Frege's Puzzles

The paired semantic challenge that co-referring expressions can differ in cognitive value and can fail to substitute truth-preservingly inside propositional-attitude reports.

Version

v3 · 2026-09-19 · History

Domain-specific #

1887

Domain group

Humanities

Origin domain

Philosophy

Subdomains

Philosophy of Language, Theories of Reference and Propositional Attitudes → Philosophy

Aliases

Frege's puzzle, Fregean puzzles, Frege's identity puzzle, Frege's puzzle about belief

Core Idea

Frege's Puzzles name a paired explanatory challenge for theories of linguistic meaning and reference. Two expressions can designate the same object, yet replacing one presentation with the other can change what an identity statement teaches or whether a propositional-attitude report is true. A semantics that assigns only the ordinary referent to each name appears to predict more equivalence than competent speakers actually display.

The identity form contrasts a reflexive statement such as “Hesperus is Hesperus” with an informative co-reference statement such as “Hesperus is Phosphorus.” Both are true because the names designate Venus. Yet the first can be recognized without astronomical discovery, whereas the second can enlarge a speaker's knowledge. Truth and referential equality therefore do not by themselves explain the difference in cognitive value.

How would you explain it like I'm…

One Star, Two Names

Imagine a superhero who is also your quiet neighbor, but you don't know they're the same person. Saying 'the hero is the hero' teaches you nothing, but finding out 'the hero is the neighbor' is big news. And you might think the hero can fly but not think your neighbor can. Thinkers wonder how two names for one person can work so differently.

The Morning-Evening Star Mystery

"Hesperus" and "Phosphorus" are two old names, and both turned out to name the planet Venus. "Hesperus is Hesperus" is obviously true, but "Hesperus is Phosphorus" took a discovery to learn, even though both are about the same planet. There is a second puzzle too. Someone might believe Hesperus shows up in the evening but not believe Phosphorus does, because they don't know it's the same planet. Frege's Puzzles ask how meaning can work if swapping names for the same thing can change what you learn or what someone believes.

Co-Reference Without Equivalence

Frege's Puzzles are two related challenges for any theory of meaning. First, the identity puzzle: "Hesperus is Hesperus" and "Hesperus is Phosphorus" are both true and about the same object, Venus, yet only the second is informative. Second, the belief puzzle: normally you can swap names for the same thing without changing truth, so if Hesperus is bright, Phosphorus is bright. Inside "believes that...", though, the swap can turn a true report false. A theory that says a name means only the object it picks out seems to predict that these sentences should be equivalent, but they are not. The puzzles are the problem itself; Frege's own answer, called "sense," is just one proposed solution.

 

Frege's Puzzles name a paired explanatory challenge for semantics: co-referring expressions can differ in cognitive value and in their behavior inside attitude reports. The identity form contrasts 'Hesperus is Hesperus', knowable without discovery, with 'Hesperus is Phosphorus', which is informative, though both are true because both names designate Venus. So truth and sameness of reference do not explain the difference. The attitude-report form starts from the substitution principle: co-designating terms can be swapped salva veritate in transparent contexts. Under verbs like 'believes', 'knows', or 'hopes', substitution can fail, since someone may believe Hesperus is visible in the evening without believing Phosphorus is. A semantics assigning each name only its referent seems to predict more equivalence than competent speakers show. The abstraction is the problem architecture, not any solution: Frege's sense/reference distinction is one response, while Russellian, direct-reference, pragmatic, metalinguistic, guise, and contextualist views relocate the explanation. A case counts when it combines referential sameness with a principled divergence in cognitive value or attitude-report truth.

Scope of Application

The home domain is philosophy of language, especially the semantics of names, identity statements, propositional attitudes, and indirect discourse. The puzzles function as adequacy tests for theories of reference and content. Fregean sense theories, Russellian propositions, descriptivism, direct-reference theories, two-dimensional semantics, pragmatic enrichment, hidden-indexical views, and mental-file or guise approaches differ partly in where they place the presentation-sensitive information.

The identity form applies when a true a=b statement has a different route to knowledge or cognitive value from a=a. It is not limited to empirical astronomy.

Clarity

A candidate case passes a six-question test:

1. Are a and b distinct expressions?
2. Is their ordinary co-reference established rather than merely suspected?
3. What baseline semantic or substitution prediction follows from that co-reference?
4. Is the divergence cognitive value in an identity statement, truth in an attitude report, or both?
5. What subject, informational state, or mode of presentation makes the divergence possible?
6. Is the analysis presenting the puzzle neutrally, or has it already assumed one contested solution?

Manages Complexity

The node compresses a large literature into a common stress test. Without it, debates about names can become catalogs of rival positions whose disagreements are difficult to compare. Frege's Puzzles provide a shared input: ordinary co-reference, a predicted equivalence, and a presentation-sensitive divergence. Competing theories can then be compared by what they preserve, what additional structure they posit, and where they locate the explanatory work.

Abstract Reasoning

Several inferences follow from the structure.

Reference is not yet an explanation of cognitive value. From [[a]]=[[b]] one may infer that the two expressions pick out one object. One may not infer that a=a and a=b are equally informative without an additional theory of content and cognition.

Knowledge Transfer

Within philosophy, the complete pattern transfers from proper names to definite descriptions and sometimes indexicals, provided the expressions co-designate and the presentation-sensitive contrast remains. Translation and quotation require extra care because they can change linguistic objects or context rather than hold reference fixed.

The node supplies a reusable review procedure for semantic theories: identify the theory's ordinary semantic value for a and b; derive its substitution prediction; test informative identity and attitude contexts; then locate any additional information invoked by the repair.

Relationships to Other Abstractions

Current abstraction Frege's Puzzles Domain-specific

Parents (1) — outgoing typed relations

• Frege's Puzzles presupposes Substitutability Prime

  Substitutability is the broader abstraction this entry instantiates.

Hierarchy paths (8) — routes to 4 parentless roots

• Frege's Puzzles → presupposes Substitutability → is a kind of Compatibility

Show alternative paths (7)

Neighborhood in Abstraction Space

Frege's Puzzles sits in a sparse region of the domain-specific corpus (83^rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

• Sense and Reference — 0.84
• Lexical Definition — 0.84
• Systemic Hypothesising — 0.82
• Identification (Psychology) — 0.81
• The Saying and the Said — 0.80

Computed from structural-signature embeddings · 2026-10-25