navstokgap

The Ising gap gives a conditional local parent Hamiltonian

Markdown source · PDF

The square-root Gibbs transform of C124 is a positive, three-site local, frustration-free Hermitian operator with unique ground state \(\sqrt{\pi_{N,b}}\) and exact gap \(a[1-\tanh(2b)]\). Multiplication by a supplied action constant gives an energy-unit Hamiltonian. This construction preserves the chosen stochastic clock and does not select physical dynamics or action units. It resolves G04’s finite-volume transfer decision.

1. Object, domain and transform

Use C124’s model: \(N\ge3\) periodic spins, \(0\le b<\infty\), per-site refresh rate \(a>0\) and stationary law \(\pi(\sigma)=Z^{-1}\exp(b\sum_i\sigma_i\sigma_{i+1})\). The generator \(Q=a\sum_i(E_i-I)\) acts on functions in complex \(L^2(\pi)\); \(E_i\) is conditional expectation over spin \(i\). Its time parameter is the auxiliary sampling clock and \(a\) has inverse-time units.

Let \(D=\operatorname{diag}\pi\) and \(U:L^2(\pi)\to\mathcal H_N=(\mathbb C^2)^{\otimes N}\) be \((Uf)_\sigma=\sqrt{\pi(\sigma)}f(\sigma)\). The codomain has the standard counting inner product; \(U\) is unitary between these two inner-product spaces. All operators below have the entire finite-dimensional space as domain. Set

\[ A_N=-UQU^{-1}=-D^{1/2}QD^{-1/2}. \]

Detailed balance makes this matrix Hermitian. Its quadratic form is

\[ \langle\psi,A_N\psi\rangle =\frac12\sum_{\sigma,\eta}\pi(\sigma)Q_{\sigma\eta} \left|\frac{\psi_\eta}{\sqrt{\pi(\eta)}}- \frac{\psi_\sigma}{\sqrt{\pi(\sigma)}}\right|^2\ge0, \]

where diagonal terms vanish. Irreducibility implies a one-dimensional kernel, spanned by the normalized vector

\[ |\Omega\rangle=Z^{-1/2}\sum_\sigma e^{(b/2)\sum_i\sigma_i\sigma_{i+1}}|\sigma\rangle. \]

Its squared basis amplitudes reproduce the Gibbs probabilities. Regarding those amplitudes and other vectors as physical quantum states requires a state/measurement interpretation in addition to this Hilbert-space map.

2. Explicit local terms and their positivity

Write \(Z_i|\sigma\rangle=\sigma_i|\sigma\rangle\) and let \(X_i\) flip spin \(i\). Put \(\theta=\tanh(2b)\), \(s=\operatorname{sech}(2b)\), \(c_+=(1+s)/2\), \(c_-=(s-1)/2\). Then

\[ A_N=\sum_i h_i,\qquad h_i=\frac a2\left[I-\frac\theta2Z_i(Z_{i-1}+Z_{i+1}) -(c_+I+c_-Z_{i-1}Z_{i+1})X_i\right]. \]

Derivation. For fixed neighbours with sum \(S=\sigma_{i-1}+\sigma_{i+1}\), the original flip rate is \(r_i(\sigma)=a[1-\sigma_i\tanh(bS)]/2\). Detailed balance gives transformed off-diagonal entry \(-\sqrt{r_i(\sigma)r_i(\sigma^i)}=-a/[2\cosh(bS)]\). The diagonal entry is unchanged. Since \(S=0,\pm2\), the off-diagonal coefficient equals \(c_++c_-\sigma_{i-1}\sigma_{i+1}\), giving the formula. The two neighbours are distinct even for \(N=3\); all indices are cyclic.

Alternatively \(h_i=a(I-UE_iU^{-1})\). Each \(E_i\) is an orthogonal projection, so \(h_i\ge0\), \(h_i^2=ah_i\) and \(h_i|\Omega\rangle=0\). Thus the sum is frustration-free, although its overlapping terms need not commute. Each term has norm \(a\) and support in the three sites \(i-1,i,i+1\); the global density transform has not introduced interactions of growing range. For a fixed neighbour block the local ground vector has components proportional to \(e^{bS/2},e^{-bS/2}\), which also checks the square-root and sign conventions. At \(b=0\), \(h_i=a(I-X_i)/2\) and the ground state is the product of plus-\(X\) vectors. These statements are direct finite-dimensional derivations.

3. Exact gap, normalization and limits (C126)

Unitary equivalence and C124 give

\[ \operatorname{gap}(A_N)=a[1-\tanh(2b)]. \]

The corresponding nonzero eigenvector is \(U M\) with \(M(\sigma)=\sum_i\sigma_i\), orthogonal to \(|\Omega\rangle\) by spin-flip symmetry. This is an all-mode spectral statement, inherited from C124 rather than inferred from locality or the ground state alone.

For a supplied constant \(K>0\) with action units, define \(H_N=K A_N\). Its ground energy is zero and its excitation gap in energy units is

\[ \Delta_N=Ka[1-\tanh(2b)]. \]

For \(K_N\ge K_0>0\), \(a_N\ge a_0>0\), \(0\le b_N\le B<\infty\), \(\Delta_N\ge K_0a_0[1-\tanh(2B)]\) uniformly in \(N\). These are bounds on a sequence of finite-volume Hamiltonians. They do not construct an infinite-volume Hilbert representation or establish a continuum gap. Fixed total refresh rate replaces \(a\) by \(a/N\) and closes the gap. Growing \(b_N\) or vanishing \(K_N\) can close it as well when the other parameters stay fixed.

4. Which physical premise has been supplied?

The exact semigroup identity is

\[ e^{-tH_N/K}=Ue^{tQ}U^{-1}. \]

It identifies the transformed sampling semigroup with imaginary-time evolution of the constructed operator, in the same time units. Defining the unitary \(e^{-itH_N/K}\) is mathematically possible but is an additional choice of real-time evolution; it is not the original stochastic transition law.

Input Use What remains to identify physically
Positive Gibbs law and detailed balance Unitary weighted-space transform and Hermiticity Physical state and observable interpretation
Single-site heat-bath generator Three-site terms and exact inherited gap Why this generator governs physical imaginary time
Per-site rate \(a\) Overall spectral rate scale Conversion from sampling time to physical time
Supplied action constant \(K\) Converts inverse time to energy and sets phase units Positive universal action normalization
Uniform \(K,a,b\) bounds Finite-volume energy-gap lower bound Their physical origin and any further limit construction

The classical energy function \(-J\sum_i\sigma_i\sigma_{i+1}\) used to prepare \(\pi\) is not the constructed \(H_N\): the latter has off-diagonal spin flips and coefficients depending on \(b=\beta J\) and \(Ka\). Knowledge of \(J\) and \(\beta\) fixes neither the refresh clock nor a universal action constant. For the same \(\pi\), replacing \(Q\) by \(\lambda Q\), \(\lambda>0\), preserves the ground state and locality while multiplying every excitation gap by \(\lambda\). Holding \(Q\) fixed and varying \(K\) independently gives the other normalization freedom. Thus the map yields a conditional stochastic-parent Hamiltonian; an independent physical identification is still a premise.

5. Evidence and strategic consequence

Written proof and literature status are recorded separately in the ledger. The B71 audit matches the established detailed-balance parent mapping in Henley and Castelnovo et al.; the explicit Ising term is a derived specialization. The written review checks its coefficients and inherited gap. No novelty is asserted.

G04 stops at the explicit finite-volume operator. Additional Ising spectra would not select its clock or action scale. A further operator task requires an independently specified physical dynamics whose imaginary-time generator can be compared with this one. The quantum track’s reversible-descent premise is selected next: test whether a specified interacting mechanical orientation flow closes on C125’s moment coordinates. STATE owns that bounded decision.