On Reference — Definite Descriptions
18 Jan 2025This post was machine translated and has not yet been proofread. It may contain minor errors or unnatural expressions. Proofreading will be done in the near future.
4. Definite Descriptions
Unlike names, the primary question raised concerning definite descriptions is whether definite descriptions actually refer. Frege held that definite descriptions, like names, possess both sense (the content of the definite description) and reference (the object that uniquely satisfies that content), whilst Russell maintained that definite descriptions do not have referents.
4.1. Russell’s Theory of Descriptions
Russell’s Claim
According to Russell, definite descriptions do not refer to objects. Rather, definite descriptions are semantically equivalent to second-order predicates. In other words, just as we can think of a predicate P as
P: object → truth value
when D is a definite description, we have
D: (object → truth value) → truth value (i.e. predicate → truth value)
According to this view, “the present King of Britain” does not refer to Charles III, but rather is a second-order predicate that returns true for “x is male”, “x is British”, “x is the son of Elizabeth II”, etc., and returns false for “x is female”, “x is American”, “x is the son of James I”, etc.
The strength of Russell’s theory of descriptions lies in its ability to explain how definite descriptions lacking referents can be used meaningfully in sentences such as “The present King of France is bald” (according to Frege this sentence would be meaningless, but according to Russell this sentence is simply false), and to resolve Frege’s puzzle.
Furthermore, Russell claims that all names except ‘this’ and ‘I’ are disguised definite descriptions. Whilst this claim has the advantage of making the semantics of language elegantly simple, it fails to adequately explain Kripke’s modal and semantic arguments. Therefore, this essay shall examine only Russell’s claim that “definite descriptions do not refer”.
Remark. However, there remains doubt as to whether Russell’s theory of descriptions truly resolves Frege’s puzzle. If we adopt an extensional definition for predicates, then ‘the morning star’ and ‘the evening star’ become exactly the same (second-order) predicate. The arguments of Quine and others regarding intensionalism and extensionalism are also worth considering.
Strawson’s Theory
Consider the following sentence:
The table is covered with books.
This sentence appears to be true when, for instance, the speaker is in a room, there is exactly one table in the room, and that table is covered with books. However, according to Russell’s theory of descriptions, since there is more than one table in the universe, the sentence is false.
In response, Strawson argues that there are cases where definite descriptions genuinely refer. When a definite description is used referentially, the definite description refers to the object amongst those satisfying the descriptive content that is most salient in the given context. Conversely, when a definite description is used attributively, the description functions as in Russell’s theory.
Donnellan’s Theory
Consider the following sentence:
Who is the man drinking a martini?
Suppose that the man the speaker intends to refer to is actually drinking water, not a martini. In this case, according to Strawson, “the man drinking a martini” cannot be used referentially (since there is no man drinking a martini in the relevant context), nor can it be used attributively (since there are multiple men currently drinking martinis worldwide).
Therefore, Donnellan argues that definite descriptions can also refer to objects that do not satisfy their descriptive content. What matters is the speaker’s intention to designate a particular object with the definite description. Donnellan’s theory appears to fall into the problem of rigidity concerning definite descriptions, and being conscious of this, Donnellan appeals to Gricean theories of meaning.
Kripke’s Theory
Kripke argues that if we understand sentences such as “Who is the man drinking a martini?” pragmatically, we can maintain a strong Russellian position that definite descriptions never refer under any circumstances.
Kripke appears to hold that reference must be achieved purely through convention (though it might be more appropriate to understand this as an algorithm). For instance, Kripke’s theory of names precisely specifies, given a name $N$, the object that this name designates through causal theory. However, given that strong conventionalism about names fails to properly explain names with multiple bearers or the principle of impure indexicals, Kripke’s conventionalism lacks solid foundation, and consequently Kripke’s argument against Donnellanism is also weakly grounded.
Definite Descriptions v. Indefinite Descriptions
Recently, the very distinction between definite descriptions (the F) and indefinite descriptions (a F) has come under question. Not only has precisely defining the difference between them proved to be extremely challenging, but numerous languages have been reported that do not grammatically distinguish between the two. If the distinction between definite and indefinite descriptions proves to be unfounded, this could strengthen Russellian views that definite descriptions do not refer. (Alternatively, one would have to advance the highly controversial claim that indefinite descriptions also refer.)
5. Referential Scepticism
Descriptivism, Millianism, and indexical theories agree in their position that reference is a philosophically worthy subject of enquiry and that meaningful analysis thereof is possible. However, there also exist so-called “referentially sceptical” positions that argue reference is an essentially meaningless or vacuous concept.
Quinean Scepticism
Quine argues through the indeterminacy of translation thesis that referential expressions such as ‘gavagai’ are incomprehensible. Regarding the indeterminacy of translation thesis, the present author has written previously, which may be consulted.
The Problem of the Many
The problem of the many concerns how to define the boundaries of the object that a referential expression refers to. For example, I can refer to the present Gyeongbokgung Palace with ‘Gyeongbokgung’. Now imagine that one day an incident occurs in which the roof tiles of Gyeongbokgung are burnt away. Even after the fire incident, it seems I can still refer to Gyeongbokgung (now without roof tiles) with ‘Gyeongbokgung’. Can we then say that ‘Gyeongbokgung’ before and after the fire incident has the same referent? How much of Gyeongbokgung must be burnt away before it can no longer be the referent of ‘Gyeongbokgung’? This problem raises doubts about the general schema that referential expressions have referents.
Davidson’s Theory of Truth
The reason philosophers pursue theories of reference is that they are necessary for establishing the semantics of language. That is, the referent of each expression appearing in a sentence must be determined in order to determine the truth value of that sentence.
However, Davidson argues that this puts the cart before the horse. Davidson claims that merely providing a theory of truth for sentences suffices to establish the semantics of the relevant language, and that reference theory is therefore unnecessary. Philosophical discussions of truth shall be organised in a separate series hereafter.