Page 29 - Encyclopedia of Philosophy of Language

pose of identifying an object. For Strawson, when a definite description fails to apply to anything a basic presupposition for successful communication has failed and no truth-value can be assigned to the result- ing assertion (Russell had given the value false to sentences with failed references). Strawson's analysis was contributory to the enormous subsequent interest in presupposition. Donnellan (1966), pursuing the idea of reference as grounded in contexts of utterance, argued that there are at least two fundamental uses of definite descriptions which he called referential and attributive, the latter being somewhat akin to Straw- son's conception, but the former, controversially, giv- ing definite descriptions a role more like that of proper names. (For an account of this highly influential theory, and its ramifications, see Names and Descrip- tions; Russell, Bertrand; Identity.)
3.2 Proper Names
Russell and Wittgenstein (in his early work) developed the idea of a 'genuine' or 'logically' proper name the sole function of which is to denote a single and simple object, such that the object itself constitutes the mean- ing of the name (if there were no such object the name would be meaningless). Both were constrained by the search for an ideal logical language. In contrast, Frege argued that even in an ideal language denotation by means of a proper name must be mediated by some- thing he called Sinn (sense). Philosophers of language have long debated the merits of these opposing views, and versions of them have proliferated. Even Russell admitted that his conception of a logically proper name did not fit ordinary proper names in natural languages (Socrates, London, etc.), which he took to be 'truncated descriptions' analyzable according to the Theory of Descriptions. However, Saul Kripke (1972) and others have attempted to revive something like the pure denotation view applied to ordinary pro- per names, rejecting both Fregean Sinn and Russell's 'descriptivism.' This has led to so-called 'causal theo- ries' of names, where what determines the reference of a name is not the 'mode of presentation' embodied in the name's sense or whatever object satisfies an implied description but rather a direct causal link back to the object initially 'baptized' with the name (see Devitt and Sterelny 1987). In Kripke's terminology names are 'rigid designators,' in the sense that they designate the same object in all possible worlds where the object exists. Definite descriptions do not des- ignate rigidly because in different possible worlds different objects will satisfy the descriptions. Versions of causal theories are probably in the ascendancy, though increasingly sophisticated accounts attempt to reassimilate something like Fregean sense (see Moore, 1993). Searle (1983) presents a powerful case against causal theories, locating reference in the philosophy of mind and arguing for the central place of inten- tionality in an account of language (see Intentionality).
3.3 Indexicals
One particularly perplexing class of referential devices are indexical expressions (e.g., /, here, now, this, etc.) whose reference or extension is determined by context of utterance. The reference of T changes from person to person, It is late now changes truth-value when uttered at different times, and so forth. Once again it was Frege who set up the modern debate by proposing that sentences containing indexicals do not express a 'complete thought' until supplemented by indications of time, place, and other contextual determinants; while sentences contain indexical terms, thoughts (i.e., propositions) do not, for thoughts, if true, are time- lessly true. Other philosophers have attempted by different strategies to eliminate indexicals. Perhaps the most systematic and widely regarded modern treat- ment is by David Kaplan (1989). An important dis- tinction drawn by Kaplan, which has applications and consequences elsewherein philosophy of language, is between two aspects of sense: 'content' and 'charac- ter.' The character of an expression such as T is most naturally associated with its 'meaning,' conceived as a rule which roughly identifies the use of T to refer to the speaker. Its content is given only relative to a context of use and is associated with what is said in that context: for Kaplan, when two people use the sentence / am tired in different contexts not only will the truth-value be affected but the content will differ also. Kaplan's conception of indexicals as 'directly referential,' not mediated by Fregean sense, places him broadly in the same camp as the 'causal theorists' of reference. The article Indexicals explores philo- sophical approaches (notably Kaplan's), while Deixis presents the issues from the perspective of theoretical linguistics.
3.4 Natural Kind Terms
There is one further area where the distinction between sense and reference has come under pressure and where debate closely parallels that between descriptivists and causal theorists: this concerns the meaning of certain kinds of general terms, such as lemon, tiger, gold, water, which, reviving scholastic vocabulary, are said to stand for 'natural kinds'. On one view, stemming from Locke, the meaning of such terms is given by specifying salient properties of the natural kinds in question: 'a yellow citrus fruit with a bitter taste...,' and so on. This meaning will then determine the reference (extension) of the term, in the sense that the term will have in its extension, by definition, all those things which satisfy the descrip- tion. However, the very foundations of this view have come under attack, principally by Hilary Putnam and Kripke. Putnam (1975) argues that natural kind terms should be treated as rigid designators whose extension is determined not by clusters of descriptions but by the very structure of nature, as investigated by science. The initial application of a term—parallelling the
Philosophy of Language
7