Uniqueness of Participant (UP): each event has at most one θ-filler. eq. (43): θ(x,e) ∧ θ(y,e) → x = y.
Equations
- Semantics.Events.Krifka1998.UP θ = ∀ (x y : α) (e : β), θ x e → θ y e → x = y
Instances For
Cumulative theta (CumTheta): θ preserves sums. eq. (44): θ(x,e) ∧ θ(y,e') → θ(x⊕y, e⊕e'). This is the relational analog of IsSumHom.
Equations
- Semantics.Events.Krifka1998.CumTheta θ = ∀ (x y : α) (e e' : β), θ x e → θ y e' → θ (x ⊔ y) (e ⊔ e')
Instances For
Mapping to Events (ME): object parts map to event parts. eq. (45): θ(x,e) ∧ y ≤ x → ∃e'. e' ≤ e ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.ME θ = ∀ (x : α) (e : β) (y : α), θ x e → y ≤ x → ∃ e' ≤ e, θ y e'
Instances For
Mapping to Strict subEvents (MSE): proper object parts map to proper subevents. eq. (46): θ(x,e) ∧ y < x → ∃e'. e' < e ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.MSE θ = ∀ (x : α) (e : β) (y : α), θ x e → y < x → ∃ e' < e, θ y e'
Instances For
Uniqueness of Events (UE): each object part maps to a unique event part. eq. (47): θ(x,e) ∧ y ≤ x → ∃!e'. e' ≤ e ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.UE θ = ∀ (x : α) (e : β) (y : α), θ x e → y ≤ x → ∃ e' ≤ e, θ y e' ∧ ∀ e'' ≤ e, θ y e'' → e'' = e'
Instances For
Mapping to Objects (MO): event parts map to object parts. eq. (48): θ(x,e) ∧ e' ≤ e → ∃y. y ≤ x ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.MO θ = ∀ (x : α) (e e' : β), θ x e → e' ≤ e → ∃ y ≤ x, θ y e'
Instances For
Mapping to Strict subObjects (MSO): proper subevents map to proper object parts. eq. (49): θ(x,e) ∧ e' < e → ∃y. y < x ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.MSO θ = ∀ (x : α) (e e' : β), θ x e → e' < e → ∃ y < x, θ y e'
Instances For
Uniqueness of Objects (UO): each event part maps to a unique object part. eq. (50): θ(x,e) ∧ e' ≤ e → ∃!y. y ≤ x ∧ θ(y,e').
Equations
- Semantics.Events.Krifka1998.UO θ = ∀ (x : α) (e e' : β), θ x e → e' ≤ e → ∃ y ≤ x, θ y e' ∧ ∀ z ≤ x, θ z e' → z = y
Instances For
Generalized Uniqueness of Events (GUE): each object participates in at most one event. eq. (52): θ(x,e) ∧ θ(x,e') → e = e'.
Equations
- Semantics.Events.Krifka1998.GUE θ = ∀ (x : α) (e e' : β), θ x e → θ x e' → e = e'
Instances For
Strict Incrementality (SINC): the conjunction of MSO, UO, MSE, UE plus a non-degeneracy condition requiring extended entities. eq. (51):
SINC(θ) iff (i) MSO(θ) ∧ UO(θ) ∧ MSE(θ) ∧ UE(θ) (ii) ∃x,y∈U_P ∃e,e'∈U_E [y < x ∧ e' < e ∧ θ(x,e) ∧ θ(y,e')]
Condition (i) guarantees a bijective correspondence between the part structure of the object and the part structure of the event. Condition (ii) ensures θ actually applies to extended entities (ones with proper parts), ruling out degenerate cases where θ only relates atoms. This is the key property of incremental-theme verbs like "eat", "build", "read".
- mso : MSO θ
Proper subevents map to proper object parts.
- uo : UO θ
Each event part has a unique object counterpart.
- mse : MSE θ
Proper object parts map to proper subevents.
- ue : UE θ
Each object part has a unique event counterpart.
Non-degeneracy: θ applies to at least one extended entity pair. eq. (51.ii): the relation must involve entities with proper parts, not just atoms.
Instances For
MSE implies ME: weaken strict to non-strict. : if proper parts map to proper subevents, then parts (including the whole) also map, taking e' = e when y = x.
MSO implies MO: weaken strict to non-strict (dual of me_of_mse).
SINC implies ME (via MSE).
SINC implies MO (via MSO).
VP predicate formed by existential closure over the object argument. eq. (53): VP_θ,OBJ = λe.∃y[OBJ(y) ∧ θ(y,e)]. "eat apples" = λe.∃y[apples(y) ∧ eat.theme(y,e)].
Equations
- Semantics.Events.Krifka1998.VP θ OBJ e = ∃ (y : α), OBJ y ∧ θ y e
Instances For
CUM propagation: cumulative θ transmits CUM from NP to VP. §3.3: CumTheta(θ) ∧ CUM(OBJ) → CUM(VP θ OBJ).
"eat apples" is CUM because:
- APPLES is CUM (mass/bare-plural NPs are cumulative)
- EAT's incremental theme is cumulative (CumTheta)
- Therefore VP = λe.∃y[apples(y) ∧ eat(y,e)] is CUM
Proof: Given VP(e₁) and VP(e₂), we have y₁ with OBJ(y₁) ∧ θ(y₁,e₁) and y₂ with OBJ(y₂) ∧ θ(y₂,e₂). By CumTheta, θ(y₁⊔y₂, e₁⊔e₂). By CUM(OBJ), OBJ(y₁⊔y₂). So VP(e₁⊔e₂).
QUA propagation: UP + MSO + UO transmit QUA from NP to VP. §3.3: UP(θ) ∧ MSO(θ) ∧ UO(θ) ∧ QUA(OBJ) → QUA(VP θ OBJ).
"eat two apples" is QUA because:
- TWO-APPLES is QUA (quantized NPs have no P-proper-parts)
- EAT's incremental theme is SINC + UP
- Therefore VP = λe.∃y[two-apples(y) ∧ eat(y,e)] is QUA
Proof: Suppose VP(e) via ⟨y, OBJ(y), θ(y,e)⟩ and e' < e. We must show ¬VP(e'). Suppose for contradiction VP(e') via ⟨z, OBJ(z), θ(z,e')⟩. By UP, z = y' for any other filler of e'. By MSO, ∃ y' < y with θ(y',e'). By UP, z = y'. So OBJ(y') with y' < y. But QUA(OBJ) says ¬OBJ(y'). Contradiction.
Functional case: When θ is a function (not a relation) with
IsSumHom + Function.Injective, this reduces to
qua_of_injective_sumHom in Core/Mereology.lean via
qua_pullback. The relational UP + MSO conditions collapse to
functional injectivity + monotonicity.
QUA propagation from SINC + UP (convenience wrapper). In practice, incremental-theme verbs satisfy both SINC and UP.
Graduality (GRAD): more object measure entails more event measure. @cite{krifka-1989}: GRAD(θ, μ_obj, μ_ev) ⇔ ∀x,y,e,e'. θ(x,e) ∧ θ(y,e') ∧ μ_obj(x) < μ_obj(y) → μ_ev(e) < μ_ev(e'). GRAD captures the intuition that eating more food takes more time.
Equations
- Semantics.Events.Krifka1998.GRAD θ μ_obj μ_ev = ∀ (x y : α) (e e' : β), θ x e → θ y e' → μ_obj x < μ_obj y → μ_ev e < μ_ev e'
Instances For
SINC + extensive measures + measure proportionality → GRAD. @cite{krifka-1989}: strictly incremental themes with extensive measures and a constant consumption rate exhibit gradual change.
Proof: if μ_ev(e) = c · μ_obj(x) and μ_ev(e') = c · μ_obj(y), then μ_obj(x) < μ_obj(y) implies c · μ_obj(x) < c · μ_obj(y) since c > 0, giving μ_ev(e) < μ_ev(e').
The GRAD square extends LaxMeasureSquare with strict incrementality
(SINC), enabling the derivation of GRAD (gradual change). The non-SINC
theorems (MereoDim, QUA pullback, lax commutativity) are inherited from
LaxMeasureSquare.
- laxComm : MeasureProportional θ μ_obj (dur ∘ τ_fn)
- ext₁ : ExtMeasure α μ_obj
- ext₂ : ExtMeasure γ (dur ∘ τ_fn)
- sinc : SINC θ
Strict incrementality of the thematic role.
Instances For
The defining equation of the GRAD square: for any θ-pair (x,e), the temporal measure equals the rate times the object measure.
GRAD follows from the square via grad_of_sinc.
The object arm is a MereoDim (via ExtMeasure).
The temporal arm (composed path) is a MereoDim (via ExtMeasure).
QUA pullback through the temporal path: QUA on ℚ pulls back to
QUA on events via the composed measure dur ∘ τ_fn.
Incrementality annotations for verbs. §3.2–3.7: verbs differ in which thematic role properties their incremental theme satisfies.
- sinc : VerbIncClass
Strictly incremental: consumption/creation verbs (eat, build). SINC + GUE on object role. Krifka §3.3.
- inc : VerbIncClass
Incremental: allows backups (read, write). INC but ¬SINC on object role. Krifka §3.6.
- cumOnly : VerbIncClass
Cumulative only: non-incremental (push, carry). CumTheta but ¬MSE on object role. Krifka §3.2.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- Semantics.Events.Krifka1998.instBEqVerbIncClass.beq x✝ y✝ = (x✝.ctorIdx == y✝.ctorIdx)
Instances For
Meaning postulates for verb incrementality. These axiomatize which verbs have which incrementality properties on their object/theme role.
- eat_sinc : SINC eat
"eat" has a strictly incremental theme: consumption creates a bijection between food parts and eating subevents.
- eat_gue : GUE eat
"eat" satisfies GUE: each portion of food is eaten at most once.
- push_cum : CumTheta push
"push" is cumulative but not strictly incremental: the cart doesn't have parts that correspond to pushing subevents.
- build_sinc : SINC build
"build" has a strictly incremental theme: creation mirrors consumption.
- build_gue : GUE build
"build" satisfies GUE: each part is built at most once.
- read_cum : CumTheta read_
"read" is incremental but not strictly so: allows re-reading.
Instances
General Incrementality (INC): θ is the closure of some strictly incremental θ' under sum formation. eq. (59): θ is incremental iff there exists a SINC θ' such that θ(x,e) holds iff (x,e) can be decomposed into θ'-pairs that sum to (x,e). This handles "read the article" (allows re-reading/backups): reading events are built from strictly incremental reading-parts that can overlap in their object coverage.
Formulated inductively: the smallest relation containing θ' and closed under componentwise sum.
- base
{α : Type u_1}
{β : Type u_2}
[SemilatticeSup α]
[SemilatticeSup β]
{θ' : α → β → Prop}
{x : α}
{e : β}
: θ' x e → IncClosure θ' x e
Base: anything in θ' is in the closure.
- sum
{α : Type u_1}
{β : Type u_2}
[SemilatticeSup α]
[SemilatticeSup β]
{θ' : α → β → Prop}
{x₁ x₂ : α}
{e₁ e₂ : β}
: IncClosure θ' x₁ e₁ → IncClosure θ' x₂ e₂ → IncClosure θ' (x₁ ⊔ x₂) (e₁ ⊔ e₂)
Sum: if (x₁,e₁) and (x₂,e₂) are in the closure, so is (x₁⊔x₂, e₁⊔e₂).
Instances For
General Incrementality: θ is the IncClosure of some SINC θ'.
Equations
- Semantics.Events.Krifka1998.INC θ = ∃ (θ' : α → β → Prop), Semantics.Events.Krifka1998.SINC θ' ∧ ∀ (x : α) (e : β), θ x e ↔ Semantics.Events.Krifka1998.IncClosure θ' x e
Instances For
SINC + CumTheta implies INC: a strictly incremental θ that is cumulative is trivially its own closure. CumTheta ensures the reverse direction: IncClosure θ ⊆ θ.
Bridge: CUM propagation via CumTheta + CUM(OBJ). CumTheta(θ) ∧ CUM(OBJ) → CUM(VP θ OBJ). In CEM models, SINC implies CumTheta, so this covers the "SINC verb + CUM noun → CUM VP" case from §3.3.
Bridge: QUA propagation from SINC + UP (renamed for API).
Bridge: RoleHom (functional θ from Mereology.lean) implies CumTheta (relational θ). A sum-homomorphic function θ : β → α induces a cumulative relation λ x e, θ(e) = x.
With injectivity, we additionally get StrictMono via
IsSumHom.strictMono_of_injective (Core/Mereology.lean §10),
which enables qua_of_injective_sumHom — the functional
QUA propagation theorem.