Context
The Sumerian King List starts pinch 8 kings who ruled earlier the flood. Their reign lengths are tremendous and unusually regular. Three examples are 28,800 years, 36,000 years, and 43,200 years. Most are integer multiples of 3,600, and each 8 are multiples of 600. Their full is 241,200 years.
The input series tested present is the ETCSL composite antediluvian list:
| 1 | Alulim | 28,800 years |
| 2 | Alalgar | 36,000 years |
| 3 | Enmenluana | 43,200 years |
| 4 | Enmengalana | 28,800 years |
| 5 | Dumuzid | 36,000 years |
| 6 | Ensipadzidana | 28,800 years |
| 7 | Enmenduranna | 21,000 years |
| 8 | Ubara-Tutu | 18,600 years |
| Total | 241,200 years |
The transliterations travel the composite matter cited below. Like the post-flood sections of the King List, the antediluvian database is simply a textual contented pinch variants, truthful the array is an study input alternatively than a declare astir literal humanities reigns.
One speculative mentation treats these numbers arsenic a distorted representation of prehistory. Under this hypothesis, the reign boundaries encode existent ambiance shifts, eruptions, impacts, aliases sea-level changes. After rescaling and anchoring the database to a projected flood date, the boundaries should coincide pinch dated events successful the geological record.
Here, I trial that thought pinch an exploratory analysis. The explorer rescales each chronology to a fixed 241.2 ka span, anchors 1 boundary, and compares the resulting dates pinch a catalog of Quaternary events. I usage 11.6 ka BP, aliases astir 11,600 years ago, arsenic an analyst-chosen anchor adjacent the Younger Dryas termination. The King List does not supply that day aliases propose this paleoclimate interpretation.
Finding matches is easy. With 9 boundaries and state to alteration the anchor aliases bandwidth, chance alignments are common. The applicable mobility is whether the observed Sumerian reign bid scores unusually precocious nether a intelligibly defined null model.
Results
- In the superior paleoclimate catalog, the Sumerian series does not show a statistically important alignment. At the fixed 11.6 ka anchor and a kernel bandwidth of σ = 1.60 ka, the permutation p-value is 0.350. After accommodation for aggregate comparisons, q = 0.622.
- Expanding the study to each 103 usable catalog entries increases the number of evident matches but does not alteration the conclusion. At the aforesaid bandwidth, the wide-catalog p-value is 0.148 and the adjusted q-value is 0.430.
- The smallest earthy p-value for the Sumerian series occurs successful the catastrophic exploratory catalog astatine σ = 1.60 ka: p = 0.021. This was the champion consequence recovered successful a larger exploratory search. After accommodation for each comparisons, q = 0.222, truthful the consequence does not support the hypothesis.
- As a secondary check, I besides count really galore events autumn wrong a fixed region of a boundary. The kernel people is the main measurement because it gives little weight to events farther distant alternatively of utilizing an abrupt cutoff.
What Would Count arsenic Evidence?
A reliable alignment would request to meet 3 conditions.
First, the matter must find which bound to anchor. The Sumerian King List places the flood aft the eighth reign, truthful I anchor the extremity of the antediluvian sequence. The matter does not delegate that bound an absolute date. I usage 11.6 ka BP arsenic an analyst-chosen day adjacent the Younger Dryas termination.
Second, the Sumerian series should look different adjacent to different ancient chronologies. I comparison it pinch a series derived from Biblical patriarchal ages, 7 deity and demigod reigns from the Manethonian part preserved successful the Excerpta Latina Barbari, and the first 8 listed Kish I rulers. These are examples for comparison, not independent statistical controls, and I do not trial whether 1 chronology outperforms another. They show really readily unrelated ancient sequences tin nutrient evident matches nether the aforesaid procedure.
Third, a reliable consequence should stay important aft accounting for the tested chronologies, catalog tiers, and bandwidths. None does.
Interactive Explorer
Use the selectors beneath to move chronology, catalog tier, and bandwidth. The superior paleoclimate catalog defines the main analysis. The wide and catastrophic catalogs supply exploratory sensitivity analyses. The fixed-anchor conception uses p-values precomputed successful Rust. The sliding-anchor conception recomputes the anchor expanse successful the browser.
Sumerian King List alignment explorer pinch fixed-anchor p-values and exploratory anchor sweeps.
Chronology
Catalog
Bandwidth
Fixed-anchor analysis: Gaussian kernel pinch anchor fixed astatine 11.6 ka (precomputed successful Rust)
Kernel score
-
at fixed anchor
Permutation p-value
-
raw p; BH-adjusted q
Binary hits
-
sensitivity measure
Adjustment for aggregate comparisons crossed each fixed-anchor permutation tests: nary consequence has q < 0.05.
Exploratory anchor search: champion anchor from 10 to 13 ka (live)
Best anchor successful window
-
maximum kernel people from 10 to 13 ka
Maximum kernel score
-
relative to the fixed 11.6 ka anchor
Monte Carlo exceedance estimate
-
3,000 windows pinch centers azygous from 0 to 100 ka
Kernel people arsenic the anchor slides from 0 to 30 ka BP
Anchor (ka BP)11.6
Timeline astatine existent anchor
-
How to Read the Numbers
The main statistic is simply a Gaussian kernel proximity score:
S = Σᵢ Σⱼ exp(-(tᵢ - bⱼ)² / 2σ²)
Here tᵢ is an arena day and bⱼ is simply a reign boundary. A adjacent arena contributes almost 1 point, while the publication of a distant arena approaches zero. This avoids the abrupt discontinuity of a difficult cutoff, wherever an arena conscionable wrong the model counts and 1 conscionable extracurricular does not.
The 2 statistical questions are different. The simplest measurement to spot the quality is to inquire what is allowed to move:
| Exhaustive permutation | 11.6 ka anchor, catalog tier, bandwidth, and the aforesaid group of reign lengths | The bid of the reign lengths | Fraction of each branded orders that people astatine slightest arsenic precocious arsenic the observed order | Primary p-value |
| Random-anchor Monte Carlo | Reign order, catalog tier, and bandwidth | The anchor, aliases the section anchor window | Fraction of sampled anchors aliases windows that people astatine slightest arsenic high | Secondary sensitivity check |
The 2 bandwidths are calibrated to person the aforesaid full weight arsenic difficult windows extending 1 ka and 2 ka connected either broadside of a boundary. Using σ = τ·√(2/π) gives σ ≈ 0.80 ka and σ ≈ 1.60 ka. The browser truncates the kernel astatine 4σ for speed, matching the Rust precomputation.
The superior study fixes the anchor astatine 11.6 ka BP and reports the observed kernel score, a earthy permutation p-value, and a q-value adjusted for aggregate comparisons. The p-value comes from an exhaustive permutation test, not Monte Carlo sampling. The programme checks each imaginable reign order. With the anchor, catalog tier, bandwidth, and group of reign lengths held fixed, the p-value is the fraction of those orders whose kernel people is astatine slightest arsenic ample arsenic the observed score. Under this null model, each reign bid is treated arsenic as plausible. The trial does not relationship for choosing the anchor, catalog, bandwidth, aliases people aft inspecting the data.
The superior catalog includes the Younger Dryas termination utilized to motivate the 11.6 ka anchor, truthful the terminal lucifer is built into the setup. Its publication is changeless crossed reign-order permutations and is not grounds that the soul series is unusual.
The secondary anchor hunt asks really the consequence changes erstwhile the anchor is optimized aft inspecting the data. It compares the maximum successful the fixed 10 to 13 ka model pinch maxima from 3,000 windows whose centers are sampled uniformly from 0 to 100 ka. Both the fixed and random windows usage a 3 ka width and 0.1 ka anchor spacing. The displayed Monte Carlo estimate adds 1 to some the exceedance count and the proceedings count truthful that a finite simulation ne'er reports a probability of precisely zero. It is not the article’s superior p-value.
The fixed-anchor consequence paper besides reports a abstracted random-anchor sensitivity value. For each chronology, catalog tier, and bandwidth, the Rust programme compares the people astatine 11.6 ka pinch scores from six cardinal anchors sampled uniformly from 0 to 100 ka. This calculation is chopped from the browser’s 3,000-window maximum-score estimate. Neither worth is the superior p-value.
Data and Comparators
The root catalog contains 104 dated events from the Quaternary period, grouped into 11 categories. The superior study uses a narrower paleoclimate catalog containing Heinrich-event dates, Greenland Interstadial onset dates, and selected Holocene climate-event dates. The resulting superior catalog contains 39 events wrong the 0 to 260 ka BP study window.
The wide exploratory catalog contains each 103 usable entries, including the 39 superior entries. It besides includes meltwater pulses, Marine Isotope Stage boundaries, ample volcanic eruptions, effect structures and contested effect hypotheses, geomagnetic excursions, utmost star proton events, megafauna extinction nodes, and awesome taste transitions. The catastrophic and deep-time tiers supply overlapping sensitivity analyses. Lonar is kept successful the root catalog but excluded from the dashboard because it falls extracurricular the study window.
Each introduction records a selected date, an approximate uncertainty, root information, and whether the arena is contested. The uncertainty values were estimated successful different ways, truthful they are not straight comparable and should not each beryllium publication arsenic modular errors. The existent study uses only the selected dates. The wide catalog is still exploratory, moreover though each introduction has a documented source.
For the Biblical comparator, the first 8 values are patriarchal lifespans and the last 600 is Noah’s property astatine the flood. For the Excerpta Latina Barbari comparator, I usage the 7 named deity and demigod reigns listed successful that fragment: Hephaestus 680, Helios 77, Sosinosiris 320, Orus 28, Typhon 45, Anubes 83, and Amusis 67 years. For Kish I, I usage the first 8 rulers successful the ETCSL composite text; that contented has manuscript variants, truthful the comparator should beryllium publication descriptively alternatively than arsenic a fixed humanities chronology.
Each chronology is transformed successful the aforesaid way. Reign lengths are rescaled truthful the full span equals the Sumerian antediluvian full of 241.2 ka, past converted to absolute dates by anchoring 1 bound and accumulating durations backward successful time. This normalization removes full long arsenic a facet and compares only the comparative spacing of the boundaries.
The Rust programme sounds the aforesaid events.json catalog published pinch this article. It runs six cardinal random-anchor tests per chronology, catalog tier, and bandwidth, runs exhaustive permutation tests, applies the Benjamini-Hochberg correction to the permutation p-values, and writes the consequence array utilized by the page. The TypeScript explorer loads the nationalist catalog and consequence table, past calculates and displays the unrecorded anchor sweep.
Implementation Note
The costly activity runs offline successful Rust alternatively than successful the browser. The superior trial is mini capable to enumerate exactly: the Sumerian and Kish sequences each person 8! = 40,320 branded orders, the Excerpta Latina Barbari series has 7! = 5,040, and the Biblical comparator has 9! = 362,880. Repeated reign lengths are still treated arsenic branded positions, truthful the null exemplary asks whether this bid is different among each reorderings of the aforesaid values.
The halfway calculations are definitive successful the source: Gaussian kernel scoring, Heap’s algorithm for permutations, the random-anchor sensitivity check, the Benjamini-Hochberg step-up adjustment, and the deterministic JSON writer. The random-anchor sensitivity checks usage a fixed seed, truthful rerunning cargo tally --release regenerates the published consequence record alternatively than producing a caller simulation each time.
Caveats
A Benjamini-Hochberg accommodation for aggregate comparisons covers the 32 fixed-anchor permutation tests. It does not correct for earlier experimentation pinch arena inclusion, anchor choices, people definitions, aliases comparator construction. Because this is an exploratory study and the afloat group of tests was not specified successful advance, the q-values summarize this hunt alternatively than corroborate a finding. No comparison has q < 0.05.
The catalog was assembled editorially alternatively than done a systematic reappraisal aliases an inclusion protocol defined successful advance. Several entries picture related parts of the aforesaid ambiance sequence, truthful the 39 superior dates should not beryllium interpreted arsenic 39 independent observations. The permutation trial keeps this catalog fixed; it cannot make the catalog independent aliases complete.
Catalog density varies substantially complete time. Recent events are much galore and mostly amended dated, truthful chronologies pinch Holocene boundaries person much opportunities to lucifer an event. The permutation trial partially addresses this imbalance by holding the anchor and arena catalog fixed while changing only the reign order.
The deep-time sensitivity tier removes the crowded caller grounds by utilizing only dates older than 60 ka. The Sumerian series scores astatine aliases beneath chance astatine some bandwidths.
The study does not transportation making love uncertainty done the calculations. Several entries person uncertainties wider than the fixed matching windows, and the Younger Dryas effect presumption remains disputed. Treating each introduction arsenic an nonstop day overstates the precision of the result.
The reported p-values and q-values do not afloat relationship for expert choices, including decisions made aft inspecting the matches. The interface exposes immoderate of these choices, but the study remains exploratory.
Data and Code
The arena catalog and generated consequence array are nationalist and machine-readable:
- Event catalog pinch root metadata
- Precomputed dashboard consequence JSON
- Fixed-anchor Rust precompute, browser script, catalog, and results merchandise v0.2.1
The arena catalog records the study tier, day estimate, uncertainty, root URL, contested flag, and editorial notes for each entry. The consequence record contains the values generated by the Rust programme and displayed by the browser explorer.
Selected Sources
- The Electronic Text Corpus of Sumerian Literature, Sumerian King List
- CDLI grounds for the Sumerian King List tradition
- Hemming 2004, Heinrich events
- Rasmussen et al. 2014, INTIMATE arena stratigraphy for Greenland crystal cores
- Lisiecki and Raymo 2005, LR04 benthic stack
- Earth Impact Database
- The Fragments of Manetho, Book I, including Excerpta Latina Barbari
Next Steps
A early type should put the making love uncertainties connected a much accordant ground and transportation them done the calculations. Repeatedly sampling plausible dates for each arena would nutrient a scope of scores and deed counts alternatively of a azygous worth for each.
A stronger study would usage a catalog assembled independently and defined earlier immoderate chronology is scored. Repeating the process pinch respective specified catalogs would trial whether the consequence depends connected the underlying arena information aliases connected editorial curation.
This study explores a circumstantial alignment claim. It finds nary statistically important grounds that the Sumerian reign bid encodes the paleoclimate catalog utilized here.
English (US) ·
Indonesian (ID) ·