The Crisis in Academic Publishing: Symptoms, Diagnoses, and Potential Solution Concepts

Abstract

Academic publishing bundles four functions — administration, certification, development, and recognition — into a single, slow, gatekeeping step. We document the resulting failures: two or three referees cannot reliably judge importance (inter-rater reliability near 0.3); submission volume is surging under generative AI; and the signals that actually distinguish important work emerge only over years. We argue the crux move is to decouple certification from recognition and to break the constraint of simultaneity. We describe the Academic Ledger, a neutral non-profit record in which work is certified up front and importance accrues over time through measured impact and named, signed endorsement, organized in three tiers — Certified, Refereed, and Canon — and we sketch a stylized two-stage model that makes a low early entry bar optimal under noisy signals and heavy-tailed importance.

Keywords: scholarly publishing · peer review · evaluation · inter-rater reliability · citation dynamics · large language models · mechanism design

Inside the QaL estimate — the signals behind it

Weights are tunable and public; the score is reproducible from the raw signals below. No single number confers Canon — a field panel does, under published criteria, with this dossier in view.

All metrics

Endorsements — 27 named, signed, on the record

Discussion — 89 comments · 64 discussants

Submission history

Comments: 28 pages, 5 figures. Seminar version prepared for the OID Department, 9 July 2026.