Resiliency, propensities, and causal necessity

Abstract

This paper analyzes the concept of ‘resiliency’. It argues that statistical laws confer propensities on systems to behave in probabilistic ways. This is not simply stating that the relative frequency of some event is a certain proportion, as this may not hold in all cases due to random fluctuations. The paper proposes a measure called ‘resiliency’ to quantify the stability of a probability claim. Two types of resiliency are defined: probabilistic resiliency and resiliency for conditional probabilities. It argues that statistical laws with high resiliency are better supported. Furthermore, resiliency provides a way to address classic problems like the ‘paradox of ideal evidence’ and the impact of hidden variables on propensities mentioned in quantum mechanics – AI-generated abstract.