Abstract

In this paper I will defend an approach to the thought experiment known as ‘Maxwell’s Demon’ based on a Maxwellian conception of statistical mechanics. Instead of assuming that thermodynamic descriptions depend reductively on the dynamics of molecular components, I will adopt a conception of thermophysics as a ‘resource theory’ in the Maxwellian line recently defended by Myrvold (2011) and Wallace (2017). From this interpretative stance, Maxwell’s demon would not lead directly to the plausibility of violating the second law of thermodynamics, but to show the pragmatic impossibility of knowing deterministically the exact microscopic of thermal systems.