While proof is often presented to mathematics undergraduates as a well-defined mathematical object, the proofs students encounter in different pedagogical contexts may bear salient differences. In this paper we draw on the work of Dawkins and Weber (2017) to explore variations in norms and values underlying a proof across different pedagogical contexts, including the lecture, the textbook, home assignments and received feedback. We conclude that shifts in norms of proof within instruction and assessment, and the consequent difficulties in student proof learning, may be the result of an inherent tension between mathematicians' values and pedagogical goals.