"Reference" is usually conceived as the central relation between language or thought and the world. To talk or think about something is to refer to it. Twentieth-century philosophy found such relations particularly problematic.
One paradigm of reference is the relation between a proper name and its bearer. On a more theoretical conception all the constituents of an utterance or thought that contribute to determining whether it is true refer to their contributions (as, for example, a predicate refers to a property).
In analytic philosophy discussion of reference was dominated until the 1960s by the views of Gottlob Frege and Bertrand Russell and modifications of them (such as those by P. F. Strawson). Criticisms of assumptions common to those views then provoked a revolution in the theory of reference. The alternatives include causal and minimalist theories.
Objections to Descriptivism
|Objections to Descriptivism|
One model of reference is that of descriptive fit. The paradigm is a definite description (such as "the tallest tree") that refers to whatever it accurately describes. Frege and Russell assimilated the reference of ordinary proper names to this case by supposing that speakers associate them with descriptions.
Similar accounts were later given of mass terms (such as "blood"), natural-kind terms ("gorilla"), and theoretical terms in science ("inertia"). It was conceded that most terms are associated with vague and context-dependent clusters of descriptions and that reference might be to whatever they least inaccurately described, but such liberalizations did not challenge the underlying idea that descriptive fit determines reference.
However, Keith Donnellan, Saul Kripke, and Hilary Putnam proposed counterexamples to that idea. Suppose, for instance, that speakers associate the name "Jonah" with the Bible story. Traditional descriptivism concludes that the sentence "Although Jonah existed, those things happened only to someone else" is untrue. For if one person satisfied the relevant descriptions, "Jonah" would refer to him.
But then descriptivism proves too much, for philosophical reflection cannot show that the Bible story is not a mere legend that grew up about a real person; if those things really happened to someone else, of whom no word reached the biblical writer, the name "Jonah" would still refer to the former, not the latter. Similarly, traditional descriptivism permits someone who thinks of gorillas primarily as ferocious monkeys to conclude falsely that the sentence "Gorillas exist, but they are not ferocious monkeys" is untrue.
A second criticism was this. Say that a term t rigidly designates an object x if and only if t designates (refers to) x with respect to all possible circumstances (except perhaps for circumstances in which x does not exist).
Most descriptions designate nonrigidly: "the tallest tree" designates one tree with respect to present circumstances, another with respect to possible circumstances in which the former is outgrown. The descriptions that traditional descriptivists associated with names were nonrigid.
However, names designate rigidly: Although we can envisage circumstances in which the Danube would have been called something else instead, we are still using our name "Danube" to hypothesize circumstances involving the very same river. Thus, most descriptions do not behave like names.
The second criticism was met by a modification of descriptivism. The descriptions associated with a name were rigidified by a qualifying phrase such as "in present circumstances." "The tallest tree in present circumstances" rigidly designates what "the tallest tree" nonrigidly designates.
The first criticism is less easily met. Some descriptivists used deferential descriptions such as "the person referred to in the Bible as 'Jonah.'" A more general strategy is to exploit the success of any rival theory of reference by building that theory into the associated descriptions.
However, such moves jeopardize the connection between reference and speakers' understanding (a connection that descriptivism was intended to secure) as the descriptions that speakers supposedly associate with names become less and less accessible to the speakers themselves.
It is in any case clear that, as Russell recognized, not all reference is purely descriptive. If the sentence "It is hot now" is uttered at different times in exactly similar circumstances, associated with exactly the same descriptions, those descriptions are not what determines that it changes its reference from one time to the other.
The reference of a token of "now" is determined by the time of its production and the invariant linguistic meaning of "now," the rule that any such token refers to the time of its production. Similarly, the presence of an object to the speaker or thinker plays an ineliminably nondescriptive role in the reference of demonstratives such as "this."
THE KRIPKE-PUTNAM PICTURE. Kripke and Putnam proposed an alternative picture. Something x is singled out, usually demonstratively ("this river," "this kind of animal"). A name n, proper or common, is conferred on x ("Danube," "gorilla").
The name is passed on from one speaker to another, the latter intending to preserve the former's reference. Such intentions are self-fulfilling: n continues to refer to x. The beliefs that speakers would express in sentences containing n play no role in making n refer to x, so it can turn out that most of them are false.
The picture involves two kinds of deference. Synchronically, there is division of linguistic labor: Ordinary speakers defer to experts (as in deciding which animals "gorilla" refers to). Diachronically, later speakers defer to earlier ones in a historical chain.
Thus, reference typically depends on both the natural environment of the initial baptism (to fix the demonstrative reference) and the social environment of the later use. An individual speaker's understanding plays only a minor role. The account may be generalized (as to many adjectives and verbs).
The picture needs qualification. Gareth Evans pointed out that a name can change its reference as a result of misidentification, even if each speaker intends to preserve reference. What matters is not just the initial baptism but subsequent interaction between word and object. Such concessions do not constitute a return to descriptivism.
CAUSAL THEORIES. The Kripke-Putnam picture is often developed into a causal theory of reference, on which for n to refer to x is for a causal chain of a special kind to connect n to x. Such a theory goes beyond the original picture in at least two ways.
First, although that picture required later uses of n to depend causally on the initial baptism, it did not require the initial baptism to depend causally on x. Kripke allowed reference to be fixed descriptively (not just demonstratively), as in "I name the tallest tree 'Albie'"; he merely insisted that the description did not give the meaning of the name.
There is no causal connection between the name "Albie" and the tree Albie. Second, Kripke and Putnam did not attempt to define the notions they used in causal terms; the notion of an intention to preserve reference is not obviously causal.
Causal theories are often motivated by a desire to naturalize linguistic and mentalistic phenomena by reducing them to the terms of physical science. Such theories are therefore not restricted to proper names. Causal theorists will postulate that our use of the words "tall" and "tree" is causally sensitive to tallness and trees respectively, hoping thereby to explain the reference of "Albie."
One problem for causal theories is that any word is at the end of many intertwined causal chains with different beginnings. It is extremely difficult to specify in causal terms which causal chains carry reference. For this reason, causal theories of reference remain programmatic.
DIRECT REFERENCE. Consonant with the Kripke-Putnam picture, but independent of causal theories of reference, is the theory of direct reference developed by David Kaplan. A term t directly refers to an object x in a given context if and only if the use of t in that context contributes nothing to what is said but x itself. For Kaplan, proper names, demonstratives, and indexicals such as "now" refer directly. Ruth Barcan Marcus had earlier made the similar suggestion that proper names are mere tags.
The reference of a directly referential term may be determined relative to context by its context-independent linguistic meaning, as for "now"; the claim is that what "now" contributes to the proposition expressed by an utterance of "It is hot now" is not its invariant linguistic meaning but the time itself.
Although all direct reference is rigid designation, not all rigid designation is direct reference: "the square of 7" rigidly designates 49, but the reference is not direct, for the structure of the description figures in the proposition expressed by "The square of 7 is 49." On one view all genuine reference is direct, sentences of the form "The F is G" being quantified on the pattern of "Every F is G" (as Russell held); "the F" is neither a constituent nor a referring term.
If "Constantinople" and "Istanbul" have the same direct reference, the proposition (C) expressed by "Constantinople is crowded" is the proposition (I) expressed by "Istanbul is crowded," so believing (C) is believing (I), even if one would not express it in those words. Similarly, when a term of a directly referential type fails to refer, sentences in which it is used express no proposition. The view is anti-Fregean.
In suitable contexts Frege would attribute different senses but the same reference to "Constantinople" and "Istanbul" and a sense but no reference to an empty name; for him the sense, not the reference, is part of what is said or thought. Russell held that logically proper names are directly referential but concluded that ordinary names are not logically proper.
|mechanisms of reference|
The challenge to defenders of the direct-reference view is to explain away the appearance of sameness of reference without sameness of thought and absence of reference without absence of thought, perhaps by postulating senselike entities in the act rather than the content of thought. The theory of direct reference concerns content, not the mechanisms of reference.