In 1889, as part of excavations in the ancient city of Koptos in Egypt, archaeologists claimed to have made a remarkable discovery. As one of the Roman-era houses was being unearthed, the scientists noticed that one of its walls was hollow. Inside was a book, made of papyrus, which had been carefully and deliberately concealed there hundreds of years before. The book contained the writings of the Greek-speaking Jewish philosopher Philo of Alexandria. Even more astonishing, leaves of the Gospel of Luke were found among the waste papyrus used to stuff the leather cover of the book.
Scholars who examined the text deduced that the writings of Philo were from the third century; the scraps of Luke, according to mid-20th-century papyrologist Colin Roberts, thus must also have been copied even earlier than the manuscript of Philo’s works. Because the book had apparently been hidden away in a niche in a house, Roberts argued, it must have been placed there either when the emperor Diocletian destroyed the city in 292 CE or during Diocletian’s persecution of Christians in 303-305. In either case, this book had somehow survived.
When examining some manuscripts held at Magdalene College, Oxford, Roberts made another, equally interesting discovery. There he found more fragments of papyrus, this time from the Gospel of Matthew. Roberts identified these as coming from the same codex (the academic term for an ancient book with pages) as the scraps of Luke from Koptos. It was not unusual for books to be divided up and dispersed like this in antiquity. Soon other scholars joined the conversation. Some fragments of Matthew found in Barcelona were assigned to the same hand but now dated earlier, to the second century. Roberts pronounced that they were all evidence for a book that had contained at least two Gospels (Matthew and Luke) and that must have been written sometime before part of the book was repackaged and buried in Egypt during the Diocletianic persecution.
“What he found is poised to upend what we thought we knew about the history of the New Testament.”
On Christmas Eve 1994, the Times of London ran the story of the manuscript’s discovery on the front page with the title, “Oxford Papyrus is eyewitness record of the life of Christ.” It was based in part on the arguments of a German scholar, Carsten Thiede, who thought that the Oxford and Barcelona fragments were from the first century. Despite the objections of many of his fellow scholars, the story gathered steam: Thiede published a popular book based on his theory and his work was featured in The New York Times and other prominent media outlets.
At the same time, more well-respected and less controversial scholars argued that the book was originally a second-century collection of all four canonical Gospels. If we had the full copy, it would be the earliest collection of the Gospels and the most important manuscript of the Bible ever discovered. Even in its fragmentary state people saw (and continue to see) it as evidence that as early as the second century, some early Christians had a set collection of the canonical Gospels that they used in their worship.
This story, however, is based on some completely unverified “facts.”
The tale of this manuscript is told and dismantled in the recently released God’s Library (Yale, 2018), a tour de force from Brent Nongbri, a renowned expert on early Christian manuscripts. In writing his book, Nongbri “went back to the very beginning” and asked, in many cases for the first time, what do we know about the manuscripts that are supposed to serve as early evidence for the life of Jesus and the practices? Were they really found where we are told they are found? And are they really dated correctly?
Related in Travel

The New Room With a View: Madrid’s Gran Hotel Inglés
What he found is poised to upend what we thought we knew about the history of the New Testament.
In the case of the Philo codex from Koptos that contains the fragments of Luke, it was first acquired by Jean-Vincent Scheil, a scholar of the ancient Near East, during a trip to Egypt. Scheil stated in his 1893 publication that the pages of the Gospel of Luke were loose pages found between pages 88-89 of the codex. They weren’t stuffed in the binding at all. As Nongbri says, it’s likely “they were slipped in [the Philo codex] for safekeeping.” We’ve probably all done that.
It might seem like a small thing, but the dating of the pages of the Gospel of Luke was based on the idea that they were used to construct the leather cover of the buried book of Philo. Then things then got even stranger. Looking at some of Scheil’s correspondence, Nongbri realized that Scheil did not discover the book himself. He had actually bought it in Luxor in 1891. The story of the book’s discovery came from the book dealer who sold it to him. The fragments that Roberts identified at Oxford had also been purchased in Luxor.
“None of this is good news for those who, for religious reasons, want to use the existence of early manuscripts of the New Testament as evidence for the accuracy of scripture…”
All of which means that no one actually knows where the fragments come from. Because of the significance of ‘find stories’ to scholars, book dealers often produce narratives about where the books were unearthed. When you actually look into the facts about important ancient manuscripts, it turns out that the just-unearthed-after-centuries or discovered-by-a-goatherd stories are second- or even third-hand stories.
The first time a scholar actually laid eyes or hands on these texts was often in antiquities dealerships in Cairo, Luxor and elsewhere. These men and women were enterprising academics and amateur well-heeled collectors looking for manuscripts in the ethically cloudy world of the antiquities market. According to his secretary (who may not be an entirely reliable source), Martin Bodmer, a wealthy Swiss collector who wanted to assemble a collection of important World Literature, received one package of ancient papyri at the airport as he passed through Cairo on his way home from New Delhi.
As for the date of these Luke fragments, Nongbri noticed that while Roberts assumed they were second-century (and thus enormously important), other early experts assigned them to the third or even fourth century. And without the fictitious find story, we have no reason to assert that the fragments must be earlier that the fourth century. Any time between the second and fourth century is reasonable.
This might not seem like a huge difference but an early second-century manuscript could potentially have been copied by someone who was just barely removed from Jesus himself: an acquaintance of someone who knew the apostles, for example. Because of this, people use manuscripts like these to make grand claims about the accuracy of the stories about Jesus. A fourth-century manuscript, however, was written so much later that, as important as they are, no one can make exaggerated claims about eyewitness testimony.
Throughout the book Nongbri goes back to the basics of the discovery and dating of some of the world’s most famous and valuable manuscripts. What he found was that sometimes the reasons for dating the first copies of the New Testament are a bit “sloppy” and fail to take account of the physicality of these texts themselves. Some scholars, Nongbri told me, “find a sample of handwriting datable to the early second century that looks similar to an early Christian manuscript, and then claim that the early Christian manuscript must be of a similar date (all the while ignoring other equally similar samples of handwriting from later dates).”
“With Christian manuscripts,” he added, “there is a temptation to try to restrict a date to the early end of the range because we all would love to have artifacts that bring us closer in time to Jesus and his earliest followers. But for reasons I outline in the book, pretending that we can be so precise doesn’t really own up to all the complexities of how we assign dates to manuscripts.”
None of this is good news for those who, for religious reasons, want to use the existence of early manuscripts of the New Testament as evidence for the accuracy of scripture, the life of Jesus, or the status of the Bible in the first centuries of the Common Era. Anyone who previously thought that the manuscripts of the New Testament proved anything about the accuracy or authenticity of God’s message should read Nongbri’s book.
What Nongbri ultimately wants, however, is for people to care about the manuscripts on their own terms: “For too long, we have tended to view early Christian manuscripts almost exclusively as carriers and preservers of texts…But these manuscripts are archaeological artifacts that have their own stories… The first step toward the production of reliable knowledge about early Christian manuscripts really is just being honest about what we don’t know. From there, we can build sound hypotheses based on actual evidence rather than wishful thinking.”