Reading preferences
Optional display controls need JavaScript. All reading content and navigation work without it.
Source file content/intuitionistic-logic/soundness-completeness/soundness-completeness.tex
Editorial
This chapter collects soundness and completeness results for propositional intuitionistic logic. It needs an introduction. The completeness proof makes use of facts about provability that should be stated and proved explicitly somewhere.
Source file content/intuitionistic-logic/soundness-completeness/soundness-axd.tex
Soundness of Axiomatic derivation
Editorial
The soundness proof relies on the fact that all axioms are intuitionistically valid; this still needs to be proved, e.g., in the Semantics chapter.
Soundness theorem for the intuitionistic axiomatic calculus
Proof
We prove that if source, then source. The proof is by induction on the number source of formulas in the derivation of source from source. We show that if source, dots, source is a derivation from source, then source. Note that if source, dots, source is a derivation, so is source, dots, source for any source.
There are no derivations of length source, so for source the claim holds vacuously. So the claim holds for all derivations of length source. We distinguish cases according to the justification of source.
source is an axiom. All axioms are valid, so source for any source.
source. Then for any source and source, if source, obviously source, i.e., source.
source follows by MP from source and source. source, dots, source and source, dots, source are derivations from source, so by inductive hypothesis, source and source.
Suppose source. Since source and source, source. By definition, this means that for all source such that source, if source then source. Since source is reflexive, source is among the source such that source, i.e., we have that if source then source. Since source, source. So, source, as we wanted to show.
Source file content/intuitionistic-logic/soundness-completeness/soundness-nd.tex
Soundness of Natural Deduction
We will now prove soundness of natural deduction with regards to the relational semantics, that is, showing that if a formula is derivable from a set of assumptions then the set of assumptions entails the formula.
Soundness theorem for intuitionistic natural deduction
Proof
We prove that if source, then source. The proof is by induction on the derivation of source from source.
If the derivation consists of just the assumption source, we have source, and want to show that source. Suppose that source. Then trivially source.
The derivation ends in source: Exercise.
The derivation ends in source: Exercise.
The derivation ends in source: Suppose the premise is source, and the undischarged assumptions of the derivation ending in source are source. Then we have source and by inductive hypothesis, source. We have to show that source. Suppose source. Since source, source. But then also source. Similarly, if the premise is source, we have that source.
The derivation ends in source: The derivations ending in the premises are of source from undischarged assumptions source, of source from undischarged assumptions source, and of source from undischarged assumptions source. So we have source, source, and source. By induction hypothesis, source, source, and source. We have to prove that source.
Suppose source. Then source and since source, source. By definition of source, either source or source. So we distinguish cases: (a) source. Thensource. Since source, we have source. (b) source. Then source. Since source, we have source. So in either case, source, as we wanted to show.
The derivation ends with source concluding source. Then the premise is source, and the derivation ending in the premise has undischarged assumptions source. So we have that source, and by induction hypothesis that source. We have to show that source.
Suppose source. We want to show that for all source such that source, if source, then source. So assume that source and source. By the earlier proposition that intuitionistic truth persists along accessibility, source. Since source, source, which is what we wanted to show.
The derivation ends in source and conclusion source. The premises are source and source, with derivations from undischarged assumptions source, source. So we have source and source. By inductive hypothesis, source and source. We have to show that source.
Suppose source. Since source and source, source. By definition, this means that for all source such that source, if source then source. Since source is reflexive, source is among the source such that source, i.e., we have that if source then source. Since source and source, source. So, source, as we wanted to show.
The derivation ends in source, concluding source. The premise is source and the undischarged assumptions of the derivation of the premise are source. Then source. By inductive hypothesis, source. We have to show source.
We proceed indirectly. If source there is a model source and world source such that source and source. Since source, source. But that's impossible, since by definition, source. So source.
The derivation ends in source: Exercise.
The derivation ends in source: Exercise.
Exercise completing the natural-deduction soundness proof
Complete the proof of the natural-deduction soundness theorem. For the cases for source and source, use the definition of source in the definition of truth at a world, i.e., don't treat source as defined by source.
Exercise giving three intuitionistic nonderivability results
Show that the following formulas are not derivable in intuitionistic logic:
Source file content/intuitionistic-logic/soundness-completeness/lindenbaum.tex
Lindenbaum's Lemma
The completeness theorem for intuitionistic logic is proved by assuming source and constructing a model source and source.
In classical logic the relation of derivability can be reduced to the notion of consistency since a formula source is derivable from a set of formulas iff the set together with the negation of source is inconsistent. This is not possible in intuitionistic logic. In intuitionistic logic, if source is inconsistent, we only get that source. Since source does not hold intuitionistically in general, we cannot conclude that source.
Thus, when constructing the model source, we will need to keep track of the non-derivability of the formula source and thus we will not be able to use a complete set source to build the model source, as in every complete set source, we have source.
Instead of using a complete set source, we will us the notion of a prime set of formulas:
Definition of a prime set of formulas
A set of formulas source is prime iff
Lindenbaum's Lemma for intuitionistic logic
[Lindenbaum's Lemma] If source, there is a source such that source is prime and source.
Proof
Let source, source, dots, be an enumeration of all formulas of the form source. We'll define an increasing sequence of sets of formulas source, where each source is defined as source together with one new formula. source will be the union of all source. The new formulas are selected so as to ensure that source is prime and still source. This means that at each step we should find the first disjunction source such that:
We add to source either source if source, or source otherwise. We'll have to show that this works. For now, let's define source as the least source such that the first condition selecting a disjunction and the second condition selecting a disjunction hold.
Define source and
If source is undefined, i.e., whenever source, either source or source, we let source. Now let source
First we show that for all source, source. We proceed by induction on source. For source the claim holds by the hypothesis of the theorem, i.e., source. If source, we have to show that if source then source. If source is undefined, source and there is nothing to prove. So suppose source is defined. For simplicity, let source.
We'll prove the contrapositive of the claim. Suppose source. By construction, source if source, or else source. It clearly can't be the first, since then source. Hence, source and source. By definition of source, we have that source. We have source. We also have source. Hence, source, which is what we wanted to show.
If source, there would be some finite subset source such that source. Each source must be in source for some source. Let source be the largest of these. Since source if source, source. But then source, contrary to our proof above that source.
Lastly, we show that source is prime, i.e., satisfies conditions the consistency condition in the definition of a prime set, the deductive-closure condition in the definition of a prime set, and the disjunction condition in the definition of a prime set of the definition of a prime set.
First, source, so source is consistent, so the consistency condition in the definition of a prime set holds.
We now show that if source, then either source or source. This proves the disjunction condition in the definition of a prime set, since if source then also source. So assume source but source and source. Since source, source for some source. source appears on the enumeration of all disjunctions, say, as source. source satisfies the properties in the definition of source, namely we have source, while source and source. At each stage, at least one fewer disjunction source satisfies the conditions (since at each stage we add either source or source), so at some stage source we will have source. But then either source or source, contrary to the assumption that source and source.
Now suppose source. Then source. But we've just proved that if source then source. Hence, source satisfies the deductive-closure condition in the definition of a prime set of the definition of a prime set.
Exercise relating nonderivability of falsity to classical consistency
Show that if source then source is consistent in classical logic, i.e., there is a valuation making all formulas in source true.
Source file content/intuitionistic-logic/soundness-completeness/canonical-model.tex
The Canonical Model
The worlds in our model will be finite sequences source of natural numbers, i.e., source. Note that source is inductively defined by:
If source and source, then source (where source is source and source is the concatenation if source and source).
Nothing else is in source.
So we can use source to give inductive definitions.
Let source, source, dots, be an enumeration of all pairs of formulas. Given a set of formulas source, define source by induction as follows:
Here by source we mean the prime set of formulas which exists by Lindenbaum's Lemma applied to the set source and the formula source. Note that by this definition, if source, then source and source. Note also that source for any source. If source is prime, then source is prime for all source.
Definition of the canonical intuitionistic model
Suppose source is prime. Then the canonical model source for source is defined by:
It is easy to verify that source is indeed a partial order. Also, the monotonicity condition on source is satisfied. Since source we get source whenever source by induction on source.
Source file content/intuitionistic-logic/soundness-completeness/truth-lemma.tex
The Truth Lemma
The following lemma connects satisfaction in the canonical model with which formulas are elements of the prime set source.
Truth Lemma for the canonical model
Proof
By induction on source.
Case: source
Since source is prime, it is consistent, so source. By definition, source.
Case: source
Case: source
Case: source
source iff source and source. By induction hypothesis, source iff source, and similarly for source. But source and source iff source.
Case: source
source iff source or source. By induction hypothesis, this holds iff source or source. We have to show that this in turn holds iff source. The left-to-right direction is clear. The right-to-left direction follows since source is prime.
Case: source
First the contrapositive of the left-to-right direction: Assume source. Then also source. Since source is source for some source, we have source, and source but source. By inductive hypothesis, source and source. Since source, this means that source.
Now assume source, and let source. Since source, we have: if source, then source. In other words, for every source such that source, either source or source. By induction hypothesis, this means that whenever source, either source or source, i.e., source.
Source file content/intuitionistic-logic/soundness-completeness/completeness-thm.tex
The Completeness Theorem
Completeness theorem for intuitionistic logic
Proof
We prove the contrapositive: Suppose source. Then by Lindenbaum's Lemma, there is a prime set source such that source. Consider the canonical model source for source as defined in the definition of the canonical model. For any source, source. Note that source. By the Truth Lemma (the Truth Lemma), we have source for all source and source. This shows that source.
Exercise on formulas using only variables, disjunction, and conjunction
Show that if source only contains propositional variables, source, and source, then source. Use this to conclude that source is not definable in intuitionistic logic from source and source.
Exercise proving the disjunction property
By using the completeness theorem prove that if source then source or source. (Hint: Assume source and source and construct a new model source such that source.)
Exercise on linearly ordered relational models
Show that if source is a relational model using a linear order then source.
Source file content/intuitionistic-logic/soundness-completeness/decidability.tex
Decidability
Observe that the proof of the completeness theorem gives us for every source a model with an infinite number of worlds witnessing the fact that source. The following proposition shows that to prove source it is enough to prove that source for all finite models (i.e., models with a finite set of worlds).
Finite-countermodel theorem
Proof
Assume source is such that source and source is the set of propositional variables occurring in source. Define source by letting source where source, source be the subset relation, and source. It should be clear that source is a finite set and that source is a relational model.
It can be shown, by induction on source, that
for all formulas source with only propositional variables from source. This is left as an exercise for the reader.
Exercise finishing the finite-countermodel proof
Finish the proof of the finite-countermodel theorem by showing that source iff source for all formulas source with only propositional variables from source.
From the finite-countermodel theorem it follows that there is an algorithm to decide whether source.
Source disclosures
- TR061-SAR-006: Source dependency note. The axiomatic soundness proof uses validity of every intuitionistic axiom, while the source editorial says that supporting result is still unproved in the text. No missing proof is invented. source
- TR061-SAR-001: Reader repair. The malformed three-argument satisfaction occurrence is rendered as model M satisfying A subscript n at world w. The immutable source text and its exact anchor remain available. source
- TR061-SAR-003: Reader repair. The world argument w is moved inside the local satisfaction formula so the occurrence reads that model M satisfies B at world w. The immutable misplaced delimiter remains source-visible. source
- TR061-SAR-004: Reader repair. Delta subscript one is unioned with the singleton containing B before entailing D, matching the source's premise set on the preceding line. The unbraced source remains available. source
- TR061-SAR-005: Reader repair. Delta subscript two is unioned with the singleton containing C before entailing D, matching the source's premise set on the preceding line. The unbraced source remains available. source
- TR061-SAR-002: Source incompleteness note. The negation induction case in the Truth Lemma has an empty proof body. The edition preserves the empty case and does not invent a proof. source
- TR061-SAR-007: Source mathematical caveat. Identifying worlds solely by which atoms in P are true and then ordering all resulting profiles by inclusion can add accessibility that was absent from the original model. Consequently the asserted preservation of every intuitionistic formula is false in general. The printed construction and unsolved exercise are retained; no replacement proof is invented. source