In philosophy a natural kind is a family of "entities possessing properties bound by natural law; we know of natural kinds in the form of categories of minerals, plants, or animals, and we know that different human cultures classify natural realities that surround them in a completely analogous fashion" (Molino 2000, p.168). The term was brought into contemporary philosophy by W. V. Quine in his essay "Natural Kinds", where any set of objects forms a kind only if (and perhaps if) it is "projectible", meaning judgments made about some members of that set can plausibly be extended by scientific induction to other members. Hence "raven" and "black" are natural kind terms, because any black raven consistutes at least some evidence that all ravens are black. But "nonblack" and "nonraven" are not, because a nonblack nonraven (say, a red herring) is not evidence that all nonblack things are nonravens. Nelson Goodman's problem predicate "grue", meaning "observed before 1 January 2000 and blue or observed after 1 January 2000 and green", turns out to be inappropriate for science because it does not denote a natural kind. Quine argued that kind-hood was logically primitive: it could not be reduced non-trivially to any other relation among individuals.

Cultural artifacts are not generally considered natural kinds. As one author puts it, "they never stop changing, and terms that designate them constitute only what Wittgenstein called 'family resemblance predicates'" (ibid, p.169). This point is more diputed; John McDowell has extensively argued that this opposition between "culture" and "nature" cannot be clearly formulated, and that in any case it ought to lead us to construing cultural products not as un-natural, but as, adopting Aristotle's terminology, a kind of "second nature."

Source

• Molino, Jean (2000). "Toward an Evolutionary Theory of Music and Language", The Origins of Music. Cambridge, Mass: A Bradford Book, The MIT Press. ISBN 0262232065.