What Can Automation Tell Us About Agency?

Computerized systems for automated assessment of writing and speaking create a situation in which Burkean symbolic action confronts nonsymbolic motion. What is at stake in such confrontations is rhetorical agency. In this article, an informal survey that asked teachers of writing and speaking about automated assessment informs an analysis of agency that contrasts writing and speaking along the dimensions of performance, audience, and interaction. The analysis suggests that agency can be understood as the kinetic energy of performance that is generated through a process of mutual attribution between rhetor and audience. Agency is thus a property of the rhetorical event, not of agents, and can best be located between the two traditional ways of defining agency: as rhetorical capacity and as rhetorical effectivity. Unwillingness to attribute agency to automated assessment systems makes them rhetorically ineffective and morally problematic.