Abstract
Strawson style counterexamples to Grice’s account of communication show that a communicative intention has to be overt. Saying what overtness consists in has proven to be difficult for Gricean accounts. In this paper, I show that a common explanation of overtness, one that construes it in terms of a network of shared beliefs or knowledge, is mistaken. I offer an alternative, collectivist, model of communication. This model takes the utterer’s communicative intention to be a we-intention, a kind of intention with a distinctive content that cannot be reduced to an intention in favor of an individual action. I show that the collectivist model can explain overtness in terms of a general feature of we-intentions, namely the requirement that the participants in a shared activity are to intend to act in accordance with meshing subplans.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Note the following about the terminology. [1] The collectivist model of communication acknowledges the existence of the audience’s contribution to a communicative action and, therefore, does not identify the action of communicating with the utterance act and does not identify communicative intention tout court with the speaker’s intention. To align with the prevalent usage, however, talk of an unqualified communicative intention is to be understood as concerning the utterer’s communicative intention. On the collectivist model, this intention is a we-intention, while on the individualist model, it is an I-intention. (For an explanation of how I am using the terms we-intention and I-intention see n. 21.) I am explicit when I want to talk of the audience’s communicative intention or a shared communicative intention. [2] Shared, collective, joint, cooperative are used interchangeably in the context of discussions of shared agency.
For the concept of an essentially intentional collective action type see (Ludwig 2013), especially pp. 4–8.
(Grice 1957) There are other types of information transfer that are called communicative. For instance, on an ordinary sense of the word communicate the honeybee’s dance communicates information about the locations of patches of flowers. But this is not the sense of communication which requires the communicator to mean something by her utterance. Here, our interest is only in the form of communication that entails non-natural meaning.
Jonathan Bennett advocated such a model when he suggested that instead of looking for necessary and sufficient conditions for speaker meaning we secure an explanatory base from which we can “easily range out and capture the rest of the territory” (1976, p. 23). On the cooperative model, the base is cooperative communication. Noncooperative forms of communication are understood as the sabotaging or extending of an established practice.
Grice’s target analysandum is utterer’s occasion meaning. However, Grice is usually understood as holding that an utterer means something by an utterance just in case he intends to communicate with some audience, so I take Grice’s analysis to apply to communicative intention.
In 1989 (p. 303), Grice says: “the deficiency in that proposal [i.e. adding a clause that prohibits the utterer from having deceptive intentions] was that it gave no explanation of why this was a reasonable condition to put into an account of speaker meaning.”
Another major type of response in the Gricean tradition is the so-called reflexive approach endorsed by Grice in (1957), (Harman 1974, 1977), (Bach and Harnish 1979) among others, and criticized by e.g. (Sperber and Wilson 1996) and (Recanati 1986). I believe that this approach is mistaken because it fails to account for the distinctive commitment to truthfulness an utterer has toward one’s addressee as opposed to someone who is openly witnessing a speech act. But I leave a detailed discussion of reflexive accounts to the side for reasons of space.
This assumption restricts our discussion to assertive utterances but the conclusion is general.
For example, in the broken hair-drier case, Mary does not intend that Peter know that Mary knows that Peter knows that Mary has an informative intention.
The description of the case relies on (Chant and Ernst 2008).
We calculate P and P c as follows. G1 received zero messages, so either (i) or (ii) obtains. The prior probabilities of (i) and (ii) are ε and ε(1 − ε), respectively. Dividing the credence between (i) and (ii) so that P: P c = ε: ε(1 − ε) we get that P = 1/(2 − ε) and P c = (1 − ε)/(2 − ε).
In general, if one of the generals (say, G2) has received n messages, he knows that there are two possible outcomes consistent with that—either (i) G2’s nth message did not get through or (ii) it did but G1’s (n + 1)th message did not get through. The prior probabilities for (i) and (ii) are ε(1 − ε)n−1 and ε(1 − ε)n, respectively. This gives us that the probability of (i) given G2’s evidence is P and the probability of (ii) is P c .
Here are two examples.
[1] “Making use of epistemic logic it is provable that whenever the smallest uncertainty of message delivery is present, common knowledge via communication is impossible. (van der Grijn 1994, p. 1, my emphasis).
[2] “[The Coordinated Attack Problem] is a case in which common knowledge is required for action but in which it is impossible for the agents to elevate a piece of information to the status of common knowledge. And this is in spite of the fact that agents can communicate with each other indefinitely, have the same interests, and consciously aim to coordinate their beliefs and actions.” (Chant and Ernst 2008, p. 553, my emphasis).
Grice’s focus in (1957) is explicitly on telling. See, for instance, the discussion of Herod’s showing Salome the head of St. John the Baptist.
If we focus on messages about the time of the attack, it seems more plausible that any attempt at communication is not rational.(Chant and Ernst 2008) contains a proof that in CAP common knowledge of the time of the attack is necessary for coordination. No matter how many confirmations reach the generals, once a confirmation inevitably fails to reach one of them, each can reason to the conclusion that it is better to not attack than to attack. If so, it is plausible that a rational G1 would not even attempt to send a message (for she can calculate that it will be futile). But this does not lend plausibility to the simple account. It only shows that in this case a rational agent can calculate that the (perlocutionary) point of a certain communicative action (i.e. coordination of attack times) cannot be achieved, not that communication is impossible.
In accordance with the gloss S&W give to the notion of degrees of manifestness, I treat it as picking out a purely psychological and not an epistemic (i.e. normative) scale. The notion of manifestness involves epistemic elements, for it involves what the environment provides sufficient evidence for accepting. This entails nothing about which propositions are likely to be actually accepted by the individuals in the environment. The notion of degrees of manifestness, as I understand it, is a measure of this—that is, a measure of how likely a manifest proposition is to be accepted by actual individuals with thus-and-such cognitive make-up.
It can be proven that no matter how many messages are exchanged the informative intention cannot become mutually manifest in CAP. Note also that while there may be garden variety counterexamples to the simple account (like e.g. a message in a bottle), the formal clarity of the CAP set-up is crucial in providing counterexamples to the account that involves a weaker notion like mutual manifestness.
I read “sufficient” in S&W’s “the environment provides evidence sufficient for adoption” of a proposition p normatively throughout. It can perhaps be given a non-normative, psychological reading, but I explain below why a nonepistemic reading is implausible.
Some remarks S&W make support (b) as the correct interpretative assumption. S&W consider a case in which (it is manifest to Peter that it is manifest to Mary)n that the phone is ringing. They say that, even though more complex propositions in the progression are not likely to be entertained, they are manifest since “there is no cut-off point at which these assumptions are more likely to be false rather than true; they remain manifest throughout, even though their degree of manifestness tends asymptotically toward zero” (p. 42, my emphasis).
The example is from (Searle 1990).
Note that collective action does not have to be intentional under any description (as e.g. when we pollute the environment together but not intentionally). Agents in the first case perform a collective action of e.g. running together, but not a collective intentional action.
The assumption that a group of agents shares an intention to J just in case each member of the group has a we-intention that they J may be questioned in the case of institutional groups. But this complication is tangential to our concern here, as we will be concerned with the agency of informal, non-institutional groups, for example, the ones we form when we have a conversation, walk together, etc.
Bratman also requires that (1) and (2) be common knowledge. I discuss why I omit this requirement below. Bratman takes the conditions in his analysis of shared intention to be jointly sufficient but allows that shared intention may be multiply realizable. See e.g. Bratman (1999a, pp. 143–144; 2009, p. 155).
This is a minimalist sketch of the collectivist model. Further development must wait another occasion. The sketch, however, suffices for present purposes. It is developed enough to show how it solves an important problem and this, in turn, motivates a fuller development of it.
It might be suggested that communication can be effective even without cooperation on the part of the audience, if the utterer succeeds in capturing the audience’s attention without her will. But relying exclusively on attention capture to communicate would be difficult and inefficient, so this can hardly be a paradigmatic case of communication or a suggestion for how it is to be conceived as a going enterprise. I would suggest that this case is like a case in which you get someone to catch a ball by throwing it at him unexpectedly. This doesn’t show that playing catch isn’t a collective intentional activity. Mutatus mutandis for communication. In this paper, however, I put aside further discussion of this issue. My goal here is to show how the collectivist model helps us understand overtness of communicative intention. I leave for another occasion responding to objections of the sort just sketched (where we move away from paradigmatic communicative exchanges to other sorts that those make possible).
The former is the case if the production of the response is caused solely by the utterer’s action.
Too see this, note that the complex of intentions arrived at through Anne’s deception cannot perform the characteristic roles of shared intention. For example, perceived conflict or problems in carrying out the joint activity would not trigger the type of shared deliberation characteristic of shared agency. It could only trigger furtive sub-plan adjustment. Thus, if Bob runs out of paint, he can ask Anne for more only under false pretenses (given that his intention is that she not realize that he was looking at her leaving the paint at his door).
Grice (1969) formulates the characteristic feature of Strawsonian deception as follows. In each case there is an inference-element E such that U uttered x intending both that (i) the inference by which A reaches his response r should rely on E and (ii) that A should think U to intend that (i) be false. The description of the characteristic feature above differs from Grice’s in that it describes the deceptive plan as requiring not only that A should think that U intends that (i) be false but that A intends that U believe that (i) is false (i.e. that A intends that U believe that her deception, as U conceives of it, is successful). To see that this additional requirement is apt consider again S&W’s Strawson style example. Let us say that p is the proposition that Mary wants Peter to fix the drier. Mary’s deceptive plan is carried out just in case Peter, believing that p and believing that Mary intends that he does not believe that p, intends that Mary maintain the false belief that her plan was successful. If Peter lets Mary know that he has seen her lay out the parts and that he recognizes why she did it, Mary’s deceptive plan does not succeed. See, however, n. 36 for an explanation of how the collectivist account can accommodate deception cases in which U lacks the intention that A should intend that U believe that (i) is false.
It might be objected that Peter does not act in accordance with any plan in coming to believe that Mary wants him to fix the drier. He does not do any prior planning but simply finds himself presented with scattered parts. This is not an objection on the present use of “plan”. I understand intention as an attitude of a commitment to a plan of action. “Plan” here means simply a sequence of action types. A plan is something that can be a product of explicit planning, but need not be.
This discussion depends on treating (2) as a necessary feature of Strawson style deception. This was crucial in explaining why U does not intend that he and A act in accordance with meshing subplans, since U does not intend that the intention A has in (2) be satisfied. One may wonder whether a deceptive intention that included (1) but not (2) would be a we-intention (and a communicative intention). Consider the dryer case again. Suppose that Mary conspicuously arranges the parts, that Peter reasons in the way she intends him to, and that he as a result comes to believe that she wants him to fix the dryer. But Peter does not intend to keep it hidden from Mary that he saw her arranging the parts. He intends to tell her that he saw her and that he is happy to fix her dryer. But he simply doesn’t get to. Suppose that this state of affairs satisfies Mary’s deceptive intent. In such a case, I do not think that we should say that Mary’s intention in arranging the parts was a communicative one or that Mary and Peter acted together intentionally in getting Peter to have the belief in question. But, we cannot invoke the explanation above, as it is not the case that Mary is open to sidestepping any of Peter’s intentions. However, here Mary intends that Peter have a false belief about how the mesh is to be achieved. This suggests that parties that share an intention to J need not only to intend to J in accordance with meshing subplans but also that no party should intend that the others have a significant misconception about how she intends to achieve the mesh. (What counts as “significant” misconception may be a matter of degree. A certain amount of sneakiness may not undermine a shared intention.) Requiring something along the lines of what is proposed in (Kutz 2000, p. 6)—that the participants ought to have “dispositions favorable to mutual manifestness”—would suffice to rule out shared intention in these deception cases. Thanks to the anonymous referee for raising this issue.
My thanks to Michael Bratman and Kirk Ludwig for help with this.
This discussion treats the intended mesh as a necessary component of a shared intention. Bratman, however, claims that his conditions are merely jointly sufficient. (See n. 28.) In that case, he says, “the indicated web of interlocking intentions was one important case of shared intention” (1999a, p. 144). If intended mesh is not necessary for shared intention, this discussion is limited to the important case of shared intention identified by Bratman and exemplified by typical cases of e.g. intending to go to NYC together, take a walk together, sing a duet, etc. of which communication seems to be a paradigmatic instance.
See Chant and Ernst (2008).
Note that, with shared intentional activities in which the participants’ contributions are not simultaneous, the initiator doesn’t have to believe that it is possible that other members intend to J at the time of his contribution to J-ing. It is enough that he thinks it is possible that they will come to intend to J at the time at which they are to make a contribution. Thus, we can sing a song together in virtue of my singing the first part simply hoping that you will catch on in time to sing the second part.
Actual communicative intentions are more complex because communication is not just any (joint intentional) way of imparting a belief. For example, the appropriate way will restrict the utterances to actions that are not natural signs of the belief content. But (i1) and (i2) serve our purpose and simplify the discussion as there is no reason to think that introducing the relevant restrictions will make it impossible to impart a belief in such a way in CAP.
Bratman’s account of shared intention entails that if U and A share an intention to J, each intends that they J. It follows that, when there is no prearranged topic, A and U cannot share an intention to communicate that p (for some specific p). The audience cannot intend that they communicate that p when she does not know in advance what signal the utterer is going to produce. (Thanks to Karen Lewis for bring this up in her prepared comments at the 2012 Central APA meeting.) We bracket this concern here and assume that the generals have a time of the attack as a prearranged topic in stating the content of their intentions. This does not affect the overall point. I maintain that in cases with no prearranged topic the utterer and the audience share an intention to communicate something or other, that is, to engage in a determinate of the determinable type communicate. The formal implementation of this idea does not fall in the scope of this paper. Informally, we consider a typical communicative intention to be shared at the level of a determinable, partial plan. The utterer fills in this partial plan by making a specific utterance and that determines the audience’s part. This is exactly analogous to e.g. playing tic-tac-toe or chess (and, more generally, to engaging in EICs in which participants take turns and in which the contributions at each turn constrain subsequent contributions), where the joint intention (prior to the first move) is to play a game of the type characterized by the rules of tic-tac-toe and where each move constrains the choices of the subsequent moves for each player.
Note, however, that for Gricean accounts the main motivation for iterated epistemic conditions came from Strawson style cases. With the collectivist model there is no parallel motivation—we can rule out Strawsonian deception without introducing any epistemic conditions on the utterer’s intention. It is of course still open that the iterated conditions are a general requirement on a shared intention.
References
Bach, K., & Harnish, R. (1979). Linguistic communication and speech acts. Cambridge: MIT Press.
Bennett, J. F. (1976). Linguistic behaviour. Cambridge: Cambridge University Press.
Bratman, M. E. (1984). Two faces of intention. Philosophical Review, 93(3), 375–405.
Bratman, M. E. (1993). Shared intention. Ethics, 104(1), 97–113.
Bratman, M. E. (1999a). I intend that we J. In Faces of intention (pp. 142–163). Cambridge: Cambridge University Press.
Bratman, M. E. (1999b). Shared cooperative activity. In Faces of intention (pp. 93–108). Cambridge: Cambridge University Press.
Bratman, M. E. (2009). Modest sociality and the distinctiveness of intention. Philosophical Studies, 144(1), 149–165. doi:10.1007/s11098-009-9375-9.
Chant, S. R., & Ernst, Z. (2008). Epistemic conditions for collective action. Mind, 117(467), 549–573. doi:10.1093/mind/fzn033.
Gilbert, M. P. (1990). Walking together: A paradigmatic social phenomenon. Midwest Studies in Philosophy, 15(1), 1–14.
Grice, H. P. (1957). Meaning. The Philosophical Review, 66(3), 377–388. doi:10.2307/2182440.
Grice, H. P. (1969). Utterer’s meaning and intention. Philosophical Review, 78(2), 147–177.
Grice, H. P. (1989). Meaning revisited. In P. Grice (Ed.), Studies in the way of words (pp. 283–303). Cambridge: Harvard University Press.
Halpern, J. Y. (1986). Reasoning about knowledge: an overview. In Proceedings of the 1986 conference on theoretical aspects of reasoning about knowledge (pp. 1–17).
Harman, G. (1974). Review of Meaning by Stephen R. Schiffer. The Journal of Philosophy, 71(7), 224–229. doi:10.2307/2025349.
Harman, G. (1977). Review of linguistic behaviour by Jonathan Bennett. Language, 53(2), 417–424.
Kutz, C. (2000). Acting together. Philosophy and Phenomenological Research, 61(1), 1–31.
Ludwig, K. (2007). Collective intentional behavior from the standpoint of semantics. Nous, 41(3), 355–393.
Ludwig, K. (2013). Proxy agency in collective action. Nous. doi:10.1111/nous.12013.
Mele, A. R. (1992). Recent work on intentional action. American Philosophical Quarterly, 29(3), 199–217. doi:10.2307/20014416.
Recanati, F. (1986). On defining communicative intentions. Mind & Language, 1(3), 213–241. doi:10.1111/j.1468-0017.1986.tb00102.x.
Rubinstein, A. (1989). The electronic mail game: strategic behavior under “almost common knowledge”. The American Economic Review, 79(3), 385–391. doi:10.2307/1806851.
Schiffer, S. R. (1972). Meaning. Oxford: Clarendon Press.
Searle, J. R. (1990). Collective intentions and actions. In P. Cohen, J. Morgan, & M. E. Pollack (Eds.), Intentions in communication (pp. 401–415). Cambridge: MIT Press.
Sperber, D., & Wilson, D. (1996). Relevance: Communication and cognition. Oxford: Wiley.
Strawson, P. F. (1964). Intention and convention in speech acts. Philosophical Review, 73(4), 439–460.
Tuomela, R. (1991). We will do it: An analysis of group-intentions. Philosophy and Phenomenological Research, 51(2), 249–277.
van der Grijn, F. (1994). (Im)possibility of a coordinated attack ILLC Publications: Preprints and Other Reports: Universiteit van Amsterdam.
Acknowledgments
I would like to thank Michael Bratman, Gary Ebbs, Mark Kaplan, Karen Lewis, Matthew Stone, Steven Wagner, Joan Weiner, and an anonymous referee for this journal for helpful comments. I owe special thanks to Kirk Ludwig for discussions and written comments on multiple versions of this paper. I developed large parts of the critical portion of this paper at the International Summer School in Cognitive Sciences and Semantics in Riga, July 2012. My work there benefited from various conversations with the participants.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Jankovic, M. Communication and shared information. Philos Stud 169, 489–508 (2014). https://doi.org/10.1007/s11098-013-0205-8
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-013-0205-8