Abstract

Some strongly empiricist views of scientific knowledge advocate a rejection of metaphysics. On such views, scientific knowledge is described strictly in terms of knowledge of the observable world, demarcated by human sensory abilities, and no metaphysical considerations need arise. This paper argues that even these views require some recourse to metaphysics in order to derive knowledge from experience. Central here is the notion of metaphysical inference, which admits of different “magnitudes”, thus generating a spectrum of putative knowledge with more substantially empirical beliefs at one end, and more metaphysically imbued beliefs at the other. Given that metaphysical inference is required even concerning knowledge of the observable, the empiricist hope of avoiding metaphysics altogether is futile: knowledge of the observable simply involves metaphysical inferences that are of smaller magnitudes than others. Metaphysical inferences are required not only to distinguish veridical from non-veridical experience and to determine the quality of empirical information, but also in order to explain how we construct experience (through categorizations and classifications of objects, events, processes, and properties), how we extrapolate from empirical evidence to generalize about observable phenomena, and how we use this evidence to test and confirm hypotheses and theories.