Dimensional necessity from response

Physical premise (IDWT): Physical spacetime has infinitely many ordinary spatial dimensions plus time. Matter occupies up to ten spatial dimensions; all remaining spatial directions are unbound for every particle. Depth counts a pattern's bound spatial involvement within that space. “Observed” and “unobserved” describe an observer's access, not different kinds of dimension. Older projective and nested-geometry calculations below retain their additional hypotheses; they do not establish a separate physical manifold or compactified direction. See spatial meaning and mathematical scope.

Summary: A fixed-preparation unobserved-response rank theorem gives a conditional six-dimensional lower bound, while exact five-dimensional dynamics and averaging counterexamples identify what the existing constructions do not require.

Scope and timing

These counterfactual models test which mathematical structures require six spatial dimensions. They do not reassign the mass-formula electron index. No particular sphere or projective spatial geometry is assumed for the competing models.

The corrected orbit has an exact five-dimensional realization

Write the tested model on $\mathbb R^d$ as $q,p\in\mathbb R^d$, $C\in\mathfrak{so}(d)$, with a fixed rank-three nuclear projector $P$. Its potential depends on $|q|$ and $|Pq|$; its other operations are matrix products, commutators, and the bivector $q\wedge p$.

Let $E:\mathbb R^5\to\mathbb R^6$ include the first five coordinates, keeping all three visible coordinates. Set

$$q_6=Eq_5,\qquad p_6=Ep_5,\qquad C_6=EC_5E^T.$$

Since $E^TE=I$, $P_6E=EP_5$, and radii are unchanged, substitution into every equation gives $F_6(\iota s)=\iota F_5(s)$, where $\iota$ is the corresponding state inclusion. No force drives the omitted coordinate or its mixed source entries. Differentiation gives $DF_6(\iota s)\iota=\iota DF_5(s)$ and the same relation for the tangent flow. This holds on both nuclear-profile branches and across their continuous join, away from the remaining full-radius singularity.

The corrected orbit already lies in this subspace. Fresh five-dimensional integration, retaining the third visible perturbation, gives:

At $h=0.0128$ the growing multiplier is still $1.138471302788$. At $h=0.100$, five dimensions give maximum modulus $8.54765518819$; six dimensions give $9.18464295801$ because an additional instability occurs in the third-unobserved-coordinate block. Thus identical invariant orbits do not imply identical responses to all six-dimensional disturbances. At the weak turning-point sample neither transverse block has a resolved exponential instability; neutral splitting does not prove nonlinear stability.

This is stronger than a similar display: it preserves the interaction and the complete five-dimensional tangent dynamics. It does not reproduce an arbitrary six-dimensional preparation, every molecular configuration, or the CP3-constrained source model. The source here is a general skew matrix with a rank-four prepared spectrum, not an invertible complex structure.

The active trajectory plane has dimension four and intersects the fixed visible three-plane in two dimensions. Their span has dimension five. Deleting another coordinate cannot preserve this particular full trajectory and that visible plane isometrically. This is a minimum for this embedding, not a theorem excluding every four-dimensional model of the visible curve.

A sharp local lower-bound theorem

Let $x\in\mathbb R^3$ be visible displacements and $y\in\mathbb R^k$ the only eliminated spatial displacements. At a fixed preparation suppose

$$E_2(x,y)=\tfrac12x^TAx+x^TBy+\tfrac12y^TKy,\qquad K>0.$$

Eliminating $y$ gives $y_*=-K^{-1}B^Tx$ and

$$E_{\rm eff,2}=\tfrac12x^T(A-R)x, \qquad R=BK^{-1}B^T.$$

The factorization $R=(BK^{-1/2})(BK^{-1/2})^T$ proves

$$R\geq0,\qquad\operatorname{rank}R=\operatorname{rank}B\leq k.$$

Consequently a rank-three unobserved-mediated response at one fixed preparation requires $k\geq3$, hence $d=3+k\geq6$. The bound is sharp: any positive semidefinite rank-$r$ response has an $r$-coordinate realization by taking its positive-eigenvalue square root and $K=I_r$. To infer exactly six, rather than at least six, additionally require a minimal realization with no unobserved spectator coordinates.

This is also a local theorem for nonlinear energies: if the unobserved equilibrium is unique nearby and its Hessian $E_{yy}=K$ is positive definite, implicit differentiation gives the same Schur complement for the effective Hessian at that preparation. No globally quadratic energy is needed. Singular unobserved Hessians require a separate treatment.

For a stationary linear response with fixed coordinate coupling $B$, unobserved inertia $M$, and damping $\Gamma$, the frequency-dependent mediated matrix is

$$R(\omega)=B\left(K-\omega^2M-i\omega\Gamma\right)^{-1}B^T.$$

Where the inverse exists, its complex rank is still at most $k$; positive semidefiniteness need not survive at nonzero frequency. Direct feedthrough must again be separated. This formula does not cover a periodically rotating coupling $B(t)$: sidebands and period averaging can mix different directions, as the rank-one averaging counterexample demonstrates. An orbit's cycle-averaged susceptibility cannot simply be treated as the stationary matrix above. A response-state count also does not become a spatial count without identifying the eliminated variables as physical coordinates.

The sign matters: minimizing a stable unobserved coordinate reduces the visible energy Hessian by $R$. Positive definiteness of the total visible Hessian neither implies nor requires rank-three $R$.

A symmetry version of the theorem

Suppose the fixed preparation is rotationally invariant, the unobserved linear displacements carry a representation $U(R_0)$, and the mixed map obeys $BU(R_0)=R_0B$ for every visible rotation $R_0\in SO(3)$. Then $\operatorname{im}B$ is an invariant subspace of the ordinary visible vector representation. Every nonzero vector has a rotation orbit spanning $\mathbb R^3$, so a nonzero $B$ is surjective and $k\geq3$.

This does not assume three unobserved directions at the start. It assumes that a nonzero mixed linear coupling exists at a symmetry-preserving state. Covariance of a family of oriented states is weaker and does not give this result. At $k=3$, orthogonal unobserved representations make the nonzero coupling a scalar times an orthogonal identification: $BB^T$ commutes with every visible rotation and is therefore $b^2I_3$.

Exact counterexamples to weaker proposed tests

1. Direct visible response. For $k=1$ or $2$, choose $B=(e_1,\ldots,e_k)$ and positive diagonal $K$. The unobserved response has rank $k$, but adding the direct matrix $D=I_3-R$ gives total response $D+R=I_3$. Total rank is not unobserved-channel rank.

2. Averaging. With one unobserved coordinate and $K=1$, choose the prescribed rotating coupling

   $$b(t)=\left(\sqrt{2/3}\cos t,\sqrt{2/3}\sin t,1/\sqrt3\right)^T.$$

   Every $b(t)b(t)^T$ has rank one, but its exact period average is $I_3/3$. This example includes a prescribed changing orientation; it does not claim that its orientation clock is generated for free by that one unobserved displacement. It disproves the inference from averaged rank to instantaneous unobserved rank.

3. Additional representation variables in a comparison model. The same rank proof counts whatever variables were eliminated. Calling polarization or source-matrix entries spatial directions would change the theorem's physical meaning.

4. Additional directions of the same physical manifold. Appending uncoupled coordinates preserves a model's responses on its original invariant subspace. A lower-bound argument alone cannot exclude such higher-dimensional extensions.

These are limits of the inference, not alternative IDWT sectors. In IDWT all physical directions already belong to the one infinite manifold, and all particle sectors are nested submanifolds of it. Source-matrix components are representations of fields, not independently postulated spatial dimensions.

The current cycle-averaged molecular Hessians do not yet isolate $R$ under these conditions. Three molecular axes and petal counts therefore do not supply the missing rank-three premise.

Verification and next discriminating target

dimensional_necessity_checks.py checks exact rank and averaging identities, equivariance examples, random full-equation and derivative intertwinings, and optional fresh five-dimensional orbit integrations. The first full run completed 171 runtime checks; its evidence is dimensional-necessity.json.

The productive next target is a physically derived, preparation-resolved response with its direct contribution separated, not another unperturbed petal image. A proof that its unobserved-mediated part is necessarily nonsingular would activate the lower-bound theorem. That nonsingularity is not yet proved.

Work-block verification record

The final regression run completed 63 finite diagnostic scripts, with no failures. It excluded the two original expensive shooting searches explicitly; it included the archived-state rejection audit and the fresh ten-candidate cross-method verification. The new corrector and full-orbit stability tools were exercised separately with the saved arguments and states described in the orbit note. The complete suite output is in diagnostic-suite-final.json.

Three deliberate in-memory mutations were rejected: adding $10^{-4}$ to the five-dimensional velocity field failed equation intertwining; doubling one spinor-bivector operator failed the universal moment-norm polynomial; replacing reciprocal matrix evolution by identity failed the independent nonlinear response comparison. No mutated source was saved. The four new dimensional diagnostic scripts also passed under Python -O: their explicit runtime checks are not disabled with ordinary assertions.

These finite numerical checks do not upgrade numerical recurrence to an interval existence proof or identify the physical electron from the conditional theorems.

Related

• Five-dimensional CP3 selection and spectral discrimination — stronger operator-class no-go theorem and a CP3 counterexample developed during the same work block.
• Accurate cylindrical correction — supplies the corrected orbit and full-coordinate perturbation calculations.