Page 252 - Encyclopedia of Philosophy of Language

Reference
proper names: Homer existed, Pegasus did not exist, etc. The apparent names cannot be logically proper names—pure denoting symbols—so must be treated as abbreviated descriptions. The first step is to find some description that the name abbreviates: perhaps Homer, the author of the Iliad and the Odyssey. Then the statement that Homer exists can be analyzed as:
There is one and only one person who wrote the (6) Iliad and the Odyssey.
W. V. O. Quine has developed this analysis by sug- gesting that if no appropriate definite description can be found with which to replace the proper name, it is possible, from a technical point of view, simply to turn the name directly into a predicate. In his example, the existence of Pegasus can be denied by denying the existence of anything which 'pegasizes' (Quine 1953).
Another advantage of the Theory of Descriptions concerns identity statements. The sentence Scott is the author of Waverley clearly has a differenfcognitive value' from Scott is Scott, implying that the expression 'the author of Waverley has a semantic role over and above that of denotation. Russell maintains that the sentence's true logical form is: There is one and only onepersonwhowroteWaverleyandthatpersonisident- ical with Scott. If someone holds that Tully is Cicero is something more than a tautology, he must be using one or more of the names as disguised descriptions, in which case again he must follow the two-step process of substituting descriptions for the names and then analyzing the descriptions accordingly.
6. Criticismsof RussellonDescriptions
Amongst the vast literature of commentary on Rus- sell's Theory of Descriptions one general kind of criti- cism has been prominent, and is relevant to the debate about the relations between names and descriptions. This is the criticism that Russell's theory fails to account for the purposes for which people use 'refer- ring expressions' in their ordinary speech. The most influential attack along these lines was made by P. F. Strawson (1950) who charged that Russell failed to distinguish a sentence or expression from its 'use on a particular occasion.' The very idea of a 'logically pro- per name' is flawed, Strawson thinks, through a con- fusion of what an expression 'means' and what it is being used to 'refer' to on an occasion; and likewise not any use of a sentence containing a definite descrip- tion should be assumed to involve a true or false assertion. There might be occasions where, even though the sentence is perfectly meaningful, nothing true or false has been said. The existential implication that Russell took to be part of the 'content' of a sentence with a definite description is, according to Strawson, merely a 'presupposition' for saying some- thing true or false by means of the sentence. By seeing reference as something that speakers do, on an occasion of utterance, rather than something that
expressions in a language do, Strawson radically altered the perspective for talking about names and descriptions.
Pursuing a critique along similar lines, Keith Don- nellan (1966) has argued that there are at least two distinct uses of definite descriptions, which he called 'referential' and 'attributive' uses. He gives the exam- ple of the sentence Smith's murderer is insane. In a context where several people believe (perhaps quite wrongly) that a certain individual (Jones) murdered Smith, then the sentence could be used to make the statement that Jones is insane. This referential use of the definite description, 'Smith's murderer,' does not require for its success that it be true that Jones mur- dered Smith, or even that Smith was murdered; it is true just in case Jones is insane. By contrast, a speaker using the description attributively to claim that 'who- ever murdered Smith is insane' cannot be making a true assertion, according to Donnellan, if it turns out that Smith has not been murdered at all. Donnellan suggests that Russell's theory gives at best an account of the attributive use of definite descriptions. Part of the interest of Donnellan's work is that he identifies a use of definite descriptions which make them look far more like proper names, conforming indeed to something approaching Mill's account, in that 'ref- erential' uses do not rely essentially on their descrip- tive ('connotative,' in Mill's terms) content (for refinements and problems, see Kripke 1977; Searle 1979).
7. Description Theoriesof ProperNames
Returning to proper names, it was noticed that both Frege and Russell arrive at a similar account of ordi- nary proper names, though through different routes. For Frege, names like Aristotle have a sense which might be specified by definite descriptions of the kind The teacher of Alexander the Great ...; for Russell, proper names of that kind are really 'truncated descriptions' and should be treated, within a full logi- cal analysis, as definite descriptions in the manner of the Theory of Descriptions. The common, 'descrip- tivist,' element, which some philosophers view as amounting to a theory of proper names, has several advantages: it meets the difficulties confronting Mill's account; it explains how names become attached to particular objects (the related descriptions are true of those objects); it offers a straightforward way of determining which objects the names refer to and indeed whether they refer to anything at all.
There are problems, though, with the Frege/Russell version of the description theory. One is that it leaves unclear exactly which descriptions are supposed to be connected to which names. Is there some principal description attached to each name such that if it fails to apply, the name itself fails? But one could not expect agreement in particular cases as to what that description should be. A deeper problem is that it
230