Bible Scribe_Copyist_03

Edward D. Andrews

The first part of this chapter will cover the gist of what is most often discussed in New Testament textual criticism today. Thereafter, we will discuss what should be the primary focus of NTTC (New Testament Textual Criticism). It would seem that Bart D. Ehrman and other Bible critics of his persuasion have sent many textual scholars on a quest. These scholars have become obsessed with discussing how many variants there are, how to count the textual variants, and whether they are significant or insignificant. Below, we will cover what is being said about variants, as well as whether some are more significant than others, and then close the chapter with what actually is the most important mission in NTTC.

Some Bible critics seem, to begin with the belief that if the originals were inspired of God and fully inerrant, the subsequent copies must continue to be inerrant in order for the inerrancy of the originals to have value. They seem to be asking, “If only the originals were inspired, and the copies were not inspired, and we do not have the originals, how are we to be certain of any passage in Scripture?” In other words, God would never allow the inspired, inerrant Word to suffer copying errors. Why would he perform the miracle of inspiring the message to be fully inerrant and not continue with the miracle of inspiring the copyists throughout the centuries to keep it inerrant? First, we must acknowledge that God has not given us the specifics of every decision he has made in reference to humans. If we begin asking, “Why did God not do this or do that,” where would it end? For example, why didn’t God just produce the books himself, and miraculously deliver them to people as he gave the commandments to Moses? Instead of using humans, why did he not use angelic messengers to pen the message, or produce the message miraculously? God has chosen not to tell us why he did not move the copyists along with Holy Spirit, so as to have perfect copies, and it remains an unknown. However, it should be noted that if we can restore the text to its original wording through the science of textual criticism, i.e. to an exact representation thereof, we have, in essence, the originals.

We do know that the Jewish copyists and later Christian copyists were not infallible as were the original writers. The Holy Spirit inspired the original writers, while the most that can be said about the copyists is that they were guided by the Holy Spirit. However, do we not have a treasure-load of evidence from centuries of copies, unlike ancient secular literature? Regardless of the recopying, do we not have the Bible in a reliable critical text and trustworthy translations, with both improving all the time? It was only inevitable that imperfect copyists, who were not under inspiration, would cause errors to creep into the text. However, the thousands of copies that we have enable textual scholars to identify and reject these errors. How? For one thing, different copyists made different errors. Therefore, the textual scholar compares the work of different copyists. He is then able to identify their mistakes.

A Simple Example

Suppose 100 people were invited or hired to make a handwritten copy of Matthew’s Gospel, with 18,345 words. Further suppose that these people fit in one of four categories as writers: (1) struggle to write and have no experience as a document makers; (2) skilled document makers (recorders of events, wills, business, certificates, etc.); (3) trained copyists of literature; and (4) the professional copyists. There is little doubt that these copyists would make some copying errors, even the professionals. However, it would be impossible that they would all make the same errors. If a trained textual scholar with many years of religious education, including textual studies, and decades of experience, were to compare the 100 documents carefully, he could identify the errors and restore the text to its original form, even if he had never seen that original.

The textual scholars of the last 250 years, especially the last 70 years have had over 5,800 Greek manuscripts at their disposal. A number of the manuscripts are portions dating to the second and third centuries C.E. Moreover, more manuscripts are always becoming known; technology is ever advancing, and improvements are always being made.


Hundreds of scholars throughout the last three centuries have produced what we might call a master text, by way of lifetimes of hard work and careful study. Are there places where we are not certain of the reading? Yes, of course. However, we are considering very infrequent places in the text of the Greek NT that contains about 138,020 words, which would be considered difficult in arriving at what the original reading was. In all these places the alternative readings are provided in the apparatus. Bible critics who exaggerate the extent of errors are misleading the public on several fronts. First, some copies are almost error-free and negate the critics, who claim, “We have only error-ridden copies.”[1] Second, the vast majority of the Greek New Testament has no scribal errors. Third, textual scholarship can easily identify and correct the majority of the scribal errors. In addition, of the remaining errors, we can still say most are solved with satisfaction. Of the small number of scribal errors remaining, we can say that most are solved with some difficulty, and there remain very few errors of which textual scholarship continues to be uncertain about the original reading at this time.

400,000 to 500,000 Supposed Variants in the Manuscripts

With this abundance of evidence, what can we say about the total number of variants known today? Scholars differ significantly in their estimates—some say there are 200,000 variants known, some say 300,000, some say 400,000 or more! We do not know for sure because, despite impressive developments in computer technology, no one has yet been able to count them all. Perhaps, as I indicated earlier, it is best simply to leave the matter in comparative terms. There are more variations among our manuscripts than there are words in the New Testament.[2]

Bart D. Ehrman has some favorite, unprofessional ways of describing the problems, which he stresses without qualification, in every interview he has for a lay audience or seminary students. Below are several, the first two from the quotation above:

  • Scholars differ significantly in their estimates—some say there are 200,000 variants known, some say 300,000, some say 400,000 or more!
  • There are more variations among our manuscripts than there are words in the New Testament.
  • We have only error-ridden copies, and the vast majority of these are centuries removed from the originals and different from them, evidently, in thousands of ways. (Whose Word is It, 7)
  • We don’t even have copies of the copies of the originals, or copies of the copies of the copies of the originals. (Misquoting Jesus, 10)
  • In the early Christian centuries, scribes were amateurs and as such were more inclined to alter the texts they copied. (Misquoting Jesus, 98)
  • We could go on nearly forever talking about specific places in which the texts of the New Testament came to be changed, either accidentally or intentionally. (Misquoting Jesus, 98)
  • The Bible began to appear to me as a very human book. (Misquoting Jesus, 11)

Each of the bullet points above claimed by Ehrman can be categorized as an exaggeration, misinformation, misleading, or just a failure to be truthful. Many laypersons-churchgoers have been spiritually shipwrecked in their faith by such unexplained hype. What the uninformed person hears is that we can never get back to the originals or even close, that there are hundreds of thousands of significant variants that have so scarred the text, we no longer have the Word of God, and it is merely the word of man. How such a knowledgeable man cannot know the impact his words are having is beyond this author.

Miscounting Textual Variants

In 1963, Neil R. Lightfoot penned a book that has served to help over a million readers, How We Got the Bible. It has been revised two times since 1963, once in 1988, and again in 2003. There is a “miscalculation” in the book which has contributed to a misunderstanding in how textual variants are counted. In fact, there are several other books repeating it. A leading textual scholar, Daniel B. Wallace, has brought this to our attention in an article entitled, The Number of Textual Variants an Evangelical Miscalculation.[3] World-renowned Bible apologist Norman L. Geisler has commented on it as well.

Lightfoot wrote,

From one point of view, it may be said that there are 200,000 scribal errors in the manuscripts. Indeed, the number may well considerably exceed this and obviously will grow, as more and more manuscripts become known. However, it is wholly misleading and untrue to say that there are 200,000 errors in the text of the New Testament. (Actually, textual critics consciously avoid the word “error;” they prefer to speak of “textual variants.”) This large number is gained by counting all the variations in all of the manuscripts (over 5,800). This means that if, for example, one word is misspelled in 4,000 different manuscripts, and it amounts to 4,000 “errors.” Actually, in a case of this kind, only one slight error has been made, and it has been copied 4,000 times. But this is the procedure which is followed in arriving at the large number of 200,000 “errors.”[4]

Wallace makes this observation in his article:

In other words, Lightfoot was claiming that textual variants are counted by the number of manuscripts that support such variants, rather than by the wording of the variants. This book has been widely influential in evangelical circles. I believe over a million copies of it have been sold. And this particular definition of textual variants has found its way into countless apologetic works.” He goes on to clarify just what a textual variant is, “The problem is, the definition is wrong. Terribly wrong. A textual variant is simply any difference from a standard text (e.g., a printed text, a particular manuscript, etc.) that involves spelling, word order, omission, addition, substitution, or a total rewrite of the text. No textual critic defines a textual variant the way that Lightfoot and those who have followed him have done.

Geisler writes,

Some have estimated there are about 200,000 of them. First of all, these are not “errors” but variant readings, the vast majority of which are strictly grammatical. Second, these readings are spread throughout more than 5300 manuscripts, so that a variant spelling of one letter of one word in one verse in 2000 manuscripts is counted as 2000 “errors.”[5]

Lightfoot evidently was thought to have erred by counting manuscripts, rather than the variants in the text. In fairness to Lightfoot, it should be pointed out that he deplored the system of counting “errors” by the number of manuscripts, as the quotation above reveals. He was simply saying that critics were doing this, not that it was proper. It is difficult to see why Wallace would attribute responsibility for the system to Lightfoot. Also, Wallace cited Lightfoot’s 1963 edition that did not include the distinction between “error” and “textual variant.”

Let me offer the reader an example for our purposes. First, we should underscore a few important points raised: 1) we have so many variants because we have so many manuscripts. 2) We do not count the manuscripts; we count the variants. 3) A variant is any portion of the text that exhibits variations in its reading between two or more different manuscripts. This is more precisely called a variation unit. It is important to distinguish variation units from variant readings. Variation units are the places in the text where manuscripts disagree, and each variation unit has at least two variant readings. Setting the limits and range of a variation unit is sometimes difficult or even controversial because some variant readings affect others nearby. Such variations may be considered individually, or as elements of a longer single reading.

We should also note that the terms “manuscript” and “witness” may appear to be used interchangeably in this context. Strictly speaking, “witness” (see below) only refers to the content of a given manuscript or fragment, so the witness predates the physical manuscript on which it is written to a greater or lesser extent. However, the only way to reference the “witness” is by referring to the manuscript or fragment that contains it. In this book, we have sometimes used the terminology “witness of x or y manuscript” to distinguish the content in this way.

We begin by choosing our “base” or “standard text.” We are using the standard text (critical or master text), Nestle-Aland (NA) Greek Text (28th edition) and the United Bible Society (UBS) Greek Text (5th edition). These two critical texts are actually the same. Therefore,

Note: When the acronym NU is used, N stands for Nestle-Aland, the U for United Bible Societies, since the texts are the same. The apparatuses are different, the UBS version designed primarily for translators (more on this below).

In this writer’s opinion, the critical NU text is as close as we can get to what the original would have been like.[6] Therefore, we can use the reading in the critical text as the original reading, and anything outside of that in the manuscript history is a variant: spelling, word order, omission, addition, substitution, or a total rewrite of the text. Any difference in two different manuscripts is a variant, technically speaking.

Before going to our example, I want to emphasize that Bible critics, who grumble and repeat over and over again how there are 400,000 variants in the text of the New Testament, have only one agenda: they want to discredit the Word of God. They use the issue of variants as a misrepresented excuse for their having lost their faith, having shipwrecked their faith, or having had no faith from the start. These Bible critics are no different from the religious leaders Jesus dealt with in the first century. Jesus said of them, “Blind guides! You strain out a gnat, yet gulp down a camel!” (Matt. 23:24). They thrust aside 99.95 percent because 0.05 of one per cent is in not absolutely certain! Now let’s turn to our example, which comes from the Apostle Paul’s letter to the Colossians.

Example of a Textual Variant

Colossians 2:2 Updated American standard Version (UASV)

that their hearts may be comforted, having been knit together in love, and into all riches of the full assurance of understanding, and that they may have a complete knowledge[7] of the mystery of God, namely Christ, [τοῦ θεοῦ Χριστοῦ; tou theou Christou]

See the chart below.


Variant MSS or Versions
NU[8] of the God of Christ

Standard Text


of the God 10 MSS[9]
2 of the Christ

1 MS


of the God who is Christ 4 MSS
4 of the God who is concerning Christ



Of the God in the Christ 2 MSS
6 of the God in the Christ Jesus

1 MS


of the God and Christ 1 MS
8 Of God the father Christ



Of God the father of Christ 5 MSS
10 Of God and Father of Christ



Of God father and of Christ 4 MSS
12 Of God father and of Christ Jesus



Of God father and of Lord of us Christ Jesus 2 MSS
14 Of God and father and of Christ

38 MSS

Total 14

14 Variants in 79 MSS

79 MSS

These variants are found in 79 MSS, Thus, we have 14 variants in 79 manuscripts, not 79 variants. We do not count manuscripts, as most textual scholars know. In trying to paint a picture about the trustworthiness of the text, this author does not think talking about variants is really helpful, and it can confuse the layperson. It is important for the churchgoer to know what a variant is and the general extent of the variants, but in the long run, it is the places in the text that are affected by variants that most matter, and what we have as our text in the end.

The United Bible Society’s “A” “B” “C” and “D” ratings are fine, and the definitions by UBS, i.e., [A] certain, [B] almost certain, [C] difficulty in deciding, and [D] great difficulty in arriving at, are helpful but should be better qualified, with some numbers of what percentage of the text fall under each area.

All Variant Units (Places)

What we need to talk about is how many places there are where we find variants. What percentage is this of the entire New Testament text?

We can then discuss:

  • What percentage of the text is untouched by variants?
  • Of the percentage affected, how much can we say or surmise to be given an “A” rating, a  “B” Rating, a “C,” or “D” rating?

Variant Reading and Variation Unit

This section is based in large part on the work by Eldon Jay Epp and Gordon D. Fee, Studies in the Theory and Method of New Testament Textual Criticism (Grand Rapids, MI: Eerdmans, 1993), wherein Eldon J. Epp expands on the brief 1964 article of Ernest C. Colwell (1901–74) and Ernest W. Tune on “Variant Readings: Classification and Use.”

Again, what we need to discuss is how many variation units (places) there are where we find variations. Before doing so, let us define some terms.

SIGNIFICANT AND INSIGNIFICANT READINGS AND OR VARIANTS: Below we have what are commonly described as significant and insignificant variants. Significant would mean any reading that has an impact on the transmission history of a variant unit. For example, it would apply to how we determine the relationship of the manuscripts to one another, such as where a particular manuscript would fall in the history and transmission of the manuscripts. It would also be impactful if the reading could help the textual scholar establish the original. Therefore, insignificant would mean just the opposite, referring to a reading that has very little to no impact at all in many aspects of a transmission history. The reason we stop at “many” aspects here is that all readings in a manuscript play a role in some aspects of the transmission history, such as the characteristics of the manuscript it is in and the scribal activity within that individual manuscript.

InsignificantNonsense Reading: As Epp points out, a nonsense reading is “a reading that fails to make sense because it cannot be construed grammatically, either in terms of grammatical/lexical form or in terms of grammatical structure, or because in some other way it lacks a recognizable meaning. Since authors and scribes do not produce nonsense intentionally, it is to be assumed (1) that nonsense readings resulted from errors in transmission, (2) that they, therefore, cannot represent either the original text or the intended text of any MS or alert scribe, and (3) that they do not aid in the process of discerning the relationships among MSS.”[10] It should also be stated that the original did not contain any nonsense readings, as the writers were led by the Holy Spirit. The inspired author before publication would have corrected any error by a scribe such as Tertius or Silvanus.

InsignificantCertainty of Scribal Errors: while these errors “can be construed grammatically and make sense,” there is a certainty on the part of textual scholars that these are scribal errors. These are not nonsense readings but rather readings that make sense, which are scribal errors beyond all reasonable doubt. These would “be certain instances of haplography and dittography, cases of harmonization with similar contexts, hearing errors producing a similar-sounding word, and the transposition of letters or words with a resultant change in meaning.”[11] The problem that we sometimes encounter here is that what may be certainty of scribal error to one scholar may instead be an almost certainty to another, and even less so to another. The key element here in determining a reading that is understandable as insignificant is that it can be “demonstrated” so by the scholar making such a claim.

InsignificantIncorrect Orthography (Greek for “correct writing”): this term is used loosely to refer to the spelling of words, which (for Greek) can include breathing and accent marks. Thus, one can refer to variations in the orthography of a word, or even to incorrect orthography. When a variation in orthography is due merely to dialectical or historical changes in spelling for variant readings, the variations are often ignored in the decision process because the reading in question is identical to another reading, once the orthographical differences are factored in (mutatis mutandis). Epp writes, “Mere orthographic differences, particularly itacisms and nu-movables (as well as abbreviations) are ‘insignificant’ as here defined; they cannot be utilized in any decisive way for establishing manuscript relationships, and they are not substantive in the search for the original text. Again, the exception might be the work of a slavish scribe, whose scrupulousness might be considered useful in tracing manuscript descent, but the pervasive character of itacism, for example, over wide areas and time-spans precludes the ‘significance’ of orthographic differences for this important text-critical task.”[12]

InsignificantSingular Readings: a singular reading is technically a variant reading that occurs only once in only one Greek manuscript and is therefore immediately suspect. There is some quibbling over this because critics who reject the Westcott and Hort position on the combination of 01 (Sinaiticus) and 03 (Vaticanus) might call a reading “nearly singular” if it has only the support of these two manuscripts. Moreover, it is understood that not all manuscripts are comparable. Thus, for example, one would comfortably reject a reading found only in a single late manuscript, while many critics would not find it so easy to reject a reading supported uniquely by 03. Some also give more credit to singular readings that have additional support from versions. Singular readings that are insignificant would be nonsense readings, transcriptional errors, meaningless transpositions, and itacisms.

Significant Variants: a significant reading/variant is any reading that has an impact on any major facet of transmission history of a variant unit. One approach to identifying these is to remove the insignificant variants first: nonsense readings, determined (without doubt) scribal errors, incorrect orthography, and singular readings. Those readings that cannot be ruled out in this process are probably significant.

Number of Variants, Significant and Insignificant Variants vs. Level of Certainty

It would seem that some scholars have lost sight of the most important goal of textual criticism, namely, reconstructing the original. There is little doubt that agnostic Bible scholar Dr. Bart D. Ehrman has led the conversation on how many textual variants exist. The authors of this publication are focusing their attention on the initial goal of textual criticism, returning to the original. We believe that even now the Greek New Testament completely reliable. However, there are some 2,000 textual places within the New Testament that need to be dealt with because the witnesses and internal evidence require consideration and deliberation.

Level of Certainty

The level of certainty charts below is generated from A TEXTUAL COMMENTARY ON THE GREEK NEW TESTAMENT (Second Edition), A Companion Volume to the UNITED BIBLE SOCIETIES’ GREEK NEW TESTAMENT (Fourth Revised Edition) by Bruce M. Metzger.

The letter {A} signifies that the text is certain.

The letter {B} indicates that the text is almost certain.

The letter {C} indicates that the Committee had difficulty in deciding which variant to place in the text.

The letter {D}, which occurs only rarely, indicates that the Committee had great difficulty in arriving at a decision. In fact, among the {D} decisions sometimes none of the variant readings commended itself as original, and therefore the only recourse was to print the least unsatisfactory reading.

The word count below is taken from the Nestle-Aland Novum Testamentum Graece using Logos Bible Software.[13] While this author has compiled the numbers regarding the level of certainty of readings from Metzger’s Textual Commentary, he has not gone to the point of counting the letters or words at each variant place. We will just offer the reader the general statement that almost all textual variants in the commentary were based on a letter or a few letters in a Greek word, to two-three words. Seldom was it an entire sentence or verse, very rarely several verses like the long ending of Mark. Therefore, we have chosen three words as the average to multiply the total number of variants, so that the reader can see the truly small number of variants that are even worthy of consideration, as opposed to the total number of words in the New Testament. For example, Matthew has 18,346 words with a mere 153 places where we find variants selected for the GNT, affecting about 459 words.

We need to add and emphasize, that all of the variants counted were selected by the GNT editors as relevant for translation, and the total does not include other variant units that were not considered relevant for that purpose. A good number of these additional variants can be found in the NA apparatus, but only with considerable difficulty in many cases because the same variants are frequently handled differently in the GNT and NA apparatuses. The authors of this book do consider all variant units relevant even if a good number of them are difficult or virtually impossible to represent in translation (depending on the target language), and we recommend that the reader adjust the figures offered below by multiplying the numbers of variants by a factor of two, which should compensate for any variants that are not reported in the GNT text. We see no reason to assume a significantly different outcome in the ratings that might have been assigned to these variants if they had been included in the GNT, except possibly where no decisions might be possible in the cases of competing readings that were fully acceptable (rather than difficult).

For readers who have a working knowledge of NT Greek, it may be informative simply to select a few random pages of corresponding text from the GNT and NA and compare the apparatuses to see what is missing from the GNT relative to the NA apparatus. We believe that our suggestion of multiplying the variant figures below by a factor of two will appear more than reasonable; however, even using a factor of three or four will still leave a relatively minute percentage of “C” and “D” readings, as revealed below.

So then, if we look at Matthew and first multiply the GNT variant units by three for an average three words a variant, we have 459 words. Of the 153 variant units found in Matthew, we are certain of about 32 of them, almost certain about 70, have a little difficulty deciding on 50, and great difficulty deciding on only one variant unit. When we say that we have difficulty deciding, this does not mean that we cannot decide, as we can. Moreover, a good translation will list the alternative reading in a footnote. So in the entirety of the Gospel of Matthew, there is only one variant place (Matt 23:26) which we would count as about three out of 18,346 words, where there was great difficulty in deciding the original. As it turns out, in this case, the GNT apparatus handles it as a variant of eight words, while NA breaks it into two variants, thus illustrating our point about the difficulty of comparing the two apparatuses. Some translations have incorporated the variant (ESV, NASB, NIV, TNIV, NJB, and the NLT), viewing it as the original, while other translations (NRSV, NEB, REB, NAB, CSB, and the UASV) see the variant as an addition taken from the previous verse.

Matthew 23:26 Blind Pharisee, cleanse first the inside of the cup,[14] so that the outside of it may also become clean. (UASV)

NU has καθάρισον πρῶτον τὸ ἐντὸς τοῦ ποτηρίου, ἵνα γένηται καὶ τὸ ἐκτὸς αὐτοῦ καθαρόν “first cleanse the inside of the cup, that the outside of it may also become clean,” which is supported by D Θ f1 ita,e syrs (bold mine).

Variant/Byz WH καθαρισον πρωτον το εντος του ποτηριου και της παροψιδος ινα γενηται και το εκτος αυτων καθαρον have “first cleanse the inside of the cup [and the dish], that the outside of them may also become clean,” which is supported by א (B2) C L W 0102 0281 Maj.

 Looking at the above support alone, it would seem that the witnesses for the longer reading (“and the dish”) are weightier, making the longer reading the likely original. Then, when we consider the presence of a few manuscripts (B* f13 28 al) that are not listed for the shorter reading because they have the longer reading (“and the dish”), the weight shifts over to the shorter reading’s being the original. Why? Because these few manuscripts have the singular αυτου instead of αὐτῶν, even though they have the longer reading. This tells us that the archetype text was the shorter reading. Clearly, the copyist added (“and the dish”) from the previous verse, Matthew 23:25, which reads, “Woe to you, scribes and Pharisees, hypocrites! because you cleanse the outside of the cup and of the dish, but inside they are full of greediness and self-indulgence.”

Below, we will look at all of the numbers, the total words in the Greek New Testament, the number of A, B, C, and D variants in each book as they were selected by the GNT committee, followed by the total number of variants listed in Metzger’s textual commentary.

The Entire New Testament (138,020 Words)

{A-D} New Testament
{A} 505
{B} 523
{C} 354
{D} 10
Total Var. 1,392
Words 138,020

The Gospels (64,767 Words)

{A-D} Matt Mark Luke John
{A} 32 45 44 44
{B} 70 49 73 62
{C} 50 45 44 41
{D} 1 1 0 2
Total Var. 153 140 161 149
Words 18,346 11,304 19,482 15,635

The Acts of the Apostles (18,450 Words)

{A-D} Acts
{A} 74
{B} 82
{C} 40
{D} 1
Total Var. 197
Words 18,450

Paul’s Fourteen Epistles (37,361 Words)

{A-D} Rom 1 Cor 2 Cor Gal. Eph. Php Col.
{A} 39 21 12 16 16 10 8
{B} 19 22 17 3 11 7 12
{C} 20 15 10 8 7 3 8
{D} 1 1 0 0 0 0 0
Total Var. 79 59 39 27 34 20 28
WORDS 7,111 6,830 4,477 2,230 2,422 1,629 1,582
{A-D} 1 Th 2 Th 1 Tim 2 Tim Tit Phm. Heb.
{A} 9 3 15 2 2 2 20
{B} 2 3 2 6 1 3 11
{C} 3 2 2 1 1 0 12
{D} 0 0 0 0 0 0 0
Total Var. 14 8 19 9 4 5 43
WORDS 1,481 823 1,591 1,238 659 335 4,953

The General Epistles (7,591 Words)

{A-D} Jam 1 Pet 2 Pet 1 Jn 2 Jn 3 Jn Jude
{A} 7 21 8 18 4 1 9
{B} 12 9 7 7 1 1 0
{C} 4 7 6 4 0 0 3
{D} 0 0 1 0 0 0 1
Total Var. 23 37 22 29 5 2 13
WORDS 1,742 1,684 1,099 2,141 245 219 461

The Book of Revelation (9,851 Words)

{A-D} Revelation
{A} 23
{B} 31
{C} 18
{D} 1
Total Var. 73
Words 9,851

As noted above, the authors of this publication maintain that all variation units or places where variations occur are significant because we are dealing with the Word of God, and reconstructing the original wording is of the utmost importance. Recall Lightfoot once more. “What about the significance of these variations? Are these variations immaterial or are they important? What bearing do they have on the New Testament message and on faith? To respond to these questions, it will be helpful to introduce three types of textual variations, classified in relation to their significance for our present New Testament text. 1. Trivial variations which are of no consequence to the text. 2. Substantial variations which are of no consequence to the text. 3. Substantial variations that have bearing on the text.”[15]

Whether we are talking about the addition or omission of such words as “for,” “and,” and “the,” or different forms of similar Greek words, differences in spelling, or the addition of a whole verse or even several verses, the importance lies not with the significance of impact on the meaning of the text but rather the certainty of the wording in the original. What we want to focus on is the certainty level of reconstructing every single word that Matthew, Mark, Luke, John, Paul, Peter, James, and Jude penned.

We will use Lightfoot’s example of Matthew 11:10-23, that is, fourteen verses of 231 words; we have eleven variants in verses 10, 15, 16, 17, 18, 19(2), 20, 21, and 23(2). This may seem worrisome to the churchgoer or someone new to textual criticism. However, while all of the variants are found in the NA28 critical apparatus (2012), pp. 31–32,[16] the following sources below only covered seven of them because four are not even an issue. Why are they not an issue? We know what the original reading is with absolute certainty. The seven that have some uncertainty are mentioned in the textual commentaries below.

  • Comfort New Testament Text and Translation covers verses 15 and 19
  • Comfort Commentary on the Manuscripts and Text of the New Testament covers verses 12 and 19
  • Metzger’s Textual Commentary on the Greek New Testament covers 15, 17, 19, and 23.

Immediately we need to note that verse 12 is absolutely certain as to the original words as well. Verse 19a is mentioned in Comfort’s textual commentary because he is drawing attention to the “Son of Man” being written as a nomen sacrum (“sacred name” that is abbreviated) in two early manuscripts (א W), as well as in L. Therefore, verse 19a is absolutely certain as well. We are now down to five variants. The original readings of verses 15, 17, 19a and the two in verse 23 where variants occur are almost certain. The textual scholars on the committees for four leading semi-literal and literal translations (ESV, LEB, CSB, and the NASB) agree on ten of the eleven variants. There is disagreement on Matthew 11:15. Even so, the reader has access to the original and alternatives in the footnote.

“He who has ears to hear, let him hear.” (ESV, NASB, UASV)

The variant is ο εχων ωτα ακουειν ακουετω “the one having ears to hear let him hear,” which is supported by א C L W Z Θ f1,13 33 Maj syrc,h,p cop

“The one who has ears to hear, let him hear!” (LEB, cf. CSB)

WH and NU have ὁ ἔχων ὦτα ἀκουέτω “the one having ears let him hear,” which is supported by B D 700 itk syrs

As is usually the case in more difficult decisions, the variant readings are divided in their support between the leading Alexandrian manuscripts. One reading has 01 (Sinaiticus) on its side, the other has 03 (Vaticanus). This tends to cancel out the weight of documentary evidence.

Now, we return to the charts above. There are 138,020 words in the New Testament. Just 1,392 textual variants deemed relevant for translation have enough of an issue to even be considered in the textual commentary. Again, if we average three words per variant, this amounts only to about 3.026 percent of the 138,020 words, or about 6 percent when we compensate for variant units ignored by the GNT editors. We can also remove the 505 {A} ratings because they are certain. Then, we really have no concerns about the {B} ratings because they are almost certain as well. This means that out of 138,020 words in the Greek New Testament, we only have 364 variants (1,092 words by our average) with which we have difficulty, a mere 10 of which involve great difficulty in deciding which reading to put in the text. Our average would make these variants 0.791 percent of the text without accounting for any difficult variants not included because they were considered irrelevant for translation.

We need not be disturbed or distracted by worries of how many variants there are, or whether they are significant or insignificant. We need only to deal with the certainty of each variation unit, endeavoring to determine the original reading. We should also be concerned with the role textual criticism plays in apologetics. There is no possibility of apologetics if we do not have an authoritative and true Word of God. J. Harold Greenlee was correct when he wrote, “Textual criticism is the basic study for the accurate knowledge of any text. New Testament textual criticism, therefore, is the basic biblical study, a prerequisite to all other biblical and theological work. Interpretation, systemization, and application of the teachings of the NT cannot be done until textual criticism has done at least some of its work.”[17] We would add apologetics to that list for which textual criticism is a prerequisite. How are we to defend the Word of God as inspired, inerrant, true, and authoritative, if we do not know whether we even have the Word of God? Therefore, when Bible critics try to muddy the waters of truth with misinformation, it is up to the textual scholar to correct the Bible critic’s misinformation.

Again, it is true that Lightfoot erred if he was counting the manuscripts instead of the variants. However, we need not count variants either but rather variation units, namely, the places where there are variations. The above Colossians 2:2 example of variations that are found in 79 manuscripts was seen to have 14 variants in 79 manuscripts, not 79 variants. While this is true, it is also true that this is simply one variation unit, i.e. one place, where a variation occurs. This may sound as though we are trying to rationalize a major problem of hundreds of thousands of variants. However, it is actually the other way around. The Bible critic is misrepresenting the facts, trying to talk about an issue without giving the reader or listener all of the facts. We need to consider Benjamin Disraeli’s words on statistics: “There are three types of lies: lies, damn lies, and statistics.”

Certainty of the Original Words of the Original Authors

Virgil (70-19 B.C.E.) wrote the Aeneid between 29 and 19 B.C.E. for which there are only five manuscripts dating to the fourth and fifth centuries C.E.[18] Jewish historian Josephus (37-100 C.E.) wrote The Jewish Wars about 75 C.E., for which we have nine complete manuscripts, seven of major importance dating from the tenth to the twelfth centuries C.E.[19] Tacitus (59-129 C.E.) wrote Annals of Imperial Rome sometime before 116 C.E., a work considered vital to understanding the history of the Roman Empire during the first century, and we have only thirty-three manuscripts, two of the earliest that date 850 and 1050 C.E. Julius Caesar (100-44 B.C.E.) wrote his Gallic Wars between 51-46 B.C.E.,[20] which is a firsthand account in a third-person narrative of the war, of which we have 251 manuscripts dating between the ninth and fifteenth centuries.[21]

On the other hand, New Testament textual scholars have over 5,800 Greek manuscripts, not to mention ancient versions such as Latin, Coptic, Syriac, Armenian, Georgian, and Gothic, which number into the tens of thousands. We have many early and reliable manuscripts in Greek and the versions, a good number that cover almost the entire New Testament dating within 100 years of the originals. Therefore, reconstructing the original Greek New Testament is a realistic goal for Bible scholars. This belief and goal that we could anticipate a time when we would recover the original wording of the Greek New Testament had its greatest advocates in the nineteenth century, in Samuel Tregelles (1813-75), B. F. Westcott (1825-1901), and F. J. A. Hort (1828-92). While they acknowledged that we would never recover every word with absolute certainty, they knew that it was always the primary goal to come extremely close to the original. When we entered the twentieth century, there were two textual scholars who have since stood above all others, Kurt Aland and Bruce Metzger. These two men carried the same purpose with them, as they were instrumental in bringing us the Nestle-Aland and the United Bible Societies critical editions, which are at the foundation of almost all modern translations.

From the days of Johann Jacob Griesbach (1745-1812), to Constantin Von Tischendorf (1815-1874), to Samuel Prideaux Tregelles (1813-1875), to Fenton John Anthony Hort (1828-1892), to Kurt Aland (1915-1994), to Bruce M. Metzger (1914-2007),[22] we have been blessed with extraordinary textual scholars. These scholars have devoted their entire lives to providing us the transmission of the New Testament text and the methodologies by which we can recover the original words of the New Testament authors. They did not construct these histories and methodologies from textbooks or in university classrooms. No, they spent decades upon decades in working with manuscripts and putting their methods of textual criticism into practice, as they provided us with one improved critical edition after another. As their knowledge grew, the number of manuscripts which they had to work with fortunately grew as well.

Samuel Tregelles stated that it was his purpose to restore the Greek New Testament text “as nearly as can be done on existing evidence.”[23] B. F. Westcott and F. J. A. Hort declared that their goal was “to present exactly the original words of the New Testament, so far as they can now be determined from surviving documents.”[24] Metzger said that the goal of textual criticism is “to ascertain from the divergent copies which form of the text should be regarded as most nearly conforming to the original.”[25] Sadly, after centuries, textual criticism is losing its way, as new textual scholars have begun to set aside the goal of recovering and establishing the original wording of the Greek New Testament. They have little concern for the certainty of a reading as to whether it is the original.

In speaking of the positions of agnostic Bart D. Ehrman (author of The Orthodox Corruption of Scripture) and David Parker (author of The Living Text of the Gospels), Elliott overserved, “Both emphasize the living and therefore changing text of the New Testament and the needlessness and inappropriateness of trying to establish one immutable original text. The changeable text in all its variety is what we textual critics should be displaying.”[26] Elliott then reflects further on his goals within textual criticism: “Despite my own published work in trying to prove the originality of the text in selected areas of textual variation … I agree that the task of trying to establish the original words of the original authors with 100% certainty is impossible. More dominant in text critics’ thinking now is the need to plot the changes in the history of the text. That certainly seemed to be the consensus at one of the sessions of the 1998 SBL conference in Orlando, where the question of whether the original text was an achievable goal received generally negative responses.”[27]

We strongly disagree. The goal of textual criticism had been and still should be to restore the New Testament Greek text in every word that was originally penned by the New Testament authors, in a critical edition. If we are aiming only “to plot the changes in the history of the text,” as Elliott put it, we are unable to do so precisely at the time when we have the greatest need to see what happened, i.e. soon after the NT books were first published, if we actually deny and rob ourselves of any chance to recover the original. Then we must admit either that we can never have the complete word of God (the new position), or that any and potentially every quality Greek witness must be considered the word of God. The latter might even be said of a quality version, or at least of readings clearly inferred from such a version. In reality, however, any manuscript that departs from the original in its witness is more or less damaged goods.

We obviously do not think such pessimism is the necessary or inevitable response. In looking at the numbers above as to the certainty level of the restoration of the original Greek New Testament, we have come a long way since John Fell (1625-1686). A spot comparison of changes in ratings between GNT5 and previous GNT editions indicates that the level of certainty is increasing in most cases, and when it does not, the preference tends toward the earliest and most reliable manuscripts.[28] To set aside the primary goal of textual criticism now would be an insult to the lives of many textual scholars who preceded us, not to mention to the authors who penned the New Testament books and the Almighty God who inspired them.

[1] (Bart D. Ehrman, Misquoting Jesus: The Story Behind Who Changed the Bible and Why 2005, 7)

[2] Ibid., 89-90


[4] How We Got the Bible (Grand Rapids: Baker, 2003; p). Lightfoot says (53-54)

[5] Baker Encyclopedia of Christian Apologetics, by Norm Geisler (Grand Rapids: Baker, 1998; p. 532)

[6] It is true that some scholars, such as Philip Comfort, argue that the NU could be improved upon because in many cases it is too dependent on internal evidence, when the documentary evidence should be more of a consideration in choosing readings. It should be pointed out, however, that this is in only a relative handful of places, when one considers 138,020 words in the Greek New Testament, and it is hardly consequential. I would also mention that this writer would agree with Comfort in the matter of giving more weight to documentary evidence.

[7] Epignosis is a strengthened or intensified form of gnosis (epi, meaning “additional”), meaning, “true,” “real,” “full,” “complete” or “accurate,” depending upon the context. Paul and Peter alone use epignosis.

[8] Recall that NU is an acronym for two critical manuscripts: (1) Nestle-Aland Greek Text (28th ed.) and (2) United Bible Societies Greek Text (5th ed.)

[9] This is only a partial list of the manuscripts, as we are just offering an example, to see how we count the variants.

[10] Eldon Jay Epp and Gordon D. Fee, Studies in the Theory and Method of New Testament Textual Criticism (Grand Rapids, MI: Eerdmans, 1993), 58.

[11] Ibid. 58.

[12] Ibid. 58.

[13] Word Counts for Every Book of the Bible …, (accessed April 20, 2017).

[14] The NU (D Θ f1 ita,e syrs) has the above reading. A variant, WH and Byz (א (B2) C L W 0102 0281 Maj) add “and of the dish.” The variant is an addition taken from the previous verse.

[15] How We Got the Bibles, by Neil R. Lightfoot (Grand Rapids: Baker, 1998; p. 95-103)

[16] Eberhard Nestle and Erwin Nestle, Nestle-Aland: NTG Apparatus Criticus, ed. Barbara Aland et al., 28. revidierte Auflage. (Stuttgart: Deutsche Bibelgesellschaft, 2012), 31–32.

[17] Introduction to New Testament Textual Criticism, by J. Harold Greenlee (Peabody: Hendrickson Publishers, 1995; p. 7)

[18] Preface | Dickinson College Commentaries. (April 25, 2017)

[19] Honora Howell Chapman (Editor), Zuleika Rodgers (Editor), 2016, A Companion to Josephus (Blackwell Companions to the Ancient World), Wiley-Blackwell: p. 307.

[20] Carolyn Hammond, 1996, Introduction to The Gallic War, Oxford University Press: p. xxxii.

Max Radin, 1918, The date of composition of Caesar’s Gallic War, Classical Philology XIII: 283–300.

[21] O. Seel, 1961, Bellum Gallicum. (Bibl. Teubneriana.) Teubner, Leipzig.

  1. Hering, 1987, C. Iulii Caesaris commentarii rerum gestarum, Vol. I: Bellum Gallicum.(Bibl. Teubneriana.) Teubner, Leipzig.

Virginia Brown, 1972, The Textual Transmission of Caesar’s Civil War, Brill.

Caesar’s Gallic war – Tim Mitchell. (April 25, 2017)

[22] These textual scholars provided us with histories of the transmission of the New Testament text and methodologies. However, we have had dozens of textual scholars who have given their lives to the text of the New Testament. To mention just a few, we have Brian Walton (1600-1661), John Fell (1625-1686), John Mill (1645-1707), Edward Wells (1667-1727), Richard Bentley (1662-1742), Johann Albert Bengel (1687-1752), Johann Jacob Wettstein (1693-1754), Johann Salomo Semler (1725-1791), Johann Leonard Hug (1765-1846), Johann Martin Augustinus Scholz (1794-1852), Karl Lachmann (1793-1851), Erwin Nestle (1883-1972), Allen Wikgren (1906-1998), Matthew Black, (1908-1994), Barbara Aland (1937-present), and Carlo Maria Martini (1927-2012).

[23] Tregelles, An Account of the Printed Text of the Greek New Testament, 174.

[24] Westcott and Hort, Introduction to the New Testament in the Original Greek, 1.

[25] Metzger, The Text of the New Testament, v.

[26] J. K. Elliott, New Testament Textual Criticism: The Application of Thoroughgoing Principles: Essays on Manuscripts and Textual Variation, 592.

[27] Ibid. 592.

[28] Sample comparisons of the General Epistles in GNT5 with previous GNT editions led to this conclusion. When the level of certainty decreased–which was infrequent compared to the reverse–the trend seemed to be that more weight was being given to 03 and/or 01 in opposition to internal factors. It is also expected that certainty levels will increase with the use of the CBGM (discussed in detail below).