THIS IS AN ONLINE EVENT

In a chapter on rhetoric in his recently published book, The Software Arts (MIT Press, 2019) Warren recounts a history of demonstration.  Aristotle tells us that the strongest rhetoric is closely tied to logical demonstration.  The history of the “demo” starts in ancient Greece, when definitive demonstration was a matter of deduction as practiced in geometry. In the 17th c., Euclid’s demonstration is displaced by Boyle’s inductive demonstration made necessary when arguments began to be based on empirical data and not just derived from statements taken to be obviously true. Today, arguments are made on the basis of so much data—“big data”—that no one person could possibly read it all, much less observe its collection. This has necessitated the invention of yet another form of argumentation, which he terms “abductive demonstration.”  Computer games, simulations, the Silicon Valley “demo,” and various forms of data visualization are of this kind of rhetoric.  Curiously many of them can be understood as procedures of data compression, otherwise known as machine learning. He claims that algorithms, especially machine-learning algorithms, can be understood as arguments.  To argue against an algorithm yet another form of persuasive writing needs to be developed: the software essay, a form independently suggested by Alan Kay and Donald Knuth, two Turing Award winning computer scientists.  Today we are in dire need of rhetorical techniques for arguing against algorithms, especially machine-learning algorithms when they are now deployed so enthusiastically, pervasively, and irresponsibly.