Possible-World Semantics¶

Interpret expressions relative to possible worlds and use those world-indexed values to give truth conditions.

Version: v1 · 2026-10-31 · History
Domain-specific #: 14557
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: Mathematical Logic, Formal Semantics → Mathematics

Core Idea¶

Possible-world semantics interprets expressions relative to alternative worlds or world-like evaluation points. A model assigns expressions an extension at a world and specifies how complex expressions are evaluated from those assignments. A sentence’s world-indexed truth values can be represented as an intension: a function from worlds to truth values, or equivalently its set of true worlds. This is a semantic framework: it supplies a way to determine truth conditions, not just a claim that alternatives exist.^[1]

In standard relational semantics for normal propositional modal logic, a model M = ⟨W,R,V⟩ has a nonempty set of points, an accessibility relation, and a valuation of atomic sentences. At a point w, □φ is true exactly when φ is true at each R-accessible point; ◇φ is true exactly when φ is true at some such point. Other world-based theories use different parameters and rules, so a single binary relation is not required of all possible-world semantics.^[2]^[1]

How would you explain it like I'm…

Checking Every Pretend World

Imagine a big set of pretend worlds, each a different way things could be. To check if a sentence is true, you look at it in each world. Saying something 'must' be true means it's true in every world you're allowed to look at, and 'might' means it's true in at least one. That way of figuring out meaning is possible-world semantics.

Truth Across Possible Worlds

Possible-world semantics is a method in logic and language study for working out when sentences are true. Instead of judging a sentence only for the real world, you judge it at many 'worlds' — different ways things could be. A model says what's true at each world and how to figure out complicated sentences from simpler ones. In the most common version, 'necessarily P' is true at a world if P is true at every world that can be 'reached' from it, and 'possibly P' is true if P is true at at least one reachable world. Other versions use different rules.

World-Relative Truth Conditions

Possible-world semantics is a framework that interprets expressions relative to alternative worlds or evaluation points. A model says what each expression refers to at each world and how to evaluate complex expressions from those pieces. A sentence's pattern of truth values across worlds can be packaged as its intension — a function from worlds to true/false, or equivalently the set of worlds where it's true. In the standard version for modal logic, a model has a set of worlds W, an accessibility relation R saying which worlds can 'see' which, and a valuation V for basic sentences. Then 'necessarily φ' is true at w when φ is true at every world accessible from w, and 'possibly φ' when φ is true at some accessible world. Other theories use different setups, so not every possible-world semantics needs a single accessibility relation.

Possible-world semantics interprets expressions relative to alternative worlds or world-like evaluation points. A model assigns each expression an extension at each world and specifies how complex expressions are evaluated compositionally from those assignments. A sentence's world-indexed truth values can be represented as its intension, a function from worlds to truth values or equivalently the set of worlds at which it is true. It is thus a semantic framework that determines truth conditions, not merely a claim that alternatives exist. In standard relational (Kripke) semantics for normal propositional modal logic, a model M = ⟨W, R, V⟩ consists of a nonempty set of points W, an accessibility relation R on W and a valuation V of atomic sentences. At a point w, □φ is true exactly when φ is true at every R-accessible point, and ◇φ is true exactly when φ is true at some R-accessible point. Other world-based theories use different parameters and evaluation rules, so a single binary accessibility relation is not a requirement of possible-world semantics in general.

Structural Signature¶

World-indexed evaluation: a collection of alternatives and a point relative to which an expression is interpreted.
Interpretation: values of basic expressions at points; complex expressions receive values by the framework’s composition rules.
Truth or extension: a sentence can be true at one world and false at another; predicates may have different extensions at different worlds.
Intension: in a standard intensional treatment, a function maps each world to an expression’s extension there.
Modal selection, when specified: a relation or another selection device determines which alternatives a particular operator considers.
Explicit scope: the semantic result is relative to a model, language and interpretation; philosophical accounts dispute what worlds themselves are.^[1]

What It Is Not¶

A possible world is one point or alternative, not the full method of interpreting expressions across such points. An accessibility relation is one component of relational models, not the valuation or the truth clauses. Normal modal logics are deductively closed systems; their proof rules and the models used to interpret them are related but different objects. Semantics (logic) is broader than any single world-based framework. A model’s availability does not establish that its points are concrete parallel universes.^[1]^[2]

Scope of Application¶

Relational models interpret necessity, possibility and indexed knowledge operators by reference to relevant alternatives. In intensional semantics, world-to-extension functions distinguish expressions that happen to agree in the actual world but vary differently across alternatives. Selection-based treatments of counterfactuals use additional structure; the ordinary □ clause alone is not a complete counterfactual semantics. Context, time, individual domains or other indices can also be needed in more expressive settings.^[2]^[1]

Clarity¶

Keep four claims apart: φ holds at a point of one model; φ holds at every point of that model; φ is valid on a frame under every valuation; and φ is valid over a specified class of models. Their quantifiers range over different things. A dead-end point makes □φ true and ◇φ false under the displayed ordinary relational clauses; seriality excludes such points but is an additional frame condition, not a universal assumption.^[2]

Manages Complexity¶

A common interpretation method separates expressions from their values at alternatives. Varying the valuation while keeping the frame fixed distinguishes facts about a specific model from frame-valid principles; varying a relation can change which modal formulas hold without changing the propositional vocabulary. The economy has a limit: the analyst must specify which alternatives and evaluation rules are intended, not infer them from the term “world.”

Abstract Reasoning¶

Let W = {w,u}, R = {(w,u)} and V(p) = {u}. At w, p is false, while □p and ◇p are true because u is the sole accessible world and p holds there. At u no world is accessible, so □p is vacuously true and ◇p false. Adding (w,w) changes □p at w to false without changing V(p). The model is constructed to expose the quantifiers; it is not a historical experiment.^[2]

Knowledge Transfer¶

In epistemic logic, accessible points can represent alternatives an agent cannot rule out, and knowledge is evaluated against those alternatives under the chosen model. In intensional semantics, an expression can instead be studied through its world-indexed extension. These uses share world-relative interpretation, while the meaning of the relation, the index set and the clauses differ.^[3]^[1]

Examples¶

Information-compatible alternatives¶

Suppose an agent at w cannot distinguish w from u, with R_a(w) = {w,u}. If p is true at w but false at u, the usual epistemic clause makes K_a p false at w: at least one information-compatible point falsifies p. If new evidence removes u from this toy successor set and p remains true at w, K_a p becomes true under the updated model. This is a stipulated model, not a claim that any actual agent has ideal knowledge.^[3]

Mapped back: points → information-compatible alternatives; valuation → where p holds; selection → R_a; rule → truth at every selected point.

Two expressions agreeing locally¶

Take W = {w,u} and sentences p and q both true at w, with p true at u and q false there. Their extensions agree at w, but their intensions differ: p is true at both worlds, q only at w. A world-to-truth-value representation captures a distinction that the single local truth value cannot. This does not claim that every difference in linguistic meaning is captured by possible-world intensions.^[1]

Structural Tensions¶

A clear truth-condition model versus choice of alternatives. Once W, a valuation and an operator clause are fixed, a local semantic result follows precisely; the formalism alone does not decide which alternatives, domain constraints, or relation best model the intended interpretation. Diagnostic: Which parts are stipulated, and which are consequences of the model?
Extensional tractability versus finer meaning distinctions. World-to-extension functions make intensional contrasts explicit, but expressions agreeing at every world share an intension in this representation even if a theory needs finer distinctions. Diagnostic: Would the analysis need to distinguish expressions with the same value at every modeled world?^[1]

Structural–Framed Character¶

The repeated structure is a family of indexed interpretations plus composition rules, with possible-world, modal-operator and semantic-value roles supplied by logic and linguistics. This specialist framework is portable across modal and linguistic cases but does not by itself imply that worlds are concrete objects or that all theories use the same relation.

Structural Core vs. Domain Accent¶

The portable skeleton is index-dependent evaluation. The domain-specific identity adds alternatives interpreted as worlds, semantic assignments and rules for evaluating modal or intensional expressions. A weather table indexed by location is not thereby possible-world semantics.

A model may use Possible World as an evaluation point, Accessibility Relation as a relational selection parameter, and Formal Semantics as the broader context. These are working conceptual relations; graph placement remains for independent review.

Neighborhood in Abstraction Space¶

Possible-World Semantics sits in a sparse region of the domain-specific corpus (97^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2835 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-31

Not to Be Confused With¶

Possible-world semantics differs from one possible world, an accessibility relation, a particular normal modal proof system, a theory of the ontology of worlds, and generic possible-scenario reasoning without explicit semantic evaluation.

References¶

[1] Menzel, Christopher, “Possible Worlds”, Stanford Encyclopedia of Philosophy, §§1.1–1.2; directly inspected scholarly exposition of possible-world interpretation and intensions. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h

[2] Garson, James, “Modal Logic”, Stanford Encyclopedia of Philosophy, §§6–7, directly inspected scholarly exposition of relational frames, valuations, and modal clauses. registry ↩a ↩b ↩c ↩d ↩e

[3] Rendsvig, Rasmus, John Symons, and Yanjing Wang, “Epistemic Logic”, Stanford Encyclopedia of Philosophy, §2.4, directly inspected exposition of information-compatible alternatives. registry ↩a ↩b