This paper determines and limns Writing Aloud Voice (WAV) sequences –a particular change of footing employed in hypothetical speech (Haberland, 1986) for doing formulation. It examines a collection of video recordings of student project group meetings and was assayed with Ethnomethodological and Conversation Analytic methods. It shows that WAV sequences consist of (1) a quotative construction projecting a change of footing, (2) a WAV clause produced with a prosody resembling that of reading aloud, and (3) an exiting device forming a return to the prior footing. WAV clauses include many speech perturbations, and the paper displays how these are employed by speakers to show the ongoing work of formulating a potential text.