Abstract

Both lay people and scientists organize the world around them by categorizing particular things as belonging to kinds. Scientists speak and theorize about various kinds of things, like hydrogen, gold, and water; electron and neutron; Canis lupus and Felis catus; igneous rock, sedimentary rock, and metamorphic rock; schizophrenia, psychopathy, and autism; Caucasian, African, and Amerindian. Given this variety of scientific kind categories, one fundamental question for philosophers of science is whether any of these kinds really are natural kinds, and if so, what it is that constitutes their naturalness.nbsp;is to say, when we classify particular things as belonging to the same kind, is it ever the case that they really or naturally belong to the same kind? Do we ever carve the world at its joints, so to speak? Or, are there rather just many different ways in which we, humans, decide to classify particulars into kinds, each possible classification as conventional as any other? In this dissertation, a realist account of natural kinds is defended, according to which some kind classifications are more natural than others. It is argued that the naturalness of scientific kinds can be accounted for by the Causal Unification Theory. According to this view, the naturalness of a kind category lies in the contribution this category makes to our epistemic practices, as a result of accommodating these practices to a causally unified group of properties. Because the properties of natural kind-members are causally unified, they ground multiple causally sustained generalizations. Thus, natural kinds are the metaphysical grounds of systematic scientific research by allowing for multiple inductive generalizations and predictions, and for unified explanations. This view is developed in response to the historically most important theories about natural kinds and several more recent ones. The Causal Unification Theory of natural kinds is then put to work to elucidate the naturalness of two controversial humannbsp;classifications, namely the biological classification of human races and the social psychological classification of homophobes. The current metaphysical race debate is criticized for being too dependent on broad philosophical issues, and it is shown that asking about the naturalness of races, rather than about their existence, allows for a more nuanced debate that is to a larger extent dependent on empirical results. The example of homophobia is discussed in relation to the worry that the normative dimension of many human kinds diminishes the extent to which they can be considered natural kinds. It is argued that we should not be too pessimistic, but that there is indeed reason for concern. To illustrate this, a case is made that the measurement of homophobia is inflated as a result of the normative dimension of this kind.