Old Testament Miracles as History

Summary

The miracle stories of the Old Testament can be judged as history in part.  A historian cannot certify that God acted, but can ask three separate questions of each story:  how close the text is to the event, whether its ordinary events happened, and whether anything unusual is independently attested.  Asked case by case, those questions give very different answers.

• Method (Part 1). Hume's claim that no testimony could ever establish a miracle was called "an abject failure" by the philosopher of science John Earman.¹ Bart Ehrman's narrower point stands: history works in probabilities and cannot certify a miracle. It can still test everything around one.
• Plagues and sea crossing (Part 2). No Egyptian record exists, and the Exodus cannot yet be dated. Colin Humphreys's natural chain for the plagues and Drews and Han's wind-setdown model for the crossing are serious, but each needs a run of perfectly timed events, and the crossing model rests on a reconstructed sandbar. Santorini's eruption is a century and more too early. The Ipuwer papyrus should not be cited.
• Manna and quail (Part 3). Tamarisk honeydew matches the Bible's description of manna closely, but the scientist who identified it found that Sinai yields "several kilograms" a year.² The quail are a real migration; quail poisoning, the usual explanation of the plague in Numbers 11, belongs to the autumn passage, not the spring.
• Jericho (Part 4). A walled city did fall violently, but its excavators date the destruction to 1550-1525 BC, which fits neither Exodus date. An earthquake explains walls falling at the cost of about a 1 in 1,300 coincidence in the week of the siege.
• Joshua's long day (Part 5). The text says it is quoting the Book of Jashar, an older poem. The "NASA missing day" is a legend whose own teller admitted he could not document it. The 1207 BC eclipse proposal is real but fits the sky, the hour and the length of the story badly.
• Elijah (Part 6). The Phoenician historian Menander, quoted by Josephus, records a drought under Ethbaal, Jezebel's father, ended by a thunderstorm. It lasted one year, not three and a half.
• Sennacherib (Part 7). The strongest case. Sennacherib's own prism confirms the invasion, the blockade of Jerusalem, Hezekiah's 30 talents of gold and the failure to take the city; the Babylonian Chronicle confirms his murder by a son. Assyria's failure to depose Hezekiah was unusual, and Herodotus preserves a separate story of a night disaster. The 185,000 dead have no outside support, and the claim that Berossus confirms a plague rests on a lost quotation. The full case has its own article.
• Jonah (Part 8). Nineveh was a city of about 750 hectares, and the "three days" can be read as a visit rather than a width. The story of James Bartley, a sailor said to have survived inside a whale in 1891, is false: the captain's wife denied it, and the crew list has no Bartley. Whether Jonah is history or parable is a question of genre that history cannot settle.
• Hezekiah's shadow and the axe head (Part 9). The setting, Merodach-baladan's embassy to a sick Hezekiah, fits 703-701 BC. The proposed eclipse of 689 BC does not fit the sign, and the axe head offers nothing to test.

Verdict.  History affirms the ordinary core of Sennacherib's campaign, Elijah's drought, the fall of Jericho and Jonah's Nineveh, and shows something unusual in 701 BC.  It finds mismatches of date and size that defenders must admit, and four popular "confirmations" that should be dropped.  Every natural explanation examined works only with exact timing, and timing is the one thing history cannot measure.

Part 1: How a historian can judge a miracle story

History cannot certify that God acted.  It can test almost everything around that claim, and the tests come out very differently for different stories.

Hume's argument and what is left of it.  David Hume's essay "Of Miracles" (1748) is the source of the modern rule that miracle reports are never credible.  He defined the term:  "A miracle is a violation of the laws of nature;  and as a firm and unalterable experience has established these laws, the proof against a miracle, from the very nature of the fact, is as entire as any argument from experience can possibly be imagined."  He then gave his maxim:  "no testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavours to establish."³  He added that miracle stories "are observed chiefly to abound among ignorant and barbarous nations."³

The most thorough modern assessment came from John Earman, a philosopher of science at the University of Pittsburgh who defends no miracle report.  His verdict in Hume's Abject Failure (2000):  "I contend that 'Of Miracles' is an abject failure.  It is not simply that Hume's essay does not achieve its goals, but that his goals are ambiguous and confused.  Most of Hume's considerations are unoriginal, warmed over versions of arguments that are found in the writings of predecessors and contemporaries."¹  Read as a statement about probability, the maxim says only that testimony must make an event more probable than not, which Earman calls "the unhelpful tautology."  His dilemma for Hume:  "There is a weak version of his thesis that is surely correct, but it amounts to no more than a collection of platitudes ... There is a very strong version of his thesis that is far from platitudinous:  indeed, it is patently false."¹  The weak version, that extraordinary claims need strong evidence, survives and is used here.  The strong version, that no evidence could ever be enough, does not.

The methodological objection.  Bart Ehrman puts the modern form without Hume's metaphysics.  In his 2006 debate with William Lane Craig he said:  "Historians can only establish what probably happened in the past," and "we can't claim historically that a miracle probably happened.  By definition, it probably didn't.  And history can only establish what probably did."  He was careful to add:  "Miracles are not impossible.  I won't say they're impossible," and later:  "I don't accept Hume's argument that miracles can't happen."⁴  Not every critic agrees that the question is closed to history.  Gerd Lüdemann, who rejected the resurrection, held that a historian faced with a miracle claim "must be interested in a natural explanation of the miracle—or it must admit that even on historical grounds a supernatural explanation is more plausible."⁵

This article takes Ehrman's caution seriously and does not ask history to pronounce on God.  It asks three separate questions of each story, each of which a historian of any belief can answer.

1. Is the text early and close to the event? How many years separate the event from the writing, and does the writer claim older sources?
2. Did the ordinary part happen? Every miracle story is embedded in ordinary events: a city fell, an army withdrew, a drought came, a people ate something in the desert. Those can be checked against archaeology and foreign records.
3. Is there evidence for or against the extraordinary part? Is anything unusual independently attested? Does the evidence rule the event out? And when a natural explanation is offered, what does it cost: which facts must be assumed, which details of the text must be dropped, and how much coincidence of timing does it need?

The third question is where most popular argument goes wrong on both sides.  A natural mechanism that only works with a lucky coincidence has moved the improbability, not removed it.  And a "confirmation" that turns out to be a legend costs the believer more than having no confirmation at all.

Part 2: The plagues of Egypt and the sea crossing

No Egyptian text records the plagues, the escape or the loss of an army, and the Exodus itself cannot yet be dated.  The natural explanations on offer each fit some details well and buy that fit with a string of lucky timings, and the best physical model of the crossing was built by a researcher who declines to call it the Exodus.

The date and route belong to the site's Dating the Exodus article, which concludes that the event "cannot presently be dated":  the early date of about 1446 BC fits the Bible's own numbers better, and the late date in the 1200s fits the place names and Israel's appearance in Canaan by 1208 BC.  Neither has a settling artifact.  That leaves the score for this case thin at the start.  (a) The text's date is disputed, as the Deuteronomy article shows for the Pentateuch.  (b) The ordinary part, a group of Semites leaving the eastern delta, has circumstantial support, which that article weighs, but no direct record.  (c) The extraordinary parts can only be judged by whether a natural cause could produce them, and at what cost.

The best-developed natural account of the plagues is Colin Humphreys's.  Humphreys, a Cambridge materials scientist who accepts the narrative as history, set out the rule he works by in The Miracles of Exodus (2003):  "What made certain natural events miraculous was their timing."⁶  His chain runs from a toxic algal bloom in the Nile estuary ("the first plague occurred not on the freshwater Nile River, but in the Nile Estuary"), through frogs driven out and dying, biting midges and stable flies breeding in the dead frogs, livestock disease carried by the midges, a great hailstorm, locusts, and three days of darkness from a desert sandstorm, the khamsin.⁶  Building on Greta Hort's natural-chain account of 1957, he concludes:  "Thus we have a natural scientific explanation for all ten plagues, which follow a logical, connected sequence."⁶

Its strengths are real.  Each step is one or two links from a known organism or weather pattern of the delta, and the sequence runs in a plausible order of seasons.  Its costs are three.

1. Key links are inferred, not observed. For the first plague Humphreys concedes: "As far as I am aware, no one has looked to see what toxic algae exist today in the Nile Delta."⁶
2. The tenth plague strains it hardest, and he says so. He writes that the death of only the firstborn "seems impossible to explain by science. Why should some fatal disease afflict only the firstborn? And why only males?"⁶ His answer is a grain mould toxin, with firstborn sons given first share of the surviving grain. For the firstborn animals he proposes the same favoured feeding, on the analogy of Hebrew practice, and admits: "I have searched ancient Egyptian literature and can find no indication of whether or not firstborn livestock were special to the Egyptians."⁶ And nothing in the chain explains why the Israelites in the same region were spared.
3. The timing carries everything. Ten natural events in one year, each announced in advance, each ending on cue, is the miracle. Humphreys says so himself, which is to his credit. It also means the naturalistic version is not cheaper than the biblical one: it moves the improbability from the mechanism to the timing, as the earthquake did at Jericho.

The sea crossing has a published physical model, and it is worth reading exactly.  Exodus 14:21 says "the Lord drove the sea apart by a strong east wind all that night."⁷  A steady wind blowing along a body of shallow water piles the water at the far end and drops it at the near end;  this is "wind setdown."  Carl Drews and Weiqing Han of the National Center for Atmospheric Research modelled it in PLoS ONE in 2010.⁸  Their results:

1. The Gulf of Suez, the traditional site, does not work. The reef Nof and Paldor proposed there in 1992 "is presently too deep to emerge at these wind speeds," and with realistic seabed shape "the reef takes more than 12 hours to become fully exposed and passable."⁸
2. A lagoon in the eastern Nile delta does. At the Kedua Gap, at the eastern end of the ancient Lake of Tanis, a wind of 28 metres per second (100 km/h) blowing for 12 hours exposes a land bridge "3–4 km long and 5 km wide, and it remains open for 4 hours." They add: "A 100 km/h wind would be memorable but would not prevent travel on foot."⁸
3. The model rests on a reconstruction, and the authors say which part is weakest. "The most uncertain element of our geographic reconstruction is the long coastal sandbar running from Damietta to Pelusium," and without the sandbar the 28 m/s wind opens no crossing.⁸ At a stronger wind the passage stays open longer, but "these stronger winds may render walking too difficult for a mixed group of people."
4. The authors decline the identification. Their competing-interests statement reads: "The present study treats the Exodus 14 narrative as an interesting and ancient story of uncertain origin."⁸

Humphreys puts the crossing instead at the head of the Gulf of Aqaba, because a northeasterly wind "would have blown straight down the Gulf of Aqaba, exactly as required for wind setdown to occur," while it would blow across the Gulf of Suez.⁶  His figure is that a hurricane wind of 80 miles per hour "pushes back the sea for a distance of 800 yards," which he calls "an approximate estimate" with no stated method.⁶  For the drowning he needs a second piece of timing:  normally the wind dies over hours and the water "slowly returns to its previous position, causing no threat to humans," so he proposes that the wind stopped suddenly at dawn and the water came back as a bore at eleven miles per hour.⁶

What wind setdown costs.  The physics is sound:  Drews and Han cite an eyewitness report from Lake Manzala in 1882, when a gale drove the lake "away beyond the horizon" and "the natives were walking about on the mud."⁸  Throughput is not a problem either:  a passage 5 km wide open for 4 hours could pass any number the text implies.⁹  The costs are the location and the timing.  The working site depends on a reconstructed sandbar and on reading yam suph as a delta lagoon rather than the Red Sea of Exodus 15:4.  The wind must arrive the night the Israelites are trapped, last all night, and stop when the army is in the gap.  The text's water "forming a wall for them on their right and on their left" (Exodus 14:22) is the detail Drews and Han built their model to meet:  they chose a bend in the lagoon so as to test "the possible separation of a body of water, with the presence of water remaining on both sides of a dry passage," and note that an east wind on the north-south Gulf of Suez "will not produce a dry region with water remaining on both sides."⁸  Humphreys's Aqaba setdown pulls the water back from the head of the gulf and leaves water on one side only, so it meets the "wall" only if the word is read loosely.  Neither model produces water standing high on either hand.

The Santorini proposals fail on the date.  The eruption of Thera in the Aegean was among the largest of the Bronze Age, and since A. G. Galanopoulos in 1964 writers have tied its ash and tsunamis to the plagues and the sea.  Barbara Sivertsen's The Parting of the Sea (2009) is the fullest version:  "the 1628 B.C.E. Minoan eruption of Santorini/Thera that was the cause of most of the Exodus plagues and the impetus for the first departure from Egypt," with a tsunami drowning the Egyptians in a coastal lagoon.¹⁰  The eruption has since been dated closely.  Sturt Manning's 2022 analysis gives "~1606–1589 BCE (68.3% probability), ~1609–1560 BCE (95.4% probability)."¹¹  That is 114 to 160 years before the early Exodus date of about 1446 BC and more than three centuries before the late date.  Sivertsen accepts this by placing her Exodus in 1628, which matches neither biblical date.  A Thera explanation can keep the eruption or the Bible's chronology, not both.

The Ipuwer papyrus should not be used.  Its line "the river is blood" is often quoted as an Egyptian memory of the first plague.  The manuscript, Leiden I 344, is a copy made in the Nineteenth Dynasty, as the catalogue entry by its modern editor, Roland Enmarch, records, and the poem is a lament about the collapse of order in Egypt, not a narrative of named events.¹²  Its date of composition is disputed, but no proposed date ties it to either Exodus date.  Creation Ministries International, which once used it, now prints a box at the head of the section:  "we now caution against using the Leiden papyrus as evidence."¹³

Part 3: Manna and quail

Manna is a real desert substance that matches the Bible's description closely, and the scientist who identified it said plainly that nature supplies a few kilograms of it, not a nation's bread.  The quail are a real migration.  In both cases the natural explanation accounts for what the food was and fails on how much of it there was.

Exodus 16 describes "a thin flaky substance, thin like frost," "like coriander seed and was white, and it tasted like wafers with honey."  Kept overnight "it was full of worms and began to stink," and "when the sun got hot, it would melt."  The Israelites "ate manna 40 years."⁷

The natural explanation, in its author's words.  F. S. Bodenheimer, a Hebrew University zoologist, collected in Sinai in June 1927 and published "The Manna of Sinai" in The Biblical Archaeologist in 1947.  Two scale insects living on tamarisk excrete a sugary honeydew:  "This manna is none other than the well known honeydew excretion of so many plant lice and scale insects."  It appears in June "for a period of three to six weeks," and the Bible's season fits:  "This date agrees with the natural season of manna production."²  Ants carry it off once the ground warms, which he took to be the "worms," and he judged the melting "a late and mistaken interpolation."²

Bodenheimer priced it himself.  "Usually the annual crop does not exceed several kilograms, but one steady man may collect over a kilogram a day at the peak of the season.  This certainly does not allow for the 'bread' or daily food of the wandering Israelites."  And:  "the nutritive value of these few kilograms of manna could not have been important enough to deserve a recording in Israel's history."²

Humphreys accepts the identification and tries to close the gap by shrinking the people and enlarging the trees.  He reads the Hebrew 'elep, "thousand," in the census lists as "troop," which turns 600,000 men into "5,550," and "the total number of men, women, and children at the Exodus was probably about 20,000."⁶  He then argues that the large quantity was found only in one place, the Desert of Sin, which he puts in northwest Arabia where tamarisks grew thickly, and that "the Bible doesn’t record that they ate a lot of manna all the time."⁶  He concedes:  "We do not know which tree, and which insect, was responsible for the manna the Israelites ate in the Desert of Sin."⁶

The arithmetic.  Humphreys reads the daily ration, an omer, as "about two quarts" per person.  For his 20,000 people that is about 38,000 litres a day, over 20 tonnes if the granules pack like loose sugar.¹⁴  Bodenheimer's figure for the whole year's crop in the manna valleys of central Sinai was "several kilograms."  So 20,000 people would need, every day, several thousand times what Sinai yields in a year;  the traditional reading of 600,000 men, with their families, needs a hundred times more again.  A single collector's rate, "over a kilogram a day" at the peak, is not the limit;  the supply is.  The season is three to six weeks a year;  the text says forty years.  Humphreys's dense Arabian tamarisk groves have not been surveyed, as he says, so the gap cannot be closed with data.  The double portion on the sixth day and none on the Sabbath (Exodus 16:22-30) has no natural counterpart at all, and Bodenheimer did not discuss it.

What survives.  The match between the biblical description and tamarisk honeydew, in look, taste, season, the night fall and its spoiling, is close, and that is a point in favour of a writer who knew the Sinai.  The quantity and the duration are the miracle, and no natural explanation offered reaches them.

The quail.  Exodus 16:13 says "the quail came up and covered the camp," and Numbers 11:31-33 says a wind "brought quail from the sea," that the people gathered them for two days, and that "while the meat was still between their teeth ... the Lord struck the people with a very great plague."⁷  Common quail cross Sinai each spring and autumn, low and exhausted after the sea crossing, and have been netted in huge numbers on the Egyptian and Sinai coasts for millennia.  Humphreys quotes Meinertzhagen that "in 1920 Egypt exported over three million wild quail for food."⁶  A fall of quail in the right season needs no miracle, only a wind at the right time, which is how Numbers 11:31 describes it.

The "very great plague" after the quail has no natural explanation that fits.  The common proposal is coturnism, a muscle-destroying poisoning from eating quail that have fed on certain seeds.  It is real, but the medical literature places it in the wrong season.  A 2004 case report in the Canadian Medical Association Journal says it "is seen in rural Mediterranean areas during autumn, the migration period of the European common quail," while noting that "the Old Testament mentions the group poisoning of the Jewish people during the Exodus."¹⁵  Two Greek physicians writing on Numbers 11 in 2022 drew the conclusion:  since the biblical events "took place in the spring," the birds were northbound quail that "are not responsible for coturnism."¹⁶  They proposed food poisoning from meat left in the sun, which no study tests.  The scale of the fall is also within nature's range:  hunters in North Sinai alone take about "2.0 million quail and 0.5 million birds of other species ... during the 45 days of peak migration" each autumn.¹⁷

Part 4: The walls of Jericho

The site's Jericho, Ai, and the Conquest Narratives owns this case;  only its verdict and the part that bears on the miracle are given here.  Jericho was a walled town of about eight acres that was violently destroyed, burned and then largely abandoned.  The fight is over the date.  John Garstang and later Bryant Wood put the destruction about 1400 BC, close to the early Exodus date.  Kathleen Kenyon put it about 1550.  Short-lived grain from the destruction layer, dated by Hendrik Bruins and Johannes van der Plicht in 1995, went against the early date by name:  "The 14C dates do not support a low 15th century BC chronology for the end of the MBA, as Bimson (1978) proposed.  Further, the fortified Bronze Age city at Tell es-Sultan was not destroyed by ca. 1400 BC, as Wood (1990) suggested."¹⁸  Lorenzo Nigro's current expedition dates the final Middle Bronze destruction "in the third quarter of the 16th century BC (1550-1525 BC)."¹⁹  The sibling article shows why the question reduces to the date of the Thera eruption, and why the early date is presently behind but not dead.

On the score used here:  (a) the text's date of writing is disputed, and the Deuteronomy article covers the composition of Joshua-Kings;  (b) a walled town fell violently and burned, but at a date that does not presently match either Exodus date;  (c) nothing attests the falling of the walls at a shout.

Kenyon, the archaeologist usually cited against Joshua 6, did not say the wall-fall was false.  She wrote that the excavation "has thrown no light on the walls of Jericho of which the destruction is so vividly described in the Book of Joshua ... But as to what caused the walls to fall down flat, we have no factual evidence.  We can guess that it was an earthquake, which the excavations have shown to have destroyed a number of the earlier walls, but this is only conjecture.  It would have been very natural for the Israelites to have regarded such a visitation as divine intervention on their behalf, as indeed it can be regarded."²⁰

The earthquake explanation can be priced, and the price is the timing.  Kenyon gives the rate:  in the Jordan rift "Major earthquakes are believed to be not quite so numerous now as formerly, but they still average four a century."²⁰  At four a century, the chance that a major quake falls in any given seven-day window is about 4 ÷ 100 × 7 ÷ 365, or 1 in 1,300.²¹  Joshua 6:5 puts the collapse not in a week but at the moment of the shout.⁷  So an earthquake explains how walls can fall, which nobody doubts, at the cost of a coincidence of about one in a thousand or worse.  That is the general shape of every "natural mechanism, miraculous timing" proposal in this article:  the mechanism is cheap and the timing carries the whole weight.

Part 5: Joshua's long day

The text itself says the key lines are a quotation from an older war poem.  History has nothing to test in the literal reading, the "NASA missing day" is a legend with no source that its own author could produce, and the eclipse reading is dated and testable but fits the text badly.

Joshua 10:12-14 reads:  "O sun, stand still over Gibeon;  O moon, over the Valley of Aijalon!"  Then:  "The sun stood still and the moon stood motionless while the nation took vengeance on its enemies.  The event is recorded in the Scroll of the Upright One.  The sun stood motionless in the middle of the sky and did not set for about a full day."⁷  The "Scroll of the Upright One," the Book of Jashar, is cited once more in the Bible, in 2 Samuel 1:18, as the source of David's lament "The Bow."⁷

(a) The text.  The date of Joshua is disputed, as for Kings.  But the passage points to an older written source of its own, a collection of poems, and quotes the couplet from it.  (b) The ordinary part, a battle at Gibeon and a pursuit down the Beth Horon road, is not attested outside the Bible and belongs to the conquest question.  (c) The extraordinary part has four readings.

1. The literal reading.  Russell Grigg of Creation Ministries International takes the event as a real halt of the sun's apparent motion and lists three proposed mechanisms:  "Some form of refraction (bending) of the light from the sun and the moon ... A wobble in the direction of the Earth's axis of rotation ... A slowing of the Earth's rotation."  On the last:  "God caused the rotation of the Earth to slow down so that it made one full revolution in about 48 hours rather than 24."  He states the ground plainly:  "Miracles rest on testimony, not on scientific analyses."²²  Grigg proposes no natural mechanism, so there is nothing for physics to falsify.  For scale only, since no one claims a natural cause:  halving the earth's spin and restoring it means removing and returning about 1.6 × 10²⁹ joules of rotational energy, with the oceans and atmosphere held in place.²³  Grigg accepts that this would require God to protect the earth from the effects.  History can say only that no other ancient record of a long day has been tied to this date.

2. The "NASA missing day," a legend to retire.  The story runs that NASA computers, plotting the positions of the planets for the space program, found a day missing in elapsed time, which scientists traced to Joshua and Hezekiah.  Its trail is now clear.

1. 1890. C. A. L. Totten, an army officer teaching military science at Yale, argued by hand calculation that Joshua's day lasted 23⅓ extra hours, that "These 47⅓ hours were considered as two full days by the calendar keepers of that time," and that Hezekiah's sign later set "the Calendar thus made absolutely correct."²⁴ No computer is involved.
2. 1936. Harry Rimmer retold it in The Harmony of Science and Scripture: "Sir Edwin Ball, the great British astronomer, found that twenty-four hours had been lost out of solar time."²⁵
3. 1969 to 1974. Harold Hill, a Baltimore industrialist who said his company had contracts with the NASA program,²⁶ told the story in talks with the computers added; it was printed in the Evening World of Spencer, Indiana, in October 1969. Bryant Wood of Associates for Biblical Research prints NASA's reply to inquirers: "We know nothing of Mr. Harold Hill and in no way can corroborate the 'lost day' reference in the article."²⁷ Hill's own book of 1974 tells it with the arithmetic "Ten degrees is exactly forty minutes!" and then concedes: "And my inability to furnish documentation of the 'Missing Day' incident in no way detracts from its authenticity."²⁶

Creation Ministries International lists it among arguments not to use:  "Not promoted by major creationist organizations, but an hoax in wide circulation."  Its reason is the decisive one:  "the whole story is mathematically impossible—it requires a fixed reference point before Joshua's long day."²⁸  A computer projecting the planets backward from today shows where they would be if nothing had changed.  It cannot find a gap unless it is checked against a dated observation from before the event, and none exists.

3. The eclipse reading.  In 1918 Robert Dick Wilson of Princeton read the Hebrew verb dmm, usually "be still," by an Akkadian cognate:  "it would make good sense to render the form dom in Joshua's prayer by 'become dark,' or 'be eclipsed.'"²⁹  In 2017 Colin Humphreys and Graeme Waddington of Cambridge revived the idea in Astronomy & Geophysics.  Humphreys's own summary:  "an alternative meaning could be that the sun and moon just stopped doing what they normally do:  they stopped shining," and "the only annular eclipse visible from Canaan between 1500 and 1050 BC was on 30 October 1207 BC, in the afternoon."  From it they date Ramesses II to 1276-1210 BC.³⁰

It is the only reading with a date and a place, which is its merit.  Its costs:

1. The sky. The text puts the sun "over Gibeon" and the moon "over the Valley of Aijalon," which lies to the west. John Walton pointed out in 1994, before the Cambridge paper, that "a solar eclipse cannot take place when the sun is in the east and the moon in the west."³¹
2. The time and the darkness. The text has the sun "in the middle of the sky"; the 1207 eclipse was in the afternoon. Marinus Anthony van der Sluijs, testing the proposal in 2019, notes that it was annular, not total, and that "even during total solar eclipses the sky remains five times brighter than with a full moon." His verdict: "the explanation as an eclipse, whether annular or total, is unacceptable on calendrical, philological and physical grounds alike."³²
3. The date. 1207 BC fits only the late conquest, after Merneptah's stele names Israel in Canaan, and Emil Khalisi, recomputing the eclipses of the period in 2021, prefers the total eclipse of 30 September 1131 BC.³³ Walton had already listed the eclipses visible in central Palestine between 1500 and 1000 BC, 1157, 1131 and 1041, and observed: "None of those dates correlates even closely with the period of the conquest, no matter which of the two principal dates one assigns the Exodus from Egypt."³¹
4. The duration. An eclipse shortens daylight; the text says the sun "did not set for about a full day."

4. The omen or poetic reading.  Walton, an evangelical Old Testament scholar, reads dmm as "wait" and the prayer as a request for a favourable omen, the sun and moon in a configuration that Babylonian omen texts read as a sign against an enemy:  "It is therefore a distortion of the biblical text to suggest that Joshua was requesting that the rotation of heavenly bodies actually cease."³¹  Bryant Wood of Associates for Biblical Research, a defender of the early conquest date, calls it "A much more plausible explanation for the so-called 'missing day'."²⁷  Even Grigg grants that Joshua's words "appear to be quoted from this book" and "are in poetic form," though he holds that verse 13b "reverts to Hebrew prose to describe what happened."²²

Score.  History cannot test the literal reading, and its defenders say so.  The NASA story is false and should never be repeated.  The eclipse proposal is a real hypothesis that most of its critics, Christian and secular, find strained.  What the text itself shows is that its author was quoting a poem, which leaves room for a historical battle with a non-literal couplet, a reading conservative scholars such as Walton already hold.

Part 6: Elijah, the drought and Mount Carmel

A Phoenician record of a drought under Jezebel's father exists, and it is the best non-biblical parallel for any Elijah story.  It does not match the biblical length, and it credits the rain to the other side's gods.

1 Kings 17:1 has Elijah tell Ahab "there will be no dew or rain in the years ahead unless I give the command."  The word to end it comes "in the third year of the famine."  On Mount Carmel "fire from the Lord fell from the sky" on Elijah's altar, and then "there was a heavy rainstorm."⁷  The New Testament twice gives the drought as three and a half years.⁷

(a) The text.  Ahab reigned in the mid-ninth century BC, so Kings in its final form, after 561 BC, is about 300 years later.  Ahab himself is attested:  Shalmaneser III's Kurkh Monolith lists his 2,000 chariots at the battle of Qarqar in 853 BC, covered in the site's archaeological confirmations.  1 Kings 16:31 says Ahab married "Jezebel the daughter of King Ethbaal of the Sidonians."⁷

(b) The ordinary part, and the one outside witness.  Josephus quotes Menander of Ephesus, who, Josephus says, "translated the Tyrian records from the Phoenician language into Greek speech."³⁴  On the reign of Ithobalos, the Greek form of Ethbaal, Menander wrote:  "There was a drought in his reign, which lasted from the month of Hyperberetaios until the month of Hyperberetaios in the following year.  But he made supplication to the gods, whereupon a heavy thunderstorm broke out."  Josephus adds:  "This, then, is what Menander wrote, referring to the drought which came in Achab's reign, for it was in his time that Ithobalos was king of Tyre."³⁴

Three cautions apply, and a skeptic will raise all of them.

1. The link to Elijah is Josephus's, not Menander's. Menander names no Israelite. What he attests is a drought in Phoenicia under the king the Bible makes Ahab's father-in-law.
2. The lengths differ. Menander's drought runs one year, from one Hyperberetaios to the next. 1 Kings 18:1 has the word to end it "in the third year," which could fit a drought of a little over a year that touched three calendar years. Luke 4:25 and James 5:17 have "three and a half years," which cannot be squeezed into Menander's twelve months. Either Menander's annal records only the worst year of a longer dry spell on the coast, or the New Testament uses a round figure for a time of judgment. Neither answer is shown by evidence.
3. Menander survives only through Josephus, so it is one chain of copying, and its date is not known closely.

The Phoenician version has a feature that tells against invention by Josephus:  it ends the drought with a thunderstorm after Ithobalos "made supplication to the gods."  A Jewish writer making up support for Elijah would not have handed the rain to Baal's king.  It reads like a Tyrian court record of the same dry spell seen from the other altar.

(c) The extraordinary part.  Nothing outside 1 Kings attests fire falling on the altar, and nothing could be expected to.  Lightning is the natural explanation sometimes offered, and the story itself ends in a storm.  But the text describes fire that "consumed the offering, the wood, the stones, and the dirt, and licked up the water in the trench," after Elijah prayed "When it was time for the evening offering,"⁷ and no historical method can test that.  The score is a real drought, a real king and a real father-in-law, with a length that only partly fits, and a miracle that history cannot reach.

Part 7: Sennacherib's army before Jerusalem

This is the strongest case in the register for a historical core, and the enemy's own records supply most of it.  The full case is in Why Did Sennacherib Fail to Take Jerusalem?.  Its findings, by the three questions of Part 1:

1. The text. The Chicago Prism copies of Sennacherib's annals are dated 691 and 689 BC, ten to twelve years after the campaign,³⁵ and an earlier edition, the Rassam Cylinder, dates to spring 700 BC. Kings in its final form is after 561 BC, about 140 years later, and cites older royal annals.⁷
2. The ordinary part. Almost all of it is attested by the other side. The prism names Hezekiah, claims 46 of his cities, and says "I confined him inside the city Jerusalem, his royal city, like a bird in a cage."³⁵ It never claims the city fell. The gold of the tribute, 30 talents, matches 2 Kings 18:14. Lachish's siege is carved on Sennacherib's palace walls and survives as a siege ramp. The Babylonian Chronicle records that Sennacherib "was killed by his son" in 681 BC,³⁶ and the son's name very likely lies behind the biblical Adrammelech.³⁷
3. The extraordinary part. Something unusual is attested: of ten resisting rulers or ruling groups in Sennacherib's annals, Hezekiah alone kept his throne in his own city. Herodotus preserves a separate Egyptian story of a night disaster to the army.³⁸ The 185,000 dead have no outside support, are half again the largest Assyrian muster on record,³⁹ and cannot be a natural epidemic in one night, since even plague takes "one to seven days" to incubate.⁴⁰ "Berossus confirms the plague" rests on a quotation that has been lost from Josephus.⁴¹

A historian can say that Sennacherib devastated Judah, blockaded Jerusalem, did not take it, left Hezekiah on the throne, received his tribute, and was killed by a son twenty years later.  The cause of the withdrawal is beyond the evidence.

Part 8: Jonah and the fish

Jonah is the case where history has least to work with.  The setting is real, the size of Nineveh fits the book better than its critics allow, the fish is beyond any test, and the one modern "confirmation" Christians have repeated for a century, the swallowing of James Bartley, is a sailor's yarn that should be retired.

The book says the Lord "sent a huge fish to swallow Jonah, and Jonah was in the stomach of the fish three days and three nights," that Nineveh "was an enormous city—it required three days to walk through it," and that the city repented at his preaching.⁷  Jesus cites both the fish and the repentance in Matthew 12:40-41.

(a) The text.  The book carries no date.  2 Kings 14:25 names "Jonah son of Amittai, the prophet from Gath Hepher" as a prophet under Jeroboam II, in the first half of the eighth century BC.⁷  So a prophet of that name is placed in history by another book, but the book of Jonah itself cannot be dated closely, and the gap between event and text is unknown.

(b) The ordinary part.  Nineveh was one of the largest cities of the ancient world.  A. Leo Oppenheim of the Oriental Institute gives its occupied area as "1,850 acres," second only to Babylon.⁴²  That is about 750 hectares.  The objection usually raised is Jonah 3:3's "three days' journey."  D. J. Wiseman, an Assyriologist, tested the size readings in 1979.  Read as a diameter or circuit, "even at the modest designation of a day's walk at 17 miles per day this would be a distance of over 50 miles," while "The maximum axis of Nineveh at any period was much less than three miles."  The old answer, that three days meant the 7½-mile circuit of Sennacherib's walls, is one Wiseman had himself revived and now judged he had done "probably wrongly."⁴³  His alternative reads the phrase as a visit's protocol:  "the day of arrival in the city, followed by the customary day of visiting, business and rest, then the day of departure."⁴³  On that reading the verse makes no claim about the city's width, and Jonah 3:4, where Jonah begins with "one day's walk" into the city, describes a preaching tour rather than a crossing.  The NET's own rendering, "it required three days to walk through it," takes the size reading, so the translation a reader picks decides whether there is a problem.

The older defence fails on its own sources.  Diodorus Siculus gives Nineveh a circuit of "four hundred and eighty stades," about 55 miles, but in the same passage he says Ninus "founded on the Euphrates river a city," when Nineveh stands on the Tigris.⁴⁴  A source that misplaces the river does not confirm the walls.

(c) The extraordinary part, and the legend to drop.  For a century the standard proof that a man can survive inside a whale was the story of James Bartley, swallowed by a sperm whale off the Falkland Islands in February 1891 and cut out alive a day and a half later.  Edward Davis, a historian of science at Messiah College, traced it in 1991 for Perspectives on Science and Christian Faith.⁴⁵  The trail runs like this.

1. 1891. The earliest printing Davis found is a newspaper report "published on 22 August 1891" in the Yarmouth Mercury of Great Yarmouth, England. The New York Times reprinted it on 22 November 1896, attributing it to a nonexistent "South Yarmouth Mercury."⁴⁵
2. 1906. A. Lukyn Williams asked in The Expository Times for "the name of the ship and the exact date and place of the occurrence." Eduard König, who had cited the story in Hastings's Dictionary of the Bible, could name only "English newspapers, whose account was reproduced in the Canadian Aurora," and added: "I myself should be interested if the source and the certainty of the above narrative could be established."⁴⁶
3. 1906 to 1907. Lloyd's identified the ship as the Star of the East, which "sailed from Auckland, N.Z., on the 27th December 1890 for New York." She was a merchant ship on passage, not a whaler. The captain's wife wrote from Nova Scotia: "There is not one word of truth in the whale story. I was with my husband all the years he was in the Star of the East. There was never a man lost overboard while my husband was in her. The sailor has told a great sea yarn."⁴⁷
4. 1927 to 1928. Ambrose John Wilson revived the story in the Princeton Theological Review, assuring readers that "the whole story was carefully investigated."⁴⁸ When critics cited the captain's wife, he answered that "denial is not the same as disproof," and described her letter as amounting to "no more than the assertion that she 'had never heard of the story.'"⁴⁹ Her printed letter says something much stronger.
5. 1991. Davis obtained the ship's crew agreement: "there is no James Bartley on the list, nor anyone of similar name, either for the entire voyage or any part thereof!" An archivist told him "whaling in the Falkland Islands did not commence until 1909." His conclusion: "there never had been a whale, and that all of this was no more than a fish story, albeit a dandy."⁴⁵

Fairness requires one concession.  The ship's records put the Star of the East at sea in February 1891, and Davis notes that this is "not inconsistent with the claim that she was off the Falkland Islands."⁴⁵  So the story was built around a real ship on a real passage.  That is what makes it convincing, and it is still false:  no crewman of that name, no whaling, and a flat denial from the captain's household.

With Bartley gone, nothing outside the book bears on the fish.  A historian can say only that no human being is known to have survived three days inside any sea creature.

Genre is the real question for Jonah, and both sides have a case.  Julius Bewer's 1912 commentary in the International Critical Commentary series read Jonah as "a story with a moral, a parable, a prose poem," and wrote of its strange details:  "Surely this is not the record of actual historical events nor was it ever intended as such."⁵⁰  T. Desmond Alexander granted in 1985 that "most recent writers prefer to classify Jonah as either parable or didactic fiction," then argued from the book's form that its author "intended his work to be read as a didactic history, rather than as a didactic fiction."⁵¹  Alexander's point is that improbability is not evidence of genre:  the opening, "The Lord's message came to Jonah son of Amittai,"⁷ is the formula of the prophetic narratives in Kings.  The Christian reader also has Matthew 12:41, where "The people of Nineveh will stand up at the judgment."⁷  That reads more naturally of real people than of characters in a parable, though Bewer answered that "an illustration may be drawn from fiction as well as from actual history."⁵⁰  This dispute is about what kind of book Jonah is, and history as a method cannot settle it.

Rodger Young has suggested that the solar eclipse of 15 June 763 BC, recorded in the Assyrian eponym list, could explain why Nineveh was ready to repent.  He labels it himself:  "Such a correlation of Jonah's ministry with the Bur-Sagale eclipse and other events in Assyrian history is speculative."⁵²  No ancient text links the two.

Part 9: Hezekiah's shadow and the floating axe head

Hezekiah's sign.  When Hezekiah asked for a sign of his recovery, Isaiah offered to move a shadow, and "the Lord made the shadow go back ten steps on the stairs of Ahaz."⁷  2 Chronicles 32:31 adds that Babylonian envoys came "to inquire about the sign that occurred in the land."  The Hebrew word, ma'alot, means steps or stairs;  the Great Isaiah Scroll from Qumran, as Martin Abegg, Peter Flint and Eugene Ulrich translate it, speaks of "the steps of the upper dial of Ahaz."⁵³

(b) The ordinary part has a real anchor.  2 Kings 20:12 says that "Merodach Baladan son of Baladan, king of Babylon, sent messengers with letters and a gift to Hezekiah, for he had heard that Hezekiah was ill."  Merodach-baladan (Marduk-apla-iddina II) is well known from Assyrian and Babylonian records as Assyria's chief rival for Babylon in these years, and Cogan notes that his "get-well wishes" may have had "wider, conspiratorial connotations."⁵⁴  A Babylonian rebel courting Judah fits 703-701 BC exactly.

(c) The extraordinary part.  Two natural proposals exist.  J. W. Bosanquet argued in 1855 that the sign was a partial solar eclipse, since a large eclipse of the sun's upper edge near noon would make a shadow on steps move back:  "We look, therefore, for a large partial solar eclipse, on the upper limb, visible at Jerusalem, about twelve o'clock, on some day in the year B.C. 689."  The Astronomer Royal, George Airy, computed for him that the eclipse of 11 January 689 BC "was annular, and the annulus was rather broad," and its computed time fell about three quarters of an hour earlier than the story needs, a gap Bosanquet hoped new calculations would close.⁵⁵  The date also depends on placing Hezekiah's illness in 689.  Paul Dunbavin, checking the eclipse paths in 2023 against what he calls "the widely-accepted date of 702 BC," found that the computed paths offer "no suitable eclipse that fits the requirement."⁵⁶  Creation Ministries International prefers refraction:  "Some form of light refraction appears to have been what happened in the reign of Hezekiah."²²  No one has published the calculation of the refraction needed to move a shadow ten steps.

Harold Hill's arithmetic, "Ten degrees is exactly forty minutes," rests on a translation the Hebrew does not support.  The text says ten steps, and no one knows how tall the steps were or how much time each marked.

The axe head that floated for Elisha (2 Kings 6:5-7) has an ordinary setting, a borrowed iron tool lost in the Jordan, and nothing a historian could test.  It is included here only to say so.  Many of the shorter miracle stories in Kings are of this kind:  their truth stands or falls with the reliability of the narratives around them, not with any evidence of their own.

Part 10: The register

The table scores each case on the three questions of Part 1.  "Text distance" is the gap between the event and the written account as far as it can be measured.  "Cost" is what the best natural explanation has to assume.

Objections and responses

"Historians cannot study miracles at all, so this whole register is a category mistake."  The register does not ask a historian to certify a miracle.  It splits each story into the parts a historian can test:  how close the text is, whether the ordinary event happened, and whether anything unusual is independently attested.  Those are ordinary historical questions, and the answers differ sharply from case to case, which a blanket rule would hide.  Sennacherib's withdrawal and the Bartley whale are not the same kind of evidence, and a method that cannot tell them apart is not careful.

"An earthquake at Jericho, a wind at the sea, a honeydew in Sinai:  science has explained these, so they were not miracles."  The proposals are real and some are good physics.  But in each case the mechanism is the cheap part and the timing is the expensive part.  A major quake in the Jordan rift comes about four times a century, so one in the week of the siege is a chance of about 1 in 1,300, and one at the hour of the shout far less.²¹  Humphreys, who holds that the events were natural, makes this point himself:  for him the miracle is the timing.  Drews and Han show that a 100 km/h east wind blowing for twelve hours can open a crossing;  they do not show why it blew that night.  Explaining how something could happen does not explain why it happened when the text says it did.

"Natural explanations show the stories had a real basis, so they confirm the Bible."  Only if the explanation is evidence and not a guess.  Bodenheimer's manna is observed;  the Lake of Tanis sandbar is reconstructed and its authors call it the weakest part of their model;  Humphreys's algal bloom in the delta has never been looked for.  And several natural accounts cost the text something:  Humphreys's 20,000 Israelites require reading every census number in Numbers differently, and Bodenheimer's manna requires calling the melting "a late and mistaken interpolation."  A defender who welcomes the natural account has to accept its price.

Objections about Sennacherib's campaign (the silence of the Assyrian annals, "Berossus confirms the plague," Tirhakah and the two-campaign theory) are answered in Why Did Sennacherib Fail to Take Jerusalem?.

"NASA's computers found Joshua's missing day."  No such computation exists, and the story's own source admits it.  Harold Hill wrote that "my inability to furnish documentation of the 'Missing Day' incident in no way detracts from its authenticity."²⁶  The arithmetic comes from C. A. L. Totten's 1890 book, which used no computer.  More basically, a calculation of planetary positions backward from the present cannot reveal a lost day unless it is checked against a dated ancient observation, and none was cited.  The legend should be dropped.

"A man named James Bartley survived a day and a half in a whale in 1891."  The captain's wife wrote in 1906 that "There is not one word of truth in the whale story," the crew agreement has no Bartley, and no whaling was done at the Falklands until 1909.  This should be dropped.

"Jonah 3:3 says Nineveh took three days to cross, and no city was that big, so the book is fiction."  The size reading is one reading of a Hebrew phrase, not the only one.  D. J. Wiseman, an Assyriologist, read it as the three days of an official visit, and he gave up his own earlier attempt to make it a measurement.  On his reading the verse says nothing about width.  The genre question is real, but this verse does not settle it.

"Menander's drought proves Elijah's."  It proves a drought in Phoenicia under Ethbaal lasting a year.  The Bible's "third year" might stretch over that, but the New Testament's three and a half years cannot.  It is a real parallel with a real mismatch.

"The Bible's numbers are so large that they discredit the stories."  Some are.  The 185,000 is larger than any Assyrian muster on record, and two million people eating manna for forty years has no natural counterpart.  But large numbers in ancient texts are a known problem with known proposed solutions, including the reading of 'elep as a military unit, and a large number does not show that the event behind it did not happen.

Issues and unanswered questions

1. Humphreys and Waddington's 2017 paper and their 2019 Vetus Testamentum article were not read. Both are behind publisher challenges. Their position is stated here from Humphreys's own words in the Cambridge announcement and from their critics. They may answer Walton's sky-position objection; the article should be checked against their text before it is published.
2. Several primary sources on the plagues and the sea were read only through others. Greta Hort's 1957-1958 papers, the full model of Nof and Paldor (1992), Hans Goedicke's Thera proposals and Siro Trevisanato's book were not obtained. Humphreys and Sivertsen summarise them.
3. The date of composition of the Ipuwer poem rests here on the manuscript date only. Roland Enmarch's A World Upturned (2008) argues the composition date and was not read.
4. Simo Parpola's 1980 paper on Sennacherib's murderer is known only through Knapp (2020).
5. No spring count of quail crossing Sinai was found; every count is of the autumn passage. The coturnism argument depends on the season, so a spring case of quail poisoning would change it.
6. The 'elep reading of the census numbers is mentioned twice (Humphreys's 20,000 Israelites and the 185,000 Assyrians) but not tested here. Its strengths and costs belong to an article on the Exodus numbers.
7. The date of Kings is taken from its last event (561 BC). Earlier editions under Hezekiah or Josiah are widely proposed and would shorten the text distance for Elijah and Sennacherib; the Deuteronomy article covers that debate.
8. The Hebrew of the Great Isaiah Scroll at Isaiah 38:8 and the NET translators' notes on Joshua 10:12-13 were not checked; the readings "upper dial" and "be still" rest on the translations cited.
9. Michael Licona's 2010 treatment of methodological naturalism was not obtained. Ehrman and Lüdemann represent the two sides of the method question here; a defender of the historian's right to conclude a miracle in his own words would strengthen Part 1.
10. The 185,000 dead have no archaeological test. No mass grave of the period near Jerusalem has been identified, and none would be easy to tie to 701 BC if found.

Verdict

History can affirm the ordinary core of several of these stories, can point to something unusual in two of them, and cannot reach the miraculous part of any.

1. Ordinary core affirmed. Sennacherib's invasion of Judah, his blockade of Jerusalem, his failure to take it, Hezekiah's tribute and the murder of Sennacherib by a son are attested by Assyrian and Babylonian records, some written within twelve years. Ahab, Ethbaal and a drought in Ethbaal's reign are attested. Jericho was a walled city that was violently destroyed. Nineveh was the great city the book of Jonah says it was, and the "three days" need not mean its width. The manna of Exodus matches a real Sinai substance.
2. Independent hints of something unusual. The Assyrians left the leader of a rebellion on his throne, against their own practice, and Herodotus preserves a separate Egyptian memory of a night disaster to Sennacherib's army. Menander's Tyrian record preserves a drought under Jezebel's father. Neither confirms a miracle; both confirm that something happened at the place the Bible says.
3. Mismatches to admit. Jericho's destruction is dated a century or more before the early Exodus date and three centuries before the late one. Menander's drought is one year, not three and a half. The 185,000 dead exceed any Assyrian muster. The supply of natural manna falls short by thousands of times.
4. Legends to drop. The NASA "missing day," the swallowing of James Bartley, "Berossus confirms the plague," and the Ipuwer papyrus as a record of the plagues. Each fails on its own sources, and each has been repeated by Christians for decades. Creation Ministries International already rejects two of them.
5. What history cannot say. Every natural explanation examined here works only with lucky timing: a quake at the shout, a gale on the night of escape that stops at dawn, ten disasters in sequence on cue. The timing is the part the Bible calls the act of God, and no historical method can measure it. Ehrman is right that history cannot certify a miracle. It does not follow that it can rule one out. What it can show is whether the story is set in real places, among real kings, at a time that fits, and on that test these stories range from the well attested (Sennacherib) to the untestable (Jonah's fish), not to the uniformly legendary.

Sources