Accessibility-Loss Mortality: The Triple Barrier Before the Discoveries of Non-Native, Disabled, and Neurodivergent Scholars

Academic accessibility is usually discussed as a question of fairness: who is able to enter higher education, obtain accommodation, publish, or build a research career. This article identifies a further consequence. When scholarly institutions require an insight to pass through a narrow sequence of representational forms before its content can be evaluated, they may not merely exclude a person. They may permanently lose insights that could otherwise have become knowledge.The article identifies a triple barrier. First, a person must convert thought into conventional written language and must be able to read the formats in which scholarship is stored. Second, that writing must often be converted into academic English. Third, when artificial intelligence is used to cross either of the first two barriers through dictation, transcription, translation, editing, or polishing, the resulting text may face suspicion because it sounds machine-generated. The remedy for one barrier can therefore create the next. Evidence already establishes each component separately: researchers with disabilities report systematic accessibility barriers; scholarly PDFs remain widely inaccessible; non-native English-speaking researchers bear additional costs and higher rates of language-related rejection; and AI detectors and stylistic judgments can disproportionately misclassify non-native English writing.This article distinguishes this structure from recognition lag. A delayed discovery remains available for later recognition. An unrecorded insight may disappear when its bearer dies, loses the relevant memory, or leaves no durable and interpretable record. The article calls the life-related consequence Accessibility-Loss Mortality: avoidable loss of human life caused when a potentially valid and life-saving insight is prevented from entering a durable, assessable record by accessibility barriers and istherefore never evaluated or implemented. No magnitude is claimed. The contribution is to identify the causal structure and the institutional responsibilities it creates.Current publisher policies already provide part of the solution. Major publishers increasingly locate human authorship in verification, interpretation, final approval, correction, and accountability rather than in unaided manual text production. If a person can explain a claim orally, distinguish supporting from conflicting evidence, answer objections, reject errors, and authorize correction, the absence of unaided typing or fluent academic English should not by itself disqualify that person as an epistemic contributor. Generative AI can function as an accessibility interface. Scholarly institutions must therefore evaluate warranty and answerability directly, permit multimodal contribution and verification, and stop treating linguistic style as evidence of cognitive origin.