Researchers who have attempted to make sense of silence in data have generally considered literal silences or such things as laughter. We consider the analysis of veiled silences where participants speak, but their speaking serves as ‘noise’ that ‘veils’, or masks, their inability or unwillingness to talk about a (potentially sensitive) topic. Extending Lisa Mazzei’s ‘problematic of silence’ by using our performativity–performance analytical method, we propose the purposeful use of ‘unusual conversational moves’, the deployment of researcher reflexivity and the analysis of trouble and repair as methods to expose taken-for-granted normative frameworks in veiled silences. We illustrate the potential of these research practices through reference to our study on men’s involvement in reproductive decision-making, in which participants demonstrated an inability to engage with the topic. The veiled silence that this produced, together with what was said, pointed to the operation of procreative heteronormativity.