Documentation

Linglib.Phenomena.FreeChoice.Studies.Aloni2022

@cite{aloni-2022}: BSML applied to permission disjunction #

@cite{aloni-2022}

Computational instantiation of NS FC, WS FC, Dual Prohibition, Double Negation FC, and Modal Disjunction (Facts 3, 4, 5, 11, 12) in BSML+ on a 4-world deontic model. Each result is a named theorem invoking the universal substrate theorem in Theories/Semantics/BSML/FreeChoice.lean, applied to the concrete model constructed below.

Out of scope #

The universal results live in Theories/Semantics/BSML/FreeChoice.lean: narrowScopeFC, wideScopeFC, modalDisjunction, dualProhibition, doubleNegationFC, negativeFC_poss_fails_bsmlPlus. The Negative FC failure (Fact 14) on a single Unit world is proved there.

Atoms and worlds #

This file uses typed atoms FCAtom.{a, b} (from Phenomena.FreeChoice.Atoms) rather than String identifiers. The valuation routes directly through PowerSet2World.holds, eliminating the silent typo trap of the earlier String-based pattern (match p with | "coffee" => ... | _ => false).

Worlds are PowerSet2World (nothing/onlyA/onlyB/both), matching Aloni 2022 Figure 1's w_∅/w_a/w_b/w_ab. We label FCAtom.a as "coffee" and FCAtom.b as "tea" in prose only — the formal types are the typed atoms throughout.

The free-choice team: {both, onlyA, onlyB} = {w_ab, w_a, w_b}. Excludes nothing (= w_∅), the world that would supply a zero-witness substate for the disjunction enrichment.

Equations
Instances For

    The prohibition team: just nothing (= w_∅). With restrictiveModel, R[nothing] = {nothing} so the prohibition ¬◇(a ∨ b) is supported.

    Equations
    Instances For

      Deontic model: universal accessibility from every world. Indisputable on every non-empty team (R[w] = R[v] = univ trivially), but not state-based on freeChoiceTeam since R[w] = univ ⊋ freeChoiceTeam.

      Valuation: val a w = w.holds a — direct routing through the typed atom. No String fallthrough, no silent typos.

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

        Restrictive deontic model: from nothing, only nothing is accessible; from any other world, all worlds. Used for Dual Prohibition on the prohibition team {nothing}.

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

          State-based deontic model: R[w] = freeChoiceTeam for every world. Strictly stronger than indisputability; required for Modal Disjunction (Fact 3).

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

            A deontic model where indisputability fails on freeChoiceTeamonlyA and onlyB see different accessible sets. Demonstrates that the indisputable-R precondition of wideScopeFC (Fact 5) is necessary, not incidental: WS FC's conclusion may fail on this model even when its enriched premise is supported. (Aloni 2022 Figure 5(b) shape — non-indisputable R. Figure 5(a) shows the dual case where R is indisputable but enrichment fails.)

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

              ◇(a ∨ b) — narrow-scope premise (Fact 4).

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

                ¬◇(a ∨ b) — Dual Prohibition premise (Fact 11).

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

                  ◇a ∨ ◇b — wide-scope disjunction premise (Fact 5).

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

                    ¬¬◇(a ∨ b) — double-negation premise (Fact 12).

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

                      Fact 4 instantiated at the deontic model + free-choice team: enriched ◇(a ∨ b) entails ◇a ∧ ◇b. Proved by invoking the universal substrate theorem narrowScopeFC.

                      Fact 5 instantiated at the deontic model + free-choice team: enriched (◇a ∨ ◇b) entails ◇a ∧ ◇b, given indisputability. Indisputability holds trivially on this model (universal access), so this is a consequence-direction test — see aloni2022_fact5_WS_FC_fails_loose for the discriminating failure direction.

                      The WS FC enriched premise IS supported on looseDeonticModel for freeChoiceTeam — even though indisputability fails. This pairs with aloni2022_fact5_WS_FC_fails_loose to demonstrate that the implication of WS FC genuinely fails (premise holds, conclusion does not) on this model — not a vacuous failure.

                      Necessity of indisputability: on looseDeonticModel (where indisputability fails on freeChoiceTeam), the WS FC conclusion mayCoffee is not supported on the team — even though the enriched premise IS supported (aloni2022_fact5_premise_supported_loose). The substrate's wideScopeFC cannot apply (its indisputability hypothesis fails), and the implication genuinely breaks: premise holds, conclusion does not.

                      Fact 12 instantiated at the deontic model + free-choice team: enriched ¬¬◇(a ∨ b) entails ◇a ∧ ◇b (FC re-emerges under double negation). The earlier instantiation of this file (lines 117–118 in the pre-refactor version) duplicated the NS FC test rather than exhibiting the entailment from the double-negation premise; this is the correct form.

                      Fact 3 instantiated at the state-based model + free-choice team: enriched (a ∨ b) entails ◇a ∧ ◇b — without ◇ in the premise, by virtue of state-basedness.