The correspondence theory of truth holds that a statement, belief, or other truth-bearing representation is true because it agrees with reality. A belief that snow is white is true if snow is white; it is false if snow is not white. The central idea is that truth depends on how things are, rather than merely on what anyone believes, what a community accepts, or what a system of beliefs consistently maintains. More developed versions explain this dependence through relations between representations and facts, objects, properties, or states of affairs.
Correspondence is one of the oldest and most influential approaches to truth in philosophy. Its ancestry extends to ancient Greek accounts of true and false assertion, while its modern forms emerged through debates about judgment, language, and the structure of reality. It occupies a central place in epistemology, metaphysics, and the philosophy of language. Its apparent simplicity conceals difficult questions: what bears truth, what in reality makes it true, and what “correspondence” contributes beyond saying that a statement is true if things are as it states.
The central claim and its varieties
The basic correspondence intuition distinguishes a representation from what it represents. A map can misrepresent a coastline, a witness can misdescribe an event, and a scientific hypothesis can describe a process that does not occur. In each case, correctness is constrained by something other than the representation itself. Applied to truth, the thought is that asserting that a bridge has collapsed does not make the assertion true: its truth depends on the bridge’s having collapsed. Correspondence theories turn this familiar distinction into an account of the nature of truth.
A general formulation is:
Here stands for a truth-bearer, often a proposition. This formula is a starting point rather than an explanation, because both “corresponds” and “reality” require interpretation. Some theories identify a particular fact corresponding to each true proposition. Others analyze truth through the objects referred to and the properties or relations attributed to them. Still others understand correspondence broadly as the dependence of truth on how things stand, without accepting facts as a separate category of entities.
A narrow correspondence theory therefore differs from a broad correspondence conception. The narrow theory supplies an explicit relation between truth-bearers and truth-making items, perhaps with a structural account of how the two fit together. The broad conception says only that truth is determined by reality rather than by belief or justification. Many philosophers accept the broad conception while rejecting particular theories of facts or relations of correspondence. Disagreement over whether a philosopher is a correspondence theorist often reflects these different uses of the label.
Correspondence need not mean visual resemblance. A sentence about a mountain need not look like a mountain, and an algebraic description of motion need not resemble a moving body. The relevant agreement is usually representational or structural: the sentence attributes a property that the object actually possesses, or arranges names in a way that expresses an obtaining relation. Nor does correspondence require an observer-independent realm in every case. A statement about someone’s pain may be true because that person experiences pain; a statement about a legal institution may be true because certain socially constituted conditions obtain.
The theory also does not by itself provide a procedure for discovering truth. Whether a proposition is true and whether there is adequate evidence for accepting it are different questions. A belief can be true without being well supported, and a well-supported belief can be false. This separation gives correspondence its characteristic combination of objectivity and epistemic modesty: inquiry aims at a condition that inquiry itself does not automatically guarantee.
Ancient and medieval antecedents
Ancient discussions supplied influential formulations of the correspondence intuition without always presenting a distinct theory under that name. In the Sophist, Plato examines how meaningful falsehood is possible. A statement combines a name with a verb and says something about something. “Theaetetus sits” and “Theaetetus flies” can both be meaningful, although the latter is false in the situation discussed. The analysis shifts attention from isolated words to assertions and their agreement or disagreement with the condition of their subject.
The most frequently cited ancient formulation comes from Aristotle in Metaphysics, Book IV, chapter 7: to say of what is that it is not, or of what is not that it is, is false; to say of what is that it is, and of what is not that it is not, is true. This identifies truth through the agreement of assertion with being and nonbeing. In Categories, chapter 5, Aristotle also observes that a statement can change from true to false because the relevant thing changes, without the statement itself undergoing a corresponding alteration. A statement that someone is sitting becomes false when that person stands.
These passages establish important asymmetries. Things are not ordinarily as they are because a statement about them is true; rather, the statement is true because things are as it says. Yet Aristotle should not simply be credited with every feature of later correspondence metaphysics. His accounts involve predication, combination, and separation, not a fully developed modern theory in which abstract propositions stand in a uniform relation to independently individuated facts. The historical continuity lies principally in the priority of how things are over how they are asserted to be.
Medieval discussions developed this inheritance within theories of intellect, being, and divine knowledge. Thomas Aquinas, in De veritate and the Summa theologiae, characterizes truth through the conformity or adequation of intellect and thing. The familiar Latin expression adaequatio rei et intellectus denotes this relation. For human knowledge, the intellect is measured by the thing known: judgment is true when it conforms to its object. Aquinas locates truth primarily in the intellect, particularly in its acts of combining and separating, while also discussing ways in which things themselves are called true.
This account belongs to a theological framework that complicates any simple opposition between mind and world. Natural things, for Aquinas, are measured by the divine intellect, while artifacts are measured by the conception of their human makers. Human speculative understanding, by contrast, is measured by its objects. Medieval correspondence language thus covered several directions of dependence. Its contribution to later theories was not merely a slogan about matching, but a systematic distinction between the truth of judgment, the intelligibility of things, and the standards by which each is measured.
Modern philosophy and the rise of analytic formulations
Early modern philosophers often treated agreement between thought and its object as the nominal meaning of truth while disagreeing about how such agreement could be known. René Descartes investigated the reliability of clear and distinct perception within a broader account of knowledge and error. John Locke, in An Essay Concerning Human Understanding (1690), connected truth with the joining or separating of signs according to the agreement or disagreement of what they signify. These approaches retained a world-directed conception of truth while placing growing emphasis on ideas, representation, and the conditions of judgment.
Immanuel Kant accepted the nominal definition of truth as the agreement of cognition with its object in the Critique of Pure Reason (1781; second edition 1787). He denied, however, that this definition supplies a universal material criterion of truth. Such a criterion would have to disregard the differences between objects while nevertheless determining whether cognition agrees with them. Kant distinguished general logical conditions, which concern the form of thinking, from the object-specific conditions needed for knowledge. His transcendental idealism also makes his position substantially different from an unrestricted realist correspondence theory.
The modern debate took a distinctive form with the emergence of analytic philosophy. Gottlob Frege questioned whether truth could be defined as correspondence. Any proposed definition seemed to require determining whether the relevant correspondence itself truly obtained, threatening an explanatory regress. Frege also emphasized thoughts as bearers of truth rather than private mental images. His criticism helped establish that accepting objective truth does not necessarily entail accepting a relational definition of truth.
Bertrand Russell developed more explicitly correspondence-oriented accounts. In The Problems of Philosophy (1912), he argued that a theory must allow both truth and falsehood and must explain their dependence on something outside belief. His accounts changed over time, including changes in the analysis of judgment, but he repeatedly connected true belief with appropriately arranged objects and relations. In his 1918 lectures on logical atomism, facts played a central role as features of the world that make propositions true or false.
Ludwig Wittgenstein offered a structural version in the Tractatus Logico-Philosophicus, published in German in 1921 and in a German–English edition in 1922. A proposition pictures a possible arrangement of objects; it is true when that possibility agrees with reality. Its elements need not resemble objects perceptually, but their configuration must share the relevant logical form. In 1950, J. L. Austin proposed a different account in “Truth,” distinguishing conventions that connect sentence types with types of situation from conventions that connect particular statements with particular historical situations. These approaches show why correspondence is a family of theories rather than one unchanged doctrine.
Truth-bearers, facts, and the correspondence relation
A developed theory must first identify what can be true or false. Candidates include sentences, utterances, beliefs, judgments, and propositions. Sentences are publicly available linguistic objects, but the same sentence can express different contents in different contexts. “The meeting begins now” depends on who speaks, when, and which meeting is intended. Propositions are often introduced as the contents expressed across such contextual variations. Two sentences in different languages may express the same proposition, while two utterances of the same sentence may express different propositions. Choosing a truth-bearer therefore affects how correspondence is formulated.
The second question concerns what lies on the worldly side of the relation. A fact is often understood as an obtaining state of affairs: not merely a person and a property, but that person’s actually possessing that property. The fact that a particular apple is green differs from the apple considered alone. A state of affairs is frequently treated as something that may obtain or fail to obtain, whereas a fact is, by definition, obtaining. Terminology varies, and some philosophers reject the distinction or avoid both categories. The underlying issue is whether truth requires structured entities in addition to ordinary objects and their characteristics.
Structural correspondence promises an informative explanation. The proposition that a particular apple is green contains, or otherwise represents, an object and a property; the corresponding fact involves that object’s instantiating that property. Relational propositions make the attraction clearer. “Anna precedes Ben” and “Ben precedes Anna” involve the same individuals and relation but represent different arrangements. A theory must account for this difference rather than merely list matching components. The problem is especially important when propositions and facts are treated as complexes: what unifies their constituents into one representation or one obtaining configuration?
This approach also faces a problem concerning falsehood. If a true proposition corresponds to a fact, a false proposition cannot ordinarily correspond to an existing false fact. One response says that a false proposition represents a possible state of affairs that fails to obtain. Another analyzes falsehood through the failure of the relevant objects to satisfy the asserted relation. Both approaches must explain representation independently enough to allow a meaningful proposition to be false. Otherwise, the theory risks explaining only successful representation.
Finally, correspondence must be distinguished from causal connection and resemblance. A photograph may be caused by a scene but misrepresent its colors; two unrelated objects may resemble one another without either representing the other. Reference, predication, context, and linguistic convention all contribute to determining what a statement says. Correspondence theories consequently interact with theories of meaning rather than replacing them. Some related accounts identify true propositions with facts instead of relating two distinct entities. These identity theories preserve a close connection between truth and reality while rejecting correspondence understood strictly as an external relation.
Tarski, satisfaction, and formal truth definitions
Alfred Tarski transformed the study of truth by showing how truth could be defined for specified formal languages under carefully controlled conditions. His major work appeared in Polish in 1933 and in an expanded German version in 1935. Tarski aimed to articulate the classical conception according to which a sentence is true when things are as the sentence says. His resulting semantic theory of truth is closely associated with correspondence, although whether it constitutes a full correspondence theory remains disputed.
Tarski’s condition of material adequacy requires a truth definition to yield appropriate instances of what is commonly called the T-schema:
The expression on the left names a sentence, while the expression on the right uses a sentence to state its truth condition. More generally, an appropriate biconditional connects the truth of a named object-language sentence with a translation of that sentence into the metalanguage. The schema is not itself a single exhaustive definition of truth. Rather, it specifies consequences that an adequate definition for a language must deliver.
The distinction between object language and metalanguage is essential. The object language is the language whose truth predicate is being defined; the metalanguage is the language in which the definition is given. Unrestricted self-application of a truth predicate can generate the liar paradox, as in a sentence that says of itself that it is not true. Tarski’s formal treatment avoids this difficulty by restricting the expressive resources available at each level. His work thereby connects the theory of truth with limitations on what sufficiently expressive languages can consistently say about their own semantics.
For languages containing variables and quantifiers, Tarski defines truth using satisfaction. An assignment associates variables with objects in a domain. An atomic formula is satisfied when the assigned objects stand in the relation interpreted by its predicate. Satisfaction for compound formulas is then specified recursively: a negation is satisfied when its component is not; a conjunction when both components are; and a universally quantified formula when the relevant condition holds under every required variation of the assignment. For a sentence without free variables, satisfaction does not depend on the choice of assignment.
This framework explains how the truth conditions of indefinitely many sentences can arise from finite semantic rules. Yet it does not automatically settle the metaphysics of correspondence. Satisfaction clauses can describe relations between language and objects without invoking facts as special entities. Conversely, a correspondence theorist can interpret those clauses as a precise articulation of world-directed truth conditions. A model also interprets symbols, so truth in a model must be distinguished from truth under a language’s intended interpretation. Tarski supplies powerful formal resources, but additional philosophical commitments determine what those resources establish about reality.
Negative, general, modal, and other difficult truths
Simple affirmative predications make correspondence appear straightforward. The statement that a particular door is open seems to depend on that door’s condition. More difficult cases arise when no obvious positive item corresponds to the whole assertion. A negative statement such as “There are no elephants in this room” raises the question whether its truth requires a negative fact, the room’s complete positive condition, or some other explanation. Merely pointing to the objects present may seem insufficient unless their being all the relevant objects is also secured.
Universal statements create a related difficulty. “Every sample in this batch contains copper” is not ordinarily made true by one sample, and listing confirming samples does not establish universality unless the list is exhaustive. Some correspondence accounts accept general facts. Others analyze general truth through all relevant instances plus a completeness or totality condition. Existential statements are often easier: a suitable particular instance can suffice for “Some sample contains copper.” Even here, explaining the truth of the whole statement requires an account of how the instance supports its quantified content.
These problems are central to truthmaker theory, which investigates what in reality makes truths true. A truthmaker need not reproduce the structure of a proposition or stand in a relation described as correspondence. One event may make many different propositions true, and one proposition may have several truthmakers. Truthmaker maximalism holds that every truth has a truthmaker; weaker positions restrict the requirement. Negative truths, truths about the past, and modal truths are important tests because they make unrestricted truthmaker requirements ontologically demanding.
Truths involving modality and counterfactuals extend beyond what actually occurs. “This glass could have broken” cannot be explained simply by a corresponding actual breaking. Theories may appeal to dispositions, essences, laws, or possible worlds. David Lewis developed an influential possible-worlds treatment of modality and counterfactuals, but correspondence theorists need not accept his ontology of concrete possible worlds. The important point is that a commitment to reality-constrained truth leaves open substantial disagreement about what modal reality contains.
Mathematical, ethical, and fictional discourse introduce further variations. A realist about mathematics may explain mathematical truth through abstract structures, whereas a nominalist seeks an account without such entities. In metaethics, correspondence becomes substantive when paired with claims about moral facts, but it does not independently establish their existence. Fiction requires distinguishing truth within a story from literal assertions about the actual world. Statements about the past raise questions about whether past events exist, leave present truthmakers, or require a different treatment. These cases demonstrate that the general correspondence slogan underdetermines a theory’s ontology: difficult truths expose commitments that elementary examples conceal.
Objections concerning access, explanation, and representation
A persistent epistemological objection asks how a representation could ever be compared with reality except through further representations. If access to the world always involves perception, concepts, or language, the critic argues, there is no independent standpoint from which thought and reality can be placed side by side. The correspondence theorist typically replies that knowing an object through perception does not entail knowing only a representation of it. Moreover, a theory of what truth is need not provide an infallible test for recognizing it. The objection has greatest force when correspondence is presented simultaneously as a definition and as an immediately usable criterion.
This reply does not eliminate questions about evidence. Beliefs are evaluated through observation, inference, testimony, and instruments, all of which can fail. Philosophical skepticism exploits the possibility that convincing evidence might accompany widespread error. Correspondence theories generally allow that possibility rather than defining it away. They can be combined with fallibilism, according to which knowledge or justified belief need not rest on guarantees against error. Their vulnerability is therefore not simply that mistakes occur, but that they must explain how fallible practices nevertheless connect inquiry with the world.
Another objection concerns explanatory circularity. If a fact is identified merely as whatever makes a proposition true, saying that the proposition is true because it corresponds to a fact may add little. An explanation of “The door is open” that invokes “the fact that the door is open” can appear to restate the original assertion. Correspondence theorists answer by supplying independent accounts of objects, properties, relations, and representational content. Whether those accounts are successful depends on their detail; adding the word “fact” does not by itself explain truth.
A different challenge targets the individuation of corresponding facts. Arguments commonly called slingshot arguments attempt, under particular assumptions about reference, substitution, and definite descriptions, to show that all true sentences refer to the same entity or correspond to the same fact. Donald Davidson used related considerations against explaining truth through distinct facts matched to sentences. The argument’s force depends on its logical and semantic assumptions. Correspondence theorists can reject some substitutions or deny that sentences refer to facts in the required way. Nevertheless, the challenge exposes the need for a principled account of fact identity.
Finally, conceptual mediation raises the question whether reality determines one uniquely correct description. The same region can be described geographically, politically, ecologically, or physically without those descriptions necessarily competing. A correspondence account can allow multiple true descriptions, provided their contents are compatible and their conditions obtain. It need not assume that reality arrives already divided according to every vocabulary used to describe it. The harder issue is how representation remains constrained when classifications are interest-dependent, socially variable, or theoretically revisable. This connects debates about truth to disputes over reference, conceptual schemes, and interpretation.
Coherence, pragmatism, deflationism, and realist disagreement
The coherence theory of truth associates truth with membership in an appropriately coherent system of judgments. Coherence involves more than the absence of contradiction: it may include mutual support, explanatory integration, and systematic completeness. Correspondence theorists object that more than one incompatible system might satisfy such requirements, and that an internally coherent fiction need not be true. Coherence theorists, in turn, question whether correspondence can be understood independently of the networks of belief through which objects are identified. A coherence theory of truth must also be distinguished from coherentism about justification, which can coexist with correspondence truth.
Pragmatic theories of truth connect truth with inquiry, practical consequences, or warranted assertion. Charles Sanders Peirce associated truth with the opinion toward which inquiry would ultimately converge, rather than with whatever happens to be useful or currently accepted. Other pragmatist approaches emphasize different relations between truth and the conduct of inquiry. Correspondence critics of pragmatism ask whether an ideal limit of inquiry could remain mistaken or whether some truths might never become discoverable. Pragmatists question whether a truth wholly detached from possible inquiry has an intelligible role.
Deflationism about truth takes a different approach. Rather than replacing correspondence with another substantial property, it argues that the truth predicate performs a comparatively modest logical or expressive function. Saying that “Snow is white” is true allows, in an important sense, the assertion that snow is white; more generally, truth enables endorsement of statements that cannot all be individually repeated, as in “Everything the witness said is true.” Deflationists dispute whether an additional metaphysical relation is needed to explain this function. Correspondence theorists reply that explaining the use of “true” may not fully explain the dependence of representations on reality.
The boundary between correspondence and deflationism is nevertheless contested. A deflationist can accept every ordinary claim that a statement is true because things are as it says, while resisting a general theory of correspondence. Conversely, a modest correspondence theorist may reject elaborate facts and structural matching, leaving a view close to minimalist accounts. The disagreement then concerns whether world-dependence is a substantial explanatory thesis or simply follows from understanding the relevant sentences and their truth conditions. Terminological agreement about “correspondence” can therefore conceal philosophical disagreement, while terminological disagreement can exaggerate it.
Debates over realism introduce another distinction. Scientific realism concerns, among other matters, whether successful scientific theories describe unobservable features of a mind-independent world. Correspondence is a theory of truth, not by itself a commitment that any particular theory is true. Similarly, rejecting verificationism does not establish an ontology of facts. Discussions associated with Michael Dummett connected realism with truth conditions that may outrun the ability to recognize them. This approach sharpened the contrast between truth and verification, but neither correspondence nor its alternatives determine every position concerning bivalence, undecidability, or the limits of human knowledge.
Scientific inquiry, measurement, and socially constituted facts
In scientific inquiry, correspondence helps explain why successful prediction and justified acceptance remain distinguishable from truth. A theory may fit available observations yet misdescribe the mechanism responsible for them. Conversely, an initially unsupported conjecture may later prove correct. Karl Popper defended an objective conception of truth influenced by Tarski while arguing that empirical science advances through conjecture and criticism rather than final verification. Correspondence supplies a target for inquiry; methods of testing concern how proposals are assessed against that target.
Scientific representations are frequently idealized. Classical mechanics treats some bodies as point masses and some surfaces as frictionless, even though actual systems do not exactly instantiate these descriptions. This creates a distinction between literal truth, approximate truth, and usefulness within a domain. Correspondence does not automatically provide a measure of approximation. Measuring error in a predicted quantity is often straightforward, but comparing the overall truthlikeness of theories with different structures is considerably harder. Popper’s early formal proposals for verisimilitude encountered technical difficulties identified in the 1970s, illustrating that “closer to truth” requires more than an intuitive appeal to correspondence.
Measurement introduces further qualifications. An assertion that a sample has a given temperature depends on the meaning of the measured quantity, the physical conditions, and the specification of units and uncertainty. Calibration connects instrument indications with reference standards, while measurement uncertainty characterizes the dispersion of values reasonably attributable to the measurand under the adopted procedure. These practices do not redefine truth as agreement among instruments. They provide controlled ways of obtaining evidence about quantities while acknowledging limits of resolution, modeling, and inference. An exact-value claim and a claim about a stated interval have different truth conditions.
Social and institutional facts show why correspondence should not be equated with the claim that all facts exist independently of human activity. Whether a legislature passed a bill, a person holds an office, or a document counts as legal tender depends partly on rules and collective practices. John Searle analyzed institutional facts through the assignment of status functions, often expressed by the pattern that something counts as something else in a context. Once the relevant practices and events are specified, a statement about them can be true or false regardless of a particular speaker’s preferences.
This distinction also matters in history and law. Evidence can be incomplete, testimony misleading, and records destroyed, yet an assertion that a particular event occurred need not acquire its truth from the surviving evidential record. In legal settings, procedural standards determine what a court may find established; they are not automatically identical with the conditions under which the underlying allegation is true. Conversely, some legal facts, such as the validity of an enactment, partly depend on recognized procedures themselves. Correspondence accommodates both kinds of dependence only when the relevant subject matter is carefully identified: facts about physical events, evidential findings, and institutional status need not have the same structure.