Five More Likely Copies of Printed Editions in the Liste

I’ve recently been doing some work on the ending of Romans that necessitated re-examining the data from the Teststellen in Romans. In doing so, I’ve identified five additional manuscripts that most likely derive, in whole or in part, from an early printed edition of the Greek New Testament, but that I had missed in my previous survey. Each of them has the doxology as found in Erasmus, with the Byzantine verse 24 followed by 16:25–27. 

While this is (apart from the presence of verse 24) the placement found in the earliest manuscripts, these manuscripts contain no other early non-majority text readings, making it very unlikely that their doxology placement is the result of influence from early manuscripts. It should be noted that the combination of 16:24 and the doxology at 16:25–27 is exceptionally rare in the tradition. All other manuscripts that have this combination have a large number of other distinctive readings elsewhere in Romans. 

GA 1931—XVI century

Contains the Pauline Corpus only. As in printed editions (but unlike the majority of Byzantine manuscripts), it has Hebrews at the end, after Philemon. More importantly, the manuscript contains no nomina sacra, and also uses the (Western) chapter divisions contained in printed editions. Apart from the doxology placement, the only non 1 or 1/2 reading in the Teststellen for Romans is εχομεν at TT_3/5:1, which is the reading of Erasmus. I would say this is almost certainly a copy of a printed edition. 

GA 1882—Liste XV, Omont XVI Century 

Contains the Pauline Corpus, with Hebrews after Philemon, followed by the Catholic Letters (as in early printed editions). While the *Liste* dates to the XV century, Omont, who catalogued the ms, dated it to the XVI century. It contains no nomina sacra. Although it does not have chapter divisions and does use ekthesis, the paragraph breaks (at least for the first few chapters of Romans) follow precisely those of Erasmus. The use of empty lines to mark the end of paragraphs is also typical of printed editions rather than Greek manuscripts. Apart from the doxology placement, there are only two readings in the Teststellen for Romans that are neither 1 or 1/2: 

TT_3/5:1 εχομεν (as above) 

TT_19/10:15 non-addition of των ευαγγελιζομενων ειρηνην

This latter could easily be accounted for by a simple copying error, though it is possible that an early edition also was lacking this phrase—we really do need a collation of early editions! It does not contain the Comma Johanneum (henceforth CJ) at 1 John 5:7, but then again, neither did Erasmus’s first two editions. I would say this is almost certainly copied, in whole or in part, from a printed edition. 

While the situation for the next three manuscripts is more complex, I still think it is more probable than not that they are in some way dependent on a printed edition. As all three of them come from Athos, it is likely that there is a relationship between them. 

GA 1104—1702

Acts, Paul (with Hebrews after Philemon), Catholic Letters. Contains nomina sacra. Minimal text divisions of any kind, but no obvious printed features. No CJ. However, it does contain the long Erasmian retroversion from the Latin at Acts 9:5. According the ECM, this addition is otherwise only found in very late marginal additions to GA 69 and 808. The reading in GA 69 is in a hand that is known to have compared the text to that of a printed edition of the Textus Receptus.

Apart from the doxology placement, the only non 1 or 1/2 reading in the *Teststellen* for Romans is εχομεν at TT_3/5:1, which is the reading of Erasmus. While this manuscript was clearly copied by a scribe more familiar with Greek ms conventions than those responsible for GA 1931 and 1882, the evidence still seems sufficient to conclude that this manuscript goes back, in whole or in part, to a printed edition. 

GA 1869—1688

Acts, Paul (with Hebrews after Philemon), Catholic Letters. No CJ, but with Erasmus’s addition at Acts 9:5. Apart from the doxology placement, the only non 1 or 1/2 reading in the Teststellen for Romans is εχομεν at TT_3/5:1, which is the reading of Erasmus. The situation for this manuscript is very similar to that of GA 1104 above. 

GA 1903—1636

Acts, Paul (Heb after Philemon), Catholics + Revelation. According to Text und Textwert for Revelation, this manuscript was “either partially copied from the TR or heavily influenced by it.” As it has ξυλου της ζωης in Rev 22.19, it is not simply a copy of the Textus Receptus. No CJ, but it does have Erasmus’s addition at Acts 9:5. Apart from the doxology placement, the only non 1 or 1/2 reading in the *Teststellen* for Romans is εχομεν at TT_3/5:1, which is the reading of Erasmus.  The situation for this manuscript is again very similar to the two preceding manuscripts.

On the Possibility that GA 1506 May Not Have Omitted Romans 16:1–24 after All

The placement of the Romans doxology in GA 1506 has long created a good deal of confusion. This manuscript is a copy of John of Damascus’s edited collection of extracts from John Chrysostom’s Homilies on Romans. The biblical text is written in red ink, often very faint, especially in the microfilm images, while the commentary text is written in a darker ink that is much easier to read. 

The NA28 apparatus (in the variation unit on the doxology) tells us: “add p. 14,23 et p. 15,33 (sed om. 16,1–24) 1506.” This is also the reading provided in the IGNTP transcription, in Text und Textwert, and in the literature more generally. Some sources (e.g. Aland’s 1979 article “Der Schluß und die ursprüngliche Gestalt des Römerbriefs”) also point out that a blank space of half a page has been left for 16:1–24. 

Today, however, as I took another look at this “blank space,” I noticed something odd. Most of the text visible on the page is the result of ink bleedthrough from the other side of the folio. As such, the text is reversed. Yet in the margin, outside the block of transfer, a marginal note was visible, one in which the letters were facing the right direction and which therefore cannot be the result of ink bleedthrough or transfer: 

In examining other manuscripts of John of Damascus’s commentary (GA 2110 and GA 0150, identified by David Flood as members of the same family as GA 1506), I discovered that not only did they contain a form of this same marginal note but that this note was a gloss on the name Ἑρμᾶν in 16:14. 

Proportional to the overall length of 16:1–24, the placement of the marginal note in GA 1506 is about where it should be as a comment on this name, as it is in the other manuscripts. In other words, while it may be difficult (or even impossible) to identify any of the biblical text under the heavy ink transfer and ink bleed through with the microfilm images currently available, it is very likely that GA 1506, as originally copied, contained 16:1–24, just as do the other manuscripts of this commentary tradition. 

Therefore, rather than categorizing this as an omission, it is likely better classified as a lacuna due to illegibility (as Swanson already did). In any case, the presence of this marginal note at the correct place and in the same red ink used elsewhere for the lemmata makes it unlikely either that the exemplar of this manuscript was missing these verses or that the red ink used for the lemmata had run out (possibilities suggested by David Flood, “New Readings in GA 1506: A Fresh Analysis of Old Microfilm Scans”). 

The simplest explanation seems to be that the scribe did indeed copy these verses, but that the ink has now faded beyond recovery without better images. It may also be relevant that the page in question (330v) seems to be the final page of its quire, which could have resulted in more severe fading, either before the initial binding or at some subsequent rebinding.  

We really do need better images of the Athos manuscripts! 

Longacre on the larger differences between the Great Isaiah Scroll and the MT

Thanks to the still-humming ETC blogroll, I noticed that Drew Longacre just posted a short blog post at his OTTC site on the similarity of the Great Isaiah Scroll (1QIsa-a) and the MT. This question became something of the flashpoint over the last year after the now famous Wes Huff interview on Joe Rogan’s podcast where he said it was word-for-word the same. 

Many counter-apologists jumped on this as misleading and it went round and round on the interweb (Wes later clarified in a video with John Meade). Apparently the issue is not dead since Bart Ehrman jumped into the fray this week. (I do not recommend you go down the YouTube rabbit hole on this one if you value your time). 

Drew gives some of this backstory in his post. But I want to ignore all that and higlight instead Drew’s 2013 DSD article that deserves much more attention. It addresses some of the largest differences between the Great Isaiah Scroll and the MT. Here is what he says about that:

Third, I would point out a series of difficult passages in the second half of the scroll that differ considerably from the Masoretic text, often with 1QIsa-a missing entire verses of the MT. Eugene Ulrich has argued that 1QIsa-a reflects an earlier version of Isaiah in these passages and that the MT has grown with additional verses being added. I have argued in an article that most of these passages were actually missing in the exemplar from which 1QIsa-a was copied, because it was damaged along the bottom edge. Thus, every column was physically damaged and missing one or more lines. You can see this, e.g., in how the scribe originally left blank lines for the missing verses. If correct, and I’m quite convinced it is, it means that most of the biggest differences between the Great Isaiah Scroll and the MT are easily explained as copying problems in the former.

I don’t think we’ve blogged about Drew’s DSD article here before. The full title is “Developmental Stage, Scribal Lapse, or Physical Defect? 1QIsaa’s Damaged Exemplar for Isaiah Chapters 34-66” and is online here (or in free pre-pub here). Here is the abstract:

The Great Isaiah Scroll from Qumran (1QIsaa) does not generally reflect a text form earlier than the Masoretic text. Instead, the convergence in 1QIsaa of patterns of spacing irregularities, literary and textual problems, and secondary supplementations, as well as a consistent pattern of distribution, are best explained on the basis of the hypothesis of an exemplar for chapters 34–66 with a damaged bottom edge. Upon reaching the defective edge in each column of his exemplar, the scribe dealt with any lacunose or illegible text in one of two ways before continuing with the unaffected text at the top of the subsequent exemplar column. Sometimes he left blank spaces in his new copy to be filled in with the correct text from other manuscripts at a later time. At other times he attempted full or partial reconstructions of the text based on whatever text remained legible in the damaged exemplar, memory, and contextual clues.

P.S. The fact that anyone is counting spelling differences between these two texts as somehow important is hilarious to me as a NT text critic. 

F.F. Bruce on 'Kept Pure in All Ages'

Yesterday, I was looking for one of F.F. Bruce's more obscure writings, and thankfully, I found that Rob Bradshaw had digitized it. While I was there, I stumbled across something more interesting though. In 1953, Bruce wrote a short overview-type article on textual criticism for The Christian Graduate [6.4 (Dec. 1953): 135-139].

The general thrust of the article is about what you would expect from Bruce: a careful overview of why we can have confidence in the text of the Scriptures.

What was interesting to me was his conclusion, which I share here:

According to F.F. Bruce (and a lot of us who are little more than dust in the shadow of F.F. Bruce), textual criticism is one of the means by which God's providential preservation of Scripture works.

You can read the whole article here.

Unrelated, if you've ever wondered what Bruce sounded like, you can hear his glorious Scottish accent in this lecture from 1977.

Jude 24: The Byzantine Text vs. the KJV

My interest is always piqued when a reading in the Byzantine text is not what’s translated in the KJV. Famous examples where the former lacks text the latter includes are Acts 8.37 and 1 John 5.7-8. In these cases, the bulk of Greek MSS is decidedly on the side of the Byz text against the KJV. But there are other, less notable cases, that go the other direction. In these cases, the KJV abandons the Byz text in favor of the less dominant reading.

Today I came across one in the best known section of Jude, the doxology (vv. 24-25). There we find a difference in our manuscripts as to who God is said to be able to “keep/protect blameless” (φυλάξαι ... ἀπταίστους) in v. 24. The NA28 prints the direct object as ὑμᾶς and gives, in a negative apparatus, the following alternates: ημας A vg-ms sy-ph ¦ αυτους P Byz. According to Tommy’s work on Jude and Klaus Wachtel’s on the Catholic Letters, the NA reading is attested in over 100 Greek MSS; ἡμᾶς is found in less than 10; and αὐτούς in over 300.

Jude 24 in 1611 KJV

The NA reading provides the smoother reading since the previous verses are addressed to the readers who have been warned against “the garment stained by flesh.” In light of this danger, it comes as a relief to be told that there is one who is able to “protect you from stumbling.” The Byz reading, which Wachtel includes in his list of “atypical majority readings,” is less obvious. Who could “they” be and why would Jude switch to address them? The answer comes from the Byz text of v. 23 where it is those who the readers are called on to save by snatching from the fire. God is able to keep them from stumbling. You could certainly argue that this is the lectio difficilior and some have (e.g., Tommy and A. J. Brown).

But, as I said, this is not the reading of the KJV even though it is the reading found in all of Erasmus’s editions. It is also not the reading of Tyndale’s editions, the Matthew Bible, the Geneva Bible or Luther’s 1522 Septembertestament for that matter. Luther and Tyndale particularly stand out here since both used Erasmus as their source and Erasmus has “they” in both his Greek and Latin columns. One waits until Stephanus 1550 to see ὑμᾶς in the margin where it is said to be supported by 5, 6, 2298, and a now lost MS (δ, ε, ι, ια). For his part, Beza has “you” (vos / ὑμᾶς) in his first and last editions which he explains as follows (in his first edition):

Vos ὑμᾶς. Ita lego ex Veteris interpretis and quatuor veterum codicum authoritate. Quod si malimus legere ἀυτοὺς dicemus pro ἑαυτοὺς positum, ita ut eodem recidat sententia. Neque enim videtur satis commode accipi posse pro ipsos vel eos, ut convertit Erasmus.

Which can roughly be translated as

Vos ὑμᾶς. Thus I read from the authority of the Old Interpreter [Jerome, I assume] and the four old codices. But if we prefer to read ἀυτοὺς, we will say that it is placed instead of ἑαυτοὺς, so that the sentence falls back on the same thing. For it does not seem like it can be taken quite conveniently for ipsos or eos, as Erasmus translates. 

Beza, of course, could not be the source for Luther and Tyndale. Neither could Stephanus. More likely is that they have here decided to follow the Vulgate instead of their main source in Erasmus. They were then followed by others who, in time, could point to Stephanus and Beza as further support.

One final note. Given the number of minuscules that support ὑμᾶς, the Byz edition of Robinson-Pierpont has listed it as an alternate in the margin as it does whenever the Byz text is split like this. Since the expressed method in that edition is to use internal evidence when the Byz reading is split, one must assume that the editors felt αὐτούς was finally preferable on those grounds rather than on external ones. Would that we had a commentary on such a decision!

In any case, this is an interesting example where the old KJV has the same text as our modern eclectic editions over and against the dominant Byzantine text.

Vulgate Manuscripts Missing the Last Two Chapters of Romans?

I’m attempting to work on a paper on the doxology in Romans, found as 16:25–27 in the THGNT and (in brackets) in the NA28. As part of my literature review, I spent some time surveying the treatment of this question in a number of commentaries. While reading Cranfield’s classic ICC, I was very surprised to come across the following statement, for which no explicit source was cited: 

(v) In the Vulgate MSS. 1648, 1792 and 2089 the text of Romans comprises only 1:1–14:23 followed immediately by 16:24–27. [p. 1:6]

My previous understanding had been that, despite a number of theories that have postulated a shorter version of Paul’s most famous letter, there were no extant manuscripts that actually contained such a shorter text. Further investigation seemed warranted. 

The first task was to identify which manuscripts Cranfield had in mind. The reference numbers he provided do not match anything found in Wordsworth-White’s large critical edition of the Vulgate, nor do they provide any notice of a Latin manuscript with a shorter text form. Furthermore, although they mention the textual problem of the doxology in a note, they do not cite any Latin manuscripts that have it at the end of our chapter 14, where it is found in the bulk of the Byzantine tradition. 

After a little digging, I determined that Cranfield’s source (which he had previously cited) was almost certainly T. W. Manson’s 1948 article, “St. Paul’s Letter to the Romans—And Others.” Manson states that: 

To the evidence of these tables we can add that of three MSS. of the Vulgate in Gregory’s notation 1648 and 1792 (both of Munich) and 2089 (of the Monza Chapter Library). In these the text ends at xiv. 23 and is followed by a short benediction and the doxology. [p. 232]

My first query was now solved—the numbers used by Cranfield go back to Gregory’s survey of Latin manuscripts in his Textkritik and with this knowledge I was able to identify their modern shelf numbers. For this statement, Manson cites Rudolf Schumacher’s 1929 study Die beiden letzten Kapitel des Römerbriefes in Neutestamentliche Abhandlungen. Schumacher discusses the three manuscripts, all of which have the doxology, preceded by a grace wish in one form or another, after 14:23. Schumacher concludes his discussion of these manuscripts by stating: 

Nach allem Gesagten ist nicht daran zu zweifeln, daß es eine lateinische Rezension gegeben hat, welche die beiden Schlußkapitel fortließ.  [p. 15] 

In light of all that has been said, there is no doubt that there was a Latin version that omitted the two final chapters.

While this is a clear enough statement and could well be the case, what was still not clear to me was whether or not these particular manuscripts actually contained such a shorter version as Cranfield said they did. Schumacher in turn pointed to a 1911 article by De Bruyne, entitled “La finale Marcionite de la lettre aux Romains retrouvée.” 

Having already discussed the Monza manuscript (this is VL 86, a detailed discussion and full transcription of which can be found in Frede’s Altlateinische Paulus-Handschriften), De Bruyne first mentions a group of manuscripts that have the doxology after 14:23. According to De Bruyne: 

• Florence, BML, Plut 25.10 has the doxology after 14:23 written in the margin by a second hand, while a third hand has crossed it out. 
• Florence, BML, Plut 25.2 has the doxology after 14:23 in the main text. 
• Florence, Biblioteca Riccardiana, MS 222, as above.
• Madrid, Biblioteca Nacional, A.7, as above. 

De Bruyne then comes to the two Münich manuscripts originally mentioned by Cranfield, BSB, Clm 17040 and BSB, Clm 17043. De Bruyne notes that, in Clm 17040, 14:23 is followed by the text of the blessing he had previously reconstructed for the Monza manuscript, then the doxology (“il contient après le chapitre 14 le souhait suivi de la doxologie” [134]). Since this manuscript has few other archaic features, he suspected it as an interpolation, which was confirmed by the fact that it is actually written in the margin of Clm 17043, of which it seems to be a copy. 

In context, De Bruyne is concerned about using these manuscripts to reconstruct the text of the blessing and doxology found in VL 86. Although VL 86 itself contains the last two chapters of Romans, De Bruyne was already convinced that the early placement of the doxology was evidence for a shorter version of Romans in the Latin tradition. Yet while neither he nor Schumacher clearly state that any of these manuscripts actually omit the last two chapters of Romans, they also do not explicitly say that they do contain them. 

Unfortunately, neither of the Münich manuscripts are available online. However, Elisabeth Klemm provides a detailed treatment of Clm 17043 in her 2004 catalogue. In this catalogue, she explicitly refers to this discussion and says: 

Im Zusammenhang mit textkritischen Untersuchungen zu Rm 16,25-27 und der Frage nach dem originalen Schluß des Römerbriefs zieht H. J. Frede u. a. Clm 17043 heran, in welchem die Schlußverse wie in einer Reihe italienischer Hss. an Rm 14,23 anschließen (die Hss. enden aber nicht damit, es folgen trotzdem noch die beiden Schlußkapitel des Römerbriefs). [p. 89]

In connection with textual-critical studies of Rom 16:25–27 and the question of the original conclusion of the Epistle to the Romans, H. J. Frede draws upon, among other sources, Clm 17043, in which the concluding verses follow Rom 14:23, as they do in a number of Italian manuscripts (though the manuscripts do not end there; the two concluding chapters of the Epistle to the Romans still follow). 

In addition to Klemm’s statement about the Münich manuscripts, I was able to consult images of one of De Bruyne’s Italian group (Florence, BML, Plut 25.2, pictured) and determine that, while it did indeed have the doxology after 14:23, this was then followed by the remainder of the last two chapters, as well as a repetition of the doxology in its normal placement in the Latin tradition in 16:25–27.

Taken together, these data points seemed sufficient to settle my second question. While Cranfield was merely repeating Manson, Manson had misunderstood the (admittedly ambiguous) statements by Schumacher, who was, in turn, replicating the ambiguity present already in De Bruyne. While it may be the case that (as De Bruyne argued) the various placements of the doxology provide evidence for a lost shorter version of Romans, this postulated shorter version seems to be indeed lost, not only in Greek, but also in Latin.

The majority of the Latin tradition has the doxology as 16:25–27, VL86 has the doxology (though largely reconstructed) after 14:23. BML, Plut 25.2 has it in both places. (I was unable to check the rest of the witnesses mentioned by De Bruyne, but they will fall into one of these two categories). VL135 and VL77(the Latin of 012) omit it entirely. This means all four placements found in more than one Greek manuscript are also attested in the Latin tradition. (This last paragraph was edited after initial posting)