Exact preservation criterion.
ProvedMultiverseInvariantFragment.background_branches_iffaether-cataloglogic
Exact preservation criterion. A background condition b (the abstract
stand-in for "the fixed hierarchy of large-cardinal assertions still holds") is
preserved into both a p-branch and a ¬p-branch of the multiverse iff b
meets both ⟦p⟧ and its complement. This is the weakest possible preservation
hypothesis: it is not merely sufficient but necessary.
theorem MultiverseInvariantFragment.background_branches_iff(hR : Rich B) (v : α → B) (b : B) (p : BForm α) :
((∃ U : Generic B, U.mem b ∧ sat (quot v U) p) ∧
(∃ U : Generic B, U.mem b ∧ ¬ sat (quot v U) p)) ↔
(b ⊓ bval v p ≠ ⊥ ∧ b ⊓ (bval v p)ᶜ ≠ ⊥) := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Logic/Multiverse/InvariantFragment.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Logic/Multiverse/InvariantFragment.lean
import Mathlib
import Definitions.Def_Logic_Multiverse_BooleanValuedRealization
import Definitions.Def_Logic_Multiverse_InvariantFragment
/-
# Multiverse-Invariant Fragments, Exact Frame Calibration and Branch Preservation
This file continues `Catalog/Logic/Multiverse/BooleanValuedRealization.lean`, whose
Boolean-valued universe and control frame it takes as given, and settles three
further directions of the multiverse programme.
## 1. Exact calibration of the modal strength of the frame (Direction 2)
* `directed_of_dot2` — the schema `.2` (`◇□p → □◇p`), valid for *all* predicates,
**implies** directedness: `.2` is exactly directedness, so nothing weaker than a
directed frame can validate it.
* `cacc_euclidean_iff` — the forcing frame is Euclidean (i.e. validates `5`)
**iff there are no buttons at all**. One button already destroys `S5`.
* `dot3_fails` — with two *independent* buttons the frame refutes the linearity
axiom `.3`. Hence the frame sits strictly between `S4.2` and `S4.3`: it is a
genuine `S4.2` frame, not accidentally linear.
## 2. The multiverse-invariant fragment theorem (Direction 3)
`buttonFree_invariant` shows that the **button-free fragment** is invariant under
both forcing extensions and grounds; `invariant_iff_buttonFree` proves the sharp
converse: *an assertion is invariant along the button order iff it is equivalent to
a button-free assertion*. The proof substitutes `⊥` for every button atom
(`subBtnFls`) and uses the substitution lemma `csat_subBtnFls`. This is a genuine
characterization, and it is falsifiable in the intended sense: a single button atom
already fails invariance (`btn_atom_not_invariant`).
## 3. Calibration of branching by preserved backgrounds (Direction 4)
The exact preservation criterion is Boolean:
* `background_branches_iff` — a background condition `b` survives into **both** a
`p`-branch and a `¬p`-branch **iff** `b ⊓ ⟦p⟧ ≠ ⊥` and `b ⊓ ⟦p⟧ᶜ ≠ ⊥`.
Applied to the control frame this yields `button_background_branches`: every
*button-definable* background (the abstract stand-in for an indestructible
large-cardinal assertion) is preserved in both CH branches, while
`background_obstruction` exhibits the boundary — a switch-dependent background has
an empty meet with one branch and therefore cannot be preserved.
-/
open MultiverseInvariantFragment
open Multiverse BooleanValuedRealization
/-! ## 1. Exact calibration of the modal strength of the frame -/
variable {W : Type*}
variable {Btn Sw : Type*}
/-! ## 2. The multiverse-invariant fragment -/
variable {Btn Sw : Type*}
/-! ## 3. Calibration of branching by preserved backgrounds -/
variable {α : Type*} {B : Type*} [BooleanAlgebra B]Formal statement
theorem MultiverseInvariantFragment.background_branches_iff(hR : Rich B) (v : α → B) (b : B) (p : BForm α) :
((∃ U : Generic B, U.mem b ∧ sat (quot v U) p) ∧
(∃ U : Generic B, U.mem b ∧ ¬ sat (quot v U) p)) ↔
(b ⊓ bval v p ≠ ⊥ ∧ b ⊓ (bval v p)ᶜ ≠ ⊥) := by sorrySource