The main purpose of this paper is to demonstrate that classical concept theories and hybrids thereof are empirically inadequate for the terminological analysis and description of concepts in a number of sciences. Examples of the classification and definition of minerals in the field of mineralogy are used to illustrate that the defining features of mineral species are typically the attributes of prototype categories; i.e. they are, amongst others, culturally, perceptually, and bodily based, idealized and essentially interactional and functional in nature. Furthermore, it is argued that classification in mineralogy is founded on an experientialist rather than an objectivist epistemology. These factors strengthen the argument for a prototype approach to concept analysis not only in the humanities and the social sciences but also in the so-called natural and pure sciences.