\chapter{Mathematical contribution, dependencies, and limits}
\section{Established foundations and the proposed synthesis}
The soul theorem of Cheeger and Gromoll, Perelman's soul-conjecture theorem,
O'Neill's submersion equations, and the Ambrose--Singer theorem provide the
classical Riemannian background \cite{cheeger-gromoll,perelman,oneill,ambrose-singer}.
Associated-bundle constructions with large normal holonomy are already part
of that literature \cite{tapp}. The existence of nonnegative curvature
with nontrivial normal holonomy is not claimed as a new discovery here.

There is also an established quantitative study of loop length and holonomy.
Sol\'orzano's holonomic spaces explicitly retain a norm on the holonomy group
defined by the least length of a realizing loop, rather than recording only
the abstract group \cite{solorzano}. Thus the distinction between holonomy
as a group and holonomy with a length cost is not itself new. The present
problem further specifies a single marked ray, scalar observations,
finite-dimensional frame criteria, and a noise experiment.

Regular simplices and tight frames are classical \cite{benedetto-fickus}.
Finite-dimensional realization from input--output quantities has a classical
systems-theoretic history \cite{HoKalman}. The explicit Gaussian simulation
kernel is an application of the comparison-of-experiments framework
\cite{Blackwell}. The manuscript proves its precise special case rather
than attributing novelty to that framework.

The candidate mathematical contribution is the conjunction of the following
proved statements in one compatible observation model: an explicit
fixed-soul degeneration; a global bound with the factor
$\lambda(1-\lambda)$; a uniform sharp small-loop constant; an exact circle
writer; an exact conversion of the maximal scalar deficit into optimal Gram
and exterior-sensor resources; the resulting two-level sensitivity law; and
a sharp statistical cost under a fully stated Gaussian experiment. The
finite-radius comparison supplies a geometric calibration compatible with
weak signals rather than an unrelated absolute error estimate.

These claims of derivation and proof should not be read as a certified
historical-priority statement. A targeted primary-source search was performed
for this release, but it is not a comprehensive classification of all work
on holonomy control, inverse connection problems, experimental design, or
statistical comparison. Independent specialist review is still required to
assess priority and the most appropriate publication framing.

\section{Proof dependency map}
The main statements have a short dependency structure despite the number of
applications. The quotient metric is established independently of the
observation algebra. Its connection formula then feeds the transport ODE.
The two pure-gauge endpoints and the second parameter derivative give the
uniform quadratic bound. Normal coordinates, a sharp alternating-area
inequality, and the exact circle construction give the sharp local scalar
law. No numerical fit enters that chain.

The exact scalar-resource bridge uses only bounded-length compactness,
stabilizer symmetry, and cap frame optimization. It does not require an
explicit finite-length formula for the maximizing loop. Exterior singular
values then provide deterministic inverse bounds. The Gaussian theorem
uses only the interval of accessible scalar gaps and a comparison kernel;
the geometric asymptotic is inserted afterward. This separation prevents
a weakness in a numerical example from being mistaken for a dependency of
the main theorem.

\begin{longtable}{>{\raggedright\arraybackslash}p{.28\textwidth}>{\raggedright\arraybackslash}p{.31\textwidth}>{\raggedright\arraybackslash}p{.32\textwidth}}
\toprule
Statement&Essential inputs&What is not required\\
\midrule\endhead
Fixed soul and full normal holonomy&Quotient metric; curvature span; standard
submersion and holonomy results&An experimental model or scalar decoder\\
Uniform scalar asymptotic&Parameter interpolation; area bound; exact circle
writer&Finite-length global optimality of that circle\\
Exact Gram/exterior bridge&Scalar extremum; compactness; stabilizer action;
cap optimization&A reconstruction algorithm that discovers the optimizer
without calibration\\
Noisy target recovery&Known error bounds; positive true Gram matrix&Independent
entry noise or a Gaussian assumption\\
Exact adaptive risk&Two known hypotheses; reset scalar Gaussian shots;
fixed sample size&Full matrix reconstruction or a second physical ray\\
Finite-radius comparison&Explicit connection metric; totally geodesic fibers;
local convex neighborhoods&A universal detector-noise law\\
\bottomrule
\end{longtable}

\section{Precisely what is complete}
All numbered statements in the manuscript are accompanied by proofs or an
explicit reference to an earlier proved statement. The principal external
Riemannian theorems are identified rather than silently reproved. The
computational library and its executed checks are supplied in readable
source form. The manuscript, implementation, example tables, theorem ledger,
and machine-readable verification record form a complete release for these
stated results.

The evidence does not include a proof-assistant formalization, independent
peer review, an optical experiment, or a benchmark of a physical computing
system. The percentage $142/142=100\%$ describes the listed automated
checks only. It is not a numerical probability that the mathematics is
correct and not a percentage of worldwide novelty verification.

\section{Boundaries that do not obstruct the proved results}
The exact finite-length optimum $D_\lambda(L)$ is not expressed in an
elementary closed form. The circle formula gives an explicit admissible
loop and the correct sharp leading constant. The exact resource identities
remain valid in terms of $D_\lambda(L)$ without solving the separate
finite-length isoholonomic optimization problem.

The theorems are not a classification for arbitrary souls or arbitrary
normal connections. The fixed-soul family is sufficient to disprove the
specified universal conditioning claim and to derive a complete local
observation law. Additional curvature nondegeneracy, derivative, or control
assumptions could yield different positive bounds in other families.

Likewise, scalar testing does not determine every feature of an unknown
ambient manifold. The statistical null and alternative are specified
connections on the same marked experiment. Access to the ambient metric,
external frames, multiple initial vectors, or independent curvature data
would be a different experiment and might distinguish the hypotheses more
efficiently.

\section{The central conclusion}
For this explicit Riemannian family, the operational content of hidden
holonomy is not captured by the group alone. One scalar extremum links
geometric displacement, matrix conditioning, and Gaussian testing risk:
\[
 D_\lambda(L)
 \quad\longrightarrow\quad
 \Gamma_\lambda(L)=\Omega_\lambda(L)
 =\frac d{d-1}(2D_\lambda(L)-D_\lambda(L)^2)
 \quad\longrightarrow\quad
 \Phi\!\left(-\frac{\sqrt nD_\lambda(L)}{2\sigma}\right).
\]
The arrows are exact theorems under their stated cap and noise assumptions.
The sharp local coefficient in $D_\lambda(L)$ turns them into explicit
length and noise laws. This is the substantive endpoint of the manuscript:
qualitative geometric freedom can persist while every bounded-resource
scalar observation becomes arbitrarily weak, and the weakening has a
computable, optimally attainable leading scale.