Page 262 - Encyclopedia of Philosophy of Language

Reference
that meaningful observational predicates are never empty, and that coextensive observational terms are semantically interchangeable, but this position is very difficult to defend for those who admit a wider range of undefinable general terms.
4.2 Causal Theories and the Fregean Problems
Causal theories of reference may well be able to meet the Fregean problems in the same way that description theories do, by holding that a term's mechanism of reference is semantically relevant. The mechanism of reference, on the causal account, is not a description but a causal network linking term token to reference (see Devitt and Sterelny 1987). Distinct causal net- works link tokens of 'Mount Taranaki' and 'Mount Egmont' to Mount Taranaki, one network ultimately deriving from the linguistic behavior of Maori-speak- ing New Zealanders and the other, English-speaking New Zealanders. Because the two tokens refer in vir- tue of their linkage to distinct networks, Egmont is Taranaki differs in meaning from Egmont is Egmont.
For the same reason, belief sentences differing only in the substitution of one name for the other differ in meaning. A name can be meaningful but empty through being associated with a causal network similar to that of standard names, but which is not grounded in any object. The substitution of causal networks for descriptions in this revised approach to Frege's prob- lem has the advantage of being extendable to the par- allel problems with general terms, and without the restriction implicit in buck-passing theories.
The idea of deploying causal theories to reconstruct the notion of sense is very controversial; most see the causal theory as a theory of'direct reference,' a theory that dispenses with Frege's notion of sense and his agenda of problems for the semantics of singular terms (see Wettstein 1986; Almog, et al. 1989). Proponents of'direct reference' do not of course deny that there is some interesting difference between Alphonse believes that Samuel Clement wrote Huckleberry Finn and Alphonse believes that Mark Twain wrote Huckleberry Finn; they just deny that there is a semantic difference between them, for there is just one proposition—the set of worlds in which Twain wrote the book—for Alphonse to be related to. Explaining the difference between the sentences is a job for pragmatics or psy- chology or both, but not for semantics, the proper role of which is an explanation of the information load of linguistic representation (see Almog, et al. 1989). This division of labor may be appropriate, though troubles plague it. Direct reference theorists have mostly concentrated on singular reference, yet Fregean problems seem to arise for general terms as well. Even within singular reference, most attention is paid to identity and opacity. It is hard to see what account of the meaningfulness of sentences essentially containing an empty name can be given by those who abjure any surrogate for the notion of sense.
4.3 Other Problemsfor Causal Theories
However the Byzantine disputes between different factions of causal theorists might resolve themselves, causal theories can probably handle Frege's problems. But they face some of their own. The causal theory, like the description theory, is not fully general. There are a raft of examples that do not fit it. Some natural kind terms have been introduced not by encounter with samples but by prediction; 'neutrino' and 'black hole' are famous examples. The same can be true of names; 'Nemesis' has been introduced as the name for a hypothesized companion of the sun that disturbs the Oort cloud every 26 million years. Rather more prosaically, architects sometimes name their build- ings, and parents their children, before these have progressed beyond planning. At the time of their introduction, these terms do not depend for their sem- antic properties on a causal chain between term and thing. There is within language some passing of the semantic buck, and the causal theorist must accom- modate that.
Moreover the causal theories seems to deem imposs- ible something that clearly is possible: reference change. Simple versions of the theory imply that the reference of a term is fixed by the acts that introduce that term into the language; all future use depends on those introducing uses. Yet reference clearly can change;'Aotearoa' isnowwidelyused inNew Zealand as an alternative name for New Zealand, but it orig- inally named only the north island. The same is true of general terms. We use 'consumption' as a Victorian synonym of TB, but they used it for a much wider range of respiratory complaints. 'Gay' is another gen- eral term that has shifted reference vigorously in recent times. Clearly, causal theorists need a more complex account in which an expression's reference will depend not just on the introducing use but also on the ongoing uses; on not just their initial 'ground- ing' on their bearers but on re-encounters, 'reground- ings' on them. Empty names and general terms are also trouble. The introduction of a term always has some cause, yet some are empty. 'Phlogiston' does not designate oxygen even though oxygen was the usual salient cause of the experimental results culminating in the introduction of 'phlogiston.' 'Santa Claus' is empty despite a connection with a real historical figure and its 'regrounding' on assorted impostors.
No consensus has emerged; there is even debate about just what the phenomena are, let alone which theory best handles them. The most devoted defender of causal semantics could not claim that they offer a complete and satisfactory account of the phenomena in their domain. Still, sophisticated causal theories do a fairly good job with the most pressing items on their semantic agenda. The story is much less cheerful when the reduction of the semantic to the nonsemantic is considered.
240