Brasoveanu (2010): Decomposing Modal Quantification #
[Bra10] interprets discourse relative to plural information states, sets of variable assignments that store both the values of discourse referents and the dependencies between them, in a Dynamic Ty2 reformulation of [Mus96]'s CDRT. Generalized determiners store their restrictor and nuclear scope sets together with the dependencies inside them, and singular pronouns require the dref they retrieve to have one value. The modal half of the paper adds drefs for possible worlds and analyses modal verbs as the modal counterparts of determiners.
This file formalizes the system for individuals of §2. The indefinite, pronoun and proper name
translations (36), (37), (42) compose the wolf discourse (27), whose truth at the dummy state
(35) is derived for any model: A wolf came in. It ate Harvey. is true exactly when some wolf
came in and ate Harvey (wolfDiscourse_true_iff), while a pronoun out of the blue is false
there. On the matrix (64), value inclusion (65) scrambles the dependency between conventions and
women and structured inclusion (66) keeps it.
Implementation notes #
- A row is an assignment
Dref → Flat E, CDRT's canonical register structure atFlat E, and the dummy individual ★ is⊥. - Predication is interpreted distributively over the rows where its drefs have values ((30),
(34)); this is
PCDRT.atomandPCDRT.atom₂.
TODO #
- Dynamic generalized quantifiers (74)–(77), quantificational subordination and donkey anaphora (§2.5–§2.7).
- Intensional PCDRT (§3): world drefs, modal quantifiers and modal subordination.
- The entailment particles of §4.
References #
Indefinites, pronouns and proper names (§2.2) #
Equations
- Brasoveanu2010.instDecidableEqDref x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
A row of a plural info state: an individual or ★ for each dref.
Equations
- Brasoveanu2010.Row E = (Brasoveanu2010.Dref → Flat E)
Instances For
A property, type et: a dref to an update.
Equations
- Brasoveanu2010.Pred E = (Brasoveanu2010.Dref → DynamicSemantics.Update (Set (Brasoveanu2010.Row E)))
Instances For
The dummy assignment i★, which assigns ★ to every dref; {i★} is the dummy info state
(28), which holds no anaphoric information.
Equations
Instances For
(26), (30): a noun or intransitive verb is a test that its argument satisfies it.
Equations
Instances For
(36): the singular indefinite introduces u, makes it singular, and predicates its restrictor
and its scope of it.
Equations
- Brasoveanu2010.indef d P P' = ((SetRel.comp (PCDRT.intro d) (DynamicSemantics.Update.test (PCDRT.sing d))).comp (P d)).comp (P' d)
Instances For
(37): the singular pronoun requires its dref to be singular and passes it to the predicate.
Equations
- Brasoveanu2010.pron d P = SetRel.comp (DynamicSemantics.Update.test (PCDRT.sing d)) (P d)
Instances For
(42), (43): a proper name introduces a dref whose only value is its bearer.
Equations
- Brasoveanu2010.name x d P = (SetRel.comp (PCDRT.intro d) (DynamicSemantics.Update.test {I : Set (Brasoveanu2010.Row E) | PCDRT.value d I = {x}})).comp (P d)
Instances For
(45a): a transitive verb applies its object's quantifier to the relation.
Equations
- Brasoveanu2010.transVerb R Q v = Q fun (v' : Brasoveanu2010.Dref) => DynamicSemantics.Update.test (PCDRT.atom₂ R v v')
Instances For
The wolf discourse (27), (46): A^u wolf came in. It_u ate Harvey^{u'}.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The truth conditions of (27) at the dummy state (35): the discourse can update {i★} exactly
when some wolf came in and ate Harvey.
Structured inclusion (63)–(66) #
The matrix (64): the conventions α₁ … α₄ in u₁, the women β₁ … β₄ courted at them in u₂,
and two ways of storing a subset of the conventions in u₃.
Equations
- Brasoveanu2010.instDecidableEqDref₃ x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
Equations
- Brasoveanu2010.instDecidableEqInd x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
A row of (64) from its u₁, u₂ and u₃ values.
Equations
- Brasoveanu2010.row a b c Brasoveanu2010.Dref₃.u₁ = a
- Brasoveanu2010.row a b c Brasoveanu2010.Dref₃.u₂ = b
- Brasoveanu2010.row a b c Brasoveanu2010.Dref₃.u₃ = c
Instances For
Both columns store a subset of the conventions (65).
The value-inclusion column is not a structured subset (66).
The structured column is a fully structured subset (68).