Simulations for historical research ?

As part of our work in the project ModelSEN, Socio-epistemic networks: Modelling Historical Knowledge Processes, we are also interested in the Why! question. Instead of analyzing corpora or archives from the past, we would like to understand the motivation of actors in their time or how archives that we see now came into being, e.g. by controlled deletion of material in relation to power, or different likelyhoods of survival for data depending on gender.

One method we consider is the simulation of historical processes using agent-based modelling. The documentation for a software package that simulates the sending of letters can be found here: SciCom.

It allows testing hypotheses of communication ranges and updating of ideas and produces a letter ledger of all simulated communication. On a large number of such simulation outputs we can then test the effect of different kinds of “forgetting” letters on the structure of the generated social networks and compare the results with empirical findings.