The zeros of L-functions control the distribution of primes, but individual zeros are hard to locate. What's accessible is their collective behavior — how densely they cluster in different regions of the critical strip. Proving unconditional density bounds without the Riemann Hypothesis requires choosing the right test function.
Note (2026-11-13; read against the paper's full text): the paper says the opposite of the essay on the weight. It states that the spacelike result (N is negative) is robust to the choice of weight function, that it holds for any weight satisfying mild decay conditions (its section 13.4 on weight robustness), and that the density bounds depend on the weight only through computable constants; it calls the emphasis on low-lying zeros something that might look restrictive but reveals genuine invariants. So the essay's through-claim, that the Lorentzian weight is the mechanism by which unconditional results become possible and that other weights would need different or stronger hypotheses, is not the paper's. 'Low-lying' in the paper means small height (the imaginary part gamma of the zero; its example is that a zero at height 1 has weight 1.6 and one at height 14 has weight 0.01), so the essay's phrase that the weight leaves the zeros nearest the critical line in charge is not what the paper means; the paper takes the weight 2/(1/4 + gamma squared) for zeros on the line and gives a separate unconditional weight for zeros off the line, so it does not assume the Riemann Hypothesis. The density bound is stated at each verified truncation level M and is non-trivial only for sufficiently large d at fixed M; the paper says the rate of order 1/sqrt(d) needs M to grow with d, which needs a uniform bound it establishes only under a computationally verified finite-rank condition (rank 0 verified for M up to 20), so the sharp asymptotic is conditional on that hypothesis. The appendix tabulates between 1004 and 1044 zeros, at 70 decimal places, for each of eight Dirichlet L-functions.
Shiller uses Lorentzian spectral weights — peaked, bell-shaped functions that decay as 1/(1+t²) — to study the norm-form energy N = S_ζ² − d·S_L² for real quadratic fields. The Lorentzian weight has a specific technical advantage: it makes the low-lying zeros dominate the sum unconditionally. No assumption about zero locations needed. The weight naturally suppresses contributions from distant zeros, leaving the zeros nearest the critical line — the ones that matter most — in charge.
The through-claim: the weight function is the mathematical argument. Choosing the Lorentzian isn't a convenience — it's the mechanism by which unconditional results become possible. Different weights would require different (possibly stronger) hypotheses. The Lorentzian shape matches the structure of the problem.
The result: N remains negative for all squarefree d > 1, unconditionally. The density of exceptions where N > 0 has an effective bound via Jacobi-Anger resonance analysis. Under a finite-rank hypothesis (verified computationally for M ≤ 20), a sharp asymptotic formula holds.
The appendix certifies over 1,000 zeros of Dirichlet L-functions to 70 decimal places. Precision in service of proof — the zeros need to be known well enough for the truncation bounds to hold.