{
  "id": "conceptual-metaphor-theory",
  "title": "Conceptual Metaphor Theory",
  "lang": "en",
  "source_lang": "en",
  "summary": "Conceptual Metaphor Theory explains metaphor as systematic connections between domains of thought that shape language, reasoning, and other forms of expression.",
  "category": "language",
  "subcategory": "",
  "type": "concept",
  "aliases": [
    "CMT",
    "conceptual theory of metaphor",
    "cognitive metaphor theory"
  ],
  "generated_at": "2026-11-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/conceptual-metaphor-theory/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/conceptual-metaphor-theory/",
      "json": "https://aiwiki.page/zh/conceptual-metaphor-theory.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Conceptual Metaphor Theory is an approach to metaphor that treats it primarily as a pattern of thought rather than merely an ornamental use of words. Its central proposal is that people understand aspects of one domain of experience through structures associated with another: time through spatial motion, affection through warmth, or purposes through destinations. Expressions such as “a long meeting,” “a warm relationship,” and “reaching a goal” are therefore investigated as linguistic manifestations of more general conceptual relationships.\n\nThe theory became a central component of cognitive linguistics following George Lakoff and Mark Johnson’s Metaphors We Live By (1980). It challenged the separation of ordinary, supposedly literal thought from exceptional, figurative expression. Its influence extends across linguistics, psychology, literary studies, education, and the analysis of public discourse. Research now examines not only verbal expressions but also gesture, images, experimental behavior, and patterns of cultural variation.\n\nConceptual Metaphor Theory is a research framework with several versions rather than a single fully specified model. There is broad evidence that metaphorical patterns organize vocabulary and discourse, but stronger claims—that particular mappings are automatically activated during comprehension, grounded in specific bodily experiences, or implemented by particular neural mechanisms—require independent evidence and remain subjects of debate."
    },
    {
      "heading": "Intellectual origins and development",
      "text": "The relationship between metaphor and thought predates modern cognitive linguistics. Aristotle discussed metaphor in the Poetics and Rhetoric, including its capacity to reveal similarities. Later traditions often classified it among the devices of rhetoric, although this did not mean that every earlier account reduced it to decoration. In the twentieth century, I. A. Richards’s The Philosophy of Rhetoric (1936) and Max Black’s essay “Metaphor” (1954–1955) developed interactional accounts in which a metaphor’s components reshape the interpretation of one another.\n\nConceptual Metaphor Theory shifted the central explanatory question. Instead of asking only how a particular unusual sentence acquires a figurative meaning, it asked why ordinary expressions form recurring families. Lakoff and Johnson’s examples included “defending a position,” “attacking an argument,” and “winning an argument.” They proposed that these expressions reflect a systematic relationship conventionally labeled ARGUMENT IS WAR. The capitalization identifies an analyst’s formulation of a conceptual pattern, not a sentence that speakers must consciously represent or endorse.\n\nA second important background was Michael Reddy’s “The Conduit Metaphor,” published in 1979. Reddy identified expressions that present communication as putting ideas into words, sending them, and extracting their contents. This analysis suggested that everyday ways of talking about communication carry assumptions about what communication is. The importance of such examples lay in their conventionality: expressions could be highly familiar while still reflecting a patterned construal of an abstract activity through a more concrete one.\n\nSubsequent publications broadened the framework. Lakoff and Mark Turner’s More Than Cool Reason (1989) connected everyday conceptual patterns with poetry. Lakoff and Johnson’s Philosophy in the Flesh (1999) developed the theory’s relationship with embodied cognition. Zoltán Kövecses investigated emotion, cultural variation, and contextual influences, while Joseph Grady’s work in the 1990s distinguished relatively elementary experiential associations from more elaborate metaphorical systems.\n\nThe framework developed alongside, rather than replacing, other approaches to meaning. Scholars working in semantics, pragmatics, and psycholinguistics continued to investigate metaphor through lexical interpretation, contextual inference, categorization, and analogy. Consequently, “conceptual metaphor” can designate either a relatively modest claim about systematic cross-domain patterns or a stronger theory about the cognitive mechanisms underlying them. Distinguishing these uses is necessary when comparing findings: evidence for a recurrent family of expressions is not automatically evidence for every psychological or neurological claim associated with the framework."
    },
    {
      "heading": "Domains, mappings, and metaphorical inference",
      "text": "The basic analytical unit is a mapping between a source domain and a target domain. A conceptual domain is an organized body of knowledge or experience, such as journeys, economic exchange, illness, or social relationships. The source supplies structure used to understand some aspect of the target. In LOVE IS A JOURNEY, for example, travelers may correspond to partners, their vehicle to the relationship, obstacles to difficulties, and destinations to shared purposes. Expressions such as “a relationship at a crossroads” make parts of this proposed organization linguistically accessible.\n\nA cross-domain mapping is selective rather than exhaustive. A journey metaphor does not predict that every feature of transportation will apply to relationships. Fuel consumption, passport control, and tire pressure do not automatically become relevant. The analyst must establish which correspondences are supported by actual expressions or reasoning. A simplified representation is \\(M:S\\rightarrow T\\), where \\(S\\) is the relevant source structure and \\(T\\) the target structure. This notation indicates directionality; it does not imply that metaphorical mappings behave like fully defined mathematical functions.\n\nMappings matter because they can license inferences, not merely substitute vocabulary. If an activity is understood as movement toward a destination, delays may become setbacks, alternatives may become routes, and successful completion may become arrival. Nevertheless, a plausible inference is not necessarily one that a speaker actually draws. The theory distinguishes the availability of an inference from its activation in a particular context, although early analyses did not always maintain this distinction explicitly.\n\nMetaphorical structure is also partial in another sense: the same target can be understood through several sources. Time can be represented as a moving object, a landscape through which an observer moves, a limited resource, or a commodity. “The deadline is approaching” and “the team is approaching the deadline” organize relative motion differently. “Saving time” introduces an additional resource-related construal. These patterns need not form one completely consistent mental picture; each can make selected relations salient.\n\nLakoff’s invariance hypothesis attempted to constrain projection by proposing that mappings preserve relevant source-domain image-schematic structure while respecting the target’s inherent organization. A container’s interior, boundary, and exterior, for example, should retain their relational organization when projected. The proposal addresses why mappings are not arbitrary, but applying it requires independent descriptions of both domains.\n\nDirectionality is generally asymmetric rather than reversible by default. Understanding theories as buildings does not establish that buildings are ordinarily understood as theories. Reversed comparisons may be possible, especially in creative discourse, but they can have different interpretations and conventional status. Establishing a mapping therefore requires more than observing that two domains share a property."
    },
    {
      "heading": "Embodiment, image schemas, and primary metaphors",
      "text": "A major claim of the theory is that many abstract concepts draw structure from recurring bodily experience. Embodied cognition encompasses several positions concerning how perception, action, and bodily states contribute to thought. Within Conceptual Metaphor Theory, embodiment commonly explains why spatial orientation, movement, balance, containment, force, and manipulation provide productive source structures. Bodily experience is not treated simply as a stock of vivid illustrations; it is proposed as a recurring basis for conceptual organization.\n\nAn image schema is a relatively schematic pattern abstracted from perception and action. CONTAINER involves an interior, an exterior, and a boundary; SOURCE–PATH–GOAL involves a starting point, a trajectory, and an endpoint. Mark Johnson’s The Body in the Mind (1987) gave these patterns a central role. Their schematic character is important: a container schema need not resemble a detailed mental picture of a cup or room. It captures relations that can recur across many perceptual situations.\n\nSuch patterns can organize abstract expressions. “Being in trouble” construes a condition in terms of containment, while “getting out of trouble” adds boundary crossing. Describing a project as “moving forward” recruits directed motion. These examples do not establish that every use requires conscious imagery. The proposed organization may be conventional and operate without deliberate visualization, making its psychological status a matter for behavioral investigation rather than introspection alone.\n\nGrady’s 1997 work on primary metaphors sought to identify comparatively elementary mappings associated with recurrent experiential correlations. MORE IS UP is commonly related to situations in which increasing quantity also increases height, as when objects accumulate in a pile. AFFECTION IS WARMTH has been connected to early experiences in which physical closeness and warmth accompany care. These are hypotheses about developmental grounding, not demonstrations that a single experience mechanically determines a concept.\n\nPrimary metaphors are distinguished from complex systems that combine several mappings with cultural knowledge. Purposive activities, for instance, can draw on associations between purposes and destinations, actions and motion, and difficulties and impediments. A complex metaphor may therefore inherit bodily motivations while also depending on historically specific practices and institutions. “Time is money” cannot be explained solely through universal anatomy; paid labor, accounting, and economic organization are relevant to its prominence.\n\nSome versions connect these proposals to learning in neural networks, suggesting that repeated coactivation can establish links between experiential domains. Such accounts remain more specific than the general observation that bodily experience influences thought. A neural explanation must identify mechanisms and predictions beyond those supplied by linguistic examples. Neither a bodily correlation nor activation in a relevant brain region, by itself, establishes a particular conceptual mapping."
    },
    {
      "heading": "Conventionality and relationships with other meaning processes",
      "text": "Conceptual Metaphor Theory gives conventional metaphor a central position. A phrase such as “a close friendship” may be processed fluently and attract no special attention, yet it links social intimacy with spatial proximity. This challenges the assumption that only conspicuous poetic expressions merit analysis as metaphor. It also complicates the category of dead metaphor, traditionally applied to expressions whose figurative origins are no longer noticed. Lack of conscious awareness does not, within this framework, demonstrate lack of conceptual structure.\n\nConventionality, however, has several dimensions. An expression can be frequent, familiar, lexicalized, or easy to process, and these properties do not necessarily coincide. A dictionary may record a nonphysical sense of “grasp,” but the existence of that sense does not settle whether understanding it activates manual action. Analyses of polysemy therefore distinguish a word’s established senses from hypotheses about how those senses are represented and accessed. Historical derivation, present-day semantic relationships, and online processing are separate questions.\n\nCreative metaphor can extend, elaborate, combine, or challenge familiar mappings. An extended metaphor may develop a source domain across several sentences or an entire poem. Its effects depend not only on a general correspondence but also on particular word choices, narrative position, and cultural associations. Conceptual analysis can explain why an image is intelligible without fully explaining its aesthetic force or local interpretive significance.\n\nMetaphor also interacts with metonymy, which involves accessing one entity through another within an associated conceptual frame. “The White House announced” identifies an institution through a building associated with it; it does not ordinarily conceptualize the institution as a building through a systematic cross-domain comparison. In practice, the boundary can be difficult to draw. A bodily symptom can stand metonymically for an emotion and also participate in metaphorical elaboration. Some analyses consequently describe metaphor–metonymy interaction rather than forcing every expression into an exclusive category.\n\nPersonification attributes human capacities or social roles to nonhuman entities, as when a market “panics” or an institution “refuses” to change. Conceptual metaphor analysis asks which properties are projected and what explanations become available. Simile, by contrast, marks a comparison explicitly, often with “like” or “as.” Its surface form does not determine whether it expresses a conventional mapping, a novel analogy, or a limited resemblance.\n\nA related but distinct framework is conceptual blending, developed by Gilles Fauconnier and Mark Turner. It models integration across multiple mental spaces, allowing emergent structure that is not directly inherited from a single source-to-target mapping. Blending and conceptual metaphor can be complementary descriptions, but their units of analysis and explanatory priorities are not identical."
    },
    {
      "heading": "Cultural, linguistic, and historical variation",
      "text": "Conceptual Metaphor Theory has sometimes been associated with universal claims because human bodies share many perceptual and motor capacities. Its research tradition nevertheless documents substantial variation. Shared embodiment does not require identical lexical expressions, identical metaphor frequencies, or identical cultural significance. Environmental conditions, social institutions, writing practices, religious traditions, and interactional conventions can make different experiences especially salient.\n\nTemporal metaphors provide a well-studied example. Many languages use spatial expressions for time, but the organization of those expressions varies. In English, future events are often described as ahead and past events as behind. Research by Rafael Núñez and Eve Sweetser, published in 2006, examined Aymara expressions and gestures in which the known past is associated with the front and the unknown future with the back. The analysis connected temporal orientation with visibility and knowledge. It did not establish that every speaker, in every task, organizes all temporal reasoning in one fixed arrangement.\n\nWork on Mandarin Chinese has investigated both horizontal and vertical temporal expressions. Terms involving upper and lower positions occur in conventional expressions for earlier and later periods. Experiments have tested whether such linguistic patterns affect spatial–temporal judgments, but findings depend on task design, language experience, and the kinds of expressions tested. The presence of a spatial expression is not sufficient to establish an exclusive or immutable conception of time.\n\nKövecses has distinguished variation in the availability of a metaphor from variation in its elaboration and preferred use. Two communities may both conceptualize anger through heat while differing in associated bodily models, conventional expressions, social expectations, and communicative functions. Even within one language, professional groups, generations, or individual situations can favor different sources. A sport may be framed as war in one context and artistic performance in another.\n\nHistorical evidence adds another distinction. Semantic change can show how physical senses develop abstract uses, but an expression’s origin does not automatically reveal its current processing. Conversely, a historically old metaphor may remain productive if speakers create new expressions from its underlying relations. Historical linguistics and conceptual analysis therefore answer related but different questions: one reconstructs changes through time, while the other may investigate present conceptual organization.\n\nCross-linguistic comparison requires comparable evidence. Translations chosen because they resemble English examples can conceal alternative patterns, and differences in dictionaries may reflect lexicographic conventions rather than conceptual differences. Stronger studies combine naturally occurring discourse, native-speaker interpretation, gesture, historical materials, and experimental tasks. They also distinguish variation between languages from variation among speakers and situations within the same language."
    },
    {
      "heading": "Identification methods and experimental evidence",
      "text": "The earliest influential analyses relied heavily on families of intuitively related expressions. Such examples are valuable for generating hypotheses, but systematic investigation requires procedures that separate observed language from proposed conceptual structure. Corpus linguistics makes it possible to examine frequency, distribution, genre, and recurrent combinations across a text corpus. It can reveal that an apparently central metaphor is uncommon in some settings or that several competing sources organize the same topic.\n\nThe Pragglejaz Group’s 2007 Metaphor Identification Procedure, usually called MIP, was designed to improve the transparency of linguistic identification. Analysts establish a lexical unit’s contextual meaning, consider whether it has a more basic contemporary meaning in other contexts, and ask whether the contextual meaning contrasts with that basic meaning but can be understood through comparison with it. The later MIPVU procedure, developed at Vrije Universiteit Amsterdam and presented in a 2010 manual, elaborated identification conventions.\n\nThese procedures identify metaphor-related words, not conceptual mappings automatically. Deciding that “attacked” is metaphorically used in a discussion of criticism does not by itself establish the scope of ARGUMENT IS WAR. That further step requires decisions about source and target categories, recurring correspondences, and contextual alternatives. Dictionaries can support judgments about basic meanings, while independent annotation and agreement measures make disagreements visible. Neither annotation reliability nor a large corpus eliminates the need for interpretive choices.\n\nExperimental research asks different questions. Reaction-time tasks can test whether a spatial manipulation facilitates judgments about time; memory tasks can examine source-consistent reconstruction; gesture studies can investigate whether speakers organize abstract discourse spatially. Lera Boroditsky’s 2000 research on spatial and temporal reasoning helped establish this experimental program. Daniel Casasanto’s 2009 work reported associations between handedness and the spatial evaluation of positive and negative concepts, challenging accounts that assume a uniform bodily basis for all speakers.\n\nInterpretation depends on controls. A result can arise from shared vocabulary, explicit task strategies, learned associations, or general attentional effects rather than the particular conceptual mapping proposed. If a task repeatedly presents spatial diagrams before temporal judgments, it may encourage a strategy that is less prominent in ordinary conversation. Researchers therefore distinguish spontaneous use from experimentally induced recruitment and seek converging results across tasks.\n\nEvidence from the brain, including electrophysiology and neuroimaging, can investigate timing and neural participation. However, activation in sensory or motor regions does not uniquely identify conceptual metaphor: literal language, imagery, attention, and task demands can recruit overlapping systems. The strongest arguments connect clearly specified mappings to independently motivated behavioral and neural predictions rather than treating any embodiment-compatible result as confirmation."
    },
    {
      "heading": "Discourse, framing, gesture, and visual expression",
      "text": "Conceptual metaphors do more than organize isolated word senses. In extended discourse they can select participants, causal relations, goals, and possible responses. The concept of framing concerns how a presentation makes certain aspects of a situation salient while leaving others in the background. Metaphorical framing is one mechanism among others: a problem represented as a battle, a journey, or a breakdown makes different roles and interventions readily available.\n\nPaul Thibodeau and Lera Boroditsky’s 2011 experiments examined descriptions of crime as a beast or a virus. In the reported tasks, these descriptions influenced the kinds of responses participants favored. Subsequent work has investigated the robustness and conditions of such effects. The studies are significant because they move beyond identifying expressions to testing consequences for reasoning, but they do not establish that a single metaphor reliably determines real-world policy preferences.\n\nA discourse metaphor’s effects depend on context, familiarity, and competing knowledge. Readers may accept a mapping, ignore it, interpret it ironically, or explicitly reject it. Calling a debate a battle does not mean that every participant will reason consistently in military terms. Public discourse often mixes sources, and an audience can recognize the rhetorical purposes of that mixing. Analysis therefore distinguishes a text’s available implications from effects actually demonstrated in its audience.\n\nResearch on gesture supplies evidence beyond words. Speakers may place past and future events in different regions of space, represent an abstract hierarchy through vertical movement, or gesture along a path while discussing progress. Such behavior can support the claim that spatial organization is not merely a peculiarity of vocabulary. Nevertheless, gesture is shaped by discourse context and cultural convention, and a movement’s interpretation must be established rather than inferred solely from its shape.\n\nVisual and multimodal research, associated particularly with Charles Forceville’s work, investigates metaphor in advertising, cartoons, film, and combinations of image and text. A product may be represented as an animal or machine, with the viewer invited to transfer selected attributes. In multimodal metaphor, source and target may be cued through different channels, such as an image supplying one domain and a caption supplying the other.\n\nThe analysis of sign languages raises especially useful distinctions because signs can be spatially organized and iconically motivated. Iconicity involves a perceived resemblance between form and meaning; metaphor involves a relationship between conceptual domains. A sign or signed construction can involve both, but spatial form alone does not prove metaphorical structure. Careful analysis separates linguistic organization, depiction, bodily enactment, and cross-domain mapping rather than grouping every visually motivated expression under metaphor."
    },
    {
      "heading": "Translation, education, and computational applications",
      "text": "In translation, conceptual metaphor analysis helps distinguish between preserving an expression’s wording and preserving its function. A source-language phrase may have a familiar target-language counterpart, require a different metaphor, or be rendered without metaphor. These options can preserve different aspects of the original: image, evaluation, discourse continuity, or inferential structure. A literal rendering of a journey expression, for example, may sound natural in one target language but misleading or stylistically marked in another.\n\nThe framework is especially useful when a text develops a metaphor across several passages. Translating each phrase separately can obscure connections between departure, obstacles, routes, and arrival. Conversely, preserving every source image can introduce an artificial coherence that the target language would not normally support. Translation studies therefore treats conceptual patterns as one source of evidence alongside genre, audience, cultural knowledge, and the purpose of the translation.\n\nIn second-language acquisition, metaphor-based instruction has been investigated as a way of organizing figurative vocabulary and idiomatic expressions. Grouping expressions around motion, containment, or force can make relationships among meanings more explicit than presenting them as unrelated items. However, a plausible conceptual explanation does not guarantee improved learning. Outcomes depend on learners’ proficiency, the transparency of the examples, teaching time, and whether assessments measure recognition, retention, or productive use.\n\nWithin education, the theory also informs studies of scientific and mathematical explanation. Expressions such as electrical “flow” or a chemical “bond” provide access to unfamiliar relations through more familiar experience. Such models are selective: an electrical circuit is not equivalent in every respect to a water system. The distinction between productive inference and inappropriate transfer is therefore central. Research on teaching examines both what a metaphor makes intelligible and which misconceptions it may support.\n\nComputational work includes detecting metaphor-related language and identifying possible source–target relations. Detection is not the same task as interpretation: a system may classify a word as metaphorical without identifying the conceptual mapping, and it may identify a broad mapping while missing a passage’s rhetorical significance. Statistical methods can exploit differences between a word’s contextual representation and its typical literal contexts, while annotated corpora provide training and evaluation material.\n\nRecent large language models can generate plausible interpretations and clusters of expressions, but fluency is not evidence of psychological validity. Evaluation must distinguish accuracy on annotated examples, consistency across genres and languages, and the correctness of claimed conceptual correspondences. Computationally proposed mappings require scrutiny for overgeneralization, especially when systems reproduce familiar textbook labels without showing that the relevant relations are supported by the passage being analyzed."
    },
    {
      "heading": "Competing explanations and unresolved theoretical questions",
      "text": "A major criticism concerns circularity. An analyst may infer a conceptual metaphor from several expressions and then explain those same expressions by invoking the inferred metaphor. This is not necessarily illegitimate as descriptive modeling, but it supplies limited independent evidence for a mental mechanism. Stronger tests require predictions about unfamiliar expressions, nonlinguistic behavior, developmental patterns, or processing that were not used to construct the original mapping.\n\nAnother issue is analytical granularity. “Theories are buildings,” “arguments are buildings,” and “abstract structures are physical structures” describe patterns at different levels. A very general label can encompass many expressions while making weak predictions; a narrow label can fit a few examples without explaining broader organization. Researchers must specify the correspondences and conditions under which a mapping is expected to apply. Negative evidence—features that are not projected and contexts in which the metaphor is unavailable—is especially important.\n\nAlternative accounts explain some metaphorical language through categorization, analogy, or pragmatic inference. Sam Glucksberg and colleagues developed accounts in which metaphors can function as category-inclusion statements: describing someone as a “shark” may assign that person to an ad hoc category of predatory or ruthless agents. Relevance theory analyzes metaphor through context-sensitive adjustment of communicated concepts. Analogy-based accounts emphasize alignment between relational structures. These explanations need not deny recurrent conceptual patterns, but they differ over whether stored cross-domain mappings are necessary for a particular interpretation.\n\nDedre Gentner and Brian Bowdle’s “career of metaphor” proposal, developed in work published in 2005, connected comprehension processes with conventionality. Novel metaphors may invite comparison, whereas familiar metaphors can be understood through established categories. This approach highlights a question that broad conceptual analyses can obscure: the same historical source–target relationship may not be processed in the same way at every stage of conventionalization.\n\nDeliberate Metaphor Theory, associated with Gerard Steen, distinguishes metaphor in language, thought, and communication. An expression can be linguistically metaphor-related without inviting the audience to consciously consider its source domain as a separate perspective. This distinction addresses whether every conventional metaphor requires active cross-domain comparison during comprehension. It also helps separate the existence of a conceptual pattern from the communicative decision to foreground it.\n\nDevelopmental and neural accounts face related tests of explanatory precision. Children may acquire particular word senses before mastering an analyst’s proposed broad mapping, and adults may understand familiar expressions without measurable activation of detailed source knowledge. Evidence must therefore discriminate among stored lexical meaning, temporary comparison, schematic association, and richer simulation. Central unresolved questions concern when each mechanism operates, how mappings are acquired and revised, and which findings distinguish conceptual metaphor from other forms of semantic and relational learning."
    }
  ],
  "links": [
    "metaphor",
    "linguistics",
    "aristotle",
    "poetry",
    "semantics",
    "pragmatics",
    "embodied-cognition",
    "dead-metaphor",
    "polysemy",
    "personification",
    "simile",
    "english-language",
    "historical-linguistics",
    "corpus-linguistics",
    "text-corpus",
    "brain",
    "framing-effect",
    "film",
    "sign-language",
    "translation",
    "translation-studies",
    "second-language-acquisition",
    "education",
    "large-language-model",
    "relevance-theory"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}