In this paper, we give a brief overview of current, cognitive-psychological theories, which provide an account for how people explain facts: causal model theories (the predominant type of dependence theory) and mechanistic theories. These theories differ in (i) what they assume people to explain and (ii) how they assume people to provide an explanation. In consequence, they require different types of knowledge in order to explain. We work out predictions from the theoretical accounts for the questions people may ask to fill in gaps in knowledge. Two empirical studies are presented looking at the questions people ask in order to get or give an explanation. The first observational study explored the causal questions people ask on the internet, including questions asking for an explanation. We also analyzed the facts that people want to have explained and found that people inquire about tokens and types of events as well as tokens and types of causal relations. The second experimental study directly investigated which information people ask for in order to provide an explanation. Several scenarios describing tokens and types of events were presented to participants. As a second factor, we manipulated whether the facts were familiar to participants or not. Questions were analyzed and coded with respect to the information inquired about. We found that both factors affected the types of questions participants asked. Surprisingly, participants asked only few questions about actual causation or about information, which would have allowed them to infer actual causation, when a token event had to be explained. Overall the findings neither fully supported causal model nor mechanistic theories. Hence, they are in contrast to many other studies, in which participants were provided with relevant information upfront and just asked for an explanation or judgment. We conclude that more empirical and theoretical work is needed to reconcile the findings from these two lines of research into causal explanations.