# NS-54 — map, closure and heat-flow scope

Internal analytic review, pinned baseline f0178e0 (v1.53), with separately identified current v1.54 update bea0ed2. No new numerical experiment or window. These findings concern Claude's supplied essay, not a retraction of the existing exact identities.

## MAJOR — “every obstruction” is false

The map closes different constructions for different reasons. The norm-contraction witness is a **positive** direction with q/D>2.05 at lambda4 (prop:v120-norm-counterexample); the fixed tail comparison has positive q/D<7e-18 and <12e-18, contradicting the stipulated1e-8, not proving a negative Weil direction. The block comparison Rayleigh values341/250 and2093/1500 exceed1. The abstract scalar-set countermodels share scalar set extrema but have signed bottoms-1/5 and+1/10; their logical point is missing joint operator information. The v1.53 removal-rank obstruction comes from a complete block with norm residual C exp(-lambda²/2000). None is identified by a theorem with the largest local zero-count deficit or with an S(t) omega statement.

The actual v1.37 depth proof exploits a negative cutoff side lobe under RH and then a bounded-floor contradiction for unconditional divergence; its unconditional conclusion has no linear rate. A scalar amplitude or primitive cannot replace complete operator, parity/source and cross-term information.

## MAJOR — QG plus packets does not close all concentration

At pinned manuscript lines11214–11235, QG growth and independent source-admissible bounded-energy, all-Borel-dominating packets give

eta(m_a) >= [c0 QC_K(tilde m_a)-Q_*]_+.

Thus a fixed number of distinct minorant values cannot have uniformly bounded error. This closes that bounded-level class, conditional on BOTH open inputs. It says neither that arbitrarily increasing level counts fail nor that all signed weighted concentration estimates fail. The unrestricted exact adaptive-level criterion in v1.52 is equivalent to the full complement-floor target. The broad concentration node remains blocked/open to specific arithmetic estimates. The packet input is not supplied by a law of scalar symbol values. “Last purple mechanism turns red” overstates the conclusion.

## MINOR — zero motion has the wrong description

For the conventional family H_t satisfying H_t,t=-H_t,zz, the real-zero velocity law is x_k'=2 PV sum_{j!=k}1/(x_k-x_j). Each two-zero internal term is repulsive as t increases: if their separation is d>0, its contribution to d' is4/d, not negative. Other zeros contribute as well; no universal attraction description follows. Rodgers–Tao(2021), equations(4),(5), Theorem11 and discussion on printed pages4–5,27, https://arxiv.org/pdf/1801.05914 . The cited theorem is stated on the hypothetical all-real regime Lambda<t<=0; the local simple-zero identity follows directly from the backwards heat equation wherever it applies. This does not assert every individual real gap grows in the full interacting system.

## NOTE — the heat comparison is a valid open transport question, not supplied by v1.46

The t>=Lambda sum-of-squares observation is valid for the correctly normalized deformed ZERO form on a smooth compactly supported core. It is not the unchanged arithmetic prime-side window form. Version1.54/NS52 now supplies an exact q_(a,t) and convergence statement; it explicitly does not claim a closed deformed form on the full original domain or justify differentiating the infinite sum.

The proposed monotonicity must specify direction and quantifier. q_(a,t0)>=0 at a later real-zero time is useful for the original form only with a backward lower comparison, e.g. q_(a,0)[f]>=q_(a,t0)[f]-C||f||² with C uniform in the window. Forward persistence of globally real zeros does not give that inequality. Already for L_t=H_t'/H_t, the exact evolution is partial_t L_t=-partial_z²L_t-2L_t partial_zL_t. The nonlinear term prevents replacing the deformation by a linear Gaussian smoothing of the original arithmetic symbol.

NS52 gives finite-cluster identities retaining interactions with the exterior zero factor. Under the expressly unverified condition that an all-real deformed time contains a zero unmatched by the original radical transform, it excludes the displayed same-test uniform comparison. That is a conditional, scoped obstruction, not a theorem excluding every heat transport. Test transport and other nonlinear comparisons remain open.

## NOTE — novelty and map status are stale

The claim “nobody has asked” is contradicted by the already merged conclusions audit PartB3, which formulates the nonlinear signed heat comparison, and by now-merged v1.54/NS52. The broad DBN “no bridge” closure was already withdrawn; at f0178e0 its map status is blocked, and at bea0ed2 it is open. The latter map also has debranges, f1 and nb open. Thus DBN is not the unique closed alternative now warranting reopening. Scoped proved closures must remain distinct from a whole research subject.

Read the supplied essay as a research proposal needing these corrections. Its exact original-form identity and the correctly defined real-zero sum of squares survive. No full arithmetic QG, ZLD disproof, heat monotonicity, signed floor, G2 or RH result is established by it.