Hapless Human: "Siri, what is two plus two?" *Ding, ding!* Siri: "The capital of Bangladesh is Dhaka."

Carolyn R. Miller’s article “What Can Automation Tell Us About Agency?” describes an interesting thought-experiment based on ideas about rhetorical agency from the Alliance of Rhetorical Studies conference and the recent increase of automated assessment by computerized systems in American public education. Miller notes at the beginning how many tests are now standardized and corrected by machines that give instant feedback to the students. She creates the fictional AutoSpeech-Easy, a computer system in which “students deliver their speeches to a videocam connected to a computer, and the system then delivers an assessment (or score) to the student and to the instructor or placement administrator, who records it without ever having to hear or see the student’s work (139). Miller poses this development as “revolutionary technology” that will transform public speaking classes.

When Miller asked instructors of public speaking courses their thoughts on AutoSpeech-Easy, many of the responses were negative. The responses delved into the obvious, such as the fact that speaking to a video camera takes away the “public” part of public speaking, into the more subtle and nuanced aspects of public speaking’s art, such as the connection with audience and the ability to feel out kairos and take advantage of it. More importantly, according to Miller, the overall response to her thought-experiment showed that people seem to resist this automated system because they are committed to agency, especially that of the audience (141). Miller sees agency as “the kinetic energy of rhetorical performance” (147). She believes that the performance itself owns agency, rather than the audience or rhetor owning or possessing it. Agency involves the interaction between speaker and audience, the give and take of beliefs, words, nods, frowns, laughter, sighs, and all that goes on to create the connection between the two agents.

Miller concludes that the main dilemma concerning AutoSpeech-Easy is that, as an “audience,” it cannot perform (149). It lacks a living presence that is needed for a speech to be a true speech. People are not acting like Shakespeare’s Hamlet, throwing  out soliloquies left and right to determine a decision; they are mostly speaking to real, living people who have positions, perspectives and ideas to be altered. They are changeable creatures who can be persuaded; a computer (so far as I know) is not. It is a neutral, lifeless object that cannot push its emotions or reactions back onto the speaker. If Miller is right in believing that agency is a kinetic energy flowing between speaker and audience, then agency cannot occur when one talks to a computer or other object. Energy flows from the speaker, but then the object just absorbs it; the object cannot effuse the kind of energy a human audience provides.

Reading about the automated computer assessment of speeches reminded me of Siri, the so-called “intelligent personal assistant” (according to Apple) on iPhones. While it acts human, it is obviously not human. While it has a name, giving it agency (in the literary definition of the word; not the rhetorical definition of it), it still has a lot of problems that come with any automated system. I have heard my friends on multiple occasions asking Siri to give them an answer, only to receive one that is completely off-topic. When my friend asked for the Clemson football game score last Saturday, Siri answered with a fun fact about the year 1902. My other friend looked up the score using good, old-fashioned Google on his smart phone after my friend repeatedly failed at coaxing Siri to give her the score. Apple has some of the best technology in the world; however, not even this company has the means to create a machine that responds easily to human questioning and speech. It seems that, for the moment, rhetoric will have to stick between humans.