In this talk I address an objection to theories that posit ‘free' (non-linguistically mandated) pragmatic contributions to explicit utterance content. The concern is that such processes are insufficiently constrained, making impossible a systematic account of our grasp of explicit content. In response, I first show that the objection depends on working with a highly underspecified notion of what these pragmatic processes are, which fails to appreciate their context-sensitivity. Second, I discuss an alternative that aims to accommodate the optionality of the effects in question by positing optional covert linguistic structure (e.g. Marti 2006, Merchant 2010). I argue that such structure has no role in utterance processing, therefore has no syntactic reality in the linguistic system, even as part of an account of competence.