Published online by Cambridge University Press: 13 April 2010
1 See, e.g., Salmon, Nathan, Frege's Puzzle (Cambridge, MA: MIT Press/Bradford Books, 1983)Google Scholar; Barwise, J. and Perry, John, Situations and Attitudes (Cambridge, MA: MIT Press/Bradford Books, 1983)Google Scholar; and Soames, Scott, “Direct Reference, Propositional Attitudes, and Semantic Content,” in Propositions and Attitudes, edited by Salmon, N. and Soames, S. (Oxford: Oxford University Press, 1988), pp. 197–239.Google Scholar
2 Completeness would require applying this point to coreferential expressions in general.
3 See, e.g., Barwise and Perry, Situations and Attitudes.
4 See, e.g., Soames, “Direct Reference.”
5 See, e.g., Tye, Michael, “The Puzzle of Hesperus and Phosphorus,” Australasian Journal of Philosophy, 56 (1978): 219–24CrossRefGoogle Scholar, and Sainsbury, R. M., “On a Fregean Argument for the Distinctness of Sense and Reference,” Analysis, 43 (1983): 12–14.CrossRefGoogle Scholar
6 Salmon, Nathan, “Reflexivity,” in Propositions and Attitudes, p. 246.Google Scholar
7 Quine, W. V. O., “Quantifiers and Propositional Attitudes,” Journal of Philosophy, 53 (1956): 177–86.CrossRefGoogle Scholar
8 Transparent ascriptions are naturally expressed using the “believes of” locution, as in “Bush believes of Deep Throat that he behaved inappropriately when Reagan was shot.”
9 See Richard, Mark, Propositional Attitudes (Cambridge: Cambridge University Press, 1990), p. 126CrossRefGoogle Scholar, and Crimmins, Mark, Talk About Beliefs (Cambridge, MA: MIT Press/Bradford Books, 1992), p. 33Google Scholar, for earlier versions of this objection.
10 Schiffer, S., “The ‘Fido’-Fido Theory of Belief,” Philosophical Perspectives, 1 (1987): 445–80.CrossRefGoogle Scholar
11 See Peter Grice, “Logic and Conversation,” in Syntax and Semantics, Vol. 3, edited by P. Cole and J. Morgan (New York: Academic Press, 1975), pp. 41–58. It might seem that rejecting a single mechanism of pragmatically imparting information is a rather weak argument against the “naive” theory. Schiffer (“The ‘Fido’-Fido Theory of Belief,” p. 471) does suggest, however, that this will suffice because “it seems unavoidable that for the weak theorist that account [of pragmatic impartation] will follow the pattern adumbrated above.”
12 Moreover, Schiffer also argues that the sentence (1) “Florence believes that the ancient astronomers did not believe that Hesperus is Phosphorus” could not be true given the “naive” account of belief ascriptions and certain fairly uncontroversial assumptions. Since “Florence believes that the ancient astronomers believed that Hesperus is Hesperus” is, presumably, true, so is (2) “Florence believes that the ancient astronomers believed that Hesperus is Phosphorus,” according to the “naive” theory. (1) and (2) could both be true (assuming Florence is rational) only if she had distinct “modes of presentation” of the proposition that the ancient astronomers believed that Hesperus is Phosphorus. But since Florence realizes that “Hesperus” and “Phosphorus” both refer to Venus, she does not have two such distinct modes of presentation. The very data we are trying to account for seem to be undermined by the “naive” theory.
For another objection to the “naive” theory roughly in the same family as Schiffer's, see Recanati, Francois, Direct Reference (Oxford: Blackwell, 1993), p. 342.Google Scholar
13 See, e.g., Salmon, Nathan, “Illogical Belief,” Philosophical Perspectives, 3 (1989): 243–86CrossRefGoogle Scholar, and Soames, Scott, “Substitutivity,” in On Being and Saying, edited by Thomson, J. J. (Cambridge, MA: MIT Press, 1988), pp. 99–132.Google Scholar
14 Recanati, Direct Reference, p. 333.
15 Of course, this response to Schiffer's objection stands or falls on whether there are any grounds to believe that the maxim of faithfulness is a genuine conversational principle. Salmon, “Illogical Belief,” and Soames, “Substitutivity,” attempt to provide support for it.
16 Crimmins, Talk About Beliefs, pp. 33–34.
17 Actually, we may have to take f to be a function of other features of the context of utterance in addition to the complement sentence.
18 We had best not allow f(x) to be a function from complement sentences to the Russellian propositions they express if we are to have any hope of utilizing the implicated information in explanations of behaviour. After all, taking the literal meaning of ascription complements to be Russellian propositions was what got the “naive” theory in trouble in the first place.
19 A natural candidate for this function would be one whose value for a given complement sentence is its translation in the ascribee's idiolect, with appropriate adjustments if the complement contains any indexicals.
20 It is worth emphasizing that each of these latter two options still count as versions of the simple “naive” theory. What makes a theory simple or sophisticated is its account of the literal meaning of belief ascriptions. It does not depend on its account of what information is implicated.
21 Of course, most ordinary folk probably would not put it this way.
22 This latter challenge is reminiscent of Recanati, Direct Reference, p. 342.
23 Salmon, Frege's Puzzle.
24 Ibid., p. 111.
25 Strictly speaking, Salmon's analysis of “T believes that p” is “(∃x)(T grasps р by means of x & BEL [T, tpt-p, x])” (Frege's Puzzle, p. 111). If we assume that a principle like “(∃x)(BEL [T, tpt-p, x] → T grasps tpt-p by means of x)” holds for BEL, as we presumably should, then the simpler analysis I have offered is recoverable from the analysis Salmon in fact offers.
26 Salmon, “Reflexivity,” pp. 240–74.
27 As Salmon points out (“Reflexivity,” pp. 247–48), this is not quite true. If one takes propositions to be sets of circumstances, then the inference from (b) to (c) is valid, but if one takes them to be structured Russellian entities, it is not. The same problem does arise, however, for advocates of the latter sort of propositions if, as is plausible, it is true that believers tend to believe both the existential generalizations of their beliefs and certain simple conjunctions of things they believe.
28 Ibid., p. 243.
29 Of course, if Lois stood in the BEL relation to the proposition that she will directly inform Superman of Superman's danger with her note under a guise cognate with the sentence “Lois will directly inform Superman of Superman's danger with her note,” it might be reasonable to insist that there is a guise under which Lois stands in the BEL relation to the proposition that there is someone she will directly inform of his or her danger with her note.
30 According to Crimmins (Talk About Beliefs), for example, the values of f(x) are token representations in the heads of the people beliefs are ascribed to. Of course, precision would require that we take the denoted representation in a given context to be a function of contextual features in addition to the ascription complement.
31 Grice, “Logic and Conversation.”
32 Salmon, “Illogical Belief.”
33 Soames, “Substitutivity.”
34 Crimmins, Talk About Beliefs, p. 34.
35 Schiffer, S., “Belief Ascriptions,” The Journal of Philosophy, 89 (1992): 499–521CrossRefGoogle Scholar. Schiffer refers to what I have been calling the “two-factor view” as the “hidden indexical theory.”
36 Schiffer, “Belief Ascriptions,” pp. 507–508.