Reference, Understanding, and Communication

Abstract

Brian Loar [1976] observed that, even in the simplest of cases, such as an utterance of (1): ‘He is a stockbroker’, a speaker's audience might misunderstand her utterance even if they correctly identify the referent of the relevant singular term, and understand what is being predicated of it. Numerous theorists, including Bezuidenhout [1997], Heck [1995], Paul [1999], and Récanati [1993, 1995], have used Loar's observation to argue against direct reference accounts of assertoric content and communication, maintaining that, even in these simple cases, the propositional contribution of a referring expression must be more than just its referent.

I argue here that, while Loar's observation is correct, the conclusion he and others have sought to draw from it simply does not follow. Rather, his observation helps to remind us of an important Gricean insight into the nature of communicative acts—including acts of speaker-reference—namely, that there is more to understanding a communicative act than merely entertaining what a speaker is intending to communicate thereby. Once we remember this insight, we see that the phenomenon to which Loar is calling our attention should actually be expected given the direct reference theorist's assumptions, together with independently plausible Gricean principles concerning how we make our referential intentions manifest in communication. More generally, the Gricean strategy for explaining the challenge posed by Loar cases suggests a novel way to account for certain crucial anti-direct reference intuitions—one requiring no modification of the original theory (e.g., no invocation of ‘descriptive enrichments’ as in Soames [2002]), thereby allowing for a direct reference account of what is asserted in utterances of ‘simple sentences’ such as (1).

Keywords:

• reference
• content
• communication
• pragmatics

Notes

¹ The foregoing is a direct reference account of what a speaker asserts, or literally means, by an utterance of simple sentences such as (1)–(4). While some theorists might seek to identify the ‘semantic content’ of an utterance with what a speaker asserts in producing it, many would not, preferring to reserve this label for the compositionally-determined, context-invariant meaning of the sentence-type uttered (or, perhaps, such a sentence-meaning evaluated with respect to a context). The distinction between ‘what is asserted’ and ‘semantic content’ is important, as many theorists who are sympathetic to a ‘direct reference’ account of the latter would maintain that what a speaker asserts in an utterance of simple sentence cannot (typically) be identified with a singular proposition.

² See, for example, Schiffer [1981: 50–1].

³ García-Carpintero [2000] offers Loar cases in favour of his ‘presuppositional account of reference-fixing’, a view which is, in many respects, similar to Récanati's view that we will consider shortly. García-Carpintero's account does not happily fit into the simple taxonomy that we are working with in this paper between DR and Fregean accounts. If, however, we are prepared to be liberal about what counts as a ‘presupposition’, there are ways of elaborating his account that would be congenial to the DR theorist who accepts the Gricean account of Loar cases to be discussed in §5. I refer the reader to García-Carpintero's discussion for details.

⁴ Case Three is a variant of an example from Heck [1995: 95].

⁵ The descriptivist will find considerable inspiration from Soames [2002] and [2008].

⁶ See Bezuidenhout [1997] for the former suggestion, and Paul [1999: 157–61] for the latter.

⁷ See Soames [2002: 39–50] for some such worries.

⁸ See Bach and Harnish [1979: 15], as well as Bach [1987], for a nice discussion of this requirement on communicative intentions.

⁹ See Bach and Harnish's discussion [1979: 84–9]. I thank an anonymous referee for the formulation of this point in the text.

¹⁰ See Grice's discussion of these ‘features’ (as he called) them in connection with Searle's putative counterexample involving the American Soldier captured by Italian troops in World War II in Grice [1969: 161–5]. I prefer Schiffer's recent terminology, so I’ll stick with it in the text.

¹¹ Theorists who agree on this much might disagree about what is required to have de re thoughts, including de re intentions. We can, thankfully, set aside these further debates for another day.

¹² For further discussion concerning speaker-reference see Bach [2006], and Schiffer [1981; forthcoming].

¹³ Bach continues: ‘It is rather like having a justified true belief that p without knowing that p’ [2006: 524]. Bach's claim that such cases—which I shall suggest Loar cases to be instances of—are, in some respects, analogous to Gettier cases is plausible. I will not, however, pursue this analogy since I would like to avoid the difficult task of specifying exactly what kind of ‘luck’ is involved in Gettier cases. Fortunately, the Bach-inspired diagnosis of Loar cases can be given without having to first get clear on the exact sense in which the speaker's audience is ‘lucky’. Thanks to an anonymous referee for discussion of this point.

¹⁴ Note that the Gricean idea that recognizing a communicative intention requires arriving at what the speaker means (or refers to) in a certain way neither is intended, nor should be taken, to undermine the Fregean view concerning the nature of object-dependent thoughts—namely, that thinking of an object requires thinking of it under some or other MOP.

¹⁵ The strategy that we have been pursuing differs in interesting respects from what one is apt to expect given the standard invocation of ‘Gricean pragmatics’ in the literature. Typically, when a theorist appeals to ‘Gricean pragmatics’, we expect her to be invoking the theory of conversational implicature (usually in the service of an attempt to explain away certain otherwise problematic intuitions for her favoured semantic theory). This has not been our strategy, since in Loar cases the speaker's audience is not plausibly confusing what she implicated with she asserted. Rather, the speaker's audience is, I claim, confusing what she asserted with a heterogeneous mixture of what was asserted and further information in the common ground the speaker intends to be relevant to her audience's recognition of her intentions.

¹⁶ This paper was completed during my stay at the University of Barcelona, funded by the Spanish Ministry of Education, during the 2011–12 academic year. I would like to think the members of the LOGOS group in Barcelona for their hospitality, and the Ministry for its generous support. I would also like to thank Manuel García-Carpintero, Josep Macià, Bryan Pickel, Gary Ostertag, François Récanati, Mark Sainsbury, Zsófia Zvolenszky, and two anonymous AJP referees for helpful discussion of these issues.