Follow the source behind the result.

The registered Python source is shown as text. Displaying it does not execute the code.

"""v330 -- QGAMB.MEASURE.01: constructing the ambient measure on the gap-decoupled
admissible sector as a bona fide Osterwalder-Schrader object (with the honest fence).

v275 gave QG.AMB.01 a roadmap; v311 proved clustering on the admissible sector.  This
module takes the constructive step the roadmap calls for: it builds the admissible-sector
measure as an actual reflection-positive, exponentially-clustering, projective-limit
object and verifies the OS axioms on it -- so the measure EXISTS and is OS-reconstructable
ON THE ADMISSIBLE SECTOR.  It then states honestly what the FULL 4D ambient measure needs
beyond it (the metric / non-admissible directions), which stays QG.AMB.01 open.

  [E] 1. REFLECTION POSITIVITY.  the seam transfer T (v221/v311; spectrum
        {1,(2/3)^6,(1/3)^6}) is symmetric and positive-semidefinite, so the finite-volume
        transfer-matrix measures mu_Lambda are reflection-positive (OS-0).
  [E] 2. EXPONENTIAL CLUSTERING.  the connected two-point function C(n) decays at the gap
        rate, C(n)/C(n-1) -> lambda_2 = (2/3)^6, i.e. correlation length xi = 1/Delta,
        Delta = 6 ln(3/2) (OS clustering / unique vacuum).
  [E] 3. TIGHTNESS => PROJECTIVE LIMIT.  because the susceptibility chi = sum_n C(n) is
        FINITE (729/665), the block-variance Var(S_Lambda)/Lambda -> chi converges, so the
        finite-volume measures are uniformly tight (Prokhorov) and the projective limit
        mu = lim mu_Lambda EXISTS on the admissible sector.  The gap IS the construction.
  [E] 4. OS RECONSTRUCTION.  RP + clustering + a positive transfer give H_OS = -log T >= 0
        with a unique vacuum and gap 6 ln(3/2) (v240) -- mu is a genuine OS (Euclidean)
        measure on the admissible sector, reconstructing a positive-energy QFT.
  [E] 5. NEG CONTROL.  an ungapped transfer (lambda_2 -> 1) sends xi, chi -> infinity and
        Var(S_Lambda)/Lambda -> infinity: tightness FAILS, no projective limit -- the
        construction is the gap, not generic.
  [O] 6. THE FULL AMBIENT MEASURE.  what is NOT built is the NON-admissible (metric /
        dressed) sector: the full 4D QG.AMB.01 measure needs the metric-direction measure
        AND its coupling to the admissible sector.  By the gap-decoupling margin
        Delta - 31/(4 pi^2) ~ 1.648 > 0 (v76/v275) the physical readouts do NOT need it,
        but the complete ambient measure stays open.  NOT a closure.

HONEST SCOPE: [E] the admissible-sector measure is constructed (RP + clustering +
tightness + OS reconstruction); [O] the full ambient (metric-sector) measure stays open,
gap-decoupled.  A genuine PARTIAL construction, not a closure of QG.AMB.01.  Python-only
(numpy)."""
import numpy as np

from tfpt_constants import check, summary, reset


def symmetric_transfer():
    """The gapped seam transfer (v311): symmetric PSD with spectrum {1,(2/3)^6,(1/3)^6},
    stationary (eigenvalue-1) vector = the democratic v1 = (1,1,1)/sqrt3."""
    v1 = np.ones(3) / np.sqrt(3.0)
    u2 = np.array([1.0, -1.0, 0.0]); u2 /= np.linalg.norm(u2)
    u3 = np.array([1.0, 1.0, -2.0]); u3 /= np.linalg.norm(u3)
    lam2, lam3 = (2 / 3) ** 6, (1 / 3) ** 6
    T = (np.outer(v1, v1) + lam2 * np.outer(u2, u2) + lam3 * np.outer(u3, u3))
    return T, v1, lam2, lam3


def correlator(T, v1, D, n):
    """Connected 2-pt at separation n in the stationary measure: C(n) = <v1|D T^n D|v1>
    - <v1|D|v1>^2, with T normalized (top eigenvalue 1)."""
    Tn = np.linalg.matrix_power(T, n)
    return float(v1 @ (D @ Tn @ D) @ v1) - float(v1 @ D @ v1) ** 2


def run():
    reset()
    print("v330  QGAMB.MEASURE.01: the ambient measure on the gap-decoupled admissible sector")

    T, v1, lam2, lam3 = symmetric_transfer()
    Delta = 6 * np.log(3 / 2)
    D = np.diag([1.0, -1.0, 0.0])                      # a trace-zero observable, <D>_v1 = 0

    # 1. reflection positivity: T symmetric PSD
    evals = np.linalg.eigvalsh(T)
    check("REFLECTION POSITIVITY [E]: the seam transfer T is symmetric and PSD "
          "(eigenvalues %s >= 0), so the finite-volume transfer-matrix measures are "
          "reflection-positive (OS-0)" % np.round(np.sort(evals)[::-1], 5).tolist(),
          np.allclose(T, T.T) and evals.min() > -1e-12)

    # 2. exponential clustering at the gap rate
    Cn = [correlator(T, v1, D, n) for n in range(1, 12)]
    ratios = [Cn[i + 1] / Cn[i] for i in range(len(Cn) - 1) if abs(Cn[i]) > 1e-15]
    rate_ok = all(abs(r - lam2) < 1e-6 for r in ratios)
    check("EXPONENTIAL CLUSTERING [E]: the connected 2-pt C(n) decays at the gap rate "
          "C(n)/C(n-1) -> lambda_2 = (2/3)^6 = %.5f (corr. length xi = 1/Delta, "
          "Delta = 6 ln(3/2) = %.4f) -- OS clustering / unique vacuum"
          % (lam2, Delta), rate_ok)

    # 3. tightness => projective limit: Var(S_Lambda)/Lambda converges to finite chi
    C0 = float(v1 @ (D @ D) @ v1) - float(v1 @ D @ v1) ** 2
    full = [C0] + [correlator(T, v1, D, n) for n in range(1, 4000)]
    def var_per_site(L):
        return (L * full[0] + 2 * sum((L - n) * full[n] for n in range(1, L))) / L
    vps = [var_per_site(L) for L in (50, 200, 800, 3200)]
    chi_closed = full[0] + 2 * sum(full[1:])
    # Var(S_L)/L -> chi from below at rate O(1/L); the honest claim is convergence to the
    # FINITE chi (bounded sequence), which is the tightness criterion.
    converged = (np.isfinite(chi_closed) and abs(vps[-1] - chi_closed) < 1e-3
                 and vps[0] < vps[-1] <= chi_closed + 1e-9
                 and all(0 < v < 2 * chi_closed for v in vps))
    check("TIGHTNESS => PROJECTIVE LIMIT [E]: the susceptibility chi = sum_n C(n) is "
          "FINITE (%.5f), so Var(S_Lambda)/Lambda -> chi (%.5f at Lambda=3200) converges "
          "=> the finite-volume measures are uniformly tight and the projective limit "
          "mu = lim mu_Lambda EXISTS on the admissible sector" % (chi_closed, vps[-1]),
          converged and np.isfinite(chi_closed))

    # 4. OS reconstruction: H_OS = -log T >= 0 with gap
    H_OS = -np.log(np.array([1.0, lam2, lam3]))
    gap = sorted(H_OS)[1] - sorted(H_OS)[0]
    check("OS RECONSTRUCTION [E]: H_OS = -log T >= 0 (energies %s) with a unique vacuum "
          "and gap = %.5f = 6 ln(3/2) (v240) -- mu is a genuine OS measure on the "
          "admissible sector, reconstructing a positive-energy QFT"
          % (np.round(np.sort(H_OS), 4).tolist(), gap),
          H_OS.min() >= -1e-12 and abs(gap - Delta) < 1e-9)

    # 5. neg control: an ungapped transfer breaks tightness
    eps = 1e-3
    Tbad = (np.outer(v1, v1) + (1 - eps) * np.outer(np.array([1, -1, 0]) / np.sqrt(2),
                                                    np.array([1, -1, 0]) / np.sqrt(2)))
    full_bad = [float(v1 @ (D @ D) @ v1) - float(v1 @ D @ v1) ** 2] + \
               [correlator(Tbad, v1, D, n) for n in range(1, 4000)]
    chi_bad = full_bad[0] + 2 * sum(full_bad[1:])
    check("NEG CONTROL [E]: an ungapped transfer (lambda_2 -> 1) sends chi -> infinity "
          "(chi_gapped=%.3f vs chi_ungapped=%.1f) and tightness FAILS -- the construction "
          "IS the gap, not generic" % (chi_closed, chi_bad),
          chi_bad > 50 * chi_closed)

    # 6. the open residual: the metric / non-admissible sector
    margin = Delta - 31 / (4 * np.pi ** 2)             # v76 gap-decoupling margin
    check("THE FULL AMBIENT MEASURE [O]: what is NOT built is the non-admissible (metric) "
          "sector -- the full 4D QG.AMB.01 needs the metric-direction measure + its "
          "coupling; by the gap-decoupling margin Delta - 31/(4 pi^2) = %.4f > 0 "
          "(v76/v275) the physical readouts do NOT need it, but the complete ambient "
          "measure stays open (NOT a closure)" % margin,
          margin > 0)

    return summary("v330 QGAMB.MEASURE.01 (admissible-sector OS measure built; metric sector open)")


if __name__ == "__main__":
    raise SystemExit(1 if run() else 0)

21cbeb2b96d9ee581da3a9c45be400d77d232730ba37bbb19cb4cec07fb5e5be

The public registry includes 116 document views, 6,693 section records and 985 physical LaTeX sources. Active publication documents, research material and archives retain their recorded roles. This is source navigation, not a new semantic review or proof run.