Classic probability theory (CPT) is generally considered the rational way to make inferences, but there have been some empirical findings showing a divergence between reasoning and the principles of classical probability theory (CPT), inviting the conclusion that humans are irrational. Perhaps the most famous of these findings is the conjunction fallacy (CF). Recently, the CF has been shown consistent with the principles of an alternative probabilistic framework, quantum probability theory (QPT). Does this imply that QPT is irrational or does QPT provide an alternative interpretation of rationality? Our presentation consists of 3 parts. First, we examine the putative rational status of QPT using the same argument as used to establish the rationality of CPT, the Dutch Book (DB) argument, according to which reasoners should not commit to bets guaranteeing a loss. We prove the rational status of QPT by formulating it as a particular case of an extended form of CPT, with separate probability spaces produced by changing context. Second, we empirically examine the key requirement for whether a CF can be rational or not; the results show that participants indeed behave rationally, at least relative to the representations they employ. Finally, we consider whether the conditions for the CF to be rational are applicable in the outside (nonmental) world. Our discussion provides a general and alternative perspective for rational probabilistic inference, based on the idea that contextuality requires either reasoning in separate CPT probability spaces or reasoning with QPT principles.