When someone says the New Testament has been copied and changed too many times to trust, they are usually pointing at one number: roughly half a million differences among the manuscripts. A good answer starts by explaining why that number exists. Differences pile up when a book has many copies, and the New Testament has more copies than any other ancient book. Those same copies let scholars check one against another and recover what was originally written. Most of the differences change nothing. The few that matter are known, studied openly, and leave the core of the faith untouched.
This guide builds that answer one term at a time. Each term depends on the one before it, so work through them in order. You can use the guide alone over a week or with a group in one long session or two shorter ones. The material comes from a Jude 3 Project conversation between Lisa Fields and Dr. Daniel Wallace, senior professor of New Testament Studies at Dallas Theological Seminary. By the end, the goal is that you can explain the manuscript evidence in your own words, without notes, to someone who doubts it.
Begin by making some variants yourself
Before any definitions, try a short exercise. Each person copies John 4:1-6 from a King James Bible by hand. Use a pen, copy at a normal pace, and do not go back to check your work. Then trade pages with someone and mark every place where their copy differs from the printed text. Count spelling slips, skipped words, words in the wrong order, and anything added.
Working alone, copy the passage once today and again in two or three days without looking at your first attempt. Then compare both copies with the printed text.
Leaders, keep this to about ten minutes. Most groups find several differences across six verses. Do not explain the result yet. Write the total where everyone can see it, because the rest of the guide will make sense of it.
Textual criticism
The name of the field puts people on guard, so define it first. Wallace explains that “the word criticism doesn’t mean that we’re being critical of the Bible. It really is virtually a synonym for research.” Textual criticism is the work of recovering the wording of a text whose original is lost or out of reach.
This is not a special test for Christian Scripture. Wallace says scholars do the same work with Shakespeare’s plays, Mozart’s operas and Lincoln’s Gettysburg Address. He adds that no original survives for any ancient literary document. So when someone says we don’t have the originals of the New Testament, the honest reply is that this is true of every ancient book. The useful question is how well the copies let us recover the original.
Manuscript and variant
A manuscript, in Wallace’s usage, is a handwritten copy made before the printing press or made by hand rather than copied from a printed text. A textual variant is any place where manuscripts differ in wording. Wallace includes word order, an added or missing word, an added or missing verse, and even spelling.
The next idea is the one people most often get wrong. Variants are counted by the different readings at a given spot, not by how many manuscripts carry each reading. Wallace’s example is John 4:1. Some manuscripts say Jesus knew the Pharisees had heard. Others say the Lord knew. The King James Version follows the second: “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, “If they all have the same thing in that disagreement, it’s still just one variant.”
Go back to your copying exercise and count again by this rule. If three people all dropped the same word at the same spot, that is one variant, not three. Your total will likely shrink. That shrinking teaches the counting rule better than a definition can.
Why more copies produce more variants
Now the large number makes sense. Wallace grants the skeptic’s figure of about half a million variants, roughly three and a half for every word of the Greek New Testament. Then he points to where the number comes from: “If we had one manuscript, it wouldn’t disagree with itself.”
A second copy brings the first disagreements, and every copy after it can add more. “It’s virtually impossible for human beings to make two identical manuscripts that go on for hundreds of pages,” Wallace says. 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, made over about 1,500 years of hand copying. Against that much material, he argues, half a million differences is what ordinary human copying would produce.
Your exercise shows the principle. Several hand copies of six verses produced several differences. More copiers would have produced more. The differences tell you how many copies exist, not how badly the text was damaged.
The exercise only goes so far, though. Your group copied from a printed Bible that was sitting in front of you, so you could always check the original. Ancient scribes copied from copies, and the originals eventually disappeared. That is why scholars compare thousands of manuscripts with one another instead of checking a single master copy. The question is whether the copies are numerous enough and early enough to make that comparison reliable. Wallace’s numbers are meant to answer it. He counts 20,000 to 25,000 New Testament manuscripts across all ancient languages, against fewer than 15 for the average Greco-Roman author. He says the average classical author waits over a thousand years for a single surviving copy, while the New Testament has about a dozen manuscripts within a century of its completion. These counts are Wallace’s, and anyone who wants to repeat them publicly should check them against his published work.
Meaningful and viable
Counting variants tells you nothing about what they affect. For that you need two tests. Wallace concludes that “approximately one fifth of 1% of all textual variants are both meaningful and viable.” He defines the two terms together: the readings “can change the meaning of the text to some degree, and they have a good possibility of going back to the original wording.” By his estimate, about a thousand variants pass both tests.
To see how many variants fail the first test, take Wallace’s experiment with the sentence “John loves Mary” in Greek. By changing spellings, word order and the use of the article, he wrote out 384 versions. “Every single one is translated exactly the same, and it’s the same verb each time,” he says. Most of the half million are like that: spelling differences, obvious slips and word order that changes nothing.
Now look at two variants that pass both tests. Read the KJV of Mark 9:29: “And he said unto them, This kind can come forth by nothing, but by prayer and fasting.” According to Wallace, most early manuscripts lack the words about fasting, while the later manuscripts, which are more numerous, include them. That matters because this is the only verse in the New Testament that seems to make fasting necessary for certain exorcisms. He gives a precise verdict: “this is not a matter of orthodoxy, it is a matter of orthopraxy.” Orthodoxy is right belief, what Christians confess about God and salvation. Orthopraxy is right practice, how Christians live out that faith.
Revelation 13:18 is the second example. The KJV gives the number of the beast as “Six hundred threescore and six,” which is 666. Wallace reports two surviving manuscripts that read 616, one of them the oldest manuscript of that chapter. Most scholars still favor 666. He also notes that no church or seminary lists the beast’s number in its doctrinal statement next to the virgin birth and the resurrection.
Sorting practice: Draw two columns, one for variants that change meaning and one for variants that don’t. Place your group’s copying mistakes, the John 4:1 example, the “John loves Mary” versions, Mark 9:29 and Revelation 13:18. For anything in the meaning column, ask whether it touches belief or practice.
External and internal evidence
How do scholars decide which reading is original? Wallace uses two kinds of evidence. External evidence concerns the manuscripts themselves: how old they are, how many there are, which early translations agree with them, and what ancient writers said about the copies they had seen. Internal evidence concerns the wording itself. Wallace mentions style, vocabulary, syntax and theological perspective.
He applies both to the two longest variants in the New Testament, each twelve verses long: Mark 16:9-20 and John 7:53-8:11. For Mark, the two oldest Greek manuscripts of that passage, both from the fourth century, lack the ending. Several early translations lack it or carry a different ending. Wallace says Eusebius and Jerome reported they could hardly find manuscripts containing it. On internal grounds, the verses do not match the rest of Mark’s style. For the woman caught in adultery, Wallace says only one Greek manuscript from the first eight centuries contains it, and in a different form. The story also turns up in about half a dozen places, including within Luke.
Scripture observation: Read Mark 16:1-8 in the KJV and stop at verse 8. Notice where the women are and what they feel. Then read verses 9-20. Write down any shift you notice in tone, pace or subject. Then read John 7:45-52, skip to 8:12, and see whether the passage reads smoothly without the story in between. If your study Bible has a footnote at either passage, read it. These passages remain in the KJV, and their earliest manuscript support is debated. That is something to be informed about, not alarmed by.
Wallace admits the second passage is hard to let go. Most people would keep it because it shows Jesus’ mercy. His reply is direct: “We can’t come to the text on the basis of emotion. We have to come to it on the basis of evidence.” He also notes that what the long ending of Mark describes, such as people speaking in tongues and Paul surviving a snakebite, appears elsewhere in the New Testament. No essential teaching depends on either passage.
The misunderstanding to expect
Once people learn that some verses were added later, many jump to the opposite conclusion: if scribes added twelve verses here, what else did they slip in that no one noticed? Take that worry seriously, because it points to the strongest form of the challenge. The objection is about intentional, theological changes, not careless slips.
Wallace does not deny that such changes happened. He grants that some scribes changed wording to fit their theology. His question is whether those changes altered the essential teaching of the New Testament, and he answers with the record from before AD 325. He says at least 48 manuscripts survive from before that date, including P66. That manuscript contains John 1:1: “In the beginning was the Word, and the Word was with God, and the Word was God.” His conclusion is that “the early manuscripts don’t disagree with the later ones on the essentials of the faith.”
The added passages actually support this. We know about them because earlier manuscripts lack them. A text changed beyond recovery would not show where it had been changed.
Be clear about what this argument can and cannot do. It shows that we can recover what the authors wrote. It does not by itself show that what they wrote is true. That is a separate conversation, and it is worth having once the first one is settled.
Practice: explain it in two minutes
Each person writes a short explanation that uses the terms in order: textual criticism, manuscript, variant and how variants are counted, why more copies produce more variants, meaningful and viable, and external and internal evidence. Keep it to what you could say in two minutes. Use your own words wherever you can.
Then pair up. One person explains while the other raises honest objections, such as but you just admitted whole stories were added or scribes changed things on purpose, didn’t they? Switch roles. Afterward, each pair names one point they could not yet explain clearly. Leaders, collect those points. They show where the group needs to go back, and they are a better measure of the session than a recall quiz.
Questions for discussion
- When you first heard the half-million figure, what did you assume it meant? Which term in this guide changed that assumption the most?
- After your copying exercise, how did your count change when you counted by readings instead of by copies? What does that tell you about how claims built on large numbers should be checked?
- Wallace places the fasting variant in Mark 9:29 under practice, not belief. Do you agree with how he sorts it? Can you think of a variant that might be harder to sort?
- Which kind of evidence, external or internal, would persuade you more about Mark 16:9-20, and why?
- Wallace says we must come to the text on the basis of evidence, not emotion. Where might that standard cost you something personally? Answer only if you are comfortable doing so.
- A skeptic grants everything in this guide but still says the New Testament is untrustworthy. What might they actually be doubting, and how would you find out?
Something to carry with you
Lisa Fields first ran into this challenge in a college New Testament course, and she says that without it, “I would’ve never dug deeper.” The aim here is not a memorized script. It is to understand the evidence well enough that a hard question makes you want to go and look. This week, find one person who has heard the half-million claim and try your two-minute explanation on them. Then write down the question they asked that you could not answer, and bring it back to the group.