How do-or how should-we parse the world into kinds of things? Going back at least to Plato, most philosophers have done so with respect to some notion or other of natural kinds. And many analyses of natural kinds have been essentialistic-that is defining those kinds with respect to universals, or some set of intrinsic properties, or necessary and sufficient conditions. And there's a long-standing dispute between thinkers who regard scientific categories as natural kinds with essential properties fixed by nature-those that "cut nature at its joints"-and thinkers who maintain that our classifications and categories have no essence and instead merely reflect human interests and values. A typical example of the former would be "having a mass of 1.7 × 10-27,"and examples of the latter would be the categories of "ADHD," "race," or "child abuse."
Khalidi aims for an epistemic, naturalistic, non-essentialist, account of natural kinds, one which comfortably embraces not only the usual candidates favored by essentialists (e.g., elementary particles, chemical elements, biological species), but also categories in the social and behavioral sciences. Drawing on cases from many scientific fields, from fluid mechanics and polymer science to virology and psychiatry, Khalidi argues that "natural kinds are investigative or epistemic kinds, in the sense that they are the categories revealed by our systematic attempts to gain knowledge of nature" (p. 43). Moreover, he claims that natural kinds can be "fuzzy" (i.e. have indefinite boundaries), satisfy epistemic virtues to varying degrees, and be mind-dependent in a way that doesn't detract from their reality or objectivity.
Although the book is pitched for a sophisticated and philosophically informed audience (and, needless to say, too complex to be adequately summarized in a brief notice such as this), it's clearly written, nuanced, compellingly argued, and worth the effort for JSE readers curious about the unavoidable metaphysical dimensions of doing science of any kind.
Authors retain copyright to JSE articles and share the copyright with the JSE after publication.