A striking case of difference between kinds of kind terms

In "The Meaning of 'Meaning'," Putnam famously presented a largely convincing argument about natural kind terms (‘water’), which derived the necessary a posteriori: Water is H2O.

In an important court case pertaining to the rights of homosexuals to marry currently before the Supreme Court of Canada, a philosopher of language presented an argument which claimed to derive, on grounds strikingly similar to Putnam’s, the metaphysical impossibility for gays to marry. I paraphrase the argument as follows, to bring out its similarity to Putnam’s:

(1) Sociolinguistic observation (Putnam’s “indexical component”):

The word ‘water’ / ‘marriage’ has been applied (consistently and exclusively) to this liquid (pointing to instances of water) / this sexual relationship (pointing to actually married couples).

(2) Previous use imposes normative constraints:

Whether some liquid / relationship is rightly called ‘water’ / ‘marriage’ or not depends on whether or not it bears the same-(kind of)-liquid/relationship-as relation to this, the original liquid / relationship.

(3) Experts determine same-as relation (Putnam’s “pragmatic realism”):

(i) This kind of liquid / relationship is a chemical / historical kind. Its essence is individuated by its elements / origins. According to consensual chemistry / history, it is essentially a chemical / religious kind.

(ii) The elements of water / marriage according to chemistry / religion are H2O / a man and a woman.

(4) Metaphysical fact (trivial):

XYZ / A union of two gays or of two lesbians is not H2O / a union of a man and a woman.

(5) Metaphysical consequence (non-trivial):

Suppose there existed a transparent, odourless, colourless, potable liquid / an intimate, sexual, loving, committed, long-term, even child-raising relationship that was indistinguishable to us from water / marriage but was not H2O / heterosexual (say it was XYZ / homosexual). Such a liquid / relationship would unarguably we very water/marriage-like, but for all its similarity to the real thing, such a liquid / relationship wouldn’t be WATER / MARRIAGE. Seeming-to-us-as-if-it-were-water/marriage is not the same thing as being water / marriage.

Fool’s gold is not gold. A gay marriage is not a marriage.

(6) Modal consequence:

Water / Marriage is of necessity H2O / a union of a man and a woman.

(7) Normative linguistic consequence:

Our word ‘water’/ ‘marriage’, when used with its ordinary common meaning, excludes reference to XYZ / homosexual relationships. It is correctly applied (indeed, intelligible) only if it is used to refer to H2O / heterosexual relationships.

As have many besides me, I find the paraphrase frankly lamentable. But what is interesting is the fact that the arguments seem, on their face, indeed similar. Yet, judging by their widely divergent rhetorical appeal (should my judgments of their respective philosophical worth be shared), there is something drastically different between them. What that is is the subject of this paper. It points to a fundamental difference between natural kind terms and social kind terms.

Since (2), (4) and (6) are unproblematic premises, and (1) is true of water –for want of evidence of there ever having been anything different in rivers and lakes than what there currently is and given a postulate of the stability of nature-- and can be assumed true of marriage for the sake of argument, and assuming the reasoning to (5) and (7) to be valid, the difference of soundness between the arguments must then lie within (3) –that is, with the differing role of experts in both cases. I cash out this difference in terms of two complications with social kind terms that make them less amenable to well-understanding than chemical kind terms. The first is that their object does not preexist its theoreticians, and hence is not independent of them. The second is that the “indexical component” –as Putnam calls it– in each kind of kind term is different: A chemical kind term (at least in the simplest case) is indexed to an essence, a sort of thing which, by its very nature, is unchanging; whereas a social kind term (even in the simplest case) is indexed to an imagining, a sort of thing which, by its very nature, evolves over time and across cultures. A social kind term is thus essentially indexed to time, an index neutralized in the case of natural kinds terms because of the stability of nature.

These complications distinguish the natural (in this case, chemical) kind term from the social (in this case, legal) kind term in the following way. Where a chemical kind term like ‘water’ minimally means: same-liquid-as-this-canonical-instance (as individuated by experts), a social kind term like ‘marriage’ minimally means something expressible rather as: same-kind-of-relationship-as-we-communally-understand-this-canonical-instance (as individuated by experts). When we refer to water, we refer to a substance (real, or pragmatically real). When we refer to a canonical instance of marriage, we refer to a relation (real, or pragmatically real, or indeed perhaps ideal). But which such relation we refer to depends on delicate intensional cultural factors. (This is demonstrated by appeal to differences between cases of meaning –or word– changes and cases of simple extension of reference.) So social kind terms implicate a level of intensionality beyond that implicated by Putnam’s pragmatic realism in the case of natural kind terms.

In the paper, I point out the interesting and, I think, insufficiently unrecognized extent to which, as the individuation of natural (chemical) kind terms is intertwined with scientific (chemical) theory, so the individuation of social (legal or religious) kind terms overlap with our moral perspectives. Just as the chemist is not primitively compelled to exclude XYZ from the concept of water, but is compelled to do so by theory, theory is also what compels the (religious) speaker to exclude same-sex couples from marriage. (The Great Ape Project is dedicated to the inclusion of apes into the category of persons. People who would deny apes such status must do so based on their (implicit or explicit) views about apes; no conception of personhood can exclude apes all on its own.) Likewise, no conception of marriage excludes same-sex couples all on its own. What we find primitively compelling about the notion of marriage is inseparable from how we represent to ourselves the moral status of gay and lesbian relationships. As a result, the paper clarifies, I think, a thing or two about the role of Courts in our mental lives: To individuate our natural (chemical) kind terms –our concepts of natural (chemical) kinds– we defer to scientists. To individuate our social (legal) kind terms –or concepts of (legal) kinds– we defer to the Courts. Herein lies a relevant similarity in all general terms (or normative concepts).

Finally, the marriage argument could also be paraphrasable in terms of a Kripkean causal-historical picture –“the word ‘marriage’ that has come down the generations to me through an uninterrupted causal chain leading back to an original baptism (or national Constitution).” But that only reveals the paucity of the causal-historical picture’s capacity to handle important differences between kinds of kind terms. Part of what makes our concept of marriage the one it is, regardless of the accuracy of our representation of it to ourself, is the extent to which we (individually or collectively) commit our concepts about the social to the constraints imposed by our Constitution, the extent to which we will our concept of marriage to be consistent with these. And to the extent that we acknowledge the Courts as the experts in the consistency of our social concepts with constitutional provisions, just as we acknowledge chemists as the experts in the consistency of our chemical concepts with science as a whole, we defer to them. The gay marriage example nicely illustrates why any overly causal account of reference misses the essential fact about deference in the case of social kinds, one based on will.