The Custody Cycle of the Most High: Investiture, Disclosure, and the Transfer of the Supreme God from Bronze Age Canaan to Rome, with the Textual, Institutional, and Cosmic Registers of One Recurring Design

A god whose cult center burned around 1185 BCE bears a title that is today spoken daily on every continent. This paper asks what, exactly, survived, and answers with a structural claim: not a temple, not a priesthood, not a state, but an office, the custody of the Most High, transmitted across successive and mutually hostile custodians by a repeatable cycle of seven stations. The cycle is reconstructed from converging witnesses: the Ugaritic tablets and Genesis 14 for the prior custodian; the textual seams of Deuteronomy 32 and Psalm 82 for the investiture; the ancient Near Eastern grammar of divine departure for what "the death of a god" meant in period terms; the charter against manufactured form and the unimaged inner sanctuary for the custody's one governing term; the golden calf, Kuntillet Ajrud, and the prophetic lawsuits for the recurrent failure; the hiding of the face and the staged exit of the divine glory for the prosecution; the destruction of 586 BCE and the empty room Pompey found for the sentence and the interval; and the Hypsistos cults, Constantine, and the three-day kerygma for the new investiture and its self-description. Two further registers replicate the same design: the Hebrew scribal tradition, which preserved its text by annotation rather than correction and thereby kept the seams of its own history legible under four successive empires; and the comparative record of civilizations, in which no polity that chartered its own eternity reached its stated term, the sharpest charters failed fastest, and every long-endurer institutionalized its own scheduled dissolution. Beneath all four registers the paper identifies a single engine, stated in the sources themselves: dissolution, which makes every frozen form mortal, and the emptiness of each new receiver, which makes fresh custody possible. The argument is graded throughout, its comparative method controls for common sources, and it closes with four falsifiable predictions.