Title:
Predicting Without Explaining, Explaining Without Predicting
Abstract:
In this talk, we challenge the general applicability of Douglas (2009)’s view, according to which scientific explanation and prediction are functionally intertwined. Drawing on quantum mechanics—where a single, exceptionally successful predictive formalism coexists with multiple, mutually incompatible explanatory interpretations—we advance five claims. First, the pursuit of scientific explanation does not necessarily originate from deficiencies in predictive power. Second, the value of prediction does not lie in the veracity of the underlying predictive account. Third, the value of explanation is not necessarily related to a pragmatic tool for guiding interventions or practical action. Fourth, there is no connection between the scientific character of an explanation and its capacity to generate novel predictions or evidence. Fifth, predictions do not always reliably serve to test, adjudicate between, or probe the truth of competing explanatory accounts. We further defend these claims through examples across various areas of past and contemporary science beyond quantum mechanics, showing that prediction and explanation operate according to distinct principles. Altogether, our arguments advocate for a divorce between prediction and explanation.