[The current snippet belongs in Part 1 which does “… first discuss mainstream scholarly attitudes towards law, legal theory and complexity combined …”]
Most legal people that we meet have a deep-rooted aversion against mathematical stories. And most legal people that we meet expect that complexity stories are mathematical. So they tend to turn their backs on complexity theory before they have seen its face. On the other hand, most economists that we meet love mathematics and have an aversion against complexity because complex stories defy mathematics.
It is very difficult to debate the use of mathematical stories with whom believe such stories to be existential to their trade. Winning (or at least keeping his ground in) such a debate is one of Ronald Coase’s many accomplishments. Coase upheld (crudely speaking) that science aims not for improved prediction per se, but for improved comprehension. And that the use of mathematical models can easily digress into “blackboard economics,” into playing with mindless abstractions, irrelevant to our comprehension of real world economics.
In the two articles that won Coase his Nobel memorial award the use of mathematics is extremely simple, yet convincing and comprehensible. These models represent how individual agents will behave (given a small collection of possible assets, actions, circumstances, benefits and costs). Such simple mathematics on simple, direct causal models applied in what reminds one of “lab conditions” is extremely valuable — has proven to be extremely valuable to science in general — and belongs in the methodological attitude that Coase promotes.
Such simple formal mathematical models keep within the limits of first order predicate logic. Consequently, modelling the dynamics in predicates (or in the characteristics of objects) is outside these limits (we’ll discuss this claim, which is one of our operationalizations of Coase’s attitude to method, elsewhere).
So we expect that Coase would prefer studying with agent based simulations the behavioral results of complex groups, of at least somewhat realistically modeled agents, rather than to attempt and model such a system’s behavior in complicated and difficult to understand mathematical models that predominantly aim for prediction power.
Designing agent-based models, simulating their behavior and analysing the behavioral output constitute the three-pronged tool of complexity theory. The tool conserves a clear relationship between model and agent behavior and thus supports comprehension. And it prevents the use of the sort of “blackboard” mathematics that generates huge knowledge asymmetries, accompanied by the interpretative magic of who claim to understand math, yet know little about economics (or law) and thus may easily float with their theories into the domain for mindless abstractions.
We adopt Coase’s attitude to method, because it seems exceptionally appropriate when adopting complexity theory for studying the law and aiming for an audience of legal professionals. (As we will be returning to the subject a few times yet, we will refer to it as CAM [for Coase’s Attitude to Method])
[We ground our argument in the combined interpretation of seven articles: Coase (1937), Friedman (1953), Coase (1960), Coase (1978), Coase (1988), Posner (1993), Medema (1995) and Posner (2009).]