This paper treats the Linear A corpus of Minoan Crete (c. 1800–1450 BCE) as a closed, non-linguistic administrative database ("Excel in Clay"). Enforcing a strict projection-before-meaning protocol, this audit demonstrates that the century-long failure of phonetic decipherment is an architectural error rather than an informational deficit. Through forensic arithmetic and structural analysis, the paper establishes five independent tablet reconciliations (HT 122, HT 115, HT 23, HT 28, HT 101) proving rigorous internal computational consistency. Key original findings include: physical proof that inscriptions on large storage vessels (e.g., ZA Zb 3, KN Zb 27) function as sector-level warehouse manifests, as their declared volumes exceed physical vessel capacity by 66% to over 1,000%; the reclassification of the "libation formula" as a standardized six-field Stock Keeping Unit (SKU) barcode governing commodity grade and custody; and the identification of a closed-loop tablet-to-vessel tracking pipeline evidenced by the corporate identifier DI-NA-U linking central ledgers (HT 9a) to physical transit vessels. The paper connects this architectural failure to Perceptual Control Theory (PCT) and AI alignment, arguing that both machine learning models and historical linguistic efforts fail when optimizing for output plausibility rather than input verification. Part of the "Projection Before Meaning" cross-corpus research program.
Łukasz Diener (Fri,) studied this question.