Reference

3. General terms and mass terms

Just as there are description theories of names, so also there are description theories of general terms like ‘tiger’, ‘hammer’ and ‘bachelor’, and of mass terms like ‘gold’ and ‘paper’ (see Mass terms). Speakers of the language associate various descriptions with a term. One of these descriptions, or most of a cluster of them, expresses the meaning of the term and determines what it applies to. If only one description does the job, the view is analogous to the classical description theory of names. If a cluster of descriptions does, the view is analogous to the cluster theory of names.

Kripke (1980) and Putnam (1975) argued that description theories are false of general and mass terms that apply to natural kinds. So they are false of ‘tiger’ and ‘gold’. The arguments are like the three against description theories of names (see §2). First, the theories yield unwanted necessities. The description we associate with ‘tiger’ is along the lines of ‘large carnivorous quadrupedal feline, tawny yellow in colour with blackish transverse stripes and white belly’. Yet it is not necessary that a tiger has four legs and is striped: a tiger might lose a leg, or in a different environment tigers might not be striped. Second, the term ‘gold’ is a rigid designator, applying to the same kind of stuff in every possible world. In contrast, an associated description like ‘dense yellow metal’ is nonrigid. Third, people who seem perfectly able to use a term are often too ignorant or misguided about the things to which it applies to supply an appropriate identifying description. Thus, some who use ‘elm’ and ‘beech’ cannot supply descriptions that distinguish elms from beeches; many who use ‘gold’ cannot distinguish gold from fool’s gold; it was once common to associate ‘fish’ with ‘whale’.

Putnam added a further argument built around the following fantasy. Imagine that somewhere in the galaxy there is a planet, Twin Earth. Twin Earth, as its name suggests, is very like Earth. In particular, each Earthling has a doppelgänger on Twin Earth who is a cell for cell duplicate of the Earthling. Twin Earth differs from Earth in one respect, however: the stuff that the Twin Earthians who appear to speak English call ‘water’, stuff that is superficially indistinguishable from what we call ‘water’, is not H₂O but a very different compound XYZ. So Oscar on Earth and Twin Oscar on Twin Earth are referring to different stuff by ‘water’. Yet Oscar and Twin Oscar are doppelgängers, associating exactly the same descriptions with ‘water’ (which is more plausible if we place Oscar and Twin Oscar in 1750 before the chemical composition of water was known). So those associations are not sufficient to determine reference and the description theory is wrong. Indeed, nothing happening in the head is sufficient to determine reference. As Putnam put it, ‘meanings just ain’t in the head’ (see Content: wide and narrow; Putnam, H. §3).

We have considered criticisms of description theories of proper names and natural-kind words. Do these criticisms extend to description theories of other words? Putnam took the arguments to apply to almost all words, including ‘pencil’ and ‘paediatrician’. Tyler Burge (1979) took a similar line, arguing that the meanings (contents) and references of a wide range of a person’s words and accompanying thoughts are not ‘individualistic’ in that they are not determined simply by that person’s intrinsic states. To a large extent they are determined by the person’s social context. Burge’s examples include ‘arthritis’, ‘sofa’, ‘brisket’, ‘clavichord’ and ‘contract’ (see Methodological individualism).

The Twin Earth fantasy brings out an important feature of description theories in general: even if a description theory gives the right answer to our central question for some word, its answer is incomplete. Thus, consider a description theory of ‘tiger’. According to the theory, the reference of ‘tiger’ is determined by the reference of such words as ‘carnivorous’ and ‘striped’. Suppose, contrary to the arguments above, that this were so. We then need to explain the reference of those words to complete the explanation of the reference of ‘tiger’. Description theories might be offered again. But then the explanation will still be incomplete. At some point we must offer a theory of reference that does not make the reference of one word parasitic on that of other words. We need an ‘ultimate’ explanation of reference that relates some words directly to the world, if there is to be any reference at all.