What Establishes the Book’s 80,000–45,000-Year Flood Window?

At a glance

Noah’s Flood Revisited dates the Flood to somewhere in an 80,000–45,000-year window, assembled from candidates it rules out, a model-based geography of Eden, and dates for human settlement far from Mesopotamia. This response takes the assembly apart link by link and asks what each piece actually measures.

Every dating argument is a chain, and the chapter that dates the Flood in Noah’s Flood Revisited is unusually candid about its own links. The book states that the Flood cannot presently be assigned a definitive date, then assembles what it can: candidate floods it rejects, an identification of Eden that implies a time, the ages of human presence in distant lands, and an appendix arguing that the dating methods lean the right way [1].

This response examines that assembly as a chain. Each column of the book’s case is examined for what it can establish, what it assumes, and what happens to the window when a link is pulled.

The book’s case, stated as strongly as it deserves

• The candidate floods deserve rejection. The excavated flood deposits of Mesopotamia, the proposed Black Sea drowning, and the late ice-age sea-level and lake events are each the wrong event — too recent, too small, or the wrong kind of catastrophe to be the judgment Genesis describes [2].
• Eden’s location implies a time. If Eden lay in what is now the floor of the Persian Gulf, and two of its four rivers vanished when that basin flooded, then the first humans lived while the basin was dry — and the book’s reconstruction of sea levels places that exposure mainly between about 115,000 and 18,000 years ago [1].
• Distant settlement dates the human race. Human presence in Europe, Asia, and Australia is dated by the book’s sources to roughly 50,000–39,000 years ago, which — since all people descend from the ark’s survivors — requires the Flood to be older still, very likely beyond 45,000 years [1].
• The dating methods lean upward. The appendix argues that early human-history dates carry large, often unquantified systematic errors, that indirect methods tend to overestimate, and that their results are best treated as upper limits — all of which, in the book’s hands, pushes the Flood deeper into the past [3].

What the book gets right

Three acknowledgments belong on the record.

The book declines an exact date and says so plainly. That honesty is worth naming, because the rest of this response is about the limits of what the surrounding evidence can do — and the book itself flags many of those limits.

The candidate rejections are, in the main, correct — and this site agrees with them. The Kish and Ur strata, the Black Sea drowning, and the Agassiz outbursts are not Noah’s Flood. The disagreement is about what their rejection can decide: ruling out candidates narrows a list of wrong answers; it does not by itself produce a right one.

And the book never claims to have found the Flood. No securely excavated horizon anywhere in its case is identified as Noah’s Flood deposit; the window floats free of any dug-up event. Readers should hold onto that as they watch the numbers move.

The candidates, and what rejecting them can decide

Start with Mesopotamia, where the record is thickest. The flood strata excavated at Ur, Kish, and Fara belong to the third millennium BC — local inundations on the Mesopotamian plain. The book sets them aside as later regional events [2], and the sources themselves support a much more modest reading than readers might assume: Woolley, who dug Ur, described his deposit in print as “not universal, but a local disaster” confined to the lower valley [4]. The dating frame of 3500–2600 BC belongs to the book’s synthesis, not to the excavators’ reports.

Two details sharpen the point. The excavators’ own speculation ran toward connection, not away from it — Langdon tentatively suggested one Kish flood “may have been the flood mentioned in the Bible,” which later scholarship quietly dropped. And the book’s stated reason for rejection — that these were local events after humanity had dispersed — is an inference, not a finding: it imports the very dispersal premise whose dating is under examination. On this site’s reading, these are ordinary post-Flood inundations; on the book’s, post-dispersal local floods. Either way they are not the Flood — and so nothing in them dates it.

The Black Sea is the same story at larger scale. Ryan and Pitman’s hypothesis — the book reports — identified a catastrophic refilling of the Black Sea with Noah’s Flood itself; the authors of that proposal said so directly [5]. The subsequent literature has trended smaller and slower, and the dispute remains live on both sides; the book reads the current evidence as favoring a gradual reconnection and rejects the identification [2]. Grant the rejection — the site has no stake in that hypothesis — and notice again what remains: a proposal ruled out, and no date produced.

The ice-age events follow the same pattern. The Sunda Shelf flooding, the Sicilian channel drownings, and the Lake Agassiz outbursts are real geological episodes of the late ice age, studied in their own terms; none of the sources, as verified, attributes human mortality to any of them, and several are modeled scenarios. The book’s rejection — that none would kill the whole human population — is its own inference about what a Noah event would require [2]. Again: correct as a rejection, silent as a date.

Our response

Three candidate rejections, one shared outcome: the wrong answers leave the right answer unnamed. This is the pattern the next sections test in the book’s remaining columns, where the window does its actual work.

Eden in the Gulf: a location built from models

The book’s second column turns on a place. If two of Eden’s four rivers cannot now be found flowing on the modern map, the book reasons, they may survive as buried channels beneath the Persian Gulf — which would place Eden on the basin’s floor, in the window when that floor stood dry [1].

As a location proposal, this is interesting; as a dating instrument, it stacks three interpretive moves. Genesis 2 names rivers; it does not require that every river remain traceable today — the identification of the two “missing” rivers with specific buried channels is a proposal about geography, not a datum the narrative supplies. Satellite imagery shows channels beneath the Gulf; what it cannot show is that these are Eden’s rivers. And the exposure window — mainly 115,000 to 18,000 years ago — is not measured; it is read off a sea-level reconstruction at a proposed depth for the basin floor, an exercise the book itself cautions is affected by local tectonic and isostatic variation [1].

That caveat matters more than it looks. The window’s lower edge is set by sea levels in an interval where ice-volume reconstructions remain contested; a vertical uncertainty of tens of meters moves the exposure window by thousands of years. The book’s own caution is therefore also this response’s: a location proposal built from a model can imply a date the model can also un-imply. It cannot carry the window.

The dispersal dates: what they measure

The heaviest column is the dates of people far from Mesopotamia. The band of 50,000–39,000 years comes from regional anchors, and the strongest of them are genuine: the cold-steppe occupation at Ranis, Germany, is dated by its excavators to about 45,000 years ago, described in careful, method-specific terms [6]. Similar finds mark South and Southeast Asia, Australia, and Siberia in the same broad era; the book’s own note says regional ranges differ and are retained individually [1].

What the anchors measure is human presence at sites — occupations, not populations, and certainly not the Flood. To convert any of them into a constraint on the Flood’s date, three premises must hold at once: that the dated ages are true ages; that all humans everywhere descend from the ark’s survivors, so any pre-flood population is impossible; and that the book’s reading of Genesis places the ark’s landing and the early nations in Mesopotamia. The third premise is examined in this series’ earlier responses; the second is biblical and shared here; the first is the book’s own stated weak point.

The first premise deserves its own sentence, because it is where the chain bites. Even as the book leans on these dates to require an older Flood, its appendix tells readers that such dates carry large unquantified systematic errors and should be treated as upper limits [3]. A reader cannot simultaneously be asked to trust the dates enough to reset the Flood’s age and distrust them enough to discount every shallower reading. The tension is not fatal to every date — some, like Ranis, are tightly argued — but it is fatal to the claim that the window’s lower edge is settled. And the endnotes themselves deserve a reader’s caution: at least one anchor citation in this chapter carries a misprinted year and a mistyped identifier — enough to remind anyone that apparatus errors travel even in careful books.

The language dates, read against their own sources

The book’s linguistic column offers an independent clock: language families are too old, the argument runs, for a recent Flood — Austronesian, for example, estimated at 13,000–17,000 years [1].

Checked against its own sources, the column comes apart. The chapter the book draws on for that figure — a survey of Bayesian phylolinguistics — uses the 13,000–17,000-year estimate for Austronesian as the rejected “Slow Boat” scenario; its own reported estimate for the family is far shallower, on the order of a few thousand years [7]. The figure the book cites is, in its source’s own terms, the number the method was argued to disprove. Recent Indo-European work points around eight thousand years, with the usual scholarly contestation on top. And the chapter’s own standing caution — that language change is not clocklike — cuts further against using family ages as precise constraints at all. A column that cites a rejected estimate as a constraint is not carrying the weight assigned to it.

The appendix’s cautions, turned around

The book’s final column is methodological: the appendix catalogues the ways dating can go wrong — incomplete resetting, burial history, exposure histories, calibration limits — and concludes that indirect dates should generally be read as upper limits [3].

The catalogue itself is largely fair; the caution is the kind of discipline this site wishes more dating arguments showed. But the generalization is doing something its examples cannot support. The appendix’s illustration is Jinmium, the Australian rock shelter whose first-round ages were an order of magnitude too old; the correction came from single-grain optical work and radiocarbon — better technique within the same methods families, not a verdict that the methods run systematically old [8]. One resolved episode, however instructive, is not the demonstrated rule the appendix needs; “almost always overestimate” is a characterization, not a finding.

And the rule, applied consistently, would be self-defeating. The same methods the appendix discounts — sea-level curves, radiocarbon calibration, luminescence and isotope dating — are the methods that produce every number in the book’s window. The book cannot discount them where they might raise a question and lean on them where they require the Flood to be old. A categorical claim that broad also sits awkwardly beside the appendix’s own pages, which describe uranium-thorium and argon methods operating across the very interval the appendix calls undatable [3].

Before writing: the oral-transmission claim and Babel

Two supporting arguments deserve brief notes. First, the book argues the Flood’s account could have been preserved orally across the many thousands of years before writing [9]. Possibly — but only after the window is established independently; oral preservation is a solution to a problem the long window creates, not evidence for the window itself. Its supporting material is popular-level, and the argument’s work is circular.

Second, the Babel chapter argues that Genesis 11 specifies neither instantaneous language diversification nor instantaneous dispersal, placing the process around 50,000–40,000 years ago [10]. The softness of the text is granted — Genesis does not describe the mechanism on a clock. But “need not be instantaneous” is not a date; the chapter’s own floor from linguistic science — diversification required before roughly 17,000–13,000 years at minimum — is not the 50,000-year figure either, and the gap between them is the same chain examined above.

What this leaves standing

Set the columns beside each other and the shape of the argument becomes clear. The window is the product of five links: elastic genealogies, rejected candidates, a model-based Eden, occupied continents, and an appendix that discounts the clocks. The first was examined in this series’ second response; the others here. The rejections are correct and date nothing. The Eden column is a location proposal wearing a chronology. The dispersal dates measure occupations under uncertainties the book itself emphasizes. The linguistic column cites its own source’s rejected estimate. And the methodological rule, taken at face value, would cancel the rest of the book’s numbers rather than assist them.

This is why the chapter’s honest sentence — that no definitive date can presently be given — is also the strongest claim its own evidence supports. The 80,000–45,000-year window is not established by measurement; it is constructed downward from assumptions, and it carries weight only for a reader who has already accepted the genealogical elasticity and the regional scope that the earlier responses examined. On the text itself, the Flood’s date is not a free variable standing at the end of an open chronology; it is fixed by the chronology the narrative computes. The responses that follow take up the book’s geological and mechanical arguments, where the same questions return wearing different clothes.