Teaching a computer to speculate

Advanced natural language processing techniques may well lead to a major breakthrough in computer applications. Those working in artificial intelligence are seeking ways in which the computer can be made to emulate the ability of the human mind to handle language. This article illustrates the challenges of restructuring human semantic knowledge into computer-usable forms. We discuss hierarchies, Venn diagrams, chainings, cycles, matrices, maps, networks, webs, hubs, and scripts, all of which can be used in our attempts to teach the computer to handle meaning and thereby speculate.