Forgery, Provenance, and the Antiquities Market

"The best material analysis can do, after all appropriate tests have been conducted, is to announce that nothing has been found to contradict the assumption of genuineness." -- Kipp Davis, Ira Rabin and colleagues, 2017¹

"This is not to say that the inscription on the ossuary is real and authentic and that was written two thousand years ago." -- Judge Aharon Farkash, 2012, in a defence-side translation²

Summary

Since about 1980 a handful of inscribed objects bought on the antiquities market have been announced as the first physical links to Solomon's Temple, to the kings of Judah, to Jesus' family, or to the Dead Sea Scrolls.  Most of them have since been accused of being modern forgeries, and some have been proved to be.  This article asks which of them a Christian can responsibly use in a historical argument, how specialists test such objects, and what remains genuinely uncertain.  It treats physical objects only.  Literary pseudepigraphy, writing under another person's name, raises different questions.  Whether particular letters attributed to Paul belong in that category is examined in Did Paul Write the Disputed Letters?.

1. None of these market finds supplies secure independent evidence for a biblical claim.  The six case studies below include groups of fragments and two ostraca, not just six individual objects.  The Museum of the Bible and Schoyen collections form one linked manuscript-market case.  The assessments differ:

   1. The 16 Museum of the Bible "Dead Sea Scroll" fragments are, on the convergent physical evidence, very probably modern writing on old leather. Their ink runs across cracks, off broken edges and over burial deposits that were already there, and loose minerals were scattered onto it while it was wet. Nine more fragments of the same post-2002 kind in the Schøyen Collection were judged forged, two with high confidence and seven as "most likely."¹³³
   2. The Gospel of Jesus' Wife is very probably a forgery on old papyrus. A companion fragment from the same owner, which Christian Askeland judged to be in the same hand, copies the line breaks and modern restorations of a printed edition of 1924. Its owner's provenance papers fail ordinary document checks, and its own editor, Karen King, said in 2016 that the evidence "presses in the direction of forgery."⁴⁴⁵
   3. The two Moussaieff ostraca, the "Three Shekels" and "Widow's Plea" texts, are very probably forged. An Israel Antiquities Authority (IAA) examination found a lime crust lying over a wax film, which lay over fresh tool scratches in the letters. Ten excavated ostraca from six sites showed none of these layers.⁶⁶⁶
   4. The Jehoash tablet is probably forged, but the case is cumulative and contested. Frank Moore Cross, Israel Eph'al and Edward Greenstein list many expressions they judge post-biblical, modern or misunderstood borrowings from Kings and Chronicles. The philologist Chaim Cohen answered several of them item by item. He did not claim the tablet is genuine, only that it "can not be proven philologically to be a modern-day forgery." The physical evidence is disputed between laboratories.⁷⁸⁹¹⁰
   5. The ivory pomegranate is an ancient object, probably of the Late Bronze Age, but whether its inscription is ancient is unresolved. A joint examination in 2007 settled two side points: the committee withdrew its objection to one letter, and Lemaire says it conceded an error about the word spaces. It left the central question split: whether the strokes of two letters run into an ancient break.¹¹¹¹¹¹
   6. The James ossuary is an ancient box. Whether all of its inscription is ancient is unresolved. Its letter coatings differ from those on three excavated ossuaries; counterexamples from pottery challenge a wider claim about all natural crusts but are less closely matched controls. Inside the letters of "Jesus" the evidence is mixed. Even if the whole inscription is ancient, identifying its Jesus with Jesus of Nazareth is a separate, probabilistic step.^1213132

2. Old material does not by itself date new writing.  Radiocarbon dated the Gospel of Jesus' Wife papyrus to the early Middle Ages, and the Museum of the Bible leather was probably ancient too, yet both very probably carried modern writing.¹⁴⁴³  What decided these cases was evidence that ties the writing to an event:  how layers relate (ink over damage, wax under a crust, a letter cut by or stopping at a break) and texts that copy a modern source.  The ink and papyrus tests on the Gospel of Jesus' Wife found the papyrus old and the ink a carbon black like ancient inks.  Neither result dates the writing.  A later Raman method can sometimes show that an ink does not match ancient inks of a claimed date, but its authors say it "cannot positively authenticate" any manuscript.¹⁵¹⁶

3. Courts and committees decide different questions from laboratories.  In 2012 a Jerusalem court acquitted the collector Oded Golan and the dealer Robert Deutsch of forgery, because the prosecution had not proved it beyond reasonable doubt.  In the defence side's translation, the judge said in so many words that this did not make the ossuary inscription authentic.  He did not say it was probably forged either.²¹⁷  The IAA's 2003 committees reached a unanimous finding of forgery within three months.  The same translation reports that at least one member had first judged the ossuary inscription authentic and changed his view after hearing the others.¹⁸²  A verdict and a committee report are evidence about what qualified people concluded under their own standards.  The weight of either depends on the observations behind it, and those are weighed object by object below.

4. "Unprovenanced" does not mean "forged," and "excavated" does not end all argument.  The Mesha stele was never excavated.  But the missionary F. A. Klein saw it lying intact at Dhiban on 19 August 1868, measured and sketched it, and copied an alphabet from it.  A paper squeeze of the text was taken before local men broke the stone, and the fragments recovered later were fitted back by their lines of letters.¹⁹²⁰  That chain fixes the stone's modern history.  It does not by itself prove the inscription ancient, which rests on its script and language.  The Ketef Hinnom silver amulets were excavated, yet their date was argued from the seventh century BC to the Hellenistic period, until the excavators clarified the find context and new imaging sharpened the script.²¹²¹

5. A reliable dealer or collection does not certify its next object.  The Kando family sold what the scholars who later exposed the forgeries call "probably authentic" fragments to the Schøyen Collection between 1999 and 2003.  The same family was the first seller of most of the post-2002 fragments now judged forged.¹²²

6. These cases cannot give a forgery rate.  They are the objects that became famous and were then challenged, which is a biased sample.  They show what tests can and cannot do, not what fraction of museum objects is genuine.

7. The loss to Christian apologetics is real but specific.  The ossuary was offered as the earliest physical object naming James and Jesus together.  If its inscription is modern, or names other men, that proposed independent corroboration is lost.  Josephus's statement that James was "the brother of Jesus who was called the Christ" does not depend on it,²³ and is weighed in Jesus in Archaeology and Outside Sources.  The Jehoash tablet and the ostraca would each have been unique Judean texts bearing on the First Temple.  The excavated Dead Sea Scrolls are untouched by the Museum of the Bible fakes.  The Gospel of Jesus' Wife was never evidence about the historical Jesus even on its editor's reading.  What it was proposed to show was an early Christian debate about wives and mothers as disciples.²⁴  The general lesson is to give any single unprovenanced object weight only in proportion to what tests have shown about its writing.

Part 1: Six questions that must be kept apart

Every inscribed object raises at least six different questions.  Most public arguments about forgery go wrong by answering one of them and treating it as an answer to the others.

1. Is the object itself old? The box, the stone, the sherd, the scrap of leather or papyrus.
2. Is the writing on it as old as the object, or at least ancient? A forger's cheapest material is a genuine old surface with nothing written on it.
3. Where did it come from, and who has owned it? This is provenance: the findspot, and the documented chain of possession since.
4. What does it say? Readings of damaged letters can differ even when everyone agrees the writing is ancient.
5. Who or what is it about? A genuine ossuary of a man named James, son of Joseph, with a brother named Jesus, need not belong to the James of the New Testament.
6. Is what it says true, and is it independent of the Bible? An authentic ancient statement can be propaganda, error or a later retelling of a biblical story.

The answers come in degrees, and each limits what the next can carry.  An object whose writing is probably ancient can support a historical claim with weight discounted by that "probably," and by any doubt about its reading and identification.  An object whose writing is more likely modern than ancient adds little or nothing, and one shown to copy a modern edition adds nothing at all.  The six market objects in this article are all in doubt at question 2, to different degrees.  Two comparison cases show that the provenance question, question 3, works in both directions.

An excavated object can still be disputed: Ketef Hinnom

In 1979 Gabriel Barkay excavated a burial cave on the slope of the Hinnom Valley in Jerusalem and found two tiny rolled silver plaques that carry words close to the priestly blessing of Numbers 6:24-26.  Here questions 1 and 3 are as secure as archaeology allows:  the objects were found by a professional excavator in a recorded spot.  Yet their date was still argued.  The first edition put them in the seventh century BC;  Ada Yardeni dated them to the sixth;  and in 1995 Johannes Renz argued that they "are better located chronologically in the Hellenistic period," centuries later.²¹

The excavation did not settle this by itself, because a burial repository is not a layered town mound.  Families pushed old bones and gifts to the back as new burials came in.  The 2004 re-edition argued that Renz had "misunderstood the 'stratigraphy' of a burial repository":  the few Hellenistic sherds lay at the very top near the opening, while the first amulet "was found in situ in Square D in the middle of the repository, 7 cm above the floor."²¹  New photographs taken in 1994 and computer imaging then made previously unreadable letters legible.  The editors concluded, with stated caution, that "We cannot be overly precise in dating the inscriptions," but placed them at "the end of the Judaean monarchy," the late seventh or early sixth century BC.²¹

The lesson is that excavation answers the provenance question and narrows the date, but the date of the writing still rests on reading the context and the script.  "Excavated" is the strongest provenance there is, not a guarantee against argument.

An unexcavated object can still have a secure history: the Mesha stele

The Mesha stele, a black basalt monument of Mesha king of Moab, was never excavated.  The missionary F. A. Klein described his own visit in a letter printed in 1870.  "It was on the 19th of August, 1868," he wrote, that Bedouin hosts took him to the ruins of Dhiban, ancient Dibon, east of the Dead Sea.  There the stone "was lying among the ruins of Diban perfectly free and exposed to view, the inscription uppermost."  He measured and sketched it, copied a few words and an alphabet, and reported it intact.¹⁹¹⁹  Rival offers from the Prussian and French consulates followed.  Charles Clermont-Ganneau of the French consulate had an Arab messenger take a paper squeeze, a mould made by pressing wet paper onto the letters.  In a quarrel the squeeze was snatched off the stone still wet and "torn into seven pieces."  Soon afterwards the stone itself "was broken into pieces," by heating and dousing it according to later accounts.²⁰²⁵

Several independent records fix the stone's modern history.  A European saw, measured and drew it in place in 1868.  A squeeze of the whole face existed before the breakage, though torn.  By April 1870 Clermont-Ganneau had bought twenty pieces carrying 613 of an estimated thousand letters.  He reported that each piece "can be put in place easily" by lining up the rows and columns of letters on the squeeze.²⁰  A fragment that fits both a break and the lines of text on the squeeze is a physical join, and joins of that kind are much harder to fake than a plausible-looking text.

What this chain proves, and what it does not, matters for the rule being tested.  It shows that the text read today is the text on a stone lying at Dhiban in August 1868, recorded in the field before it entered any trade.  Any forgery would have had to be carved and left in the desert before that date.  The chain does not by itself show that the inscription is three thousand years old.  That rests on internal evidence, the script, language and content, judged as for any other inscription.  So the documented chain removes the modern market from the story, and the inscription's antiquity is then argued on its own merits.

So "unprovenanced" is not one thing.  A dug object, an object seen in place and documented before it moved, and an object that appears in a dealer's shop with a story are three different evidential situations.  The six cases below are all of the third kind.  That alone does not make them forgeries, but it means the burden falls on the object itself, and on tests that can reach question 2.

Excavated finds on the same subjects

The market cases below were offered as evidence about a Judean king's building work, the kings of Judah, the temple and its priests, and the family of Jesus.  Excavated inscriptions bear on the same subjects, and their findspots were recorded by the people who found them.  None of the following passed through a dealer before it was described in place.

1. The Siloam tunnel inscription. In 1880 a pupil of the Jerusalem architect Conrad Schick fell into the water of the tunnel that carries the Gihon spring to the Pool of Siloam. "On rising to the surface, he discovered some marks like letters on the wall of rock." Schick found the letters dimmed by "a deposit of silicate that has gradually covered them in course of time."²⁶ In 1996 John Rogerson and Philip Davies asked whether Hezekiah built the tunnel at all; the team that later tested it describes their claim as putting its construction in the second century BC. In 2003 the geologists Amos Frumkin, Aryeh Shimron and Jeff Rosenbaum reported radiocarbon and uranium-thorium dates from the tunnel "proving its Iron Age II date," the period of the Judean kings.²⁷ Those dates belong to the tunnel, not directly to the carving, and the text names no king. As Gary Rendsburg and William Schniedewind put it, one would expect it to name "its presumed royal sponsor, King Hezekiah of Judah," "but this is not the case."²⁸
2. The Tel Dan stele. On 21 July 1993 Gila Cook, the surveyor of Avraham Biran's expedition, saw letters on a basalt fragment "while it was still in situ." It lay "in secondary use in a wall" at the edge of the paved square outside the Israelite city gate, and "the destruction level covering it," which Biran dated to the Assyrian conquest of 733/732 BC, gives the latest date for that reuse.²⁹ Its Aramaic text mentions a "king of Israel" and the "House of David," which Biran called the first time the phrase "has been found outside the Bible." He also warned that "any reconstruction must remain tentative."²⁹ How the broken lines should be restored is argued; where the stone came from is not.
3. The Caiaphas family tomb. In December 1990 park construction in Jerusalem's Peace Forest broke through the roof of a rock-cut tomb, and "The construction superintendent reported the find to the Antiquities Authority." The archaeologist Zvi Greenhut excavated it. Five of its twelve ossuaries were inscribed, two of them with forms of the name Caiaphas, "names never before found in ossuary inscriptions."³⁰ Whether the man in the finest box was the high priest of the Gospels is a question of identification, weighed in Jesus in Archaeology and Outside Sources. Where the box was found is not in question.
4. The Pilate inscription. The Latin stone naming Pontius Pilate was found in 1961 in the theatre of Caesarea Maritima, where it had been reused in late antiquity. The epigrapher Géza Alföldy calls the restorations in its second and third lines, which give Pilate's name and his office over Judaea, certain and disputed by no one. The rest of its reconstruction and meaning, he says, is still hotly disputed.³¹

These finds have their own debates about reading, restoration and identification, just as Ketef Hinnom had.  What they lack is the gap that makes forgery a live question for the market cases:  an unknown history between the ground and the first scholar to see the writing.  An argument that the market forgeries discredit biblical archaeology as a whole has to explain why the objects in dispute here all came through the antiquities market or private collections, while excavated inscriptions on the same subjects did not.

Part 2: What each kind of test can and cannot show

One distinction runs through every case.  A measurement describes the sample.  Any conclusion about an event, such as when a scribe wrote or a chisel cut, needs a further link, and that link has to be argued.  Radiocarbon estimates when the sampled organic material acquired its carbon from the environment.  For a short-lived papyrus plant this is close to harvest;  carbon in an old tree ring can predate the tree's cutting or burning by centuries.  It dates writing only if the dated carbon belongs to the ink itself or there is good reason to connect the writing with the material's age.  An undisturbed crust directly over an incised stroke formed after that stroke, whether the crust is natural or artificial.  It gives a minimum age for the carving only if the crust's own age and original position can be established.  These links can be strong, but they must be tested case by case.  The materials scientist Ira Rabin and her co-authors put the limit of material tests plainly:  "material analysis alone ... cannot offer scientific proof that the object is genuine.  The best material analysis can do ... is to announce that nothing has been found to contradict the assumption of genuineness."¹

Three points in the table need more explanation, because each was fought over in court or in print.

Layer relationships are the strongest physical tests, because they do not need a date.  If ink runs into a crack, the crack was there first.  If a crust of salts lies over the ink with no break, the ink was there first.  If a letter's stroke is cut off by an old, weathered break, the letter is older than the break.  This kind of argument needs only a microscope, good photographs and the honesty to report what is seen.  By itself it gives only an order of events.  It becomes evidence of modern work when one of the layers can be tied to a time:  burial deposits that take centuries to form, a material such as paraffin or synthetic glue, or ink laid on a scrap already broken to its present shape.  Its weaknesses are that two observers can disagree about what is seen, as the pomegranate case shows, and that cleaning, restoration and later handling add new layers that are easy to misread.  The same kind of evidence can also count for an object.  In 2011 Yuval Goren and Boaz Zissu judged an unprovenanced ossuary authentic because its patination "in and around the inscription" showed "a complex process that occurred over a prolonged sequence of time, which is extremely difficult, if not impossible, to replicate in laboratory conditions."³²  That goes further than the limit Rabin and her co-authors state, and it rests on the examiners' reading of the surface, but it shows that layer evidence is not only a tool for condemning objects.

Patina is not a clock.  The ostraca study itself warned that "patination by itself is a somewhat unreliable indicator of antiquity, since patina-like coating can be created in the laboratory by various methods."⁶  The isotope method used against the ossuary, the Jehoash tablet and the ostraca compares the oxygen isotopes of the crust over the letters with a range of values for crusts that formed naturally.  Its authors described it as a method that "has never been applied for authenticity tests of archaeological artifacts."¹³  Their range came mainly from cave deposits near Jerusalem.  For the ostraca they stated it as -6.5 to -3.5 per mil for crusts formed "on the surface or in shallow burials" across the Judaean Hills, Samaria, Galilee and the northern Negev.⁶

At the trial a defence expert, Aldo Shemesh of the Weizmann Institute, tested that range directly.  As the judge summarised his evidence, Shemesh scraped crusts from 56 well-documented objects from official excavations across Israel, mostly pottery.  "Most" gave values between -2 and -6.2 per mil, close to the prosecution's range.  But crusts from Tel Hazor ran from -4 to -12 and from Tel Gat from -3 to -12.5.²  In 2013 three of the defence-side scientists, Joel Kronfeld, Amnon Rosenfeld and Howard Feldman, reprinted Shemesh's 56 values as a chart with every item labelled.  They described the items as "56 well-documented artifacts obtained from formally sanctioned archaeological excavations," gave their range as -3.07 to -22.46 per mil, and concluded that the isotope test "is not a useful tool in the authentication of artifacts."³³³³  How much all this matters depends on what is being compared.

1. For pottery from tells, the class the ostraca belong to, Shemesh's samples are the right comparison, and Hazor in Galilee lies inside the region the range was claimed to cover. They show that natural crusts on excavated pottery can fall well below -6.5. The bottom of the 2013 range needs care, though. The one value below about -12.4 is labelled "Ostracon #2 Almana." Almana is Hebrew for "widow," and the IAA study calls the Widow's Plea "ostracon 2" and measured its crust at -15.8 to -23.9.⁶ Four other labels also begin "Ostracon" and name no site, and the judge's summary says Shemesh and the prosecution's isotope team had measured some of the same objects.² So the -22.46 at the bottom of the published "excavated" range appears to be the disputed ostracon itself. The 51 values labelled with a site run from about -1 to about -12.4, as the judge's summary says. They do not reach the second ostracon's values.³⁴
2. For a limestone ossuary in a rock-cut tomb, pottery crusts from open sites are a weaker comparison, but so are the cave deposits behind the prosecution's range. The ossuary study's own excavated controls were three ossuaries.¹³ The ossuary's letter values, -7.5 to -10.2, fall inside the spread Shemesh found: seven of his 51 site-labelled pottery crusts, from Tel Gat, Tel Hazor and Tel Farra, are as low. The judge held that this raised a real doubt about the range.²³⁴
3. On the evidence's own standing, Shemesh's report is unpublished. His values can now be read item by item from the 2013 chart, but the sample types, contexts and pretreatment are known only through a defence-side translation of the judge's summary. The prosecution's own study warns that fired pottery contains altered calcite with values of -9.1 to -12.4 per mil, which a scraped sample could pick up.⁶ Six of Shemesh's seven lowest site-labelled values fall inside that band. That is a possible explanation of the low readings that would have to be tested on his samples, not a reason to ignore them.

An in-range value is compatible with natural formation but does not authenticate a crust;  an artificial mixture can also fall in that range.²  The pottery results challenge a universal boundary for natural crusts.  They do not establish the natural range for a limestone box in a Jerusalem tomb, and they do not erase the difference between its letters and the three excavated ossuaries.  Conversely, that difference concerns the sampled coating, not directly the age of the carving.  The second ostracon's much lower values remain further evidence against its coating when considered with the microscopy.

Controls must be real controls.  The ossuary isotope study compared six low readings from the letters with three excavated ossuaries.¹³  The pomegranate study compared its object with three uninscribed ivory pomegranates "kindly lent to us for comparison" by a private collector, two of them "reportedly found in a tomb in Samaria":  unprovenanced objects serving as the natural baseline.³⁵  The ostraca study is better designed.  It compared the suspect pieces with ten ostraca excavated at six named sites, and the difference it found was qualitative:  no wax, no fresh scratches, a different crust texture.  That does not depend on where a boundary in per-mil units is drawn.⁶

A forger's sources can date the forgery.  When Israel Eph'al and Joseph Naveh first doubted the Moussaieff ostraca in 1998, they noted phrases matching known inscriptions almost word for word.  They then reasoned that "If indeed they are forgeries, they must have been produced after the publication of the latest epigraphic phrase which was incorporated in them."³⁶  The same logic exposed the Gospel of Jesus' Wife, below.  A text that copies a modern edition's line breaks, misprints or restorations of damaged letters cannot be older than that edition.

Part 3: The Museum of the Bible fragments and the post-2002 market

Verdict:  all 16 Museum of the Bible fragments very probably carry modern writing on scraps of old, probably ancient, leather.  Nine similar fragments in the Schøyen Collection were judged forged, two with high confidence and seven as most likely.  The finding covers these fragments, not the Dead Sea Scrolls found in the Qumran caves.

How the fragments reached the market

From 2002 a stream of small "Dead Sea Scroll" fragments appeared for sale, most of them carrying a few verses of the Bible.²²  The catalogue kept by Årstein Justnes of the University of Agder traces most of them to one first seller, William Kando.  Buyers included the Norwegian collector Martin Schøyen, the Green family, whose purchases went to the Museum of the Bible, Azusa Pacific University and Southwestern Baptist Theological Seminary.²²  The Kando family had real credentials.  Davis and his co-authors, who later exposed nine of the fragments, accept that "probably authentic fragments were sold to The Schøyen Collection by the Kando family in the period 1999–2003."  The same paper urges "serious doubts about the large number of Judaean Desert fragments that has poured onto the market since 2003, and especially since 2009."¹  A dealer who once sold genuine material is not thereby a guarantee for the next sale.

The Museum of the Bible owns 16 of these fragments, bought for Steve Green "in four lots from four separate private collectors."  Thirteen were published by a team of scholars in 2016, and "At the time of publication, no scientific examination of the Museum's scroll fragments had been carried out."³  In 2017 the museum sent five fragments to the German federal materials laboratory, BAM, which reported in 2018 that they showed characteristics "inconsistent with ancient origins."³

What the 2019 tests found

In 2019 the museum hired an independent team led by the art-fraud investigator Colette Loll, with conservators and conservation scientists, to test the whole collection.  Every fragment was imaged and examined under the microscope;  six were sampled for chemical analysis.³  The team's conclusion was unanimous:  "none of the textual fragments in the Museum of the Bible's Dead Sea Scroll collection are authentic," and each showed signs of "deliberate forgeries created in the twentieth century."³

The reasons fall into two groups of very different strength.

1. Chemistry and materials were suggestive but not decisive. With one exception the fragments were leather, not the parchment typical of the real scrolls, and all were coated in a shiny animal glue that imitates the glassy look of degraded genuine scrolls.³ The team also reported signs of lime depilation, a later technology for removing hair from hides. But the laboratory that did the infrared work warned that carbonates from a burial environment, and decay of the collagen itself, can produce similar signals. It found "no anachronistic or anomalous materials" apart from conservation materials, and noted that with carbon ink lacking modern additives its methods could draw no conclusion about age.³ Chemistry alone would have left the question open.
2. The relationships between ink and damage were decisive. "In all of the fragments that contained writing," the team "observed examples where modern ink was applied atop preexisting surface deposits and across cracks and areas of delamination already present on the repurposed material."³ Its detailed list is longer: ink "going across and into deep cracks," "skipping over the high points of the grain layer," "waterfalling off the edges of the fragments," and "lying on top of thick, granular mineral deposits." The inks were "very thick and shiny," unlike the thin, matte inks of authentic scrolls.³ On some fragments loose minerals had been scattered "after writing, and while the ink was still wet."³

Each observation fixes an order:  the cracks, the broken edges and the burial deposits came first, and the ink came after.  That order alone does not prove a modern date.  In principle an ancient scribe could have reused an old, damaged scrap.  What turns the order into evidence of forgery is the whole pattern.  Ink runs off edges that were already broken to the fragment's present small size, so these strokes were applied to scraps rather than to an intact scroll that later fell apart.  Ink over burial deposits indicates writing after burial.  Minerals adhering to wet ink are consistent with an attempt to imitate a cave find.  And the texts resemble the biblical scroll fragments that entered the modern market after the discoveries of 1947.³⁷  That market explains the combination better than ancient reuse of small, buried scraps.  The report's "twentieth century" is an inference from the combination, not a measured date for the ink.³³  The report also records undocumented restoration before sale, with acrylic and cellulose coatings.³  Restoration can introduce coatings and retouching, so a modern feature alone does not establish fabrication of an entire text.  The case here rests on repeated writing over pre-existing damage and deposits across the collection, together with the apparent efforts to imitate aged fragments.

The team's own reconstruction is that forgers took "small scraps taken from archeological deposits of leather" and wrote on them.³  A radiocarbon date could test the leather's age, but would not authenticate the writing:  the leather is probably ancient.

The Schøyen fragments

The 2017 study by Kipp Davis, Ira Rabin and six colleagues examined nine fragments that the Schøyen Collection had withdrawn from its own 2016 publication.  Their list of shared features includes papyri written across the fibres instead of along them, parchment written on the flesh side, sediments unlike those on genuine scrolls, "pen strokes and ink in regions of delamination," and unusual inks.¹  They graded their conclusions:  two papyri "can be regarded with a high degree of confidence to be modern forgeries," and the rest are "most likely ancient leather and papyrus penned in modern times."¹  That grading is a model of honest reporting:  not every suspicious fragment was declared proven.

What this does and does not touch

These findings concern objects bought from dealers after 2002.  The great majority of the Qumran texts were recovered in the caves by archaeologists or bought from Bedouin between 1947 and 1962, and published in the official series.³⁷  Accelerator radiocarbon dates on fourteen scrolls, published in 1992, "agree well, except in one case, with the paleographic estimates."³⁸  So the Museum of the Bible forgeries do not cast doubt on the Isaiah scroll or on the use of the scrolls in arguments about the text of the Old Testament, which are treated in Old Testament Text and Chronology.

They do show something about the genuine corpus as well.  Even within the Qumran collection, two documentary texts catalogued as Cave 4 proved to be datable as late as the Bar Kokhba revolt of the 130s AD.  The Arizona radiocarbon team noted that these fragments, "bought from a Bedouin, were probably mixed up with the Qumran fragments by antiquity dealers."³⁹  A dealer's findspot is a claim, even when the object is real.

Part 4: The Gospel of Jesus' Wife

Verdict:  very probably a modern forgery on old papyrus.  The laboratory tests were real, but they tested the papyrus and the kind of ink, which a forger can supply.  The case rests on convergent evidence:  a companion fragment copied from a printed book, judged to be in the same hand, and provenance papers that fail on their own details.

The claim and the tests

On 18 September 2012 Karen King of Harvard Divinity School announced a small Coptic papyrus fragment in which Jesus says "my wife."  Her full edition appeared in the Harvard Theological Review in 2014.  King herself framed the find narrowly:  "The fragment does not provide evidence that the historical Jesus was married, but concerns an early Christian debate over whether women who are wives and mothers can be disciples of Jesus."²⁴

The fragment passed an unusual battery of tests, and the laboratory reports, posted by Harvard, are more cautious than the press coverage of 2014.

1. Raman spectroscopy at Columbia found the ink "primarily based on carbon black pigments such as Lamp Black," with spectra "very similar" to carbon inks on many manuscripts of the early Christian centuries. The same report found the ink of the companion John fragment, discussed below, "similar to, but distinct from" the ink of the Gospel of Jesus' Wife.¹⁵
2. Infrared work at MIT found oxidised cellulose "consistent with their composition as ancient papyrus," but the team could "not offer any conclusions regarding the identity of the ink." It then concluded that the documents "should be considered to originate from the dates determined by other analytical methods."⁴⁰ That last step is the very inference from old material to old writing that this article warns against.
3. Radiocarbon at Arizona gave a date of 405 to 209 BC, which the laboratory itself distrusted. The sample weighed 2.27 milligrams, its cleaning had to be cut short, and its carbon-13 value was the lowest of 42 papyri compared, which the laboratory called "strong evidence for the presence of a contaminant."⁴¹
4. Radiocarbon at Harvard and Woods Hole gave 1275 years before present, plus or minus 45, calibrated to AD 659-869 with a median of AD 741. The John fragment gave AD 648-800.¹⁴

King concluded that "Current testing thus supports the conclusion that the papyrus and ink of GJW are ancient."²⁴  For the papyrus that was a fair reading.  For the writing it was not, because nothing in these tests ties the moment of writing to the age of the papyrus.  Later work by the Columbia team found that the Raman spectra of carbon inks on dated Egyptian papyri change systematically with age, "a possible basis for a new nondestructive means for estimating the dates of ancient papyri."  But its authors warn that such a method "may serve to disprove a hypothesis about authenticity but cannot positively authenticate any particular writing or manuscript as ancient."¹⁶¹⁶  The team applied the method to the Gospel of Jesus' Wife and the John fragment in a 2019 paper whose results were not available for this article.⁴²

King also put the strongest argument for authenticity in her own words.  The "gravest difficulty," she wrote, was to explain how a forger "incompetent in Coptic language with poor scribal skills" could also secure ancient papyrus, make ancient-style ink, leave no ink out of place under the microscope, "achieve patterns of differential aging" and "fabricate a paper trail of modern supporting documents."  She judged that "such a combination of bumbling and sophistication seems extremely unlikely."²⁴  That is a fair argument against forgery, built on real observations.  Its force depends on how hard each "sophisticated" step really is.  Blank old papyrus can be bought, carbon ink is still easily made, and the radiocarbon dates show the papyrus is old, not that it was written on when new.  The paper trail, which King counted as one of the hard parts, fails the checks described below.  Her microscopic point is real evidence in the other direction:  no source read here reports ink lying over damage on this fragment, as it does on the Museum of the Bible pieces.  But a forger who wrote on a sound patch of old papyrus would leave no such trace, so its absence weighs little.  What remains is that the hand is clumsy and the text odd, which fits a modern writer as well as an ancient one.

The fragment that gave it away

The owner had also lent Harvard a fragment of the Gospel of John in Coptic.  It came from the same source, was tested in the same laboratories, and its papyrus was radiocarbon-dated to the seventh to ninth centuries.  Christian Askeland, a specialist in the Coptic versions of John, showed in 2014 that it "is a crude copy from Herbert Thompson's 1924 edition of the 'Qau codex.'"⁴

His evidence was specific and checkable:

1. "all seventeen of the John fragment's line breaks coincide with those of the Qau codex exactly." The copyist took every other line of Thompson's printed page, and broke the pattern where Thompson's edition turns a page, producing an impossible text.⁴⁴
2. The fragment reproduces some of Thompson's modern restorations of holes in the codex, so it was copied from the printed edition, not from an independent ancient manuscript.⁴
3. It is written in the Lycopolitan dialect of Coptic, on papyrus dated to the seventh to ninth centuries, "at least a century and a half after the Lycopolitan dialect is believed to have disappeared."⁴
4. The two fragments share peculiar letter forms, which Askeland takes as "confirming that the same hand has created them."⁴

The case against the John fragment is very strong.  An ancient copyist could share line breaks with the Qau codex, and a modern editor might correctly restore words that once stood in an ancient exemplar.  Neither fact alone would establish modern copying.  The distinctive combination is the alternating-line pattern, its dependence on the printed page layout and the restored wording.  Askeland argues that this combination identifies Thompson's edition as the source.  If that identification is correct, the fragment was written after 1924, whatever the age of its papyrus.

The step from the John fragment to the Gospel of Jesus' Wife is a separate inference, and it is not as tight.  It rests on three things:  the two came together from the same owner;  Askeland's comparison of letter forms;  and a textual point made independently.  Within days of the announcement, Andrew Bernhard showed that almost every legible word in the Gospel of Jesus' Wife occurs, in nearly the same form and often in neighbouring lines, in the Coptic Gospel of Thomas, so that a modern writer with little Coptic could have assembled it.  He was careful to add that this "does NOT mean that it was."⁴³⁴³  The ink evidence neither helps nor hurts:  the Columbia team found the two inks similar but not identical, and two batches of an easily made carbon ink would look like that whether one writer used them or two.¹⁵  Askeland's conclusion is that "Both texts are modern forgeries written on genuinely ancient fragments of papyrus."⁴  For the Gospel of Jesus' Wife that is a very probable conclusion rather than a demonstration.  It stands on the same-hand judgment, the shared origin and the failed provenance papers, each of which could in principle be challenged, but which point the same way.

Bernhard's later analysis adds a direct argument about this fragment.  In 2015 he reported that its first line shares an accidental omission with the 2002 PDF of Michael Grondin's Coptic-English edition of Thomas.  He acknowledged that analogous mistakes occur in ancient manuscripts, so this alone is not decisive.⁴⁴  He also compared the owner's English translation, released by King, with Grondin's English.  It includes English words whose Coptic equivalents are absent from the fragment, as his earlier prediction of copying had anticipated.⁴⁴  A modern translator could use an online edition to translate an ancient fragment;  modern dependence in the translation alone would not date the writing.  Bernhard's stronger case combines that dependence with the Coptic text's corresponding omissions, awkward grammar and patchwork relationship to Thomas.  This provides a further route to forgery independent of the same-hand judgment, while still requiring the connection between the modern translation and the Coptic composition to be argued.

The papers behind it

King's edition reported what the owner had told her.  He "acquired the papyrus in 1999."  He supplied a photocopied contract with a German named Hans-Ulrich Laukamp dated 12 November 1999, with a handwritten note that the "Papyri were acquired in 1963 by the seller in Potsdam (East Germany)."  He also supplied a photocopy of a 1982 letter from the Berlin Egyptologist Peter Munro, which, King noted, "makes no mention of the GJW fragment."²⁴

In 2016 the journalist Ariel Sabar identified the owner as Walter Fritz, a former student of Egyptology at the Free University of Berlin living in Florida, and tested the papers.  His article quotes the Harvard classicist Christopher Jones:  "Perhaps the hardest thing of all to forge is provenance."⁴⁵  Laukamp's family told Sabar he had been in Berlin at his dying wife's bedside when the contract was supposedly signed in Florida.  Compared with Munro's other letters, the 1982 letter had a wrong address and postal code, typed "ss" where Munro used the German sharp s, and used a typeface and letterhead that appear in Munro's correspondence only from about 1990.⁴⁵  These are a journalist's findings with the help of consultants, not a published forensic report, and Fritz denied forging the papyrus.⁴⁵  But they are concrete, and no reply producing the original documents was found in the sources read for this article.

After reading Sabar's investigation, King said it "tips the balance towards forgery."  She added that she would need scientific proof or a confession for a definitive finding, and that the papyrus itself could in theory be authentic even if its provenance story was false.⁵  That is the correct form of a conclusion here.  The evidence against the fragment is strong and convergent, but it is circumstantial, and it is right to say so.

Part 5: The Golan prosecution: what was charged and what was decided

Three of the six objects in this article were counts in one criminal case, and a fourth, the pomegranate, was drawn into the same investigation.  The case is often misreported in both directions.  Some say the court "proved them genuine";  others treat the IAA committees as having "proved them fake."  Neither is right.

The judgment itself was not read for this article.  It runs to roughly 475 pages in Hebrew and is sold by Israeli legal databases, whose free pages show only the caption.⁴⁶  Searches for a free copy of the judgment, the indictment or a later appeal ruling found none.  What follows rests on three accounts of it, each from an interested party, and is labelled as such.  The first is the IAA's own statement on the day of the verdict.  The second is English translations of selected paragraphs made by four defence-side scientists.  The third is the report of the Biblical Archaeology Society, whose editor had championed the objects.

1. The defendants. The case, State of Israel v. Golan and others, criminal case 482/04 in the Jerusalem District Court, named five defendants: the collector Oded Golan, the dealer Robert Deutsch, the restorer Rafael Brown, Shlomo Cohen and Faiz al-Amaleh. The charges against Brown and Cohen were struck out, and al-Amaleh's case had ended before the verdict.⁴⁶
2. The counts that concern this article. In the defence scientists' account, count 1 was the ossuary inscription, count 2 the Jehoash tablet, count 3 "Ostraca 1, 2 and 3," charged against Deutsch, and count 7 the decoration of a stone oil lamp.² The Biblical Archaeology Society's report says the pomegranate "was not included in the individual counts, so the judge did not deal with it in his opinion," although another line of the same report lists it among the forgery allegations.⁴⁷ No source read here shows the pomegranate as a separate count, but the judgment itself was not read, so its status in the case remains unverified. It should not be described as acquitted or convicted.
3. The result. The IAA reported that the court acquitted Golan "on grounds of doubt" of forgery and fraud and convicted him on three counts under the Antiquities Law and of possessing property suspected of being stolen.¹⁷ Deutsch was acquitted of all forgery charges.⁴⁷
4. What the judge said the acquittal meant. In the defence translation, the judge wrote that "the prosecution failed to prove beyond reasonable doubt" that the ossuary inscription was a fake, and added: "This is not to say that the inscription on the ossuary is real and authentic and that was written two thousand years ago." He also held it "not proven" that its Jesus is the Jesus of the New Testament.² The IAA's statement agrees on the essential point: the judge "stressed that it cannot be determined that the items presented at the trial ... are not forged."¹⁷ On the Jehoash count he held that "from the standpoints of the writing, and from the standpoint of the language, and from the standpoint of the material aspect," forgery was not proven beyond reasonable doubt. He rejected the prosecution's argument "that a combination of doubts in various aspects should lead to the determination of a fake inscription."¹⁸
5. What the prosecution itself conceded. The judge quoted the prosecutor: "if the ossuary was standing alone in the trial, we probably would not continue the process."²

Two lessons follow, and they cut both ways.

An acquittal is not an authentication, and it is not a finding of probable forgery either.  A criminal court asks whether guilt is proved beyond reasonable doubt against a named person, and the judge applied that standard to each count separately.  An object can fail to be proved forged and still be more likely forged than not, or more likely genuine.  The acquittal does not say which.  The judge said explicitly that his finding did not make the ossuary inscription authentic, and he did not say it was probably forged.  He did make findings that bear on the evidence, such as his acceptance of Shemesh's data, and those are weighed with the evidence below, at the same remove of translation.  The defence scientists' own summary goes further than the passages they translate:  they write that "the artifacts according to the court are not fakes and could well be genuine."²  A 2013 paper by three of the same four authors says that the ossuary and the tablet "were determined not to be forged."³³  The translated words of the judge do not say that.

A unanimous committee is not a measurement, and it is not nothing.  The IAA set up two committees in March 2003, on the writing and on the materials, and announced unanimous findings of forgery by June.¹⁸  A committee's report records the judgment of qualified people, and that judgment is evidence.  Its weight, though, comes from the observations behind it, and the committees' own members did not all make independent observations.  In the defence-side translation, the judge found that the archaeologist Ronny Reich had first written that the ossuary inscription was authentic and changed his view after hearing the other members.  He treated that as an example of how "different experts were influenced ... from other expert's conclusions."²  As the judge summarised her testimony, the radiocarbon specialist Elisabetta Boaretto signed the materials committee's report without having had enough material to date.²  These points come through the defence's selection and translation and were not checked against the transcript.  They do not show the committee was wrong.  They do show that its unanimity was not a set of independent confirmations, and that its conclusions stand or fall with the observations of its laboratory members, chiefly Goren, Ayalon and Bar-Matthews, which the following parts weigh directly.

Part 6: The James ossuary

Verdict:  the box is ancient;  whether every letter of its inscription is ancient is unresolved.  Reports of biological film inside letters of "brother of Jesus" require an explanation, but neither the film nor an in-range isotope value directly dates the carving.  They do not establish which part, if any, was added or altered.  Even if the whole inscription is ancient, identifying its Jesus with Jesus of Nazareth is a further, probabilistic step.  Until the inscription is better tested, it can carry at most a small and conditional weight.

The object

The ossuary, a limestone bone box of the first century, was published by André Lemaire in 2002.  Its owner was the collector Oded Golan.  It was "reportedly found at Silwan in Jerusalem," but no findspot is documented.¹³  Its Aramaic inscription is not in dispute as a reading:  Ya'akov bar Yosef achui di-Yeshua, "James son of Joseph, brother of Jesus."⁴⁸  Nor is the box.  Even Yuval Goren, who judged the inscription modern, wrote that "The ossuary is authentic, based on its typology and patination."¹²

The case against the inscription

Goren's examination for the IAA committee in 2003 made three observations under the microscope.  The inscription "throughout its entire length is etched into the varnish and cuts through it," meaning the thin biological film on the box's surface.  "Only in the inscription area" was there a soft grey coating, easily removed with a wooden toothpick.  And that coating contained marine microfossils called coccoliths, a feature "never observed in the other patina samples."¹²  He concluded that the inscription "was engraved (or at least, completely cleaned) in modern times."¹²  In a fuller account the next year he wrote that "after the natural patination process a sharp tool was used to create or modify the letters," and that a fake coating of chalk dissolved in hot water was then applied.⁴⁹

The isotope study by Avner Ayalon, Miryam Bar-Matthews and Goren found six samples from the letters with oxygen values of -7.5 to -10.2 per mil.  Samples from the box's surface, from three excavated ossuaries and from cave deposits near Jerusalem lay between about -3.9 and -6.7.  They argued that the letter crust could only have formed from water at 40-50 degrees Celsius, "most probably with hot water."¹³¹³

This is a serious case, and it has stated limits.

1. The study's own text allows an innocent reading of the first observation: "The letters cut through the varnish ..., hence the entire inscription postdates it, or it has been cleaned by a sharp tool after its deposition."¹³ Goren's "or at least, completely cleaned" keeps the same alternative, and he testified that he chose that wording "for caution," because the letters might have been cleaned with a hard tool such as a nail.²
2. The isotope argument depends on its reference range. Part 2 explains why the pottery crusts reported at trial challenge a broad natural limit, while not directly establishing the range for an ossuary in a rock-cut tomb. The letter values fall inside Shemesh's wider pottery spread but outside those of the three more closely matched excavated ossuaries.¹³²
3. The isotope pattern does not line up neatly with any one story. In the defence translation the judge noted that the prosecution claimed the second half was forged, yet of four samples from that half, two fell inside the expected range and two outside.² Goren's published account identifies three sampled letters in Yeshua: the shin and vav gave -10.2 and -7.7, while "Only the last letter ('Ain') had a normal value."⁴⁹ The court account's fourth, in-range sample is a separate report; it is not an additional value in the seven-sample table published by Ayalon and colleagues.

The case for the inscription

The epigraphers who studied the writing found nothing modern in it.  Lemaire argued that the mixture of formal and cursive letter forms runs through both halves and is common on ossuaries.  He also argued that cleaning, which the earlier examinations at the Geological Survey of Israel and the Royal Ontario Museum had both noted, "is the only logical conclusion in the present state of the documentation."⁴⁸  As the trial judge summarised the testimony, the Hebrew University palaeographer Ada Yardeni and the Haifa archaeologist Ronny Reich, both prosecution witnesses, judged the writing ancient on its letters and content.²

The geologist Amnon Rosenfeld and his colleagues, who examined the box for the Geological Survey before it was published, describe a different picture of the surface.  In their account the beige crust "can be observed on the surface of the ossuary continuing into the engraving";  "The engraving clearly does not cut the patina."  Fine scratches, which they attribute to stones falling from the cave roof, cross the letters "filled with the same patina," and solution pits sit on top of several letters.⁵⁰  If correct, a patinated scratch running continuously across a letter shows that the letter was there before the scratch, and that is hard to fake.  Their paper is open about its limits:  they agree that "The inscription was cleaned," and their listed samples come from near the letters, not from inside the letters of the first half.⁵⁰⁵⁰  Two of the three authors testified for the defence, and their observations have not been independently published.  Rosenfeld and Feldman's verdict, in a 2013 paper with Joel Kronfeld that attacks the isotope test, is that the work "strengthens the contention that both inscriptions of the Jehoash Inscription and James Ossuary are probably authentic."³³  Part 2 weighs their isotope argument.

The geomicrobiologist Wolfgang Krumbein of the University of Oldenburg, invited as a defence expert, examined the box for the trial.²  His 2005 report says he "found no indisputable evidence" that the inscription was recent, and that "there is a high probability" that it was "produced many hundred years ago."  His firmest figure is a minimum:  "A period of at least 50-100 years was required" to form the patina he found inside the letters.  He also found that the inscription had been treated "using a sharp implements and application of detergent/cleansers."⁵¹  If that minimum is right, the patina inside the letters took at least half a century to form, which would rule out a forgery made in the years just before the box was announced in 2002, unless older material was worked into new letters, the possibility discussed below.  Two cautions apply.  The growth rate behind the minimum has not been tested by anyone outside the dispute.  And the same report gives the same high-probability verdict for the Jehoash tablet, whose language Cross, Eph'al and Greenstein judge modern;  a reader who judges that tablet probably forged, as Part 7 does, has reason to discount the method behind both verdicts.

The most important evidence for the inscription concerns one letter, the Ain at the end of Yeshua, "Jesus."

1. The isotope sample. The peer-reviewed isotope study itself reported that the sample "taken from the last letter of the inscription" gave -5.8 per mil, inside the natural range. The authors attributed it "most likely" to contamination from the stone beneath, which is possible but was not shown.¹³
2. The microscope. At trial Goren asked to re-examine the box after his testimony. As the judge quoted him, his conclusion was "unequivocal, in the bottom line of the letter 'Ain' there is varnish, a true biopatina."²
3. A second witness. Orna Cohen, a conservator who had sat on the IAA committee and testified for the defence, saw genuine yellowish patina inside the letters of the second half, "brother of Jesus." She judged the first half, "James son of Joseph," "probably late." She still concluded that the inscription as a whole should be treated as an attempt to deceive.²

These observations are not a direct demonstration that the letters were carved in antiquity.  A biological film is not itself a dated layer, and an in-range isotope value does not independently establish that the sampled coating grew naturally over a carved stroke.  The isotope sample and the microscopic observation concern the same letter, but they were not demonstrated to concern the same material.  A forger could incorporate an old patinated scratch into a new or deepened letter;  an ancient inscription later cleaned could also retain film in missed spots.  Distinguishing those cases requires the spatial relation of film to the complete incised stroke, not merely its presence somewhere inside a letter.  The second and third observations reach us through the defence's translation of the judgment, and Goren's own explanation of the Ain at trial was not available in his words.  The first observation is direct, but the paper itself offers host-rock contamination as its explanation.

Weighing the competing explanations

Several explanations are on the table, and each has to account for all the observations, not just the convenient ones.  They are not the only possibilities;  mixtures of them are possible too.

1. The whole inscription was cut in modern times. This explains the soft coating, the coccoliths and the low isotope values. It has to account for the reported biological film, perhaps through incorporation of an older surface feature. The normal isotope value could instead reflect host-rock contamination, as the study proposes; it does not require the same explanation as the film. Rosenfeld's patinated scratches across letters would be a further challenge if independently confirmed.
2. "Brother of Jesus" was added to an ancient "James son of Joseph." This was the prosecution's position at trial, according to the judge as translated.² The reported film and in-range values in the second half challenge a simple claim that every coating there is artificial. They do not by themselves distinguish this explanation from a wholly modern inscription: either could incorporate old surface features or produce an in-range sample. That comparison also needs equivalent observations from the first half.
3. Ancient letters were modified: some recut, deepened or "improved," with the old film surviving in places. This is close to Goren's own "create or modify" and would explain mixed evidence within single words. It is hard to test without much finer sampling.
4. The inscription is ancient and was cleaned, perhaps harshly, in modern times, leaving a residue. This explains the letters cutting the varnish, the soft residue and film surviving in some letters but not others. It has to explain the low isotope values and the coccoliths. The judge did not accept the defence suggestion that detergents caused the low values, though he found sharp-tool cleaning a further reasonable doubt.²

The present evidence does not support a secure ranking of these explanations.  The apparent survival of film matters, but its age, exact relation to the carving and possible incorporation into altered letters remain in dispute.  The low isotope values and coccoliths raise a serious question about the coating without directly dating the letters beneath it.  Several important observations also reach us through selected courtroom translations.  "Unresolved" is the honest word.  What would move the question is new, published, reproducible evidence:  high-resolution imaging and sampling inside letters of both halves, examined by laboratories with no stake in the dispute, with excavated ossuaries as controls.  A blind study would be one good design, but not the only one.

Even if it is authentic

A genuine inscription would establish that a first-century Jerusalem man named James had a father named Joseph and a brother named Jesus whom it was useful to mention.  All three names were common.  Lemaire cited an estimate by a Tel Aviv statistician that about 1.71 men in Jerusalem before AD 70 would fit that description, and called the identification "very probable."⁴⁸  That estimate rests on assumptions about population and naming that are discussed, with others' estimates, in Jesus in Archaeology and Outside Sources.  The trial judge, for his part, held it "not proven" that the Jesus named is the Jesus of the New Testament.²

The stakes for a Christian argument should be stated precisely.  If the inscription were ancient and named the family of Jesus of Nazareth, it would provide a first-century physical attestation, carved within a few decades of Jesus' death.  It would also be independent of the manuscripts through which Josephus and the New Testament reached us.  That would be a real gain, and its loss, if the inscription is modern or names other men, is a real loss of a proposed corroboration.  It does not remove the literary evidence.  Josephus names James as "the brother of Jesus who was called the Christ,"²³ in a passage whose authenticity is weighed in Jesus in Archaeology and Outside Sources.

Part 7: The Jehoash inscription

Verdict:  probably a forgery, on a cumulative case that its leading philological defender contests item by item.  Several expressions have no attested ancient parallel, and the text shares distinctive wording with more than one biblical account.  The best published defence offers ancient readings for some of them, without claiming the tablet is genuine.  The physical evidence is disputed between laboratories.  The court's acquittal does not settle it either way.

The object and the claim

The Jehoash tablet is a dark stone slab with about sixteen lines of Old Hebrew script.¹⁸  It reads as a king's account of repairing the Temple in Jerusalem, matching the story of King Jehoash of Judah in 2 Kings 12 and 2 Chronicles 24.  If genuine, it would be the first royal building inscription known from Judah.  It came to public attention through a 2002 laboratory report by geologists of the Geological Survey of Israel, Ilani and colleagues including Amnon Rosenfeld, who argued that its crust was ancient.⁸¹²  Israeli police later seized it from Oded Golan.¹⁸  The king's name itself is not on the stone.  The first line is broken, and Frank Moore Cross pointed out that the reader "would be forced to restore the name of Jehoash."⁷

The linguistic case against it

Cross, working first from press photographs and later from good ones, judged that "the inscription is demonstrably a forgery."⁷  His reasons were concrete.  The phrase ndbt lb, "generosity," "is a post-biblical expression";  in other places the grammar follows later Hebrew;  "Clearly we are not dealing in these instances with ninth-century Hebrew."  The text, he concluded, was made "by taking phrases or words, now from Kings, now from Chronicles, interlarded with the occasional use of late language or biblical words with post-biblical meanings."⁷⁷

Israel Eph'al of the Hebrew University added independent points.  The tablet's phrase in lines 10-11, literally "to make the breach," "would be an oxymoron in Biblical Hebrew."  Biblical Hebrew uses bedeq for a breach to be repaired, while the phrase as used on the tablet "occurs only in modern Hebrew, where it means 'to overhaul.'"⁸  Some of its expressions "occur only in Chronicles, not in Kings."⁸  Eph'al was careful about his limits:  "I cannot evaluate the finds of the laboratory tests ... because that is not my specialty."⁸

Stuart Irvine and Charles Isbell's survey of 2005 lists the full set of objections, then tests each.⁵²  Some they found weak.  Medial vowel letters and the spelling of "his people" have early parallels, and the evidence for ninth-century spelling is too thin for firm rules.  Others they found strong, above all the closing blessing, "May YHWH command His people with blessing," whose syntax inverts the biblical idiom.  They also found telling a pattern in the borrowed wording:  each shared phrase matches only one of the biblical accounts, including the later Chronicles and the account of Josiah's repairs.  In two places, they argue, the tablet misreads its source.  The Chronicler's tax of Moses "in the desert" becomes money collected in the desert, though Cohen translates the same words as "in the (sparsely populated) steppe."⁹  And the "faithful" paymasters of 2 Kings 12 become workers who did the work faithfully, as in 2 Chronicles 34.  Their summary is that "the strength of the forgery position is the cumulative weight of all the objections that have been raised rather than any single objection taken alone."⁵²

The philological defence

A substantial philological defence came from Chaim Cohen of Ben-Gurion University, who testified for Golan at trial.  His position needs stating carefully, because it is often misreported.  He did not claim the tablet is authentic.  "I do not know whether or not this inscription is genuine," he wrote in 2007.  "I do contend, however, that it can not be proven philologically to be a modern-day forgery."  If it is a forgery, he added, "it is a most brilliant forgery."⁹

His method was explicit.  Similarities to the Bible prove nothing either way, he argued, because a forger would copy the Bible and an authentic text might share wording with, or even be a source for, a biblical account.  The real test is the inscription's unique features.  Any that can be traced to modern Hebrew would be "smoking guns."  Any that turn out to be correct ancient usage, attested in related languages but unknown to most scholars, would count against forgery.  The tablet belongs to the genre of royal building inscription, which has no example in the Hebrew Bible but many in Akkadian, so he looked for parallels there.¹⁰

His reply to Eph'al's strongest point shows the method at work.  He accepted that bedeq means "breach."  But he read the whole sentence as one verb with seven objects:  "(Then) I renovated the breach(es) of the Temple and of the surrounding walls, and the storied structure, and the meshwork, and the winding stairs, and the recesses, and the doors."  On that reading the verb means "renovate" for all seven objects, as related verbs do in Akkadian building texts.  On his syntactic analysis the "breaches" belong to the walls as well as the Temple, a phrase for which, he notes, modern Hebrew has no idiom.  "If my translation is correct," he concluded, "the claim that the usage of this phrase ... is based on Modern Hebrew and is therefore philological proof of forgery must be rejected."⁹  Greenstein's answer was that the parallel is missing where it matters:  "There is no evidence at all, in any ancient Semitic language, of an expression 'to make a fissure' in the sense of 'to repair a fissure.'"¹⁰

The reply came from Edward Greenstein, who regarded the tablet as a forgery;  the conference report says each of the two "conceded the philological expertise of the other."  Greenstein did not dispute that many of the tablet's usages are unattested in ancient Hebrew;  the conference report records that Cohen agreed with that.¹⁰  The disagreement was over method.  For a text that "simply appears on the market," Greenstein argued, each anomaly should be measured against what is known, not against what is possible.  He illustrated with the word for "testimony" near the end of the text.  In early biblical Hebrew, he wrote, 'edut means "covenant," never "testimony."  Cohen's Akkadian parallel shows only that the word could have existed, and "Many things that are possible never occur."  An explanation that needs a separate special argument for each oddity, he concluded, is weaker than one explanation, a modern writer, that covers them all.¹⁰  The trial judge preferred Cohen's testimony on the language:  "I cannot determine that the JI is a fake inscription according to the language and the linguistic aspects."¹⁸

Both sides make a fair methodological point.  Cohen is right that the Hebrew Bible is a small sample of ancient Hebrew, that an unattested form is not thereby impossible, and that likeness to the Bible proves nothing by itself.  Greenstein and Irvine and Isbell are right that a sixteen-line text with many unattested usages, several needing separate rescue, is more simply explained as modern work, and that the misreadings of specific biblical passages are hard to explain as an older source.  Cohen's 2007 study covers lines 1 to 14.  His later treatment of the last three lines, where the blessing and 'edut stand, was not available for this article, and his answer to the blessing syntax is not assessed here.⁹  On the published evidence the linguistic case for forgery is stronger than the case against, but it is a matter of weight, not demonstration.

The physical evidence

The laboratories disagreed about almost everything.

1. The stone. Goren identified it as "low-grade metamorphic greywacke," not native to the region, with similar rocks "east of the Troodos massif in Cyprus."¹² Rosenfeld and his colleagues call it "a gray, fine-grained arkosic sandstone."¹⁸ The question matters less than it seemed, since a forger can use an old stone of any origin.
2. The crust. Goren found the coating in the letters to be "an artificial mixture of ferruginous clay, powdered chalk, carbonized matter and microscopic particles of metal (gold?)," with signs of "fresh cutting and polishing" underneath. Rosenfeld's team describe an iron-rich crust "post-dating the incised inscription" and a crack running across several lines.¹²¹⁸
3. The radiocarbon date. Rosenfeld's team reports that carbon particles in the crust "yield a radiocarbon age of approximately 2250 years BP (third century BCE)." They explain the carbon and tiny gold globules by a fire in which gold objects melted, after which both settled in the ground and "were later incorporated" into a crust growing on the buried tablet.¹⁸ The measurement and its meaning need separating. "2250 years BP" is a radiocarbon age; "third century BCE" is their rough calendar equivalent. Either way, the date belongs to the death of the plant or tree whose carbon became charcoal. It is not the date of the fire, which could be later if old wood burned, and not the date the charcoal entered the crust, which could be any time after the fire. It is also not the date of the carving. On the defenders' own scenario, the crust began to form no earlier than the charcoal, and the carving could be any amount older; the measurement cannot say how much, and it says that much only if the crust is natural, which is the point in dispute. Goren, on the other side, describes the coating as a fake made partly with "some ancient charcoal."⁴⁹ Old charcoal could be mixed into a new coating, or settle into an old one. As used here, the measurement does not discriminate between the two stories.

What follows

The physical evidence could fit either story.  The linguistic evidence is the kind a modern composition is most likely to betray, and its cumulative weight still favours forgery.  But its best-known single "smoking gun," the "modern Hebrew" phrase in lines 10-11, is no longer a clean one:  Cohen offers a serious ancient reading, and Greenstein rejects it for lack of any ancient parallel.  The argument rests on the accumulation, not on any one phrase.  The tablet should be treated as probably forged and given no weight as evidence.  If a reader wants to mention it, the honest statement is that Cross, Eph'al, Greenstein and others who have published on it judge it a forgery, and that its main philological defender says only that forgery has not been proved.  The Jehoash repairs are known from 2 Kings 12 and 2 Chronicles 24.  An authentic inscription would have been the first royal Judean building text and a direct check on them, so its probable loss is a real one.  The evidence for the Old Testament's general historical setting rests on excavated inscriptions, discussed in Archaeological Confirmations and Archaeology of the First Temple.

Part 8: The ivory pomegranate

Verdict:  the pomegranate is ancient, probably Late Bronze Age;  whether its inscription is ancient is unresolved.  A joint examination in 2007 removed two of the committee's arguments but left the central one:  whether strokes of the taw and he run into an ancient break, which the two sides still read differently.  The inscription should not be used as evidence for the First Temple until that is settled.

The object

The pomegranate is a thumb-sized carving of hippopotamus ivory, 43 mm high, with a hole in its base for a rod.  Around its shoulder runs a short Old Hebrew inscription, read as "[Belonging] to the Tem[ple of Yahwe]h, holy to the priests."⁵³  Lemaire first saw it "in an antiquities shop in the Old City of Jerusalem in July 1979."  It was later "smuggled out of Israel" and exhibited in Paris.  The Israel Museum bought it "from an anonymous collector for the sum of $550,000."³⁵  The committee report gives the purchase year as 1989 and the museum's 2019 study as 1988.³⁵⁵³  If the inscription is genuine, the object was dedicated to a temple of Yahweh, most naturally the one in Jerusalem, in the eighth century BC.

The epigrapher Nahman Avigad defended the inscription in 1990.  "The edges of several of the letters' lines are not sharp as they would be in new incisions, but rather rounded and worn, merging with the surface of the object."  He was "fully convinced" of its authenticity, and wrote that "the epigraphical evidence alone, in my opinion, is absolutely convincing."⁵⁴⁵⁴

The committee's case

In 2004 the Israel Museum and the IAA appointed a committee, whose report in 2005 judged the object ancient, "probably" Late Bronze Age, and the inscription "a recent forgery."³⁵  Its central argument was about the breaks.  A large piece of the pomegranate's body broke off in antiquity, and two smaller "new breaks" run along that old break.  The committee argued that several letters near the old break, a taw, a yod and a he, stop short of it or "fade" towards it.  In their view the engraver was working on an object that was already broken and was trying not to chip it further:  "these new fractures are most telling.  They clearly indicate that the old break already existed and was taken into consideration when the letters were distributed."³⁵  They added that the deepest parts of the letters show "fresh and sharp engravings," which they thought a careless modern cleaning would not produce.  They also found the patina-like material in the letters to be an artificial mixture containing marine carbonate and modern adhesives.³⁵³⁵  And they argued from meaning:  "no-one would have dared to donate a broken item to the Temple and then engrave upon it."³⁵  At a conference in 2007, the committee's epigrapher Aaron Demsky added that he and Shmuel Ahituv had seen the spaces between words as "the smoking gun," because in his view Hebrew scribes of the eighth century divided words with dots or strokes, not spaces.  He admitted, "The truth is that I can never be sure."¹⁰

Lemaire's reply

Lemaire re-examined the pomegranate under the stereomicroscope in September 2005 and answered point by point.

1. The breaks. He reported that the upper stroke of the taw and the standing line of the he are cut by the old break, and that the edges where they meet it "are worn and eroded with the same yellowish colour/patina, while the edges of the new breaks ... are sharp and whitish." So, he concluded, the letters were "incised before the old break" and the inscription "is, therefore, authentic."⁵⁵ The new breaks he attributed to the modern discovery and handling, calling the committee's reading of them "clearly pure conjecture."⁵⁵
2. The fresh cuts and the adhesive. Lemaire reported, through the dealer Robert Deutsch on an internet mailing list, that the restorer Rafael Brown had said he "washed and cleaned the pomegranate, injected glues in its cracks, cleaned and deepened its inscription with an iron tool, and covered the surface with a coating."⁵⁵ If that is true, fresh tool marks and modern glue in the letters are what one would expect. As Lemaire put it, a restored surface "cannot be used to refute the authenticity of the inscription. Otherwise, all restored antiquities would be considered fakes."⁵⁵ The report reaches us at third hand, and it is Brown's statement, made after he was himself named in the investigation.
3. The controls. An appendix to the rejoinder pointed out that the comparison pomegranates were "three looted ivory pomegranates," and that a tell is not a cave, so cave-based isotope ranges do not apply.⁵⁵

The joint re-examination of 2007

On 3 May 2007 Goren, Ahituv, Demsky, Lemaire and Kyle McCarter examined the pomegranate together at the Israel Museum.  Goren worked the stereo microscope, its image projected on a wall so that all could see the same thing, and photographs were taken as the participants asked.  The results appeared in the Israel Exploration Journal as a joint paper with two signed parts, one by Ahituv, Demsky and Goren and one by Lemaire.¹¹  It is the closest thing to a shared record of the evidence, and it narrowed the dispute.

1. The yod: settled in favour of the inscription. The committee had cited this letter as a sign of forgery. Ahituv, Demsky and Goren "accepted Lemaire's observation" about its lower tip, and Lemaire records that "Everybody agrees" it is "not problematic at all, but is incised as it should be."¹¹
2. The word spaces: narrowed. Lemaire records that the committee's claim that two word spaces were equal "is now conceded" to be wrong. The three still see one larger space as an anachronistic word divider; Lemaire still sees the spacing as within normal variation.¹¹
3. The taw: still split. Only the top two strokes of this x-shaped letter survive. The three wrote that it is "difficult to determine whether the left stroke arrived at the OLD break," and that the right stroke "was aborted," stopping short of a bulge before the break. Lemaire saw the left stroke reach the old break, leaving a v-shaped notch in its edge. He also argued that the letter's acute angle, which the three called "faulty," is normal in eighth-century Hebrew writing.¹¹¹¹
4. The final he: answered by one side only. Lemaire wrote that its vertical stroke meets the old break "exactly as it should be" and that his colleagues "do not mention" it.¹¹ The Biblical Archaeology Society's editor, who was present, made the same complaint.⁴⁷

The three committee members kept their verdict, giving as "The main reason for this conclusion" "the apparent caution of the engraver not to access the old break," supported by the taw's form, a "curious" mem and the spacing.¹¹  Lemaire kept his:  the "clear connection" of two letters with the old break shows the inscription was cut first.¹¹  So after a joint look under the same microscope, the two sides agree on more than before, but still disagree on what one stroke does at the break.

The museum's 2019 study

In 2019 Eran Arie of the Israel Museum surveyed more than ninety ivory and bone pomegranate and poppy-capsule heads from the Late Bronze and Iron Ages.  He stated plainly that "This article does not deal with the inscription; it is rather a study of the object itself."⁵³  His conclusion nonetheless bears on it.  The object's size and shape point to the Late Bronze Age, while the script points to the eighth century BC.  "It seems unlikely that such an object would have been in use for centuries and be inscribed with a dedicatory inscription 500 years after it was produced," and no other pomegranate of the group carries an inscription.  He concluded that "its inscription is a modern fake."⁵³

This is a contextual argument, not an observation of the inscription, and Lemaire had already answered it in advance:  precious ivories and bronzes "can be kept for centuries."⁵⁵  Neither point settles the case.  That heirlooms existed shows a late inscription on an old object is possible.  It is not evidence that this inscription is ancient.  Arie's point lowers the prior plausibility of an ancient inscription without excluding it.

Where this leaves the inscription

No source read for this article shows the pomegranate as a count at the Golan trial, and the judgment itself was not read.⁴⁷  The question is a matter of fact at a scale of fractions of a millimetre:  where exactly the strokes of the taw and he meet the old break, and whether the junction is original or altered.  Two further facts complicate the reading.  The restorer Rafael Brown is reported, at third hand, to have cleaned the pomegranate, injected glue into its cracks and "deepened its inscription with an iron tool."⁵⁵  And the "new breaks" next to the old one happened at some unrecorded time.  Treatment of this kind can erase or imitate exactly the evidence in dispute.

Two published reports from the same 2007 session still disagree, and the photographs printed with them did not end the argument.  High-resolution three-dimensional imaging of the junctions, published in full and examined by specialists from outside the original dispute, might narrow it further.  It might also show that restoration has destroyed the decisive evidence.  Until then the honest verdict is that the object is real and the inscription is unproven either way.  The forgery case now rests mainly on the taw and on the context;  the authenticity case rests on Lemaire's reading of the taw and he at the break, which his colleagues have not refuted in the sources read here.

Part 9: The Moussaieff ostraca: "Three Shekels" and "Widow's Plea"

Verdict:  very probably forged.  The sherds are genuine Iron Age pottery, but the IAA examination found fresh scratches in the letters under a wax film, with a lime crust on top, unlike the layers on ten excavated ostraca.  No published reply to that layer sequence was found in the sources searched, and a criminal acquittal does not answer it.  The case is also a clear warning that "laboratory tests showed no anomalies" can be wrong.

The texts and the first tests

Two ink-written potsherds in the collection of Shlomo Moussaieff were published in 1996 and 1998.  The first records an order of a king Ashyahu to give three shekels of "silver of Tarshish" to the House of Yahweh.  The second is a widow's petition to an official about her late husband's land.³⁶  Their editors, Pierre Bordreuil, Felice Israel and Dennis Pardee, were aware of the forgery question and argued that a forger would have had to be "a very subtle and sensitive specialist in biblical studies."⁵⁶

They also reported laboratory tests.  The owner had samples examined in 1995 by McCrone Scientific and by Mikrofokus Oy of Helsinki, using electron microscopes with X-ray microanalysis.  In a sidebar to the 1998 publication, Christopher Rollston summarised the reports.  Both laboratories judged the ink similar in composition to Dead Sea Scroll ink.  Mikrofokus described the white crust over the ink as "small isometric bipyramidal calcite crystals and amorphous silica" and concluded that it "is a natural precipitation product."  The report's conclusion was that "what is under the white material must then be genuine."  The same sidebar records that Mikrofokus found contaminants, including "fragments of bronze and steel and scales of a type of glue," which it put down to "previous handling and variable storage" rather than modern manufacture.⁵⁶  The main article accordingly said that "Laboratory analyses have shown no anomalies in either the ostracon or the ink."⁵⁶

In the same year Israel Eph'al and Joseph Naveh listed phrases in the two texts that closely match published inscriptions and biblical verses.  They declined to call the ostraca forged, saying "such an accusation should not be levelled unless it can be proven beyond any doubt."³⁶  Naveh had earlier exposed forgers who "copied texts and formulae from epigraphic publications available at the time."³⁶  Others defended them.  Chaim Cohen later listed three published opinions that the ostraca were probably authentic, by Dennis Pardee in 2002, Elisha Qimron in 1998 and himself in 2003.  All three were published before the laboratory study described next.¹⁰

The 2005 examination

During the police investigation the ostraca were lent to the IAA and examined by Yuval Goren, Avner Ayalon, Miryam Bar-Matthews and Bettina Schilman.  Their design was better than in the ossuary and pomegranate studies, because the controls were real.  They compared the ostraca with ten "legally excavated and well recorded ostraca from Arad, the City of David, Lachish, Tel Beer Sheva, Horvat 'Uza, and Tell el-Far'ah (S)."⁶  Their findings were:

1. The sherds are genuine. Their clay and temper match Iron Age pottery of Judah and the Shephelah.⁶
2. A wax layer lies over the letters. A thin translucent coating over the writing was soft, and it melted under a small flame with the smell of paraffin. In thin section it had no crystal structure. "Hence it was identified as wax (paraffin)."⁶ These are simple physical tests, not a chemical analysis such as infrared spectroscopy, so "wax" is better established than "paraffin" in particular. For the argument, what matters is that the film is a soft, low-melting, non-mineral material and where it lies.
3. Fresh scratches lie under the wax. "The letters of the two ostraca reveal fresh treatment signs by a sharp metallic tool ..., appearing underneath the paraffin coating and the patina layer."⁶
4. The white crust lies on top of the wax and looks like burnt lime. "White calcitic patina overlies the inscriptions and the paraffin layers in the two ostraca."⁶ It is made of micron-sized fibrous calcite "typical of burnt lime," mixed with plant ash. On the excavated ostraca the ink sat in the pores of the sherd, "lacking any fresh scratches by a sharp tool and coated directly by calcitic patina with no paraffin inlay."⁶⁶
5. The isotopes agree, for one ostracon. The thick crust on the second ostracon gave oxygen values of -15.8 to -23.9 per mil, far from the -3.1 to -6.8 of the excavated ostraca. The authors admitted that the thin crust on the first ostracon might be contaminated by the sherd, so "based on its oxygen isotope composition alone, it is impossible to determine" whether it was natural.⁶

They concluded that the ostraca "indicate without a doubt" modern forgery, and reconstructed the sequence:  an old sherd cleaned, written with carbon ink, the letters scratched to look worn, waxed, then coated with lime.⁶

Why this case is stronger than the isotope range

At trial the ostraca were count 3, charged against Deutsch, and Deutsch was acquitted of all forgery charges.²⁴⁷  The parts of the judgment available in translation discuss the isotope "expected range," which the defence expert's data on excavated crusts weakened.²  None of the sources read here reports what the court made of the wax.

The isotope range is the weaker part of the 2005 case.  Part 2 explains that natural crusts on excavated pottery can fall well below it, though not, in the figures reported from the trial, as low as the second ostracon's values.  The stronger part does not depend on isotopes at all.  It is a layer sequence of the kind described in Part 2:  ink, then fresh tool marks, then a wax film, then a crust of burnt lime and ash.  Ten excavated ostraca from six sites showed none of it.

The obvious innocent explanation to test is conservation:  a restorer might have waxed a fragile surface, and in 1995 Mikrofokus found traces of glue and of steel, which it attributed to handling and storage.  Wax applied over an intact ancient crust would lie on top of it.  The reported relation is the reverse, so that simple treatment history does not fit.  If the film is paraffin, a product of modern refining, the overlying crust must be modern too.  But the wax's chemical identity was not established spectroscopically, and wax in general is not exclusively modern.  Nor does a crust younger than modern wax, by itself, date the underlying ink:  a restorer might remove or add surface layers.  The stronger argument is cumulative:  the fresh scratches affecting the letters, the intervening film, a crust whose texture and ash content the authors attribute to manufactured lime, the very low values on the second ostracon, and the difference from excavated controls.  An innocent treatment history would have to account for that whole pattern.  The controls' own conservation histories are not reported, which limits how confidently that comparison excludes restoration.

The 1995 and 2005 reports are often set against each other, but they agree on more than appears.  Both describe a crust of calcite with amorphous silica, and the plant phytoliths of 2005 are themselves a form of amorphous silica.  They differ on the shape of the calcite crystals and on whether the crust is natural.  No source read here shows that the two laboratories sampled the same spot, so they need not contradict each other directly.  The 1995 reports are known only through Rollston's summary, and the 2005 study is published with photographs and with its controls named.⁵⁶⁶  Neither commissioning, one through the owner and one through the prosecuting authority, is a reason to reject the work.  The 2005 study is preferred here because its observations are more specific, its controls were excavated, and its key relation, crust over wax, has not been disputed in any reply found in the routes searched.  The earlier published defences of authenticity, listed above, did not have that evidence before them.  Rollston himself later questioned the ostraca, in a 2003 article that the 2005 study cites but that was not available for this article.⁶

Part 10: The register, and what follows for Christian arguments

The table gathers the six questions of Part 1 for each object.  The units differ:  one row covers 16 fragments, one nine studied fragments, one a pair of ostraca, and the rest single objects.  "Writing" gives the graded assessment of when the writing was made, and "Current assessment" uses the same grade;  neither is more certain than the case discussion above.

What the forgeries cost, and what they do not

The forgeries matter most to arguments built on one spectacular object.  Taking each object in turn:

1. James ossuary. If its inscription is modern, or genuine but naming other men, a proposed first-century physical attestation of Jesus' family is lost. That is a real loss: an inscribed first-century box would have been independent of the manuscript tradition of Josephus and the New Testament. What remains is the literary evidence that James the brother of Jesus existed and was put to death in Jerusalem, from Josephus and Christian writers, weighed on its own terms in Jesus in Archaeology and Outside Sources. Excavated first-century inscriptions still name people in the Gospel story, the Caiaphas family ossuaries and the Pilate stone described in Part 1, though neither names Jesus or his family.
2. Ivory pomegranate. If its inscription is modern, a proposed inscribed object from the cult of the First Temple is lost. What remains for the First Temple is the biblical description, architectural parallels and excavated finds from Jerusalem and elsewhere, set out in Archaeology of the First Temple, which already sets the pomegranate aside.
3. Jehoash tablet and Moussaieff ostraca. These would have been the first royal Judean building inscription and two vivid administrative texts mentioning the House of Yahweh. Their loss removes direct, if unprovenanced, checks on 2 Kings 12 and on temple donations. The general historical outline of the monarchy did not rest on them. It rests on excavated material such as the Siloam tunnel inscription, cut into a tunnel that radiocarbon and uranium-thorium dates place in the Iron Age, the Tel Dan stele, Assyrian annals and the Lachish and Arad letters, catalogued in Archaeological Confirmations. Those texts have their own questions of reading and interpretation, which that article addresses.
4. Museum of the Bible fragments. The excavated Qumran caves and the scrolls tied to them are untouched. What the forgeries damaged was the scholarly datasets they entered, and the reputations of institutions that bought before testing.
5. Gospel of Jesus' Wife. Even on King's reading, the fragment was evidence about a later Christian debate, not about whether Jesus married. Its exposure is a reminder that forged discoveries appear on every side of the debate.

These cases do not show that most market objects are fake, because they are a selection of famous challenged objects rather than a representative sample.  They do not establish a general verdict on biblical archaeology:  the comparators show how independent context and documentation can strengthen particular finds, while each still has to be read and weighed.  Nor do they show that unprovenanced objects must be ignored.  The Mesha stele was never excavated, and the same family that supplied forged fragments also sold fragments the critics accept as probably genuine.  The right rule is graded:  the weaker the provenance, the more the object must establish about its own writing through other evidence.

Objections and responses

1. "Biblical archaeology is riddled with forgeries, so none of it can be trusted."  The cases here are a selection of famous objects that were challenged, so they cannot measure how common forgery is.  They also share one feature:  none was excavated or recorded in place before it reached the market.  The excavated Ketef Hinnom amulets and the documented Mesha stele show the difference that provenance makes, though each still had to be argued on its own evidence.  On the same subjects as the market cases stand inscriptions recorded where they were found:  the Siloam tunnel inscription on the tunnel wall in 1880, the Tel Dan stele under a destruction layer in 1993, the Caiaphas family ossuaries in a tomb opened by builders in 1990, and the Pilate stone in the theatre at Caesarea in 1961.^26302931  Their readings and identifications are argued over, but their letters were recorded in place before any dealer could handle them.  The proper conclusion is narrower:  single market objects with no findspot should carry weight only in proportion to what tests have shown about their writing.

2. "Christian collectors and institutions bought these because they wanted them to be true."  Some buyers were Christian, and buying before testing was a real failure.  But motive is not evidence about an object, and the record cuts across camps.  The Museum of the Bible paid for the study that condemned its own fragments and published it.³  The Gospel of Jesus' Wife, whose text cut against traditional belief, was announced and defended at Harvard.  Forgers make what buyers of every kind want.

3. "The court acquitted Golan and Deutsch, so the objects are genuine and the IAA conducted a witch hunt."  In the defence-side translation the judge said the opposite of the first claim:  the acquittal "is not to say that the inscription on the ossuary is real and authentic."²  A criminal court asks whether forgery was proved beyond reasonable doubt against a named person.  Scholars ask which explanation is more probable.  The trial did expose weaknesses in the IAA's case, notably the isotope reference range and the way the committees reached unanimity, and those weaknesses are real.²²  Neither they nor the acquittal show the IAA's motives, which no source read here documents.

4. "The IAA proved the James ossuary a forgery in 2003."  The committee's materials findings were serious.  But its own isotope paper reported a normal value in the last letter, and its microscopist later testified, in the defence translation, to genuine biological film in the same letter.  The proposed general isotope boundary was challenged by excavated pottery, while three excavated ossuaries remain more closely matched controls; its central observation allowed "cleaned" as an alternative to "engraved."¹²¹³²  "Unresolved" is the accurate word.

5. "The James ossuary is authentic, and the scientists who doubted it were biased."  Bias is not shown by disagreement, and the defence side has interested parties too.  One fair test of bias is what the same scientist did with a similar object.  In 2011 Yuval Goren, the committee's microscopist, and the archaeologist Boaz Zissu published an ossuary the IAA had acquired, "purportedly from a burial cave in the area of the ʾElah Valley," inscribed "Miriam daughter of Yeshua son of Caiaphas."  They examined its patina "with emphasis on the inscribed area" and concluded that "the patina and the inscription should be considered authentic beyond any reasonable doubt."³²  So the same microscopist judged an unprovenanced ossuary bearing a Gospel name genuine when he read its surface that way.  That tells against a blanket bias.  It does not show that his James observations were right, and a critic of patina tests can turn it around:  it shows how much such verdicts rest on one examiner's reading of a surface.  The question remains open because the evidence inside the letters is mixed and no published study by laboratories outside the dispute has sampled and imaged the letters of both halves with proper controls.  Even a genuine inscription would name three common names, and the judge held the link to Jesus of Nazareth unproven.⁴⁸²

6. "These objects passed laboratory tests."  So did the Moussaieff ostraca in 1995, when two laboratories called their crust natural.⁵⁶  The Gospel of Jesus' Wife papyrus tested as old and its ink as a carbon black like ancient inks, and one laboratory went on to say the writing should be dated by the papyrus.¹⁵⁴⁰  Tests of material answer questions about material, and reach the writing only through a stated link.  As the scientists who exposed the Schøyen fragments put it, the best such tests can do is find "nothing ... to contradict the assumption of genuineness."¹

7. "A forger could not combine such skill with such clumsiness."  This was King's argument for the Gospel of Jesus' Wife and Pardee's for the ostraca.⁸²⁴  It assumes that the hard parts, old papyrus or sherds, convincing ink and convincing papers, are hard to obtain.  Old blank material can be bought, carbon ink is easily made, and the text can be copied from a published edition, as Askeland showed for the John fragment.⁴  The papers behind the Gospel of Jesus' Wife failed simple checks, and the ostraca's crust was made with burnt lime.  The clumsy part, the writing itself, is exactly where the forger's modern sources show through.

8. "The isotope method has been discredited, so all the forgery verdicts collapse."  The trial challenged the proposed breadth of its natural reference range.  That limits its use as a universal test, while leaving the matched-control and coating observations to be assessed individually.  But the verdicts on the Museum of the Bible fragments, the Gospel of Jesus' Wife and the Jehoash tablet do not rest on it.  On the ostraca, the lime crust lying over wax is independent of it, and the second ostracon's values fall far below every value the defence's expert reported for an object from a named excavation site.  The one comparable value in the chart of his results that defence scientists published in 2013 carries a label, "Ostracon #2 Almana," that appears to name the widow ostracon itself.⁶⁶²³⁴

9. "If expert palaeographers were fooled, palaeographic dates for genuine manuscripts mean nothing."  Palaeographers were fooled by unprovenanced objects whose texts had been copied from published sources, which is exactly the situation in which handwriting proves least.  Where texts come from known contexts, palaeography can be checked independently, and for the Dead Sea Scrolls the first radiocarbon tests largely agreed with it.³⁸  Handwriting is evidence, not proof, and it is strongest alongside provenance.

10. "The pomegranate's inscription has been shown to run into the ancient break, so it is genuine."  That is Lemaire's observation, repeated in the joint report of 2007, and the editor of the Biblical Archaeology Society reports the same.¹¹⁴⁷  At the same session the committee members found it "difficult to determine" whether one stroke reached the break and saw another stop short of it.¹¹  Opposed readings of one microscopic junction, after a joint look, are a reason for better imaging, not for declaring victory.  The joint examination did settle one point for Lemaire:  the yod once cited as evidence of forgery is now accepted as normal.¹¹

11. "The Israel Museum's own 2019 study shows the pomegranate inscription is fake."  That study expressly did not examine the inscription.  Its argument is that an inscription five centuries younger than the object is unlikely.⁵³⁵³  Precious objects were sometimes kept for centuries, so the argument lowers the plausibility without deciding the case.⁵⁵

12. "Unprovenanced objects should simply be ignored, since buying them rewards looting."  The ethical point is serious, and it is a reason not to buy.  But as a rule of evidence, total exclusion would discard the Mesha stele and the probably genuine fragments that passed through the same dealers' hands.²⁰¹  The weaker the provenance, the more the object must prove for itself.

13. "No forger would write fifteen lines;  the first rule of forgery is to keep it short."  This is the Biblical Archaeology Society editor's argument for the Jehoash tablet.⁴⁷  Length does raise a forger's risk, and that is how the tablet's critics think it was caught:  the more Hebrew a modern writer composes, the more chances for modern idiom.  Much of its text is also built from phrases of Kings and Chronicles, which made length cheap.⁷  A long text is a test the forger must pass, not evidence that he could not have tried.

14. "Chaim Cohen showed the Jehoash tablet's Hebrew is genuine."  Cohen's own claim is narrower:  he did "not know whether or not this inscription is genuine," and contended only that it could not be proved a forgery by philology.⁹  His work does weaken the most quoted single argument, the "modern Hebrew" phrase in lines 10-11, by offering a reading that fits ancient building language.⁹  It does not make the tablet's many unattested usages attested, and Greenstein's point that each needs its own rescue remains.¹⁰

15. "The Mesha stele proves that unprovenanced objects are reliable."  It proves less and more than that.  Its documented chain, from Klein's measured sketch of 1868 to the squeeze and the rejoined fragments, rules out a dealer's forgery after 1868.¹⁹²⁰  Its antiquity then rests on its script and language.  Few market objects have a chain like it, which is exactly why it is a useful comparator rather than a precedent for trusting market objects.

Issues and unanswered questions

1. The judgment was not read. Everything said here about the court's reasoning comes from the IAA's statement, a defence-side partial translation and the Biblical Archaeology Society's report. The Hebrew judgment is sold by Israeli legal databases for a small fee. Reading it would settle which objects were counts, what the court made of the wax on the ostraca, Goren's own explanation at trial of the film in the Ain, and the full reasoning on the ossuary. Until then every court quotation here carries the caveat of a party's selection and translation.
2. The excavated isotope data are only partly published. Aldo Shemesh's 56 measurements, which did most to undermine the isotope method at trial, can be read item by item from a chart that defence scientists printed in 2013, but his report itself is unpublished. That chart's caption calls all 56 items excavated, while five of its labels name no site and one appears to be the disputed widow ostracon.³³ Published with sample types, contexts and pretreatment, the data would show how far natural crusts really range and whether the low pottery readings come from fired calcite.
3. An independent study of both halves of the ossuary inscription was not located. The most useful new evidence would be high-resolution imaging and sampling inside letters of both halves by laboratories outside the dispute, with excavated ossuaries as controls. Rosenfeld's claim of patinated scratches crossing letters also needs independent confirmation.
4. A full three-dimensional study of the pomegranate junctions was not located. The 2007 joint report narrowed the dispute but did not end it, and the effect of the reported restoration on the evidence is unknown.
5. Several primary studies were not obtained. These include Goren and colleagues' 2004 paper on the Jehoash tablet in Tel Aviv, the Geological Survey's 2002 report and later defence studies of the tablet, Chaim Cohen's 2012 study of the tablet's last three lines, Christopher Rollston's 2003 and 2004 articles on unprovenanced inscriptions in Maarav, and the 2019 Raman study of the Gospel of Jesus' Wife ink. The Jehoash and ink conclusions above are limited accordingly.
6. The lime-depilation evidence for the Museum of the Bible fragments is ambiguous. The team's own laboratory noted that burial carbonates and collagen decay can mimic it.³ The verdict stands on the ink relationships, but the chemistry deserves a cleaner test.
7. Most other post-2002 fragments remain untested individually. The pattern of their market origin is suspicious, but each fragment's status should be decided by examining it, not by association.
8. Other relics are outside this article. The Talpiot tomb, which concerns statistics more than forgery, is treated in Jesus in Archaeology and Outside Sources; relics and pilgrimage sites belong to Sacred Sites, Relics, and Pilgrimage Memory.

Verdict

Laboratory science is good at saying whether a material is old and what it is made of.  It reaches the writing on an old material only through a link:  the writing and the material's damage or crusts seen in sequence, a foreign material in the wrong layer, or a text that betrays a modern source.  The cases that were solved were solved that way.  The ink in the Museum of the Bible fragments runs across cracks, off broken edges and over burial deposits.  The Gospel of Jesus' Wife was very probably written by the same hand as a copy of a 1924 book.  On the Moussaieff ostraca a lime crust lies over wax over fresh scratches.  The Jehoash tablet's case is weaker in kind:  it rests on many unusual Hebrew usages and source-like borrowings.  The critics add these up to forgery, and its main defender answers some of them one by one without claiming the tablet is genuine.

The two most famous objects remain open.  The James ossuary's box is ancient.  Inside its letters a soft coating and low isotope values stand against natural film and a normal value in the last letter of "Jesus," in the very half that the prosecution called forged.  The pomegranate is ancient, and after a joint examination its inscription still turns on what two strokes do at an old break.  Neither should carry more than conditional weight in an argument for Christianity.  The court and the committees reached real decisions, but under standards that do not settle these questions:  a criminal acquittal on reasonable doubt, and committee votes whose weight depends on the observations behind them.

For apologetics the result is a discipline with a real cost.  Arguments that leaned on these objects lose proposed corroboration, most of all the ossuary's.  The historical case for the Bible's setting and for the existence of Jesus and James did not rest on them.  It rests on excavated finds and well-transmitted texts, each of which has to be weighed on its own evidence.  The lesson is to treat any single, sensational, unprovenanced object as a hypothesis until its writing, not only its material, has been tested.

Sources