Documentation

Studies.HarrisPotts2009

Harris and Potts (2009): Perspective-shifting with appositives and expressives #

This file formalizes [HP09]'s account of whose commitments an appositive or expressive expresses. The two-dimensional meaning of [Pot05] is kept, but the conventional-implicature content carries a free orientation variable, resolved in discourse to the speaker or to another participant (CIItem.resolve). Orientation is public commitment: an utterance commits the speaker to the at-issue content and whoever the item is oriented to to the implicature (CIItem.utter), so under a non-speaker-oriented reading the speaker's commitment set narrows by the at-issue content alone (contextSet_utter_speaker_other). Against the configurational hypothesis, on which non-speaker orientation is semantic binding by an attitude predicate (10a), the paper sets the contextual one (10b): speaker orientation is a default that pragmatic factors override in any syntactic position. The attested readings and the items of the two experiments are the rows of Data/Examples/HarrisPotts2009.json, and not_configurational is the paper's refutation: (11) and (12) are epithets in matrix clauses read from another's perspective. Experiment 1 extends this to appositives, read as the attitude subject's in 68% of unembedded and 86% of embedded trials, so that embedding facilitates the shift without being required for it; Experiment 2 finds a negative context raising subject-oriented readings of a matrix epithet from 7% to 17%; and the corpus study finds speaker orientation dominant, in 32 of the 34 embedded appositive relatives with agreed textual evidence.

Implementation notes #

The commitment state is Commitment.State over Orientation, so the values of the orientation variable are exactly the possible committers. The Both response of the experiments, which the paper reads as pragmatic enrichment of one attribution, is recorded in the rows but not modelled, and the regression models of §3.3 and §4.3 stay in prose. The rows of the experimental items carry their conditions, not the per-item response counts, which the paper reports only in aggregate.

References #

inductive HarrisPotts2009.Orientation (Person : Type) :

Whose commitment a conventional implicature expresses: the speaker, or another discourse participant.

Instances For
    @[instance_reducible]
    instance HarrisPotts2009.instDecidableEqOrientation {Person✝ : Type} [DecidableEq Person✝] :
    DecidableEq (Orientation Person✝)
    Equations
    @[instance_reducible]
    instance HarrisPotts2009.instReprOrientation {Person✝ : Type} [Repr Person✝] :
    Repr (Orientation Person✝)
    Equations
    def HarrisPotts2009.instReprOrientation.repr {Person✝ : Type} [Repr Person✝] :
    Orientation Person✝ → ℕ → Std.Format
    Equations
    • One or more equations did not get rendered due to their size.
    Instances For
      structure HarrisPotts2009.CIItem (Person W : Type) :

      An appositive or expressive: at-issue content, and conventional-implicature content that depends on the value of the orientation variable, resolved in discourse.

      • ciFor : Orientation Person → W → Prop

        The conventional implicature under each orientation.

      • atIssue : W → Prop

        The at-issue content, which does not depend on the orientation.

      Instances For
        def HarrisPotts2009.CIItem.resolve {Person W : Type} (item : CIItem Person W) (o : Orientation Person) :

        The two-dimensional meaning once the orientation variable is resolved to o.

        Equations
        Instances For
          def HarrisPotts2009.CIItem.utter {Person W : Type} (item : CIItem Person W) (o : Orientation Person) (K : Commitment.State (Orientation Person) W) :

          Orientation is public commitment (§1): uttering the item with its orientation resolved to o commits the speaker to the at-issue content and o to the conventional implicature.

          Equations
          Instances For

            A speaker-oriented reading commits the speaker to both tiers.

            A non-speaker-oriented reading narrows the speaker's commitment set by the at-issue content alone: the appositive of (8) or the expressive of (9) is no commitment of the speaker's.

            The participant the item is oriented to is committed to the implicature and to nothing at-issue.

            The configurational and the contextual hypothesis (10) #

            Whose view the content was taken to be: the response categories of (13) and (16).

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

                An attested reading: whether the item sits in the complement of an attitude or speech predicate, and to whom its content was attributed.

                Instances For
                  def HarrisPotts2009.instDecidableEqReading.decEq (x✝ x✝¹ : Reading) :
                  Decidable (x✝ = x✝¹)
                  Equations
                  • One or more equations did not get rendered due to their size.
                  Instances For
                    def HarrisPotts2009.instReprReading.repr :
                    Reading → ℕ → Std.Format
                    Equations
                    • One or more equations did not get rendered due to their size.
                    Instances For

                      Read a reading off a row that records an attribution.

                      Equations
                      Instances For

                        The attested readings of §2 and §5.

                        Equations
                        Instances For

                          (10a): non-speaker orientation arises only by semantic binding under an attitude predicate, so an unembedded item is read as the speaker's.

                          Equations
                          Instances For

                            (11) and (12): an epithet in a matrix clause read from another's perspective.

                            The configurational hypothesis is refuted by the attested readings; what remains is the contextual hypothesis of (10b), on which the orientation variable is resolved by pragmatic factors in any syntactic position.