Human cumulative culture in the laboratory: effects of (micro) population size

Traditionally, experiments on social learning (in both humans and nonhumans) involve dyads, with an experimenter or experimenter-trained conspecific serving as the demonstrator and the participant as the observer. But social learning in nature often involves multiple potential models, and the models themselves were once learners. We discuss our studies of social learning by adult humans in interactive group settings in the absence of formal demonstrations by experimenters, which tracked transmission over multiple learner generations. In these experiments, we found evidence for cumulative learning over generations. This has allowed us to manipulate learning conditions in order to test hypotheses regarding the necessary conditions for cumulative culture. We also report results from a further experiment using similar methods, which compared conditions of varying cohort size. Participants were given the task to build a paper airplane to fly as far as possible. Contrary to expectations, there was no advantage for larger cohort sizes, in terms of the cumulative effects observed.