Prior research across disciplines has established the value of dialogic, whole-class discussions. Previous studies have often defined discussions in opposition to the notorious triadic pattern called recitation, or IRE/F, focusing on variations to the teacher's initiating question or evaluative follow-up on students' responses. Recent scholarship has also identified variations on recitations and dialogic discussions that suggest these categories might be flexible, containing types of interaction associated with particular contexts. However, research remains to be done on how such types, or genres, of dialogic, whole-class discussion emerge and develop over time. In this article, I take up this line of inquiry, analyzing the classroom discourse of five lesson excerpts generated by a prospective teacher and his students in a US secondary History classroom between October and March. I illustrate how, over time, teacher and students repeatedly renegotiated the nature of a recitation-style textbook review activity using similar patterns of language that suggested an emergent discourse genre. These five interactions did not all lead to dialogic, whole-class discussions; I explain their relative success or failure in terms of how they constructed participants' relationships to historical and classroom events. I argue that even failed attempts at generating dialogic discourse may be part of a developing genre.