Linguistic Meaning is Conventional
To define the meaning of a sentence as the message or messages that the sentence is, or can be, used to convey is inadequate, because too inclusive. In order to exclude the innovative or idiosyncratic features of language use, we might reach for the notion of a rule of language: What it is for a sentence to mean that p is for there to be a rule saying that the sentence is to be used (or may be used) to convey the message that p. However, if a rule is something that is formulated explicitly (in language), then the proposal may just reintroduce the notion of linguistic meaning; and that would be unsatisfactory if the project is to define or analyze the notion of linguistic meaning in other terms. So, instead of the notion of an explicitly formulated rule we can make use of the notion of a convention, defined as a rationally self-perpetuating regularity (Lewis, 1969). The resulting proposal is that what it is for a sentence S to mean that p in the language of a given population is for there to be a convention in that population to use utterances of S to convey the message that p.
Linguistic Meaning is Compositional
The term theory of meaning can be applied to two very different kinds of theory. On the one hand, there are semantic theories that specify the meanings of the expressions of some particular language; on the other hand, there are metasemantic theories that analyze or explain the notion of meaning. We should expect the idea that meaning is compositional to be reflected in semantic theories. The way in which the meanings of sentences depend on the meanings of words and phrases should be revealed in a semantic theory by having the meaning specifications for whole sentences derived logically from more basic principles that specify the meanings of words and phrases.
Many features of the messages conveyed by the use of a sentence will not be seen simply as the results of contributions to meaning made by the words in the sentence—contributions that would be repeated in other sentences—but rather as the products of interaction between the meaning of the sentence and other background assumptions. (The study of this interaction is called pragmatics. See Davis, 1991.) It is true, for example, that a letter of reference that says only, "Mr. X's command of English is excellent, and his attendance at tutorials has been regular" is likely to convey the message that Mr. X is not a talented philosopher (Grice, 1975). But this message is not the logical product of the meanings of the words and phrases used. Rather, the letter writer is able to convey that message by relying on shared assumptions about what information would be relevant in the circumstances. (See Grice's early  proposals about pragmatics.)
Two Approaches to the Study of Meaning
These ideas, that meaning is conventional and compositional, can be seen at work in two important approaches to the study of linguistic meaning, on which this article focuses. One is Herbert Paul Grice's program for analyzing the concept of literal linguistic meaning in terms of psychological notions such as belief and intention (Grice, 1989). The other is Donald Davidson's project of illuminating the notion of meaning by considering how to construct compositional semantic theories for natural languages (Davidson, 1984).
Grice's Analytical Program
The Gricean analytical program can be regarded as having two stages (for overviews, see Avramides, 1989; Neale, 1992). The first stage aims to characterize a concept of speaker's meaning that corresponds, roughly, to the idea of conveying, or attempting to convey, a particular message (Grice, 1957, and other papers, 1989). The second stage then aims to use the concept of speaker's meaning, along with the notion of a convention, to build an analysis of literal linguistic meaning. (In fact, Grice himself did not introduce the notion of convention, but used a slightly different idea. See Grice, 1989; Lewis, 1969, 1975; Schiffer, 1972.)
The basic idea of the first stage of the program is that an agent who is attempting to convey a message—perhaps the message that it is time for tea—makes an utterance (which might or might not be linguistic in nature) with the intention that the hearer should come to believe that it is time for tea and should believe it, at least in part, in virtue of recognizing that this is what the utterer intends him or her to believe. The analysis of speaker's meaning was refined and complicated in the face of counterexamples (Grice, 1989; Strawson, 1964; Schiffer, 1972), but it retained the crucial feature of not itself importing the notion of literal meaning. This feature is shared by the analysis of convention as a rationally self-perpetuating regularity, and so the prospects are good that the analysis of meaning resulting from Grice's program can meet the requirement of noncircularity.
problems with grice's program
Grice's program does, however, face a number of serious objections. One problem concerns the application of the program to sentences that are never used at all—perhaps because they are too long or too implausible. Clearly, the Gricean analysis of literal meaning cannot be applied directly to these sentences. If we want to say that there is, nevertheless, a fact of the matter as to what unused sentences mean, then we seem bound to appeal to the meanings of the words and phrases from which unused sentences are built. But now we come to the most serious problem for the program, namely, how to analyze the notion of meaning as it applies to subsentential expressions.
Parties to a convention know what the relevant regularity is, and their belief that they and others have conformed to the regularity in the past gives them a reason to continue conforming to it. Thus, the Gricean program involves crediting speakers of a language with knowledge about regularities of use. While this is plausible in the case of the use of complete sentences, it is problematic when we move to subsentential expressions. Words and phrases are used in complete sentences, and they make a systematic contribution to the meanings of the sentences in which they occur. Regularities of use for words and phrases are regularities of contribution to the messages that sentences are used to convey. But spelling out in detail how words and phrases (and ways of putting them together) contribute to the meanings of complete sentences is a highly nontrivial project. So, it is not plausible that every speaker of a language knows what these regularities of contribution are.
The problem for the Gricean program is that it seems bound to attribute to ordinary language users knowledge that they do not really have. It may be that we can deal with this problem by invoking some notion of tacit (Chomsky, 1986) or implicit (Dummett, 1991, 1993) knowledge (Loar, 1981). But the dominant consensus—and the view of one of the most authoritative exponents of Grice's program (Schiffer, 1987)—is that the project of analyzing literal meaning in terms of intentions and beliefs cannot be completed.
Davidson and Truth-Conditional Semantics
Any metasemantic theory can be used to provide conditions of adequacy on semantic theories. Thus, consider the Gricean metasemantic proposal:
Sentence S means that p in the language of population G if and only if (iff) there is a convention in G to use utterances of S to convey the message that p.
And suppose that a semantic theory for a particular language L delivers as one of its meaning specifications:
Sentence S1 means (in L) that wombats seldom sneeze.
Then, according to the metasemantic proposal, one necessary condition for the correctness of the semantic theory is that there should be a convention in the population of L-speakers to use utterances of S1 to convey the message that wombats seldom sneeze.
This kind of transposition can be carried out in the opposite direction too. Any condition of adequacy on semantic theories can be reconfigured as a partial elucidation of the concept of meaning—or of whatever other concept plays a key role in the semantic theory—and a great deal of philosophical work on the concept of meaning proceeds by considering constraints on semantic theories. Davidson's work (1984) provides an important example of this approach.
the truth-conditional format
As we introduced the notion, a semantic theory is a theory that tells us what expressions mean. It is natural to suppose, then, that the key concept used in a semantic theory will be the concept of meaning, and that the format of the meaning specifications for sentences will be either:
The meaning of sentence S = m
Sentence S means that p
according as meanings are or are not regarded as entities. But Davidson (1967) rejects both these formats, and argues instead for the truth-conditional format:
Sentence S is true if and only if p.
His argument comes in two steps.
The first step is intended to rule out the idea that, to each word, each phrase, and each sentence, there should be assigned some entity as its meaning. This step proceeds by showing that, under certain assumptions about the assignment of entities, all true sentences would be assigned the same entity. (The argument that is used here is sometimes called the Frege argument.) Clearly, no such assignment of entities could be an assignment of meanings, since not all true sentences have the same meaning. However, it is possible to resist this first step by arguing that an assignment of meanings would not conform to the assumptions that are needed to make the Frege argument work.
Even though the first step is controversial, the second step in Davidson's argument remains important for anyone who begins by favoring the format:
Sentence S means that p.
We said that, given the compositionality of meaning, we should expect that, in a semantic theory, the meaning specifications for whole sentences will be derived from more basic principles that specify the meanings of words and phrases. But Davidson points out that the logical properties of the "means that p " construction raise problems for the formal derivation of meaning specifications for sentences. In contrast, the truth-conditional format is logically well understood. And from the work of Alfred Tarski on certain formal languages (1944, 1956) we can carry over methods for deriving truth-condition specifications for sentences from axioms that assign semantic properties to words and phrases.
conditions of adequacy
If what a semantic theory tells us about each sentence of a language is to be cast in the truth conditional format:
Sentence S is true if and only if p
then what are the conditions of adequacy on semantic theories? We have already seen an adequacy condition on the internal structure of a semantic theory; namely, that it should reveal how the truth conditions of complete sentences depend on the semantic properties of words and phrases. But what conditions must the truth condition specifications themselves meet, in order to be correct?
Tarski imposed, in effect, the condition that the sentence that fills the "p " place should translate (or else be the very same sentence as) the sentence S. (This is Tarski's Convention T .) This condition of adequacy can be transposed into a partial elucidation of the concept of truth in terms of the concept of translation. The concept of translation is sufficiently closely related to the concept of meaning that we can move from here to a partial elucidation of truth in terms of meaning:
If a sentence S means that p then S is true iff p.
But we cannot shed any light on the concept of meaning itself without bringing in extra resources.
The key notion that Davidson introduces is that of "interpretation." We imagine using the deliverances of a semantic theory to help interpret the linguistic behavior of speakers. For these purposes, we can abstract away from the details of the format, and use deliverances in the schematic form:
Sentence S __________ p
to license the redescription of utterances of a sentence S as linguistic acts of saying or asserting that p. Now, by providing a way of understanding speakers' specifically linguistic behavior, a semantic theory can play a part in the project of interpreting, or making sense of, them. So, any constraints on the project of overall interpretation of people can be reconfigured as partial elucidations of the key concepts used in semantic theories.
Two suggestions for overarching constraints on interpretation emerge from Davidson's work. One possible constraint is that speakers should be so interpreted that what they say and believe about the world turns out to be by and large correct. This is the "principle of charity" (Davidson, 1967, 1973). The other possible constraint—widely reckoned to be more plausible—is that speakers should be so interpreted that what they say and believe about the world turns out to be by and large reasonable or intelligible. This is sometimes called the "principle of humanity" (see Wiggins, 1980).
In the imagined project of interpretation, the deliverances of a semantic theory are used in schematic form. For these purposes, at least, it does not matter whether the semantic theory uses the "means that p " format or the "is true if and only if p " format. So we can, if we wish, say that the constraints on interpretation shed light on the concept of meaning and thence—by way of the connection between meaning and truth—on the concept of truth.
Meaning and Use
We began from the vague idea that meaning has something to do with use, and have focused on two approaches to the study of meaning, both of which lay stress upon such notions as conveying the message that p, saying that p, and asserting that p. Both approaches take the basic way of specifying the meaning of a sentence to involve a "that p " clause, and both permit the straightforward connection between meaning and truth. However, there are other ways to develop the idea of a link between meaning and use. For example, we might regard knowing the meaning of a sentence as knowing how to use it appropriately. Or we might say that knowing the meaning of a sentence is knowing under what circumstances a speaker would be warranted in using the sentence to make an assertion. Many of these ways of linking meaning with use do not lead to specifications of meaning by way of a "that p " clause, and so do not support the direct transfer of elucidation from the concept of meaning to the concept of truth. It is to metasemantic theories of this kind that the term "use theory of meaning" is usually applied. Use theories of meaning are often coupled with the claim that there is nothing substantive to be said about the concept of truth (see Field, 1994; Horwich, 1990, 1995).
See also Chomsky, Noam; Davidson, Donald; Dummett, Michael Anthony Eardley; Frege, Gottlob; Grice, Herbert Paul; Intention; Philosophy of Language; Pragmatics; Reference; Semantics; Strawson, Peter Frederick; Tarski, Alfred; Truth.
Avramides, A. Meaning and Mind: An Examination of a Gricean Account of Language. Cambridge, MA: MIT Press, 1989.
Chomsky, N. Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger, 1986.
Davis, S. Pragmatics: A Reader. New York: Oxford University Press, 1991.
Dummett, M. The Logical Basis of Metaphysics. Cambridge, MA: Harvard University Press, 1991.
Dummett, M. The Seas of Language. Oxford: Clarendon Press, 1993.
Field, H. "Deflationist Views of Meaning and Content." Mind 103 (1994): 249–285.
Grice, H. P. "The Causal Theory of Perception." Proceedings of the Aristotelian Society, supp. vol. 35 (1961): 121–152. Reprinted in Grice, 1989.
Grice, H. P. "Logic and Conversation." In Syntax and Semantics, Vol. 3: Speech Acts, edited by P. Cole and J. Morgan, 41–58. London, 1975. Reprinted in Grice, 1989.
Grice, H. P. "Meaning." Philosophical Review 66 (1957): 377–388. Reprinted in Grice, 1989.
Grice, H. P. Studies in the Way of Words. Cambridge, MA: Harvard University Press, 1989.
Horwich, P. "Meaning, Use and Truth." Mind 104 (1995): 355–368.
Horwich, P. Truth. Oxford: Blackwell, 1990.
Lewis, D. Convention. Cambridge, MA, 1969.
Lewis, D. "Languages and Language." In Language, Mind and Knowledge, edited by K. Gunderson. Minneapolis: University of Minnesota Press, 1975.
Loar, B. Mind and Meaning. Cambridge, U.K.: Cambridge University Press, 1981.
Neale, S. "Paul Grice and the Philosophy of Language." Linguistics and Philosophy 15 (1992): 509–559.
Schiffer, S. Meaning (1972). Oxford: Clarendon Press, 1988.
Schiffer, S. The Remnants of Meaning. Cambridge, MA: MIT Press, 1987.
Strawson, P. F. "Intention and Convention in Speech Acts." Philosophical Review 73 (1964): 439–460.
Tarski, A. "The Concept of Truth in Formalized Languages." In Logic, Semantics, Metamathematics, edited by A. Tarski, 152–278. Oxford: Clarendon Press, 1956.
Wiggins, D. "What Would Be a Substantial Theory of Truth?" In Philosophical Subjects, edited by Z. van Straaten, 189–221. Oxford: Clarendon Press, 1980.
Martin Davies (1996)
In the social sciences, the concept of meaning has been understood theoretically in three different ways, which can be called the mirroring, constructivist, and production perspectives.
The mirroring perspective derives from the work of Max Weber (1864–1920), who called attention to the subjective interpretations that people create to describe or explain their social experience. Whether or not these meanings are conveyed to others, they are social to the extent that they are broadly characteristic of those who share a similar position within the structure of social relationships. For example, according to a study of television audiences in Israel, the meanings viewers ascribed to the American television series Dallas had much more to do with ethnicity than with whatever meanings might have been intended by the show’s scriptwriters and directors. Arabs saw the show as a moral drama exemplifying how American immorality inexorably led to chaos in interpersonal relationships, while Jews who recently immigrated from Russia saw it as a social drama portraying the evils of the capitalist system (Liebes and Katz 1990). Similarly, in the United States and United Kingdom, men and women have been repeatedly found to view corporate leadership roles differently. For most men, effective leadership means setting up a forceful and consistently applied system of rewards and punishments. For most women, it means showing others how to transform one’s individual goals into group goals so that teams can work more effectively and harmoniously (Alimo-Metcalfe 1995). In the mirroring perspective, the social meanings that people are found to have about the social world are seen to stem from their differing experience of a social structure in which prestige, wealth, and power are unequally distributed. Although shared social meanings can foster a sense of shared identity, when viewed through the lens of the mirroring perspective, they are essentially epiphenomenal and close to irrelevant for explanatory purposes. For example, studies of corporate leadership styles have largely failed to demonstrate that these differing meanings are related to the style female leaders actually choose.
The constructivist perspective, in contrast, emphasizes how people employ agency in order to create and deploy social meanings within fields of social relations that are contentious, fluid, and imperfectly defined: a world of becoming rather than being. This perspective derives from a variety of 1960s theoretical innovations, including symbolic interactionism, phenomenology, sociolinguistics, anthropological linguistics, ethnomethodology, structuralism, social movement theory, pragmatics, and social constructivism. The constructivist perspective turns the mirroring perspective on its head: Social structure is seen as the consequence of the constant and contentious efforts people make to create meanings and, by means of communicative behaviors, to frame social situations in terms of those meanings. In its most radical forms, constructivists deny that what Émile Durkheim (1858–1917) would call “social facts”—objectively measurable patterns of social behavior that are external to individual motivations—have any a priori existence whatsoever. For example, medicine has struggled unsuccessfully for decades to define a “self-neglect syndrome” in objective, medical terms. If a self-neglect syndrome can be said to exist in an objective sense, it is only because, within the field of medical relations, some patients have emphasized “self-neglect” behaviors in order to frame the situation in light of their understanding of it; alternatively, a medical organization has identified certain patient behaviors as “self-neglect” in order to alter the patient’s position and rights within a medical organization.
The production perspective, the most recent of the three, derives from the work of British sociologist Anthony Giddens (1993) and, more recently, the flourishing subfield of cultural sociology. It seeks to strike a balance between the mirroring and constructivist perspectives by emphasizing the dialectic between, on the one hand, social forces that really are larger than the individual (such as institutionalized discrimination), and, on the other hand, the attempts by individuals and groups to try to frame situations to their advantage. In order to do so, they draw on the stock of meanings at hand, but they also create new ones or repackage the old ones in new ways. From the production perspective, the interpretation of Dallas offered by Arab citizens of Israel is part of a broader meaning-formation strategy that is being pursued by Sunni Arab communities throughout the Middle East: It involves a rejection of the secularism and perceived immorality of Western societies in favor of a radically reformulated, fundamentalist version of Islam known as Wahhabism. In contrast to constructivist approaches, which almost invariably emphasize interpersonal interaction as the medium in which meanings are defined, the production perspective is equally attuned to the institutions, technologies, and media that meaning-producers appropriate as they attempt to frame the meaning of situations. For example, the rise of Islamic moral and religious fundamentalism in Sunni Arab communities is clearly related to Saudi Arabia’s lavishly funded and influential system of religious schools and to the rise of Arabic media, including satellite television stations, that are independent of Western control and censorship.
SEE ALSO Anthropology; Constructivism; Durkheim, Émile; Giddens, Anthony; Identity; Islam, Shia and Sunni; Linguistic Turn; Media; Phenomenology; Representation; Television; Weber, Max
Alimo-Metcalfe, Beverly. 1995. An Investigation of Female and Male Constructs of Leadership and Empowerment. Women in Management Review 10 (2): 3–8.
Giddens, Anthony. 1993. New Rules of Sociological Method: A Positive Critique of Interpretative Sociologies. 2nd ed. Stanford, CA: Stanford University Press.
Weber, Max.  1968. Economy and Society. Trans. Guenther Roth and Claus Wittich. New York: Bedminster.
The concept of meaningful action is most closely associated with Max Weber, who distinguishes it from behaviour; that is, from merely physical movement to which the actor does not attach a meaning (for example blinking). Meaningful social action, by contrast, is action directed towards others and to which we can attach a subjective meaning. In this sense, praying alone in a church is meaningful action, as is participating in a church service. Most later theorists accept that these distinctions cannot easily be maintained and take the view that attaching a meaning to something (‘mere behaviour’) is itself an action; and, moreover, a social action because it draws on a socially constructed and accepted language. See also ACTION THEORY; INTERPRETATION.
mean·ing / ˈmēning/ • n. what is meant by a word, text, concept, or action: the meaning of the word “supermarket” | it was as if time had lost all meaning. ∎ implied or explicit significance: he gave me a look full of meaning. ∎ important or worthwhile quality; purpose: this can lead to new meaning in the life of older people. • adj. intended to communicate something that is not directly expressed: she gave Gabriel a meaning look. PHRASES: not know the meaning of the word inf. behave as if unaware of the concept referred to or implied: “Humanity?” You don't know the meaning of the word!DERIVATIVES: mean·ing·ly adv.
1. The purport or message conveyed by WORDS, PHRASES, SENTENCES, SIGNS, SYMBOLS, and the like: ‘Semantics’ means ‘the study of meaning’; A red traffic light means drivers have to stop and a green one means they can go.
2. Signification, sense, interpretation, as in The Meaning of Meaning, the title of a book on SEMANTICS by C. K. Ogden & I. A. Richards (1923); any instance of these, as in What is the meaning of the word ‘semantics’?
3. What a speaker or writer intends: What do you mean?; They don't mean any harm. See LANGUAGE, SEMIOTICS.