The collapse of the Neo-Assyrian Empire in 612 BCE — the most powerful state the ancient world had yet produced — has been debated for two and a half millennia. Two competing hypotheses dominate the literature: the slow institutional decay hypothesis, which holds that the empire was weakening internally before the Babylonian-Median coalition delivered the final blow, and the sudden external catastrophe hypothesis, which holds that a fundamentally healthy empire was overwhelmed by an unusually powerful and coordinated external force. This paper provides the first quantitative evidence bearing on this debate. We apply positional formula analysis to the complete State Archives of Assyria (SAA, 3,154 tablets, 20 volumes) — the most authoritative collection of Neo-Assyrian administrative documents ever assembled — measuring administrative formula consistency chronologically across reigns from Tiglath-Pileser III (745 BCE) through Ashurbanipal (627 BCE). Three principal findings emerge. First, correspondence formula adherence increased significantly from 54% under Tiglath-Pileser III to 84% under Sargon II (chi-square = 66.09, p < 0.001, non-overlapping bootstrap confidence intervals), confirming the known peak of Assyrian administrative efficiency under Sargon II with a specific quantitative measurement for the first time. Second, a trend analysis across seven chronological data points spanning 120 years shows statistically significant increasing formula standardization (Mann-Kendall p = 0.0069, permutation test p = 0.0023) — the administrative scribal tradition was becoming more consistent over time, not less. Third, scholarly correspondence formula adherence is perfectly stable across 100 years (88.1% vs 88.2%, Cohen's d ≈ 0, p = 0.97), demonstrating that the scribal educational tradition was preserved with essentially perfect fidelity across four generations. These findings are inconsistent with the slow institutional decay hypothesis and consistent with the sudden external catastrophe hypothesis. The Neo-Assyrian administrative machinery showed no measurable decay in the periods observable in the SAA corpus. The empire's bureaucratic core was functioning — and by measurable criteria improving — when the coalition armies arrived. All results are independently reproducible from the publicly available ORACC catf corpus (github.com/oracc/catf). No proprietary data, special tools, or institutional access are required. The analysis is produced using the Molina Methodology — a positional formula analysis framework currently being applied across multiple ancient administrative corpora. Priority deposits for related work are available on Zenodo under the Molina Methodology designation, pending peer-reviewed publication.
Juan Gabriel Molina (Fri,) studied this question.