ABSTRACT
The sciences of the deep past use models not only to inform us about actual history, but possibility as well. Most philosophical analyses of palaeoscientific knowledge focus on ‘trace-based reasoning’: inferences are drawn from observations of ‘traces’ (for instance, fossils) to the past, in virtue of background theory concerning how traces form. Such accounts have trouble accommodating the ubiquity of modelling practices in these sciences. Models are vital because access to past actuality requires knowledge of past possibility, and models grant this. Furthermore, historical scientists often adopt ‘phenomena-driven’ modelling strategies, in which models are in intimate dialogue with the historical record. This serves to ensure that the possibilities explored are relevant to the investigation at hand.
