This post is free and available to anyone who wants to look.  On the Bart Ehrman Blog, there are substantial posts on interesting topics like this five days of the week, going back well over seven years.  If you belonged to the blog, you could get these posts, and access to all the archives.  The membership fee is extremely low, given the value; and every penny goes to charity.  So why not join?


It makes sense that scholars of the New Testament are predominantly committed Christians interested in knowing as much as they can about the Christian Scriptures.  That includes, of course, textual critics, the scholars who devote their research to trying to establish, to the best of their abilities, what the authors of the New Testament (whoever they were) actually wrote – what words they used, passage after passage, verse after verse.   The majority of these textual critics, as it turns out, are conservative in their religious/theological views.  The reality is that Baptists are more interested in this kind of question than Episcopalians.  (I’m not knocking either one: I’ve been both!)

Quite often you will hear such textual scholars say crazy things.  I suppose all of us say crazy things.  But when a well-established and widely-published scholar says them, they somehow sound more convincing and people usually don’t dig very deep in order to see why in fact what originally made sense is actually non-sense.

A lot of textual critics are very concerned to show that even though there are lots of variations in our manuscripts, there is nothing to worry about.  We know what the authors originally wrote – if not with 100% certainty, then, at least something pretty close to that.  This concern that non-scholars not be dismayed by the fact that we don’t have the originals of the New Testament but only later copies, all of them with mistakes, has been with us since the early 18th century.

In 1707 an Oxford scholar named John Mill published the first major critical edition of the New Testament which noted lots of differences in our manuscripts.  In fact, Mill’s edition cited 30,000 places where the manuscripts differed from each other.  This, uh, proved unsettling.   You mean there are 30,000 places where different manuscript have different readings?  Even more disturbing, Mill was forthright: he hadn’t included all the ones he found, only the ones that he thought were interesting or important for one thing or another.

Yikes.  He created his lists of differences by examining the readings of something like 100 manuscripts.  Today we have over 5600.  Uh…   As a result, we have (estimates vary), something like 400,000-500,000 variants.

Naturally people who want to know what the authors of the New Testament really wrote, down to the very word, find this, well, disturbing.  Sometimes deeply disturbing.

And so textual critics sometimes try to calm the jitters by saying crazy things.  My own view, as you’ll see in subsequent posts, is that it does in fact seem *likely* that most of the time we have a pretty or even a very good idea of what the authors wrote.  But not all the time.  And it’s only “likely.”  Fact be told, there are some passages – comforting assurances to the contrary – where we really don’t know and are unlikely ever to know.  For most of us that’s not a huge problem.  We have a similar problem when it comes to the Iliad and the Odyssey, the Aeneid, the dialogues of Plato, the plays of Euripides – and every other author from antiquity.  On the other hand, no one is hanging on to the words of Homer or Euripides for their eternal life.  So for many people the differences in the New Testament manuscripts matter.

And so scholars standing within the Christian tradition try to assure readers that there is almost precisely zero to worry about.  We know the original words.   And to make their point, they say some crazy things.   One that has always struck me as particularly crazy involves the use of off-the-cuff statistics meant to sound reassuring.

Here’s one of the most common.  If I had a nickel for every time I’ve heard this one, or a variation of it, I could buy a beach house in Acapulco.  Scholars (and the people who have heard this from scholars) claim that even though there are differences in our manuscripts, we are certain about the text 99% of the time.  99%!!   Conclusion: there’s nothing to worry about.

I am not *about* to concede this claim, for reasons you’ll see in a second.  I think it is indisputably unprovable.  It simply is an invented, made up claim that has ZERO basis.   But just for the sake of argument, suppose it were true.   Let’s take a hypothetical passage, say a passage that is giving its readers injunctions about activities they were supposed to avoid, a passage of, say, 200 words, telling the readers not to murder, commit adultery, steal, lie, cheat on their taxes, or say crazy things without any sound basis just to make people feel better.

OK, if we *know* that 99% of these words are what the author actually wrote, then only two of the words are disputed.  But what if the two words that are disputed are two instances of the word “not”?   Would that have any effect on the meaning of the passage?  Uh….

But here’s why the claim is crazy and completely and utterly unfounded.   How does one arrive at the statistic: 99%?  Or suppose someone says 96%?  Or 98.7%?  or 99.9%?  Or, say, 20%?  How does one come up with the number?

Think about it for a second.  These apologists are saying that 99% of these words (or 96% or 20% — doesn’t matter what number they use) are just like the words of the original.  We can know that!  And how do we know?  Well, uh…

Here’s the only way we would know.  Suppose I had two of you make hand-written copies of the first chapter of the Gospel of John from the New Revised Standard Version of the Bible.   I could then take your copies and compare them to one another carefully, word for word.   And I could, on the basis of a detailed letter-by-letter analysis show that in, say 83% or 97% of the words you copied, you copied them exactly the same.  That would be incontrovertible.  It’s an experiment that could be repeated by anyone who wanted to do the comparison.

I could *also* calculate how many times each of you copied the NRSV edition *accurately*.  That is, I could take the first chapter of John in the NRSV and compare each of your copies to it, to see if you got 92% or 99% or even 100% right.  And again it would be incontrovertible: I simply need to compare your copy with the original.

But what if I don’t *have* the original?   Then I could do the first kind of comparison – no problem: I could compare your copies to each other and tell you how many times they differ from each other.

But, here is the key point, I would NOT be able to tell you how close either of them was to the New Revised Standard version of John 1, unless I actually *had* the NRSV to compare your copies to.  All I would have is your copies.  And if you had lots of differences, I wouldn’t have any way of knowing which or how many of those differences represent what the NRSV has.

Suppose I had 5600 of you make copies.  How would I know that 99% of the time we have the right words?  The only way to establish a statistic of percentage agreement is to have the model that the copy is being based on, and compare the copy with the model.  If you don’t have the model, then you can’t name a number to give a statistic.  You can take a guess.

That’s what we do.  We take guesses.  Some of the time these are educated guesses, as in – yeah, most of the time we probably have a good idea of the wording.

But 99% is not an educated guess.  It’s a statistic drawn out of the hat.  It is meant to provide assurance.  It is not rooted in a statistical reality.  It is rooted in a hope and an assumption.  In fact, in a lot assumptions.

Now in theory, you *could* say something that is demonstrably true, that scholars over the past 100 years have not engaged in sustained debates about 97% of the words in this passage.  But even there you’d have to be more precise, e.g., that 93% of the scholars who have written about this chapter in the following academic journals and books published by the following academic presses have agreed on 98% of the words of the passage – something like that.

But that doesn’t mean we actually *know* 99% (or 97% or whatever) of the words.  It means that most scholars haven’t been debating about most of the words.  There are lots of things we don’t debate about.   We used to not debate about whether the earth was the center of the universe.  We all knew it was.  That doesn’t mean we were right.

But my overarching point is that the statistic is made up.  You can’t say 99% are certainly what the original said unless you have the original to compare the words to.  That is, it’s a crazy claim, with no basis.

On the other hand, is it reasonable to think that most of the time we have a pretty good idea what the authors original wrote?  Sure.  And I’ll try to show why that’s a reasonable thing to think, or at least assume.  But that’s a different question, one that’s not open to wild statistical statements of “fact.”