Here we have used slashes to indicate minor word-order reversals.

Although there are many differences in omission and substitution of words and phrases, what remains is an astounding agreement in both vocabulary, syntax, and exact word order between these two texts, in a mere 12 verses.

This seems so trivial, that we tend ignore this, or forget about it completely. Of course it is a truism that almost any two random portions of scripture will have this kind of agreement between any two manuscripts. But we focus so often and so much on the details of the variants (differences) that we are in danger of losing sight of a fundamental fact in this case especially.

This agreement has a cause: LITERARY DEPENDANCE. This loaded term needs to be teased out here for our immediate purpose. Literary dependance means, specifically, that two texts are directly dependant, either one on the other, or both stemming from a common original, THROUGH WRITTEN TRANSMISSION.

This phenomena of detailed and exacting word-for-word content and order means exactly this, one or both texts copied a written source, perhaps even each other.

To put this in perspective, imagine two high-school students handing in exam papers only a page long, but each showed the same amount of agreement, both in exact wording, and in the very order of the words and phrases too. Both students would naturally be taken aside, and one or the other would be certainly found guilty of cheating. Until either student was found guilty, they would both remain under extreme suspicion, their records tainted.

Why have we belaboured the point here? Of course these texts show literary interdependance: They are copies of a written Gospel some 300 years older, namely the book of John.

But that is the point in this case: Some have claimed, without any credible evidence, that this particular passage is not a written tradition at all, but an oral one, inserted just prior to the manufacture of Codex Bezae (D).

Our observation is not an attempt to disprove a distant oral source: that is far too ambitious and cannot be sustained on the basis of the data here. The goal is much more modest:

We add as a 5th axiom:

(5) Both texts probably derive from a much earlier written source, just like the other portions of John do in these two texts.

The importance of this, and the reason for revisiting it is not just because of various theories of oral sources, (for indeed there must have been oral sources when the Gospels were composed): But that this observation puts reasonable constraints on timing and distance for any theory of "insertion of an oral tradition".

But this also is too ambitious an investigation at this early juncture.

What we want from this axiom is again something far more modest. It simply means we will be looking for mechanical explanations for possible error that are based upon the same copying process that accounts for such errors in other portions of John.

To our collections of axioms and observations, we can add a few more. These flow naturally from supplementary, non-controversial, historical information:

We note that St. Jerome observed (based on his own research prior to 392 A.D.) that "many copies, both Greek and Latin", contained the passage in its normal place in John's Gospel.

Jerome has become a rather colourful character, as he has been studied more thoroughly in recent times. Some have doubted his honesty and accuracy regarding many text-critical statements, and he has been accused of everything from conspiracy with the new Roman/Christian regime (c. 400 A.D.) to Judaizing the Greek O.T.

However, many things Jerome has said have since been independantly verified: For instance, Jerome was the only one to mention a second ending for Mark's Gospel, which was assumed by skeptics to be sheer fantasy, until the remarkable Codex W was discovered. This manuscript has the secondary ending just as Jerome described, corroborating Jerome's unique claims.

In a very similar manner, the Codex Bezae (D) itself corroborates Jerome's testimony that in his time "many copies, both Greek and Latin" were circulating prior to the production of his famous Latin Vulgate NT (c. 392 A.D.).

This instance will be the SECOND time a previously unique text-critical observation by Jerome will have been corroborated by independant evidence of the most clinching kind.

Our 6th axiom/observation will be this:

(6) Since Codex Bezae sports a pre-Vulgate text, Codex Bezae corroborates the testimony of Jerome that there were "many copies, both Greek and Latin" containing the Pericope de Adultera circulating c. 380 A.D.

(7) Codex D and Jerome together then, fix the date of any possible insertion of the PA to pre- 380 A.D., in order for such MSS to have been circulating in plentitude in both Greek and Latin at the time Jerome observes.

(8) This observation (7) combined with (5), namely the literary interdependance between the Byz-text and D in turn fixes a Byzantine or pre-Byzantine text form containing the PA to the same approximate period, namely the mid to late 4th century at the latest.

(9) From this we conclude that the position of the PA in John at 7:52 is contemporaneous with D, and the particular Byzantine text-type form containing the PA may also be contemporaneous with D.

Since these two sources, the Byzantine text-type and Codex D are the two earliest direct witnesses to both the position and the text of the PA, the nature of their interdependance must be investigated next, and this must be carried out on internal and indirect evidence of an independant nature.

As we said, we must now turn to the question of the interdependance of Maj and D.

Was the Maj-text derived from D? or vise versa? or were the two independantly descended from a common original?

As it turns out, both texts appear to have secondary features, indicating a subsequent independant history after the parting of the ways. The most reasonable model is one of a common ancestor, and independant lines of transmission.

How can we arrive with confidence at this conclusion? The answer depends upon intelligent sorting of the variants between the two texts.

The first separation into categories will be based on the question of whether the mechanism of the variation was deliberate or accidental.

With deliberate changes, it can be difficult to nail down the directionality of the change. This is true because an editor can arbitrarily alter a reading for a variety of reasons. He may make a text more readable, or update its language. Or he may attempt to "improve the style" of expressions according to a preconceived notion of correctness. Or he may choose to mimick a spelling or grammatical morphology in order to make a text appear "classical" or more ancient.

Thus deliberate changes can have the appearance of reversing their direction, as when an archaism is adopted and inserted into a text, or a given dialect of Greek is adopted. These tendencies became more frequent in the very period when both Codex D was produced and when the Majority text-type was finalized and stabilized.

Gratefully, accidental errors are more easily understood, and their directionality is more confidently determined where it exists. Thus it is to accidental errors that we turn for more reliable clues as to the relative dependancy of readings and text-types.

Another benefit of sorting variants by probable cause is that this can also separate layers of change, and group readings by the source and date of their occurance.

For instance, because different groups of readings can be attributed to different causes, they can be further separated from each other and the likely original text.


Verb Form Alterations:

Turning to the text at hand, we can separate out the grammatical changes having to do with verb forms. These are mostly not accidental, but reflect the activity of at least one (and probably only one) editor, intervening in the history subsequent to when the two texts parted ways.

Thus the verb variants in 7:53, 8:2, 8:3, 8:4, 8:5 (2x), 8:4/6, 8:9, are deliberate. We can leave the direction open while at the same time noting that a single editor is probably responsible for all these.


Particles, Prepositions, Cases

Codex D shows a propensity for adding verb prefixes, as at 8:4, 8:6/8, 8:11.
Codex D also eliminates prepositions and pronouns, such as at 8:4, 8:10, and even Jesus's name at 8:9, 8:11. These again are not accidental, but involve idiomatic alterations of word-order in the immediate vicinity, as in 8:10, 8:11.


Smoothings, Simplifications

Many expressions are omitted by Codex D, which would appear superfluous to an editor looking to simplify and smooth the story, e.g.,
8:6 (2x) Τοῦτο δὲ ἔλεγον..., μὴ προσποιούμενος.
8:9 (2x), ἀκούσαντες, καὶ..., εἷς καθ' εἷς,
8:10 (2x). καὶ μηδένα θεασάμενος πλὴν..., ἐκεῖνοι οἱ κατήγοροί σου;

In other cases, simplifications in D also become explanatory, such as the truncation and migration of a clause from 8:6 to 8:4 (ἵνα ἔχωσιν κατηγορίαν αὐτοῦ),
and the substitution at 8:11 (ὔπαγε ἀπὸ τοῦ νῦν).


Late Insertions and Doubtful Explanations

Some of these D readings are clearly secondary, such as at:
8:4 (οἱ ἱερεῖς - "the priests"),
8:9a, (ἕκαστος δὲ τῶν Ἰουδαίων - "all of the Jews") and
8:3 (ἐπὶ ἁμαρτίᾳ - "taken in sin") [!], although certainly ancient (cf. Papias in Euseb. E.H., c. 100 A.D.!).

Cumulatively, the nature of these deliberate variants indicate a relatively recent editor has worked on Codex D's text (after the Byz-text has branched off from their common ancestor).


Ambiguous Variants

This helps to fix the probability of the direction of other deliberate edits, like
8:9 (ὥστε πάντας ἐξελθεῖν),
8:11 (κακείνη),
and the difficult choices in the various verb-forms.

Again, some word-order changes would be superfluous for a native Greek speaker, but would plainly assist a Latin reader:
8:5 (Μωϋσῆς) Subject precedes verb and modifying phrase,
8:7 (βαλέτω λίθον) Object follows verb, & def. art. omitted.



Secondary Features of the Byzantine Text

As usual, the Byzantine text shows little deliberate editing, and has preserved many phrases lost by both deliberate editing and accidental causes. However, if there is one area where the Byz-text appears secondary, it is probably in the uniform verb-forms, and standardized spelling, bespeaking the later copying period. This however has little effect on the content of the text, or translation into other languages. What we get is a surprisingly clear view of the original text, albeit through the rose-coloured goggles of the later standardized form of Byzantine Greek.


Conclusion

Ultimately, Codex D's text in this passage appears secondary in many places, with the Majority text being much more credible and convincing in preserving the original form of the story.

Codex D here has all the look and feel of a later "lectionary" version of the story, made for easy comprehension and didactic purposes.

Both texts have the more natural "continous-text" morphology at the beginning, which tends to suggest that this form of 7:53 reaches back to a common ancestor penned before the lectionary-style pericopizing has taken place.

Both texts then contain very old readings, but the Majority text seems to have preserved the original text better, with less omissions and more complex and difficult forms of many expressions.

Although we are comparing two roughly contemporary texts, we need also to keep in mind the difference in nature and kind between these two.

Codex Bezae is after all, only a single manuscript, with a fixed probable date of manufacture. As such, it is an excellent witness to a text-type and content circulating in the late 4th to early 5th century. It delineates a limit, closes a window as to the latest possible time in which the passage (7:53-8:11) could have been added to John's Gospel and introduced into the copying stream, if indeed it was an addition. Its evidence is particularly strong because of its agreement and corroboration of the testimony of early fathers like Jerome, Ambrose, and Augustine regarding its existance and circulation. But its textual readings, particularly its singular, unique readings, must be regarded with extreme caution and suspicion.

The Traditional Text (the reading of the majority of MSS) is on the other hand, an 'averaged' text-type represented by some 2,400 later copies ranging from the 8th to the 14th century. As such it is a picture of the entire surviving manuscript stream of transmission. It does not contain "singular readings" or unique errors in the same sense as any individual manuscript. Any unique readings so strongly supported must reach back to the origin of this text-type, and quite likely to a time earlier than the date of Codex Bezae.

Since these two independant texts are the main sources for readings, and other secondary sources (such as 'versions' - early translations) are usually viewed as later in date and dependant upon Greek sources, its the contrast between these two texts which reveals any accumulated errors in either branch of transmission. Our main methodology is simple comparison.

On this basis, mistakes can be identified in both branches, but confidence in identification is far stronger in dealing with the singular reading of an individual manuscript. Such errors are far more concrete and convincing when found in Codex Bezae than when suspected to exist in the Majority text-type, represented by such an array of independant witnesses. Master-copies are checked more carefully and more frequently, so mistakes othen get found even when they get into the copying stream.


Types of Error

Most copying is done by eye, rather than oral dictation, and most mistakes arise due to reading or comprehension errors.

(1) Spelling Errors: One type of error, mis-spellings, although common, will not concern us here, since it does not affect the meaning, and is more difficult to judge in early texts where spelling varies greatly.

(2) Mis-readings: these mistakes are much more rare, and only one serious candidate appears in our two texts, the variant found in Bezae/Maj at 8:5: (σὺ δὲ νῦν / σὺ οὖν ), which seems to be a misreading/conjectural emendation on the part of Bezae.

(3) Insertion of Marginal Notes: here marginal notes are mistaken for corrections and pulled into the text. Mistaken cases are much rarer than suggested by the literature, with only a few clear examples known. Nonetheless it is possible, and may explain occasional longer readings, especially if they can be traced back to the earliest period. Such a mechanism might account for the addition of (μὴ προσποιούμενος) at 8:6 in the Maj. text, although this would have to concede a very early introduction, with the implication that here the Maj. text itself goes back further than some critics may be willing to admit.

The other three main types of accidental errors, dittography (repetitions), homoeoteleuton (omissions), Word-order Reversals (W.O.R., - words, phrases, sections copied in wrong order), all stem from the same basic error, what Colwell describes as "slips of the eye".

(4) Dittography: (simple) repetitions are easy to spot and easily corrected while ink is still wet, or simply by crossing them out later. They are also caught in re-copying, and rarely get past a single generation of copying. Dittography can usually be blamed on the copyist responsible for the manuscript in question. In our case, no examples appear in Codex Bezae, and none will likely be found in the Majority text itself.

(5) homoeoteleuton (omissions caused by a skip to a similar ending of line or phrase, & homoeoarcton, similar beginning of line) are far more frequent for two reasons: they are difficult to spot when the text still makes good sense, and slips of the eye slip ahead about 4 to 10 times as often as backward. This seems to be a combination of the mechanical forward copying action, and the psychological habit to continually look ahead.

(6) Word-Order Reversals: these are actually a variation of (5), but with the omission partly corrected. The scribe skips a word or phrase, copies something further ahead, then notices the error and doubles back, restoring the dropped text after the mis-copied word or phrase. This was an easy solution in Greek, because it is a highly inflected language, less dependant upon word-order. Often word order has no real effect on the meaning or sense of the text in Greek. This enabled the 'easy fix' of simply adding in the missing words a bit later and carrying on, without having to erase the word skipped to.

When a single word is skipped, and not replaced until a distance later, it is rather obvious which order is original.

Such is the case for the reading of Bezae, where the copyist's eye skipped ahead to (ἐλάλησεν) in 8:12, then lazily copied (αὐτοῖς ὁ Ἰησοῦς) and continued, without bothering to erase and rewrite (ἐλάλησεν).
Again, we see the editor's eye in verse 8:14 skipping ahead to (μου), after being distracted with adjusting the preceeding verb ending - (ἀληθείνη). He doesn't even erase this short word, but simply follows with the interceding (ἐστιν ἡ μαρτυρεία).

Slip of the eye is again the original cause of the insertion of (πρὸς αὐτὸν) at the end of 8:2 in Bezae. His eye skips ahead to the same phrase in 8:3, probably appearing at the beginning of a line in his master-copy. He copies it, then catches himself, lazily continuing from the beginning of 8:3, but now covering his booboo by paraphrasing the second instance of (πρὸς αὐτὸν). Nothing could be lamer or more transparent.

The originator of the Bezae-text is plainly reluctant or unable to erase his blunders.

When a single word is skipped, and immediately re-inserted it is difficult to tell which order was original. But independant factors can come to our aid here: for instance, the cases at
8:5 (Μωϋσῆς) - Subject precedes verb and modifying phrase,
8:7 (βαλέτω λίθον) - Object follows verb, & def. art. omitted.
are easily accounted for as Latinisms by the bilingual scribe.

All these examples of singular blunders and adjustments seem to fall on the side of the editor of the Bezae text, if not the copyist responsible for the MS itself. Some seem to originate in an even earlier master-copy.

Again, the preponderance of the evidence favours the purity and primitiveness of the Majority text, rather than that of Bezae. But if there was time for all these accumulated alterations, even multiple layers in the transmission line responsible for Bezae, this suggests that the text, in the Majority-form, is far older than the text of Bezae itself.

We can reasonably say that on the basis of internal grammatical evidence alone, and scribal habits, that the Majority text originated at a time significantly earlier than the Bezae text and the Bezae MS itself. This viewpoint is in harmony with the corroboration via quotations of the PA found in Ambrose, the Didaskalia, the Apostolic Constitutions.

A tenth Axiom might then be:

(10) The Majority text of the PA is more primitive than the developed and corrupted text found in Codex Bezae which was circulating in the late 4th or early 5th century. The Majority text of the PA and its position in John stems from the 4th century or earlier.