Premise sets #
Kratzer's premise semantics: a premise set is a list of propositions over an index type, a
proposition follows from it when it holds throughout the set's intersection, and the set is
consistent when that intersection is inhabited. Must and can in view of a conversational
background f are consequence from, and compatibility with, f i (Definitions 5 and 6);
when f i may be inconsistent they are restated over its consistent sublists (Definitions 7
and 8), and the two pairs agree on consistent backgrounds. Nothing here commits to what an
index is: worlds, situations, or times.
Main definitions #
propIntersection A: the indices satisfying every member ofA, Kratzer's⋂A.followsFrom p A,isConsistent A,isCompatibleWith p A.mustInView,canInView: Definitions 5 and 6.mustInView',canInView': Definitions 7 and 8, overconsistentSublists.
Main results #
mustInView_iff_mustInView'_of_consistent,canInView_iff_canInView'_of_consistent: the revised operators agree with the original ones on a consistent premise set.canInView_iff_not_mustInView_not,canInView'_iff_not_mustInView'_not: each pair of operators is dual, can being the negation of must not.canInView'_of_mem,mustInView'_of_forall_isCompatibleWith: Definition 8 holds of a premise itself, and Definition 7 of a premise compatible with every consistent sublist.
References #
- [A. Kratzer, What 'must' and 'can' must and can mean (1977)][kratzer-1977]
- [A. Kratzer, Modals and Conditionals (2012)][kratzer-2012]
Primitives on premise sets #
The intersection of a list of propositions: indices satisfying all of them.
Equations
- Modality.Kratzer.propIntersection props = {i : W | ∀ p ∈ props, p i}
Instances For
A proposition p follows from a premise set A iff ⋂ A ⊆ {i | p i}
([kratzer-1977] p. 31).
Equations
- Modality.Kratzer.followsFrom p A = (Modality.Kratzer.propIntersection A ⊆ {i : W | p i})
Instances For
A premise set is consistent iff ⋂ A is non-empty ([kratzer-1977] p. 31).
Equations
- Modality.Kratzer.isConsistent A = (Modality.Kratzer.propIntersection A).Nonempty
Instances For
A proposition p is compatible with A iff A ∪ {p} is consistent.
Equations
- Modality.Kratzer.isCompatibleWith p A = Modality.Kratzer.isConsistent (p :: A)
Instances For
The intersection of a premise set is contained in each of its members.
Duality: p is compatible with A iff ¬p does not follow from A.
Definitions 5 and 6: must and can in view of #
Def 5 ([kratzer-1977]): must p in view of f at index i
iff p follows from the premise set f i.
ν(p, f) = {i : ⋂(f i) ⊆ p}
Equations
- Modality.Kratzer.mustInView f p i = Modality.Kratzer.followsFrom p (f i)
Instances For
Def 6 ([kratzer-1977]): can p in view of f at index i
iff p is compatible with the premise set f i.
μ(p, f) = {i : ⋂((f i) ∪ {p}) ≠ ∅}
Equations
- Modality.Kratzer.canInView f p i = Modality.Kratzer.isCompatibleWith p (f i)
Instances For
Definitions 7 and 8: must and can over consistent sublists #
The set of consistent sublists of a premise set: X_A = {B ⊆ A : consistent B}.
Kratzer's revised definitions quantify over these to handle inconsistent A.
Concretely: a sublist B of A such that B is consistent.
Equations
- Modality.Kratzer.consistentSublists A = {B : List (W → Prop) | B ∈ A.sublists ∧ Modality.Kratzer.isConsistent B}
Instances For
Def 7 ([kratzer-1977]): the revised necessity operator that handles possibly inconsistent premise sets.
must p in view of f at i iff for every consistent subset B of f i,
there exists a consistent subset C ⊇ B such that p follows from C.
Original notation:
ν(p, f) = {i : ∀B[B ∈ X_{f(i)} → ∃C[C ∈ X_{f(i)} ∧ B ⊆ C ∧ ⋂C ⊆ p]]}
Equations
- Modality.Kratzer.mustInView' f p i = ∀ B ∈ Modality.Kratzer.consistentSublists (f i), ∃ C ∈ Modality.Kratzer.consistentSublists (f i), B ⊆ C ∧ Modality.Kratzer.followsFrom p C
Instances For
Def 8 ([kratzer-1977]): the revised possibility operator that handles possibly inconsistent premise sets.
can p in view of f at i iff there exists a consistent subset B of f i
such that for every consistent subset C ⊇ B, the set C ∪ {p} is consistent.
Original notation:
μ(p, f) = {i : ∃B[B ∈ X_{f(i)} ∧ ∀C[(C ∈ X_{f(i)} ∧ B ⊆ C) → consistent(C ∪ {p})]]}
Equations
- Modality.Kratzer.canInView' f p i = ∃ B ∈ Modality.Kratzer.consistentSublists (f i), ∀ C ∈ Modality.Kratzer.consistentSublists (f i), B ⊆ C → Modality.Kratzer.isCompatibleWith p C
Instances For
Monotonicity #
The reduction theorems below need three monotonicity facts about the premise algebra. They are proved here once and reused.
propIntersection is anti-monotone in the premise list: more premises
can only shrink the set of indices satisfying all of them.
followsFrom is monotone in the premise list: more premises only add
consequences.
isCompatibleWith is anti-monotone in the premise list: removing
premises can only make a proposition easier to be compatible with.
A consistent premise list is itself a member of its own consistent sublist powerset.
Every element of consistentSublists A is a ⊆-subset of A.
Reduction to Definitions 5 and 6 #
When the premise set f i is itself consistent, Kratzer's revised definitions
collapse to the original Defs 5–6: there is no "inconsistency to repair." The
witness for both directions is f i itself — it is a sublist of itself, it
is consistent by hypothesis, and B ⊆ f i for every B ∈ consistentSublists (f i).
When f i is consistent, the revised necessity operator coincides with
the original.
When f i is consistent, the revised possibility operator coincides with
the original.
Duality #
Can is the negation of must not under both pairs of definitions, since compatibility with a premise set is the failure of the negation to follow from it.
Sufficient conditions for Definitions 7 and 8 #
A premise in a consistent sublist is possible under Definition 8: every consistent extension still contains it.
The head of the premise set is necessary under Definition 7 when it is compatible with every consistent sublist: a sublist without it extends by it, one with it entails it.