On June 9, 2013, Bill Hart-Davidson, associate professor of WRAC and co-director of WIDE-MATRX, and Ryan Omizo, MATRIX post-doctoral research fellow in computational rhetorics, presented at Computers and Writing 2013 in Frostburg, Maryland. Their presentation, "Computational Rhetoric: An Invitation to the Laboratory," discussed a series of protocols that could be used to transform a natural language text into objects fit for statistical and graphable methods. This transformation is demonstrated by the accompanying image, which is a network graph of discussion board posts.
The presentation provided the audience with a first-hand glimpse into the work being done by the WIDE-MATRIX Computational Rhetorics Research Group, whose members also include WIDE-MATRIX Co-Director, Dr. Jeff Grabill; MATRIX Director of User Experience Projects, Dr. Liza Potts; and MATRIX Director, Dr. Dean Rehberger. The group's aim is to develop techniques and software that will train computers to apply principles of rhetorical theory to large-scale datasets— approximating the interpretive abilities of human readers, but on a much larger scale.
"What I think we are seeing for the first time is the 'big picture of rhetoric,'" Omizo says. "Rather than examining pieces of discourse and relating those pieces to pre-existing themes or forms, we can discern the shape and direction of massive corpora globally and empirically. Like most people in the field of rhetoric, we are building models of the ways in which people use language; but our computational models are being informed by much more data— and all at once."
The Computational Rhetorics group will also be presenting work at the upcoming Digital Humanities Conference in Lincoln, Nebraska in July. "One of the exciting things about this research is that it has appeal across a wide variety of disciplines, and we are eager to hear feedback from others using similar analytic techniques," says Hart-Davidson. "Some of the tools in our kit will be familiar, but we believe we are putting them to novel use looking for broader patterns of meaning such as our work that attempts to distinguish scientific reasoning from, say, political argument."