A defence of Fregean propositions

Stephen Schiffer 2003 presents six arguments against the Fregean model of propositions, according to which propositions are (a) the referents of that-clauses and (b) structured entities made out of concepts. Schiffer advances an alternative view: propositions are unstructured pleonastic entities. My purpose is to argue in favour of the main tenets of the Fregean model by countering each of Schiffer’s arguments and sketching the guidelines for a theory of concepts as basic components of propositions. 1. Schiffer (2003: 24–27) presents six arguments against the Fregean model of propositions. The Fregean model holds three main theses: (A) That-clauses are singular terms referring to propositions. (B) Propositions are structured entities made out of the referents of the expressions forming the that-clauses and the syntactic structure of those clauses. (C) Expressions occurring in that-clauses have concepts as their referents. Schiffer attacks the Fregean model by the following arguments: (1) We lack a theory of concepts as basic components of propositions. (2) The Fregean model does not explain how propositions get their truth conditions. (3) There are cases where it seems that certain singular terms occurring in that-clauses cannot but refer to their ordinary referents. (4) The Fregean model generates a hierarchy of concepts: concepts, concepts of concepts, concepts of concepts of concepts and so on. It is to be explained what such concepts are and what it is to grasp them. (5) Sometimes reference to concepts as building blocks of propositions seems highly implausible, because it implies that there are concepts shared by most, if not all, thinkers. (6) The sixth argument goes like follows: 1 In this paper I use ‘concept’ as synonymous with ‘sense’. Massimiliano Vignolo 40 (i) If the Fregean semantics is correct, then (a) ‘Fido’ occurs in ‘Ralph believes that Fido is a dog’ as a singular term whose referent is a concept of Fido. (ii) If (a), then the following inference (Inf) is valid: (Inf) Ralph believes that Fido is a dog ∴ ∃x(x is a concept & Ralph believes that x is a dog). (iii) But the inference is not valid; given the truth of the premise, the conclusion is also true only in the unlikely event that Ralph mistakes a concept for a dog. (iv) ∴ The Fregean semantics is not true. In what follows I will try to counter each of these arguments. I will start commenting on argument (6), then arguments (4), (1), (3), (5) and (2) in this order. 2. In this section I will respond to argument (6). My first remark is that the premise of (Inf) is ambiguous. It allows for a de dicto reading and for a de re reading. My claim is that in both cases argument (6) is not sound. If we construe the premise of (Inf) as a de dicto belief, then step iii) is false. If we construe the same premise as a de re belief, then step i) is false. If we construe the premise as a de dicto belief, then ‘Ralph believes that Fido is a dog’ is true if and only if Ralph stands in the believing relation to the proposition that Fido is a dog. According to the Fregean model, this proposition is made out of the concept of Fido and the concept of being a dog. Accordingly, the logical form of the premise of (Inf) is Bel(Ralph, 〈CFido, Cbeing a dog〉). The logical form of the conclusion of (Inf) is ∃x(x is a concept and Bel(Ralph, 〈x, Cbeing a dog〉). The conclusion is true if and only if there is a concept that, together with the concept of being a dog, forms the proposition that Ralph believes. In quantifying in that-clauses of de dicto beliefs, variables range over concepts. Therefore, according to the Fregean model, if it A Defence of Fregean Propositions 41 is true that Ralph believes de dicto that Fido is a dog, it is true that there is a concept that, together with the concept of being a dog, forms the proposition that Ralph believes. The advocate of the Fregean semantics can reply to Schiffer’s objection by the following counter-argument: (i*) If the Fregean semantics is correct, then (a) ‘Fido’ occurs in ‘Ralph believes that Fido is a dog’ as a singular term whose referent is a con-