*Abstract* Fox (2014) argues that the Gricean, pragmatic approach to scalar implicatures cannot generate scalar exhaustification implicatures. Fox (2014) assumes that the pragmatic approach might rely on the Gricean Maxim of Quantity. Fox (2014) gives a thought experiment to show that there can be an exhaustification implicature even if the Maxim of Quantity is deactivated. He concludes that the pragmatic approach is inadequate. I will defend the pragmatic approach against this particular objection. Ostensibly uncooperative speech is still subject to norms of rational behavior. There is still some conversational goal shared by the speaker and her audience even if it is common ground that the speaker will not share all the information that the audience would like to have. I will argue that, on the pragmatic approach, we should still expect an exhaustification implicature to arise in Fox’s thought experiment. My approach will, however, involve denying that the Gricean Maxim of Quantity is always responsible for the generation of scalar implicatures. But this is acceptable on a broadly Gricean picture of pragmatics. So I contend. It will be crucial to my response that we accept the Maxim of Secrecy. *Maxim of Secrecy*: Do not reveal information such that were you to share it, the goal of the conversation would be undermined. Finally, I discuss “might” implicatures that can arise in Fox’s thought experiment. I model a Gricean calculation of these implicatures in the setting of a bimodal logic that validates what I call ‘Axiom 5*’.