Two-Dimensional Semantics
Origin. Kaplan (1977), Stalnaker, Chalmers. Context and circumstance. A priori vs necessary. Primary and secondary intensions. Foundation for epistemology of modality.
Models. Expressions evaluated at two indices. Character maps a context to a content; the content maps a circumstance to a truth value. Matrix semantics.
Formalism.
Two indices: c: context (world of utterance) w: circumstance (world of evaluation) ⟦φ⟧(c, w): value at context c, world w.
Diagonal (primary intension): λc. ⟦φ⟧(c, c) Kaplan's character is λc. λw. ⟦φ⟧(c, w); the diagonal collapses the two indices. A priori necessity.
Content (secondary intension): λw. ⟦φ⟧(c*, w) Fix actual context c*. Metaphysical necessity.
Matrix representation: w₁ w₂ w₃ c₁ = w₁ T F F c₂ = w₂ F T F c₃ = w₃ F F T
A priori: Diagonal all true. True in every context at that context. e.g., "I am here now."
Necessary: Row (for actual context) all true. True at every world of evaluation. e.g., "Water is H₂O."
Contingent a priori: Diagonal true, some row false. "The meter stick is one meter."
Necessary a posteriori: Diagonal not all true, row true. "Water is H₂O" (if water existed). Discovered, not knowable a priori.
Indexicals: "I": speaker at context. "Now": time at context. Content varies with context.
Symbols.
| Symbol | Unicode | Name | Meaning |
|---|---|---|---|
| c | — | Context | World of utterance, the first index |
| w | — | Circumstance | World of evaluation, the second index |
| c* | — | Actual context | The context actually occupied |
| λ | U+03BB | Greek small lambda | Abstraction over an index |
| λc | — | Diagonal | Primary intension: collapse the two indices |
| λw | — | Content | Secondary intension: fix the context |
| ⟦ ⟧ | U+27E6, U+27E7 | Mathematical white square brackets | Semantic value of an expression |
| @ | — | Actually | Evaluates at the context's own world |
| □ | U+25A1 | White square / Box | Necessity: along a row |
| ◇ | U+25C7 | White diamond | Possibility |
| F@ | — | Fixedly actually | Davies–Humberstone operator: @ holding across all contexts |
| ↔ | U+2194 | Left right arrow | Material biconditional, as in φ ↔ @φ |
| ₁, ₂, ₃ | U+2081, U+2082, U+2083 | Subscript digits | Indices of contexts and circumstances in the matrix |
| T, F | — | Truth values | Entries of the two-dimensional matrix |
Metatheory. The formal content of the framework is that validity and necessity come apart on the diagonal. "I am here now" and the schema φ ↔ @φ are true at every context taken at its own world, hence valid, yet false at some context-circumstance pair, hence not necessary; so diagonal consequence is not closed under the rule of necessitation, and the two-dimensional setting is where that failure is exhibited rather than assumed. Crossley and Humberstone (1977) drew the corresponding distinction between real-world and general validity for the logic of the actuality operator, and Davies and Humberstone (1980) added the "fixedly actually" operator F@, which is not definable from □ and @ alone and which they used to make the contingent a priori a matter of logical form. Segerberg (1973) gave a complete axiomatization of two-dimensional modal logic. The actuality operator is likewise not definable in the basic modal language, since no formula built from □ alone is sensitive to which world is the point of utterance.
Applies to. Philosophy of language. Epistemology. Consciousness (Chalmers). Indexicality.
Limitations. Primary intensions are not compositional: the diagonal of a compound is fixed by the full matrix of its parts, not by their diagonals, so two expressions with the same primary intension can be substituted into a context and change its primary intension. The failure of necessitation means the logic of diagonal validity is not normal, so the standard modal metatheory — canonicity, Sahlqvist correspondence, the usual completeness proofs — does not transfer to it without modification. Kaplan's prohibition on monsters, operators that shift the context index, is a stipulation about English rather than a constraint the semantics imposes, and languages with context-shifting operators falsify it. And the epistemic reading requires that every epistemic possibility correspond to a scenario describable in semantically neutral vocabulary; that assumption carries the philosophical weight of the framework and is not delivered by the formalism.
© 2026 Lingenic LLC