{ "title": "Mathematical and exposition review", "manuscript": "Three simplifications in the Alpöge–Claude argument for Dirichlet GRH", "version": "21 November 2026", "text": "Review of the 21 November 2026 manuscript\nThree simplifications in the Alpöge–Claude argument for Dirichlet GRH\n\nMathematics\n\nThe authors checked the three paper arguments against their stated inputs.\nIn Proposition 2.1 the normalized input corank is the actual fraction of\nlow slopes before integration. The extremal two-point distribution\nattains the bound for the clipped mean shortfall. Substitution in the\nreleased arithmetic comparison preserves the pre-existing upper budget.\n\nIn Theorem 3.1 the average over all characters at a ramified prime gives\ntwice the uniform probability measure on the Frobenius coset. The\none-sided change of weight uses its positivity. The translation average\nin the arithmetic input is uniform. The distribution of tops in the final\ncharacter sum is weighted by their differentiation orders. In Corollary\n4.1, Chebyshev's bound absorbs a relative thick-weight width of order\nx^(-1/2 + epsilon) with fixed coefficient. This relaxation assumes the\nrest of the arithmetic and character estimates at the stated sizes.\n\nThe Mellin transform in Section 5 has positive residue at a real family\nzero, and the positively weighted multiplicities preclude cancellation at\nany family zero in the strip. The one-sided pole displacement preserves\nthe supremum of the real parts even when it is unattained. The frequency\nof sign changes used in Proposition 5.1 is the power-interval assertion\nin Kaczorowski and Pintz, Theorem 1, equation (2.2), and Corollary 1.\nFor the integral version, logarithmic averaging shortens the interval\nbut preserves the poles and the sign of their positive real residues.\n\nAn independent Codex collaborator checked the slope calculation and its\nuse in the released arithmetic inequality. The collaborator also checked\nthe single-scale comparison and projection, the local inertia masses, the\ncyclic character count and the distinction between a character and its\nfamily sum against the released definitions. In Section 5 the same\nreviewer checked cancellation at one, the residue formula, the displacement\nof real poles, the hypotheses of the Kaczorowski–Pintz theorem (including\nan unattained rightmost abscissa), and the integral criterion against the\nreleased window hypothesis. These checks concern proofs on paper.\nThe existence of the theta-lift data and the construction of simultaneous\narithmetic intervals are external inputs to Theorem 1.1.\n\nExposition\n\nThe intended audience is graduate readers of pure mathematics familiar\nwith Dirichlet L-functions and characters of finite abelian groups.\nThe introduction starts with the centered zeta prime count and identifies\nwhy the location of excursions matters. The quadratic case shows which\nfunction is fixed while the host field varies. An independent reading\nof Sections 1–3 of the combined draft checked the examples and the link\nfrom the geometric comparison to the prime sum. Equation (3.3) makes\nthis link at the actual weights of each translate. The two-order example\nthen illustrates the change of shares needed to evaluate at a common\nargument. The abstract states the stronger weight range and identifies\nthe arithmetic inputs to the final implication.\n\nFinite examples\n\nThe two Python programs in checks/ run in exact arithmetic. Their commands\nappear in README.txt. One verifies 2,048 local character-kernel identities,\n64 weighted projections and 210 directional ramp comparisons. The other\nindependently verifies 192 local kernels, 30 translate counts, 32 complex\nFourier coefficients and four ideal shares on its chosen examples." }