Current task: —

Updated 08-17-2026

D-016 — Q-3.4a: does the unconditional n₊ bound shrink Groskin's indeterminate zone?

Date: 2026-08-17 | Session: 002 | Task: 3.4 (part a) | Status: complete Tags: [VERIFIED] (scale analysis) / [ANALYSIS] (verdict)

Question

Groskin (2607.02828): truncated Weil form at archimedean cutoff T, band N, prime cutoff c certifies: eigenvalue ≥ 0 ⟹ cutoff-free positivity; eigenvalue < −B_T ⟹ off-line zero exists; [−B_T, 0) certifies NOTHING, with B_T ~ (2N+1)ρ log T/(π²T), ρ = 2π/log c. Does the Anthropic paper's unconditional positive-index/rank lower bound (N^s_0 ≥ (2/3−o(1))N ⟹ n₊ of the truncation ≥ (2/3−o(1))N) shrink the indeterminate zone?

Analysis

  1. Scale of the zone. B_T = O(N log T/T) = o(1) — the zone lives at eigenvalue scale → 0 (microscopic, tail event: at most O(1)-many eigenvalues can live there for bounded p... in fact an eigenvalue in the zone corresponds to an off-line pair at depth δ ≲ B_T·L-ish — sub-log scale from the line).
  2. Scale of our input. The n₊ ≥ (2−R)N bound is a BULK statement (macroscopic spectral mass). The Phase 1 wall analysis (W4/W5, G-001): all 2-moment/second-moment inputs are blind to o(N) spectral events — the D–H/Epstein insensitivity. The indeterminate zone is exactly an o(N)-scale question (is there ANY eigenvalue in [−B_T, 0)?).
  3. Zero-density attempt: eigenvalues ≤ −ε (fixed ε > 0) need off-line pairs at depth ≳ ε·L; Selberg-type density bounds N(σ,T) ≪ T^{1−c(σ−1/2)} log T give n₋(ε) = o(N) for fixed ε — but at the zone's scale ε ~ B_T = o(1) the depth threshold δ ~ B_T·L/T... density bounds are vacuous there (T^{1−cδ} ≈ T(1−cδ...) loses nothing). [Same invisibility as G-001: depth→0 pairs are unreachable.]

Verdict

No. The unconditional n₊ lower bound cannot shrink the indeterminate zone: bulk second-moment data and microscopic tail certification are decoupled by exactly the wall mapped in Phase 1 (W4/W5 mechanism). The zone can only be closed by input sensitive to individual shallow off-line pairs — i.e. RH-strength information at o(1/log T) depth. Consistent with D-013's necessary condition 2 for ANY fix.

Salvage (positive by-products)

  1. Unconditional negative-index bound (count form): since n₋(truncation) ≤ p ≤ (N − N^s_0)/2, Theorem A gives n₋ ≤ (1/6 + o(1))N — an explicit unconditional upper bound on the negative index of the truncated Weil form. Restatement of Thm A in Bombieri's language; possibly new as a stated corollary (check: Bombieri counts, we BOUND the count — worth a literature check for the paper's "known results" section). [CANDIDATE MINOR RESULT — log.]
  2. Quantitative detection direction: an off-line pair at depth δ produces a negative eigenvalue of magnitude ~ δ·(frame scale): the truncated form is a DETECTOR with depth-to-eigenvalue calibration — combining with Groskin's B_T gives "no eigenvalue < −B_T ⟹ no pairs at depth ≳ f(B_T)" — a conditional exclusion statement. To be quantified if Track C continues.
  3. Q-3.4a is closed; the scale-mismatch argument is reusable for ANY proposed "bulk input improves tail criterion" idea — a general filter: does the input see o(N) events? If no, it cannot touch positivity-type certifications.