Documentation

Linglib.Theories.Pragmatics.RSA.Implementations.ScopeFreezing

RSA Model for Scope Freezing #

RSA model showing that "frozen" scope readings can be rescued by world priors.

Architecture #

  1. Phenomena/ScopeFreezing: Provides examples with observed availability
  2. Grammar (Minimalism/CCG): Both predict freezing for possessor, double-object, etc.
  3. RSA: Takes grammar's parse, shows world priors can rescue frozen readings

The Key Claim #

Grammar theories say: frozen = grammatically impossible (categorical) RSA says: frozen = strongly dispreferred (gradient, rescuable by context)

This module provides the domain types and meaning functions for the scope freezing RSA model.

World state for two-quantifier sentences. w0: neither scope true, w1: inverse-only true, w2: both true

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

          Rescue prior: strongly favor inverse-only world

          Equations
          Instances For

            Phenomena-Dependent Content #

            Definitions and theorems that depend on Phenomena types (Availability, Example, possessor_baseline, possessor_frozen, etc.) have been moved to Linglib.Phenomena.Quantification.Studies.RSAScopeFreezing.

            This includes: interpPriorFromAvailability, interpPriorFromExample, baseline, frozen, grammarInterpPrior, and all associated theorems.