Blog Archives

The transferable belief model in film and games studies

This post is another draft paper, this time focusing on the difference between ergodic and non-ergodic texts and so is of relevance to games studies as well as film studies. I use an approach that I don’t think has been applied to either of fields yet: the transferable belief model, which is a mathematical theory of evidence. Hopefully soon I will be able to outline how this model will be of use in film studies in more depth, along with considering its relationship to Bayesian approaches to modelling viewer behaviour. The idea is to apply these models to the empirical analysis of the beliefs of real spectators so that it may be possible to make some statements about how we understand films that are more than theoretical but which have a solid evidential basis. The article can be downloaded as a pdf file here:Nick Redfern – Credal and pignistic reasoning in ergodic and non-ergodic texts.


This paper discusses the difference between ergodic and non-ergodic texts by considering the different levels of reasoning required of an agent in each case. The difference indentified between such texts is based on the distinction between credal and pignistic reasoning in the transferable belief model. It is argued that non-ergodic texts require an active agent to reason about the state of the world, and thus operate at the credal level; while ergodic texts require that the belief function of an agent be transformed into a probability function for the purposes of decision making, and therefore entail both credal and pignistic reasoning. The difference between ergodic and non-ergodic texts considered in these terms is illustrated through comparing narratives from the CSI: Crime Scene Investigation franchise.