Skip to main content

Composite Systems, Correlations and Gap

Two simultaneously represented holons use a tensor-product state space. This gives exact tools for marginals and correlations. It does not by itself identify correlation with empathy or seven internal axes with spacetime. This revision of 2026-10-03 replaces the former ambiguous 14-dimensional block model and distinguishes proven matrix statements from proposed physical bridges.

1. Composite coherence matrix​

For HA=HB=C7\mathcal H_A=\mathcal H_B=\mathbb C^7,

ΓAB∈D(HA⊗HB)=D(C49),ΓA=Tr⁡BΓAB,ΓB=Tr⁡AΓAB.\Gamma_{AB}\in\mathcal D(\mathcal H_A\otimes\mathcal H_B)=\mathcal D(\mathbb C^{49}),\qquad \Gamma_A=\operatorname{Tr}_B\Gamma_{AB},\quad\Gamma_B=\operatorname{Tr}_A\Gamma_{AB}.

The marginals are positive and trace one [T]. Matrix entries have two subsystem indices on each side:

γij,kl=⟨iAjB∣ΓAB∣kAlB⟩.\gamma_{ij,kl}=\langle i_Aj_B|\Gamma_{AB}|k_Al_B\rangle.

A correct block notation is a 7×77\times7 array of 7×77\times7 blocks BikB_{ik} acting on BB; then (ΓA)ik=Tr⁡Bik(\Gamma_A)_{ik}=\operatorname{Tr}B_{ik} and ΓB=∑iBii\Gamma_B=\sum_iB_{ii}. There is no canonical single 7×77\times7 off-diagonal block containing all inter-system correlations.

Tensor product versus direct sum​

The direct sum HA⊕HB\mathcal H_A\oplus\mathcal H_B represents two alternative sectors of one system. A block-diagonal density matrix is pΓA⊕(1−p)ΓBp\Gamma_A\oplus(1-p)\Gamma_B, not the trace-two matrix ΓA⊕ΓB\Gamma_A\oplus\Gamma_B. Coherences between alternative sectors may be defined, but they are not bipartite entanglement relative to A⊗BA\otimes B. The tensor product represents the two subsystems simultaneously and admits both classical correlations and entanglement.

A conditional compression is not a marginal map​

Choose ground vectors ∣0A⟩,∣0B⟩|0_A\rangle,|0_B\rangle inside the seven-dimensional factors. The subspaces HA⊗∣0B⟩\mathcal H_A\otimes|0_B\rangle and ∣0A⟩⊗HB|0_A\rangle\otimes\mathcal H_B intersect in the vacuum line. Their sum has dimension 7+7−1=137+7-1=13; removing the vacuum leaves a 12-dimensional single-excitation space. It is not a 14-dimensional direct sum.

For a projector Π\Pi onto a selected sector, Γ↦ΠΓΠ\Gamma\mapsto\Pi\Gamma\Pi is positive but generally trace-decreasing. The normalised postselected state ΠΓΠ/Tr⁡(ΠΓ)\Pi\Gamma\Pi/\operatorname{Tr}(\Pi\Gamma) is conditional and nonlinear, and changes generic marginals. For example, ∣1A1B⟩⟨1A1B∣|1_A1_B\rangle\langle1_A1_B| has nonzero marginals and zero compression to the ground/single-excitation sector. A 14-dimensional single-excitation space requires seven excitations plus an extra ground state in each factor, giving 8⊗88\otimes8 before compression. The former claim that a 14-dimensional projection of the 49-dimensional state preserves its marginals is retracted [✗].

Product, separable and entangled states​

TypeExact criterion
UncorrelatedΓAB=ΓA⊗ΓB\Gamma_{AB}=\Gamma_A\otimes\Gamma_B
SeparableΓAB=∑xpxρAx⊗ρBx\Gamma_{AB}=\sum_xp_x\rho_A^x\otimes\rho_B^x for a probability distribution
EntangledNo such separable decomposition exists

Products are separable; separable states can be correlated. The example

ΓAB=12∣00⟩⟨00∣+12∣11⟩⟨11∣\Gamma_{AB}=\tfrac12|00\rangle\langle00|+\tfrac12|11\rangle\langle11|

is separable but not the product of its marginals. Thus “no entanglement iff product” is retracted [✗].

Complete correlation certificate [T]​

Let T0=I/7,T1,…,T48T_0=I/\sqrt7,T_1,\ldots,T_{48} be an orthonormal Hermitian operator basis with TaT_a traceless for a>0a>0. The connected coefficients

Cab=Tr⁡[ΓAB(Ta⊗Tb)]−Tr⁡(ΓATa)Tr⁡(ΓBTb),a,b>0C_{ab}=\operatorname{Tr}[\Gamma_{AB}(T_a\otimes T_b)]- \operatorname{Tr}(\Gamma_AT_a)\operatorname{Tr}(\Gamma_BT_b),\quad a,b>0

vanish simultaneously exactly when the state is a product. Proof: expand ΓAB−ΓA⊗ΓB\Gamma_{AB}-\Gamma_A\otimes\Gamma_B in the tensor operator basis. All coefficients with an identity factor vanish by the marginal identities; the remaining coefficients are precisely CabC_{ab}. The full state has 492−1=240049^2-1=2400 real parameters: 96 local parameters and 2304 connected coefficients. A selected set of 49 channels cannot encode arbitrary correlations.

2. A typed inter-system Gap observable​

A phase profile needs a specified complex observable. One possible definition [D] selects seven bounded operators FiA,FjBF_i^A,F_j^B and uses the connected array

Kij=Tr⁡[ΓAB(FiA⊗(FjB)†)]−Tr⁡(ΓAFiA)Tr⁡(ΓBFjB)‾.K_{ij}=\operatorname{Tr}[\Gamma_{AB}(F_i^A\otimes(F_j^B)^\dagger)]- \operatorname{Tr}(\Gamma_AF_i^A)\overline{\operatorname{Tr}(\Gamma_BF_j^B)}.

Then GapABF(i,j)=∣Im⁡Kij∣/∣Kij∣\mathrm{Gap}_{AB}^{F}(i,j)=|\operatorname{Im}K_{ij}|/|K_{ij}| for Kij≠0K_{ij}\ne0, with a declared zero convention and detection floor. The operator family, frame, amplitudes and support mask are part of the definition. This is a 49-coordinate selected observable, not the whole composite matrix. If all FiF_i are Hermitian, KK is real and this phase detector is identically zero; nontrivial phase detection requires an appropriate complex operator family.

For products K=0K=0. A nonzero connected coefficient certifies some correlation; it does not certify entanglement. Conversely, a vanishing selected array need not imply a product unless it spans the full connected operator space. A small phase alone does not establish access to another system's experience.

Rank and representation types​

The internal Im⁡Γ\operatorname{Im}\Gamma is real skew-symmetric and has rank 0,2,4,60,2,4,6; at most three skew blocks is not rank at most three. Im⁡K\operatorname{Im}K is generally a real 7×77\times7 array of rank at most seven. Neither rank is an empathy or isolation theorem.

If the chosen operator family is supplied with a standard-vector transformation rule, K↦RAKRBTK\mapsto R_AKR_B^T, then under independent G2×G2G_2\times G_2 it carries the external product (7,7)(7,7) of dimension 49. Under the diagonal subgroup the decomposition is

7⊗7=1⊕27⏟Sym⁡2(7)⊕7⊕14⏟Λ2(7).7\otimes7=\underbrace{1\oplus27}_{\operatorname{Sym}^2(7)}\oplus\underbrace{7\oplus14}_{\Lambda^2(7)}.

The 14 is the adjoint representation of G2G_2. This representation statement does not identify a gauge field or spacetime tensor. An invariant trace is available under the diagonal action, not under arbitrary independent rotations. Primary algebraic context: Baez, The Octonions.

3. Holevo bound: precise information task​

Theorem [T]. For a declared ensemble {px,ρx}\{p_x,\rho_x\} on a dd-dimensional register and a measurement outcome YY,

I(X:Y)≤χ:=S(ρˉ)−∑xpxS(ρx)≤S(ρˉ)≤log⁡d,ρˉ=∑xpxρx.I(X:Y)\le\chi:=S(\bar\rho)-\sum_xp_xS(\rho_x)\le S(\bar\rho)\le\log d,\qquad\bar\rho=\sum_xp_x\rho_x.

It bounds information about the encoded classical label in the stated preparation/measurement task. The logarithm convention fixes the units. Source: Watrous, Theory of Quantum Information, §5.3.3.

The bound does not prohibit tomography using many identically prepared systems, establish permanent inaccessibility of an inner aspect, or prove that empathy bypasses measurement limitations. A restriction on reconstructing another system requires the ensemble, number of uses, available observables and operational notion of understanding to be specified. The former “complete understanding only through a shared inner map” corollary is retracted [✗] as an inference from Holevo's theorem; it may be proposed philosophically [I].

4. Correlation measure and entanglement​

The former quantity EGap\mathcal E_{\mathrm{Gap}} is exactly quantum mutual information:

I(A:B)=S(ΓA)+S(ΓB)−S(ΓAB)=D(ΓAB∥ΓA⊗ΓB).I(A:B)=S(\Gamma_A)+S(\Gamma_B)-S(\Gamma_{AB}) =D(\Gamma_{AB}\|\Gamma_A\otimes\Gamma_B).

Hence I≥0I\ge0, with equality exactly on products [T]. It measures total correlation, including classical correlation. The separable example above has I=log⁡2I=\log2; the pure Bell state has I=2log⁡2I=2\log2. In general

I(A:B)≤2min⁡{S(ΓA),S(ΓB)},I(A:B)\le2\min\{S(\Gamma_A),S(\Gamma_B)\},

by the Araki–Lieb inequality. The old normalisation min⁡(SA,SB)\min(S_A,S_B) is not a general maximum. Zero marginal entropy also makes ratio-based normalisations singular and needs a declared convention.

A phase Gap and mutual information are distinct statistics. Real product states and real entangled states can both have zero pairwise phase. The former “mutual understanding” inequalities relating nonzero lower Gap to small II, without a stated operator/encoding bridge, are retracted [✗]. A genuine relation needs declared observables, an ensemble/noise model and a proof, not a change of terminology.

5. Collective transitions​

A collective state of NN holons lives on (C7)⊗N(\mathbb C^7)^{\otimes N}. Correlation and collective viability require a joint model and an aggregation map. Passing individual gates does not by itself produce a new subject; see Collective consciousness.

The former universal formula

Tccoll=Tcindiv(1+(N−1)σˉ2/μ2)T_c^{\mathrm{coll}}=T_c^{\mathrm{indiv}}\left(1+(N-1)\bar\sigma^2/\mu^2\right)

is retracted [✗] as a theorem about arbitrary interacting holons. It needs a specified statistical Hamiltonian, coupling sign, ensemble, order parameter and approximation. Also Tr⁡(K2)\operatorname{Tr}(K^2) is not a nonnegative strength for an arbitrary real array; a positive strength is ∥K∥F2=Tr⁡(KTK)\|K\|_F^2=\operatorname{Tr}(K^TK). Interaction can stabilise or destabilise an ordered phase depending on the model.

6. Empathy as an operational proposal​

An amplitude-supported small phase can be stipulated as a phase-alignment test [D]:

∣Kij∣>δ,GapABF(i,j)<ϵ.|K_{ij}|>\delta,\qquad\mathrm{Gap}_{AB}^F(i,j)<\epsilon.

Calling this empathy requires a separate behavioural/phenomenal bridge [H]/[I]. Nonzero mutual information does not mean “the systems cannot be separable”. Classical communication and correlated learning can support prediction of another agent without quantum entanglement.

A stronger measurable proposal tests an agent's predictions of a partner's independently observed responses across held-out interventions, compares to a marginal-only baseline, and controls shared inputs. Tolerances, readouts and the resource model are declared. The former universal four-condition necessity theorem for empathy is retracted [✗]; its phase and viability conditions may be tested as hypotheses, not assumed to follow from quantum information.

7. Holonomy and the arrow of time​

A nonmaximally mixed state does not determine a connection or imply nontrivial holonomy. A nontrivial holonomy need not differ from its inverse (an element of order two is a counterexample), and reversible parallel transport alone does not establish an entropy-production arrow. Cubic terms, associators and PT breaking need a specified action and transformation law. The former unconditional four-step phenomenology → holonomy → time arrow → octonions chain is retracted [✗] in this formulation. Its proposed bridges are tracked in Octonionic derivation and Lindblad operators.

8. RG flow: an action and scheme are required​

A beta function belongs to a specified field theory: fields, kinetic term, dimension, renormalisation scheme and dimensionless coupling definitions must be supplied. Seven labels or 21 matrix pairs do not specify one-loop combinatorial coefficients.

The former displayed beta functions and Wilson–Fisher conclusion are retracted [✗] as established UHM results. They are internally inconsistent: setting λ3=0\lambda_3=0 in the displayed βλ4=63λ42/(4π2)\beta_{\lambda_4}=63\lambda_4^2/(4\pi^2) leaves only λ4=0\lambda_4=0 as a zero, while the proposed λ4∗=4π2/63\lambda_4^*=4\pi^2/63 gives a nonzero beta function. A Wilson–Fisher calculation in an epsilon expansion would require a dimension term and a declared action. No microscopic/macroscopic time-arrow conclusion follows from these unverified coefficients.

9. Algebraic decomposition and spacetime​

Fixing a unit imaginary octonion gives a stabiliser SU(3)⊂G2SU(3)\subset G_2. As real representations,

R7≅R⊕C3;\mathbb R^7\cong\mathbb R\oplus\mathbb C^3;

its complexification restricts as 7C=1⊕3⊕3ˉ7_\mathbb C=1\oplus3\oplus\bar3. The complex three-space has six real dimensions. The metric dτ2−∑a=13∣dza∣2d\tau^2-\sum_{a=1}^3|dz_a|^2 therefore has signature (1,6)(1,6), not (1,3)(1,3); a Kähler structure does not halve the real dimension.

A real spatial three-plane and a Lorentzian metric require extra data and covariance conditions. Current physical bridges, including the tangent-spinor hypothesis (P), are stated in Physics correspondence. The former unconditional 3+1 derivation from the SU(3)SU(3) split is retracted [✗].

A torsion-free G2G_2 structure on an actual seven-dimensional manifold is geometric data, not supplied by a finite density matrix. Holonomy contained in G2G_2 needs the appropriate differential conditions; equality to G2G_2 is stronger. Identifying the internal axes with compact extra dimensions is a proposed physical bridge [Pr]. The former entrywise metric ansatz gij∝∣γij∣2+Gap(i,j)2g_{ij}\propto|\gamma_{ij}|^2+\mathrm{Gap}(i,j)^2 need not even be positive definite: for a state with diagonal 1/71/7 and a small purely imaginary γ12\gamma_{12}, that ansatz gives an off-diagonal entry near one with diagonal entries 1/491/49, yielding a negative principal determinant.

No measured cosmological constant is derived here from O-channel opacity. A quantitative relation requires an action, units, normalisation and a renormalisation prescription. Current proposals: Emergent geometry and Cosmological constant.

10. Curvature needs a geometric bridge​

Riemann curvature belongs to a metric connection on a manifold. A scalar phase array is not such a connection. A proposed projection to spacetime must specify the base manifold, bundle, connection, metric, pullback/projection maps, tensor symmetries and dimensional factors. Summing internal labels alone is not a covariant tensor construction.

The former curvature equalities and unconditional Einstein-equation derivation from Gap are retracted [✗] in this form. Zero scalar curvature does not imply a flat Riemann tensor. A spectral-action or variational derivation needs the declared geometric/spectral data and its approximations. See Einstein equations and Physical correspondence.

11. Topology and an actual energy barrier​

Correct scope of T-69​

Group-theoretic theorem [T]. For the compact simply connected G2G_2 and its maximal torus T2T^2,

π2(G2/T2)≅Z2.\pi_2(G_2/T^2)\cong\mathbb Z^2.

The long exact sequence of T2→G2→G2/T2T^2\to G_2\to G_2/T^2 uses π2(G2)=0\pi_2(G_2)=0 and π1(G2)=0\pi_1(G_2)=0. The former can also be obtained through SU(2)→SU(3)→S5SU(2)\to SU(3)\to S^5 and SU(3)→G2→S6SU(3)\to G_2\to S^6. This classifies homotopy classes of maps from a sphere into the orbit, not individual finite-matrix configurations. A soliton interpretation requires a spatial field and boundary conditions. The stabiliser T2T^2 itself is an extra vacuum-model assumption (SV), not a consequence of a frame phase profile.

Energy-separation theorem [T]. Let V:K→RV:K\to\mathbb R be continuous on a nonempty compact configuration space KK, and let a nonempty compact forbidden set F⊂KF\subset K be disjoint from the set of global minima M0M_0. Then

Δ=min⁡FV−min⁡KV>0.\Delta=\min_FV-\min_KV>0.

Every path from a minimum to FF reaches energy at least min⁡KV+Δ\min_KV+\Delta. Proof: the minimum on FF is attained; if it equalled the global minimum, FF would intersect M0M_0. This theorem requires no winding-number claim. It gives no universal numeric value of Δ\Delta.

A positive Hessian gives local stability, not a global barrier 6μ26\mu^2 or 9μ29\mu^2. A quantitative bound needs a uniform Hessian estimate along the relevant region and a positive separation distance, or direct minimisation of the full potential. Pairwise Gap is discontinuous at zero coherence, so compactness arguments must use a continuous observable or impose an amplitude floor. The former universal T-69 numeric barriers and unavoidable topological transition to any zero-Gap pair are retracted [✗] from this proof; they need stronger specified assumptions beyond (SV).

12. Connections and local dynamics​

For every trace-preserving local channel,

Tr⁡A[(TA⊗idB)(ΓAB)]=ΓB.\operatorname{Tr}_A[(\mathcal T_A\otimes\mathrm{id}_B)(\Gamma_{AB})]=\Gamma_B.

It follows from the dual identity TA∗(IA)=IA\mathcal T_A^*(I_A)=I_A. For state-selected local channels this remains a pointwise marginal identity if a valid extension is given. Full no-signalling with nonlinear updates additionally requires a specified measurement prescription; see Physics correspondence §8.