The Development of the New Testament Text

by Jim Snapp II

Autographs. A inside a circle.The history of the development of the New Testament begins (of course) with the original text. The first document to contain the original text is called the "autograph." That is what the "A" in this box stands for. By "autograph," I do not necessarily mean the text written by the main author and only the text written by the main author. If I did, we would probably have to toss out John 21:24 and Romans 16:22 -- since John 21:24 seems to be written by someone other than John, and Romans 16:22 obviously originated with Tertius, not Paul. By "autograph" I refer to the first manuscript to contain the text in its original, definitive, published form. Copies of the autographs were hand-copied in different locations.

But not everyone treated the text the same way when they made copies. Just as there are differences between British English and American English, there were differences between Greek dialects, and adjustments were made accordingly. There were also different ideas -- held by different people in different areas -- about how the text should be treated. These ideas influenced the contents of the copies, in at least three different ways.

As copies were made and re-copied in the first three centuries of the church, at least three different types of text arose. Each text-type had its own characteristics, which were determined in part by the approach taken by copyists to the text.

The Western Text [pictured in red] contained many expansions, harmonizations, and paraphrases. The "Western" approach was basically, "Make the text clear by adding to it." The message, not the words themselves, seems to have been the main thing to the Western copyists. The Western Text omits some phrases, too; some of these omissions are attributable to sloppy copying. Also, some "Western" manuscripts show signs of doctrinal tampering, although the question of exactly who was behind the tampering has not been entirely resolved.

The Alexandrian Text [in blue] contained grammatical refinements and sometimes "pruned" the text, removing repetitions and words and phrases which the copyists considered superfluous. The "Alexandrian" approach was basically, "Make the text clear by refining it." Because of the nice climate of Egypt, most ancient manuscripts which contain the Alexandrian Text have been discovered in Egypt.

The Caesarean Text [in green] seems to have mostly been a result of copyists – in a place where there were both Western and Alexandrian manuscripts (like Caesarea) – comparing Western manuscripts and Alexandrian manuscripts, and picking and choosing between the two. Sometimes, though, the Caesarean Text features a variant which is neither Alexandrian nor Western. The Caesarean Text is a text of the Gospels.

Now I want to jump forward in time from this diagram, which depicts a theory of how the New Testament text had developed in the period from the first century to around A.D. 250, to a different diagram, which will depict an important theory (proposed in 1881) about how another text-type originated.

Wescott and Hott's Theory

This diagram depicts the Theory of the Lucian Recension.

Before proceeding further, let me identify a couple of new things: notice these two white areas, overlapping part of both the Western and Alexandrian Texts. The one closer to the center represents the Old Latin translations. There was quite a bit of diversity in the text of the ancient Latin versions, partly because there were several of them, and partly because they were based on, and/or were rapidly influenced by, different texts. The Old Latin is generally "Western" but sometimes has "Alexandrian" features too. The outer block represents the Vulgate, a translation which was made by a scholar named Jerome (and his assistants) in the 380's, with the hope that it would provide a standard Latin text for church-use. Eventually it did so; however, along the way copies of the Vulgate were influenced by copies of the Old Latin(s).

The big orange arc in the diagram represents the Byzantine Text. This text-type is the Greek text displayed (with some variation) by 95% of all Greek New Testament manuscripts. This is the text-type on which the King James Version of the New Testament was based. It is characterized by harmonizations (similar to the Western Text, but less drastic) and liturgically useful adaptations.

Notice that in this diagram, the Byzantine Text has no direct link to the autographs. That was proposed in 1881 by two scholars, Brooke Foss Westcott and Fenton John Anthony Hort (we'll just call them Westcott and Hort or "W&H"). They made the following proposals: The Byzantine Text is completely secondary to the Alexandrian and Western Texts. The Byzantine Text was not used by anyone prior to the year 250 or so. It sometimes combines the readings of the Alexandrian and Western Texts, and sometimes forms a new reading. Where it combines (or "conflates") readings found in the other two text-types, and where it has a new reading, its contents are completely unattested prior to 250, indicating that the Byzantine Text is a late and derivative text. In no instances should its unique readings be regarded as original.

W&H proceeded to dismiss the Western Text as a wildly expanded text, so they generally concluded that the Alexandrian Text – especially when exhibited in agreement by two important manuscripts, Vaticanus and Sinaiticus -- was very very close to the text of the autographs. Since the days of W&H, this view has dominated the field of New Testament manuscript-analysis – also known as Textual Criticism (in this context, "criticism" means "scientific analysis;" it is not a negative term). There is no historical account of a thorough revision of the New Testament text being undertaken as W&H propose. And it would have to have been distributed and accepted speedily, since some church leaders who wrote around the year 300 seem to have used Byzantine manuscripts. Nevertheless, W&H's theory was, and still is, widely accepted. Although W&H only suggested that the Byzantine Text was a revision made by a church-leader named Lucian of Antioch, that idea has been strongly promoted, and is slightly implied by W&H's use of the term "Syrian Text" to describe the Byzantine Text.

A couple of decades after W&H introduced their theory, a scholar named Kirsopp Lake and others affirmed the existence and antiquity of the Caesarean Text. However, the manuscripts which were claimed to demonstrate a "Caesarean" origin frequently contained Byzantine readings. This was accounted for by the theory, depicted here, that the "ancestors" of these manuscripts had originally been heavily Caesarean, but as the text was copied and re-copied, it was influenced by Byzantine readings. So by the time it emerged in the present extant manuscripts, the text was "Mixed" – partly Caesarean, and partly Byzantine. The discovery of the Caesarean Text thus adjusted, but did not drastically alter, W&H's theory or its plausibility.

But then, something happened. Researchers began to discover ancient papyrus copies of parts of the New Testament – documents exhibiting the text used in the third century, and sometimes even earlier. Generally, the texts confirmed the early existence of the Alexandrian Text (which effectively destroys the claim made by some that the Alexandrian Text was produced by a writer named Origen in the 200's). But scattered among the Alexandrian readings were some Byzantine readings! Quite a few readings that had previously been classified as Byzantine (and therefore unoriginal and late) were suddenly discovered to pre-date W&H's proposed date for the birth of the Byzantine Text. How does one explain a uniquely Byzantine variant existing before the revision which (according to W&H) produced all unique Byzantine readings? There were not many unique Byzantine readings in the early papyri, but according to W&H's theory, there should not have been any.

Ordinarily, when textual critics observe, in one manuscript, some variants of one text-type, and some variants of another text-type, they conclude that the copyist, or the copyists of some "ancestors" of the manuscript, used manuscripts which exhibited different text-types. Textual critics were hesitant to take that approach in regard to the papyri, though. If it were affirmed that a Byzantine "ancestor" of the papyri existed, W&H's theory would fall completely apart.

Instead, text-critics almost immediately noticed that the papyri did not contain "block-mixing." This means that although Byzantine readings popped up here and there, there were no large portions of text exhibiting Byzantine variant after Byzantine variant after Byzantine variant – which is the sort of thing that would prove that the copyist was using a manuscript with a genuinely Byzantine Text. So the papyri could be interpreted to indicate that, at early stages, part of the Byzantine Text existed – the parts that are exhibited in the papyri – but the rest did not. The Byzantine Text may thus be seen as a gradually developing text; the Byzantine variants in the papyri do not necessarily imply the antiquity of the entire Byzantine Text.

However, every text-type was a gradually developing text. The Byzantine Text just took longer to reach a standardized form than the Alexandrian Text did. If we treat the Western and Alexandrian Text-types in the same way the Byzantine readings in the papyri have been treated, it is clear that some Western and Alexandrian readings do not have early attestation. Nevertheless antiquity is frequently assumed for all Western and Alexandrian (or Proto-Alexandrian) variants, simply because those text-types, in general, are ancient.

Westcott & Hort's theory that the entire Byzantine Text is late and derivative does not fully interlock with the available evidence. But that does not mean that we should consider the Byzantine Text to be identical to the original text and ignore the Alexandrian, Western, and Caesarean manuscripts. There's a better way to evaluate the evidence.

Textual Araneum

Here is how I think the evidence should be viewed. None of the text-types were created overnight. The Byzantine, Western, and Alexandrian text-types all developed gradually – in other words, the Byzantine, Western, and Alexandrian texts all gradually detoured from the autographic text, and developed their own variants and characteristics. The Caesarean Text is sort of an X-factor, which developed at a location where the Alexandrian and Western Texts (and later, the Byzantine Text) intersected. But each text-type can be filtered against the others, removing the non-original readings that are unique to each text-type, to reconstruct the text of the archetype of all copies. In some cases it is clear that the Byzantine Text contains a variant for which the attestation is late. In some cases, something similar will be true of the Alexandrian, Western, and Caesarean Texts. But in no case should a Byzantine reading be dismissed simply because it is Byzantine. In every case there is a possibility that the Byzantine reading was a late development, but in every case there is also a possibility that the Byzantine reading is early, and that it preserves the original text.

This elicits a desire to evaluate readings without pre-judging the Byzantine readings to be late, and without a reflex to favor the agreement of Vaticanus and Sinaiticus (which tend to share one early text-type -- the Alexandrian Text). In many cases, this approach yields results which oppose the readings adopted in the modern Critical Text and support the originality of readings in the Majority/Byzantine Text.

Textual Araneum of the BibleHere is the full Textual Araneum (Web-structure), which I present to give some idea of how the influence of the various text-types spread, and how they have affected both English and non- English translations. This arrangement shows the "genealogy" of the various branches of the manuscript evidence.

Two final notes:

In the Araneum, I have separated the Textus Receptus from the Byzantine Text. This is because when one closely compares the two, they are different in many points – not so much as to form a different message, but more than enough to disprove the claim that the King James Version New Testament (which is based on the Textus Receptus) agrees completely with the Byzantine Text. In some cases (Ephesians 3:9, for instance) the underlying text of the KJV contains a late, unoriginal reading.

Also, the Araneum is just an estimate. The picture would vary somewhat depending on whether one wanted to show the relationship of manuscripts of the Gospels, Acts, the Pauline Epistles, the General Epistles, or Revelation. Nevertheless I hope it helps you visualize the development of the New Testament text.

Manuscript Evidence

Now that some of the Big Ideas about the manuscript-evidence have been reviewed, let's get acquainted with the manuscript-evidence itself.

The evidence that is used to establish the New Testament text is categorized into four groups:

  • Greek manuscripts
  • Early versions (especially early Latin and Syriac translations)
  • Patristic evidence (quotations from early church writers)
  • Lectionaries (worship-manuals containing extensive quotations from the New Testament).

The evidence from versions, early church writers, and lectionaries is very important -- especially since the earliest attestation of many passages is in the patristic evidence, rather than in manuscripts (For instance, a quotation made by a church leader named Irenaeus around the year 177 confirms the early existence of the last 12 verses of Mark). But much of the most important evidence is contained in the Greek manuscripts.

Greek manuscripts may be categorized into four sub-groups:

  1. PAPYRI - texts written on papyrus. Numbering over 100, these include the earliest known copies of Scripture:
    1. P-52, the John Rylands Papyri, from about A.D. 125, a fragment containing text from John 18.
    2. P-66, from about 200 (some say slightly earlier), a damaged copy of the Gospel of John.
    3. P-45 and P-46, from about 225.
    4. P-45 mainly consists of portions of Mark, Luke, and Acts.
    5. P-75, from about A.D. 200, a copy of Luke and John.
  2. UNCIALS - texts written on parchment in ancient capital-letter script with little punctuation and generally no spaces between words. These include six major codices (books, that is -- a "Codex" is a book), each of which has its own symbol:
    1. Codex Vaticanus (B) - a damaged Bible containing the text of most of the books of the New Testaments, probably produced around 325.
    2. Codex Sinaiticus (א [Aleph]) - a nearly-complete Bible containing every book of the New Testament, probably produced around 350-375. It is likely that Sinaiticus and Vaticanus were produced by copyists who had been trained at the same location.
    3. Codex Bezae (D) - a Greek-Latin copy of the Gospels, Acts, and Third John, probably produced in the 400's (though later dates have been assigned).
    4. Codex Alexandrinus (A) - a damaged Bible (most of the book of Matthew is missing) produced around A.D. 450.
    5. Codex Ephraemi (C) - a heavily-damaged text produced in the 400's (and subsequently re-used), containing extensive portions of the Bible.
    6. Codex Washingtonensis (W) - a copy of the Gospels from about A.D. 400.
  3. MINUSCULES - texts written in cursive lettering (typical of later manuscripts). There are hundreds of cursives. Some of the most famous ones are:
    1. 33 - a New Testament (missing Revelation) from about the year 850.
    2. 565 - the four Gospels, from about 900, written in gold letters on purple parchment.
    3. 700 - the four Gospels, from about 1100.
  4. MISCELLANEOUS - a minor sub-group consisting of everything not included in the others (such as amulets and tomb-inscriptions).

The hand-copying of manuscripts was not an easy task. Copyists could, and sometimes did, accidentally skip words, phrases, and entire lines of text. Some copyists made their work very fancy (see, for example, the Lindisfarne Gospels, a Latin text of the four Gospels from the 700's). After the invention of the printing press in the 1450's, when Johann Gutenberg produced the Gutenberg Bible (containing the text of the Vulgate), hand-copying gradually ceased.

Some manuscripts are more valuable than others. A manuscript that is in good condition, contains a lot of text, and that is perceived (mainly via handwriting-analysis, technically known as palaeography) to come from the early centuries of the church, is clearly more valuable than a manuscript that is badly damaged, fragmentary, and was written in the Middle Ages. But besides the condition, extent, and age of a manuscript, another important thing to consider is a manuscript's "genealogy." Manuscripts can be classified into family-like groups, according to the variants they have in common (and other shared features, such as book-arrangements and headings).

When a manuscript-analyst, or "textual critic," begins to evaluate a manuscript, several things need to be done: first, the provenance of the manuscript should be determined as much as possible: where was it found? Where and when was it written? Then readings unique to the manuscript should be identified, and readings which are manifestly mistakes made by the copyist should be filtered out, in an attempt to ascertain the text of the exemplar (master-copy) of the manuscript. The next step is to ascertain the type of the manuscript's text. Some manuscripts have a "mixed" text, which suggests that their genealogy includes, somewhere along the line, manuscripts from more than one text-type.

Once the text-type(s) of a manuscript has been identified, the textual critic's next goal is to determine the archetype of all manuscripts which belong to that text-type. After the archetypical text of each text-type is determined, the next step is to discern the archetype of all text-types (that is, the original text) by engaging in a case-by-case analysis of the points at which the text-types contain divergent readings.

This is quite a task, because even though the basic message of the New Testament is not drastically altered by any text-type, the words themselves have been altered quite a bit. Centuries ago, the preferred Greek text was the Textus Receptus (which means "Received Text"), a branch of the Byzantine Text-type which was promoted during the Protestant Reformation by Desiderius Erasmus (who made printed editions of it, beginning in 1514), William Tyndale (who made an English translation based on it, in 1525), Robert Stephanus (who re-edited it and added verse-divisions to it in 1550-1551), and others. The King James Version was based on the Textus Receptus.

Currently, the text which most scholars consider the closest representative of the original text is the "Critical Text" -- i.e., the text that is produced by an analysis of many manuscripts and text-types. The standard edition of the Critical Text is the Nestle-Aland Novum Testamentum Graece (which contains the same text as the United Bible Societies' 4th edition of the Greek New Testament.) Its Introduction explicitly denies that it is a "definitive" work; nevertheless it is widely used as a standard text.

The Textus Receptus and the Critical Text are significantly different. How different? Some scholars have said that they have over 5,000 differences. Some of those differences merely involve word-order and spelling. Perhaps a more meaningful way of depicting the differences is as follows: Stephanus' 1550 edition of the Textus Receptus includes 4,168 words which the 27th edition of the Nestle-Aland text does not include, and the Textus Receptus omits 1,672 words which the Nestle-Aland text does include. (I arrived at those totals via a manual comparison.)

A third type of Greek text which has been published is the Majority Text -- the editors of which aspired to simply adopt the reading that is found in the majority of all Greek manuscripts. In effect, this produces the archetype of the Byzantine Text-type, free of the late alterations which are contained in the Textus Receptus. Some scholars consider the Majority Text to be congruent to the original text. Most, however, regard it only as a useful tool for the study of the Byzantine Text.

Much of the work of New Testament textual criticism has already been done. Although some well-respected textual critics would disagree with me, I believe that currently there are only about 1,400 points in the New Testament text where the evidence is closely balanced between two or more variants. (Variants in the book of Revelation are not included in this count.) Often either reading may be adopted with no appreciable difference in the meaning of the text. But in some cases the difference significantly affects the meaning.

The field of New Testament Textual Criticism merits more attention from conservative Christian scholars, not only with the goal of resolving the remaining problems, but to increase the stability of the text, to equip the church to respond to questions about translation-differences which involve differences in the translations' underlying Greek texts, and to be prepared to cogently handle any new manuscript discoveries which the future may hold.

To learn more about Textual Criticism and early manuscripts of the New Testament, explore the following links: