Relevance theory is a cognitive account of communication developed by Dan Sperber and Deirdre Wilson. Its central claim is that human cognition tends to seek information that yields worthwhile cognitive effects for the effort required to process it, and that overt attempts to communicate create expectations of their own relevance. These expectations guide an audience in identifying what a communicator intends to convey. The theory is a major approach within pragmatics, the study of how utterances and other communicative acts receive interpretations in context.
First presented systematically in Relevance: Communication and Cognition in 1986 and revised in 1995, the theory treats linguistic communication as a combination of decoding and inference. An expression supplies evidence about a speaker’s meaning, but does not normally encode that meaning completely. Its interpretation depends on contextual assumptions, recognition of intentions, and expectations concerning the likely benefits and costs of processing. The framework extends beyond spoken language to gestures, pictures, writing, and other deliberately communicative behavior.
The theory has generated influential accounts of explicit meaning, implicit communication, metaphor, irony, and lexical interpretation. It also makes psychological claims about attention and comprehension, giving it connections with experimental research on language processing and social cognition. Its explanatory scope, and the difficulty of independently measuring relevance, remain important subjects of debate.
Intellectual origins and development
Relevance theory emerged from the inferential account of communication associated with the philosopher H. P. Grice. In his 1957 paper “Meaning,” Grice distinguished meaning conveyed through a communicator’s intentions from natural indications such as smoke indicating fire. His William James Lectures of 1967, published in part as “Logic and Conversation” in 1975, developed an account of how audiences recover meanings that speakers do not explicitly state. These ideas helped establish the modern distinction between linguistic meaning and speaker meaning.
Grice’s cooperative principle and conversational maxims explained how hearers could derive conversational implicatures by assuming that speakers were contributing appropriately to an exchange. The maxims concerned quantity, quality, relation, and manner. Sperber and Wilson retained the central importance of intention recognition but questioned whether a collection of conversational norms provided the best explanation of comprehension. They proposed that a cognitively grounded principle of relevance could account for many of the same phenomena without requiring hearers to reason through separate maxims.
Their first edition of Relevance: Communication and Cognition appeared in 1986. It combined an account of inferential communication with a theory of how information interacts with existing assumptions. The second edition, published in 1995, added a postface that clarified and developed the framework, including the distinction between the cognitive and communicative principles of relevance. Later presentations refined the account of optimal relevance and emphasized that comprehension is constrained by the communicator’s abilities and preferences.
The theory subsequently developed through work on specific interpretive problems. Robyn Carston’s Thoughts and Utterances (2002) examined the contribution of pragmatic inference to explicitly communicated content. Diane Blakemore’s Semantic Constraints on Relevance (1987) developed the distinction between conceptual and procedural meaning. Ernst-August Gutt’s Translation and Relevance (1991) applied the framework to translation, while work by Deirdre Wilson, Dan Sperber, and other researchers elaborated accounts of irony, metaphor, and lexical adjustment.
These developments made relevance theory more than a replacement for one conversational maxim. Its distinctive ambition is to explain linguistic pragmatics within a broader account of cognition and communication. It therefore differs both from approaches that concentrate on the formal properties of sentences and from approaches that primarily describe the social organization of interaction. The framework accepts the importance of linguistic structure and social circumstances, but asks how their contribution is represented and used by an individual interpreter.
Cognitive relevance: effects, effort, and context
In its technical sense, relevance is a property that an input can have for an individual at a particular time. Inputs include utterances, perceptions, memories, and internally generated thoughts. An input is relevant when processing it produces positive cognitive effects: worthwhile changes in the individual’s representation of the world. Relevance is therefore not simply topical relatedness, novelty, interest, or factual importance considered independently of a thinker.
The theory distinguishes several central kinds of cognitive effect. An input can combine with existing assumptions to yield a new conclusion, called a contextual implication. It can strengthen an existing assumption by supplying further evidence. It can also contradict an existing assumption and lead to its revision or abandonment. If a traveler believes that a train normally leaves from platform 4, an announcement assigning it to platform 7 can produce an important corrective effect. Information about an unrelated train may be entirely new yet produce little benefit for that traveler.
Processing effort provides the other major dimension. Perceiving a signal, identifying words, resolving ambiguity, retrieving contextual assumptions, and drawing conclusions all require cognitive resources. Other things being equal, greater positive cognitive effects increase relevance, whereas greater processing effort decreases it. The framework does not define relevance by a universally applicable numerical ratio. Its central claims are comparative: two inputs may differ in expected relevance because one offers more useful effects or requires less effort under otherwise comparable conditions.
The cognitive principle of relevance states that human cognition tends to be geared toward the maximization of relevance. This is a claim about the organization of attention and processing, not a claim that people consciously calculate utility whenever they hear a sentence. Nor does it entail that actual processing always reaches the objectively best result. Fatigue, misleading evidence, inaccessible knowledge, and mistaken expectations can all affect which inputs receive attention and which interpretations appear satisfactory.
Context is consequently not treated as a fixed collection of facts supplied in advance. It consists of assumptions used in interpreting a particular input, potentially drawn from perception, memory, preceding discourse, and general knowledge. For example, “The lights are still on” may support an inference that somebody remains in an office, that electricity is being wasted, or that a building has not lost power. Which assumptions become accessible helps determine the effects the utterance produces.
This account overlaps with questions in psycholinguistics about attention, memory retrieval, and processing difficulty. Its distinctive proposal is that the selection of context and the interpretation of an input are interdependent: hearers do not normally identify a complete context first and interpret the utterance only afterward.
Ostensive communication and communicative intentions
Relevance theory distinguishes ordinary evidence from evidence deliberately presented as communicative. Wet pavement may provide evidence that it has rained without anyone intending to convey that information. A person who deliberately points at the wet pavement to answer a question, however, makes an overt attempt to direct another person’s attention. Such behavior is ostensive: it makes an intention to communicate recognizable.
The framework analyzes ostensive-inferential communication through two related intentions. An informative intention is an intention to make some set of assumptions manifest or more manifest to an audience. A communicative intention is an intention to make it mutually manifest to the audience and communicator that the communicator has this informative intention. “Manifest” here concerns what an individual is capable of mentally representing and accepting as true or probably true, rather than only what is currently being consciously considered.
The distinction explains why intentionally providing evidence is not always equivalent to openly communicating. Someone might leave a document where a colleague will discover it while concealing the fact that its placement was deliberate. This can be an attempt to affect the colleague’s beliefs without openly presenting an informative intention. By contrast, handing over the document with a meaningful gesture makes the communicative act itself recognizable.
Ostension gives rise to the communicative principle of relevance: every act of ostensive communication communicates a presumption of its own optimal relevance. This presumption has two parts. The stimulus is relevant enough to be worth the audience’s processing effort, and it is the most relevant one compatible with the communicator’s abilities and preferences. The second qualification matters because communicators may lack information, have difficulty expressing themselves, wish to conceal something, or pursue interests that differ from those of their audience.
Optimal relevance does not mean maximal informativeness, perfect clarity, or universal usefulness. A short warning such as “Behind you!” may be optimally relevant in an urgent situation precisely because it avoids a longer description. Conversely, an elaborate explanation may be appropriate when the audience needs conceptual distinctions that cannot be conveyed economically. What counts as sufficient depends on the audience’s expectations and circumstances.
The account is related to, but not identical with, common ground. Sperber and Wilson introduced mutual manifestness partly to avoid requiring an indefinitely nested chain of actual shared knowledge. Communicators need not establish that each knows that the other knows every relevant assumption. They need a cognitive environment in which relevant assumptions and communicative intentions can be recognized. This makes room for communication under uncertainty, while preserving a distinction between merely causing a response and openly inviting an interpretation.
The relevance-guided comprehension procedure
Relevance theory proposes a practical interpretive procedure for an audience encountering an ostensive stimulus: follow a path of least effort in computing cognitive effects, testing interpretive hypotheses in order of accessibility, and stop when expectations of relevance are satisfied. This procedure concerns how interpretations are considered, not a requirement that the final interpretation always be the shortest or most literal one.
Consider an exchange in which one person asks whether a colleague can join a meeting and receives the reply, “She is on a flight to Madrid.” Decoding the reply does not directly provide a yes-or-no answer. The hearer retrieves assumptions about flights, availability, and the meeting’s timing, and may infer that the colleague cannot attend. A less accessible interpretation—such as an indirect suggestion to move the meeting to Madrid—would require additional contextual support. The comprehension procedure explains why the ordinary interpretation can be reached without exhaustively considering every logically possible meaning.
Interpretation proceeds through the mutual adjustment of several components. The hearer develops a hypothesis about explicit content, accesses or constructs contextual premises, and derives contextual conclusions. These processes constrain one another. If the expected answer concerns attendance at a meeting happening now, “on a flight” is likely to be understood as describing the colleague’s present situation. If the meeting is next month, the same words may require a different temporal interpretation or further contextual information.
This model does not require the hearer first to establish a complete literal proposition and only then search for additional meaning. Word recognition, reference resolution, contextual enrichment, and implication can interact during comprehension. Linguistic form nevertheless constrains the process. Syntax and encoded meaning provide structured evidence; relevance is not a license to substitute any interpretation that happens to be useful.
The procedure also explains the possibility of misunderstanding. A hearer may stop at an interpretation that satisfies expectations of relevance but differs from what the speaker intended. Different background assumptions, an incorrectly identified referent, or an unexpected communicative goal can all produce such an outcome. The theory therefore predicts a tendency toward efficient interpretation, not guaranteed communicative success.
Conscious deliberation becomes especially important when ordinary interpretation fails, when irony is uncertain, or when the audience suspects deception. The general framework can accommodate reasoning about motives and reliability, but its central comprehension heuristic concerns the interpretation of a communicative act. Determining what someone intends to convey remains distinct from accepting that content as true. An audience can understand an assertion fully while rejecting it.
Explicit meaning, explicatures, and implicatures
One of relevance theory’s most influential claims is that pragmatic inference contributes to what is explicitly communicated, not merely to additional implications. A decoded linguistic representation often leaves unresolved who is being discussed, which time is intended, or what relation holds between mentioned events. Relevance theory calls a communicated assumption an explicature when it is developed from the logical form encoded by an utterance.
Reference assignment is a straightforward case. Understanding “She left it there” requires identification of the relevant person, object, and location. Deictic expressions and pronouns supply instructions or constraints, but the sentence does not independently specify their referents. Ambiguity resolution is similarly necessary when a word has several conventional senses. Pragmatic development also includes less overtly signaled enrichment, as when “I have eaten” is understood as meaning that the speaker has eaten recently enough not to need the meal currently under discussion.
An explicature differs from an implicature, which is communicated without being a development of the utterance’s encoded logical form. In the meeting example, the colleague’s being on a flight can form part of explicit content, whereas her inability to attend is an implicated conclusion. An implicated premise, such as the assumption that someone traveling on that flight cannot join the meeting, helps support the conclusion. Such premises need not be explicitly stated or consciously articulated.
Relevance theorists also distinguish basic explicatures from higher-level explicatures. The latter embed a proposition under a description of an attitude or communicative act. A hearer may understand an utterance not only as presenting a proposition but as expressing a belief, a doubt, or a warning concerning it. These representations help explain how interpretation includes the speaker’s stance as well as a description of a possible state of affairs.
The boundary between explicature and implicature has generated extensive debate. The theory rejects the assumption that everything affecting truth conditions must be determined by linguistic decoding alone. At the same time, its distinction is not simply identical to the distinction between truth-conditional and non-truth-conditional content. Development of encoded logical form is the principal criterion for explicature, although the precise limits of “development” remain disputed.
These issues connect relevance theory with truth-conditional semantics and compositionality. A central question is how far sentence structure determines a proposition before contextual inference contributes. Relevance theory generally allows extensive pragmatic contribution while maintaining that linguistic encoding places real constraints on interpretation. Disagreement concerns both particular examples and the architecture of the interface between language and thought.
Lexical adjustment and procedural meaning
Relevance-theoretic lexical pragmatics examines how a word’s conventional meaning is adjusted in context. Its central proposal is that an interpreted concept need not coincide exactly with the relatively stable concept associated with a lexical item. Hearers can construct an ad hoc concept suited to the speaker’s communicative purpose, narrowing or broadening the encoded concept as appropriate.
Narrowing occurs when a word is understood more restrictively than its conventional meaning alone requires. In a discussion of abstinence, “He does not drink” is ordinarily interpreted as concerning alcoholic beverages rather than all liquids. Broadening occurs when the communicated category extends beyond the encoded category. Calling an approximately circular arrangement “a circle,” or describing a very quiet room as “silent,” need not commit the speaker to strict geometrical or acoustic conditions.
This account treats approximation, hyperbole, and many uses of metaphor as points along a continuum rather than as entirely separate interpretive mechanisms. “The road is a ribbon” can guide the construction of a concept that captures relevant features such as narrowness and continuity without representing the road as a literal strip of fabric. An interpretation is constrained by which properties yield relevant implications in the context, not by a requirement to preserve all features of the encoded concept.
The approach differs from theories that require a literal interpretation to be completed and rejected before figurative meaning can be recovered. It also differs from conceptual metaphor theory, which places greater emphasis on systematic mappings between conceptual domains. The approaches can address overlapping data while offering different explanations: one focuses on local inferential adjustment, the other on recurrent conceptual organization. Neither the presence of conventional patterns nor the possibility of innovative usage, by itself, settles the issue.
A second important distinction concerns conceptual and procedural meaning. Conceptual expressions contribute representations that can figure in thought. Procedural expressions instead constrain aspects of interpretation, for example by directing a hearer toward a certain inferential relation. Blakemore’s early work argued that discourse connectives such as “so” and “after all” could be analyzed as guiding the inferential processing of associated propositions rather than merely adding another descriptive proposition.
Procedural accounts have subsequently been explored for pronouns, tense, discourse markers, and intonation, although particular analyses vary. The distinction should not be equated with a simple division between content words and function words. Nor does procedural meaning automatically coincide with non-truth-conditional meaning. The central issue is what kind of information an expression encodes and how that information constrains the inferential processes involved in comprehension.
Irony, echo, and weak communication
Relevance theory analyzes many cases of verbal irony as echoic uses of language. An utterance is echoic when it represents an attributed thought or utterance and conveys the speaker’s attitude toward it. The attributed content can come from a particular person, an earlier remark, a socially recognizable expectation, or a thought that someone might reasonably have entertained. It need not reproduce an actual statement word for word.
Suppose an outdoor event is disrupted by heavy rain after somebody predicted excellent weather. A participant says, “A perfect day for a picnic.” On an echoic interpretation, the speaker evokes the earlier prediction or expectation and expresses a dissociative attitude toward it. The point is not merely to assert the logical opposite of the sentence. Depending on context, the utterance may communicate amusement, disappointment, skepticism, or criticism of the original prediction.
The theory distinguishes descriptive use, in which a representation describes a state of affairs, from interpretive use, in which a representation resembles another representation in relevant respects. Echoic communication is a form of interpretive use. Reported speech, paraphrase, and some questions likewise involve relations between representations, although they do not necessarily express irony. Echo alone is insufficient: the dissociative attitude is crucial to the echoic account of irony.
This analysis helps explain why an ironic statement can retain the wording and perspective of an attributed voice. It also accommodates ironic questions and expressions whose interpretation cannot be reduced neatly to asserting a contradictory proposition. Competing accounts emphasize pretense, enacted perspectives, or breaches of expectations. Debate concerns whether echoic interpretation captures all irony and how the relevant attitudes should be identified independently.
The same framework distinguishes strong from weak communication. A strongly communicated implication is one the audience is specifically expected to recover. Weak communication gives the audience greater responsibility for selecting among a range of implications. A metaphorical description of a city as “waking from a long sleep” may evoke changes in activity, public life, or confidence without making every such implication separately determinate.
This notion is important in relevance-theoretic work on poetry and literary interpretation. Additional effort can be rewarded by a rich range of weak effects rather than by one easily paraphrased message. Weakness here does not mean insignificance, carelessness, or failure. It concerns the degree to which the communicator takes responsibility for a particular interpretation. The account offers a way to discuss suggestiveness while distinguishing textually supported interpretation from associations supplied primarily by an individual reader.
Experimental investigation and cognitive architecture
Because relevance theory makes claims about comprehension, it invites investigation through behavioral experiments and other methods used in language research. Researchers examine reading times, response times, eye movements, interpretation choices, and sensitivity to contextual information. Such evidence can test particular processing hypotheses, although no single task directly measures the theory’s entire construct of relevance.
A prominent research area concerns scalar implicatures, such as the interpretation of “some” as “some but not all.” Classical logical meaning permits “some” to be true when all members satisfy a description, whereas ordinary utterances often suggest an upper limit. Contextualist approaches associated with relevance theory predict that the strengthened interpretation depends on its usefulness in context rather than being obligatorily generated whenever “some” occurs.
Work by Ira Noveck, including a 2001 study, helped establish developmental differences in responses to underinformative scalar statements. Children often accepted logically true statements that adults were more likely to reject as pragmatically inadequate. Subsequent research showed that performance varies with task design, contextual support, and the alternatives made salient. These findings cannot be reduced to a simple claim that children lack pragmatic inference; they require separating linguistic understanding from the demands of the experimental task.
Studies by Noveck and Andrés Posada in 2003, and by Lewis Bott and Noveck in 2004, examined processing associated with scalar interpretation. Results showing additional costs in some conditions supported challenges to automatic, cost-free enrichment. Later experimental work complicated the contrast between “default” and “contextual” processing: supportive contexts can facilitate strengthened interpretations, and different experimental measures can reveal different stages of processing. Processing cost alone does not establish a unique theoretical explanation.
The framework also connects communication with theory of mind, or the capacity to represent mental states. Sperber and Wilson have argued that comprehension may involve a specialized inferential mechanism adapted to communicative intentions, rather than relying exclusively on unrestricted general reasoning about other minds. This proposal concerns cognitive architecture and should be distinguished from the more general claim that communication involves intentions. The existence and precise organization of such specialized mechanisms remain debated.
Experimental assessment faces a recurring methodological problem: effort and cognitive effects can be affected by vocabulary, familiarity, task goals, and expectations at the same time. Strong tests therefore require explicit predictions about particular contrasts, not retrospective claims that whichever interpretation occurred must have been most relevant. Relevance theory supplies a broad framework, while experimentally testable accounts usually add assumptions about lexical access, context accessibility, or the structure of a specific task.
Translation, literature, and multimodal communication
In translation studies, relevance theory has been influential through Ernst-August Gutt’s account of translation as interpretive use across languages. A translation represents an earlier utterance or text rather than simply describing the same external situation independently. Its relation to the original can therefore be analyzed in terms of shared explicatures and implicatures, considered relative to the target audience’s cognitive environment.
Gutt distinguished direct and indirect translation within this framework. Direct translation aims at complete interpretive resemblance when interpreted against the context envisaged for the original; indirect translation claims resemblance in respects relevant to its audience and purpose. These are theoretical categories rather than simple equivalents of word-for-word and free translation. A formally close rendering can fail to preserve an implication, while a structurally different rendering can preserve an important aspect of interpretation.
The approach highlights the consequences of unequal background knowledge. An allusion immediately accessible to the original audience may require substantial effort or remain opaque to another audience. Adding an explanation can make some implications accessible but also change pacing, emphasis, and the distribution of explicit and implicit content. Relevance theory describes these trade-offs without guaranteeing that every feature of a text can be reproduced simultaneously. Its relation to functional approaches such as skopos theory concerns, among other things, how communicative purpose constrains acceptable resemblance.
Literary applications examine how style guides inference. Repetition, unusual word choice, narrative perspective, and deliberate ambiguity can impose effort while yielding effects unavailable from a plain paraphrase. Relevance-theoretic stylistics, including work by Adrian Pilkington, has used weak communication to explain why literary significance often exceeds a list of determinate propositions. The framework also distinguishes the intentions attributable to narrators, characters, and authors, rather than treating every sentence of a narrative as a direct authorial assertion.
The theory’s conception of ostension allows applications to visual and multimodal communication. In advertising, political cartoons, film, and face-to-face interaction, an image, gesture, caption, or sound can direct attention toward assumptions that support an interpretation. Charles Forceville and other researchers have explored connections between relevance theory and visual or multimodal meaning. Linguistic decoding is absent or only one component in such cases, making them useful tests of the framework’s broader inferential claims.
These applications require attention to medium-specific conventions. A camera movement, a pointing gesture, and a written connective do not encode information in the same way. Relevance theory supplies a general account of communicative expectations, but an adequate analysis must also explain how audiences recognize the particular representational resources and conventions of the medium.
Theoretical disputes and unresolved questions
A central criticism concerns explanatory precision. If relevance depends jointly on cognitive effects and processing effort, and neither is directly measurable in a uniform way, there is a risk of explaining an interpretation only after it has occurred. Relevance theorists respond by emphasizing comparative predictions under controlled conditions and the theory’s explicit stopping procedure. The continuing issue is how independently specified measures of accessibility, effort, and benefit can constrain explanations rather than merely redescribe successful comprehension.
A second dispute concerns the extent of pragmatic contribution to explicit meaning. Some approaches preserve a relatively minimal, linguistically determined proposition and place contextual additions elsewhere in the analysis. Relevance theorists typically argue that ordinary understanding involves pragmatic development well beyond reference assignment and ambiguity resolution. Disagreement often turns on how to identify the proposition intuitively judged to have been asserted and whether that judgment should determine the boundary between semantics and pragmatics.
Socially oriented approaches question whether an account centered on individual cognition gives sufficient explanatory weight to interaction. Status, identity, politeness, institutional roles, and turn-by-turn negotiation affect what speakers do and what audiences infer. Relevance theory can represent social expectations as contextual assumptions and communicative goals, but critics argue that this may not explain how those expectations are produced or maintained. The dispute is partly about explanatory level: a cognitive theory of interpretation and an account of social practice need not have identical objects of analysis.
There are also challenges from probabilistic and game-theoretic models. These approaches can model speakers and listeners as reasoning about alternative expressions, meanings, and communicative goals. They share with relevance theory an emphasis on inference and expectations, but may provide more explicit quantitative predictions. Their utilities, priors, and costs do not automatically correspond to relevance-theoretic cognitive effects and effort. Establishing a principled relationship requires more than observing that both approaches favor informative communication.
The treatment of indirectness and politeness raises a related question. A longer or less direct expression may preserve a relationship, avoid an unwanted commitment, or signal deference. Such outcomes can explain why the shortest available utterance is not always preferred. However, analyses must distinguish the communicator’s goals from the audience’s cognitive benefits and explain how social consequences enter interpretation without rendering optimal relevance unrestricted.
Finally, the theory must specify its domain carefully. The presumption of optimal relevance concerns recognized ostensive communication, not every noise, behavior, or linguistic token. Audiences may encounter accidental inscriptions, overhear speech not addressed to them, or recognize communication aimed at several groups with different knowledge. These cases test how communicative intentions, audience design, and contextual accessibility interact, especially when no single interpretation is equally available to all recipients.