Spectral theory · OWR-3389-016
The gap bound
is never attained.
Nonattainment in the Harrell–Stubbe
Dirichlet gap inequality
The result
For every nonempty bounded open Ω ⊂ ℝn, n ≥ 1, and every finite J ≥ 1. Eigenvalues are Dirichlet eigenvalues counted with multiplicity. Connectedness and smooth boundaries are unnecessary.
Here M1 = ((n+2)/n) mean(E1, …, EJ) and M22 = ((n+4)/n) mean(E12, …, EJ2).
A short equality argument
- The classical spectral identity expresses the Yang deficit as a sum of nonnegative terms.
- Equality would make a coordinate times an eigenfunction a finite sum of eigenfunctions. Self-adjointness rules this out.
- The strict Yang inequality and its quadratic roots give the strict gap bound. Repeated ground eigenvalues are handled directly.
The paper proves strict Yang inequality at every real threshold above the lowest eigenvalue.
What was checked
Two fresh adversarial AI reviews checked the preprint; the final round found no actionable issue. Exact rational checks cover 1,312 certified box/index cases, including 600 zero gaps and 12 repeated-ground cases.
The 7,016 checks audit algebra and explicit spectra. They do not machine-prove the general analytic theorem.
Preprint review · Verification report · Priority audit · Source audit
The question and its scope
This gives a negative answer to the corrected saturation question in OWR-3389-016. The 2009 report prints a missing square on M2; the manuscript documents why the squared-means expression is the intended bound. The neighboring growth-bound questions are outside this result.
The scoped literature audit located no earlier explicit nonattainment proof. Ashbaugh’s 2002 survey explicitly left strictness of Yang’s first inequality undecided. The underlying trace identities and non-strict bounds are classical. The audit is not an exhaustive priority certificate.
Reusable research package
Ready for a Zenodo deposit
The kit includes the paper, source archive, metadata, checksums, and fields ready to copy. No DOI has been minted.