Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 23 additions & 3 deletions frontiers/cap_first_exit_lean_20261002/ALIGNMENT.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@

| Source | Path | Role |
|---|---|---|
| SP | `reviews/d1_cap_elder_partner_second_pass_claude_20260930/REVIEW.md` | the line-numbered reconstruction of CAP §§2–5. L13, L16, L17 and L19–L22 are proved here. L7, L11, L12 and L18 enter `ridge_capHyp` as hypotheses. |
| SP | `reviews/d1_cap_elder_partner_second_pass_claude_20260930/REVIEW.md` | the line-numbered reconstruction of CAP §§2–5. L13, L16, L17 and L19–L22 are proved here. Since v1.5, L1, L2, L5, L7 (maximum and uniqueness), L8, L11's chain rule and L18 are also proved, from SP's normalization and L4's output (`ridge_capHyp_normalized`, `ridge_capHyp_sp`, `ridge_capHyp_of_slices`). Inputs that remain: L3–L4's derivation of `δ`, L7's existence and regularity of `h`, and L12. |
| CAP | `imports/lifetime_parent_20260925/MARKED_CYLINDER_CAP_PROOF.md` | the deterministic theorem that SP reconstructs |

The source setting (SP §2) is as follows:
Expand Down Expand Up @@ -51,6 +51,20 @@ The source setting (SP §2) is as follows:
| `ridge_profile` | on `D = [-2r, 2r]`: `g' = F`, `F(-r/2) = F(r/2) = 0`, `ConvexOn D (F - (x + r/2)(x - r/2)/8)` ⇒ `g ≤ g(-r/2)` on `[-2r, r/2]`; `g(r/2) ≤ g` on `[-r/2, 2r]`; `g(-2r) ≤ g(-r/2) - 9r³/32`; `g(r/2) + 9r³/32 ≤ g(2r)` | SP L13 (sign pattern), L16 (ceiling along the ridge), L17 (both exact integrals `9r³/32`), L20 (ridge minimum) | SP states L12 as `F'' − 1/4 > 0`; Lean takes the weaker convexity of `F − q`. L17's integrals come from comparing `g` with `Q = x³/24 − r²x/32`, `Q' = q`. |
| `ridge_capHyp` | `CapHyp f C M z b s`, with `C = Icc (-2r) (r/2) ×ˢ closedBall 0 (2r)`, `M = (-r/2, h(-r/2))`, `z = (2r, h(2r))`, `b = g(-r/2)`, `s = g(r/2)` | SP L16, L17, L19, L20 | Hypotheses, all local as in SP: `hT` = L7's `f(x, y) ≤ g(x)` on `C`; `hg` = L11 on `[-2r, 2r]`; `hh` = continuity of `h` on `[-r/2, 2r]`; `hFa`, `hFc` = critical pins; `hconv` = L12 on `[-2r, 2r]`; `hside` = L18 (non-strict suffices); `hM` (`h(a) = 0` in SP); `hgap` (SP: `r³/6 < 9r³/32`). The ridge path does not need to stay in `C`. |
| `ridge_saddle_strict` | under `hTs` (the unique transverse maximizer, L7) and a strict `hside`, every frontier point other than `S = (r/2, h(r/2))` has `f < s` | SP L19, "only `S` attains it" | The left face is strict through `hgap`; the right face is strict through uniqueness. |
| `strongConcave_le_of_hasFDerivAt` | `StrongConcaveOn K δ φ`, `p ∈ K`, `HasFDerivAt φ L p`, `y ∈ K` ⇒ `φ y ≤ φ p + L (y - p) - δ/2 ‖y - p‖²` | SP L7 ("strong concavity about the maximizer"), L8, L18 | The engine of the slice rows. `p` need not be interior: only the segment from `p` to `y` is used. Mathlib's `StrongConcaveOn K δ` is uniform concavity with modulus `δ/2 · r²`. In an inner product space it is equivalent to the concavity of `φ + δ/2 ‖·‖²` (Mathlib's `strongConcaveOn_iff_convex`), i.e. SP's `D_y² f ≤ -δ I`. |
| `slice_le` | `L = 0`: `φ y ≤ φ p - δ/2 ‖y - p‖²` | SP L7 (unique transverse maximizer), L16 (`f(x, y) ≤ g(x)`), L19 (`f(c, y) ≤ s - (δ/2)‖y‖²`) | With `δ > 0`, equality only at `y = p`. |
| `slice_ridge_bound` | `0, p ∈ K`, `p` critical, `w` = derivative at `0` ⇒ `δ ‖p‖ ≤ ‖w‖` | SP L8 (`‖h‖ ≤ ‖w‖/δ`) | Two applications of the engine replace SP's strong-monotonicity step. Needs no sign of `δ`. |
| `slice_curved` | on `sphere 0 R`: `φ y < s`, given `p ∈ B̄(0, R)` critical, `‖w‖ ≤ ω ≤ Rδ`, `φ p ≤ b` and `2δ(b - s) < (Rδ - ω)²` | SP L18 (`R = 2r`) | SP: `f ≤ g(x) - (δ/2)‖y - h‖²`, `‖y - h‖ ≥ 2r - ‖h‖`, `‖h‖ ≤ ‖w‖/δ`. Since `(δ/2)(2r - ω/δ)² = (2rδ - ω)²/(2δ)`, the stated condition is exactly SP's comparison, written without division. |
| `slice_hasFDerivAt`, `ridge_hasDerivAt` | the slice's derivative is `D ∘ inr`; if `D ∘ inr = 0` at `(x, h x)` and `h` has derivative `h'`, then `g = f(·, h ·)` has derivative `D (1, 0)` | SP L11 (`F := g' = f_x(x, h(x))`; "the `∇_y f · h'` term vanishes on the ridge") | `D (1, 0)` is `f_x` at `(x, h x)`. |
| `ridge_inputs_of_joint` | joint derivatives `D x` along the ridge, `HasDerivAt h (h' x) x` and `D x ∘ inr = 0` on `[-2r, 2r]` ⇒ `hh`, `hg` with `F x = D x (1, 0)`, and the slice critical points `hcrit` on `[-2r, r/2]` | SP L7 (the ridge equation `∇_y f(x, h(x)) = 0`, `h ∈ C³`), L11 | Differentiability of `h` is L7's implicit-function output. Only first derivatives are used. |
| `ridge_capHyp_of_slices` | the conclusions of `ridge_capHyp` and `ridge_saddle_strict`, with `hT`, `hTs`, `hside` and `hM` replaced by: `hconc` (each slice `δ`-strongly concave on `B̄(0, 2r)` for `x ∈ [-2r, r/2]`, `δ > 0`), `hball` and `hcrit` (`h x ∈ B̄(0, 2r)` is a critical point of its slice), `hw` and `hω` (the slice derivative `w x` at `0` has `‖w x‖ ≤ ω`), `hωr` (`ω ≤ 2rδ`) and `hcurv` (`2δ(b - s) < (2rδ - ω)²`) | SP L4 (`hconc`), L5 (`hω`), L7, L8, L16, L18, L19 | SP takes strong concavity for `x ∈ [-2r, 2r]`; Lean needs it only on the cap's `x`-range. `w x` is a continuous linear functional; with `E` Euclidean, its operator norm is SP's `‖∇_y f(x, 0)‖`. |
| `convexOn_add_node`, `two_node_bound` | `u + (m/2)t²` convex and `u − (m/2)t²` concave on `D`, `u(a) = u(c) = 0` with `a < c` in `D` ⇒ `|u x| ≤ (m/2)|(x − a)(x − c)|` for every `x ∈ D` | SP L5 (two-node interpolation remainder, "valid outside the pin interval too") | The convexity pair is the weak form of `|u''| ≤ m`. Proved by comparing `±u + (m/2)(t − a)(t − c)` with the `convex_zeros_*` lemmas, with no mean-value point `ξ`. |
| `slice_gradient_bound` | for `w : ℝ → E →L[ℝ] ℝ` vanishing at `a < c`, with the convexity pair for each `t ↦ w t v` at modulus `m‖v‖`: `‖w x‖ ≤ (m/2)|(x − a)(x − c)|` on `D` | SP L5 (`‖w(x)‖ ≤ (m/2)|x² − r²/4|`, with `‖w''‖ ≤ m`) | SP takes `e = w(x)/‖w(x)‖`; Lean bounds every `w x v` and uses `opNorm_le_bound`. With `E` Euclidean this is the same norm. |
| `ridge_capHyp_sp` | the conclusions of `ridge_capHyp_of_slices` under `m ≥ 2`, `δ > rm(8m − 5)`, `hconc`, `hball`, `hcrit`, `h(∓r/2) = 0`, `hw`, the convexity pair for `w` on `[−2r, r/2]`, `hh`, `hg`, `hFa`, `hFc`, `hconv` and `b − s ≤ r³/6` | SP §2 normalization (`b − s = r³/6`), L2, L4, L5, L7, L8, L11, L12, L16–L20 | `w(∓r/2) = 0` follows from `h(∓r/2) = 0` and uniqueness of the derivative. `‖w‖ ≤ 2mr²` follows from `slice_gradient_bound`, since `|x² − r²/4| ≤ 4r²` on the cap; `hωr` and `hcurv` then follow from `sp_curved_side_constants`. |
| `hermite_gap_le` | `g' = G` on `[a, c]`, `G(a) = G(c) = 0`, the convexity pair for `G` at modulus `m` ⇒ `g(a) − g(c) ≤ m(c − a)³/12` | SP L1 (`g_0(c) − g_0(a) = −(1/2)∫ φ g_0'''`, `∫ φ = r³/6`) | No integral: `two_node_bound` gives `G ≥ −(m/2)(t − a)(c − t)`, so `g + (m/2)Φ` with `Φ' = (t − a)(c − t)` is monotone, and `Φ(c) = (c − a)³/6`. The pair is the weak form of `|g_0'''| ≤ m`. |
| `hermite_m_ge_two` | on `[−r/2, r/2]`, `g(−r/2) − g(r/2) = r³/6` ⇒ `2 ≤ m` | SP L2 (`m ≥ 2`) | SP's `φ`-weighted average of `f_xxx` is exactly 2, so `m ≥ 2`. Lean proves this inequality, not the average identity, and does not need L2's second statement (a point with `f_xxx = 2`). |
| `ridge_capHyp_normalized` | the conclusions of `ridge_capHyp_sp`, with `hm` replaced by: `g_0 = f(·, 0)` has derivative `G0` on `[−r/2, r/2]`, `G0(∓r/2) = 0`, the convexity pair for `G0` at modulus `m`, and the gap exactly `r³/6` | SP §2 normalization, L1, L2, then as `ridge_capHyp_sp` | `G0(∓r/2) = 0` are the critical pins along `y = 0`. `h(∓r/2) = 0` identifies `b − s` with `g_0(−r/2) − g_0(r/2)`. |
| `sp_curved_side_constants` | `r > 0`, `m ≥ 2`, `δ > r m (8m - 5)`, `0 ≤ ω ≤ 2 m r²` ⇒ `ω ≤ 2rδ` and `2δ · r³/6 < (2rδ - ω)²` | SP L2 (`m ≥ 2`), L4 (`δ > rmk`, `k = 8m - 5`), L5 (`‖w‖ ≤ (15/8) m r² < 2 m r²`), L18 (`4400r³/121 > r³/6`) | So SP's own constants meet `hωr` and `hcurv` at SP's gap `b - s = r³/6`. |
| `CapHyp.congr` | from `g = f` on `closure C` and along one ridge `γ₀`, `CapHyp g C M z b s` | the audit's exterior-invariance corollary (Math-#193 deliverable 2) | `closure C ⊇ C ∪ frontier C`. In SP, `C` is closed. |

## 3. Non-vacuity and necessity, proved in Lean
Expand Down Expand Up @@ -78,6 +92,10 @@ Each necessity theorem is one conjunction, so it certifies its own premises, not

`ridgeToy_capHyp` and `ridgeToy_saddle_strict` instantiate `ridge_capHyp` and `ridge_saddle_strict` at SP's normalization, with `r = 1`, `E = ℝ`, `f(x, y) = x³/3 − x/4 − y²`, `h = 0` and `F = x² − 1/4`. The gap there is exactly `r³/6`, so all ridge hypotheses are jointly satisfiable.

`ridgeToy_joint` and `ridgeToy_slices` run the same toy through the slice route (v1.5). The joint derivative along `y = 0` is `D x = (x² − 1/4) · dx`, its transverse part vanishes, and each slice `y ↦ x³/3 − x/4 − y²` is `2`-strongly concave. With `w = 0`, `ω = 0` and `δ = 2`, the curvature condition reads `2/3 < 16`. Every input of `ridge_inputs_of_joint` and `ridge_capHyp_of_slices` is therefore jointly satisfiable, and the two theorems re-derive (C1)–(C3) and (C2⁺).

The ridge toy has `δ = 2`, below SP's `δ > rm(8m − 5) = 22`. So `spToyF (x, y) = x³/3 − x/4 − 12y²` is added: there `m = 2` (`f_xxx = 2`), `δ = 24`, `w = 0` and the gap is exactly `r³/6`. `spToy_capHyp` shows that every hypothesis of `ridge_capHyp_sp` holds there, so SP's normalized hypotheses are jointly satisfiable. `spToy_normalized` does the same for `ridge_capHyp_normalized`, with `G0 = x² − 1/4`, `|G0''| = 2 = m` and gap exactly `1/6`.

The toy lives on `ℝ`, which is not compact. So the older-peak theorems (`older_peak`, `elder_death_level_peak`) have
no concrete instance exhibited in Lean. Their extra hypotheses (compact, locally connected, continuous `f`) hold on the
torus.
Expand All @@ -91,7 +109,7 @@ illustrates the hypotheses; it is not a model of the planar cap.
|---|---|---|---|
| I1 | Coordinates: SP's cap is `[-2r, r/2] × B̄(0, 2r)` in a frame `φ` of the periodic lift, and SP's "boundary" is the topological frontier. The transport to the torus (`CapHyp.toTorus`) and the face list (`frontier_cylinder`) are proved; the transport needs no embedding radius | SP §2 | Only the identification of SP's objects with these Lean objects remains. |
| I2 | Persistence bookkeeping: in H0 superlevel persistence with distinct critical values, the class born at a local maximum `M` dies at the largest level at which its component contains a strictly higher local maximum (the elder rule) | SP L22, P §8 | This is the definition of the elder pairing. On a compact, locally connected space, Lean proves that this level is `s` (`elder_death_level_peak`). Only the identification of the persistence module's pairing with that definition is informal. |
| I3 | The multi-variable analytic steps L1–L12 and L18, which supply `hT`, `hTs`, `hg`, `hFa`, `hFc`, `hconv`, `hside` and `hgap` of `ridge_capHyp` from `G_r` | SP L1–L12, L18; CAP §§2–5 | Hermite identity, increments, strong concavity, implicit function theorem, ridge identity, `F''` bound and curved-side estimate, with exact constants checked by SP's script, not in Lean. L13, L16, L17, L19 and L20 are proved in Lean from these outputs. |
| I3 | The analytic steps that produce the inputs of `ridge_capHyp_normalized` from `G_r`: L3–L4 (increments, and the transverse Hessian bound giving `δ > rmk` in `hconc`); `m` as a bound on the third derivatives, which gives the convexity pairs for `G0 = ∂_x f(·, 0)` (L1's input `|g_0'''| ≤ m`) and for `w` (L5's input `‖w''‖ ≤ m`); the critical pins `G0(∓r/2) = 0` and `h(∓r/2) = 0`; L6 (the `w'` bound); L7's existence of the transverse maximizer `h(x)` in the ball and the implicit-function regularity of `h`; and L9–L12 (`‖h'‖` and the `F''` bound giving `hconv`) | SP L3, L4, L6, L7, L9–L12; CAP §§2–5 | These are checked with exact constants by SP's script, not in Lean. Since v1.5, Lean proves L1, L2, L5, L7's maximum and uniqueness, L8, L11's chain rule and L18 from those outputs, and checks SP's constants in the curvature step. L13, L16, L17, L19 and L20 were already proved in v1.4. |
| I4 | Theorem A, measurability, `Q^W(G_r^c) ≤ C r³` | P §§7–8, C1, G1 | probability |

## 5. Review contract
Expand All @@ -106,5 +124,7 @@ names:
For each row of §§1–2, the record gives ALIGNED, MISALIGNED (with the corrected statement), or NOT ASSESSED. For each
row of §4, it says whether the interface is stated correctly.

A changed Lean source, toolchain or pin makes an earlier review stale. A same-provider read, including any Claude
A changed Lean source, toolchain or pin makes an earlier review stale.

v1.5 is additive: it inserts the sections `Slice` and the slice-route toy, and updates the file's header comment. The 53 v1.4 declarations and their statements are unchanged, as the diff against v1.4 shows. The v1.4 alignment review ([5396110365](https://github.com/d6g8k5htny-coder/Math-/pull/246#pullrequestreview-5396110365), `ALIGNED_SCOPED`) therefore still covers the v1.4 rows once a reviewer confirms by diff that they are unchanged. A delta review assesses the new rows: the fourteen slice rows of §2, the v1.5 toy paragraphs of §3 and the revised I3. A same-provider read, including any Claude
read, gives zero independence credit. A build or merge is not an alignment review.
Loading
Loading