The Textual History of the Books of the New Testament

Contents: Introduction
The Books:
The Gospels: Matthew, Mark, Luke, John
Paul: Romans, 1 Corinthians, 2 Corinthians, Galatians, Ephesians, Philippians, Colossians, 1 Thessalonians, 2 Thessalonians, 1 Timothy, 2 Timothy, Titus, Philemon, Hebrews
Catholic Epistles: James, 1 Peter, 2 Peter, 1 John, 2 John and 3 John, Jude


The history of the New Testament text cannot be written based on our present knowledge. We do not know, and likely will never know, how the original text was transmuted into the forms found in our present manuscripts.

And yet, knowing textual history is important for criticism. The more we know about it, the better we are able to reconstruct the original text. And there are certain things which all critics will agree on -- e.g. the existence of the Alexandrian and Byzantine text-types, and the broad nature of (though not the reason for or significance of) their differences.

This article attempts to briefly outline what little we know about the history of the various New Testament books. Much of what is said here parallels the material in the article on Text-Types, but the emphasis is different. The discussion is concerned primarily with major changes and deliberate (recensional) activity.

The sections which follow are organized by corpus, and then by book within the corpus. In general this document does not attempt to give a definitive history, but merely to outline the questions while allowing the student to form conclusions.

The Gospels

Most of the evolution of the gospels took place after they were gathered into a single corpus. Of the four widely-acknowledged text-types, three (the Alexandrian, Byzantine, and "Western") are universally agreed to be found in all four gospels. This is less certain in the case of the "Cæsarean" text, which has been studied primarily in Mark -- but if it exists at all, it almost certainly exists for all four gospels.

Both the Alexandrian and "Western" text-types appear to date back to at least the second century. In the case of the Alexandrian text, this is based on the age of the early papyri, most of which, including P66 and P75, have Alexandrian texts. The age of the "Western" text is based on the witness of early writers such as Irenæus.

The date of the "Cæsarean" text is uncertain. It is often described as a combination of the Alexandrian and "Western" texts, but this is not true. (If it were, it would imply that the "Cæsarean" text is the result of recensional activity. But the type is not unified enough for this.) Rather, it has a combination of readings characteristic of those text-types (this is inevitable, since most variants are binary), with some variants of its own (e.g. "Jesus Barabbas" in Matt. 27:16-17; also a very high number of harmonizing variants, at least in Mark). If those who champion the text-type are correct, it was in existence by the third century, when Origen used it.

The earliest Greek witness to the Byzantine text is the uncial A, of the fifth century. The Peshitta Syriac is also largely (though not overwhelmingly) Byzantine; its date is uncertain though it is usually ascribed to the fourth century (and can hardly be later than this).

Hort thought that the Byzantine text was recensional (i.e. that someone, perhaps Lucian of Antioch, assembled it). Certainly it is more unified than any of the other text-types. But it is now generally believed that even the Byzantine text evolved naturally. There is thus no evidence of recensional activity in the gospels as a whole.


Of the gospels, Matthew shows the fewest signs of recensional activity. There are no changes in writing style and few truly major variants. Unlike in Luke, the text of Codex Bezae appears to have evolved naturally. This is perhaps not surprising; Matthew is usually the first and most-quoted gospel. It influenced the others rather than being influenced by them. It would seem likely that we have it very nearly as it was written (c. 80 C.E.?).


If Matthew has suffered the least textual activity, Mark has probably suffered the most. Generally held in low esteem and rarely quoted, it is always vulnerable to assimilation to Matthew or Luke.

The other side of this is that scribes have been less concerned with the text of Mark. Since no one used it, why bother correcting it to the prevailing text? This means that there are a number of manuscripts -- among them D Y 28 565 -- which are much more interesting in Mark than elsewhere.

But though minor changes in Mark are common, they seem to have happened almost at random. Few serious attempts seem to have been made to edit the book (probably because it was so little used). There is only one place in Mark where recensional activity has clearly taken place. This is in the ending of the book (the material following 16:8). In some texts, the book ends here; in others, we find either of two possible endings, often combined.

The earliest Alexandrian text, as represented by Aleph and B plus one manuscript of the Sahidic Coptic, clearly had no ending. It is possible that the prototype of the "Cæsarean" text ended here, as many of the oldest Armenian manuscripts and the two best Georgian manuscripts omit, while Family 1 and others have critical signs around the passage.

From the only surviving African Latin witness, k, comes the so-called "short ending," three dozen words obviously written to round off a defective manuscript.

Originating perhaps with the "Western" text (D ff2, etc.; b is defective here) is the well-known "long ending," found in most editions and supported by the entire Byzantine text. (It is, however, by no means certain that all European Latin manuscripts support this reading; the most important of these manuscripts, a, is defective here; the pages have been removed and replaced by a vulgate text. Space considerations seem to indicate that there was not room for the longer ending; its lack may explain why the pages were removed.)

Finally, in many late Alexandrian witnesses (L Y 083 099 579 and many Coptic and Ethiopic manuscripts) we find the longer and shorter endings combined, often with critical notations.

It should be noted that the style of the common ending, "16:9-20," does not match that of the rest of Mark. It also seems to be derived from materials in the other gospels and even the Acts.

The conclusion seems clear: Whether by accident or design, the published gospel of Mark ended at 16:8. (It is barely possible that Matthew had access to the real ending; it is even less likely that Luke had this ending). This lack was severe enough that at least two attempts were made to mend the gospel. The more minimal of these is the short ending of k, which cannot be original. The longer ending is better supported, but textual and stylistic considerations also argue against it.


If the gospels of Matthew shows no evidence of recensional activity, and that of Mark shows it only at the end, there is clear evidence of editorial work in Luke. (In fact, we know that Marcion used this gospel, and only this, and edited it in his own way. However, that edition has perished completely; our comments are based solely on the differences observed between surviving documents.) The differences between the Alexandrian and "Western" texts are so pronounced that they can hardly have arisen entirely by accident. Many examples can be offered, but the two best are offered by Luke's genealogy of Jesus and by the so-called "Western Non-Interpolations." The list below summarizes these variants, with the UBS/Alexandrian/Majority reading first (with a summary of supporters), followed by the "Western" reading (with a complete list of supporters):

The overall effect of this is to make it effectively certain that either D or the Alexandrian/Byzantine text has been edited. And the fact that D uses Matthew's genealogy strongly argues that D is the edited document. Does this mean that the entire "Western" text is an editorial production? This is not clear. It will be seen that none of the other Latin witnesses support D's genealogy of Jesus, and even the "Western Non-Interpolations" have only partial support from the Latin, Syriac, and Georgian witnesses. Kurt Aland has argued that the "Western" text, as a type, does not exist. The evidence for his view (in the Gospels) is significant -- but not overwhelming; the final decision must be left to the student. (We should note, however, that there is clearly a Greek/Latin type in Paul.)


Literary problems swirl around the Gospel of John: Who wrote it? When was it written? In what location? What is its relationship with the Synoptic Gospels?

Textual criticism can shed little light on these questions. (Though the manuscripts demolish Baur's proposal for a late date. Two important papyri of John -- P52 and P66 -- date from the second century, and more follow soon thereafter. Thus the book cannot be much more recent than 100 C.E. With this in mind, we can turn to the state of the book itself.)

The textual problems in John revolve around two sections: The story of the Adulteress ("John 7:53-8:11") and the entirety of Chapter 21.

Internal evidence would make it appear that Chapter 21 is an addition. The ending of Chapter 20 reads like the end of the book -- and then we find Chapter 21, a seeming afterthought, with perhaps the purpose of explaining the death of the "Beloved Disciple."

But there is not the slightest textual evidence for this. Every known manuscript contains chapter 21. (Philip Wesley Comfort has argued that neither P5 nor P75, which are single-quire codices, contained enough leaves to hold John 21. This is possible, but by no means convincing. Both codices break off well before John 21; it is possible that the scribes would have condensed their writing to save space. And if that proved insufficient, they could have added additional leaves at the end. All Comfort's calculations prove is that we cannot be certain these documents contained Chapter 21.) Chapter 21 may well be an addition to the book, but if so, it was almost certainly added before the gospel entered widespread circulation.

The case of the Adulteress is rather different, as here there is variation in the manuscripts. But this case is not parallel to, say, Mark 16:9-20, where the text-types disagree. Here almost all the evidence is hostile to the passage.

Taking the internal evidence first, we observe that the language is clearly non-Johannine. This likely will be evident to any who read the passage in Greek, but we can put it on an objective basis. In this passage of twelve verses, there are no fewer than four words hapax legomena, and four other words (one of them used twice in the passage) which occur only two to four times in the NT. By comparison, in the 52 legitimate verses of John 7 there are five hapax, and five other rare words. In the following 48 verses of John 8, there are no hapax and only three rare words. In fact, John as a whole (867 verses) contains only 58 hapax, or one every fifteen verses. It is not impossible that an author who used such a simple vocabulary could manage to insert so many rare words into such a short passage -- but it's not very likely, either.

In addition, the story shows every sign of being unassimilated folklore (for discussion, see the article on Oral Tradition). It is true that many other parts of the gospel rest on oral tradition -- but in all cases it has been assimilated: smoothed out and placed in an outside context. The Adulteress has not been placed in context, which is exactly what we would expect of folklore.

The external evidence argues strongly against its inclusion. Even if it is accepted as scriptural, it appears in no fewer than five different places in the manuscripts:

Thus the evidence clearly indicates that the story of the Adulteress is an addition to John, and probably not an original part of any of the gospels. If it is to be included in Bibles at all, it should be treated as an independent incident.

The Acts

Of all the books of the Bible, none shows such intense textual variations as Acts. There are thousands of differences between the texts of B and D -- often so substantial as to significantly change the meaning of the passage.

This leads to three questions: First, is the D text actually representative of the "Western" text? Second, is the "Western" text recensionally different from the Alexandrian, or did the differences arise naturally? Third, if the two are recensionally different, which recension is original?

To address the first question, we provide the following table illustrating differences between D and other so-called "Western" witnesses. The table tabulates all readings of D in the Nestle-Aland text which are not shared by either the Alexandrian or the Byzantine texts (defined in this case as readings of D which are not shared with any of the group P74 Aleph A B or M/pm). The number of agreements with each of the most important so-called "Western" witnesses is listed, followed by the percent of the time each agrees with D. Chapters are grouped in blocks of four. Note: Family 1739 is defined as the reading of 1739, or at least two of the group 323 630 945 1891 against 1739 if 1739 is Byzantine. Family 2138 is defined by any non-Byzantine member of the group, here represented by 614 1505 2495. A "Unique reading of D" is defined as a reading of D for which Nestle shows no Greek or versional support and no more than one patristic supporter.

rdgs of D
of D
Chapters 1-412829 (23%)10 (8%)1 (1%)6 (5%)36 (28%)27 (21%) 13 (10%)31 (29%)
Chapters 5-810324 (23%)14 (14%)1 (1%)8 (8%)18 (17%) 9 (9%)10 (10%)38 (37%)
Chapters 9-126413 (20%)2 (3%)1 (1%)3 (5%)14 (22%) 15 (23%)15 (23%)20 (31%)
Chapters 13-1617064 (38%)14 (8%)4 (2%)21 (12%) 40 (24%)15 (9%)36 (21%)25 (15%)
Chapters 17-2016658 (35%)6 (4%)3 (2%)18 (11%) 35 (21%)12 (7%)22 (13%)--
Chapters 21-226129 (48%)1 (2%)0 (0%)1 (2%) 20 (33%)5 (8%)2 (3%)--
Totals:692217 (31%)47 (7%)10 (1%)57 (6%) 163 (24%)83 (12%)98 (14%)114/395 (29%)

The above numbers should instantly demolish Von Soden's contention that 1739 is "Western" in Acts! The question is, can any of the other texts listed here be considered to belong to this type? Note that fully 31% of D's readings are singular, and none of the other witnesses agree with more than 30% of its readings.

It is the author's opinion that D is not a proper representative of the "Western" text; rather, it is an edited text based on "Western" materials. (This is similar to the views of Kurt Aland, except that Aland does not offer an explanation for the other "Western" texts.) Still, this is a point upon which scholars will differ, and in any case there is still a "Western" text -- which must be reconstructed, laboriously, from the Latins and copG67 and other witnesses (it is by no means clear, however, that Family 2138 is part of the "Western" text.)

This brings us to the question, Is the "Western" text recensionally different from the Alexandrian and Byzantine? If we subtract D, this is a difficult question. With no reliable Greek witness to the type, some of the variations may be translational.

Given the state of the evidence, we cannot make a certain statement. The sundry "Western" witnesses do not appear to form a true unity, so they cannot form a recension. But our evidence is imperfect. It would seem that the "Western" witnesses attest to an influence, similar to but not actually derived from D. Many of the readings of this text differ recensionally from the Alexandrian text, but by no means all.

Under the circumstances, it would appear that -- here if nowhere else -- the Alexandrian/Byzantine recension is clearly superior. But much remains uncertain. Some scholars have proposed, e.g., that Luke produced two editions of his work -- with the Alexandrian being probably the "official" edition, but the other survived because copies were so difficult to produce. In this case, how does one decide which reading is "original?" Questions such as this must be left as an exercise for the student.

The Pauline Epistles

The textual theory of Westcott and Hort held that the text-types in Paul were the same as in the Gospels: Alexandrian (Aleph A C 33 etc.), "Western" (D F G Old Latin), and Byzantine (K L 049 etc.), with B being mostly Alexandrian with "Western" readings.

Two discoveries changed this: P46 and 1739. 1739 united the semi-Alexandrian witnesses M (0121+0243), 6, 424c. P46 was even more significant, because it showed that the peculiar text of B is not peculiar. Zuntz later showed that P46 and B formed the key witnesses to a separate textual grouping. Zuntz called this group "Proto-Alexandrian" (implying that the later Alexandrian text evolved from it), and listed 1739, the Sahidic Coptic, and the Bohairic Coptic as additional witnesses. All this may be questioned; in particular, it appears that the mainstream Alexandrian text (Aleph A C 33 81 1175) is not actually descended from P46 and B; also, 1739 appears to head its own group. Still, it can be regarded as established that there are additional text-types beyond the traditional three.

It is also noteworthy that the "Western" text of Paul shows none of the peculiarities of Codex Bezae. The "Western" of Paul is clearly not a recensional product; its readings are relatively restrained (this is particularly true of the readings of D-F-G together; the close relatives F and G have many peculiarities of their own which likely derive from a common ancestor). Thus a careful scholar will have to take four non-Byzantine groups into account in examining the text of Paul: the Alexandrian (Aleph A C 33 81 1175), the P46/B/Sahidic group, the "Western" text (D F G (629) Old Latin), and the 1739 group (1739 0243 0121 1881 6 424c and (in Romans-Galatians only) 630+2200).


Of the legitimate Pauline epistles, Romans has perhaps the most complex textual history. There are two reasons for this: The nature of the manuscripts and the complicated nature of the literary tradition -- especially with regard to the sixteenth chapter and the doxology ("16:25-27").

Treating the problem of the manuscripts first, it is worth noting that very many manuscripts change their nature in Romans. The most glaring example is 33. In the other epistles, it is a strongly Alexandrian witness, falling closer to Aleph than any other document. In Romans, however, we have a text from another hand, which is largely if not entirely Byzantine.

Much the same is true of 1175 (though the degree is less); in Romans it is Byzantine; in the other Paulines it is mostly Alexandrian. 2464, too, is Byzantine in Romans but Alexandrian/Byzantine mix elsewhere. (On the other hand, a few minuscules, such as 1852 and 1908, probably have better texts in Paul than elsewhere.)

More important, however, is the case of 1739. The colophon claims that the text of Romans is taken, as far as possible, from Origen's commentary on that book, while the other epistles come from an old Origenic manuscript but not from Origen himself. It appears that this is not true -- 1739 shows no clear change in textual affiliation between Romans and 1 Corinthians -- but the possibility must be taken into account that the manuscript has some alien readings here. (There is a bare possibility that this colophon derives from one of 1739's ancestors, and that this ancestor, taken partly from the commentary and partly from another manuscript, became the ancestor of Family 1739.)

And, finally, there is P46. Although no rigorous study has been done, the text of that papyrus appears to be much more wild (and rather less affiliated with B) in Romans than in any other part of Paul.

Thus, in examining the textual history of Romans, one must be very careful to assess the evidence based on its affiliations in this book rather than elsewhere.

Which brings us to the questions of Chapter 16 and the Doxology -- linked problems, as it is the location of the Doxology which causes us to question the origin of Chapter 16. It is true that Chapter 16 seems unlikely in a letter to Rome -- how could Paul, who had never visited Rome, know so many people there? But the question would not be as difficult if it were not for the question of "16:25-27." (The related question of whether or not to include "16:24" need not detain us; even in the unlikely event that this verse be thought original, it merely adds slightly to the uncertainty about 16:25f.)

Although these verses are 16:25-27 in the Textus Receptus, this is not their place in the Byzantine text. In the majority of manuscripts, including L Y 0209vid 6 181 326 330 424 451 614 1175 1241 1505 1881 1912 2492 2495 mvid dem hark geo2 slav, the verses fall at the end of chapter 14. In most of the Alexandrian and "Western" witnesses, however, the verses fall at the end of chapter 16 (so Aleph B C D 048 81 256 263 365 436 630 1319 1739 1852 1962 2127 2200 a b d* f r am ful pesh pal sa bo eth). Some witnesses, usually mixed, have the verses in both places (so A P 0150 5 33 88 104 459 2805 arm geo1). Others omit the doxology altogether (F G 629 dc-vid). P46, astonishingly, places the verses at the end of chapter 15. Even more astonishingly, the minuscule 1506 (which ordinarily has an Alexandrian text) has the verses after both chapter 14 and chapter 15, but omits chapter 16. We are also told (by Origen) that Marcion omitted chapters 15 and 16 of Romans (this testimony should, however, be used with great caution). The capitulations of certain Latin manuscripts also seem to imply that Chapters 15-16 were not part of their texts. (Harry Gamble has speculated that the original text of Family 1739 omitted chapter 16, but the evidence of the family, combined with that of 1506, argues strongly against this.)

What does this mean? This question continues to exercise scholars. Are 16:25-27 any part of Romans? If so, where did they originally belong? The level of support for the location after chapter 16 is extraordinarily strong -- but internal evidence favours the location after chapter 14. Why would any scribe, finding the verses after chapter 16, where they fit, move them after chapter 14, where they interrupt the argument and serve no useful purpose? It has been speculated that the doxology came to be placed after chapter 14 as a result of Marcion's mutilation of Romans, but this is a rather long chain of suppositions. (Not least of which is the supposition that Origen actually knew Marcion's text. Chapter 14 is a strange place to truncate the epistle, as the argument extends to 15:13.)

Did shorter forms of Romans circulate, lacking either chapter 16 or chapters 15 and 16? Gamble, in The Textual History of the Letter to the Romans, offers a good synopsis of the internal evidence (though his data on the external evidence is questionable). But neither sort of evidence allows us to reach a firm conclusion. Apart from the Marcionite product, there is no evidence of a 14-chapter form in Greek, although there may once have been a Latin version. That a 15-chapter form of Romans circulated is proved by 1506, and the evidence of P46 implicitly supports this (as well as implying that this edition was very early). It probably was not widespread, however.

As for the location of the doxology, we simply don't have enough evidence to be dogmatic. My personal opinion is that it is an addition, appended to the end of one edition of the letter and then later moved to the other positions. If this is the case, then the most likely position is perhaps after chapter 14. But this is so uncertain as to amount to speculation.

1 Corinthians

The textual history of 1 Corinthians appears quite simple. It is a single writing, preserved without real evidence of alteration. There are variations, but (with possibly a single exception) all appear accidental.

The exception is in 14:34-35. These verses are found in this position in P46 Aleph A B K L 0150 0243 6 33 81 104 256 330 365 436 451 629 1175 1319 1505 1739 1881 1962 2127 2492 M am bam cav ful hub harl theo tol pesh hark pal s bo fay ar geo eth slav -- but in D F G 88* a b d f reg Ambrosiaster Sedulius-Scottus we find the verses placed after 14:40. It has been supposed by some that the verses were originally lacking; there is, however, absolutely no direct evidence for this; the verses are found in every witness. Only the location varies. It is equally possible that they were moved an attempt at a clarification; it is also possible that a careless scribe omitted them, then someone reinserted them in the wrong place. In any case, a single reading implies very little about the history of the text.

2 Corinthians

The literary history of 2 Corinthians is exceedingly complex. It is possible that it contains fragments of six letters; that it contains portions of at least two is almost certain (the various sections are as follows: 1:1-6:13, a friendly letter to Corinth; 6:14-7:1, on marriage with unbelievers; 7:2-16, rejoicing at word from Titus; Chapters 8 and 9, on the collection for the saints, but possibly two separate discussions on the subject; 10:1-13, Paul's defense of his ministry. The first and last sections can hardly have been in the same letter, and the four intermediate sections may have come from anywhere).

This combination of fragments, however, clearly took place before the text was published, since there are no relevant variants in the tradition. Every known manuscript contains the entirety of all sections of the combined document. Thus these literary factors do not affect the textual criticism of the epistle.


There is little to be said, textually, about Galatians. It is clearly a literary unity, and there is no evidence of editorial tampering. The closest thing to an interesting variant is the alternation (in 1:18, 2:9, 11, 14) between "Cephas/Kephas" and "Peter."


The question of the textual history of Ephesians is closely tied in with the matter of its authorship. We can hardly address the latter here (though I freely admit that the style of Ephesians is so unlike Paul that I cannot believe Paul wrote the letter). But this makes the question of the destination of the letter, in 1:1, crucial. The words "in Ephesus" are found in Alephc A Bc D F G 33 81 104 256 365 436 1175 1319 1881 M a b d f r vg pesh hark sa bo arm geo eth slav, but P46 Aleph* B 6 424c 1739 omit. It seems clear that this is an editorial difference -- and that the form lacking "in Ephesus" is at least as old as the form with it, probably older.

This variation has led to much speculation about the nature and origin of this letter (so clearly linked to Colossians), but this does not affect the textual history, so we leave the problem there.


Until recently, scholarly consensus held that Philippians was a unity. In modern times, though, some have held that the abrupt break in 3:1 (between 3:1a and 3:1b, or between 3:1 and 3:2) indicates a discontinuity, and that Philippians actually consists of two (or perhaps three) letters. In this they are bolstered by Polycarp's remark that Paul had written "letters" to the church in Philippi.

Whether Philippians is a unity or not, it seems clear that it was published as a single letter. There is no evidence of recensional activity in the text.


Textually, Colossians is an unusual case: Of all the epistles, it has suffered the most from assimilation of parallels. It is generally agreed that it is a unity, although some have questioned its Pauline authorship (on insufficient grounds, to my mind). But the great problem of Colossians is its relationship to Ephesians.

That these two letters are dependent cannot seriously be denied. The author of one worked from the other (even if Paul wrote both, it is not impossible that he would have used one as a template for the other -- though, frankly, I find it inconceivable that Paul could have written Ephesians). In all probability, Colossians is the earlier letter.

But it is also the weaker letter (at least textually). Shorter, placed later in the cannon, with less development of its themes, it was almost inevitable that it would in many places be contaminated with wording from Ephesians.

Examples of this assimilation of parallels are so frequent that they simply cannot be detailed here; the matter will be left for the commentaries. It does appear, however, that this assimilation was not deliberate or recensional; scribes simply wrote the more familiar form, as they so often did in the gospels.

1 Thessalonians

As with most of Paul's letters, there is no real evidence of editorial activity in this book.

2 Thessalonians

As in 1 Thessalonians, there is no sign of editorial activity in this book.

1 Timothy

The textual situation in the Pastoral Epistles differs slightly from the rest of Paul. This is not due to editorial activity but to the state of the manuscripts. B does not exist for these books, and. P46 apparently never included them. Thus we are missing a whole text-type.

This might possibly be significant, as these books are among the most questionable of the Pauline Epistles. It is, of course, widely though not universally held that these books are not by Paul, though they may be based on his notes. But as far as we know, this is not a textual question; there are no signs of editorial work in our surviving text-types.

2 Timothy

2 Timothy operates under the same restrictions as 1 Timothy: The book's authorship is in question, and P46 and B lack the book. Of the Pastoral Epistles, it gives the strongest signs of composite authorship, with the personal sections having the genuine Pauline touch while the sections on church order have show all the symptoms of being later than the apostle. But, as in 1 Timothy, there is no reason to believe that the text has been edited since it was published; all the work of combining the Pauline and non-Pauline material preceded publication.


The situation in Titus is exactly the same as in 1 Timothy, and the shortness of the book makes it even less likely that it has been edited.


With a book as short as Philemon, it is difficult to form textual theories. There simply aren't enough variations to work with. But there is no reason to believe that the book has been edited in any way.


Hebrews is unique among the Pauline corpus in many ways. The obvious way is that it is not by Paul. But more noteworthy is the fact that it was not universally recognized as canonical.

The surviving witnesses almost universally include the book (indeed, Hebrews is the only one of the Pauline Epistles for which we have two substantial papyri --P13 and P46); the only manuscripts which lack it are F and G, and this may be because it was missing in their exemplar (we note that these two manuscripts actually ignored lacunae in mid-book). Even so, it is likely that relatively few copies of Hebrews circulated in the second and third centuries, and some of those were probably separate from the rest of the Pauline corpus.

What effect this may have had on the text, if any, is not immediately evident.

The Catholic Epistles

In recent years, the Catholic Epistles have been subjected to many detailed examinations -- due most likely to the fact that their brevity makes them relatively easy to analyse. Scholars such as Amphoux, Richards, and Wachtel have all undertaken studies of the text-types in these books.

In the Catholic Epistles, the "Western" text seems to disappear. There have been various attempts to find it, but these cannot be considered convincing. There are few Old Latin texts of the Catholics, but we find extravagant readings in certain of the Vulgate witnesses (these are detailed in the descriptions of the individual books). These, presumably, are " Western" -- but they simply do not match any of the Greek texts.

The text-type most often associated with the "Western" text (so, e.g. Amphoux) is Family 2138. This large group (Wachtel's Hkgr; Richards's A1) includes, among others, 206, 429, 522, 614, 630, 1505, 1611, 1799, 2138, 2200, 2412, and 2495, as well as the Harklean Syriac. Despite Amphoux, however, this type is not close to the Old Latin, and in Acts it is not overly close to D. (See the table in the section on Acts).

In addition, Amphoux and Richards (though not Wachtel) identify two groups within what has traditionally been called the Alexandrian text. One of these the is Alexandrian text proper (P72? Aleph A B? Y 33 81 436 bo etc.), the other is Family 1739 (1739 1241 1881 323 945 etc.). C seems to be a mix of these two types, but closer to Family 1739.

Richards views these three types -- Alexandrian group, Family 1739, and Family 2138 -- as subgroups of the Alexandrian text. This is, however, clearly incorrect (even Richards is unable to define the differences between the types). Amphoux, who regards the three is distinct types, is correct.

Of the three types, Family 2138 is the most affected by the Byzantine text. Even the best members of the type (2138+1611, 1505+2495, 2412+614, 630+2200) have lost about 20-30% of their family readings to Byzantine influence. As, however, the influence is different in each of the subgroups, it is often possible to determine the original text of the family. Of course, the fact that our witnesses are so late (none except the Harklean Syriac precedes the tenth century, and the Harklean is one of the weaker representatives of the type) may mean that there are additional corruptions we cannot recover.

The Alexandrian text is much earlier and purer. Family 1739 consists of late witnesses (except for C), but its similarity to Origen and its relative closeness to the Alexandrian text, as well as its general freedom from Byzantine readings (at least in the leading witnesses, 1739 C 1241), indicates that it too is early and pure. Thus our tools for reconstructing the text of the Catholic Epistles are perhaps better than for any other section of the New Testament.

Balancing this is the fact that the books became canonical at widely differing dates. While a corpus of Paul must have been compiled early, it was not until quite late that unified editions of the Catholic Epistles would be circulated. This point will be taken up under the individual books.


James was the last of the longer Catholic Epistles to be accepted by the church. Eusebius, for instance, describes it as disputed (III.25; also II.23). It appears in all our manuscripts, however (except P72, which is a special case), and is included in the Peshitta. It clearly circulated widely in the early church. There do not seem to be any particular problems associated with its text; the variations it displays are the sort one would expect in the ordinary course of transmission.

1 Peter

1 Peter was one of only two general epistles to be recognized as canonical "from the beginning" (1 John was the other). It is found in every witness, including P72. Its text is in good condition, and shows little evidence of recensional activity.

The one exception to this is in the Latin tradition. This contains a number of substantial variations. After 1:19, for instance, a few Latin witnesses add "ipse ergo qui et praecognitus est ante constitutionem mundi et novissimo tempore natus et passus est epse accepit gloriam quam deus verbum semper possedit sine initio manens in patre." More important, because better attested, is the addition in 3:22, "deglutiens mortem ut vitae aeternae heredes efficeremur" (z am cav fulc hub sang theo tol val Aug Cass; all Greek witnesses, supported by ful* juv, omit). These readings likely derive from the now-lost "Western" text of 1 Peter, perhaps indicating that it showed some of the same sort of extreme readings we find in the Bezan text of Acts. (The fact that these readings do not occur in Greek is further evidence that the "Western" text is not represented by Family 2138 or any of our other witnesses.) As, however, we have no continuous "Western" texts, there is very little we can do about this problem.

2 Peter

Unlike 1 Peter, 2 Peter did not gain instant recognition as canonical. Moderns see many reasons for this -- it does not read like 1 Peter, it is dependent on Jude, it's much too wordy for a simple Galilean fisherman. How much of this was apparent to the early Christians is not clear, but the fact is that the book was not universally recognized until well into the fourth century. The Peshitta, for instance, omits it. We find it in P72 -- but of course P72 contains sundry non-canonical materials.

Despite this, there is little evidence of deliberate editorial work in 2 Peter. Textually, the most noteworthy thing about this epistle is its relationship to Jude. For the most part, 2 Peter influenced Jude rather than the reverse (2 Peter is longer, more respected, and comes earlier in the canon), but the influence may sometimes have gone the other way as well (see, e.g., the discussion on 2 Peter 2:13).

1 John

1 John is the second of the Catholic Epistles to have been universally accepted as canonical. Since it also has a simple and straightforward text, there seems to have been little temptation to alter it.

The one exception is, of course, 1 John 5:7-8. Priscillian seems to have been responsible for the explanatory Latin gloss "in caelo: Pater, Verbum, et Spiritus Sanctus, et hi tres unum sunt. Et tres sunt, qui testimonium dant in terra" (though Priscillian had the reading in a noticeably different form). This worked its way into certain Latin manuscripts (l r (cav) harl (tol) (valmarg); am bam dubl ful hub juv* mon sang val* omit), from there into a bare handful of Greek manuscripts ((61) 88marg 221marg 429marg (629) 636marg 918, 2318 all with variations), and from there, by an absurd twist of fate, into the Textus Receptus. (A similar Latin expansion, not found in the Textus Receptus, occurs in 2:17, where p t Cyprian Lucifer Augustine samss add variations on "quomodo (et) (sicut) (deus) (ipse) manet in aeternum"; this, however, affected the Vulgate only minimally and the Greek not at all. Similarly in 5:9 Beatus and a handful of Latin manuscripts add "quem misit salvatorem super terram, et fulius testimonium perhibuit in terra scripturas perficiens, et nos testimonium perhibemus quoniam vidimus eum ad adnuntiamus vobis ut credatis, et ideo." A final example occurs in 5:20, where t Speculum (Hilary) Julianus-Toledo add "et carnem induit nostri causa et passus est et resurrexit a mortuis adsumpsit nos.")

The Byzantine text offers a handful of other interesting readings:

Both of these, however, appear to be simple scribal errors that never were corrected.

2 John and 3 John

2 John and 3 John are the shortest books in the New Testament. They are so short that no textual history can be written, and no textual analysis should be undertaken on the basis of their few dozen verses of text. We truly cannot tell their history; recensional activity is possible, since they were adopted into the canon late (and separately). Still, there is clear sign of editorial activity; the most noteworthy variant (2 John 2, omit dia thn alhqeian Y 6 323 614 630 1241 1505 1611 1739 1852 2138 2412 2495 ful hark), despite its strong attestation, appears to be the result of haplography. There is also a typically Latin insertion in 2 John 11, with variations on "ecce praedixi vobis ut in die(m) domini (nostri Iesu Christi) non confundamini" (so pc Speculum and the Sixtine Vulgate, but not am cav ful hub sang tol val etc.).


The book of Jude is a leading candidate for the title of "most textually damaged." Certainly no other epistle is in such poor condition. There are many reasons for this. It was one of the last books to be canonized. It is rather dense and difficult. It parallels 2 Peter, and falls after that book in the canon, meaning that it has suffered more heavily from harmonization.

The witnesses strongly reflect this problem. The Alexandrian text shatters in Jude; the manuscripts show no particular pattern of agreements. The papyri are of little help. P72 has been called "wild" in this book, the fragmentary P78, of about the same date, manages to have two singular readings despite preserving parts of only four verses. We find important omissions and/or additions in almost every major manuscript. A few samples (this list could be multiplied several times over):

This confusion does not mean that Jude has ever been edited; it will be observed that these odd readings are found in all sorts of texts. They simply mean that the text of Jude is in very bad condition, and that its recovery is a difficult and unreliable process. No witness, not even B, can be considered to be very reliable.


The textual evidence for the Apocalypse is the weakest of any part of the New Testament. The surviving manuscripts represent only about a third of the number found for the Epistles and a tenth that for the Gospels. It is not found in the Lectionary. Some early versions, such as the Peshitta, omit it, and we can only speculate about types such as the Old Syriac and Old Georgian.

The good side of this is that it is possible to examine the manuscript tradition approximately in its entirety, as was done by Josef Schmid in Studien zur Geschichte des griechischen Apokalypse-Textes. Given the completeness of this work, we will only briefly outline its contents.

Schmid finds four text-types:

Both the Byzantine group and the Andreas group are very large; where they divide (as they frequently do), it is not really possible to speak of the Majority Text. Both of these groups, as might be expected, break down into smaller subgroups.

It might be noted that Andreas's text is, in effect, a recension. It is not really the result of editorial work, but the intricate relation of text and commentary has ensured that this particular type of text maintains its independent identity. Due to their differing forms of presentation, mixed Andreas/Byzantine manuscripts are relatively rare. It should be noted that the Textus Receptus derives from an Andreas text, and has readings characteristic of the type (and, in fact, a handful derived from the commentary itself, where Erasmus could not tell text from margin in 1r).