Lachlan Walmsley (Australian National University)
Models are essential scientific tools that scientists use to make inferences and build theories and that policy-makers use to inform their decisions. But models are imperfect representations of their real-world targets, containing simplifications and distortions called idealizations. According to Michael Weisberg there are three kinds of idealization: Galilean idealization, minimalist idealization, and multi-model idealization (MMI). In this paper, I argue that Weisberg's notion of minimalist idealization is too minimal and that his notion of MMI mischaracterizes an important instance of multi-modelling in science today, that is, ensemble modelling in climate science, and should not be considered a kind of idealization at all.
