Martin Reznick
Villagers in Late Roman Egypt left us the histories of their misfortunes on hundreds of papyrus petitions. They wrote in order to motivate the local magistrates to deliver them from injustice. Magistrates apparently needed the extra motivation: Benjamin Kelly writes that although officials did their best, obstruction abounded and cases caromed around within the justice system. Kelly can't escape the depressing conclusion that the system was so badly engineered that litigants often just gave up. (78) I argue that this is no bug of the system; this is a feature. Equipped with the concept of performance and a simple model, I reverse engineer this system and show that frustration and arbitrariness might have been essential. They plausibly induced villagers to petition less often than they otherwise might have. Individual choices to petition less reduced the overall volume. The bureaucracy, limited by manpower, would have been able to answer a larger share of petitions received than otherwise possible without individual self-regulation. Answering a larger share of petitions implied broader confirmation of the validity of the political order. Deceptively poor architecture might have served a very useful purpose. I show that self-limitation discouraged villagers from abusing the legal system for some ulterior motive, like prosecuting feuds. The petitions themselves and recent scholarship on Roman justice inform my reasoning.
I accept Clifford Ando's argument that petitions were performative communication. With each petition a citizen redeemed Rome's implicit claim to be the guarantor of social and political stability. In Ando's words, petitions “establish intersubjective recognition of normative validity claims.” (78) Via petitions government and governed recognized each other as such and accepted their mutual obligations. Kelly observes an “ethos of careful and prompt execution” among bureaucrats in Late Roman Egypt. (77) This ethos based on performance motivates bureaucrats in my model to answer as large a share of petitions received as possible. But self-interest constrains good intentions. Persuading subjects that the social order is normative might benefit Roman rule in general, but it is unclear how much the good of the group would have overcome natural resistance to the hard work of intervening in village problems. One magistrate might be tempted to let some other strategos worry about universal intersubjective recognition while he takes a long weekend. If all of them think like this, justice grinds to a halt. From the villagers' perspective, too much abstraction exsanguinates the petition process entirely. The villager whose son was beaten nearly to death probably cared more about retribution than getting feedback that the social order is normative. (BGU 1 45) Individual conflicts mattered.
My analytical approach takes both political performance and individual interests into account. It may be useful to think of the “legal system” as the aggregate of the conscious choices of villagers and magistrates. We can learn much about this aggregate by characterizing individual choices. The choices of villagers (do I send a petition or not?) and magistrates (do I intervene or stonewall this petition?) each depend on what the players expect that others will do. In other words, the petition process was strategic. Strategic situations lend themselves to game theoretic modeling. (Myerson 1-2) I develop a simple mathematical model to make precise arguments that require very few and very reasonable assumptions. We do not have to imagine any universal social rules for Roman Egypt to explain the decision to invoke the law. (Hobson 1993) This approach is useful because it connects phenomena we understand in non-obvious ways that might have resisted connection without a model. History is messy, but our arguments don't have to be.