ABSTRACT In order to explore the possibilities of affirmative ethics and algorithmic judgment, this article puts machinic rhetoric in conversation with classical imitation pedagogy. Taking a machine-learning chatbot as my example, I examine how imitation and repetition in a restrictive economy of rhetorical models produces a limited affirmative ethics through dialectical relations. Drawing on Hannah Arendt's concept of representative thinking to theorize a procedure for algorithmic judgment, I argue that rhetorical training requires the affirmation of a plurality of models if (...) it is to generate not only versatility but also ethical repetitions. (shrink)