Self-adaptive systems (SASs) have the ability to reconfigure their behavior to respond to changing external conditions. A key element of a SAS, therefore, is how to monitor the environment so that appropriate adaptations can be triggered. In complex systems, monitoring the environment in its entirety is either impossible or too expensive. As a result, some adaptations are not possible because there is no monitor in place to trigger them. This paper discusses the role of human input, given as speech or text, as a way to provide environmental information to a SAS. The idea is that, given the limitations of monitoring the environment in full, human commentary can potentially be used to build up a more complete picture of the operating context of a SAS. The paper describes existing technology that could be used to realize this idea and describes a number of scenarios where the idea could be useful.