In response to Robyn Carston's (2010) plea for explicature as terminable, this article explores what Grice meant by ‘cancellable’. It presents reasons for assuming that only generalised conversational implicatures are cancellable, not relevance theory's explicatures or particularised conversational implicatures. It concludes by obelizing relevance theory's re-analysis of Grice's GCIs as explicatures.