Anyone who has attempted to keep a diary knows that faithfully recording history is nearly impossible, even in very small number settings. There is much that must be left out because of space and time constraints. Both deductive and inductive approaches to history face similar problems. Just as every historical narrative can potentially be more complete, so can every model. Indeed, many, perhaps most, debates among historians, political scientists, and economists arise from disagreements about whether variable x or event y has been inappropriately neglected or focused on by others. In this, the preceding chapters are no different from other models and alternative historical narratives and so are naturally open to such criticism. The historical narratives of Part II suggests that the models of Part I can shed useful light on the emergence of Western democracy, but no proof is possible.

For many readers, the predictions of the models will seem so evident in the historical narratives that further empirical analysis seems unnecessary. On the other hand, the historical narratives were written with the models in mind, and it is possible that the factors focused on were given greater prominence than they deserved. It is also possible that the pattern of reform was less regular and predictable than the narratives make them appear, because so much has been left out. It is possible that constitutional bargaining played a smaller role in constitutional developments than the narratives suggest.