This development reduces the classification of low-value rational planes in five coordinates to a finite, exact certificate. The mathematical input is the four-speed gap, stated as an explicit hypothesis. The plane reduction, normalization, finite tests, and conversion to the five-speed spectrum are proved separately.
Statements
For a rational subspace E of ℚ⁵, let realSpan E be its span in ℝ⁵. Its value is the supremum, over all points x in that real span, of
CoordinateFree E means that each coordinate is nonzero at some point of E. Different coordinates may use different witnesses.
FourSpeedGap states the following. If four distinct positive integer speeds have value strictly below 2/9, their list is a permutation of a positive integer multiple of either
This predicate is a hypothesis of the plane theorem. It is neither an axiom declaration nor a result asserted by the development.
The main plane theorem, critical_product_of_low_value, has these inputs:
FourSpeedGap;- a coordinate-free rational subspace
Eofℚ⁵; dim E = 2;spaceValue (realSpan E) < 2/9.
It proves that a signed coordinate permutation sends E to
where e_j is the jth coordinate axis and the four coordinates of q outside j are exactly one of the two critical quadruples. Thus those four coordinates lie on one critical line, and coordinate j is free.
The theorem spectrum_in_critical_product supplies the next step. If an integer vector w belongs to such a product plane, every coordinate of w is nonzero, and its speed value is below 1/5, then
The speed list for an integer vector uses its distinct absolute coordinates. The proof also uses the full five-entry list, including repetitions, when applying the verified critical-family theorem.
Two integer directions in the plane
The starting point is coordinate coalescence. A coordinate-free rational plane contains an integer direction whose coordinates are all nonzero and whose absolute coordinates have at most four distinct values.
The strengthened coalescence theorem constructs a second such direction outside the first direction's line. It does this by choosing a hyperplane of the exact form
whose restriction to the plane remains coordinate-free and which avoids a prescribed vector with all coordinates nonzero. Clearing denominators gives the second integer vector. Both vectors lie exactly in the original rational plane, and they span it over ℚ.
The avoiding condition comes from fixing the functional used to normalize the coordinate forms: evaluate each form at the prescribed vector. Orient the selected forms so their evaluations are positive. Their signed sum then has positive evaluation, so its kernel avoids that vector. An ordering argument selects two forms whose signed sum is not parallel to any original coordinate form. Its kernel therefore remains coordinate-free. The resulting coordinate coefficients are exactly 1 and -1; normalization is used to select the pair, not to replace those coefficients by arbitrary weights.
The verified lower bound for at most three distinct positive speeds is 1/4. Each direction in a plane of value below 2/9 therefore has exactly four distinct absolute speeds. Applying FourSpeedGap makes both directions critical.
Complete normalization
There are eight positive canonical five-coordinate patterns:
| Critical support | Repeated coordinate | Pattern |
|---|---|---|
1,2,3,4 |
1 | 1,1,2,3,4 |
1,2,3,4 |
2 | 1,2,2,3,4 |
1,2,3,4 |
3 | 1,2,3,3,4 |
1,2,3,4 |
4 | 1,2,3,4,4 |
1,3,4,7 |
1 | 1,1,3,4,7 |
1,3,4,7 |
3 | 1,3,3,4,7 |
1,3,4,7 |
4 | 1,3,4,4,7 |
1,3,4,7 |
7 | 1,3,4,7,7 |
critical_direction_normalization proves that every critical integer direction with five nonzero coordinates is a positive integer dilation of a signed permutation of one of these patterns. The proof derives the single repeated coordinate from list length and support; it does not assume that repeated coordinates have been placed in a particular order.
Apply one signed coordinate permutation to make the first direction a positive canonical pattern, after removing its scalar. Transform the second direction by the same map. Its absolute speed support is unchanged. Normalize its positive scalar and its overall sign so its first coordinate is positive. Its four remaining signs are encoded by a number from 0 through 15.
normalized_low_plane proves that this procedure covers every plane in the theorem. The normalized pair spans the transformed plane exactly. Signed coordinate permutations preserve the actual supremum defining the plane value.
The finite test
Write the normalized directions as q and c. For every pair a < b and every sign ε = ±1, form the integer vector
This vector belongs to the plane and satisfies u_a = ε u_b. If every coordinate of u is nonzero, its distinct absolute speeds number at most four. The actual low-value plane hypothesis and FourSpeedGap therefore force its absolute support to be a positive multiple of one of the two critical quadruples.
The Boolean support test computes the smallest absolute coordinate m and checks equality of supports with m * [1,2,3,4] or m * [1,3,4,7]. It checks both inclusions. A line with a zero coordinate is skipped. admissibleTest_of_low_plane proves that every plane under the theorem's hypotheses passes this precise Boolean test.
The conclusion of each finite test is a free coordinate j for which:
- deleting coordinate
jfromqleaves one critical quadruple; - every two-coordinate minor outside
jvanishes:q_i * c_k = q_k * c_ifori,k ≠ j.
The certificate covers eight choices for q, eight patterns for c, all 120 coordinate permutations, and all 16 sign masks. This gives 122,880 indexed cases. The repeated coordinate makes each distinct second direction occur twice, so there are 61,440 distinct ordered vector pairs.
The 120-entry permutation table is proved equal to Mathlib's complete permutation enumeration. A separate coverage theorem applies the table to every coordinate permutation. Sign coverage is also proved in Lean. The finite check therefore has no unverified normalization or coverage step.
From accepted minors to the spectrum
Vanishing minors show that all four coordinates outside j are proportional to the corresponding coordinates of q. Hence the plane lies in span {q,e_j}. Both spaces have dimension two, so they are equal. canonical_product_structure proves this equality.
For an integer vector in this product plane, choose the coordinate of the critical quadruple that equals 1. Its integer value determines the common scalar on the other three critical coordinates. Taking absolute values gives a positive integer dilation of the entire critical quadruple in the five-entry speed list. The already verified critical-family classification then gives s/(5s+1) below 1/5. Signed permutation invariance transfers the result to the original integer vector.
Modules
| Module | Role |
|---|---|
LonelinessCoalescenceAvoiding |
Coalescence avoiding a prescribed vector |
LonelinessTwoDirections |
Concrete coordinate hyperplane with avoidance |
LonelinessPlaneDirections |
Two integer directions spanning the plane |
LonelinessPlaneGap |
Exact four-speed input and its use on both directions |
LonelinessCriticalNormalization |
Eight-pattern classification of directions |
LonelinessSignedCoordinates |
Signed permutation and value invariance |
LonelinessPlaneNormalization |
Complete normalized input domain |
LonelinessPlaneCertificateDefs |
Exact Boolean tests |
LonelinessPlanePermutationTable |
Complete permutation table |
LonelinessPlaneWitness |
Optional short rejection witnesses |
LonelinessPlaneWitnessSoundness |
Test soundness and table coverage |
LonelinessPlaneCertificate00 through 77 |
The 64 complete finite blocks |
LonelinessPlaneCertificate |
Universal finite-pair theorem |
LonelinessPlaneAdmissibility |
Actual plane hypothesis implies test acceptance |
LonelinessPlaneProduct |
Accepted test implies the product plane structure |
LonelinessPlaneClassificationCore |
Classification with exact finite coverage as a parameter |
LonelinessPlaneClassification |
Low-value plane classification |
LonelinessProductSpectrum |
Spectrum of integer directions in product planes |
generate_plane_certificates.py writes the 64 finite blocks. Its default output asks Lean to reduce each complete block in the kernel. The optional --rows form proposes shorter rejecting-line witnesses and checks each row separately. Correctness depends on the Lean proofs, including coverage, rather than on the Python generator.
The semantic composition is separated from the generated evidence. CandidatePairsChecked is the exact proposition that all normalized candidate pairs pass pairTest. The core classification takes this proposition as an explicit parameter. The final classification supplies it using the complete 64-block theorem. No assertion of finite coverage is introduced as an axiom or left implicit.
Verification status
All four support modules and all 64 finite blocks passed source checking, stock leanchecker replay, and the transitive logical-axiom whitelist in hosted run 44446557652 at commit bac947b. All eight row receipts were retrieved. The 64 source hashes, generator hash, and four support hashes match the local files.
The geometric normalization, admissibility, product semantics, and explicit-coverage proof cores also passed source checks, the 53-declaration transitive axiom audit, and separate stock kernel replays. The final aggregate and classification wrappers passed source checking, their seven-declaration audit, and separate kernel replays against the imported, hash-verified certificate artifacts. These checks allow only propext, Classical.choice, and Quot.sound.
FourSpeedGap remains an explicit mathematical input. The finite plane reduction is complete under that hypothesis. The full five-speed result also requires the four-speed classification and the remaining finite classification of the original speed vectors.