Natural Kinds

views updated


Whether engaged in high-level scientific activity or in the ordinary business of living, we spend a great deal of our time sorting the objects we come across into kinds. Philosophers are concerned with the kinds of kinds into which we sort these objects, and with the principles that distinguish one kind of kinds from another. One kind of kinds that has loomed large in recent philosophical discussions is that of so-called natural kinds. And one conception of natural kinds has dominated discussion in the contemporary philosophies of science, language, and mind, and this conception will concern us here. But first, some background.

Historical discussions of natural kinds (Ayers 1981) usually start with Aristotle and his conception of the individuals that are members of a kind of substance in virtue of the fact that they share a certain property (an essence ) with all and only the other members of that kind. This essence can be specified in a real definition in terms of two of the five predicables : genus, species, difference, properties, and accidents. To give the most famous example, the species human being is part of the genus animal, and is distinguished from other animals by the difference rational; thus the essence of human beings is that they are rational animals. This essence determines the properties human beings possess (language, for example), although some members of the kind will also possess further properties that are not so determined, and these are the accidents (high intelligence, say, or lustrous skin).

In reaction against this Aristotelian vision, John Locke offered the distinction between real and nominal essences. Locke distinguished between the real essence ("the being of anything whereby it is what it is") and the nominal essence ("the abstract idea which the general, or sortal name stands for") (1975 [1689]). He argued that when we use general terms, we refer to kinds whose definition can be given entirely in terms of their nominal essence. He maintained, first, that the members of a kind share a real essence in virtue of sharing some property concerning their microstructure, and, second, that because we lack "microscopical eyes," we can never know if an entity has this property or not. He then claimed that the features constitutive of the nominal essence of an entity are nonproblematically open to our view, and that as a result only these features are capable of ensuring that our use of a term refers to the kind in question.

This view rests upon some questionable assumptions. First, it is not obvious that our reference to an entity must be secured by features that are unproblematically open to our view. Second, it is not obvious that we cannot know that an entity possesses some microstructural feature simply because that feature is not observable. The modern view of natural kinds rejects both of these assumptions.

This modern view was inspired by the writings of Saul Kripke (1980) and Hilary Putnam (1975). The numerous advances in natural-scientific knowledge since Locke's time have greatly increased our sense that we are able to know about the microstructures of things, and these advances helped lead Kripke and Putnam to reject the second of Locke's epistemological assumptions. Far more radical, however, was their rejection of his first assumption. They insisted that the reference of a natural-kind term is secured by the real essence of the kind, even if no one has any idea what this essence is. A connection with the mental lives of those who use the relevant natural-kind term remains, but is secured insteadby the requirement that they use the term with the intention of referring to entities of the relevant kind. More specifically, when people learn the meaning of a natural-kind term, they are presented with a sample of the kind, and their competent use of the term is then (partly) a matter of their using it with the intention of referring to anything whose nature is the same as the relevant sample. This idea of a nature has clear Aristotelian resonances, and like Aristotle, Kripke and Putnam took the nature of an entity to be identical to its real essence.

Putnam analyzed the meaning of a natural-kind term into the following four components: a syntactic marker (the part of speech to which it belongs, obviously "noun"), a semantic marker (in the case of "water," this would be "liquid"), a stereotype (in effect, the nominal essence, the range of observable features commonly associated with the term; in this case, "colorless, tasteless liquid," for instance), and an extension (the things in the world determined by the real essence of the kind, whatever that may be). In the Kripke and Putnam picture, the stereotype provides guidelines for the use of the term, but it does not fix the reference of the term in a sentence containing it. Nonetheless, use of a term that is guided by the stereotype is still genuine usepartly in virtue of the intention to refer, and partly in virtue of what Putnam called "the division of linguistic labor" (the idea that a competent user of the term would defer to relevant experts on the matter of whether something actually is a member of the relevant kind).

This account of natural-kind terms has numerous advantages. It seems to provide an easy solution to the apparent problem of incommensurability, for example. Formulations of this problem start from an assumption characteristic of logical-empiricist accounts of scientific terms, namely, that the reference of a natural-kind term is fixed by the theoretically informed general beliefs of those who use that term. Consequently, when those beliefs change to a certain degree, so does the reference of the term. In the light of Thomas Kuhn's idea that science undergoes massive revolutions in the theoretical beliefs of scientists, it seems, once this assumption is granted, that past scientists who used the term "electrons" were not speaking about the same things that present-day scientists speak about when they use this term. By insisting that reference is fixed by the real nature of kinds and not the transitory beliefs of scientists, the theory of Kripke and Putnam allows scientific terms to refer to the same entities over time even though the relevant beliefs of scientists change massively.

In more recent years, some philosophers have started to become suspicious of attempts to extend this account to all natural-kind terms (Dupré 1993). It is questionable how far the theory is capable of handling the kinds that biologists appear to speak of, for instance. Terms of ordinary language such as "frog," "toad," "rabbit," "hare," "onion," "garlic"terms that one might assume are both natural kind terms and of relevance to biologistsare deployed by the latter in ways that radically diverge from how they are deployed by ordinary speakers. When ordinary speakers use these terms, it seems that their intention is not to refer to the putative real essence of, say, "garlic," but rather to something that serves a certain function ("garlic" refers to that which serves a certain culinary purpose, for instance). One obvious response at this point is to say that these terms refer not to natural kinds but to functional kinds (Wiggins 2001), and that their reference is fixed by some description available to the users of the term. This possibility of diverging intentions suggests that one kind term might be a natural-kind term among a group of scientists (given how they use it) and a functional-kind term among a group of lay persons (given how they use it).

There are interesting questions as to whether this account of natural-kind terms contravenes or accords with a Fregean view of meaning (Evans 1973). There are also questions about the exact role that an appeal to natural-kind terms should play in arguments for an externalist account of mental content (that the content of a mental state is determined by suitably 'external' features). In addition, if the arguments for externalism that rely on the Kripke and Putnam account of natural-kind terms are sound and if a term such as "garlic" denotes a natural kind on the lips of a scientist but a functional kind on the lips of a layperson, and if the reference of functional kind terms is fixed by a description, then we seem to be saddled with the idea that scientists have a greater number of broad, externally determined mental states than laypersons.

See also Aristotle; Essence and Existence; Kripke, Saul; Kuhn, Thomas; Laws of Nature; Locke, John; Meaning; Natural Law; Proper Names and Descriptions; Properties; Putnam, Hilary.


Ayers, Michael R. "Locke versus Aristotle on Natural Kinds." Journal of Philosophy 78 (1981): 247272.

Dupré, John. The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge, MA: Harvard University Press, 1993.

Evans, Gareth. "The Causal Theory of Names." Proceedings of the Aristotelian Society, supp. vol. 47 (1973): 187208.

Kripke, Saul. Naming and Necessity. Cambridge, MA: Harvard University Press, 1980.

Locke, John. An Essay concerning Human Understanding, edited by P. H. Nidditch. Oxford: Oxford University Press, 1975 (1689): bk. III, ch. 3, sec. 15, p. 417.

Putnam, Hilary. "The Meaning of 'Meaning.' " In his Mind, Language, and Reality. Vol. 2 of Philosophical Papers, 215271. New York: Cambridge University Press, 1975.

Wiggins, David. Sameness and Substance Renewed. Cambridge, U.K.: Cambridge University Press, 2001.

Adrian Haddock (2005)

John Dupré (2005)