Almost half of all graduate students leave their doctoral programs without finishing. Who leaves, taking which skills and strengths with them, is still poorly understood, however, because it is hard to measure exactly what graduate students learn in their doctoral programs. Since the expertise required of a PhD holder is highly dependent on discipline, the development of a better understanding of graduate education and attrition requires studying the process at the departmental level.This is a qualitative study of the cultural values and norms of academic astronomy, as transmitted through the socialization of graduate students into giving talks, asking questions, and participating in departmental speaking events. This study also looks at the conflicts that arise when implicit cultural norms, which are practiced but remain unacknowledged, are inconsistent with the official, explicit values and norms for speaking in astronomy.Doctoral students and faculty members in a single astronomy department, at a large western university, filled out a short survey about the stakes involved in astronomy speaking events. A subset of these individuals was interviewed in-depth about the goals of, and their experiences with, five departmental speaking events: Coffee Hour, Journal Club, research talks, Thesis defense talks, and Colloquia. These interviewees were: (1) graduate students who had given a verbal presentation at one of these events, and (2) graduate students and faculty members who were in the audience at a graduate student's presentation.The desired outcomes which were expressed for these speaking events included: (1) lively, informal discussion among all participants, (2) increasing graduate student verbal participation in these events as they "learn to speak like astronomers," and (3) the utility of these events in helping graduate students learn and practice their speaking and reasoning skills related to astronomy research. In practice these goals were not achieved due to: (1) the ubiquitous, but unacknowledged practice of judging others' speech performance to come to negative conclusions about those individuals' intentions, intellectual abilities or efforts, (2) a lack of feedback for graduate students on their verbal performances, and (3) a lack of faculty members making explicit their own solutions to the inherent dilemmas of academic speaking.