The Automatist Storytelling System: Putting the Editor's Knowledge in Software
by Michael Murtaugh

Master of Science, MIT Media Lab, June 1996.


This thesis presents the Automatist Storytelling System -- an "editor in software" or "narrative engine" -- a system that produces dynamic and responsive presentations from an extensible collection of keyword-annotated materials. Sequencing decisions are made on the basis of association, and the overall structure and meaning of an experience emerges from the interactions of individual material presentations. In this highly decentralized model, viewers are consistently integrated participants, who exert varying degrees of influence or control over the construction of the experience. The viewers' role is considered primarily extradiegetic; viewers' actions influence the process of the storytelling rather than altering actual events in the story world. By making both the viewing experience and authoring process variable and extensible, the Automatist Storytelling System supports new story forms such as the "Evolving Documentary."

This thesis presents two systems, ConTour and Dexter, as examples of Automatist Storytelling Systems. These systems were developed and are described in terms of, respectively, the stories: Boston: Renewed Vistas and Jerome B. Wiesner: A Random Walk through the Twentieth Century.

Thesis Supervisor: Glorianna Davenport

Get the entire thesis in HTML form.