Intuitionistic Logic

Semantics

Reading preferences

Optional display controls need JavaScript. All reading content and navigation work without it.

Source file content/intuitionistic-logic/semantics/semantics.tex

Editorial

This chapter collects definitions for semantics for intuitionistic logic. So far only Kripke and topological semantics are covered. There are no examples yet, either of how models make formulas true or of proofs that formulas are valid.

Source file content/intuitionistic-logic/semantics/introduction.tex

Introduction

No logic is satisfactorily described without a semantics, and intuitionistic logic is no exception. Whereas for classical logic, the semantics based on valuations is canonical, there are several competing semantics for intuitionistic logic. None of them are completely satisfactory in the sense that they give an intuitionistically acceptable account of the meanings of the connectives.

The semantics based on relational models, similar to the semantics for modal logics, is perhaps the most popular one. In this semantics, propositional variables are assigned to worlds, and these worlds are related by an accessibility relation. That relation is always a partial order, i.e., it is reflexive, antisymmetric, and transitive.

Intuitively, you might think of these worlds as states of knowledge or “evidentiary situations.” A state w′w'source is accessible from wwsource iff, for all we know, w′w'source is a possible (future) state of knowledge, i.e., one that is compatible with what's known at wwsource. Once a proposition is known, it can't become un-known, i.e., whenever A!Asource is known at wwsource and Rww′Rww'source, A!Asource is known at w′w'source as well. So “knowledge” is monotonic with respect to the accessibility relation.

If we define “A!Asource is known” as in epistemic logic as “true in all epistemic alternatives,” then A∧B!A \land !Bsource is known at wwsource if in all epistemic alternatives, both A!Asource and B!Bsource are known. But since knowledge is monotonic and RRsource is reflexive, that means that A∧B!A \land !Bsource is known at wwsource iff A!Asource and B!Bsource are known at wwsource. For the same reason, A∨B!A \lor !Bsource is known at wwsource iff at least one of them is known. So for ∧\landsource and ∨\lorsource, the truth conditions of the connectives coincide with those in classical logic.

The truth conditions for the conditional, however, differ from classical logic. A→B!A \lif !Bsource is known at wwsource iff at no w′w'source with Rww′Rww'source, A!Asource is known without B!Bsource also being known. This is not the same as the condition that A!Asource is unknown or B!Bsource is known at wwsource. For if we know neither A!Asource nor B!Bsource at wwsource, there might be a future epistemic state w′w'source with Rww′Rww'source such that at w′w'source, A!Asource is known without also coming to know B!Bsource.

We know ¬A\lnot !Asource only if there is no possible future epistemic state in which we know A!Asource. Here the idea is that if A!Asource were knowable, then in some possible future epistemic state A!Asource becomes known. Since we can't know ⊥\lfalsesource, in that future epistemic state, we would know A!Asource but not know ⊥\lfalsesource.

On this interpretation the principle of excluded middle fails. For there are some A!Asource which we don't yet know, but which we might come to know. For such a formula A!Asource, both A!Asource and ¬A\lnot !Asource are unknown, so A∨¬A!A \lor \lnot !Asource is not known. But we do know, e.g., that ¬(A∧¬A)\lnot(!A \land \lnot !A)source. For no future state in which we know both A!Asource and ¬A\lnot !Asource is possible, and we know this independently of whether or not we know A!Asource or ¬A\lnot !Asource.

Relational models are not the only available semantics for intuitionistic logic. The topological semantics is another: here propositions are interpreted as open sets in a topological space, and the connectives are interpreted as operations on these sets (e.g., ∧\landsource corresponds to intersection).

Source file content/intuitionistic-logic/semantics/relational-models.tex

relational model

In order to give a precise semantics for intuitionistic propositional logic, we have to give a definition of what counts as a model relative to which we can evaluate formulas. On the basis of such a definition it is then also possible to define semantics notions such as validity and entailment. One such semantics is given by relational models.

Definition of an intuitionistic relational model

A relational model for intuitionistic propositional logic is a triple M=⟨W,R,V⟩\mModel{M} = \tuple{W, R, V}source, where

  1. WWsource is a non-empty set,

  2. RRsource is a partial order (i.e., a reflexive, antisymmetric, and transitive binary relation) on WWsource, and

  3. VVsource is a function assigning to each propositional variable ppsource a subset of WWsource, such that

  4. VVsource is monotone with respect to RRsource, i.e., if w∈V(p)w \in V(p)source and Rww′Rww'source, then w′∈V(p)w' \in V(p)source.

Inductive definition of truth at a world

We define the notion of A!Asource being true at wwsource in M\mModel{M}source, M⊩A[w]\mSat{M}{!A}[w]source, inductively as follows:

  1. Case: A≡p!A \ident psource

    M⊩A[w]\mSat{M}{\indfrm}[w]source iff w∈V(p)w \in V(p)source.

  2. Case: A≡⊥!A \ident \lfalsesource

    not M⊩A[w]\mSat{M}{\indfrm}[w]source.

  3. Case: A≡¬B!A \ident \lnot !Bsource

    M⊩A[w]\mSat{M}{\indfrm}[w]source iff for no w′w'source such that Rww′Rww'source, M⊩B[w′]\mSat{M}{!B}[w']source.

  4. Case: A≡B∧C!A \ident !B \land !Csource

    M⊩A[w]\mSat{M}{\indfrm}[w]source iff M⊩B[w]\mSat{M}{!B}[w]source and M⊩C[w]\mSat{M}{!C}[w]source.

  5. Case: A≡B∨C!A \ident !B \lor !Csource

    M⊩A[w]\mSat{M}{\indfrm}[w]source iff M⊩B[w]\mSat{M}{!B}[w]source or M⊩C[w]\mSat{M}{!C}[w]source (or both).

  6. Case: A≡B→C!A \ident !B \lif !Csource

    M⊩A[w]\mSat{M}{\indfrm}[w]source iff for every w′w'source such that Rww′Rww'source, not M⊩B[w′]\mSat{M}{!B}[w']source or M⊩C[w′]\mSat{M}{!C}[w']source (or both).

We write M⊮A[w]\mSat/{M}{!A}[w]source if not M⊩A[w]\mSat{M}{!A}[w]source. If Γ\Gammasource is a set of formulas, M⊩Γ[w]\mSat{M}{\Gamma}[w]source means M⊩B[w]\mSat{M}{!B}[w]source for all B∈Γ!B \in \Gammasource.

Exercise relating negation to a conditional

Show that according to the definition of truth at a world, M⊩¬A[w]\mSat{M}{\lnot !A}[w]source iff M⊩A→⊥[w]\mSat{M}{!A \lif \lfalse}[w]source.

Proposition that truth is monotone

Truth at worlds is monotonic with respect to RRsource, i.e., if M⊩A[w]\mSat{M}{!A}[w]source and Rww′Rww'source, then M⊩A[w′]\mSat{M}{!A}[w']source.

Proof

Exercise.

Exercise proving truth monotonicity

Prove the proposition that truth at worlds is monotone.

Source file content/intuitionistic-logic/semantics/semantic-notions.tex

Semantic Notions

Definition of global truth, validity, and entailment

We say A!Asource is true in the model M=⟨W,R,V⟩\mModel{M} = \tuple{W,R,V}source, M⊩A\mSat{M}{!A}source, iff M⊩A[w]\mSat{M}{!A}[w]source for all w∈Ww \in Wsource. A!Asource is valid, ⊨A\Entails !Asource, iff it is true in all models. We say a set of formulas Γ\Gammasource entails A!Asource, Γ⊨A\Gamma \Entails !Asource, iff for every model M\mModel{M}source and every wwsource such that M⊩Γ[w]\mSat{M}{\Gamma}[w]source, M⊩A[w]\mSat{M}{!A}[w]source.

Two consequences of entailment

  1. If M⊩Γ[w]\mSat{M}{\Gamma}[w]source and Γ⊨A\Gamma \Entails !Asource, then M⊩A[w]\mSat{M}{!A}[w]source.

  2. If M⊩Γ\mSat{M}{\Gamma}source and Γ⊨A\Gamma \Entails !Asource, then M⊩A\mSat{M}{!A}source.

Proof

  1. Suppose M⊩Γ\mSat{M}{\Gamma}source. Since Γ⊨A\Gamma \Entails !Asource, we know that if M⊩Γ[w]\mSat{M}{\Gamma}[w]source, then M⊩A[w]\mSat{M}{!A}[w]source. Since M⊩Γ[u]\mSat{M}{\Gamma}[u]source for all every u∈Wu \in Wsource, M⊩Γ[w]\mSat{M}{\Gamma}[w]source. Hence M⊩A[w]\mSat{M}{!A}[w]source.

  2. Follows immediately from the first part of the proposition relating satisfaction and entailment.

Definition of a model restricted to a world

Suppose M\mModel{M}source is a relational model and w∈Ww \in Wsource. The restriction Mw=⟨Ww,Rw,Vw⟩\mModel{M}_w=\tuple{W_w, R_w, V_w}source of M\mModel{M}source to wwsource is given by:

Ww={u∈W:Rwu},Rw=R∩(Ww)2, andVw(p)=V(p)∩Ww.W_w & = \Setabs{u \in W}{Rwu},\\ R_w & = R \cap (W_w)^2, \text{ and}\\ V_w(p) & = V(p) \cap W_w.source

Proposition characterizing restriction

M⊩A[w]\mSat{M}{!A}[w]source iff Mw⊩A\mSat{M_w}{!A}source.

Exercise proving the restriction proposition

Prove the proposition characterizing restriction to a world.

Proposition deriving local entailment from global model truth

Suppose for every model M\mModel{M}source such that M⊩Γ\mSat{M}{\Gamma}source, M⊩A\mSat{M}{!A}source. Then Γ⊨A\Gamma \Entails !Asource.

Proof

Suppose that M⊩Γ[w]\mSat{M}{\Gamma}[w]source. By the the proposition characterizing restriction to a world applied to every B∈Γ!B \in \Gammasource, we have Mw⊩Γ\mSat{M_w}{\Gamma}source. By the assumption, we have Mw⊩A\mSat{M_w}{!A}source. By the proposition characterizing restriction to a world again, we get M⊩A[w]\mSat{M}{!A}[w]source.

Source file content/intuitionistic-logic/semantics/topological-semantics.tex

Topological Semantics

Another way to provide a semantics for intuitionistic logic is using the mathematical concept of a topology.

Definition of a topology

Let XXsource be a set. A topology on XXsource is a set O⊆℘(X)\Top{O} \subseteq \Pow{X}source that satisfies the properties below. The elements of O\Top{O}source are called the open sets of the topology. The set XXsource together with O\Top{O}source is called a topological space.

  1. The empty set and the entire space are open: ∅\emptysetsource, X∈OX \in \Top{O}source.

  2. Open sets are closed under finite intersections: if UUsource, V∈OV \in \Top{O}source then U∩V∈OU \cap V \in \Top{O}source

  3. Open sets are closed under arbitrary unions: if Ui∈OU_i \in \Top{O}source for all i∈Ii \in Isource, then ⋃{Ui:i∈I}∈O\bigcup \Setabs{U_i}{i \in I} \in \Top{O}source.

We may write XXsource for a topology if the collection of open sets can be inferred from the context; note that, still, only after XXsource is endowed with open sets can it be called a topology.

Definition of a topological model and its propositions

A topological model of intuitionistic propositional logic is a triple X=⟨X,O,V⟩\mModel{X} = \tuple{X, \Top{O}, V}source where O\Top{O}source is a topology on XXsource and VVsource is a function assigning an open set in O\Top{O}source to each propositional variable.

Given a topological model X\mModel{X}source, we can define ⟦A⟧X\Prop{X}{!A}source inductively as follows:

  1. ⟦⊥⟧X=∅\Prop{X}{\lfalse} = \emptysetsource

  2. ⟦p⟧X=V(p)\Prop{X}{p} = V(p)source

  3. ⟦A∧B⟧X=⟦A⟧X∩⟦B⟧X\Prop{X}{!A \land !B} = \Prop{X}{!A} \cap \Prop{X}{!B}source

  4. ⟦A∨B⟧X=⟦A⟧X∪⟦B⟧X\Prop{X}{!A \lor !B} = \Prop{X}{!A} \cup \Prop{X}{!B}source

  5. ⟦A→B⟧X=Int((X∖⟦A⟧X)∪⟦B⟧X)\Prop{X}{!A \lif !B} = \Interior{(X \setminus \Prop{X}{!A}) \cup \Prop{X}{!B}}source

Here, Int(V)\Interior{V}source is the function that maps a set V⊆XV \subseteq Xsource to its interior, that is, the union of all open sets it contains. In other words,

Int(V)=⋃{U:U⊆V and U∈O}.\Interior{V} = \bigcup \Setabs{U}{U \subseteq V \text{ and } U \in \Top{O}}.source

Note that the interior of any set is always open, since it is a union of open sets. Thus, ⟦A⟧X\Prop{X}{!A}source is always an open set.

Although topological semantics is highly abstract, there are ways to think about it that might motivate it. Suppose that the elements, or “points,” of XXsource are points at which statements can be evaluated. The set of all points where A!Asource is true is the proposition expressed by A!Asource. Not every set of points is a potential proposition; only the elements of O\Top{O}source are. A⊨B!A \Entails !Bsource iff B!Bsource is true at every point at which A!Asource is true, i.e., ⟦A⟧X⊆⟦B⟧X\Prop{X}{!A} \subseteq \Prop{X}{!B}source, for all XXsource. The absurd statement ⊥\lfalsesource is never true, so ⟦⊥⟧X=∅\Prop{X}{\lfalse} = \emptysetsource.

How must the propositions expressed by B∧C!B \land !Csource, B∨C!B \lor !Csource, and B→C!B \lif !Csource be related to those expressed by B!Bsource and C!Csource for the intuitionistically valid laws to hold, i.e., so that A⊢B!A \Proves !Bsource iff ⟦A⟧X⊂⟦B⟧X\Prop{X}{!A} \subset \Prop{X}{!B}source? We require ⊥⊢A\lfalse \Proves !Asource for any A!Asource, which is satisfied because ∅⊆U\emptyset \subseteq Usource for all UUsource. Since B∧C⊢B!B \land !C \Proves !Bsource, we require that ⟦B∧C⟧X⊆⟦B⟧X\Prop{X}{!B \land !C} \subseteq \Prop{X}{!B}source, and similarly ⟦B∧C⟧X⊆⟦C⟧X\Prop{X}{!B \land !C} \subseteq \Prop{X}{!C}source. The largest set satisfying W⊆UW \subseteq Usource and W⊆VW \subseteq Vsource is U∩VU \cap Vsource. Conversely, B⊢B∨C!B \Proves !B \lor !Csource and C⊢B∨C!C \Proves !B \lor !Csource, and so we require that ⟦B⟧X⊆⟦B∨C⟧X\Prop{X}{!B} \subseteq \Prop{X}{!B \lor !C}source and ⟦C⟧X⊆⟦B∨C⟧X\Prop{X}{!C} \subseteq \Prop{X}{!B \lor !C}source. The smallest set WWsource such that U⊆WU \subseteq Wsource and V⊆WV \subseteq Wsource is U∪VU \cup Vsource.

The definition for →\lifsource is tricky: A→B!A \lif !Bsource expresses the weakest proposition that, combined with A!Asource, entails B!Bsource. That A→B!A \lif !Bsource combined with A!Asource entails B!Bsource is clear from (A→B)∧A⊢B(!A \lif !B) \land !A \Proves !Bsource. So ⟦A→B⟧X\Prop{X}{!A \lif !B}source should be the greatest open set such that ⟦A→B⟧X∩⟦A⟧X⊂⟦B⟧X\Prop{X}{!A \lif !B} \cap \Prop{X}{!A} \subset \Prop{X}{!B}source, leading to our definition.

Source disclosures