Friday, March 25, 2011

The Basis for a Positive h.t. Identification

The following is culled and developed from a recent discussion at Fighting Fundamental Forums, which resulted in a useful exchange and explanation of the methodology and procedure for identifying probable homoeoteleuton, homoeoarcton and other similar accidental errors.

Q:   Aren't the so-called errors simply presumptions made by those who have a preference for the Textus Receptus (TR) and Byzantine manuscripts?

 The question speaks as if the "deductions" were simply opinions or fashions, and that the decisions about the nature of the variants were made on the basis of some predisposition or preference for the TR.

This is certainly not the case for the 75 homoeoteleuton cases examined in the Aleph/B text. Here's the simple reason why:

You can't turn these 75 homoeoteleuton cases inside-out on the basis of another opinion or some other preference, and have any kind of coherent explanation for the omissions/additions, or equally plausible explanations for the other variants. One of the most basic 'canons' of TC is a preference for the reading that explains the arising of the other variants.   That is, the explanation and temporal sequence is preferred which accounts best for the physical features of the whole Variation Unit.

These cases are judged to be homoeoteleuton cases on the basis of their physical features alone, not predetermined preferences for, or predisposition to favour a given view or text-type.

Homoeoteleuton cases in the Byzantine text-type are also in the same boat.

No preferences or opinions can change their physical features.  Its these physical features that class them as possible homoeoteleuton errors, and nothing else.

Q:   What is the standard text which makes these cases errors?  For an error to be deducted, a strong standard needs to be established.  Haven't you erected the Traditional Text as the standard against which you judge the Alexandrian?

If this really were the case, then one could claim a methodological weakness, or at least a presumption.  But that is not how textual criticism is legitimately being done here.

(1) Textual Evidence: Legitimate Use

The texts themselves can't speak or recommend readings.

The textual evidence (External Evidence) can only be used legitimately in certain ways: Textual witnesses are used to establish areas of variation (Variation Units), and the variants themselves.

The manuscripts (MSS), 'versions" - early translations (ETs), early Christian writers (ECWs), these witnesses are collated in order to find the boundaries where a variation exists, and list support for each variant reading.  The apparatus so generated establishes each reading's geographical and temporal extent, and its earliest appearance. That is pretty much all textual evidence can do.

Next it is up to textual critics to evaluate and interpret these variants, using rational and impartial scientific methods and principles, to create a plausible and probable history that can explain how the variants arose, and what the original text was.

Possible homoeoteleuton errors are identified by using a combination of textual evidence and internal evidence, using the following methodologies:

(2) Internal Evidence: Legitimate Use

The Internal Evidence is a different kind of evidence entirely. It breaks down into two basic categories:

(a) Transcriptional Evidence (Scribal Habits):  Since we look at the lost originals through the lens of copyists, we must understand thoroughly how the coloring of this lens affects the text.   This first of all comes from examinations of individual manuscripts, the work of the scribes themselves.

(1)  Singular Readings:   For maximum reliability of findings, singular readings (unique readings not found in other manuscripts) are used to evaluate copyists.   These have the highest probability of being accidental errors or quirky edits by the actual copyist of the manuscript.

(2)  Accidental Readings:   Additionally, singular readings are sorted into probable accidents (where the unique reading makes no sense, or less sense, and where an accident best explains the alternatives), and possible deliberate edits (readings which make sense, have theological or historical value, and cannot be explained as accidents).

By concentrating on probable errors, we can identify patterns and probable causes, as well as general habits and tendencies of copyists.  This in turn helps to identify other variants which don't have the appearance or probability of being mistakes.

(3)  Transcriptional Probabilities:   From examining hundreds of MSS and the copying habits displayed therein, the features of and general probabilities for various types of errors (and scribal modifications) are established.

Please note: The knowledge of transcriptional errors is not established by comparing text-types or groups to one another. 
The features and probabilites of transcriptional errors are instead established by collecting unambiguous instances of each type of error in individual manuscripts, not those in text-types or groups.  By this approach, the common features that all instances share can be noted, the common causes of the errors understood, and reliable statistics generated.

The effect of scribal habits upon text-types or groups is based on the accumulation of individual scribal habits in the process of transmission.  Thus our knowledge of groups or text-types can only come from understanding the effects of the individual copyists and editors on the text.   We must first study individual copyists to understand text-types. 

This is why individual manuscripts are studied first, and the most general habits and tendencies are established, to build a solid foundation for the study of transmission and text-types.

This knowledge is carefully built up before any further application of Transcriptional Evidence can be applied to Variation Units or text-types.
Important examples of such preliminary studies are given below:
E.C. Colwell (1969): Haplography - & P45, P66, P75
Jongkind (2005): א - tests Singular Readings Method!
J. Hernandez (2006): Errors of א in Rev - singular OMs
J. Royse (2008): Scribal Habits - P45,46,47,66,72,75
J. Royse (2008) homoeoteleuton - singular omissions
From such studies, more reliable observations, and solid canons can be established, as in the following discussions:
H. Gamble (1977): Interpolation - Identifying Marks
L. Haines (2008): Scribal Habits - 'Shorter Reading'?
J.Royse (2008) Shorter Reading? - & Griesbach

When a given Variation Unit is examined, the text-types involved and their textual support are not relevant for discovering and evaluating generalized transcriptional features.

Remember that Transcriptional Evidence has to do with the kinds of errors and changes that ALL copyists are vulnerable to, independent of time-period, location, or text-type. Text-types are not internal evidence.
When we identify a Variation Unit as a possible instance of homoeoteleuton, we don't do this on the basis of text-type, nor by comparing a given variant inside the Unit to another variant as if one were a standard and the other a mistake.

We identify whole Variation Units as homoeoteleuton instances by the features that the whole Variation Unit presents, regardless of text-type or opinion regarding individual variants within the Variation Unit.

It is not the variants that are identified as homoeoteleuton, but the entire Variation Unit itself which is classified as possible homoeoteleuton. This is done on the basis of its intrinsic features, which are independent of text-type or manuscript support, and its not done on the basis of favoring a specific text-type within the Variation Unit.

 The Classification 'homoeoteleuton' (h.t.) belongs to the Variation Unit, not the individual variants in it.  The Classification of 'text-type' belongs to individual readings, not the Variation Unit, which is an over-arching structure involving all text-types and groups.

Variation Units identified as possible or probable homoeoteleuton cases occur in all text-types, and in all manuscripts and witnesses. They are not 'text-type' specific, and they are not defined or determined by choosing any text-type as a standard. That is just nonsense.

All the probable homoeoteleuton cases we have identified have been identified on the basis of their own intrinsic features as shown, and not on the basis of their agreement or disagreement with the Textus Receptus or the UBS text (or any other text). The only subsequent process imposed upon the full group of homoeoteleuton cases was this:

On certain occasions we chose to talk about those cases mistakenly adopted by the UBS text, and this necessarily involved selecting those cases as a sub-set of the complete list.   

We can certainly provide other lists of probable homoeoteleuton cases in the TR, or the Western Text, or the critical texts of Griesbach, Lachmann, Tregelles, or Tischendorf, or Hodges/Farstad if you like. But again all of these examples will be established by their actual features, not by evaluating textual support or favoring text-types.

We can group homoeoteleuton errors according to text-type or geographical extent only AFTER we have already found them.

(b) Intrinsic Probability (Author's Intent):  This is another category of evidence discussed by Hort and others.  It refers to what it was that the author was most likely to have written.

Intrinsic Probability involves grammatical evidence (vocabulary and syntax), literary evidence (content and structure), and sometimes theological evidence (what the author believed or knew, based on what he shows elsewhere and what is historically known or plausible) But AGAIN it is not based on text-types or favoring one form or source over another. It is also a kind of INTERNAL evidence, not TEXTUAL per se.   But that is another subject.

Q: seems that Nazaroo's text-critical philosophy is a thoroughgoing eclecticism. If all that external evidence can do is give you various readings, with geneaological & geographical aspects concerning those readings, then you fall very much in line with J. K. Elliott.

This is really a basic misunderstanding.  We are not advocating any particular philosophy or method of Textual Criticism.  We can find flaws with all of them.    The point is rather that certain specific tasks in TC require solid techniques.  These do not then become the 'only method', or a generalized philosophy.  They remain limited specialized techniques, only justifiable with clear and specific applications in view.

Q:    In order to identify the cases in question as specific examples of probable homoeoteleuton, with resultant dropout of text, isn't an actual exemplar containing the "missing" text presumed to exist or have existed?

Again the question assumes we are creating hypothetical entities, when in actual fact we are limiting ourselves to previously documented textual variants found within Variation Units.   It is well-recognized that almost all variants are 'old', and stem from the early centuries (1st - 3rd cent. A.D.)
The variants themselves are already attested, although in some cases by later manuscripts, or appear as minority readings.   We don't postulate any new archetypes or lost exemplars.  We simply stick to known units in the standard apparatus and give all well-attested variants the possibility of being an original reading.
That is the natural starting point for all textual criticism.  We assume that we don't know the original reading with absolute certainty, and evaluate variants based on textual and transcriptional evidences we find.

(1) Since we start with textual evidence, compiling Variation Units from variants actually found among real manuscripts, we don't need to conjecture any texts out of thin air, and in fact we reject entirely any conjectures which lack actual manuscript support. We limit ourselves to variants supported by good textual evidence.

(2) The only 'conjecture' or 'presumption' involved is the openmindedness to consider any well-attested variants as possible cases, and check to see if the Variation Units involved have the physical features required.

(3) The evaluation is conducted on the basis of the most generalized and well-known scribal habits and tendencies (the most reliable kind), and is not dependent upon text-types or peculiar local practices, or temporary trends found in particular eras, such as "Alexandrian editng techniques" or 'Western tendencies of conflation'.

(3) At this stage, no preference for text-types or geographical/temporal witnesses need be considered, nor should it be. It would be far more 'presumptive' in that sense if we pre-selected and favoured text-types like the Alexandrian as more probably "original".

(4) Remember that we are investigating all text-types, and all eras, and the only focus or 'bias' will be our attention upon the earliest and most reliable textual evidences. The Variation Units we use are composed and filled out utilizing data from all text-types, and will often naturally group the evidence largely by text-type in many cases. But that is not a free choice or preference on the part of those collating manuscripts for the apparatus. That is just the way the evidence naturally falls, and organizes itself.

(5) We ourselves have no hesitation in using for instance the good data collected and organized in the UBS4 apparatus, or that found in any other good critical apparatus, like Tregelles, Tischendorf, or von Soden.

(6) It would be far more biased to focus only on the errors of a certain preselected text-type or group of manuscripts, and thereby imply or give the impression that other text-types were superior, or focus on a supposed superior text-type, and presume others were inferior.

(7) But what we really have done is to look at all text-types and witnesses, and to use independent data and evaluations of scribal habits, to categorize Variation Units found in all text-types and manuscripts.

Again we recap, that we don't use conjectural reconstructions of non-existant texts, but instead restrict ourselves to known and well-attested textual variants supported by real manuscripts, versions and text-types.

Nor do we begin with presumptions or preferences for text-types or manuscripts. We deliberately put those aside and appeal to independent data on scribal habits and errors, culled first of all from hard textual evidence, such as the singular readings and corrections found in individual manuscripts of all types.

We apply already accumulated knowledge about scribal errors, knowledge which has stood the test of time, passed peer review, and has been accepted by textual critics of all viewpoints. We stick to the most well-understood, well-known and reliable data on scribal habits, such as the mechanisms of omission and dittography due to homoeoteleuton and homoeoarcton features of the texts.

We have investigated homoeoteleuton in both the reconstructed archetypes of manuscripts like Aleph/B, and actual singular errors found in individual manuscripts.

It should be understood that WE did not reconstruct the archetypes of Aleph/B, but that other textual critics (Lachmann, Tregelles, Tischendorf, Hort, Nestle, Aland etc.) reconstructed Aleph/B on the direct basis of "agreement in text", independent of and regardless of the mechanisms for changes.

We do not dispute at all that the UBS4 text substantially represents the archetype of Aleph/B, and that textual critics have done a relatively good job at reconstructing that archetype. We don't doubt the essential genealogical tree or the fact that the UBS text is an Alexandrian text which circulated earlier than either Aleph or B.

The point is, WE did not create or conjecture an ancestor for Aleph/B. It definitely exists and conforms to the UBS text. This is not in serious dispute by any textual critic, and this reconstructed text is NOT considered "conjectural" or a mere "presumption".

But nor is the existance of the 2nd century Western text in serious dispute, or the 4th century Byzantine, or the 3rd century Old Latin or the 4th century Vulgate. These texts are as real as the manuscripts that support them, and they all reach back into the 2nd century.

The texts and readings we are using and the Variation Units are well-known, documented and accepted by textual critics of all persuasions, and are all found in the UBS4 Apparatus.

We avoid entirely any conjectural texts proposed by others, or any emendations of our own to the Variation Units.

 Q:    Wouldn't your resultant or corrected text be a lengthier, "fuller" text than that of any Critical Text (CT), or probably even than the Traditional Text (TT)?

This is not actually true, and logically incorrect.

Any text corrected from a subset of probable homoeoteleuton errors would only put back some of the omissions found among Variation Units, and even if these were removed from texts and apparatus, a substantial part of the basic differences between the UBS4 text and the Majority Text would remain: Some 120 omissions/additions would still be in dispute.

The text created from my data would not make a longer text than the Majority Text.


Wednesday, March 23, 2011

Codex W: Matt. 4:21-22 - h.t./h.a. Combo

Here is the first example Sanders gave as a first-hand homoeoteleuton blunder in Codex W:

Matt. 4:21-22
Click to Enlarge: Backbutton to return

The standard text in the master copy probably ran like this:
[πετρον] και ανδρεαν τον αδελφον αυτου βαλλοντας αμφιβληστρον εις
την θαλασσαν ησαν γαρ αλιεις * και λεγει αυτοις δευτε οπισω μου
και ποιησω υμας αλιεις ανθρωπων οι δε ευθεως αφεντες τα δικτυα
ηκολουθησαν αυτω   και προβας εκειθεν ειδεν αλλους  δυο αδελφους
ιακωβον τον  του ζεβεδαιου   και ιωαννην  τον αδελφον αυτου  εν τω
πλοιω μετα ζεβεδαιου του πατρος αυτων καταρτιζοντας τα δικτυα 
αυτων και εκαλεσεν αυτους οι δε ευθεως αφεντες το πλοιον και τον PAR (πατερα)
αυτων ηκολουθησαν αυτω και περιηγεν ολην την γαλιλαιαν ο IS (ιησους)
διδασκων εν ταις συναγωγαις αυτων ...
This exemplar probably had either 50 or perhaps 25 characters per line.

One can see the comedy of errors unfolding.  The scribe looks from his copy to the master, searching for the line-end "τα δικτυα" but slips 3 lines lower by an homoeoteleuton (similar line end).  He now writes "αυτων", but is distracted once more: he again looks back to his master-copy for the line beginning with "αυτων", and now commits the complimentary line-skip, via homoeoarcton (similar beginning).   Perhaps a younger monk had pestered him with some question, and the exchange was enough to cause the double-fumble.   It was some kind of Kodak moment, or it was just a bad-hair day for the scribe of W.

On the fly, the scribe of W expands some of the lesser known contractions (Nomina Sacra) such as anthropos (man),  probably for readability.  His text now looks like a swiss-cheese:
και ανδρεαν τον αδελφον αυτου 
βαλλοντας αμφιβληστρον εις την 
θαλασσαν ησαν γαρ αλιεις  και λεγει 
αυτοις δευτε οπισω μου και ποιη-
σω υμας αλιεις ανθρωπων οι δε ευ
θεως αφεντες τα δικτυα.../...αυτων  
 .../... ηκολουθησαν αυτω 
και περιηγεν ολην την γαλιλαιαν ο  
IS  διδασκων εν ταις συναγωγαις αυ-
των ...
Needless to say, other scribes who may have occasionally used Codex W as an exemplar would have probably rolled their eyes at this zinger, and quietly ignored this reading in favor of any other handy copy.


Tuesday, March 22, 2011

Codex W: h.t. singular list - from Sanders

Sanders (1912) the editor of the facsimile photos of Codex W, in his detailed account of the text gives a list of cases where he is certain that the reading is a result of homoeoteleuton (like endings) errors by the copyist of the manuscript itself (i.e., not copied from the exemplar).
"Omissions by carelessness or because of like endings, which can be definitely assigned to our scribe:
Matt.   4:21-22    [h.t./h.a. combo]
Matt.  15:18         [h.t. ]
Matt.  16:2-3
Mark   6:23
Mark   7:13
Mark  11:15
Luke    8:31
Luke  15:19
Luke  15:24
Luke  17:35   [h.t.]  (A most interesting case; the scribe himself corrected his mistake after writing 3 words)
John   5:11-12 (perhaps from a parent [h.t., replacement quire])
John   21:4.

The regular scribe is relatively free from such errors, if we consider the rapid style of his script and the length of the MS ... Only in the case of those characteristics which run through the MS without change can we assume our scribe responsible. Among these I venture to draw deductions only from those pronounced characteristics which are rare to non-existent in other Biblical MSS. In this class we may enumerate: tendency towards aspirated consonants, αλλα before all vowels, κα for και , and the decided tendency towards Attic or other old forms [of Greek]. The most of these find their nearest parallels in the early papyri and the oldest uncials of Egyptian origin, thus confirming the supposed Egyptian origin and suggesting an early date....
On the preceding page I have listed 12 cases of longer omissions by our scribe; 9 of these were due to like endings [homoeoteleuton] and 3 to like beginnings [homoeoarcton] of successive phrases.  We may assume that these omissions would more easily occur if the parallel parts stood at the beginnings or ends of neighboring lines, and thus may draw inferences as to the length of line in the parent MS. 
The 3 omissions in Matthew are respectively 214, 44, and 36 letters long, indicating a line of either 20 or 40 letters in the parent.  As W has about 30, it seems quite certain that the parent did not agree [in line length]. 
In Mark the 3 omissions are of 36, 30, and 14 letters each.  These lengths might be consistent with a line length similar to W, but seem to point to a line of about half the length. 
In Luke, the lengths of the 4 omissions are 17, 27, 22, 65 letters, which would seem to suggest the short line attributed to the parent of Matthew.  
In John there are 2 omissions of this type; one comes between the first and second quires and is 69 letters long; the other, at  Jn 21:4 is 49 letters long.  We are also assisted by a repetition 139 letters long, covering 5 lines in the repeated form and 5 lines and 8 letters in its first form.  If we may unite the evidence of these three, the parent MS would seem to have had a line from 23 to 25 letters in length, i.e., again a different length, and so indicating a different parent.

The average amount of text written on a 16-page quire of the MS is 10 and a half pages of the Oxford 1880 edition.   Yet the first quire of John has about eleven and a half pages, and the last two full quires of Luke (crowded writing noted above, p. 7) contain nearly 12 pages of text each.  It is easy enough to explain large quires toward the end of a gospel, if crowding would have saved an extra small quire, but such is not the case here, as Luke ends on a four-page quire.  This looks like a hint that the parent MS had larger quires.  The larger first quire of John suggests a similar guess for that gospel as well.  We shall find this though confirmed in our study of the text affiliations later." 

Of those listed by Sanders, we have already looked at a few analyzed by Schmid ("Reassessing the Palaeography..", see the links above in our list).  We will have a closer look at a few others as well shortly.


Friday, March 18, 2011

Codex Bezae - Matt 23:34 h.t. blunder

Here is another rather transparent h.t. omission, either by the scribe of D or an ancestor:

Click to Enlarge

This could be a double-line at about 22 characters per line in the master-copy, or a single line skip at 42-44 cpl.  This area of text appears especially prone to h.t. errors, and does not disappoint.

Codex D:
Codex D: Matt 23:34  -  Click to Enlarge

Thursday, March 17, 2011

John 5:12 - Codex W: ...h.t. damage caused by a book-repair!

The chain by which our attention was drawn to this remarkable story is relatively long linkage in itself.  I noted Mr. Scrivener's post on codex W, in which Dr. Tim Finney in a comment drew attention to Urlich Schmid's article contained in Dr. Larry Hurtado's book, The Freer biblical manuscripts (2006).  There Mr. Schmid cites a discovery by Henry Sanders in volume 9, The NT MSS in the Freer Collection Pt 1: the Washington MS of the 4 Gospels (1912), p. 135-6, who in turn quotes Sir F. Kenyon in passing.    Well, thats enough name-dropping.  Lets cut to the chase.

First the essential facts.  Codex W (or the book of John therein) somehow lost the first quire from the book of John.  It had quires of 8 leaves (4 sheets) each, although a few leaves are missing, with the opposing leaves belonging to those sheets sown back in during a re-binding.   The current 1st quire of John is a replacement quire, by a different hand.    Although it appears older from deterioration, this could simply be because of poorer quality parchment.    The handwriting difference between the main book and the replacement quire is obvious:

Scribe W (main text in gospels):

General features
good slant, smooth straight lines of text, good spacing between lines,
most letters about the same size and on the line.
Occasional outdented letters same size as normal text.

Letters are elegant, but not fancy.
Phi (φ) - not oversized, Omicron (ο) - stout, often pear-shaped, Xi (ξ) - unique. Omega (ω) - angular, plain.  Psi (ψ) - Unusual, straight bar.  Epsilon (ε) - stout, substantive. Alpha (α) - often angular, but varies.  

note Xi (ξ) in bottom rightPsi (ψ) - Unusual, straight bar.
Phi (φ), overhanging KaiOmicron (ο) - stout, pear-shaped

Scribe JnQuire1: (replacement quire)

Oversize  Phi, K in midline,
less slant
Round Omega, Upright Alpha,
std Upsilon
Enlarged Outdent,
uneven crowded lines
Oversized phi, angular epsilon,
oval omicron.

Such examples establish that the scribe of Quire 1 (Jn) tried to imitate W but was not as skilled or consistent.  Certain features of his own style (e.g., enlarged phi) overrode his concern or ability to match the original. 

The Seam between Quire 1(scribe Q1) and Quire 2 (scribe W)

Now lets turn to what happened, as the new scribe tried to match up his quire:
- replacement page, Original John continues...

Here is Sanders' original description of the action:

"...we may thus with safety date the whole MS as not later than the early part of the 5th century [A.D.].  But does this also apply to the first quire of John?  Dr. Kenyon (op.cit.) thinks not and dates it tentatively in the 7th or 8th century, on the basis of the writing, which he classes as a Slavonic sloping uncial [script].   It seems impossible to separate so far the two parts of the MS, and fortunately we do not have to rely entirely on the comparison of styles of writing.  It is certain that this strange quire was written to fill a gap, to supply a lost quire.  On the last page of it the text is stretched and ends of lines left vacant after each sentence, so as to come out just even;  The three preceding pages were just as plainly crowded, an extra line even being added on each page.  It must be admitted that the writer was both inexperienced and had before him a copy quite different in size of page [layout].   Yet with all his care to make his quire come out even he omitted nearly a verse at the end.    This not only emphasizes the difference in form of the MSS from which and for which he was copying, but proves conclusively that one was not the parent of the other.   In other words, he was not copying an injured or wornout quire, but was striving to arrange in a quire a certain amount of text.   His task was to copy as far as the words  κραβατον σου και περιπατει of Jn 5:12, but he stopped with the same words in verse 5:11.    This might have been an omission in the parent text and be explained as due to 'like endings' [h.t.], but the fact that the omission falls exactly at the end of the quire seems sufficient proof that it was first made in copying this inserted quire." (Sanders, p. 135-6)
A few remarks are needed at this point.  Even though this took place in the replacing of a quire, right on a seam, it was still a homoeoteleuton error, an eye-skip by the copyist.  The difference is that it is unlikely to have happened without the 'opportunity' of the repair, and was unlikely to have been present in the original quire.

Sanders goes on to try to argue that the quire itself is older than Codex W[!]  The only 'evidence' he has of this is the condition of the replacement quire, but that can be better explained as the result of poorer quality vellum (improperly prepared), and different inks.   He claims to have seen an erased letter "a" above a slightly displaced quire number, but no other scholar has found any evidence of this.

We stop our discussion here, because our interest is only in this interesting case of yet another way a homoeoteleuton error can and did find its way into a surviving copy of the Gospels. 


Thursday, March 10, 2011

Matthew 10:37 - Codex B/D: Early h.t. Vanishes from UBS4!

This is a remarkable reading, not because it has any credibility as a reading (it is an obvious homoeoteleuton error, acknowledged by most critics and editors), but because of its strange history in the UBS text.

Here is the omission in Codex Vaticanus with its marginal correction:

Click to Enlarge: backbutton to return

It was duly and seriously noted in the apparatus of the UBS2 (1968, = NA26 etc.) and is indeed an important reading, because it appears to be a clear case of a previous error by a very early scribe, copied in independent lines to both Codex Vaticanus (B, and corrected in the margin there), and Codex Bezae (D).   That is, this was probably not committed coincidentally by both scribes, but by an ancient common ancestor, and duly copied (or cross-pollenated long before the 4th century).  Matt. 10:36-38:
  ..................... ..και εχθροι τ-
ου ανθρωπου οι οικιακοι αυτ-
ου ο φιλων πατερα η μητερα 
υπερ εμε ουκ εστιν μου αξιος 
και ο φιλων υιον η θυγατερα

υπερ εμε ουκ εστιν μου αξιος
και ος ου λαμβανει τον σταυ
ρον αυτου και ακολουθει οπ
ισω μου ουκ εστιν μου αξιος
 The similar ending extends to 1 1/2 lines, even at 23 characters per line.  This is an old error, from the 2nd or early 3rd century when papyrus copies of individual gospels carried only one or two columns per page.

The Master-Copy may have looked something like this:
Click to Enlarge

The UBS4 Fiasco:

But why did UBS4 (4th ed.  1993) remove it from the apparatus?    Surely not because it would damage the reputation of either Codex Vaticanus (B) or Codex Bezae.  Bezae is already well-known as a quirky and often unreliable text.   And if this is an error from a previous common ancestor, it cannot harm the reputation of the careful and skillful copyists of Vaticanus.

The answer is in the word "error".    Why?  Because it is a clear example of the careful copying of an ancient error by Codex Vaticanus.  An error of homoeoteleuton.  And it brings ALL such possible errors into sharp focus, especially those Variation Units where Vaticanus shares the omission with Codex Sinaiticus (א).

Because at least 75 of these probable h.t. errors, supported by (א/B) have been adopted as if they were original readings by the Hortian editors of the UBS text.

A large number of these readings are supported by earlier (2nd-3rd cent.) papyri, such as P66 and P75.  This was taken to mean the readings were original.  But the evidence can be more easily taken as proof of the obvious:  That most of these early h.t. errors do indeed go back to the 2nd century, but this fact merely reflects the poor state of the text and copying practice in that era, and not the purity of the 'Alexandrian' transmission stream behind (א/B).

So why delete from the apparatus this obviously significant h.t. error?  Because it, along with many other embarrassing cases, detracts from the reputation of the Alexandrian text-type, and the wisdom of following it.


Sunday, March 6, 2011

Codex Bezae - semi-singulars in Luke 24

These minority readings seem to be a perennial favorite among modern critics who prefer the world's shortest text, because they have garnered some support from the Old Latin manuscripts.  This support is rather suspect however, in the first place because the Latin was a translation from Greek, and so rather dependent upon the quality of early Greek MSS, and also secondary as a translation.   This extends further when we note (as it isn't very often in the apparatus) that the Old Latin MS "δ" is simply the Latin side of the very same manuscript "D" (Codex Bezae 5th cent.), i.e., it is really only a single witness, since the Greek and Latin have been consciously cross-harmonized to agree with each other.  Thus the citation, It-d  is still D

The original F.H.A. Scrivener describes the parallel Latin column in Bezae as follows:
"its own parallel Latin translation is too servilely accommodated to the Greek text to be regarded as an independent authority"  (Plain Introduction, vol.1,  p. 103)

That a few other Old Latin MSS have the same reading will not be surprising if Codex D was at one time used as a reference in the Latin West.

The three omissions of interest here are:

Luke 24:36, 24:40, and 24:51.

These have the unusual feature of the letter count being a multiple of 24-25 letters.  That is, the column-width of the master-copy from which the blunder was made was about 25 letters wide.

As it turns out, two out of three of these have homoeoteleuton features, and one appears to be a simple eye-skip (of which Codex Bezae sports many).

Click to enlarge: backbutton to return

The last example is particularly strong, with both homoeoteleuton and homoeoarcton features.

 Diagrams courtesy of Mr. Scrivener


Tuesday, March 1, 2011

Codex W: h.t. singulars - James Royse (cont.)

We continue our brief review of James Royse's article ("The Corrections in the Freer Gospels Codex") in the collection (book) edited by Dr. Hurtado, The Freer biblical manuscripts: fresh studies of an American treasure trove, (2006).
"Matt. 15:18-19:   The scribe made not quite the same leap as did א*,
33vid, bo-ms.  The latter witnesses leapt from της καρδιας εξερχεται of 15:18 to της καρδιας εξερχονται of 15:19 and proceeded with ονται διαλογισμοι κτλ.  The scribe of W, however, must have made the leap from εξερχεται of 15:18 to εξερχονται of 15:19 (abetted by the general similarity of the words, of course) and then proceeded with διαλογισμοι κτλ.  Thus W* wrote the singular verb of 15:18.  It is remarkable that the corrector simply shifted the verb rather than restoring the omitted words: clearly he also lost his way in moving back and forth between exemplar and copy." (- Royse, p. 203)
Codex W: Matt. 15:18-19 - h.t. singular (click to enlarge)

Codex W displays the result, an uncorrected homoeoteleuton error (accidental omission).