Documentation

Linglib.Theories.Semantics.Causation.Resultatives

Resultatives as Concealed Causatives #

@cite{baglini-bar-asher-siegal-2025} @cite{goldberg-jackendoff-2004} @cite{levin-2019} @cite{martin-rose-nichols-2025} @cite{nadathur-lauer-2020}

Connects the resultative construction to the causative semantics infrastructure:

  1. Causal Dynamics: causative resultatives modeled as concrete CausalDynamics with structural sufficiency proofs and CC-selection constraints (Baglini & Bar-Asher @cite{baglini-bar-asher-siegal-2025})
  2. Tightness via Causal Necessity: @cite{levin-2019} concealed causative constraint — intervening causers with independent energy sources disrupt necessity under counterfactual intervention, formalized through completesForEffect (not graph-structural checks)
  3. Thick/Thin Convergence: three independent paths (@cite{martin-rose-nichols-2025}, @cite{goldberg-jackendoff-2004}, @cite{embick-2009}) converge on .make/makeSem
  4. Aspect: resultative telicizes activity verbs via bounded RP
  5. ChangeOfState: constructional BECOME maps to CoSType.inception
  6. Müller decomposability: all subconstructions decompose into universal schemata

Causal Dynamics (@cite{nadathur-lauer-2020}; Baglini & Bar-Asher @cite{baglini-bar-asher-siegal-2025}) #

The constructional CAUSE in causative resultatives maps to @cite{nadathur-lauer-2020} causal sufficiency: the verbal subevent is sufficient for the constructional result. We build concrete CausalDynamics and prove structural sufficiency/necessity results.

CC-selection (Baglini & Bar-Asher @cite{baglini-bar-asher-siegal-2025}) #

The resultative construction performs CC-selection: it constrains which condition from a causal model can fill the cause role. Causative resultatives select via completion of a sufficient set: the verbal subevent must be the final condition that makes the result inevitable. Combined with the temporal constraint (G&J Principle 33), this gives the BBS2025 completion event analysis.

Causal models for causative resultatives #

Each causative resultative maps to a concrete CausalDynamics where the verbal subevent variable causally determines the result variable.

Sufficiency proofs (@cite{nadathur-lauer-2020} Def 23) #

Necessity proofs #

Single-pathway resultatives: verbal subevent is both sufficient and necessary.

Hammering is causally necessary for flatness. Under @cite{nadathur-2024} Def 10b, the background must NOT include the cause (precondition rejects it).

Noncausative resultatives = empty dynamics #

"The river froze solid": RESULT relation, no CAUSE.

Agreement with Boolean flags #

CC-selection (Baglini & Bar-Asher @cite{baglini-bar-asher-siegal-2025}) #

Different causative constructions constrain which condition from a causal model can be selected as "the cause." BBS2025 call this CC-selection.

How a causative construction selects its cause (BBS2025).

  • memberOfSufficientSet : CCSelectionMode

    Overt "cause": subject is any member of a sufficient set (BBS2025 §4.1)

  • completionOfSufficientSet : CCSelectionMode

    CoS/resultative: subject completes a sufficient set (BBS2025 §4.2)

Instances For
    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      BBS2025 completion event: sufficient + but-for necessary in context. Uses simple counterfactual but-for rather than @cite{nadathur-2024} Def 10b's supersituation necessity: the cause must be needed in the CURRENT background, not resilient against all possible supersituations. This is the right granularity for completion events — intermediate variables in causal chains are passive (no independent source), so simple but-for captures tightness correctly.

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For

        Temporal constraint as completion event (G&J Principle 33) #

        Causal completion + temporal ordering.

        Equations
        • One or more equations did not get rendered due to their size.
        Instances For

          CausativeBuilder bridge #

          @cite{levin-2019}'s Tightness via Causal Necessity #

          Resultatives require tightness: no intervening causer with its own energy source. Tightness ≡ completesForEffect (sufficiency + necessity). The necessity component does the explanatory work: causallyNecessary runs normalDevelopment with cause=false. An intermediate with an independent energy source still fires without the cause, so necessity fails.

          ExampleChainIndep. source?completesForEffect
          "hammer metal flat"directtrue
          "drink teapot dry"passive chainNotrue
          "kick door open" (direct)directtrue
          *"kick door open" (via ball)active chainYesfalse

          hasDirectLaw (graph inspection) is insufficient: it rejects "drink teapot dry" (passive chain, no direct law, but tight).

          "Drink the teapot dry": passive chain (Levin §4). drinking → tea_removal → teapot_dry. Tea removal has no independent energy source.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For

            *"Kick the door open" via ball (UNAVAILABLE, Levin §7). kick → ball_motion → door_open, plus ball_energy → ball_motion. The ball has an independent energy source.

            Equations
            • One or more equations did not get rendered due to their size.
            Instances For

              "Kick the door open" (direct, available reading). kick → door_open.

              Equations
              • One or more equations did not get rendered due to their size.
              Instances For

                Tightness = completesForEffect #

                Each theorem runs normalDevelopment under counterfactual intervention (cause=true for sufficiency, cause=false for necessity).

                Tight despite being a chain: removing drinking leaves tea_removal undetermined (no independent source), so teapot_dry doesn't fire.

                Intervening causer = independent source #

                theorem Causative.Resultatives.independent_source_disrupts_tightness (cause intermediate effect indepSrc : Core.StructuralEquationModel.Variable) (hci : cause intermediate) (hce : cause effect) (hie : intermediate effect) (hsc : indepSrc cause) (hsi : indepSrc intermediate) (hse : indepSrc effect) :

                For a chain a → b → c, an active independent source for b disrupts completesForEffect for a → c. The independent source fires b even with a=false, so a is not necessary.

                Contiguity (Levin §4) #

                Nonselected-NP resultatives require contiguity between verb's object and affected entity. All types involve passive intermediates (no independent source), so completesForEffect holds.

                Spatial contiguity preserving tightness in nonselected NP resultatives.

                Instances For
                  Equations
                  • One or more equations did not get rendered due to their size.
                  Instances For

                    Three-Way Convergence: Thick ↔ P-CAUSE ↔ Resultative #

                    Three independently-defined modules converge on the same prediction:

                    1. ProductionDependence.lean: Thick manner → P-CAUSE → .make builder
                    2. Resultatives.lean: Resultative CAUSE → .make builder
                    3. ThickThin.lean: Thick manner → ASR-compatible (empirical data)

                    This convergence is non-trivial: each path was formalized independently following different papers (@cite{martin-rose-nichols-2025}, @cite{goldberg-jackendoff-2004}, @cite{embick-2009}). The fact that they agree validates the cross-module architecture.

                    Aspect: activity + bounded RP → accomplishment #

                    ChangeOfState: BECOME = inception (¬P → P) #

                    Müller Decomposability #

                    All four resultative subconstructions are fully abstract and therefore decomposable into Müller's universal schemata.

                    Causative subconstructions → [HS, HC, HC] (same as parent resultative) Noncausative subconstructions → [HS, HC] (intransitive, fewer complements)

                    Cross-linguistic Resultative Parameters (@cite{tay-2024}) #

                    Resultative constructions vary cross-linguistically along dimensions orthogonal to the G&J subconstruction typology. @cite{tay-2024} identifies three: realization (how the result predicate is morphosyntactically encoded), orientation (whether DOR holds), and phase grammaticalization.

                    How the result predicate is morphosyntactically realized.

                    Instances For
                      Equations
                      • One or more equations did not get rendered due to their size.
                      Instances For

                        Whether the result predicate targets the subject or object. English enforces DOR (Direct Object Restriction): the result must predicate of the direct object. Mandarin allows subject-oriented resultatives productively (kū-lèi "cry-tired").

                        Instances For
                          Equations
                          • One or more equations did not get rendered due to their size.
                          Instances For

                            Phase complements: a grammaticalized closed-class subset of Mandarin V2 resultatives with fixed change-of-state semantics. Maps to CoSType from ChangeOfState.Theory.

                            Instances For
                              Equations
                              • One or more equations did not get rendered due to their size.
                              Instances For