Witnesses Found After the Manuscript

A word about the word medieval

Everyone who writes about the Gospel of Barnabas, this site included, calls the surviving text medieval. It helps to say exactly what that means here.

The manuscript in Vienna is a sixteenth-century object. Its Italian was written in the late Middle Ages or the early Renaissance, and it talks the way a book of that time talks. It knows a bank and a gold coin. It says a jubilee now cometh every hundred years. It dresses a condemned man in a fool’s white robe.^[3] None of this is surprising, and none of it is in dispute. A translation always speaks the language of the day it was made. A French Bible printed in 1530 sounds like 1530, whatever the age of the Hebrew behind it.

So when this site says medieval, it means the copy and the translation we hold — the outer layer. The pages If Fabricated? and A Historical Hypothesis argue that this outer layer does not settle the age of what lies under it. In the words of the second page, the surviving text shows a sixteenth-century feel to parts of its language and cultural detail, sitting alongside theological material and a narrative framework that may preserve a much older tradition.^[4] This page puts that reading to a hard test. If the framework is old, it should sometimes agree with old books in ways the translator could not have arranged. If it is not, the agreements should all melt away into things a medieval Italian could have read.

Why a book found later makes a fair test

Suppose Barnabas and a book found in 1896 share a detail, and no earlier text has it. There are only three ways that can happen. Two writers can reach the same idea on their own. A text that carried the detail can have existed and then disappeared, leaving no trace but the two books that drew on it. Or the detail can belong to an older tradition that stands behind the Italian text. The third possibility is the one that matters for this site. It is also the easiest one to assert loosely. That is why the test has to be run strictly — and why most of the candidates below do not survive it.

One rule governs everything that follows. For every candidate we look for the earliest route by which the detail could have reached a reader in the Latin West. Where there is none, we say so. Where there is one, we also say what century it comes from. An earlier route that turns out to be a second-century or fourth-century book is not the same kind of answer as a route that runs through a fourteenth-century Italian bookshop. The first keeps the detail old even while it takes Qumran out of the picture. The second does not.

How we ran it

We laid the 3,464 sentences of the Gospel of Barnabas beside 116 texts: Jewish books written under the names of biblical figures, early Christian writings left outside the New Testament, and the Church Fathers. A computer pass flagged every pair of sentences that shared real wording. Then came the slow part — reading each flagged passage in full, quoting it word for word, and asking the two questions above.

Each candidate ends with one of three results. Open means no route before 1709 could be found. Ancient route means the detail is explained by a text written before about 600 CE. Medieval route means a late-medieval reader could have held the source in his hands.

Two headlines, then the cases. First: not one candidate requires a text that reached Europe after 1709. Second: nine candidates keep an older-than-medieval explanation alive — either because no earlier route exists, or because the earlier route is itself ancient. We begin with the case that has been in the scholarly literature longest.

The retreat to Damascus

In chapter 139 Jesus slips away when the people of Nain try to make him king. His disciples find him, and he explains: there shall rise against me the chief priests with the elders of the people, and shall wrest authority to kill me from the Roman governor, because they shall fear that I wish to usurp kingship over Israel. Then the group gathers again, in Syria: The day following there came, two by two, thirty-six of Jesus’ disciples; and he abode in Damascus awaiting the others.^[5] Damascus returns in the last pages of the book. After the arrest, there went all, saving twenty-five of the seventy-two disciples, who for fear had fled to Damascus.^[6]

In 2001 Rodney Blackhirst, a scholar at La Trobe University in Australia, laid this scene beside two of the Dead Sea Scrolls.^[7] The Damascus Document remembers a community that went out of the land of Judah to sojourn in the land of Damascus. It calls them those who entered into the new covenant in the land of Damascus. The sect called its founder the Teacher of Righteousness and its enemy in Jerusalem the Wicked Priest, and the Habakkuk Commentary tells how that priest pursued after the Teacher of Righteousness… even to his place of exile.^[8] A teacher, a group under threat, a withdrawal to Damascus, a pursuer from the priesthood: Blackhirst read the Barnabas scene as a possible echo of that older pattern. He did not read it as proof that the book is ancient. In a 2011 interview he called the surviving work clearly a medieval document, and said that whether it owes a debt to an earlier document is a matter of debate.^[9] His Damascus study is about exactly that debt.

Any honest weighing starts with what a medieval reader could reach. The Acts of the Apostles already has disciples living in Damascus before Paul arrives there (Acts 9:2, 10, 19). And the Recognitions of Pseudo-Clement — a Latin story about the apostle Peter, something like an early Christian novel, copied all through the Middle Ages and printed in 1504 — tells very nearly the story Barnabas tells. After a riot in the Temple, the believers slip away: Then before daylight we went down to Jericho. Three days later comes word that that enemy had received a commission from Caiaphas, the chief priest, that he should arrest all who believed in Jesus, and should go to Damascus with his letters… because he believed that Peter had fled to Damascus.^[10] Fear, flight, Damascus, the high priest’s letters — every piece Barnabas uses is here, in a book an Italian reader could hold. The Damascus Document’s Damascus, by contrast, is a place of covenant and exile, wrapped in a vocabulary Barnabas never touches. On the first question the answer is plain: Qumran is not needed.

Now the second question. The Recognitions is a fourth-century book. It stands in the Decretum Gelasianum’s own list of rejected writings, as the Itinerary in the name of Peter the apostle, which is called the nine books of the holy Clement. A few lines below comes the Gospel in the name of Barnabas.^[11] And it is the book in which Paul walks on stage unnamed, as that enemy — the Jewish-Christian memory of Paul that this site’s page Paul in Jewish Texts describes. So both candidate sources for the Damascus scene are ancient, and both grew in the same soil — a Jewish or Jewish-Christian community at odds with the priests of Jerusalem. Blackhirst’s road leads to a Jewish sect of the second century BCE. The other road leads to a Jewish-Christian story of the fourth century CE, condemned by the same decree that condemned Barnabas. The only reading that keeps the scene medieval is Acts 9 plus the writer’s own imagination. That is possible. But it explains less than either old road does, because neither the fear nor the pursuit is in Acts.

This is the pattern the whole page keeps meeting. The test fails to reach Qumran — and still refuses to stop in the Middle Ages.

Six Gospel readings the Italian harmonies do not have

The most concrete agreements with later-found books are tiny, and they sit inside Barnabas’s Gospel quotations. To see them you need one piece of background. In the second century a Syrian Christian named Tatian stitched the four Gospels into a single continuous story, the Diatessaron. The book itself is lost, but its wording survived in the harmonies — blended Gospels — that were copied in many languages for a thousand years afterwards.^[12] In 2002 Jan Joosten showed that where Barnabas quotes the Gospels, it often agrees with these harmonies rather than with the plain Latin Bible. The page If Fabricated? weighs what that agreement can and cannot prove. Here we add the count.

We checked all sixteen of Joosten’s examples in our text. Ten match the Venetian and Tuscan harmonies — Italian manuscripts of the thirteenth to fifteenth centuries that an Italian writer could have used. Six match only witnesses published after 1709: a Middle Dutch harmony from Liège (printed 1835), the Old Syriac Gospels (Syriac is a dialect of Aramaic; these manuscripts came to light in 1858 and 1892), and a Persian harmony (1951).^[13]

Here are the six, with the Barnabas wording beside the Vulgate, the Latin Bible of the medieval West. You do not need to follow every detail; the point of each is the same. Barnabas agrees with a book that reached Europe after 1709, and with nothing an Italian writer could have held. In Luke 2:4 Joseph departs from Nazareth, a city of Galilee (di nazaret citta di galilea), where the Vulgate has the reverse order, from Galilee, from the city of Nazareth. Barnabas’s order is the Old Syriac order — and, we found, the order of the Arabic Diatessaron published in 1888: from Nazareth, a city of Galilee.^[14] In Matthew 15:2 the scribes complain about the traditions of our elders (nosstri uechij), where the Vulgate and the Greek have plain the elders; Liège and the Old Syriac have our. In Matthew 17:7, after the Transfiguration, Jesus went down and raised up his disciples, saying: ‘Fear not’ — the Old Syriac has Jesus raising them, where the Greek has him touching them.

In Luke 15:22 the father calls for new robes (uestimenti noue), where the Vulgate has the first robe; only Liège has a new garment. In Luke 15:13 the prodigal wasted all his substance with harlots; Liège and the Old Syriac put the harlots here, while the Vulgate mentions them only in the elder brother’s complaint. And in Matthew 21:34 the owner sends his servants when the time was come to collect the wine (racholgere il uino); Liège has the wine, where the Vulgate has the time of fruits.

What do six small readings prove? Joosten’s answer is a lost Italian harmony, older than the ones that survive, which kept the oldest Latin wording — older than Jerome’s Bible, though Latin harmonies went on carrying it for centuries after him — that later copies corrected away. Two other scholars, August den Hollander and Ulrich Schmid, replied in 2007 that a late and free-handed text is a poor witness to second-century readings, and that several of these agreements could arise on their own.^[15] Both points stand. Anyone can carry harlots from the elder brother’s speech into the prodigal’s story, and the time to gather the wine is what a vineyard parable naturally says. The first four readings are harder to produce by accident.

Here too the second question changes the picture. A lost harmony with that older Latin wording is not a guess without a shelf to stand on. Victor, bishop of Capua, had a Latin Gospel harmony copied in 546; it is still in Fulda, and scholars call it the Codex Fuldensis.^[16] Latin harmonies descended from the Diatessaron were circulating, then, in the same century as the Decretum. The readiest explanation of these six readings is a Latin stage in the history of the text, old enough to carry Old Latin wording. A route, not a proof — but a route already open in the sixth century. And it means that the harmony evidence, the strongest card in the internal case for a late forgery, points past the Italian harmonies it is usually said to prove.

Four angels with four jobs

Barnabas names four angels, Gabriel, Michael, Raphael and Uriel, who carry Jesus out of the world and later stand guard in his mother’s house. In chapter 220 Jesus explains their duties: Gabriel, who announceth God’s secrets; Michael, who fighteth against God’s enemies; Rafael, who receiveth the souls of them that die; and Uriel, who will call every one to the judgment of God at the last day.^[17]

The same four names open a famous scene in the Book of Enoch: And then Michael, Uriel, Raphael, and Gabriel looked down from heaven and saw much blood being shed upon the earth. Enoch also gives Raphael a job close to the one in Barnabas: Raphael, one of the holy angels, who is over the spirits of men.^[18] And the Apocalypse of Peter, in the Ethiopic version first published in 1910, gives Uriel almost exactly the Barnabas role. For him hath God set over the rising again of the dead at the day of judgement.^[19]

Two of the three agreements have an earlier route, and both routes are old. The four names, Uriel included, stand in the Etymologies of Isidore of Seville, a seventh-century encyclopedia and one of the most copied books of the Middle Ages. They stand in the Jewish bedtime prayer, printed in early modern prayer books. And they were venerated together in the devotion to the seven archangels that spread out from Palermo after 1516.^[20] Uriel’s job at the judgment is in the second book of the Sibylline Oracles, a Christian-Jewish poem of the second or third century, where Uriel breaks the bolts of Hades and brings every soul to judgment; the Greek text was printed at Basel in 1545.^[21] Only Raphael, who receiveth the souls of them that die, has no Christian witness before 1709. Enoch 20:3 is the nearest text we found.

There is also a nearer comparison. In Muslim tradition the four great angels are Jibril, who brings revelation; Mikail, who brings provision; Israfil, who blows the trumpet of judgment; and Azrail, the angel of death, who takes souls. Barnabas has the Latin names and, for two of the four, the same duties: Uriel does Israfil’s work, Raphael does Azrail’s. The agreement is in the duties, which are old in Enoch and the Sibyl, not in the names, which are Latin — and on this site’s reading that is the expected shape: older material that a later tradition could confirm, not a later tradition’s insertion. Where does that leave the four angels? The names are ancient and explained. Uriel’s role has a 1545 route back to a second-century poem. Raphael’s role stays open, and the open mark has to be a fair one. Of its two nearest witnesses, Enoch was out of reach of an Italian desk; Muslim tradition was within reach of a translator who knew Islam — the translator this site’s own model assumes. So the item counts for neither side.

Abraham, the idol-seller, and the clouds

Barnabas gives the boy Abraham a father who makes idols for a living: every day I make a god, which I sell to others to buy bread.^[22] Then it has the boy reason his way to God: looking round upon the stars, the moon, and the sun, he thought that they had been God. But after considering their variableness with their movements, he said: ‘It needs must be that God move not, and that clouds hide him not.’^[23]

The order — stars, then moon, then sun — is the Qur’an’s (6:76–78). But the Qur’an’s objection to each of them is that it sets. The clouds come from somewhere else. The nearest text is the Apocalypse of Abraham, a Jewish work of the first or second century that survived only in Slavonic and was first printed in 1863. There the boy says: But even it [the sun] I do not call god, because at night and by clouds its course is obscured.^[24] Two ideas, movement and cloud cover, appear together only in these two books among everything we read. The idol-selling father is older than Islam too. The Apocalypse of Abraham opens with Terah’s idols being carried to market, and the midrash Genesis Rabbah — a rabbinic storytelling commentary of the fifth century — tells the same tale; a medieval Latin reader could meet it in Raymond Martini’s Pugio fidei of 1278.^[25]

A sharp mind can reach clouds hide it, so it is not God on its own. That keeps the clouds an open case rather than a strong one. It is still the single most specific agreement with a later-found book in the whole Abraham story, and the idol-market beside it has a route that is rabbinic and ancient.

The shorter cases

The painless birth (chapter 3). The Odes of Solomon, a Syriac hymnbook published in 1909, say the Virgin bore the Son but without pain… she did not require a midwife. But the Latin Gospel of Pseudo-Matthew, a medieval favourite, already has both of Barnabas’s details together: no spilling of blood in his birth, no pain in bringing him forth, and a light that filled the cave. The motif goes back to the Protevangelium of James, a second-century book about Mary’s childhood.^[26] An ancient route.

Adam and circumcision (chapters 22–23). The Gospel of Thomas, found at Nag Hammadi in 1945, has the disciples ask whether circumcision is useful; Barnabas has the same question with a different answer.^[27] A shared question and two different answers prove nothing.

The weeds pulled up (chapter 132). Thomas and Barnabas both say pull up where the Gospels say gather. So does the Venetian harmony, and the Vulgate already uses uproot (eradicetis) in the next verse. Joosten himself lists the Persian harmony and Thomas here, as If Fabricated? notes.^[28] A medieval route.

The sevenfold hell (chapter 135). The Apocalypse of Peter punishes sinners according to their sins, and so does Barnabas. But Barnabas builds its hell out of seven regions for the seven deadly sins — the list Gregory the Great fixed around 590 — with a punishment shaped to each sin. That plan comes from the Latin Vision of Paul, which the Decretum lists as the Revelation which is called Paul’s. The seven gates are shared with the Qur’an (15:44). The sevenfold plan itself is Gregory’s and the Vision of Paul’s; the shared gates this site reads as it reads the other agreements with Muslim tradition — older material a later tradition could confirm, with the direction left open.^[29] This is the structure the Dante comparison, examined on If Fabricated?, fails to match: Dante’s hell has nine circles, and his sevenfold scheme belongs to Purgatory. Nothing specific is shared with Peter either. An ancient route, with the seven gates shared by the Qur’an.

Herod, and a king on a seat (chapter 217). The Gospel of Peter seats the mocked Jesus on the seat of judgment. Justin Martyr already had the seat in the second century, and Barnabas does not follow Peter at any point where Peter is distinctive.^[30] An ancient route, through Luke 23 and Justin.

The Lives of the Prophets. Barnabas’s story of Haggai, Hosea and Obadiah (chapters 185–188) shares nothing with this ancient collection. Its nearest match is the legend of a fourth-century Egyptian monk — the first case on the companion page The Company Barnabas Keeps: Two Shelves, which sorts the text’s closest matches by age.^[31]

Where this leaves us

Nothing in the Gospel of Barnabas requires knowledge that reached Europe only after 1709. That is the first result, and it stands.

The second result is the one the usual argument never reaches, because the usual argument stops at the first question. Of the nine items that could, in principle, point past the Middle Ages, three stay open: Raphael’s role, the clouds over Abraham’s sun, and the four harder Diatessaron readings. The rest resolve into routes that are themselves ancient — the Recognitions of the fourth century, the Sibyl of the second or third, Isidore of the seventh, the Protevangelium of the second, Gregory and the Vision of Paul of the sixth, and a Latin harmony tradition on the shelf in 546. The test keeps failing to reach Qumran. It also keeps failing to stop in a fourteenth-century Italian bookshop. Whichever road the Damascus scene travelled, it travelled an old one.

This is what the rest of the site’s argument predicts. If Fabricated? shows that the internal case for a late forgery rests on a few small cards — a Dante echo, a jubilee, a handful of harmony readings — and that the Dante card does not survive inspection. This page takes the strongest card, the harmony readings, and finds that four of them point to a Latin stage older than the Italian harmonies. It takes the Dante-like hell and finds Gregory the Great and a Latin apocalypse the Decretum knew by name. A Historical Hypothesis describes a surviving text with a sixteenth-century surface over a narrative framework that may preserve a much older tradition. The nine items here are the kind of material that split predicts: a framework whose nearest relatives, wherever they can be dated, come from the first six centuries. And the Decretum — the earliest memory of a Gospel under Barnabas’s name — turns out to list, beside that name, the very books that stand behind two of the scenes examined on this page.

Keep the limits in view. None of this proves that the ancient Gospel of Barnabas survived into the Italian text. The name of the Prophet where a title may once have stood, the bank in chapter 16 and the coins in chapters 54 and 214 remain what this site has treated them as throughout: marks of the hands the text passed through on its way into Italian. Where the text agrees with Muslim tradition in substance, as it does on the angels’ duties, this site reads the agreement the other way round — as the kind of older material a later tradition could confirm — and the text’s own differences from the Qur’an, on the heavens and on the Messiah’s title, are the reason. What this page shows is narrower, and for that reason harder to argue away. Wherever the surviving text shares a scene or a reading with a book it should not know, there is almost always an older book that explains the agreement — and that older book is almost never medieval.

Set this page beside the rest of the site’s evidence and it stays what it was at the start: a helper, not the main witness — and the test the forgery theory should have won easily and did not. Hold the two findings together one last time. The copy is medieval. The company it keeps is not. What that weighs beside the rest of the file — the decree, the tomb, the Latin under the Italian — is set out in The Case as It Stands, the site’s summing-up.