In this dictionary, every word has its' own definition page, listing all the available information: meaning, pronunciation, example sentences, and relations to other words.
As explained in the Introduction to HyperDic, this release was derived from version 3.0 of the original WordNet ® lexical database. The different semantical relations in WordNet are explained below.
An entry word can often span several senses and grammatical categories. When a word has more than one sense, its' page starts with an overview section, with links to each different sense.
The pronunciation of most words is indicated by a phonetical transcription derived from The Carnegie Mellon Pronouncing Dictionary, version 0.6, Copyright 1998 by Carnegie Mellon University. Being an American dictionary, it mostly reflects American pronunciation, which is sometimes quite different from the British one.
The phonetical transcription is expressed in an "arpabet", i.e. an alphabetical representation of the sounds, where a quote denotes primary stress.
Each word entry starts with a short section giving a definition of the word. Many words are illustrated by English sentences, showing actual usage.
For verbs, there is a special section: the generic frames. These are not sentences, but abstract patterns, showing the entities (persons or things) which can be associated with the verb. Verbal agreement is not taken into account.
The next section lists the relations between word entries and other terms. Each word can be related to other words in a number of ways:
Hyponyms are more specific nouns or verbs. Before WordNet version 3.0, hyponymy provided only a partial ordering of the nouns, because there were 9 different nouns at the top of the conceptual hierarchy.
This has changed with WordNet version 3.0, where the most general noun is entity, and every noun in the dictionary is transitively a hyponym of entity, so that hyponymy now provides a total ordering of the nouns.
The traditional definition of a reflexive relation is that for every X, the relation holds between X and X itself. Since nothing can be broader than itself, strict hypernymy and hyponymy are irreflexive: a word can neither be a hypernym nor a hyponym of itself.
Hypernymy and hyponymy are inverse relations: whenever X is a hypernym of Y, then Y is a hyponym of X.
Some other relations, such as synonymy and antonymy are symmetric, meaning that everytime the relation holds between A and B, it also holds between B and A. For example, saying that A is a synonym of B, is the same as saying that B is a synonym of A.
In HyperDic, WordNet antonymy is expanded from a relation between words, to a relation between word sets. This produces many new antonym pairs, and introduces a new property: the set of antonyms of a word's synonyms is now equal to the set of synonyms of that word's antonyms.
Another interesting property is transitivity: whenever A is broader than B, and B is broader than C, then A is also broader than C. Since hypernymy and hyponymy are transitive, they provide a partial ordering of the English vocabulary. For example, in version 3.0 of WordNet, the longest transitive hypernymy derivation has 19 levels, going from the most specific rock hind (a.k.a. Epinephelus adscensionis), to the broadest entity.