Checkability and the Editorial Gate: Why Scientific Publishing Exempts Its Own Judgements From Scientific Standards

The editorial gatekeeping of scientific publication is a chain of decisions: the editor's desk rejection, the recommendations of external reviewers, and the editor's adjudication of those recommendations, all shaped by journal metrics. The subjectivity of this process is well documented. This paper asks how discretion and accountability are distributed across the chain.

Conceptual and normative analysis organised around checkability, the standard science applies to its own claims: that a judgement's basis be recorded, its reliability measurable, and the judgement answerable beyond the say-so of whoever made it. Each stage of the chain is assessed against this standard, using published evidence on editorial and reviewer agreement.

The editorial gate is exempt from checkability, and exempt by the greatest margin where it exercises the most discretion. The desk rejection, which disposes of most submissions at many journals, is a single editor's rapid decision on criteria acknowledged to be subjective, usually without a stated basis, external check or appeal. Reviewer recommendations whose agreement barely exceeds chance are aggregated into a single outcome. The adjudication of discordant reviews is again individual discretion, lightly checked. Existing codes govern editors' conduct, not the epistemic basis of their decisions.

The wrong is not subjectivity, which is unavoidable, but an inversion: a system that demands checkability of the science it admits exempts from checkability the judging that admits it. The remedy is neither eliminating editorial judgement nor reviewing every submission, but holding the gate to the standard of checkability science demands of itself.