The original version of the Visual Thesaurus made its debut in February 1998 as an experimental interface to the English language. Since its inception, the goal of the project has been to make the semantic structure that underlies language both visible and interactive.

The interface to the Visual Thesaurus, now in its third version, is simple. You type in a word, and all the words and meanings related to that word animate in and smoothly arrange themselves around the word you typed in. Click any of these related words, and it moves to the center, along with all of its related words.

Because the Visual Thesaurus contains more than 271,000 words, 115,000 meanings, and 725,000 relationships, the primary challenge in designing the third version was to create a system that can automatically arrange “word maps” given any central word or meaning, since every map could not be “designed.” This was accomplished by creating a self-organizing system in which each word or meaning displayed onscreen is a sort of “actor” that is “thinking for itself.” When hundreds or thousands of these “actors,” each of which follows a simple set of rules that define how it should react to the actors around it, are combined, the result is a complex choreography that allows the design to be ultimately flexible. Because the rules that determine how each word or meaning behaves are derived from natural phenomena like magnetism, elasticity, viscosity and wind, the animation and movement seems fluid and natural.

Another challenge was to add features that both enhance the functionality and expand the audience—such as roll-over tips, context-sensitive menus and other devices that maintain the “see a word you like and click on it” metaphor that has always been central to the Visual Thesaurus—without sacrificing the interface’s simplicity.