Trivalent Exhaustification #
@cite{spector-sudo-2017}
Benjamin Spector and Yasutada Sudo, "Presupposed Ignorance and Exhaustification: How Scalar Implicatures and Presuppositions Interact." Linguistics and Philosophy 40, pp. 473–517.
Core Operators #
Two trivalent exhaustification operators extend bivalent EXH (@cite{fox-2007}) to handle presupposition-bearing sentences:
- EXH¹ (weak negation):
~ψ = truewhenψis undefined → does NOT import presuppositions from alternatives - EXH² (strong negation):
~ψ = #whenψis undefined → DOES import presuppositions from alternatives
Both reuse the same innocently excludable (IE) alternatives computed by Fox's algorithm on the classical truth conditions.
Connection to Presupposition Projection #
@cite{wang-davidson-2026} show that this distinction is empirically consequential for presupposition filtering across disjunction:
- EXH² + any projection theory predicts that exclusive disjunction increases projection (Type A)
- EXH¹ + any projection theory predicts no effect of exclusivity on projection (Type B)
Their Mandarin experiment finds no effect of exclusivity on filtering, consistent with Type B (EXH¹).
Design #
This file is generic infrastructure, not a paper replication.
The IE computation reuses Exhaustification.innocent.excluded
(mathlib-canonical Finset version). The trivalent layer wraps the
bivalent IE result with Truth3 semantics.
Extract the classical (bivalent) truth conditions from a
trivalent proposition: true maps to true; false and
indet both map to false.
The IE computation operates on these classical truth conditions — entailment, consistency, and maximality are all defined bivalently. The trivalent layer applies only after IE is computed.
Pointwise lift of Truth3.toBoolOrFalse.
Instances For
Trivalent EXH¹ (weak negation).
@cite{spector-sudo-2017}'s weak-negation operator (reproduced as (4)/(5) in @cite{wang-davidson-2026}):
- Presupposes whatever φ presupposes: φ(w)=# → EXH¹(w)=#
- Asserts φ and weakly negates all IE alternatives
- Weak negation:
~# = true, so alternatives' presuppositions do NOT project through EXH¹
Type B in @cite{wang-davidson-2026}: predicts no effect of exclusivity on presupposition filtering.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Trivalent EXH² (strong negation).
@cite{spector-sudo-2017}'s strong-negation operator (reproduced as (6)/(7) in @cite{wang-davidson-2026}):
- Presupposes whatever φ presupposes AND whatever the negated IE alternatives presuppose
- Asserts φ and strongly negates all IE alternatives
- Strong negation:
~# = #, so alternatives' presuppositions DO project through EXH²
Type A in @cite{wang-davidson-2026}: predicts that increasing exclusivity reduces presupposition filtering.
Equations
- One or more equations did not get rendered due to their size.
Instances For
EXH¹ preserves the presupposition of the prejacent: if φ(w) = #, then EXH¹(φ)(w) = #.
EXH² also preserves the prejacent's presupposition.