Cognitive scientists are interested in explanation because it provides a window into the cognition that underlies one's understanding of the world. We argue that the study of explanation has tended to focus on what makes an explanation "bona fide" as opposed to the processes involved in how the explanation is generated. In the current study, we asked participants to respond to the request for an explanation within a novel domain after we manipulated their initial exposure to the domain, and thus the background of the request. In two experiments, we found evidence that the background shaped participants' interpretations of the prompt for the explanation and that this, in turn, influenced whether they used a causal or functional style of explanation when responding to the prompt. We also asked participants to evaluate a number of explanations and found that the manipulation of the background did not have the same effect on the evaluative tasks. Our data support a pragmatic approach (e.g. The scientific image. Oxford University Press, Oxford, 1980) to the study of explanation generation, a philosophical approach which argues that the background influences the interpretation of the question, the development of a relevance relation which connects the question and explanation, and the identification of some set of candidate answers. We also suggest there is an important difference between the process of generating an explanation and evaluating an explanation, a difference that has escaped the attention of cognitive scientists thus far.