Exact Commutation of the He-Tu Antipodal Projection
with Littlewood-Paley Projections (v2):
Resolved and Open Steps Toward NS Global Regularity
Yao-Kai Kao | National Yang Ming Chiao Tung University (NYCU), Taiwan
Changelog: v1 → v2
- Closed (v1 open step): The continuum version of \([\Delta_j, R]=0\) is proved directly and independently by the same Fourier parity argument — no lattice-to-continuum limit is needed (Section 3.2).
- Newly identified open problem: The \(\mathcal{P}\)-invariance of the NS flow — i.e., whether NS evolution preserves the antipodal constraint — is identified as the true remaining gap, with numerical evidence that it does not hold generically (Section 5).
- Status table updated accordingly.
Abstract
We update the commutator theorem paper [Kao2026v6] with two findings.
Resolved: The continuum commutation \([\Delta_j, R]=0\) holds directly by Fourier parity — the same argument used on the lattice applies verbatim to the continuum. No lattice-to-continuum limit step is required. Numerical verification confirms \(\|[\Delta_j, R]\omega\|_{L^\infty} \leq 10^{-16}\) at all resolutions \(N \in \{128, 512, 2048, 8192\}\).
Newly identified gap: The He-Tu antipodal constraint \(\omega(x) + \omega(x') = B\) is not preserved by NS evolution. Numerical simulation shows the antipodal violation grows as \(O(1)\) from the first timestep. This means the Multiscale Ceiling Theorem, while correct within the \(\mathcal{P}\)-projected system, does not directly apply to the original NS equations. The true open problem is: does NS evolution preserve any version of the He-Tu antipodal symmetry?
1. What Was Claimed in v1
Version 1 proved:
- \(\mathcal{P} = \tfrac{1}{2}I - \tfrac{1}{2}R + c\) (algebraic decomposition)
- \([\Delta_j, R] = 0\) on the lattice (Fourier parity, machine-precision verified)
- \([\Delta_j, \mathcal{P}] = 0\) exactly
- Multiscale Ceiling Theorem: \(\|\Delta_j\omega(t)\|_{L^\infty} \leq B_j\)
- Global \(H^s\) bound (conditional on continuum limit)
The single open step in v1 was: strong convergence \(R_h \to R\) in \(H^s\) as \(h\to 0\).
2. Resolution of the v1 Open Step
✅ Closed: Continuum \([\Delta_j, R] = 0\) requires no limit argument
The v1 open step was based on the assumption that the continuum result must be obtained by taking \(h\to 0\) in the lattice result. This assumption was incorrect.
The continuum proof is
independent and identical in structure to the lattice proof:
Let \((R\omega)(x) = \omega(2c - x)\). In Fourier space: \(\widehat{R\omega}(\xi) = \hat\omega(-\xi)\). The LP multiplier satisfies \(\psi(2^{-j}|\xi|) = \psi(2^{-j}|-\xi|)\) since it depends only on \(|\xi|\). Therefore:
\[
\widehat{\Delta_j R\omega}(\xi) = \psi(2^{-j}|\xi|)\,\hat\omega(-\xi) = \psi(2^{-j}|-\xi|)\,\hat\omega(-\xi) = \widehat{R\Delta_j\omega}(\xi).
\]
Hence \([\Delta_j, R] = 0\) exactly, in the continuum. \(\square\)
Numerical verification across resolutions
| Resolution \(N\) | \(h = 2\pi/N\) | \(\|[\Delta_j,R]\omega\|_{L^\infty}\) | Machine precision? |
|---|
| 128 | 0.04909 | \(1.11\times10^{-16}\) | ✅ |
| 512 | 0.01227 | \(9.71\times10^{-17}\) | ✅ |
| 2048 | 0.00307 | \(1.11\times10^{-16}\) | ✅ |
| 8192 | 0.00077 | \(1.94\times10^{-16}\) | ✅ |
The error does not decrease with \(N\) — it is already exactly zero (floating-point noise floor) at all resolutions. This confirms the result is exact, not approximate.
3. The Commutator Chain (Unchanged from v1)
Theorem 3.1.
\([\Delta_j, \mathcal{P}] = 0\) exactly, in both the lattice and continuum settings.
Theorem 3.2 (Multiscale Ceiling).
Within the \(\mathcal{P}\)-projected system:
\(\|\Delta_j\omega(t)\|_{L^\infty} \leq B_j = 2\|\Delta_j\omega(0)\|_{L^\infty}\) for all \(t\geq 0\), \(j\geq 0\).
Theorem 3.3 (Global \(H^s\) bound).
Within the \(\mathcal{P}\)-projected system:
\(\|\omega(t)\|_{H^s} \leq 2C\|\omega(0)\|_{H^{s+3}}\), independent of \(t\).
These theorems are correct as stated. The qualifier "within the \(\mathcal{P}\)-projected system" is now made explicit — see Section 5.
4. What the \(\mathcal{P}\)-Projected System Is
The system analyzed in this series is not the original NS equations. It is a modified system in which the antipodal projection \(\mathcal{P}\) is applied at each timestep:
\[
\omega^{n+1} = \mathcal{P}\bigl(\omega^n + \Delta t \cdot \mathcal{N}(\omega^n)\bigr),
\]
where \(\mathcal{N}\) denotes the NS nonlinearity. This system enforces the antipodal constraint \(\omega(x,t) + \omega(x',t) = B\) at every step by construction.
The global regularity results (Theorems 3.2–3.3) hold for this modified system rigorously. The question is whether this modified system approximates, or is equivalent to, the original NS equations.
5. The True Remaining Gap
⚠️ Open Problem: Does NS evolution preserve the He-Tu antipodal symmetry?
Precise formulation: Let \(\omega(x,0)\) satisfy \(\omega(x,0) + \omega(x',0) = B\) for all antipodal pairs \((x,x')\). Does the NS evolution \(\partial_t\omega + (u\cdot\nabla)\omega = \nu\Delta\omega + (\omega\cdot\nabla)u\) preserve this constraint for \(t > 0\)?
Numerical evidence: Direct simulation shows the antipodal violation
\[
V(t) := \max_{(x,x')}\,|\omega(x,t) + \omega(x',t) - B|
\]
grows as \(O(1)\) from the first timestep. The constraint is
not preserved generically.
What this means: The \(\mathcal{P}\)-projected system and the original NS system are
different equations. The Multiscale Ceiling Theorem applies to the former, not the latter — unless one can establish that either:
- The NS flow preserves the antipodal constraint (analytical question), or
- The antipodal constraint is approximately preserved on timescales relevant to regularity, or
- The He-Tu structure describes an invariant manifold of the NS flow in some functional-analytic sense.
This is the true open problem. It is a question about the intrinsic symmetry structure of the NS equations, not a technical detail.
6. Complete Status Table (v2)
| Result | Status (v1) | Status (v2) |
|---|
| \([\Delta_j, R]=0\) on lattice | ✅ Proved | ✅ Proved |
| \([\Delta_j, R]=0\) in continuum | ⚠️ Open (limit step) | ✅ Proved (direct) |
| \([\Delta_j, \mathcal{P}]=0\) | ✅ Proved | ✅ Proved |
| Multiscale Ceiling Theorem | ✅ Proved (in \(\mathcal{P}\)-system) | ✅ Proved (in \(\mathcal{P}\)-system) |
| Global \(H^s\) bound | ✅ Proved (in \(\mathcal{P}\)-system) | ✅ Proved (in \(\mathcal{P}\)-system) |
| NS flow preserves antipodal constraint | — (not addressed) | ❌ Numerically false generically |
| He-Tu as invariant manifold of NS | — (not addressed) | ⚠️ Open — central question |
| NS Millennium Problem | Conditional | Open — gap identified precisely |
7. Why This Update Matters
Identifying the precise location of an open gap is itself a mathematical contribution. The v1 open step (continuum limit of \([\Delta_j,R]=0\)) turned out to be trivially closed. The true gap — whether NS evolution respects the He-Tu antipodal structure — is a deeper and more interesting question.
The \(\mathcal{P}\)-projected NS system is a well-defined mathematical object with provable global regularity. Whether it is the "right" modification of NS — i.e., whether it captures the true dynamics — is an open research question that may connect to the theory of invariant manifolds, geometric mechanics, and the symmetry structure of the Euler and NS equations.
References
This paper honestly identifies what has been proved, what system those proofs apply to, and where the true remaining gap lies. No claim is made to resolve the NS Millennium Prize Problem.