Skip to content

Sense and Reference

Frege’s semantic distinction between an expression’s objective mode of presentation and the object or truth-value it designates.

Version
v2 · 2026-09-06 · History
Domain-specific #
2753
Origin domain
philosophy of language
Subdomain
Fregean semantics
Aliases
Sinn and Bedeutung, Sinn–Bedeutung distinction, Fregean sense–reference distinction

Core Idea

Sense and Reference is Gottlob Frege’s two-level semantic distinction between how an expression presents something and what, if anything, the expression designates. In the terminology of his 1892 essay Über Sinn und Bedeutung, an expression expresses a sense (Sinn) and designates or refers to a reference (Bedeutung). “Reference,” “denotation,” “meaning,” and “significance” are competing translations of Bedeutung; the canonical English title fixes one translation convention without resolving every exegetical question.[1][2][3]

For a singular term, the reference is normally the object designated. The sense is the object’s mode of presentation: the way the referent is given through the expression. The descriptions “the morning star” and “the evening star” both refer to Venus, but they present Venus differently. Discovering that they co-refer can therefore be informative. Reference alone predicts no semantic difference between the two designations; sense accounts for their different cognitive value.[1][4]

Frege extends the architecture beyond proper names. The sense of a complete assertoric sentence is the thought it expresses, whereas its customary reference is a truth-value, the True or the False. The distinction also supports his treatment of opaque or indirect contexts: inside a report such as “Ada believes that …,” an expression does not simply contribute its customary referent; Frege says its indirect reference is its customary sense. This reference shift explains why substitution of co-referring names can preserve truth in an ordinary extensional sentence but fail inside a belief report.[1][3][5]

The abstraction is not the claim that every expression successfully refers, that a sense is a private mental image, or that every sense is a definite description. Its stable core is a stratified semantic relation: expression → objective mode of presentation → possible referent, with sentence-level and indirect-context consequences.

Structural Signature

The Fregean distinction coordinates nine roles:

  • expression (e) — a name, description, predicate, sentence, or other semantically significant expression;
  • customary sense σ(e) — the objective, shareable mode in which a reference is presented;
  • customary reference ρ(e) — the designated object, function, concept, or truth-value, when one exists;
  • presentation relation — the sense determines or provides a route toward at most one customary reference;
  • cognitive value — what a competent thinker can grasp or learn by encountering one mode of presentation rather than another;
  • co-reference relation — distinct expressions can have ρ(e₁)=ρ(e₂) while σ(e₁)≠σ(e₂);
  • compositional environment — the senses and references of parts contribute to the sense and reference of a larger expression;
  • sentence-level pair — a sentence expresses a thought as sense and, ordinarily, designates a truth-value as reference;
  • contextual reference level — quotation, indirect speech, and propositional-attitude contexts can shift which semantic level an embedded expression designates.

A compact diagnostic representation is

\[ e \xrightarrow{\text{expresses}} \sigma(e) \xrightarrow{\text{presents}} \rho(e), \]

where ρ can be partial: an expression may have an intelligible sense without a customary referent. For co-referring expressions (e_1,e_2),

\[ \rho(e_1)=\rho(e_2) \;\not\Rightarrow\; \sigma(e_1)=\sigma(e_2). \]

The recognition invariant is not merely “words have meaning.” It is that reference and cognitive presentation are distinct semantic roles, so sameness of reference does not guarantee sameness of thought or informativeness. A complete instance names the expression, identifies the proposed sense or route of presentation, identifies or explicitly withholds the referent, and states the compositional or cognitive consequence.

What It Is Not

Sense is not the referent. Venus is a planet; “the brightest celestial body visible near sunrise” is one way of presenting it. Confusing the object with its presentation makes informative co-reference inexplicable.

Sense is not a speaker’s private idea, image, or association. Frege explicitly distinguishes objective sense, which different people can grasp, from subjective representation (Vorstellung), which varies with individual experience. A personal memory triggered by “Venus” may accompany understanding but is not thereby the expression’s sense.[1][4]

The distinction is not identical to Saussure’s signifier–signified duality. Saussure analyzes a sign’s material or acoustic form and conceptual content, and his framework characteristically brackets or separates the external referent. Frege distinguishes two semantic roles on the content/world side while presupposing an expression: sense and reference are not simply form and concept.

It is not Millian connotation and denotation, though later terminology often encourages comparison. Nor is it the intension/extension distinction without qualification. A possible-worlds intension can model a function from worlds to extensions, but two expressions can share such a function while differing in Fregean cognitive significance. Modern intensional semantics is a successor formalism, not an automatic exact translation of Frege’s sense.[6]

It is not necessarily a descriptivist theory of names. Some influential Frege–Russell readings identify a name’s sense with descriptive content, and Kripke’s modal and epistemic arguments target descriptivist theories. McDowell and Evans develop Fregean approaches that resist equating every sense with a definite description. The sense/reference distinction can be stated without settling that interpretive dispute.[7][8][9]

It is not a causal-historical or direct-reference theory, a theory of indexicals, a theory of truth alone, or a generic claim about interpretation.

Scope of Application

The home domain is philosophy of language and philosophical logic. The distinction applies centrally to singular terms and definite descriptions, informative identity statements, sentence meaning, truth-value semantics, substitution, indirect discourse, and reports of belief or other propositional attitudes.[1][3]

The first problem is cognitive significance. If the semantic contribution of a name were exhausted by its referent, then “Hesperus is Hesperus” and “Hesperus is Phosphorus” would contribute the same objects in the same identity relation. Yet the first is trivial for a competent speaker while the second records an astronomical discovery. Different senses allow the same reference to be presented in epistemically different ways.

The second problem is substitution in opaque contexts. If Clark Kent is Superman, then substituting one name for the other preserves truth in “Clark Kent is a reporter” only when both names contribute the same person in a transparent context. “Lois believes that Clark Kent is a reporter” need not have the same truth-value as “Lois believes that Superman is a reporter.” Frege’s indirect-reference apparatus makes the embedded expressions contribute their customary senses instead of simply their customary referent.[5]

The framework also reaches formal semantics. In Frege’s function–argument analysis, complete expressions can designate objects and incomplete expressions can designate functions or concepts. Sentences receive truth-values as their references, supporting compositional extensional logic, while thoughts remain their senses. Modern formal semantics retains descendants of the sense/reference separation in intension/extension, content/semantic value, and opacity analyses, although none should be attributed to Frege without translation notes.

The node is historical and theoretical, not an endorsement of every Fregean thesis. Direct-reference, causal-historical, Russellian, neo-Fregean, possible-worlds, and two-dimensional semantic theories revise different roles. Their disagreement demonstrates that the distinction is a durable object of inquiry rather than making it disappear.

Clarity

A proposed use qualifies as Sense and Reference only if five questions can be answered:

  1. What expression is analyzed? A name, description, sentence, or embedded clause must be identifiable.
  2. What is its customary reference? Name the object, concept/function, or truth-value, or state that no referent is supplied.
  3. What is its sense? State the objective mode of presentation or cognitive content, not a private association.
  4. What contrast does the distinction explain? Typical tests are informative identity, meaningful emptiness, sentence-level thought versus truth-value, or substitution failure in indirect contexts.
  5. Which context level is active? Ordinary, quoted, or indirect/propositional-attitude occurrence must be kept separate.

Two common mistakes fail this test. First, saying “the sense is whatever the speaker pictures” substitutes subjective psychology for Frege’s publicly graspable content. Second, saying “sense is just a description” treats one contested implementation as the constitutive definition.

The term Bedeutung requires a translation label. A source using “meaning,” “denotation,” or “significance” may be discussing the same Fregean role that another translation calls “reference.” Conversely, ordinary English “meaning” may refer to the whole semantic phenomenon. Terminological mismatch should not be mistaken for theoretical disagreement.

Manages Complexity

One-level referential semantics is economical: map each expression to an object and compute sentence truth from those objects. It struggles with cognitive differences among co-referring expressions, meaningful discourse involving empty terms, and substitutions under belief or indirect speech. Sense and Reference manages these cases by splitting semantic labor.

Reference supports extensional evaluation. If two terms co-refer in a transparent extensional context, replacing one with the other should leave the sentence’s truth-value unchanged. Sense supports cognitive and intensional discrimination. Two terms can guide a thinker to the same object through different information and therefore contribute differently to what is understood or believed.

At the sentence level, many different thoughts can share one of only two truth-values. Treating the thought as sense explains why true sentences are not all synonymous merely because they designate the True. Treating the truth-value as reference preserves Frege’s compositional account of extensional logic.[10]

The price is additional structure. The theory must individuate senses, explain how they determine reference, handle sense without reference, and prevent a hierarchy of indirect senses from becoming unmanageable. Later philosophers dispute whether Frege gave adequate criteria for sameness of sense and whether his treatment of empty names, names, demonstratives, and attitude reports succeeds. The distinction organizes the problems; it does not make every solution costless.

Abstract Reasoning

The role structure licenses several characteristic inferences.

Co-reference without cognitive equivalence. If (e_1) and (e_2) designate the same object through different senses, an identity (e_1=e_2) can be informative even though both sides have the same reference. Referential equality is therefore weaker than sameness of sense.

Transparent substitution. In an ordinary extensional sentence, replacing a term with a co-referring term preserves the sentence’s customary reference, its truth-value, provided no contextual level shift intervenes. It need not preserve the expressed thought.

Opaque substitution warning. Inside a belief, indirect-speech, or similar context, co-reference at the ordinary level does not license substitution salva veritate. One must compare the expressions’ indirect references, which in Frege’s account are their customary senses.[5]

Meaningfulness without designation. A term can present a condition or route without a successful object. “The largest integer” is intelligible although there is no largest integer. Within the standard reading of the 1892 framework, it can have sense without customary reference; a containing assertoric sentence may consequently lack an ordinary truth-value. This empty-name doctrine and its interpretation remain disputed, so it should be labeled rather than generalized dogmatically.[1][8]

Sentence stratification. If two sentences share a truth-value, it does not follow that they express the same thought. Their component senses and cognitive values may differ. Conversely, different truth-values entail a difference in sentence-level semantic reference and normally in thought.

No private-image substitution. Agreement in sense does not require identical mental imagery. Public communication depends on multiple speakers grasping a common thought through language even when their associated images differ.

Knowledge Transfer

The complete structure transfers within analytic philosophy among theories of naming, definite descriptions, identity, propositional attitudes, truth values, and compositional semantics. The target remains Fregean whenever the transfer retains the expression, objective sense, possible referent, cognitive-significance contrast, and context-sensitive reference level.

Formal semantics often recasts the distinction as intension and extension. This is disciplined transfer only if the mapping’s losses are stated. A possible-worlds intension tracks how extension varies across worlds; Fregean sense is also meant to explain cognitive value. Co-intensional expressions can still be cognitively distinct, so “sense = intension” is a modeling choice rather than an identity guaranteed by the historical theory.[6]

Knowledge representation and programming use related distinctions between an identifier’s presentation and the entity it denotes. Two database keys, URIs, or expressions can co-designate while exposing different information. This is a useful analogy and may instantiate broader Representation or Alias-to-Authority Mapping, but it is not automatically Fregean Sense and Reference unless cognitive presentation and semantic-level roles are deliberately modeled.

The structural residue—separate a representation’s access path from its target—travels widely. The Encyclopedia already contains Representation, Symbolic Representation, Interpretation, Indexicality, and Signifier–Signified Duality for those broader patterns. The historically specified Fregean apparatus remains domain-specific.

Examples

Morning star and evening star. “The morning star” and “the evening star” designate Venus. Before their co-reference was known, a competent speaker could understand both descriptions yet learn something from the identity statement. The shared planet is the reference; the morning-visible and evening-visible presentations differ in sense.[1]

Arithmetic presentation. The expressions “4” and “8/2” designate the same number through different computational presentations. “4 = 4” and “4 = 8/2” share the truth-value True, but the latter can carry a different cognitive route. The example isolates presentation difference without relying on astronomy.[3]

Sentence thought versus truth-value. “Snow is white” and “2 + 2 = 4” are both true. On Frege’s account, they have the same reference, the True, but express different thoughts and therefore have different senses. Truth-value equality is not sentence synonymy.

Belief report. Suppose Lois knows Clark Kent as a reporter but does not know he is Superman. “Lois believes that Clark Kent is a reporter” can be true while substitution yields “Lois believes that Superman is a reporter,” which can be false. The names co-refer ordinarily, but the belief context is sensitive to how the person is presented. This is a modern editorial illustration of Frege’s indirect-reference mechanism, not Frege’s own example.[5]

Empty term. “The largest integer” supplies a comprehensible mode of presentation but no integer satisfies it. It illustrates sense without customary reference under the standard reading. It does not show that every fictional or failed name has a stable sense; the relevant route must actually be graspable.

Non-example: an indexical. The word “I” changes referent with speaker. That is an Indexicality problem about contextual reference fixing. A sense/reference theory may need an account of such expressions, but speaker dependence alone is not the Fregean distinction.

Structural Tensions

Cognitive adequacy versus extensional economy. A single referent supports clean truth-functional composition; sense explains cognitive differences. Adding sense gains explanatory power but demands criteria for individuating and comparing senses.

Objectivity versus accessibility. Fregean senses are not private images, yet they explain what thinkers can know or fail to know. The theory must connect objective content to cognitive access without reducing one to the other.

Reference determination versus empty sense. Sense is often said to determine reference, but some senses lack referents. “Determines” must mean at most one object meets the presentation, not that every sense guarantees existence.

Compositionality versus opacity. Ordinary reference composition supports substitution; belief and indirect discourse trigger level shifts. The framework protects cognitive content at the cost of more elaborate contextual semantics.

Historical text versus successor formalisms. Intension/extension and possible-world semantics make the distinction mathematically tractable. Treating them as literal translations can erase Frege’s cognitive-value and objective-thought commitments.

Fregean core versus descriptivist reading. Descriptions offer a concrete account of presentation, while Kripkean and Evans–McDowell critiques expose problems with identifying all name senses with descriptions. The node preserves the level distinction without adjudicating every theory of what senses are.

Structural–Framed Character

Sense and Reference is predominantly structural, with an approximate framed-character score of 0.25.

  • Vocabulary travels: 0.25. Expression, reference, composition, substitution, and truth-value are technical; “sense” and “meaning” retain translation and ordinary-language ambiguity.
  • Evaluative weight: 0.00. Neither level is intrinsically better or worse.
  • Institutional constitution: 0.00. Philosophical usage stabilizes the terminology but does not create instances by authority.
  • Human-practice dependence: 0.25. Linguistic conventions and cognitive grasp matter, while the proposed semantic relations are treated as objective.
  • Import versus recognition: 0.75. Applying the Fregean labels requires an interpretive theory of the expression and its presentation route.

The result remains structural because the defining work is performed by explicit semantic roles and substitution consequences. It remains domain-specific because those roles encode a particular philosophical theory rather than a neutral cross-substrate law.

Structural Core vs. Domain Accent

The portable core is target identity can remain fixed while mode of access changes. One representation presents an object under one route, another under a different route, and equality of targets does not collapse the information carried by those routes.

The domain accent supplies the identity: linguistic expressions; Sinn and Bedeutung; objective cognitive significance; names and descriptions; sentence thoughts and truth-values; Fregean composition; and customary versus indirect reference. Without those roles, the residue becomes Representation, Interpretation, Alias-to-Authority Mapping, or a generic intensional/extensional distinction.

Conversely, not every interface alias, database identifier, visualization, or alternative proof is literally a Fregean sense. Such systems may model presentation and target, but the full node requires a semantic interpretation and a stated cognitive or compositional consequence.

Symbolic Representation supplies the closest live structural parent. Frege’s framework presupposes expressions that signify under a language and then decomposes their semantic contribution into sense and reference. Because Sense and Reference is a semantic architecture operating within symbolic representation rather than a subtype of sign vehicle, the proposed DAG relation is composition, not strict subsumption.

Signifier–Signified Duality is a crucial contrast. It separates material form from conceptual content in a Saussurean sign. Sense and Reference separates an expression’s objective mode of presentation from its designated object or truth-value. The two frameworks can be compared, but their endpoints and theoretical commitments differ.

Interpretation describes recovery of meaning from a representational substrate. Indexicality explains reference through causal, existential, or contextual connection. Representation supplies the broad medium–target relation. Alias-to-Authority Mapping captures many expressions routing to one maintained referent. Each illuminates a subset of the pattern; none supplies Frege’s complete semantic stratification.

Relationships to Other Abstractions

Local relationship map for Sense and ReferenceParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Sense and ReferenceDOMAINPrime abstraction: Symbolic Representation — is part ofSymbolicRepresentationPRIME

Current abstraction Sense and Reference Domain-specific

Parents (1) — more general patterns this builds on

  • Sense and Reference is part of Symbolic Representation Prime

    Symbolic Representation supplies the closest live structural parent.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Sense and Reference sits in a sparse region of the domain-specific corpus (88th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Reference alone — the designation relation; it omits mode of presentation and cognitive value.
  • Meaning in ordinary language — an umbrella term that may cover sense, reference, use, implication, or speaker intention.
  • Signifier–Signified Duality — form versus conceptual content, not Frege’s sense versus designated object/truth-value.
  • Intension and extension — a common successor formalization whose possible-world intensions can be coarser than cognitive sense.
  • Connotation and denotation — Millian terminology with a different theory of names and general terms.
  • Descriptivism — one family of accounts of name meaning/reference; not constitutive of every Fregean sense theory.
  • Direct reference or causal-historical reference — rival or revisionary accounts emphasizing designation without Fregean descriptive mediation.
  • Indexicality — context or causal-existential linkage fixing a referent, not the general two-level distinction.
  • Polysemy — one expression with multiple related senses; Sense and Reference centrally allows multiple senses to converge on one referent.
  • Private association or mental image — subjective representation, which Frege distinguishes from objective sense.
  • Use–mention distinction — difference between using an expression and referring to the expression itself.
  • The essay Über Sinn und Bedeutung — the foundational publication; the abstraction is the semantic distinction developed in it, not the document as an artifact.

References

[1] Gottlob Frege, “Über Sinn und Bedeutung,” Zeitschrift für Philosophie und philosophische Kritik 100 (1892): 25–50; scan of the original. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g

[2] Gottlob Frege, “Sense and Reference,” translated by Max Black, The Philosophical Review 57, no. 3 (1948): 209–230; accessible translation. registry

[3] Edward N. Zalta, “Gottlob Frege,” Stanford Encyclopedia of Philosophy, especially sections 3.1–3.2 on Frege’s puzzles and sense/denotation. registry ↩a ↩b ↩c ↩d

[4] Jeff Speaks, “Theories of Meaning,” Stanford Encyclopedia of Philosophy, section on Fregean semantics. registry ↩a ↩b

[5] Michael Nelson, “Propositional Attitude Reports,” Stanford Encyclopedia of Philosophy, sections on Fregean indirect reference and substitution. registry ↩a ↩b ↩c ↩d

[6] Melvin Fitting, “Intensional Logic,” Stanford Encyclopedia of Philosophy, sections on Fregean sense/reference and later intensional semantics. registry ↩a ↩b

[7] John McDowell, “On the Sense and Reference of a Proper Name,” Mind 86, no. 342 (1977): 159–185. registry

[8] Gareth Evans, The Varieties of Reference, edited by John McDowell (Oxford: Clarendon Press, 1982); see also McDowell’s Routledge Encyclopedia summary. registry ↩a ↩b

[9] “Reference,” Stanford Encyclopedia of Philosophy, sections on descriptivist and causal-historical theories of names. registry

[10] Vladimir Shramko and Heinrich Wansing, “Truth Values,” Stanford Encyclopedia of Philosophy, section 1 on Fregean truth-values. registry