Semantics

Semantics is the branch of linguistics that studies meaning in language: what words mean, how meanings combine into larger units, and how context and use shape what is communicated.

13 frameworks, from 1880 to today · Linguistics · Humanities

still a working view ended · each leads to its entry below

Overview

Semantics is the branch of linguistics that studies meaning in language: what words mean, how meanings combine into larger units, and how context and use shape what is communicated. It matters because meaning is what language is for—the sounds and structures of a language are only interesting because they carry significance. The history of semantics is a history of attempts to make meaning tractable: to find a level of description at which the slippery, context-dependent, apparently mental phenomenon of meaning can be studied systematically. That ambition has produced several competing traditions, and the tensions among them remain unresolved.

From Philology to the Science of Meaning

Systematic reflection on meaning is ancient—the Sanskrit grammarian Pāṇini (roughly 4th century BCE) distinguished word meaning from sentence meaning, and the Stoics in Greece debated whether meaning was a property of words or of utterances. But semantics as a modern linguistic subfield emerged in the late 19th and early 20th centuries, when scholars began treating meaning change as a systematic phenomenon. The French philologist Michel Bréal coined the term sémantique in the 1880s and 1890s to describe the study of how meanings evolve1. This early work was largely historical: it catalogued regular patterns of semantic change—narrowing, broadening, pejoration, amelioration—and treated meaning as something that words "have" and gradually lose or acquire.

The Swiss linguist Ferdinand de Saussure, whose lectures (published posthumously in 1916 as the Cours de linguistique générale) founded structuralism, reframed the problem23. For Saussure, a linguistic sign was a union of a signifier (sound pattern) and a signified (concept), and meaning was not a positive property of a word but a matter of difference: what a word means is determined by what it contrasts with. This relational view made meaning a property of the language system rather than of individual words or mental images, and it set the agenda for structuralist semantics.

Structuralist Lexical Fields and Componential Analysis

In the mid-20th century, structuralist linguists developed the idea of the lexical field: a set of words that cover a conceptual domain and divide it up in language-specific ways. The German linguist Jost Trier, working in the 1930s, argued that words like wisdom, knowledge, and skill in Middle High German formed a field whose internal boundaries shifted over time. The approach was extended by others, notably in the analysis of kinship terms and color vocabulary. The key claim was that the meaning of a word is determined by its position in a field—by the words it contrasts with—rather than by any intrinsic content.

A related method, componential analysis, broke word meanings into smaller features12. The anthropologist Ward Goodenough and others analyzed kinship terms in the 1950s by decomposing them into features like [±male], [±lineal], [±senior]. The linguist Jerrold Katz and Jerry Fodor applied a similar approach to English in the 1960s, proposing that the meaning of bachelor could be represented as [human], [male], [adult], [never married]. Componential analysis was elegant but ran into trouble: it worked well for tightly structured domains like kinship and color, but many words resisted decomposition, and the features themselves seemed to require interpretation. The approach also struggled to explain how context affects meaning—bachelor can describe a knight, a young seal, or a person with a degree.

Truth-Conditional Semantics and the Formal Turn

The most consequential development in 20th-century semantics came from logic and philosophy. In the 1930s and 1940s, Alfred Tarski and Rudolf Carnap developed formal theories of truth and reference. The Polish logician Alfred Tarski showed how to define truth for formal languages in terms of satisfaction conditions, and the philosopher Donald Davidson proposed in the 1960s that a theory of meaning for a natural language could take the form of a Tarski-style truth theory: to know the meaning of a sentence is to know the conditions under which it is true.

This truth-conditional approach was radical: it treated meaning as a relation between language and the world, not between language and the mind. It also provided a rigorous method. The philosopher Richard Montague, working in the late 1960s and early 1970s, showed that a substantial fragment of English could be given a precise model-theoretic semantics45. His Montague grammar treated English expressions as having types (entities, truth values, functions between them) and used the machinery of intensional logic to handle meaning15. The sentence Every student reads a book could be assigned a precise truth condition, and the grammar could be extended compositionally—the meaning of a complex expression is a function of the meanings of its parts and their syntactic combination146.

Montague's work was technically demanding but enormously influential. It founded formal semantics as a linguistic discipline, and by the 1980s it had become the dominant approach in the English-speaking world. The key insight was that natural language meaning could be modeled with the same precision as the meaning of formal logic, and that syntax and semantics could be linked systematically.

Presupposition and the Limits of Truth Conditions

Truth-conditional semantics had a problem: it seemed to miss something7. Consider the sentence The king of France is bald. If there is no king of France, is the sentence false, or does it fail to have a truth value? The philosopher Peter Strawson argued in 1950 that such sentences presuppose the existence of their referents, and that presupposition failure is different from falsity. The linguist and philosopher Gottlob Frege had made a similar point earlier, but Strawson's formulation sparked a long debate.

In the 1970s, linguists like Lauri Karttunen and David Lewis developed a systematic theory of presupposition. They showed that certain words—stop, regret, again, even—carry presuppositions that survive negation and questioning. John stopped smoking presupposes that John used to smoke; John didn't stop smoking carries the same presupposition8. This was a problem for truth-conditional semantics, which treated negation as a simple truth-function. Presupposition theory showed that meaning involves more than truth conditions: it involves background assumptions that speakers take for granted.

Grice and the Pragmatic Turn

The philosopher H. P. Grice made a different kind of intervention9. In his 1957 paper "Meaning" and his 1967 lectures (published as "Logic and Conversation" in 1975), Grice distinguished sentence meaning from speaker meaning. He argued that communication works through implicature: when a speaker says something that appears to violate a conversational norm, the hearer infers an intended meaning10. If someone says Some of the students passed, the hearer infers that not all passed—not because some means "not all," but because if the speaker had known that all passed, they would have said so. This scalar implicature is a classic case.

Grice's cooperative principle and its maxims (quantity, quality, relation, manner) provided a framework for explaining how meaning arises from use911. The Gricean implicature approach was a major influence on pragmatics and on the semantics-pragmatics boundary. It showed that truth-conditional semantics alone could not explain how speakers communicate, and it opened the door to a more context-sensitive view of meaning.

Cognitive Semantics: Meaning in the Mind

In the 1970s and 1980s, a group of linguists and cognitive scientists—George Lakoff, Mark Johnson, Ronald Langacker, Charles Fillmore, and others—developed an alternative to formal semantics. Cognitive semantics treats meaning as a mental phenomenon, grounded in embodied experience and conceptual structure. Lakoff and Johnson's Metaphors We Live By (1980) argued that metaphor is not a literary device but a fundamental mode of thought: we understand argument as war (he attacked my position, she defended her claim), and this shapes how we reason12.

The psychologist Eleanor Rosch's work on prototype theory in the 1970s showed that categories have graded structure: some members are better examples than others13. A robin is a better example of a bird than a penguin, and this affects how people reason and how words are used. The linguist Charles Fillmore's frame semantics (developed from the 1970s onward) argued that words evoke frames—structured background knowledge—and that understanding a word means understanding its frame. The word buy evokes a commercial transaction frame involving a buyer, a seller, goods, and money; you cannot understand buy without understanding the frame14.

Construction grammar, developed by Fillmore, Adele Goldberg, and others from the 1980s, extended this to syntax: grammatical constructions are themselves meaningful units, not just formal templates. The way-construction (He elbowed his way through the crowd) has a meaning—"to move by means of the action denoted by the verb"—that is not predictable from its parts. Cognitive semantics and construction grammar share a commitment to usage-based linguistics: meaning is learned from and shaped by actual language use, not by innate rules.

Dynamic Semantics and Context

In the 1980s and 1990s, formal semanticists began to take context seriously. Dynamic semantics, developed by Irene Heim, Hans Kamp, and others, treated meaning as a relation between contexts: a sentence updates the context in which it is uttered115. This was a way to handle pronouns and definiteness. In A man walked in. He sat down, the pronoun he refers to the man introduced by the indefinite a man—but only because the first sentence updates the context to make the man available for reference. Dynamic semantics modeled this as a process of context change, and it provided a formal treatment of presupposition and anaphora that static truth-conditional semantics struggled with.

Distributional Semantics and the Computational Turn

A different tradition emerged from computational linguistics and information retrieval. Distributional semantics rests on the idea—often summarized by the linguist John Firth's 1957 remark that "you shall know a word by the company it keeps"—that words with similar distributions in text have similar meanings. In the 1990s and 2000s, researchers like Hinrich Schütze and the developers of latent semantic analysis (LSA) used statistical methods to represent word meaning as vectors in a high-dimensional space, derived from co-occurrence patterns in large corpora.

This approach was computationally powerful but linguistically controversial: it captured similarity and association but not compositionality or truth conditions. The WordNet database, developed by George Miller and colleagues at Princeton from the 1980s, took a different computational approach: it organized words into synsets (sets of synonyms) linked by relations like hyponymy (is-a) and meronymy (part-of)16. WordNet became a standard resource for natural language processing, and its relational model influenced later lexical databases.

The Present: Pluralism and Open Questions

Semantics today is pluralistic. Formal semantics remains dominant in theoretical linguistics, with ongoing work on intensionality, quantification, and the semantics-pragmatics interface. Cognitive semantics and construction grammar have a strong following, particularly in Europe and among usage-based researchers. Distributional semantics has been transformed by neural word embeddings and large language models, which learn meaning representations from massive text corpora—raising new questions about what such representations capture and whether they constitute meaning in any linguistically relevant sense.

The old tensions remain: between meaning as truth conditions and meaning as mental representation; between meaning as a property of the language system and meaning as a product of use; between formal rigor and psychological plausibility1. Presupposition, implicature, and context-dependence are now central topics, and the boundary between semantics and pragmatics is actively debated. What unites the field is the conviction that meaning is systematic enough to be studied rigorously, and that the study of meaning is essential to understanding language.

References

  1. Semantics (Definition and related fields, History, Others, Formal semantics, Compositionality, introduction), Wikipedia
  2. Structural semantics, Wikipedia
  3. Structural linguistics (introduction, History), Wikipedia
  4. Formal semantics (natural language) (History), Wikipedia
  5. Montague grammar, Wikipedia
  6. Principle of compositionality, Wikipedia
  7. Truth condition, Wikipedia
  8. Presupposition, Wikipedia
  9. Paul Grice, Wikipedia
  10. Implicature, Wikipedia
  11. Cooperative principle, Wikipedia
  12. Conceptual metaphor (introduction, Background), Wikipedia
  13. Prototype theory, Wikipedia
  14. Frame semantics (linguistics), Wikipedia
  15. Dynamic semantics, Wikipedia
  16. WordNet (introduction, History and team members), Wikipedia

The frameworks of Semantics

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1880-1930
  2. 1930-1970

    Structural semantics

    How is meaning determined by contrasts within a language system?

    Fly to its tower in Nousfarer

  3. 1950-1975

    Componential Analysis

    Can word meanings be broken into smaller features?

    Fly to its tower in Nousfarer

  4. 1965-Present

    Truth-conditional semantics

    What are the truth conditions of a sentence?

    Fly to its tower in Nousfarer

  5. 1967-Present

    Gricean Pragmatics

    How do speakers convey more than they say?

    Fly to its tower in Nousfarer

  6. 1970-Present

    Montague grammar

    Can natural language meaning be modeled formally?

    Fly to its tower in Nousfarer

  7. 1970-Present

    Presupposition

    What background assumptions do sentences carry?

    Fly to its tower in Nousfarer

  8. 1975-Present

    Cognitive Semantics

    How is meaning grounded in the mind and body?

    Fly to its tower in Nousfarer

  9. 1975-Present

    Frame semantics

    What background knowledge do words evoke?

    Fly to its tower in Nousfarer

  10. 1980-Present

    Dynamic Semantics

    How does meaning change as discourse unfolds?

    Fly to its tower in Nousfarer

  11. 1985-Present

    Construction Grammar

    Are grammatical constructions meaningful units?

    Fly to its tower in Nousfarer

  12. 1985-Present

    Lexical-Relational Resources

    How are words related to each other in the lexicon?

    Fly to its tower in Nousfarer

  13. 1990-Present

    Distributional Semantics

    Can meaning be inferred from patterns of word use?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Historical-Philological SemanticsStructural semanticsComponential AnalysisTruth-conditional semanticsGricean PragmaticsMontague grammarPresuppositionCognitive SemanticsFrame semanticsDynamic SemanticsConstruction GrammarLexical-Relational ResourcesDistributional Semantics

Beyond the field

Where Semantics's ideas came from in other fields, and what they led to there: the strongest 12 of 40 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Semantics is a planet in Nousfarer

Its city has a tower for each of these 13 frameworks. Land, walk in, and climb one concept at a time.

Land on Semantics

Related fields

Fields its ideas are linked to

More in Linguistics

See where Semantics sits on the Map of Human Knowledge