Gerald Graff and Cathy Birkenstein's writing textbook, "They Say / I Say," has triggered important debates among writing professionals. Not included within these debates, however, is the empirical question of whether the textbook's templates reflect patterns of language use in actual academic discourses. This article uses corpus-based discourse analysis to examine how two particular "moves" discussed in the textbook are realized in three large corpora of professional and student academic writing. The analysis reveals important differences between the textbook's wordings and those preferred by student and professional writers. It also uncovers differences in use of "interpersonal" functions of language by experienced and less experienced writers. In offering this detailed analysis of academic prose, I aim to extend calls to recenter language in writing research and instruction. I conclude with implications for discussing academic argumentation with students.