The charge is easy to state and serious if true: copyists changed the New Testament so many times over so many centuries that no one can know what its authors wrote. The manuscript record does not support it. The copies do contain hundreds of thousands of small differences. That large number exists because so many copies survive. The same abundance lets scholars compare the copies, see where they differ, and recover the original wording with a confidence no other ancient book allows.
That is the case Dr. Daniel Wallace makes. He has debated Bart Ehrman, the best-known scholarly skeptic of the New Testament text, and he describes the nonprofit he leads, the Center for the Study of New Testament Manuscripts, as working to digitize every Greek New Testament manuscript and make the images public. His argument can be tested the way any claim about evidence should be tested. Ask what we would expect to find if the corruption charge were true. Then look at what is actually there.
What would a corrupted New Testament look like?
If the text had been reshaped beyond recovery, certain things should show up in the record. The surviving copies would be few and late, so nothing earlier could check them. The earliest copies would disagree with later ones on major matters. Core beliefs would appear only after some convenient political moment. And the differences between copies would be large and scattered, with no clear way to tell an original reading from an alteration.
If the text had been transmitted faithfully, the record would look different. There would be many copies, some of them early. Most differences would be small and easy to explain. The disputed places would be known, limited, and openly discussed. Early and late copies would agree on the substance.
The question is which picture the evidence matches. Wallace starts where the work itself starts. The field is called textual criticism, which sounds hostile. Wallace says otherwise: “the word criticism doesn’t mean that we’re being critical of the Bible. It really is virtually a synonym for research.” The task is recovering the wording of a document whose original is gone. That applies to every ancient literary work, he notes, and even to more recent texts such as Shakespeare’s plays and Lincoln’s Gettysburg Address. The New Testament is not under special suspicion here. It faces the same problem as everything else from antiquity.
Where does the half-million figure come from?
A textual variant is any place where manuscripts differ in wording. Wallace counts a changed word order, an added or dropped word, a whole verse added or missing, and even a spelling difference. All of them count.
The method of counting matters. Variants are counted by the different readings, not by how many manuscripts carry each one. Wallace’s example is John 4:1. Some manuscripts say that Jesus knew the Pharisees had heard. Others say the Lord knew. The King James Version follows the second reading: “When therefore the Lord knew how the Pharisees had heard that Jesus made and baptized more disciples than John” (John 4:1). If a thousand manuscripts read “the Lord,” Wallace says, that is still one variant. He also says apologists have miscounted variants for about fifty years. That is a correction aimed at his own side of the argument, and it is a sign he is reporting the evidence rather than tuning it for defense.
Using that method, Wallace puts the best estimate of variants in the Greek New Testament at about half a million. That works out to roughly three and a half variants for every word of the text. Skeptics put the emphasis here, and Wallace grants the number. He even says Ehrman’s claim that there are more variants than words in the New Testament is, if anything, an understatement.
Then he explains where the number comes from. “If we had one manuscript, it wouldn’t disagree with itself.” Variants only appear when there is a second copy to compare. Hand copying never produces perfect duplicates over a long text. “It’s virtually impossible for human beings to make two identical manuscripts that go on for hundreds of pages,” Wallace says, and he knows of no two identical New Testament manuscripts. By his count there are nearly 6,000 Greek manuscripts, averaging more than 450 pages each. That comes to about two and a half million pages of handwritten text spread over fifteen centuries of copying. Half a million differences across that much material is what ordinary human copying would produce.
This is the center of the matter. The variant count measures how much evidence survives, not how badly the text decayed. A book with only a handful of copies would show very few variants and very little ability to check anything.
What kind of differences are they?
The count alone says nothing about what the variants affect. So Wallace tested how trivial a variant can be. He spent eight hours one day writing out the ways a Greek writer could say “John loves Mary.” He varied the spellings of the names, the word order, and whether the article appears. He reached 384 versions. Every one translates the same way into English. With other verbs and particles, he says, the number climbs to about 1,200. Most of the half million are of this kind: spelling differences, nonsense readings that are obvious slips, and word order changes that alter nothing.
Scholars spend their time on a much smaller set. Wallace has concluded that “approximately one fifth of 1% of all textual variants are both meaningful and viable.” He defines the terms carefully. A meaningful and viable variant is one where the different readings “can change the meaning of the text to some degree, and they have a good possibility of going back to the original wording.” That leaves about a thousand variants that deserve real study. Well over 99 percent affect nothing.
Two of his examples show what “meaningful” looks like in practice. In Mark 9:29, Jesus tells his disciples that a certain kind of demon comes out only by prayer. The KJV reads: “And he said unto them, This kind can come forth by nothing, but by prayer and fasting” (Mark 9:29). According to Wallace, the later manuscripts, which are the majority, include “and fasting,” while most early manuscripts do not. This is the only place in the New Testament that seems to require fasting for certain exorcisms, so the variant matters. But Wallace’s verdict is precise: “this is not a matter of orthodoxy, it is a matter of orthopraxy.” The question is about practice, not about who God is or how people are saved.
The second example is the number of the beast. The KJV reads: “Let him that hath understanding count the number of the beast: for it is the number of a man; and his number is Six hundred threescore and six” (Revelation 13:18). Wallace reports that two surviving manuscripts read 616, and that second-century writers knew of others. One of those manuscripts, he says, came to light less than twenty years before the conversation and is the oldest copy of Revelation 13. Most scholars still favor 666. Either way, no church or seminary puts the antichrist’s number in its statement of faith alongside the virgin birth and the resurrection. Disputed variants like these are real, and they sit at the edges of the faith.
How does this record compare with other ancient books?
The corruption charge often holds the New Testament to a standard no other ancient text could meet. Wallace insists on comparing like with like. According to him, the average Greco-Roman author survives in fewer than 15 manuscripts. For the New Testament he cites nearly 6,000 Greek manuscripts, over 10,000 in Latin, and another 5,000 to 10,000 in other ancient languages, for a total of 20,000 to 25,000.
The dates matter as much as the totals. Wallace counts about a dozen manuscripts within a century of the New Testament’s completion and over 60 within three centuries, including copies of the complete New Testament. For the classical author, the gap is far longer. “The average classical author is waiting over a thousand years before we get a single copy,” he says. “If we had to wait like that for the New Testament, the skeptics would have a field day.” Within a thousand years of its writing, the New Testament has over 1,800 Greek copies by his count.
Wallace adds a qualification that cuts against overconfidence in either direction. Skeptics say we lack the originals and even the early copies of copies. Wallace says that is probably true, but he also points to a recent study of ancient libraries suggesting that literary papyri commonly lasted two or three hundred years. If so, some originals may have survived for decades and been copied directly. He thinks the originals were gone by the end of the second century, worn out precisely because they were copied so often. These figures are Wallace’s, and readers who want to press them should go to his published work. What they establish is the scale of the gap. The New Testament is not the weak case among ancient texts. It is the strongest.
What about deliberate changes?
The most serious form of the charge is not about spelling slips. It is the claim that scribes changed the text on purpose to support their theology. Ehrman argued this in Misquoting Jesus, which Wallace dates to 2005. Wallace does not deny the evidence. He grants that Ehrman showed some scribes did alter wording to fit their theological views. The real question, he says, is whether those changes altered the essential teaching of the New Testament.
A popular claim offers a test case. In Dan Brown’s novel The Da Vinci Code, Emperor Constantine invents the deity of Christ at the Council of Nicaea in AD 325. If that were true, earlier manuscripts should lack the passages that call Jesus divine. Wallace reports at least 48 manuscripts from before 325. One of them, the papyrus known as P66, contains John 1:1, which reads in the KJV: “In the beginning was the Word, and the Word was with God, and the Word was God” (John 1:1). Wallace says every manuscript in every language has that wording. He dates P66 at least 125 years before the council and says Constantine had no hand in defining Christ’s deity. In Wallace’s summary, “the early manuscripts don’t disagree with the later ones on the essentials of the faith.”
Ehrman also argued that early scribes were amateurs who made more mistakes than later professionals. Wallace points to research by a former student, later published, which found that professional training did not predict careful copying. His example compares two early papyri. The scribe of P75, an early third-century manuscript, wrote plain, unattractive letters one at a time and produced one of the most faithful copies known. The scribe of P66 wrote elegant letters but left out words, by Wallace’s account because he was absorbed in making the letters beautiful.
Accidental errors also tend to correct themselves. Wallace compares them to someone copying the preamble to the Constitution and writing that the people formed a more perfect onion. Any reader who knows the context knows the word should be union. Later scribes caught slips like that easily.
Wallace also reports something Ehrman said himself. In an appendix to the early paperback edition of Misquoting Jesus, Wallace says, Ehrman was asked why he believed the core doctrines had been put in jeopardy by scribes. Ehrman answered that he did not believe that, and that no essential Christian doctrine is jeopardized by the variants. Wallace notes that later editions may have dropped the appendix. When Wallace pressed him in their third debate, Ehrman held that we cannot recover the original wording of Plato’s Republic either. Wallace replies that most classical scholars believe they largely have recovered their texts. If Ehrman’s standard holds, it removes knowledge of all ancient literature, not just the Bible.
Where is the record genuinely uncertain?
An honest investigation has to name where the evidence goes against the familiar text. Wallace does. Two passages stand apart: the long ending of Mark, Mark 16:9-20, and the story of the woman caught in adultery, John 7:53-8:11. Each is twelve verses long. Wallace says the next longest variant runs about two verses. Only about two dozen variants run one or two verses, and the rest involve part of a verse or a single word.
On Mark’s ending, Wallace points out that the two oldest Greek manuscripts of the passage, both from the fourth century, lack it. When those two agree, he says, their reading goes back deep into the second century. Several early translations also lack it or carry a different ending. He reports that Eusebius in the early fourth century and Jerome around the turn of the fifth, both with wide access to manuscripts, said they could hardly find copies containing these verses. The style, vocabulary and theology of the passage also differ from the rest of Mark. Wallace’s explanation is that scribes found an ending at 16:8, with frightened women and no appearance of Jesus, hard to accept, so they supplied several endings.
The case against the woman caught in adultery is stronger still, by Wallace’s account. He says no Greek manuscript of the first eight centuries contains it except one, in a quite different form. No church father wrote a commentary on it for over a thousand years, and it was absent from the Greek church’s lectionary readings until later. The story also appears in about half a dozen locations, including within Luke and at the end of the four Gospels. In Wallace’s words, “it’s a story that scribes wanted to put in the Bible and they just weren’t sure where to put it.”
Wallace admits this is hard. He asks audiences which of the two passages they would keep, and nearly everyone picks the woman caught in adultery because it shows Jesus’ mercy. Its evidence is weaker. “We can’t come to the text on the basis of emotion. We have to come to it on the basis of evidence,” Wallace says. Lisa Fields, who said she has wrestled with this passage herself, mentions pastors who preach it while telling their congregations about its textual problem. That is the honest course. Both passages remain in the KJV, and readers should know their manuscript support is debated. No core belief rests on them. Wallace notes that what the long ending of Mark describes, such as resurrection appearances, speaking in tongues and Paul surviving a snakebite, appears elsewhere in the New Testament.
The uncertainty here argues against the corruption charge. These additions are known because the early manuscripts expose them. A text altered beyond recovery would not show its seams this clearly.
What the evidence establishes, and what it does not
The manuscript record shows that the wording of the New Testament can be recovered with very high confidence, that its real disputes are known and few, and that early copies agree with later ones on the essentials. It does not show that every word is settled. Wallace’s thousand meaningful and viable variants remain open, and his specific counts are one scholar’s careful estimates, not facts anyone should repeat without checking.
The record also answers a narrower question than some readers expect. Textual reliability shows that we have what the authors wrote. Whether what they wrote is true, whether Jesus rose, whether the Word was God, is a separate question that needs its own evidence. The manuscripts clear the ground for that question. They do not decide it.
For Lisa Fields, that ground was not theoretical. She first met Ehrman’s book in a New Testament course in 2007 and said that “it really challenged me.” She has said that without that challenge she would never have dug deeper. The corruption charge works best on people who never go and look. Once you look, the half million variants turn out to be visible for one reason: there is so much evidence to compare. That is what a well-preserved text looks like.
Quick answers
Do half a million variants mean the New Testament is unreliable?
No. Wallace explains that more manuscripts produce more variants. Well over 99 percent affect nothing, and only about a thousand are both meaningful and likely to be original.
How many manuscripts survive?
Wallace counts nearly 6,000 in Greek and 20,000 to 25,000 across all ancient languages. He compares that with fewer than 15 for the average Greco-Roman author.
Did scribes ever change the text on purpose?
Some did, and Wallace grants it. He argues that those changes did not alter the essential teaching, and that early manuscripts agree with later ones on the essentials.
Did Constantine invent the deity of Christ?
The manuscripts say no. Wallace cites at least 48 manuscripts from before AD 325, including P66, which contains John 1:1 at least 125 years before Nicaea.
Are the long ending of Mark and the woman caught in adultery original?
The evidence Wallace presents points strongly against both, and more strongly against the John passage. No essential Christian belief depends on either one.