Grice's "Implicature" and Literary Interpretation:
Background and Preface
Twentieth Annual Meeting
Midwest Modern Language Association
Minneapolis, Minnesota, 2-4 November 1978

A vast literature has complicated the theory of implicature since 1978. Anyone pursuing such questions now should consult Penelope Brown and Stephen C. Levinson, Politeness: Some Universals in Language Use (Cambridge: Cambridge Univ. Press, 1988); Dan Sperber and Deirdre Wilson, Relevance: Communication and Cognition (Cambridge: Harvard Univ. Press, 1986); Georgia M. Green, Pragmatics and Natural Language Understanding (Hillsdale, NJ: Erlbaum, 1989), 87-125; and Grice's own Studies in the Way of Words (Cambridge: Harvard Univ. Press, 1989), which includes his articles and lectures, discussed below.

Grice died in 1988.

Michael Hancher, Department of English
University of Minnesota, Minneapolis MN 55455

These remarks are preliminary to the forum,

H. P. Grice (b. 1913), formerly a Fellow of St. John's College, Oxford, now teaches philosophy of language at the University of California, Berkeley. Two aspects of Grice's work are particularly relevant to literary interpretation: his theory of nonnatural meaning, and his theory of conversational implicature.

MEANING. In a series of influential and controversial papers (Grice 1957, 1968, 1969), Grice has argued that the meaning of a word (or nonnatural sign) in general is a derivative function of what speakers mean by that word in individual instances of uttering it. That is, the universal "type" meaning, or set of such meanings, for a given word is an abstraction from the "token" meanings that speakers mean for the word in specific instances of use.

Among other things, this account opposes the formalist orthodoxy in semantic theory, according to which the universal conventional meaning (or set of meanings) of a word predetermines what that word might mean in any given instance of use. The conventional theory discourages inquiry into what a particular speaker might mean by a word in a particular utterance; to understand the utterance it is enough to know what the word "means" tout court. But Grice holds that what a word "means" derives from what speakers mean by uttering it; and he further holds that "what a particular speaker or writer means by a sign on a particular occasion . . . may well diverge from the standard meaning of the sign" (Grice 1957: 381).

Grice's analysis of verbal meaning in terms of the speaker's intentions has become increasingly elaborate, as he and others have revised it (Strawson 1964, Schiffer 1972). This elaboration has become one ground of objection (Black 1973, MacKay 1972). Ziff (1967) has published an influential critique of the theory, which Patton and Stampe (1969) subject to a detailed rebuttal. Searle (1969: 42-50) incorporates Grice's analysis into his own speech-act model, but only after radical revision to suppress the unconventional aspect.

Grice's arguments on behalf of the speaker's intentions have an obvious bearing on some perennial questions in literary theory; see R. Brown (1974), Hancher (1972), Hirsch (1975).

IMPLICATURE. A few years after publishing his original paper on meaning, Grice sketched out a theory of pragmatic implication, distinct from semantic implication, as a tool for resolving certain linguistic problems in the theory of perception (Grice 1951). Consider an utterance such as:

(1) That box looks red to me.

Grice denied that it is simply by virtue of the semantics of the phrase "looks to me" that uttering (1) implies the acknowledgement of some doubt or denial of the box being red Rather, such an implication arises from "a general feature or principle of the use of language." Grice roughly formulated that principle as, "One should not make a weaker statement rather than a stronger one unless there is a good reason for so doing."

It is the hearer's tacit knowledge of such a principle governing the speaker's use of language, rather than of any peculiar semantic features (or, for that matter, pragmatic features) of the phrase "looks to me," that enables him to infer, on hearing the speaker say (1), that the speaker means to acknowledge by implication that some doubt has been cast on the box's being red.
Similarly for an utterance such as:

(2) Rudy is either in Minneapolis or in St. Paul.

It is the hearer's tacit knowledge of the general pragmatic principle discouraging "weak" statements, rather than of any special semantic or pragmatic features of the word "or," that enables him to infer, on hearing the speaker say (2), that the speaker means to imply that he does not know in which of the two cities Rudy is. For, if the speaker had known which city, he "ought" to have said which, according to the proposed quasi-ethical principle of language use. By not saying which city, the speaker implies (whether truly or falsely) that he does not know which.

Such general pragmatic implication is "cancellable": that is, the speaker can go on to say something that cancels the apparent implication. You can't do that with semantic implication. For example, to say that someone has "stopped" doing something is to imply, through the semantics of the verb, that he once did it; and that implication is not cancellable. You can't say:

(3) * John has stopped smoking, though he never did smoke.

But the general pragmatic implication of (2) is cancellable. You can say, for example, (4) Rudy is either in Minneapolis or St. Paul; I know which, but I won't tell you.

The uncertainty implied by (2) is cancelled in (4) without anomaly.

Grice filled out this sketch of general pragmatic implication in "Logic and Conversation," the series of seven William James Lectures that he delivered at Harvard University in 1967-68. These lectures, which circulated widely in typescript (Grice 1967), have strongly influenced recent work in linguistics and the philosophy of language; representative essays can be found in Cole and Morgan (1975), Rogers, Wall, and Murphy (1977), Cole (1978), and the annual volumes published by the Chicago Linguistics Society. Gazdar, Pullum, and Klein (1977: E10-12) list additional items. Gradually portions of the lectures themselves are appearing in print. Grice (1969) includes substantial portions of the fifth and sixth lectures. Grice (1975a) and Grice (1975b) each present the pivotal second lecture; Grice (1978) is based on the third. After revision the whole series is supposed to be published by Harvard University Press.

In the second lecture, Grice (1975a, 1975b) proposes that participants in conversation understand the following general "Cooperative Principle" (abbreviated CP) to be in force: "Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged." This superordinate principle comprises the following subordinate rules or "maxims":

I. Maxims of Quantity: 1. Make your contribution as informative as is required (for the current purposes of the exchange). [Note: this is the strongest-statement-possible principle first proposed in Grice (1961), and cited above.] 2. Do not make your contribution more informative than is required.

II. Maxims of Quality: Supermaxim: Try to make your contribution one that is true. 1. Do not say what you believe to be false. 2. Do not say that for which you lack adequate evidence

III. Maxim of Relation: Be relevant.

IV. Maxims of Manner: Supermaxim: Be perspicuous. 1. Avoid obscurity /of expression. 2. Avoid ambiguity. 3. Be brief (avoid unnecessary prolixity). 4. Be orderly.

Because conversation is a cooperative and social enterprise, children are instilled with these imperatives as part of the process of socialization and language acquisition. Grice would argue further that observing the CP and maxims is "reasonable (rational)" behavior, because it tends to benefit the speaker's interest. In any case, the ability to realize these imperatives is an important part of a speaker's communicative competence (Bates 1976). The result is that a violation of any of these maxims will be linguistically aberrant, or "marked," and literally "remarkable." (Once the violation is detected, that is; some violations are surreptitious.)

Faced with a speaker's violation of a maxim, a competent hearer will draw one of several possible conclusions, depending on the particular case:

A. The speaker is openly "opting out" from the operation of the maxim and the CP. A famous case is Gordon Liddy's persistent violation of the first maxim of Quantity, and repudiation of the CP along with it.

B. The speaker is deliberately and secretly subverting the maxim and the CP, for some usually selfish end Lying (covertly violating the first maxim of Quality) is one example of this.

C. The speaker means to observe the CP, but fails to fulfill a particular maxim through ineptitude. For example, he may ineptly use words too technical for the audience and occasion, inadvertently violating the first maxim of Manner. (Grice alludes to this general kind of violation only in passing.)

D. The speaker presumably means to observe the CP, and yet he obviously is violating a maxim; if he is not inept, he must mean something additional to what he is merely saying. For example, when asked what she thinks of a new restaurant, a woman who replied, "They have handsome carpets," would appear to be flouting the first maxim of Quantity. If there is no reason in her case (unlike Gordon Liddy's) to doubt that she means to be observing the CP and is capable of doing so, then her remark must mean something other than what it literally asserts--such as, for example, that the food there is at best mediocre.

E. The speaker presumably means to observe the CP, and yet he obviously fails to fulfill a maxim. Perhaps he could not fulfill both it and another maxim as well; that is, perhaps there is a "clash" of maxims in these particular circumstances. Thus the speaker of (2) fails to fulfil the first maxim of Quantity, because to do so (i.e., to say which city Rudy is in) would, under the circumstances of his not knowing which, infringe the second maxim of Quality. So the speaker of (2), by violating one maxim, invokes another, and implies thereby that he lacks "adequate evidence" to say which city Rudy is in. (A few words about terminology: Grice indifferently uses the term "violate" to characterize, in particular, the activity described in B above, and also, in general, any failure to fulfill a maxim [Grice 1975a: 49-52]; I use it in the latter sense throughout. And the notion of "invoking" the CP or a maxim is implicit in Grice, but the term is not his.)

These last two kinds of maxim-violation, which convey an unstated but meant meaning, are two kinds of what Grice calls "conversational implicature." By judiciously relying on the CP and maxims in such ways, speakers often succeed in communicating, by "implicating," more than what they say.

As in D, some implicatures flout a maxim so as to invoke the CP as a ground of interpretation. It is also possible to flout a maxim on the literal level (what is said) so as to invoke the same maxim at a figurative level (what is implicated). Grice (1975a: 49, 52) joins these two maneuvers in one general kind: each "exploits" a maxim. Irony and metaphor are two standard forms of maxim-exploiting implicature.

As in E, some implicatures flout a maxim so as to invoke another maxim as a ground of interpretation. There is a third general kind of implicature, which involves no maxim-violation at all, but simply invokes a maxim as a ground of interpretation. Thus if you say "I am out of gas," and I say "There is a gas station around the corner," my saying so im- plicates, by invoking the maxim of Relation, that I think it possible (at least) that the station is open and has gas to sell.

Besides these three kinds of "conversational implicature," Grice identifies a category of "conventional implicature," independent of the CP and its associated maxims; see Grice (1975a: 45), Kempson (1975: 145), and Katz (1972: 445-46) for discussions of this notion. He also distinguishes conversational implicatures that depend heavily on context or occasion ("particularized conversational implicatures") from those that do not ("generalized conversational implicatures"). The examples in D and E, respectively, happen to differ in this regard--though not because one turns on a clash, and the other on an exploitation.

Grice (1975: 49-50) outlines the general line of reasoning by which the hearer should be able to recover the "implicatum" (thing implicated) in any given case of conversational implicature. Evidently the conversational implicatum will be determinate (determined by the intentions of the speaker) in every case. But Grice acknowledges in passing (p. 58) that in some cases the hearer may be unable to rule out one or more possible interpretations; in that sense a particular implicatum may be indeterminate.

Grice's theory of conversational implicature has been variously attacked, defended, and revised by others. Keenan (1974), citing anthropological data, claims that Grice's conversational maxims are parochial, not universal; P. Brown and Levinson (1978: 298-99) argue to the contrary. Gordon and G. Lakoff (1971) try to formalize Grice' s theory so as to fit it within a generative-semantics grammar. R. Lakoff (1973, 1975, 1977) and P. Brown and Levinson (1978) would place it within a larger model of sociolinguistic "politeness."

The implications of Grice's model for literary and rhetorical theory have only begun to be explored. At the most basic level, Griffin (1977) notes that many reading-impaired children and adults have trouble reading because they fail to recognize conversational implicatures on the printed page as readily as they would if the words were spoken. That is, they can read the words, but not between the lines. No doubt the same thing can happen to more sophisticated readers of more sophisticated texts. Although Pratt (1977) does not specifically discuss failing to grasp a literary implicature, she does apply Grice's basic two-person model to the four-person structure of reported speech or fiction (author, reported or fictional speaker, reported or fictional hearer, reader), and explores the many ways in which the author of a literary text can implicate meanings through what he has his characters say. Hancher (1977: 1095-96) makes the further suggestion that much omniscient narration, by flouting the second maxim of Quality, implicates that the narrative is fictional.

As regards rhetoric, Grice himself notes that exploitative implicatures involve "something of the nature of a figure of speech." His own analysis of metaphor (1975a: 53) could use elaboration. R. Brown (forthcoming) provides such a full-dress analysis of irony.

Cooper (1977) proposes that the occurrence of conversational implicature is a variable feature of literary style, which can distinguish one literary genre from another, and one literary work from another. She also relates the playwright's device of dialogic plot-exposition to Grice's second maxim of Quantity, which it usually violates.

Finally, van Dijk (1976: 44-49) would define "literature" itself as discourse that systematically subverts Grice's Cooperative Principle and all its maxims. Plus ça change . . . .


Bates, Elizabeth. 1976. Language and Context: The Acquisition of Pragmatics. New York: Academic.

Black, Max. 1973. "Meaning and Intention: An Examination of Grice's Views." New Literary History, 4, 257-79.

Brown, Penelope, and Stephen Levinson. 1978. "Universals in Language Use: Politeness Phenomena." Questions and Politeness: Strategies ln Social Interaction, ed. Esther N. Goody (Cambridge: Cambridge Univ. Press), pp. 56-310.

Brown, Robert L., Jr. 1974. "Intentions and the Contexts of Poetry." Centrum, 2, 56-66.

---. Forthcoming. "The Pragmatics of Verbal Irony." Language Use and the Uses of Language, ed. Roger Shuy and Ralph Fasold (Washington, D.C.: Georgetown Univ. Press).

Cole, Peter, ed. 1978. Syntax and Semantics, 9: Pragmatics. New York: Academic Press.

---, and Jerry L. Morgan, eds. 1975. Syntax and Semantics, 3: Speech Acts. New York: Academic Press.

Cooper, Marilyn. 1977. "Implicatures in Fictional Conversations from Days of Our Lives and Mary Hartman, Mary Hartman." Centrum, 4, 5-19.

Gazdar, Gerald, Geoffrey K. Pullum, and Ewan Klein. 1977. "A Bibliography of Pragmatics." Pragmatics Microfiche, 2:5, D11-G13.

Gordon, David, and George Lakoff. 1971. Papers from the Seventh Regional Meeting of the Chicago Linguistics Society, 63-84. Rpt. in Cole and Morgan (1975), pp. 83-106.

Grice, H. P. 1957. "Meaning." Philosophical Review, 66, 377-88.

---. 1961. "The Causal Theory of Perception." Aristotelian Society, supp. vol. 35, 121-52. Rpt. in Perceiving, Sensing, and Knowing, ed. Robert J. Swartz (Garden City, N Y.: Doubleday--Anchor, 1965), pp. 438-72.

---. 1967. "Logic and Conversation" (William James Lectures, Harvard University). Typescript.

---. 1968. "Utterer's Meaning, Sentence-Meaning, and Word-Meaning." Foundations of Language, 4, 25-42.

---. 1969. "Utterer's Meaning and Intention." Philosophical Review, 68, 147-77.

---. 1975a. "Logic and Conversation." In Cole and Morgan (1975), pp. 41-58.

---. 1975b. "Logic and Conversation." The Logic of Grammar, ed. Donald Davidson and Gilbert Harman (Encino, Cal.: Dickenson), pp. 6475.

---. 1978. "Further Notes on Logic and Conversation." In Cole (1978), pp. 113-27.

Griffin, Peg. 1977. "Reading and Pragmatics: Symbiosis." Linguistic Theory: What Can It Say About Reading?, ed. Roger W. Shuy (Newark, Del.: International Reading Association), pp. 123-42.

Hancher, Michael. 1972. "Three Kinds of Intention." MLN, 87, 827-51.

---. 1977. "Beyond a Speech-Act Theory of Literary Discourse." MLN, 92, 1081-98.

Hirsch, E.D., Jr. 1975. "Stylistics and Synonymity." Critical Inquiry, 1, 559-79. Rpt. in Hirsch's The Aims of Interpretation (Chicago: Univ. of Chicago Press, 1976), pp. 50-73.

Katz, Jerrold J. 1972. Semantic Theory. New York: Harper.

Keenan, Elinor O. 1974. "The Universality of Conversational Postulates." Studies in Linguistic Variation, ed. Ralph W. Fasold and Roger W. Shuy (Washington, D.C.: Georgetown Univ. Press), pp. 255-68. Also in Language in Society, 5 (1976), 67-80.

Kempson, Ruth M. 1975. Presupposition and the Delimitation of Semantics. Cambridge: Cambridge Univ. Press.

Lakoff, Robin. 1973. "The Logic of Politeness; or, Minding Your p's and q's. Papers from the Ninth Regional Meeting, Chicago Linguistics Society, 292-305.

---. 1975. Language and Woman's Place. New York: Harper.

---. 1977. "What You Can Do With Words: Politeness, Pragmatics, and Performatives." In Rogers, Wall, and Murphy (1977), 79-105.

MacKay, Alfred F. 1972. "Professor Grice's Theory of Meaning." Mind, 81, 57-66.

Patton, T. E., and D. W. Stampe. 1969. "The Rudiments of Meaning: On Ziff, on Grice." Foundations of Language, 5, 2-16.

Pratt, Mary Louise. 1977. Toward a Speech Act Theory of Literary Discourse. Bloomington: Univ. of Indiana Press.

Rogers, Andy, Bob Wall, and John P. Murphy, eds. 1977. Proceedings of the Texas Conference on Performatives, Presuppositions, and Implicatures. Arlington, Va.: Center for Applied Linguistics.

Schiffer, Stephen R. 1972. Meaning. Oxford: Clarendon Press.

Searle, John R. 1969. Speech Acts: An Essay ln the Philosophy of Language. Cambridge: Cambridge Univ. Press.

Strawson, P. F. 1964. "Intention and Convention in Speech Acts." Philosophical Review, 73, 439-60.

van Dijk, Teun A. 1976. "Pragmatics and Poetics." Pragmatics of Language and Literature, ed. Teun A. van Dijk (Amsterdam: North-Holland), pp. 23-57.

Ziff, Paul. 1967. "On H. P. Grice's Account of Meaning." Analysis, 28, 1-8. Rpt. in Understanding Understanding (Ithaca, N.Y.: Cornell Univ. Press, 1972), pp. 178-89.