Australasian Mathematical Psychology Conference 2019

The coevolution of artifacts and ideas: An inference-based model of cultural evolution

Vanessa Ferdinand
Melbourne School of Psychological Sciences, The University of Melbourne

Learning is rarely, if ever, an unbiased process. As cultural artifacts replicate by being passed from individual to individual, among social groups, and across generations, the cognitive biases involved in the perception, processing, and production of these artifacts can operate as selection pressures on them, causing certain forms to increase in number at the expense of others. Here, I will discuss the similarities between replicator dynamics (a general model of evolution) and Bayesian inference (a general model of learning) and utilize their mathematical equivalence to specify a model where cultural artifacts and learners’ hypotheses about those artifacts co-evolve. Culture is a special evolutionary system that is composed of two types of replicators: public structures in the world, such as artifacts and behaviors, and private structures in the mind, such as brain states or hypotheses (Sperber, 1996). The most interesting part of this model is the interpretation of fitness for both types of replicators. The fitness of public replicators is given by their likelihood under the population of hypotheses in learners’ minds, and the fitness of private replicators is dictated by their likelihood under the population of artifacts in their environment. Both of these replicators can place constraints on one another as culture evolves and drive the system to unexpected places when fitness values are asymmetric.