Conversational implicature, communicative intentions, and content

Abstract

At the core of the Gricean account of conversational implicature is a certain assumption concerning the phenomenon that its proponents hope to explain and predict – namely, that conversational implicatures are, essentially, cases of speaker meaning. Heck (2006), however, has argued that once we appreciate a distinctive kind of indeterminacy characteristic of particularized implicatures, we must reject this assumption. Heck's observation is that there are cases where it is clear a speaker has conversationally implicated something by her utterance, but where there is no particular proposition – other than what the speaker said – that we can plausibly take the speaker to have meant, or intended to communicate. I argue that although Heck's observation is ultimately not in conflict with the core Gricean assumption, it is in tension with the widely held thesis that the things we mean and implicate are propositions. I sketch an alternative account of the things we mean and implicate – one that that accommodates the fact that in many cases of successful communicative exchanges, there is no particular proposition that the speaker intends to communicate.

Keywords::

• conversational implicature
• content
• communication
• meaning
• pragmatics

Notes

^1. For discussion of the role of this assumption in Grice's work on conversational implicature, see Neale (1992), and Davis (1998) and (2007). For skepticism concerning whether Grice in fact held the assumption see Saul (2001) and (2002). Though I disagree with Saul's interpretation of Grice for the reasons given in Davis (2007), I am sympathetic to the case she makes for the utility of a ‘normative’ notion of implicature that she develops in those papers. See fn. 14.

^2. A particularized implicature is a case in which a speaker implicates p by saying q, but where normally saying q is not a way of implicating p.

^3. For a discussion of such worries, see Davis (1998, 122–124). A central issue in Davis's critical discussion of Grice's account concerns an underlying disagreement regarding the kinds of intentions required for speaker-meaning. Davis argues that Griceans confuse speaker-meaning with intending to communicate and this confusion has led to undesirable consequences for the Gricean account of conversational implicatures. As against the Gricean, Davis (2003) argues that S meant that p just in case S performed a publicly observable action e with the intention that e be an undisguised indication that S has an occurrent belief that p (see (ibid. 54–56) for some important qualifications). Though I am sympathetic to Davis's suggestion, in Buchanan (2012) I argue that he has failed to provide sufficient conditions for speaker-meaning, and I offer some reasons for thinking that what is needed is an account that is more along the lines of (Meaning) below. See Davis (2013) for a response. This disagreement notwithstanding, even proponents of Davis's account of speaker-meaning will be able to find something of interest in what follows concerning the nature of our communicative intentions.

^4. The conditions in (Meaning) are inspired by the excellent discussion of the problems and prospects for Grice's original attempts at analyzing speaker-meaning in Neale (1992).

^5. Of course, the relevant feature could be just about anything, especially in non-linguistic cases. To borrow an example from Schiffer (1982), one agent might growl at another intending to communicate that she is angry, in part, in virtue of the fact that the sound she produced resembles that of an angry dog.

^6. See Grice (1957, 1989, 86–117) and Schiffer (1972, 1982). For critical discussions of the conditions in (Meaning) see Sperber and Wilson (1995, Chapter 2), Davis (2003, Chapters 4 and 5), and Schiffer (1987, Chapter 9).

^7. Proponents include Grice (1971), Harman (1976), Ross (2009), and Velleman (1989). See Setiya (2009), Section Five, and Holton (2009), Chapter Two for an overview of some of the problems and prospects for the B-I Constraint.

^8. A similar point holds for the expression of intentions in action: Q: ‘What are you doing?’; A: ‘I am reaching for a martini’.

^9. The B-I Constraint might also help explain why ‘it is rationally impermissible to intend an end while failing to intend what one regards as a necessary means to this end’ (Ross 2009, 243). See Ross for an excellent discussion of ‘cognitivist’ accounts of intentions and practical reason.

^10. See the discussion in Section Three of communicative ‘endeavorings’ inspired by Holton's critical discussion of the B-I Constraint in his (2009). As I understand the thesis to be a conceptual truth regarding intending, I take Holton's critical discussion of the B-I Constraint as establishing the need for intention-like states not constrained by belief in the way suggested. In what follows, I selectively borrow from Holton's discussion, but I encourage the reader to consult his book for his views on the topics. See footnote 19.

^11. See Bach and Harnish (1979) and Bach (2011) for discussion of this point.

^12. Sperber and Wilson (1986a, 1986b, 1987, 1995) also use the term ‘weak-implicatures’, but they mean something different by that expression than does Heck. At the core of their account is the principle that (roughly) the correct interpretation of a speaker's utterance is one that optimizes its relevance, where ‘an ostensive stimulus is optimally relevant to an addressee if and only if it has enough contextual effects to be worth his attention and puts him to no unjustified processing effort in accessing them’ (Sperber and Wilson 1987, 743). When S's audience ‘has to recover [p] in order to satisfy himself that the speaker has observed the principle of relevance’, S has ‘strongly’ implicated p (Sperber and Wilson 1986a, 383). However, ‘an utterance…can be given an interpretation consistent with the principle of relevance on the basis of different… sets of premises and conclusions has a wide range of weak implicatures’ (Sperber and Wilson 1986b, 549). For critical discussion of the relevance-theoretic approach, see Davis (1998, 73–74, and 98–107) from which I borrow the foregoing quotes (ibid. 73). Also see fn. 22.

^13. Notice, for example, that (a) could be true, and (c) false, even though in the actual world the dishes for their guest are the dishes they have been working on all afternoon.

^14. Cf. Saul's (2001) and (2002) discussion of ‘audience-implicature’. In her (2001) and (2002), Saul makes an important case in favor of also allowing for a more normative notion of ‘implicature’ – roughly put, a speaker implicates p in this normative sense if she has done what is required of her in uttering what she did (i.e., being cooperative, etc.) to make p available to her audience (even if they fail to entertain p).

^15. Here I am in agreement with McDowell's discussion of Heck's proposal; see (2006, 46–47).

^16. I am assuming that in order for an agent to entertain a disjunction, she must entertain each of the disjuncts.

^17. What is the relationship between this practical attitude of partial-belief and the quantitative notion of credence discussed by Bayesians and others? Holton claims that we might ‘think of credences as the partial beliefs that an agent would have if they were quite unconstrained by cognitive limitations; or, more plausibly, as the partial beliefs that they do have when their cognitive limitations are irrelevant’ [2009, 33–34]. Holton suggests that cases of simple, idealized betting behavior illustrate one kind of circumstance in which such cognitive limitations are ‘largely irrelevant’ (ibid., 34). Here, Holton has in mind the kind of idealized betting behavior that is sometimes appealed to in trying to illicit an agent's credence function – for example, an agent's responses to question concerning how much she would be willing to pay for chance to win a given prize, if a certain outcome obtains.

^18. Like Holton, I am somewhat skeptical that we can analyze or give a reduction of all-out belief in terms of formal, quantitative models of credence [2009, 25–28]. I am not, however, fully convinced that it cannot be done. See Sturgeon [2008], for example, for an encouraging discussion of the ‘Lockean thesis’ according to which all-out believing that p is analyzed in terms of having a credence in p that exceeds some (contextually determined, and potentially vague) threshold r.

^19. Holton (2009) suggests that there are two ways we might understand the notion of a ‘partial intention’:

We might say that an agent has a partial intention whenever they merely have a partial belief in its success. Or we might say that it is essential to partial intentions that they be only a proper part of an overall plan, i.e., that they be accompanied by alternative partial intentions to achieve the same end. (ibid., 35)

Holton pursues the latter strategy; we are pursuing the former. I suspect that we will ultimately need both notions in the theory of meaning and communication. In work in progress, I argue that appealing to Holton's favored notion of a partial intention helps us better understand the relationship between a speaker's illocutionary intentions in making an utterance and her overarching perlocutionary plan.

^20. I borrow the term ‘endeavoring’ from Bratman's critical discussion of the ‘simple view’ according to which doing A intentionally requires intending to do A (Bratman 1987). Bratman offers an example concerning an ambidextrous video game player, S, who is simultaneously playing two games, one with each hand, where the goal of each is to hit on a target on the screen with a missile. The games are linked in such a way that it is impossible to simultaneously hit the targets on screen A and on screen B, and S knows this. If S succeeds in hitting target A, we would say that she did so intentionally; likewise, for B. If the simple view were correct this would entail that S acted with the intention of hitting A, and also with the intention of hitting B. But if one can rationally intend to F & rationally intend to G, only if one can rationally intend to F & G, we should conclude that S rationally intended both to hit A and to hit B. But as she knows it impossible to hit both, she can't rationally intend any such thing. Bratman concludes that S, despite acting intentionally, acted with neither the intention to hit A or the intention to hit B. Though she didn't intend to hit A, and she didn't intend B, she ‘endeavored’ to hit both.

^21. Notice that this suggestion is compatible with the fact that in those cases in which your communicative intentions are open, your audience will likely entertain numerous propositions of the relevant type as a result of your utterance. Moreover, when your audience is appropriately sensitive to the openness of your communicative intentions, he will be sensitive to the fact that there are multiple, non-equivalent ways in which he might correctly interpret your utterance. In this sense, your audience's interpretation of your utterance should (in the relevant cases) also exhibit openness. I am inclined to think that when a speaker produces an utterance with open meaning intentions, her audience must be sensitive to the openness if she is to understand the utterance, but I will not argue for that here.

^22. The view sketched in this section is inspired (in part) by Sperber and Wilson's seminal (1986a and 1995) account of communicative intentions on which:

S means a set of propositions {I} by u iff, for some audience A, S produces u intending (i) to make {I} manifest, or more manifest, to A, and (ii) to make (i) mutually manifest to S and A. (ibid., 58 and 61)

where (1) a proposition is ‘manifest’ to an agent to the extent that she is ‘capable at that time of representing it mentally and accepting its representation as true or probably true’ (ibid.: 39), and (2) a speaker must have a representation of {I}, but not necessarily each of the members thereof – ‘any individuating description will do’ (ibid., 58). On this account, we give up both (Content) and (Meaning). While I sympathize with the motivations Sperber and Wilson offer for this account, their proposal is too weak. Suppose, for example, you ask me what there is to do in Budapest, and I throw a guidebook, which I have never read, into your lap. There is a set of propositions – the ones in the Budapest guidebook – such that by my production I have knowingly made it (and its members) more representable to you, and more mutually representable to us both. But in what sense, if any, did I meanthe propositions in the guidebook? Likewise, take some proposition in the book – say, that the hand of the first king of Hungary is in the Basilica in Pest. Is that proposition among the things I meant? What is needed, I submit, is that each proposition in the relevant set {I} be a ‘candidate’ in the sense sketched in Section Four, and that (as such) the speaker must (minimally) endeavor that each such proposition be entertained, and not merely entertainable.

^23. Earlier versions of this paper was presented were presented at the University of Barcelona, the University of St. Andrews, the University of Manitoba, and the University of Toronto during 2012–2013. I would like to thank the audiences on those occasions for extremely helpful feedback. In particular, questions from Derek Ball, Jacob Beck, Herman Cappelen, Esa Diaz-Leon, Nate Charlow, Josep Macià, Aidan McGlynn, Francois Recanati, Jennifer Saul, and Chris Tillman were especially helpful in me to get clearer on these issues. I am also deeply indebted to the editors of this volume, David Hunter and Gurpreet Rattan, as well as David Beaver, Josh Dever, Sinan Dogramaci, Hans Kamp, Manuel García-Carpintero, Bryan Pickel, Gary Ostertag, Mark Sainsbury, Neil Sinhababu, David Sosa, and Zsófia Zvolenszky for very helpful discussion. I am especially thankful to both Gary and Zsófia for the extensive comments they gave me on an earlier draft of this paper. All of the normal qualifications and disclaimers are in order. Much of this paper was completed during my stay at the University of Barcelona, funded by the Spanish Ministry of Education, during the 2011–2012 academic year. I would like to thank the members of the LOGOS group in Barcelona for their hospitality and the Ministry for its generous support.

Additional information

Notes on contributors

Ray Buchanan

Since receiving his PhD from New York University in 2008, Ray Buchanan has been an assistant professor in the Department of Philosophy at the University of Texas, Austin. His work to date has primarily focused on the questions of how, and to what extent, we can express our thoughts by our actions - linguistic, or otherwise.