Documentation

Eval.Projection.ChemlaSchlenker2012

Scoring the accounts on Chemla & Schlenker (2012) #

Chemla and Schlenker's inferential experiments rate, for if-, or- and unless-sentences with too, whether the conditional or the unconditional presupposition follows. Each cell is scored in four steps. The fact is the inference endorsed more in the cell (observed). Its parse is the target sentence, a formula of the propositional fragment that every theory reads (target). The theory says which contexts accept that formula (Theory.accepts). The linking hypothesis, that subjects infer the presupposition a sentence is predicted to carry (the paper's §2.3.1), turns the acceptance conditions into a predicted inference: the unconditional one if the theory predicts it, otherwise the conditional one if the theory predicts that, and nothing if it predicts neither (predicted). Each experiment scores a theory cell by cell (experiment1, experiment4), for a registered account or for a theory under development alike.

Every account accepts a target sentence iff the context entails if p, q, except that the incremental accounts, which coincide on the fragment (Eval.Projection.edges), require the unconditional q in the inverse order (accepts_target_iff). The conditional inference is endorsed more in every tested cell of both experiments (observed_experiment1, observed_experiment4). So every incremental account is wrong on each inverse-order sentence and right on each canonical one, and every other account, the symmetric ones and Limited Symmetry, is right on all (verdict_experiment1, verdict_experiment4).

References #

The incremental accounts.

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

    Acceptance of the target sentences #

    The target sentences carry one trigger.

    theorem Eval.Projection.ChemlaSchlenker2012.accepts_target_iff {W : Type} (I : Fin 3 → Set W) {C : Set W} (a : Account) (k : ChemlaSchlenker2012.Construction) (o : ChemlaSchlenker2012.Order) :
    a.Accepts I C (ChemlaSchlenker2012.target 0 1 2 k o) ↔ C ⊆ if a ∈ incremental ∧ o = ChemlaSchlenker2012.Order.inverse then I 1 else {w : W | w ∈ I 0 → w ∈ I 1}

    Every account accepts a target sentence iff the context entails the conditional presupposition if p, q, except that an incremental account requires the unconditional q in the inverse order.

    Predictions, observations and verdicts #

    The linking hypothesis of the experiments as a rule. A theory predicts that subjects endorse the unconditional inference q from the target sentence of a cell if it predicts that inference, and otherwise the conditional inference if p, q if it predicts that one. A theory that predicts neither predicts nothing about the cell.

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

      The inference endorsed more in a cell, among rows recording the mean rating of an inference in a cell, if the cell was tested and the two means differ.

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

        The inference endorsed more in a cell of Experiment 1 (Table 4).

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

          The inference endorsed more in a cell of Experiment 4 (Table 7).

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

            Experiment 1 of [CS12], whose inferences are rated in Table 4.

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

              Experiment 4 of [CS12], whose inferences are rated in Table 7.

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

                By Table 4, the conditional inference is endorsed more in every tested cell, and the unless-sentence in the canonical order was not tested.

                By Table 7, the conditional inference is endorsed more in every tested cell, and the unless-sentence in the canonical order was not tested.

                On Experiment 1, every incremental account is wrong on each inverse-order sentence and right on each canonical one, and every other account is right on all.

                On Experiment 4, as on Experiment 1, every incremental account is wrong on each inverse-order sentence and right on each canonical one, and every other account is right on all.