Pith. sign in

IndisputableMonolith.Cosmology.PolarizedBirthInterfaceCost

IndisputableMonolith/Cosmology/PolarizedBirthInterfaceCost.lean · 367 lines · 22 declarations

show as:
view math explainer →

open module explainer GitHub source

Explainer status: pending

   1import Mathlib
   2import IndisputableMonolith.Cosmology.LatticeBallEdges
   3import IndisputableMonolith.Cost
   4import IndisputableMonolith.Constants
   5
   6/-!
   7# The recognition COST of the coarsening ledger: carried bulk is free, the interface is paid
   8
   9`LatticeBallEdges` (Phase 54) closed the edge-*count* ledger: every adjacency of the polarized birth
  10field is either a carried monochromatic edge or a forced bichromatic interface edge, and it counted
  11each side exactly (`total = interface + carried`). This module weights that ledger by the actual
  12recognition cost the engine posts, the forced cost `J(x) = (x + x⁻¹)/2 - 1` (`Cost.Jcost`, unique by
  13`Cost.FunctionalEquation`) evaluated at the phi-rung gap between two cells' charges.
  14
  15The charge of the polarized field is `sign(x) ∈ {+1, 0, -1}` (`PolarizedBirthDomains.polarized`), and
  16the cost of one ordered adjacency is `J(φ^(charge p - charge q))`. Two facts pin it down:
  17
  18* a **carried** (monochromatic) edge has equal charges, so it spans `0` phi-rungs and costs
  19  `J(φ^0) = J(1) = 0` exactly (`Cost.Jcost_unit0`). The bulk is carried for *literally zero*
  20  recognition cost.
  21* an **interface** (bichromatic) edge has one endpoint on the `x = 0` spine and the other at
  22  `x = ±1` (`edge_structure`), so the charges differ by exactly `±1`: it spans one phi-rung and costs
  23  `J(φ^(±1)) = J(φ)` exactly (`Cost.Jcost_symm`, reciprocal symmetry).
  24
  25So the entire recognition cost of the field is the interface count times `J(φ)`:
  26
  27* 2D diamond: `totalCost t = (8t - 4) • J(φ)`, `carriedCost t = 0` (`Diamond`).
  28* 3D octahedron: `totalCost t = (8t² - 8t + 4) • J(φ)`, `carriedCost t = 0` (`Octahedron`).
  29
  30with `J(φ) = (√5 - 2)/2 > 0` (`Jcost_phi_pos`), a genuine positive cost. This is the literal,
  31cost-unit statement of the north star's compute-watch law: cost scales with the recognition activity
  32(the codimension-1 interface), not with the bulk volume the engine carries coarse for free.
  33
  34THEOREM over `ℝ` and `ℕ` with 0 `sorry` and only the three standard axioms (`propext`,
  35`Classical.choice`, `Quot.sound`).
  36-/
  37
  38namespace IndisputableMonolith
  39namespace Cosmology
  40namespace PolarizedBirthInterfaceCost
  41
  42open Finset
  43open scoped BigOperators
  44
  45/-! ## §0. The phi-rung cost, shared by both dimensions -/
  46
  47/-- The recognition cost of an ordered adjacency whose two cells' charges differ by `d` phi-rungs:
  48`J(φ^d)`. The forced cost `J` is `Cost.Jcost`; `φ` is the forced golden ratio `Constants.phi`. -/
  49noncomputable def Jpow (d : ℤ) : ℝ := Cost.Jcost (Constants.phi ^ d)
  50
  51/-- A zero-rung gap (a carried, monochromatic edge) costs nothing: `J(φ^0) = J(1) = 0`. -/
  52lemma Jpow_zero : Jpow 0 = 0 := by
  53  rw [Jpow, zpow_zero, Cost.Jcost_unit0]
  54
  55/-- A one-rung-up gap costs `J(φ)`. -/
  56lemma Jpow_one : Jpow 1 = Cost.Jcost Constants.phi := by
  57  rw [Jpow, zpow_one]
  58
  59/-- A one-rung-down gap costs `J(φ)` too, by reciprocal symmetry `J(φ⁻¹) = J(φ)`. -/
  60lemma Jpow_neg_one : Jpow (-1) = Cost.Jcost Constants.phi := by
  61  have h : Constants.phi ^ (-1 : ℤ) = Constants.phi⁻¹ := by
  62    rw [zpow_neg, zpow_one]
  63  rw [Jpow, h, ← Cost.Jcost_symm Constants.phi_pos]
  64
  65/-- Either a single rung up or a single rung down costs exactly `J(φ)`: the recognition cost of a
  66forced interface distinction. -/
  67lemma Jpow_of_abs_one {d : ℤ} (h : d = 1 ∨ d = -1) : Jpow d = Cost.Jcost Constants.phi := by
  68  rcases h with h | h <;> subst h
  69  · exact Jpow_one
  70  · exact Jpow_neg_one
  71
  72/-- `J(φ) > 0`: a forced interface distinction has a genuine, strictly positive recognition cost.
  73`J(φ) = (φ - 1)²/(2φ) = (√5 - 2)/2`. -/
  74lemma Jcost_phi_pos : 0 < Cost.Jcost Constants.phi := by
  75  rw [Cost.Jcost_eq_sq Constants.phi_ne_zero]
  76  apply div_pos
  77  · have hne : Constants.phi - 1 ≠ 0 := sub_ne_zero.mpr Constants.phi_ne_one
  78    positivity
  79  · have := Constants.phi_pos; linarith
  80
  81/-! ## §1. The 2D diamond cost ledger -/
  82
  83namespace Diamond
  84
  85open InterfaceComponentBound.Diamond
  86open PolarizedBirthDomains.Diamond (polarized)
  87open PolarizedBirthInterface.Diamond (B interface_card_eq)
  88open LatticeBallEdges.Diamond (E carried)
  89
  90/-- **The level gap across a bichromatic edge is exactly `±1`.** Adjacent cells of different charge
  91have one endpoint on the `x = 0` spine (charge `0`) and the other at `x = ±1` (charge `±1`), because
  92a unit step changes `x` by at most one and so cannot cross from `+1` to `-1`. Hence the charge
  93difference is `+1` or `-1`. -/
  94theorem level_diff (t : ℕ) (a b : Vtx t) (hadj : adj a.val b.val)
  95    (hpol : polarized t a ≠ polarized t b) :
  96    polarized t a - polarized t b = 1 ∨ polarized t a - polarized t b = -1 := by
  97  unfold adj at hadj
  98  simp only [polarized] at hpol ⊢
  99  split_ifs at hpol ⊢ <;> omega
 100
 101/-- The recognition cost the engine posts across one ordered adjacency of the polarized diamond:
 102`J` at the phi-rung gap between the two cells' charges. -/
 103noncomputable def edgeCost (t : ℕ) (p : Vtx t × Vtx t) : ℝ :=
 104  Jpow (polarized t p.1 - polarized t p.2)
 105
 106/-- A carried (monochromatic) edge costs exactly `0`: equal charges span no phi-rung. -/
 107theorem edgeCost_carried_zero (t : ℕ) (p : Vtx t × Vtx t) (hp : p ∈ carried t) :
 108    edgeCost t p = 0 := by
 109  rw [carried, Finset.mem_filter] at hp
 110  have hd : polarized t p.1 - polarized t p.2 = 0 := sub_eq_zero.mpr hp.2
 111  rw [edgeCost, hd, Jpow_zero]
 112
 113/-- A forced interface (bichromatic) edge costs exactly `J(φ)`: its charges differ by one phi-rung. -/
 114theorem edgeCost_interface (t : ℕ) (p : Vtx t × Vtx t) (hp : p ∈ B t) :
 115    edgeCost t p = Cost.Jcost Constants.phi := by
 116  simp only [B, Finset.mem_filter, Finset.mem_univ, true_and] at hp
 117  obtain ⟨hadj, hpol⟩ := hp
 118  simp only [edgeCost]
 119  exact Jpow_of_abs_one (level_diff t p.1 p.2 hadj hpol)
 120
 121/-- The total interface recognition cost: the sum of `edgeCost` over the bichromatic edges. -/
 122noncomputable def interfaceCost (t : ℕ) : ℝ := ∑ p ∈ B t, edgeCost t p
 123
 124/-- The total carried recognition cost: the sum of `edgeCost` over the monochromatic edges. -/
 125noncomputable def carriedCost (t : ℕ) : ℝ := ∑ p ∈ carried t, edgeCost t p
 126
 127/-- The total recognition cost of the field: the sum of `edgeCost` over every adjacency. -/
 128noncomputable def totalCost (t : ℕ) : ℝ := ∑ p ∈ E t, edgeCost t p
 129
 130/-- **Carried cost is exactly zero.** The whole bulk the engine carries coarse costs no recognition. -/
 131theorem carriedCost_eq_zero (t : ℕ) : carriedCost t = 0 := by
 132  simp only [carriedCost]
 133  apply Finset.sum_eq_zero
 134  intro p hp
 135  exact edgeCost_carried_zero t p hp
 136
 137/-- The interface cost is the interface edge count times `J(φ)`. -/
 138theorem interfaceCost_eq_card (t : ℕ) :
 139    interfaceCost t = (B t).card • Cost.Jcost Constants.phi := by
 140  simp only [interfaceCost]
 141  rw [Finset.sum_congr rfl (fun p hp => edgeCost_interface t p hp), Finset.sum_const]
 142
 143/-- **The total cost equals the interface cost**, because the carried bulk contributes nothing. -/
 144theorem totalCost_eq_interfaceCost (t : ℕ) : totalCost t = interfaceCost t := by
 145  have hsplit := Finset.sum_filter_add_sum_filter_not (E t)
 146    (fun p => polarized t p.1 ≠ polarized t p.2) (edgeCost t)
 147  have hBeq : (E t).filter (fun p => polarized t p.1 ≠ polarized t p.2) = B t := by
 148    rw [E, B, Finset.filter_filter]
 149  have hMeq : (E t).filter (fun p => ¬ (polarized t p.1 ≠ polarized t p.2)) = carried t := by
 150    rw [carried]
 151    apply Finset.filter_congr
 152    intro p _
 153    simp
 154  rw [hBeq, hMeq] at hsplit
 155  have hzero : ∑ p ∈ carried t, edgeCost t p = 0 := by
 156    apply Finset.sum_eq_zero
 157    intro p hp
 158    exact edgeCost_carried_zero t p hp
 159  rw [hzero, add_zero] at hsplit
 160  simp only [totalCost, interfaceCost]
 161  exact hsplit.symm
 162
 163/-- **The exact interface recognition cost is `(8t - 4) • J(φ)`** (`t ≥ 1`). -/
 164theorem interfaceCost_card (t : ℕ) (ht : 1 ≤ t) :
 165    interfaceCost t = (8 * t - 4) • Cost.Jcost Constants.phi := by
 166  rw [interfaceCost_eq_card, interface_card_eq t ht]
 167
 168/-- **The total recognition cost of the polarized diamond field is `(8t - 4) • J(φ)`** (`t ≥ 1`):
 169the carried bulk is free, and the whole cost sits on the `8t - 4` interface edges. -/
 170theorem totalCost_card (t : ℕ) (ht : 1 ≤ t) :
 171    totalCost t = (8 * t - 4) • Cost.Jcost Constants.phi := by
 172  rw [totalCost_eq_interfaceCost, interfaceCost_card t ht]
 173
 174/-- The total cost in real-multiplication form, `(8t - 4) * J(φ)` (`t ≥ 1`). -/
 175theorem totalCost_mul (t : ℕ) (ht : 1 ≤ t) :
 176    totalCost t = ((8 * t - 4 : ℕ) : ℝ) * Cost.Jcost Constants.phi := by
 177  rw [totalCost_card t ht, nsmul_eq_mul]
 178
 179/-- **The run-total recognition cost over a full forward run** from radius `1` to `T` is exactly
 180`8(T - 1) • J(φ)`. The cost the engine posts to grow the whole world is `Θ(T)`, strictly
 181sub-extensive against the brute-force volume-times-ticks `Θ(T³)`: the compute-watch law in cost
 182units, integrated over the run. -/
 183theorem runCost_growth (T : ℕ) (hT : 1 ≤ T) :
 184    totalCost T - totalCost 1 = ((8 * T - 8 : ℕ) : ℝ) * Cost.Jcost Constants.phi := by
 185  rw [totalCost_mul T hT, totalCost_mul 1 le_rfl]
 186  have h4 : 4 ≤ 8 * T := by omega
 187  have h8 : 8 ≤ 8 * T := by omega
 188  have e1 : ((8 * T - 4 : ℕ) : ℝ) = 8 * (T : ℝ) - 4 := by
 189    rw [Nat.cast_sub h4]; push_cast; ring
 190  have e2 : ((8 * 1 - 4 : ℕ) : ℝ) = 4 := by norm_num
 191  have e3 : ((8 * T - 8 : ℕ) : ℝ) = 8 * (T : ℝ) - 8 := by
 192    rw [Nat.cast_sub h8]; push_cast; ring
 193  rw [e1, e2, e3]; ring
 194
 195/-- **The per-cycle recognition cost increment** of advancing the diamond by one cadence cycle
 196(`t → t+1`) is the constant `8 • J(φ)` (`t ≥ 1`): the differential form of the compute-watch law.
 197In 2D the recognition-active interface is a `1`-dimensional curve whose length gains a constant `8`
 198ordered edges per shell, so the cost the engine posts each cycle is constant, `O(1)`, independent of
 199how large the world already is. -/
 200theorem costIncrement (t : ℕ) (ht : 1 ≤ t) :
 201    totalCost (t + 1) - totalCost t = 8 * Cost.Jcost Constants.phi := by
 202  rw [totalCost_mul (t + 1) (by omega), totalCost_mul t ht]
 203  have e1 : ((8 * (t + 1) - 4 : ℕ) : ℝ) = 8 * (t : ℝ) + 4 := by
 204    rw [Nat.cast_sub (by omega)]; push_cast; ring
 205  have e2 : ((8 * t - 4 : ℕ) : ℝ) = 8 * (t : ℝ) - 4 := by
 206    rw [Nat.cast_sub (by omega)]; push_cast; ring
 207  rw [e1, e2]; ring
 208
 209/-- **2D headline (Phase 55).** For a polarized diamond of radius `t ≥ 1`: the carried bulk costs
 210exactly zero recognition, the total cost is `(8t - 4) • J(φ)` (the interface count times the
 211one-rung cost), and `J(φ) > 0` is a genuine positive cost. "Carry the bulk free, pay only for the
 212interface," in exact cost units. -/
 213theorem t55_cost_ledger (t : ℕ) (ht : 1 ≤ t) :
 214    carriedCost t = 0
 215    ∧ totalCost t = (8 * t - 4) • Cost.Jcost Constants.phi
 216    ∧ totalCost t = interfaceCost t
 217    ∧ 0 < Cost.Jcost Constants.phi :=
 218  ⟨carriedCost_eq_zero t, totalCost_card t ht, totalCost_eq_interfaceCost t, Jcost_phi_pos⟩
 219
 220end Diamond
 221
 222/-! ## §2. The 3D octahedron cost ledger -/
 223
 224namespace Octahedron
 225
 226open InterfaceComponentBound.Octahedron
 227open PolarizedBirthDomains.Octahedron (polarized)
 228open PolarizedBirthInterface.Octahedron (B interface_card_eq)
 229open LatticeBallEdges.Octahedron (E carried)
 230
 231/-- **The level gap across a bichromatic edge is exactly `±1`** (3D). Same argument as 2D: a unit
 2326-neighbour step changes `x` by at most one, so the only way two adjacent cells differ in charge is
 233one on the spine `x = 0` and one at `x = ±1`. -/
 234theorem level_diff (t : ℕ) (a b : Vtx t) (hadj : adj a.val b.val)
 235    (hpol : polarized t a ≠ polarized t b) :
 236    polarized t a - polarized t b = 1 ∨ polarized t a - polarized t b = -1 := by
 237  unfold adj at hadj
 238  simp only [polarized] at hpol ⊢
 239  split_ifs at hpol ⊢ <;> omega
 240
 241/-- The recognition cost of one ordered adjacency of the polarized octahedron. -/
 242noncomputable def edgeCost (t : ℕ) (p : Vtx t × Vtx t) : ℝ :=
 243  Jpow (polarized t p.1 - polarized t p.2)
 244
 245/-- A carried (monochromatic) octahedron edge costs exactly `0`. -/
 246theorem edgeCost_carried_zero (t : ℕ) (p : Vtx t × Vtx t) (hp : p ∈ carried t) :
 247    edgeCost t p = 0 := by
 248  rw [carried, Finset.mem_filter] at hp
 249  have hd : polarized t p.1 - polarized t p.2 = 0 := sub_eq_zero.mpr hp.2
 250  rw [edgeCost, hd, Jpow_zero]
 251
 252/-- A forced interface (bichromatic) octahedron edge costs exactly `J(φ)`. -/
 253theorem edgeCost_interface (t : ℕ) (p : Vtx t × Vtx t) (hp : p ∈ B t) :
 254    edgeCost t p = Cost.Jcost Constants.phi := by
 255  simp only [B, Finset.mem_filter, Finset.mem_univ, true_and] at hp
 256  obtain ⟨hadj, hpol⟩ := hp
 257  simp only [edgeCost]
 258  exact Jpow_of_abs_one (level_diff t p.1 p.2 hadj hpol)
 259
 260/-- The total interface recognition cost (3D). -/
 261noncomputable def interfaceCost (t : ℕ) : ℝ := ∑ p ∈ B t, edgeCost t p
 262
 263/-- The total carried recognition cost (3D). -/
 264noncomputable def carriedCost (t : ℕ) : ℝ := ∑ p ∈ carried t, edgeCost t p
 265
 266/-- The total recognition cost of the octahedron field (3D). -/
 267noncomputable def totalCost (t : ℕ) : ℝ := ∑ p ∈ E t, edgeCost t p
 268
 269/-- **Carried cost is exactly zero** (3D). -/
 270theorem carriedCost_eq_zero (t : ℕ) : carriedCost t = 0 := by
 271  simp only [carriedCost]
 272  apply Finset.sum_eq_zero
 273  intro p hp
 274  exact edgeCost_carried_zero t p hp
 275
 276/-- The interface cost is the interface edge count times `J(φ)` (3D). -/
 277theorem interfaceCost_eq_card (t : ℕ) :
 278    interfaceCost t = (B t).card • Cost.Jcost Constants.phi := by
 279  simp only [interfaceCost]
 280  rw [Finset.sum_congr rfl (fun p hp => edgeCost_interface t p hp), Finset.sum_const]
 281
 282/-- **The total cost equals the interface cost** (3D). -/
 283theorem totalCost_eq_interfaceCost (t : ℕ) : totalCost t = interfaceCost t := by
 284  have hsplit := Finset.sum_filter_add_sum_filter_not (E t)
 285    (fun p => polarized t p.1 ≠ polarized t p.2) (edgeCost t)
 286  have hBeq : (E t).filter (fun p => polarized t p.1 ≠ polarized t p.2) = B t := by
 287    rw [E, B, Finset.filter_filter]
 288  have hMeq : (E t).filter (fun p => ¬ (polarized t p.1 ≠ polarized t p.2)) = carried t := by
 289    rw [carried]
 290    apply Finset.filter_congr
 291    intro p _
 292    simp
 293  rw [hBeq, hMeq] at hsplit
 294  have hzero : ∑ p ∈ carried t, edgeCost t p = 0 := by
 295    apply Finset.sum_eq_zero
 296    intro p hp
 297    exact edgeCost_carried_zero t p hp
 298  rw [hzero, add_zero] at hsplit
 299  simp only [totalCost, interfaceCost]
 300  exact hsplit.symm
 301
 302/-- **The exact interface recognition cost is `(8t² - 8t + 4) • J(φ)`** (`t ≥ 1`). -/
 303theorem interfaceCost_card (t : ℕ) (ht : 1 ≤ t) :
 304    interfaceCost t = (8 * t ^ 2 - 8 * t + 4) • Cost.Jcost Constants.phi := by
 305  rw [interfaceCost_eq_card, interface_card_eq t ht]
 306
 307/-- **The total recognition cost of the polarized octahedron field is `(8t² - 8t + 4) • J(φ)`**
 308(`t ≥ 1`): the 3D bulk is carried free, the cost sits on the `8t² - 8t + 4` interface edges. -/
 309theorem totalCost_card (t : ℕ) (ht : 1 ≤ t) :
 310    totalCost t = (8 * t ^ 2 - 8 * t + 4) • Cost.Jcost Constants.phi := by
 311  rw [totalCost_eq_interfaceCost, interfaceCost_card t ht]
 312
 313/-- The total cost in real-multiplication form, `(8t² - 8t + 4) * J(φ)` (`t ≥ 1`). -/
 314theorem totalCost_mul (t : ℕ) (ht : 1 ≤ t) :
 315    totalCost t = ((8 * t ^ 2 - 8 * t + 4 : ℕ) : ℝ) * Cost.Jcost Constants.phi := by
 316  rw [totalCost_card t ht, nsmul_eq_mul]
 317
 318/-- **The run-total recognition cost over a full forward run** from radius `1` to `T` is exactly
 319`8T(T - 1) • J(φ)`. In the dimension `T8` forces (D=3) the cost the engine posts to grow the whole
 320world is `Θ(T²)`, strictly sub-extensive against the brute-force `Θ(T⁴)`: the compute-watch law in
 321cost units, integrated over the run. -/
 322theorem runCost_growth (T : ℕ) (hT : 1 ≤ T) :
 323    totalCost T - totalCost 1 = ((8 * T ^ 2 - 8 * T : ℕ) : ℝ) * Cost.Jcost Constants.phi := by
 324  rw [totalCost_mul T hT, totalCost_mul 1 le_rfl]
 325  have hTT : 8 * T ≤ 8 * T ^ 2 := by nlinarith [hT]
 326  have e1 : ((8 * T ^ 2 - 8 * T + 4 : ℕ) : ℝ) = 8 * (T : ℝ) ^ 2 - 8 * (T : ℝ) + 4 := by
 327    rw [Nat.cast_add, Nat.cast_sub hTT]; push_cast; ring
 328  have e2 : ((8 * 1 ^ 2 - 8 * 1 + 4 : ℕ) : ℝ) = 4 := by norm_num
 329  have e3 : ((8 * T ^ 2 - 8 * T : ℕ) : ℝ) = 8 * (T : ℝ) ^ 2 - 8 * (T : ℝ) := by
 330    rw [Nat.cast_sub hTT]; push_cast; ring
 331  rw [e1, e2, e3]; ring
 332
 333/-- **The per-cycle recognition cost increment** of advancing the octahedron by one cadence cycle
 334(`t → t+1`) is `16t • J(φ)` (`t ≥ 1`): the differential form of the compute-watch law in the
 335dimension `T8` forces (D=3). The recognition-active interface is now a `2`-dimensional surface whose
 336area gains `16t` ordered edges per shell, so the cost the engine posts each cycle grows `Θ(t)`, the
 337honest 3D statement (not the constant `O(1)` of 2D), still strictly sub-extensive in the `Θ(t³)`
 338bulk. -/
 339theorem costIncrement (t : ℕ) (ht : 1 ≤ t) :
 340    totalCost (t + 1) - totalCost t = (16 * t : ℝ) * Cost.Jcost Constants.phi := by
 341  rw [totalCost_mul (t + 1) (by omega), totalCost_mul t ht]
 342  have hle1 : 8 * (t + 1) ≤ 8 * (t + 1) ^ 2 := by nlinarith [Nat.le_add_left 1 t]
 343  have hle2 : 8 * t ≤ 8 * t ^ 2 := by nlinarith [ht]
 344  have e1 : ((8 * (t + 1) ^ 2 - 8 * (t + 1) + 4 : ℕ) : ℝ)
 345      = 8 * (t : ℝ) ^ 2 + 8 * (t : ℝ) + 4 := by
 346    rw [Nat.cast_add, Nat.cast_sub hle1]; push_cast; ring
 347  have e2 : ((8 * t ^ 2 - 8 * t + 4 : ℕ) : ℝ) = 8 * (t : ℝ) ^ 2 - 8 * (t : ℝ) + 4 := by
 348    rw [Nat.cast_add, Nat.cast_sub hle2]; push_cast; ring
 349  rw [e1, e2]; ring
 350
 351/-- **3D headline (Phase 55).** For a polarized octahedron of radius `t ≥ 1`: the carried bulk costs
 352exactly zero recognition, the total cost is `(8t² - 8t + 4) • J(φ)` (the interface count times the
 353one-rung cost), and `J(φ) > 0`. The dimension `T8` forces (D=3) version of "carry the bulk free, pay
 354only for the interface." -/
 355theorem t55_cost_ledger (t : ℕ) (ht : 1 ≤ t) :
 356    carriedCost t = 0
 357    ∧ totalCost t = (8 * t ^ 2 - 8 * t + 4) • Cost.Jcost Constants.phi
 358    ∧ totalCost t = interfaceCost t
 359    ∧ 0 < Cost.Jcost Constants.phi :=
 360  ⟨carriedCost_eq_zero t, totalCost_card t ht, totalCost_eq_interfaceCost t, Jcost_phi_pos⟩
 361
 362end Octahedron
 363
 364end PolarizedBirthInterfaceCost
 365end Cosmology
 366end IndisputableMonolith
 367

source mirrored from github.com/jonwashburn/shape-of-logic