In the case of a " empty " but meaningful term like the word unicorn, we show, following Frege, that there is indeed an entity to which this general term refers, namely the property of being a unicorn, even though there is no object that possesses this property. In other words, theextension of the term is empty, but this does not mean that the expression is devoid of reference.
A second type of apparent counterexample to the " Fido "-Fido theory concerns true negative existential statements, such as Vulcan does not exist. Either there is an object to which the proper name " Vulcan " refers, and the statement (that this object doesn't exist) should be false. Or there is no object, but then the statement should be meaningless (since meaning reduces to reference, in the " Fido "-Fido theory, and the subject term fails to refer). Russell's response to this type of counterexample can be reformulated by means of the notion of coercion derived from contemporary semantics: in the context of an existential statement, the proper noun that (normally) designates an object acquires the semantic value of a description, so that, in Russell's theoretical framework, the counterexample disappears.
This theory of disguised descriptions applies to other putative counterexamples to the " Fido "-Fido theory, such as the use of singular terms devoid of reference in statements attributing thoughts or words (Le Verrier thought the discovery of Vulcan would make him famous); Frege's analysis of such statements can itself be reformulated in terms of coercion. Although these analyses, both Russell's and Frege's, violate the principle of " semantic innocence " accepted by many contemporary philosophers of language, it is shown that this principle must be rejected since, both in existential statements and in statements attributing thoughts or words, the use of meaningless " non-words " (such as the word borogrove in Lewis Carroll's poem Jabberwocky ) does not prevent the overall statement from having meaning (Borogroves don't exist ; Jean believes that borogroves grow in the garden). The word borogrove in these statements is, as it were, enclosed in quotation marks (The " borogroves " do not exist ; John believes that in the garden grow " borogroves "). Instead of saying, following Russell, that borogrove acquires a metalinguistic descriptive content through coercion (a thing called " borogrove "), we can appeal here to the theory of " polyphony ", according to which the use of a term is sometimes reinterpreted as the simulation of someone else's use of that term. This theory, applied to the use of empty nouns in existential statements and propositional attitude statements, also preserves the " Fido "-Fido theory: the coexistence of several points of view within a single statement is a complication that the " Fido "-Fido theory entirely disregards, and therefore cannot be held against it.
