Abstract

This article highlights and discusses the Ehrenfests’ use of toy models to explore irreversibility in statistical mechanics. In particular, we explore their urn and P–Q models and highlight that, while the former was primarily used to provide a simple counter-example to Zermelo’s objection to Boltzmann’s statistical mechanical underpinning of the Second Law of Thermodynamics, the latter was intended to highlight the role and importance of the Stoßzahlansatz as a cause of the tendency of systems to exhibit entropy increase. We also explain the sense in which these models are toy models and why agents can use them, as the Ehrenfests’ did, to carry out this important work, despite the fact that they do not represent any real system.