Abstract
This essay examines the orderly process of rhetorical composition and inquires whether that system can be implemented as an expert system on a computer or word processor. A prototypical expert program is offered based on the Burkian pentad. Based on that system, the question of whether speeches can be composed by a computer employing artificial intelligence is raised. The basic problems in using AI or expert systems for speech composition are: 1) our inability to define and describe the systematic behaviors humans use in the process of preparing speeches; 2) our inability to code the process humans use in delivery of speeches; 3) our tendency to resist technology with the claim that speech composition is a matter of humanistic artistry. A partially tested expert program is offered for consideration.