesaurus is a network of related meanings. However, with a few exceptions (e.g., WordNet) the relationships don't mean anything and are not consistent. There are no definitions in a thesaurus; the assumption is that you know what the word means, and you're looking for a similar word. The organization of a thesaurus, such as Roget's Thesaurus, is a rich taxonomy of concepts in the English language.

• Encyclopedia—An encyclopedia is meant to tell you a great deal about a concept, once you know the concept. You might refer to a dictionary to find out what a badger is, but you would refer to an encyclopedia to find out as much as you can about badgers (their habits and ranges, their diet, etc.).

An ontology is an organized body of knowledge. It is a graph (as opposed to a hierarchy) of concepts. The concepts have terms associated with them, because that is how we know and communicate concepts. The ontology helps us refine the identification of a concept (such as "rule in" and "rule out" criteria for membership) so that we can find similar terms and express everything we know about a concept. In that way an ontology is a combination of a dictionary, a thesaurus, and an encyclopedia, expressed more precisely.

^[*]http://www-ksl.stanford.edu/kst/what-is-an-ontology.html

^[*]Possibly a record for the most times the word "term" was used in a single sentence.