Emotions have become an important field of study in historical research. Looking at a few works held as important in this area, this article investigates the issue of order, that is, of the way historians of feelings organize the material they have selected. The texts in my corpus display two main models. First, arranging the data synchronically (e.g., Sanders, Envy and Jealousy in Classical Athens), they take the form of thematic tableaux accounting for the state of one or a set of emotions at a specific time and in a specific geographical area. Second, arranging the data diachronically (e.g., Reddy, The Navigation of Feeling), they take the form of stage narratives, that is, of narratives that proceed not from event to event, but from phase to phase. Whatever the model selected might be, these texts pose questions frequently asked in the epistemology of history. Are the models found in the data or constructed? If they are constructed, are historians free to go about that con­struction as they please? Debating the work of Jouni-Matti Kuukkanen and post­narrativist theorists of history for whom these questions are obsolete, the article argues that they in fact are still worth posing. A distinction, however, must be made between verifiability and acceptability, as only acceptability can be produc­tively discussed at the level of the whole text.