Mapping the Landscape of our Ignorance

Abstract

AbstractHow can we deal with the problem of equifinality when modelling the Roman economy? In this chapter I describe one approach I built with Scott Weingart that uses agent-based simulation as a way of generating potential social networks as an artefact of an economy. We simulate our understanding of the Roman economy—not the economy itself, but rather, our stories, our theories, about the economy—and use the results of the simulation to guide our sense of what we might reasonably expect to find in the archaeology if our story about the economy is correct. We can explore the archaeology of the Roman economy to find archaeological networks, and see which of these match (or not) the outputs of the model. However, the models we build have to be ‘stupid’ in the sense that we only put in them what is there in the verbal model we as archaeologists describe in our essays and monographs. If we find that we have to add elements to make the model ‘work’, then that is an indication that our original verbal models are eliding critical points. Thus, the process of operationalizing someone’s description of the economy, of building a model, holds just as much value as (if not more than) the ostensible results of the simulation itself, in that it provides a method for iterating through different theoretical conceptions of the Roman economy.