Documentation

Studies.AnandNevins2004

Anand & Nevins 2004: shifty operators in changing contexts #

In Zazaki every indexical may shift under vano 'say', and in Slave first-person indexicals shift under hadi, yet indexicals in one speech-context domain never take their values from different contexts (Shift-Together); in Slave the embedding verb decides which indexicals shift and whether they must; and under two stacked reports a shifted indexical in the intermediate clause blocks the lowest indexical from reaching the utterance context. The account: Kaplan's context and index parameters are made the same type, an attitude verb quantifies over indices, and a language's context-shifting operators overwrite the context from the index — all of it in Zazaki, the author coordinate alone in Slave — so shifting is literal overwriting, whence Shift-Together and the loss of the utterance context under an intermediate shift; lexical entries pair verbs with operator options.

Main definitions #

References #

Two parameters of the same type #

@[reducible, inline]
abbrev AnandNevins2004.Params (W : Type u_5) (E : Type u_6) (P : Type u_7) (T : Type u_8) :
Type (max (max (max u_8 u_7) u_6) u_5)

Kaplan's context and index, both of context type (§3.1).

Equations
Instances For
    @[reducible, inline]
    abbrev AnandNevins2004.Expr (W : Type u_5) (E : Type u_6) (P : Type u_7) (T : Type u_8) (R : Type u_9) :
    Type (max (max (max (max u_9 u_8) u_7) u_6) u_5)

    An expression interpreted relative to the two parameters.

    Equations
    Instances For
      def AnandNevins2004.I {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
      Expr W E P T E

      The indexicals read the context parameter (24).

      Equations
      Instances For
        def AnandNevins2004.you {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
        Expr W E P T E
        Equations
        Instances For
          def AnandNevins2004.here {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
          Expr W E P T P
          Equations
          Instances For
            def AnandNevins2004.now {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
            Expr W E P T T
            Equations
            Instances For
              def AnandNevins2004.proSubj {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
              Expr W E P T E

              Subject-controlled PRO reads the author of the index (41); the logophors read its author and addressee (43).

              Equations
              Instances For
                def AnandNevins2004.logAddr {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
                Expr W E P T E
                Equations
                Instances For

                  The context-shifting operators (2): none, Zazaki's all overwriting every coordinate of the context with the index, and Slave's auth overwriting the author coordinate alone.

                  Instances For
                    @[instance_reducible]
                    Equations
                    def AnandNevins2004.Op.act {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :
                    Op → Params W E P T → Params W E P T

                    The rewriting an operator performs on the parameters.

                    Equations
                    Instances For
                      def AnandNevins2004.attitude {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (acc : E → Reference.Context W E P T → Set (Reference.Context W E P T)) (x : E) (φ : Expr W E P T Prop) :
                      Expr W E P T Prop

                      (23b): an attitude verb quantifies over the indices compatible with the attitude and leaves the context parameter untouched.

                      Equations
                      Instances For
                        theorem AnandNevins2004.attitude_all {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (acc : E → Reference.Context W E P T → Set (Reference.Context W E P T)) (x : E) (φ : Expr W E P T Prop) (p : Params W E P T) :
                        attitude acc x (φ ∘ Op.all.act) p ↔ ∀ j ∈ acc x p.2, φ (j, j)

                        (26)–(28): with all as sister of say, the complement is evaluated at the reported context alone, so every indexical in it shifts together.

                        theorem AnandNevins2004.I_auth {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (p : Params W E P T) :
                        I (Op.auth.act p) = p.2.agent

                        (30): under auth the first person reads the reported author while the second person keeps the utterance addressee.

                        theorem AnandNevins2004.you_auth {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (p : Params W E P T) :
                        theorem AnandNevins2004.I_all_eq_proSubj {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} :

                        A shifted first person is read de se: under all it coincides with PRO (§5.1).

                        Lexical entries and readings #

                        @[reducible, inline]

                        A verb's operator options (39): Zazaki vano and Slave TELL take all optionally, Slave WANT takes auth optionally, Slave SAY takes auth obligatorily, and the other Zazaki attitude verbs take none (fn. 3).

                        Equations
                        Instances For
                          def AnandNevins2004.readings {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} {R : Type u_5} (e : Entry) (φ : Expr W E P T R) (c j : Reference.Context W E P T) :
                          List R

                          The values an expression takes under a report with context c and reported context j, one per operator option.

                          Equations
                          Instances For
                            theorem AnandNevins2004.readings_vano_I_you {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c j : Reference.Context W E P T) :
                            readings vano (fun (p : Params W E P T) => (I p, you p)) c j = [(c.agent, c.addressee), (j.agent, j.addressee)]

                            (13): the first and second person under vano read from the same context — the two mixed pairs are not among the readings.

                            theorem AnandNevins2004.mixed_notMem_readings_vano {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} {c j : Reference.Context W E P T} (ha : c.agent ≠ j.agent) (hb : c.addressee ≠ j.addressee) :
                            (c.agent, j.addressee) ∉ readings vano (fun (p : Params W E P T) => (I p, you p)) c j ∧ (j.agent, c.addressee) ∉ readings vano (fun (p : Params W E P T) => (I p, you p)) c j
                            theorem AnandNevins2004.readings_slaveSay {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c j : Reference.Context W E P T) :
                            readings slaveSay (fun (p : Params W E P T) => (I p, you p)) c j = [(j.agent, c.addressee)]

                            (17), (38b): under Slave SAY the first person is the reported author and the second the utterance addressee, obligatorily.

                            theorem AnandNevins2004.readings_slaveTell {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c j : Reference.Context W E P T) :
                            readings slaveTell (fun (p : Params W E P T) => (I p, you p)) c j = [(c.agent, c.addressee), (j.agent, j.addressee)]

                            (36): under Slave TELL both persons shift together.

                            theorem AnandNevins2004.readings_slaveWant {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c j : Reference.Context W E P T) :
                            readings slaveWant (fun (p : Params W E P T) => (I p, you p)) c j = [(c.agent, c.addressee), (j.agent, c.addressee)]

                            (37), (38a): under Slave WANT the first person shifts optionally and the second never.

                            Multiple embedding #

                            theorem AnandNevins2004.opAll_forgets {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c c' j : Reference.Context W E P T) :
                            (Op.all.act (c, j)).1 = (Op.all.act (c', j)).1

                            Overwriting loses the utterance context (§3.4): after all, the context no longer depends on what it was.

                            def AnandNevins2004.readings₂ {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} {R : Type u_5} (e₁ e₂ : Entry) (φ : Expr W E P T R) (c j₁ j₂ : Reference.Context W E P T) :
                            List R

                            Two stacked reports with reported contexts j₁ (intermediate) and j₂ (lowest): the values of an expression in the lowest clause, one per pair of operator options.

                            Equations
                            Instances For
                              theorem AnandNevins2004.readings₂_vano_I {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} (c j₁ j₂ : Reference.Context W E P T) :
                              readings₂ vano vano I c j₁ j₂ = [c.agent, j₂.agent, j₁.agent, j₂.agent]

                              (33): with nothing forcing a shift in the intermediate clause, the lowest first person may read the utterance author.

                              theorem AnandNevins2004.lowerI_of_shifted_you {W : Type u_1} {E : Type u_2} {P : Type u_3} {T : Type u_4} {c j₁ j₂ : Reference.Context W E P T} {o₁ o₂ : Op} (h₁ : o₁ ∈ vano) (h₂ : o₂ ∈ vano) (hyou : you (o₁.act (c, j₁)) = j₁.addressee) (hne : c.addressee ≠ j₁.addressee) :
                              I (o₂.act ((o₁.act (c, j₁)).1, j₂)) ∈ [j₁.agent, j₂.agent]

                              (32): a shifted second person in the intermediate clause diagnoses all there, after which the lowest first person can only be the intermediate or the lowest reported author — never the utterance author.

                              The paper's examples #

                              def AnandNevins2004.Op.shifts (o : Op) (coord : Reference.Context Bool Bool Bool Bool → Bool) :
                              Bool

                              Whether an operator makes a coordinate read the reported context: the coordinate's value after the operator on the distinguishing pair of contexts.

                              Equations
                              • One or more equations did not get rendered due to their size.
                              Instances For
                                theorem AnandNevins2004.rows_track_entries (r : Datum) :
                                r ∈ Examples.all → r.feature? "pragmatic_clash" = none → ∀ e ∈ ((r.feature? "entry").bind Entry.ofString?).toList, ∀ coord ∈ ((r.feature? "indexical").bind coordOfString?).toList, ∀ x ∈ r.readings, (x.1 = "reported context" → (x.2 = Judgment.acceptable ↔ ∃ o ∈ e, o.shifts coord = true)) ∧ (x.1 = "utterance context" → (x.2 = Judgment.acceptable ↔ ∃ o ∈ e, o.shifts coord = false))

                                Every recorded reading of a single embedded indexical is available exactly when some operator option of the embedding verb's entry gives it: the reported-context reading when an option shifts the coordinate, the utterance-context reading when one does not.

                                theorem AnandNevins2004.rows_shift_together (r : Datum) :
                                r ∈ Examples.all → r.feature? "test" = some "shift_together" → ∀ x ∈ r.readings, x.1.startsWith "mixed" = true → x.2 = Judgment.unacceptable

                                Shift-Together (16): wherever the paper lists mixed readings, they are unacceptable.