Monday, September 19, 2016

Hand-to-Hand Combat: P72 vs 6 - The Final Fight

          Minuscule 6 has beaten Papyrus 72 twice in hand-to-hand combat, showing that a manuscript produced in the 1200’s can contain a text that is more accurate from a manuscript produced in the late 200’s or early 300’s.  The second contest, though, was closer than the first one – and if one were to set aside textual variants that involve vowel-exchanges (itacisms), it was virtually a tie.  Today, Papyrus 72 and minuscule 6 meet one last time:  we will compare their texts in Jude verses 17-25.  Will Papyrus 72 finally prevail?

17 – 6 reads προειρημενον ρηματων after the first των (transposition)
17 – 6 does not have the second των (-3)
18 – 6 reads των χρονων instead of του χρονου (+4, -4)
18 – 6 reads ελευσονται instead of εσονται (+3)
19 – 6 reads εαυτους after αποδιοριζοντες (+7)
20 – 6 reads τη αγιωτατη ημων πιστει εποικοδομουντες εαυτους instead of εποικοδομουντες εαυτους τη αγιωτατη υμων πιστει (transposition) (+1, -1) [The microfilm is not clear.]
21 – no differences
22 – 6 reads ελεγχετε instead of ελεατε (+4 -3)
23 – 6 reads εσπιλομενον instead of εσπιλωμενον (+1, -1) [The microfilm is not clear; this letter is at the end of a line.  It probably reads εσπιλωμενον but I made the call against it just to be strict.] 
24 – 6 reads ασπιλους και after απταιστους και (+11)
24 – 6 reads γαλλιασει instead of αγαλλιασει (-1) [Again, the microfilm is not clear; the letter is probably present but since I could not see it, I made the call against it.]
25 – 6 reads και  after Θω (+3)
25 – 6 does not have του after παντος (-3)
25 – 6 does not have Αμην (-4)

          Thus, in Jude verses 17-25, minuscule 6 has 34 non-original letters, and is missing 20 original letters, for a total of 54 letters’ worth of corruption.  (This may be reduced to 33 non-original letters present and 17 or 18 original letters absent, for a total of 50 or 51 letters’ worth of corruption, if, as I suspect, the original letters in question are present in the manuscript but obscured in the microfilm-image.)

          Now we turn to Papyrus 72.

17 – no differences
18 – P72 does not have του after εσχατου (-3)  
18 – P72 reads εμπεκτε instead of εμπαικται (+2, -4)
18 – P72 reads επειθυμιας instead of επειθυμιας (+1)
18 – P72 reads ασεβιων instead of ασεβειων (-1)
19 – no differences
20 – P72 reads Υμις instead of Υμεις (-1)
20 – P72 reads τη εαυτων αγιοτητι πειστι ανυκοδομεισθαι instead of εποικοδομουντες εαυτους τη αγιωτατη υμων πιστει (transposition) (+15, -20)
20 – P72 reads εαυτοις at the end of the verse (+7)
21 – P72 reads τηρησωμεν instead of τηρησατε (+4, -3)
21 – P72 reads εις ζοην ημων Ιηυ Χρυ instead of ημων Ιυ Χυ εις ζωην (transposition) (+1, -1)
22 – P72 does not have Και (-3)
22/23 – P72 reads εκ πυρος αρπαζατε διακρινομενους instead of ελεατε διακρινομενους ους δε σωζετε εκ πυροσ αρπαζοντες (transposition) (+3, -20)
23 – P72 reads ελεειτε after δε instead of ελεατε (+2, -1)
23 – P72 reads μεισουντες instead of μισουντες (+1)
23 – P72 reads εσπειλωμενοι instead of εσπιλωμενον (+2, -1)
24 – P72 reads στηριξαι ασπειλους αμωμους αγνευομενους απεναντι της δοξης αυτου instead of
φυλαξαι υμας απταιτους και στησαι κατενωπιον της δοξης αυτου αμωμους (transposition) (+29, -25)
24 – P72 reads αγαλλιασι instead of αγαλλιασει (-1)
25 – P72 does not have σωτηρι (-6)  
25 – P72 reads αυτω after ημων (+4)
25 – P72 reads δοξα κρατος τιμη before δια (+8) [κρατος appears further along in the text so I considered its presence at this point a transposition.  Δοξα is repeated further along in the text.] 
25 – P72 reads αυτω δοξα και instead of δοξα (+7)
25 – P72 reads μεγαλοσυνη instead of μεγαλωσυνη (+1, -1)
25 – P72 does not have και εξουσια προ παντος του αιωνος (-28)   
25 – P72 reads τους παντας εωνας instead of παντας τους αιωνας (transposition) (+1, -2)

          Thus, in Jude verses 17-25, the text of Papyrus 72 includes 87 non-original letters, and 121 original letters are absent.  This yields a total of 208 letters’ worth of corruption in Papyrus 72’s text of Jude verses 17-25.  If NA28 is used as the standard of comparison, the text of P72 does not improve:  in verse 18, P72’s score decreases by three due to the non-inclusion of οτι and increases by three via the non-inclusion of του.    
          In this particular contest, minuscule 6 does not merely win.  It crushes and humiliates.  Its 54 letters’ worth of corruption (at most), acquired in a transmission-stream 1,100 years long, amount to only 26% of the amount of corruption acquired in Papyrus 72’s transmission-stream in the course of about 230 years.  The text of the younger manuscript, in this case, is not just better than the text in the much more ancient manuscript.  The text of minuscule 6 in Jude verses 17-25 is four times better than the text of Jude verses 17-25 in Papyrus 72.

        Now let’s consider these results together with the previous two contests between minuscule 6 and Papyrus 72, to see how the texts of these two manuscripts compare in the entire Epistle of Jude:
In verses 1-10, minuscule 6 has 15 non-original letters, and 26 original letters are absent.  41.
In verses 11-16, minuscule 6 has 31 non-original letters, and 31 original letters are absent.  62.
In verses 17-25, minuscule 6 has 34 non-original letters, and 20 original letters are absent.  54.

Totals for minuscule 6:  80 non-original letters present; 77 original letters absent.  Total:  157.

In verses 1-10, Papyrus 72 has 38 non-original letters, and 50 original letters are absent. 
In verses 11-16, Papyrus 72 has 24 non-original letters, and 79 original letters are absent.   
In verses 17-25, Papyrus 72 has 87 non-original letters, and 121 original letters are absent.

Totals for Papyrus 72:  149 non-original letters present; 250 original letters absent.  Total:  399.

          Thus, minuscule 6 has only 39% as much corruption in the Epistle of Jude as Papyrus 72 has.  Or to put it another way:  the ratio of corruption in minuscule 6 compared to Papyrus 72 is almost exactly 150:400, or 15:40, or 3:8.  If anyone still imagines that a simple appeal to “the oldest manuscripts” is decisive and persuasive, let that person carefully consider this data.

(Readers are invited to check the data and math in this post.)

Sunday, September 18, 2016

Papyrus 72 versus Minuscule 6: Rematch!

          In today’s hand-to-hand combat, Papyrus 72 has returned to the ring for a rematch with minuscule 6.  We recently observed that in verses 1-10 of Jude, the text of minuscule 6 (produced in the 1200’s) was far more accurate than the text of Papyrus 72 (produced around 300).  But what happens in the six verses after that?  We will find out today!     
          This time we will examine minuscule 6 first, in Jude verses 11-16.  The same rules are in play that were used for the previous contest; each manuscript's text is compared to the Nestle-Aland-27 compilation. 

11 – no differences.
12 – 6 reads ευωχιαις instead of απαταις (+5, -4)
12 – 6 reads υμιν before αφοβως (+4)
13 – 6 reads τον before αιωνα (+3)
14 – 6 reads Προεφητευσε instead of Προεφητευσεν (-1)
14 – 6 does not have και after δε (-3)
14 – 6 reads ηλθε instead of ηλθεν (-1)
15 – 6 reads του at the beginning of the verse (+3)
15 – 6 reads παντας instead of πασαν (+4, -3)
15 – 6 reads ασεβεις instead of ψυχην (+7, -5)
15 – 6 does not  have ασεβειας αυτων after εργων (-13)
15 – 6 reads λογων after σκληρων (+5)
16 – 6 reads εισι instead of εισιν (-1)

          Thus in Jude verses 11-16 in minuscule 6, there are 31 non-original letters, and 31 original letters are absent, for a total of 62 letters’ worth of corruption.  (Three letters’ worth of corruption are movable-nu variants.)  
          Now let’s look at the text of this passage in Papyrus 72: 

11 – P72 reads Βαλαακ instead of Βαλααμ (+1, -1)
11 – P72 reads μεισθου instead of μισθου (+1)
11 – P72 reads αντιλογεια instead of αντιλογια (+1)
12 – P72 reads σπειλαδες instead of σπιλαδες (+1)
12 – P72 reads συνευχομενοι instead of συνευωχουμενοι (-2)
12 – P72 reads πυμενοντες instead of ποιμαινοντες (+2, -4)
12 – P72 reads νεφελε instead of νεφελαι (+1, -2) 
13 – P72 reads θαλασης instead of θαλασσης (-1)
13 – P72 reads απαφριζοντα instead of απαφριζοντα (+1, -1)
13 – P72 reads πλανητε instead of πλανηται (+1, -2)
13 – P72 does not have ο after οις (-1)
13 – P72 reads εωνα instead of αιωνα (+1, -2) 
13 – P72 reads τετηρητε instead of τετηρηται (+1, -2)
14 – P72 reads Επροφητευσεν instead of Προεφητευσεν (+1, -1)
14 – P72 reads αγιων αγγελων instead of αγιαις (+9, -3)
14 – P72 does not have αυτου after μυριασιν (-5)
15 – P72 reads ελεγξε instead of ελεγξαι (+1, -2)
15 – P72 reads περει instead of περι (+1)
15 – P72 does not have των εργων ασεβειας αυτων ων ησεβησαν και περι παντων after the second παντων.  The copyist’s line of sight apparently drifted from one occurrence of παντων των to the next one.  (-44)
16 – P72 reads γογγυστε instead of γογγυσται (+1, -2)
16 – P72 reads εαυτω instead of εαυτων (-1)
16 – P72 reads πορεομενοι instead of πορευομενοι (-1)
[Note:  The copyist of P72 skipped from μεμψιμοιροι to και in the text, but he corrected this mistake by writing κατα τας επιθυμιας εαυτω πορεομενοι in the lower margin of the page.]
16 – P72 does not have το after και (-2) 
16 – P72 reads ωφελιας instead of ωφελειας (-1)
16 – P72 reads χαρειν instead of χαριν (+1)

          Thus, in Jude verses 11-16 in Papyrus 72, there are 24 non-original letters, and 79 original letters are absent, for a total of 103 letters’ worth of corruption.  (Thirty-six letters’ worth of corruption involves itacisms or movable-nu.) 

          Once again, the text of minuscule 6, though far from optimal, resembles the text in the Nestle-Aland compilation more closely than does the text of Papyrus 72.  Compared to Papyrus 72, minuscule 6 has 39% less corruption in this passage.  In Jude verses 11-16, the transmission-stream of minuscule 6 – over 1,100 years long – resisted corruption significantly better than the 230-year-long transmission-stream of Papyrus 72.    

(Readers are invited to check the data and math in this post.)

Friday, September 16, 2016

Hand-to-Hand Combat: P72 versus Minuscule 6

          Papyrus 72 is one of the participants in today’s hand-to-hand contest.  It is possibly the oldest substantial manuscript of the book of Jude, being usually assigned a production-date in the late 200’s or early 300’s.  (Papyrus 78 may be slightly older but it is a fragment, containing less than half of Jude’s 25 verses.)  Papyrus 72 contains much more than the Epistle of Jude; it also contains First Peter and Second Peter and some other compositions – but today we will focus on its text of Jude, and not the entire book, but just the first 10 verses.  You can access page-views of P72 at the website of the Vatican Library
          In the opposite corner, we have minuscule 6, from the 1200’s.  Minuscule 6 is one of the manuscripts known to European researchers in the 1500’s; it was cited by Stephanus as witness #5, that is, εʹ, in his 1550 edition of the Greek New Testament.  Microfilm-images of the pages of minuscule 6 can be viewed at the website of the National Library of France; it is catalogued as MS Grec. 112.  The text of Jude begins on page-view 137.  Minuscule 6 is not your typical manuscript; it sometimes shares unusual readings with the important minuscules 1739 and 1881. 
          It is not uncommon to read claims to the effect that the manuscripts used by researchers in the 1500’s were “late and inferior.”  So you might expect that an Alexandrian manuscript made around 300 will prove to be far more accurate than such a medieval manuscript.  Will that be what we observe when P72 and minuscule 6 square off in the ring?            
          Let’s find out.  Today’s battleground consists of the first 10 verses of the Epistle of Jude.  The rules used for the previous contests apply here:  the Nestle-Aland compilation will be the standard of comparison.  Nomina sacra and other decipherable contractions are not considered variants as such.  Transpositions are mentioned but are not included in the totals.  Words in brackets in the NA compilation will be treated as part of the text.  In addition, I have defined P72’s text as the text that left the copyist’s hand; that is, the text of P72 that is evaluated here includes corrections made by the copyist as he wrote.         

Papyrus 72, compared to NA27:

1 – no differences
2 – P72 reads πληθυνθιη instead of πληθυνθειη (-1)
3 – P72 reads ποιησαμενος instead of ποιουμενος (+3, -2) 
3 – P72 reads του before γραφειν (+3)
3 – P72 reads γραφιν instead of γραφειν (-1)
3 – P72 reads περει instead of περι (+1)
3 – P72 reads επαγωνιζεσθε instead of επαγωνιζεσθαι (+1, -2)
3 – P72 reads πειστει instead of πιστει (+1)
4 – P72 reads παλε instead of παλαι (+1, -2)
4 – P72 reads προγεγραμενοι instead of προγεγραμμενοι (-1)
4 – P72 reads χαρειτα instead of χαριτα (+1)
4 – P72 does not read νομον; the copyist wrote this word but then crossed it out. (-5)
4 – P72 reads κν ιην χρν ημων instead of κν ημων ιην χρν (transposition)
5 – P72 reads Υπομνησε instead of Υπομνησαι (+2, -1)
5 – P72 does not have υμας after ειδοτας (-4)
5 – P72 reads απαξ παντα οτι θς χρς instead of παντα οτι ο κς απαξ (transposition) (+4, -2)
5 – P72 reads εγ instead of εκ (+1, -1)
5 – P72 reads Εγυπτου instead of Αιγυπτου (+1, -2)
5 – P72 reads πειστευσαντας instead of πιστευσαντας (+1)
6 – P72 reads απολειποντας instead of απολιποντας (+1)
6 – P72 reads αειδειοις instead of αιδιοις (+2)
7 – P72 reads Γομορα instead of Γομορρα (-1)
7 – P72 reads ε instead of αι (+1, -2)
7 – P72 reads περει instead of περι (+1)
7 – P72 reads απελθουσε instead of απελθουσαι (+1, -2)
7 – P72 reads τερας instead of ετερας (-1)
7 – P72 reads προσκειντε instead of προκεινται (+2, -2)
7 – P72 reads διγμα instead of δειγμα (-1)
7 – P72 reads εωνιου instead of αιωνιου (+1, -2)
8 – P72 does not have μεν after σαρκα (-3)
8 – P72 reads μειενουσιν instead of μιαινουσιν (+3, -3)
8 – P72 reads αθετουσι instead of αθετουσιν (-1)
8 – P72 reads βασφημουσιν instead of βλασφημουσιν (-1)
9 – P72 reads Μιχαης instead of Μιχαηλ (+1, -1)
9 – P72 reads Μουσεως instead of Μωυσεως (+1, -1)
9 – P72 reads επειτειμησαι instead of επιτιμησαι (+2)
10 – P72 reads υδασιν instead of οιδασιν (+1, -2)
10 – P72 reads επειστανται instead of επιστανται (+1)
10 – P72 reads φθιρονται instead of φθειρονται (-1)

The text of Jude verses 1-5
in minuscule 6
Thus when we examine the contents of verses 1-10 of the Epistle of Jude in Papyrus 72, using NA27 as the standard of comparison, we find that 38 non-original letters are present, and 50 original letters are absent, for a total of 88 letters’ worth of corruption, most of which are spelling-related. 

Now let’s compare the text of minuscule 6 to NA27:

1 – 6 reads χυ ιυ (transposition)
1 – 6 reads εθνεσι instead of εν Θεω (+5, -4)  [Regarding this variant, which is not mentioned in Metzger’s Textual Commentary, or in the NET, or in the NKJV’s footnotes, see Robert Waltz’s comments.]
1 – 6 reads ηγιασμενοις instead of ηγαπημενοις (+3, -3)
2 – no differences
3 – 6 reads υμων instead of ημων (+1, -1)
3 – 6 reads εχων instead of εσχον (+1, -2)
4 – 6 reads χαριν instead of χαριτα (+1, -2)
5 – 6 reads ουν instead of δε (+3, -2)
5 – 6 does not have υμας after ειδοτας (-4)
5 – 6 reads Ις instead of ο Κς (+1, -2)
6 – no differences.
7 – 6 reads τουτοις τροπον instead of τροπον τουτοις (transposition)
8 – 6 reads μιαινουσι instead of μιαινουσιν (-1)
8 – 6 reads αθετουσι instead of αθετουσιν (-1)
9 – 6 reads Μωσεως instead of Μωυσεως (-1)
9 – 6 reads ετολμησε instead of ετολμησεν (-1)
9 – 6 reads αλλ instead of αλλα (-1)
10 – 6 reads οιδασι instead of οιδασιν (-1)

Thus in minuscule 6 in the first 10 verses of Jude, 15 non-original letters are present, and 26 original letters are absent, yielding a total of 41 letters’ worth of corruption. 

The score of minuscule 6 improves when the text of NA28 is the standard of comparison.  One of the newly adopted readings in NA28 is in verse 5; instead of “πάντα ὅτι [ο] κύριος απαξ,” the text of NA28 reads  απαξ παντα οτι Ιησους.  This implies a transposition in the text of minuscule 6, but it also brings minuscule 6’s total amount of corruption down to 14 non-original letters present and 24 original letters absent, yielding a total of 38 letters’ worth of corruption.  When P72’s score is compared to NA28, its score also improves, by a single letter, to 87.    

Thus, in Jude verses 1-10, when NA28 is used as the standard of comparison, minuscule 6 has only 44% as much corruption as P72 does.  Papyrus 72, our earliest complete manuscript of Jude, had a transmission-stream only 230 years long (positing its production in 300 and the composition of the Epistle of Jude around the year 70).  Minuscule 6, one of those “late and inferior” manuscripts known to Reformation-era scholars in the mid-1500’s, had a transmission-stream that was about 1,050 years long.  In the first 10 verses of Jude, the copyists in P72’s transmission-line in Egypt introduced twice as much corruption in one-fourth as much time.

(Readers are invited to check the data and math in this post.)

Sunday, September 11, 2016

Fifty Manuscripts at the Vatican Library

          The Vatican Library – officially known as the Bibliotheca Apostolica Vaticana, or BAV – contains a lot of manuscripts, including some New Testament manuscripts.  The Polonsky Foundation Digitalization Project aims to digitalize page-views of manuscripts in the Vatican Library and the Bodleian Library – with a priority on Bibles and Biblical commentaries. 
          Here are some of the Biblical manuscripts which can presently be viewed online, along with brief descriptions and notes.  You can use the embedded links to go directly to the page-views.  (This is not an exhaustive list.  There are many Biblical manuscripts in Latin not mentioned here.)

Papyrus 75:  Extant in Luke 3-24 and John 1-15, the text of this early (c. 225) manuscript closely resembles the text of Codex Vaticanus. 

Papyrus 72:   includes the text of First Peter, Second Peter, and Jude, from the late 200s or early 300s. 

Codex Vaticanus (B, 03), produced around 325, is regarded by many textual critics as the most important of all New Testament manuscripts. 

Codex S (028), also called Codex Guelpherbytanus B, is an uncial manuscript of the Gospels, made in 949.  Elephants are among the animals accompanying the hollow-uncial text of Ad Carpianus near the beginning of the manuscript.  The pericope adulterae begins on 197r.  It is given its own title in the upper margin.  After John 7:53 the lector is instructed to skip to 8:12 and resume reading there.  A large red asterisk is in the side-margin beside the beginning of John 8:3.  The entire pericope adulterae is accompanied by asterisks.  The Gospels are followed by Gospel-lections for Easter-week.

Codex Basilianus (046) (Vat. Gr. 2066), is an uncial manuscript of Revelation, produced probably in the 800s.  It also contains some patristic compositions.  The book of Revelation begins on 259r.   

GA 137 (Vat. Gr. 756) is a minuscule manuscript of the Gospels, produced probably in the 1100s.  The text of Mark is accompanied by the Catena Marcum attributed to Victor of Antioch.  The claim that “asterisks follow v. 8 in 137” is refuted by consulting 150v, where a red “+” appears at the beginning of 16:9, intended to draw the reader’s attention to the note (a normal part of the Catena Marcum) at the foot of 151v, which is also accompanied by a red “+”.  Matthew 1:1 is on 14r.

GA 150 (Pal. Gr. 189) is a Gospels-manuscript from the 1000s.

GA 151 (Pal. Gr. 220) is a manuscript of the Gospels with commentary-material in outer margins, produced probably in the 900s.  A composition by Eusebius, Answers to Questions about the Gospels asked by Stephanus & Marinus begins on 61r.  A transcription of the text of this composition is on the even-numbered pages in Roger Pearse’s Eusebius of Caesarea – Gospel Problems & Solutions, pages 6-128.  This is followed by the chapter-list for Mark and a miniature of Mark; the text of Mark begins on 100r.  Notably, “Isaiah the prophet” is read in Mark 1:2.  Luke begins on 133r.  John begins on 186r.  

GA 157 (Urb. Gr. 2) is one of the most important of all minuscule copies of the Gospels, produced in 1122 for the family of the Byzantine emperor.  It has the Jerusalem Colophon after each Gospel.  

GA 162 (Barb. Gr. 449) is a manuscript of the Gospels, written in strong black ink with red initials at the beginnings of sections.  Luke 11:2, on 151v, features a notable textual variant.

GA 389 (Ott. Gr. 297) is a Gospels-manuscript from the 1000s.  Matthew 1:1 is on 7r; Mark 1:1 is on 56r; Lk. 1:1 is on 88r; Jn. 1:1 is on 142r.  The pericope adulterae begins on 157r.

GA 390 (Ott. Gr. 381) is a manuscript of the New Testament, except Revelation, made in 1281 or 1282.  The Acts and the Epistles appear before the Gospels:  Acts (9r), Romans (51r), First Cor (66v), Second Cor (81b), etc., Hebrews (131r), James (143v), First Peter (148r), Second Peter (152v), First John (156r), Jude ends on 163r.  Matthew begins on 190r; Mark begins on 232r; Luke begins on 261r; John begins on 304r.  

GA 629 (Ottobianus Gr. 298) is a Latin-Greek manuscript from the 1300s or 1400’s known for the presence of the Comma Johanneum (without its final phrase) in Latin and in Greek, on 105v.

GA 850 (Barb. Gr. 504) is mostly a commentary by Cyril of Alexandria, but it includes text from John 7:25-10:18

GA 880 (Ott. Gr. 208) is a manuscript of the Gospels from the 1400s.  Mark 1:1 is on 103r, Luke 1:1 is on 170r, and John 1:1 is on 281r.

GA 2195 (Ross. Gr. 135-138) is in four volumes:  GA 2195 – MatthewGA 2195 – MarkGA 2195 – Luke, and GA 2195 – John.  The pericope adulterae begins on 32v.

Lectionary 35 (Vat. Gr. 351), an uncial lectionary from the 900’s, containing only 25 lections, is a model of elegant penmanship.  

Lectionary 37 (Borg. Gr. 6) begins with the Heothina lections (from Mt. 28, Mk. 16, Lk. 24, and Jn. 20-21),

Lectionary 120 (Vat. Gr. 1156) features lots of gold, plus the Evangelists’ icon-miniatures.  This is truly a deluxe manuscript.  Notice the little ascension-scene on 52r, and the passion-scenes on 194v, and the intricate headpiece on 242r.

Lectionary 123 (Vat. Gr. 1522) was produced in the 900s.  It is written in large uncials, with titles written in gold; simple framework is also gold.  It features full-page miniatures of John, Luke, Matthew, and Mark.  Mark 16:9ff. begins on 177r.

Lectionary 130, part 1 (Ott. Gr. 2) is an uncial Gospels lectionary.  Lectionary 130, part 2 contains more of the same Gospels lectionary.  The Heothina begin on 330v; Mk. 16:9ff. is on 332r.

Lectionary 131 (Ott. Gr. 175) is a minuscule Gospels lectionary.

Lectionary 132 (Ott. Gr. 326) contains readings for the twelve major feasts.  It is written in white (and gold, especially for initials) on a black-dyed background.  

Lectionary 135 (lower writing) and lectionary 136 (upper writing) are two layers of a palimpsest; lectionary 135 (Barb. Gr. 472) consists of text from Matthew 24-25 and John 19, from the 700’s.

Lectionary 379 (Vat. Gr. 357) is an uncial Gospels-lectionary from the 800’s.

Lectionary 549 (Vat. Gr. 1523), produced around 1300, is a neatly written Gospels-lectionary with ornate headpieces. 2627 includes pages from an uncial lectionary (15r-16v).

OLD TESTAMENT MANUSCRIPTS (with some New Testament extracts)

The Barberini Psalter:   Most pages of this Psalter have illustrations, with a generous use of gold.  This manuscript, like Gospels-manuscript 157, was prepared for the family of Byzantine emperor Alexius Comnenus.  Ernest DeWald has written about this manuscript’s background.   Among the Odes at the end of the Psalter, Luke 1:46ff. begins on 266r; Luke 2:68ff. begins on 266v, and Luke 2:29ff. is on 271v.  On fol. 3, written in a much later hand than the main text, is John 1:1-17; this part of the manuscript has its own Gregory-Aland number; it is 2359. 

The Leo Bible, a volume of the Old Testament in Greek (Genesis-Psalms, with Odes at the end of Psalms), written in minuscule but with uncial Table of Contents.  On 564r, the Magnificat is given as Ode #9, extracted from Luke 1:46-55.  On 564r-564v, the prayer of Zechariah is given as Ode #10, extracted from the Gospel of Luke 1:68-80. 

Psalms with commentary, with gold-grounded pictures all the way through.

Psalms with commentary, continued.  An imperial manuscript.  It also has the Odes with extracts from Luke; see 485r & ff.

 An illustrated copy of the books of Kings (beginning with First Samuel).

Greek Old Testament, Part 1 and Part 2, with pictures and commentary. 


VL 12 (Codex Claromontanus) (Vat. Lat. 7223) is a manuscript of the Old Latin Gospels (Matthew from the 400’s; Mark, Luke, and John from the 600’s.  

The Ripoll (or, Farfa) Bible (Vat. Lat. 5729) is a Latin Bible with unusual illustrations.  Matthew begins on 371r.

Vat. Lat. 41 is a Latin manuscript of the Gospels. 

Barb. Lat. 637 is a Latin Gospels manuscript which features a very early capitula system.

Pal. Lat. 502 is a Latin lectionary.

Arch. Cap. S. Pietro D 154 is a manuscript of the Vulgate Gospels.

The Manfred Bible, a medieval Vulgate Bible with sumptuous historiated initials at the beginnings of books. On 399v, in Mark 1:1, “filii dei” is in the margin rather than in the text.  Extra books and an index of names follow Revelation.

The Wigbald Gospels (Barb. Lat. 570), an artistically executed Vulgate Gospels manuscript from the late 700’s, is comparable in some ways to the Book of Kells.

A Coptic manuscript of Acts (with text from chapters 16, 17, and 27). 

A Syriac copy of the Gospels (Vat. Sir. 12) produced in 548.

 A Syriac copy of the Gospels (Vat. Sir. 13) produced in 736.

The book of Psalms in five languages (Barb. Or. 2):  Ethiopic (Ge’ez), Syriac, Bohairic, Arabic, and Armenian. 

An Arabic manuscript (Vat. Ar. 18) of the Gospel of Luke. 

A Bohairic/Arabic manuscript of the Gospels (complete with Ad Carpianus and Eusebian Canon-tables at the beginning, plus book-introductions and icons before each Gospel) made in 1205.  Cross on 20v.  Mt. 1:1 on 23r.  Mk. 1:1 on 147r.  (Mark 16:9-20 is included after 16:8.)  Lk. 1:1 on 237r.  Jn 1:1 on 389v.  PA on 431r.  433r repeats part of 7:52 (at the same point where the text begins on 431r) before continuing with 8:12.

A liturgical scroll made of dyed parchment, from around the year 1100. 

A Greek bestiary from the 1500’s.  A manticore is on 27r; a unicorn is on 27v; a chameleon is on 35v; a dragon is on 39r; a squid is on 54v. 

          Page-views of each manuscript can be selected by using the slider at the bottom of the page; a vertical slider on the right of the page provides magnification.  If greater detail is needed, one can press + and Control simultaneously.
          There are many more manuscripts yet to be digitized!  Thanks are expressed to the Polonsky Foundation, the BAV, and other institutions which assisted in the task of making these page-views available to the public.

(All digitized images at the BAV are under copyright and may not be reproduced without permission from the BAV.)

Friday, September 9, 2016

Interview with Maurice Robinson - Part 3

          Today we conclude our interview with Dr. Maurice Robinson, one of the compilers of the Byzantine Textform.

Q:  Dr. Robinson, what would you say to someone who said, “I want to follow the Majority Text, but I want to follow the majority text in manuscripts up to the year 900, and set aside the late manuscripts, most of which are Byzantine”?

Robinson:  Should others desire to adopt a different type of Byzantine preference theory, they certainly are welcome to do so. Clearly, a theory for the Gospels based on a Byzantine uncial consensus (e.g., that of A E F G H K M S U V Ω, and parts of W) would approximate closely the results obtained when the wider minuscule consensus is included.
          The problem would be more severe, however, in the Acts and Epistles, where the available uncial manuscripts are fewer, particularly those of Byzantine type that predate the 9th century.  The consensus base for those New Testament books would seem to require inclusion of later minuscule testimony, else the resultant text of those New Testament books will be less Byzantine than what the similar process might produce among the Gospels.  Further, the situation in Revelation would be far worse, since the competing Byzantine groups there primarily depend upon minuscules made after the 800’s (generally representing the Majority-Andreas and Majority-Koine forms of the text) — leave these out of consideration, and it would be difficult to say what the resultant text might become (most likely quite non-Byzantine in nature).
          For the Gospels, a pre-9th century Byzantine consensus would be as strong and even more viable than portions of the Hodges-Farstad theory where they appealed to less-than-Byzantine minority groups to settle instances of textual division; certainly such a method also would be far more reasonable than adopting a recensional form of text found only among late manuscripts (such as the Family 35 subgroup).  Other possible approaches that would result in a basically Byzantine form of text could include following the archetype of Family Π/Ka or (perhaps with less likelihood of success) von Soden’s K1 group. From my perspective, however, none of these alternatives appear superior to the present Byzantine-priority hypothesis, methodology, and obtainable results.

Q:  Any comments on Nestle-Aland 28?

Robinson:  Particularly I am disappointed with one aspect of the new format, namely the editors’ decision to eliminate mention of fluctuating degrees of minority Greek manuscript support for variant readings (the “pc” and “al” designations). This move leaves users of the NA28 apparatus unclear as to the relative amount of support a given variant reading might have, and  even worse  readers might presume that only the manuscripts cited for a particular NA28 reading actually support such  and this even though the editors explicitly claim the new format supposedly should prevent such.
          Although the editors claim that “pc and al cannot be used in a precisely defined way, because full collation of all the manuscripts would yield more witnesses for known variants,” such special pleading appears peculiar, particularly when the full collation data of Text und Textwert are compared against readings designated pc or al in the former NA27 apparatus.  The Text und Textwert data regularly validate the propriety of the pc and al designations within a concise, more limited apparatus. From my perspective, those designations ought to be reinstated in future NA editions, along with re-inclusion of at least some of the previous consistently cited witnesses from NA26/27 that no longer appear in NA28.
        On a positive note, the new typeface is nice, and the NA28 regularization of some orthographic forms was long overdue. Similarly, I consider the elimination of conjectural suggestions from the apparatus beneficial, although an appendix listing the more important of these could be informative in terms of understanding scholarly views on the matter.

Q:  In the approach you describe in “The Case for Byzantine Priority,” a prohibition on conjectural emendation is Rule #1. What do you think of NA28’s introduction of a conjectural emendation into the text of Second Peter 3:10?

Robinson:  Given that the UBS/NA editions long have had a conjecture at Acts 16:12, the inclusion of a new conjecture at Second Peter 3:10 (dating from at least the time of Tischendorf — see his 8th edition’s apparatus) is unsurprising, particularly since the basic Alexandrian reading in that location — found in the main text of previous critical editions dating back to Tregelles and W-H — simply makes no good sense (kai ta en auth erga eureqhsetai, literally “and the works in her shall be found”). The point is well illustrated in the translational circumlocutions that appear among those English versions based on the critical text.  Consider the following, grouped according to how they render eureqhsetai:

Lexham:     “and the deeds done on it will be disclosed.”
HCSB:       “and the works on it will be disclosed.”
NRSV:       “and everything that is done on it will be disclosed.”
NIV:          “and everything done in it will be laid bare.”
NET:          “and every deed done on it will be laid bare.”
CEB:          “and all the works done on it will be exposed.”
ESV:          “and the works that are done on it will be exposed.”
GW:           “and everything that people have done on it will be exposed.”
ISV:           “and everything done on it will be exposed.”
NCV:         “and everything in it will be exposed.”
TEV:          “with everything in it will vanish.”
NIrV:         God will judge the earth and everything done in it.”
Mess:         “and all its works exposed to the scrutiny of Judgment.”
NLT:          “everything on it will be found to deserve judgment.”
Voice:        “and all the works done on it will be seen as they truly are.”
NAB:         “and everything done on it will be found out.”

         Of these, only the Roman Catholic New American Bible comes close to the base meaning of the problematic construction. So certainly, the NA28 conjectural inclusion of ouk before eureqhsetai makes far better sense without requiring alteration of the proper meaning of the word in the process (thus NA28 in conjecture: “shall [not] be found”). The proposed conjecture, therefore, is quite good, and similar in quality to what Rendel Harris suggested for First Peter 3:19, where the main text en w kai should be supplemented by the conjectural addition of Enwc (thereby reading, “in which also Enoch”) — a brilliant conjecture; yet equally without manuscript evidence, and equally recognized by most scholars as non-original, just as they ought to regard the current NA28 conjecture at Second Peter 3:10.
          Put simply, researchers — particularly those involved in the study and use of actual manuscript testimony — should not invent or prefer readings that have no known existence among the Greek manuscript base merely because such might “make better sense” than an otherwise problematic preferred reading (not that I consider eureqhsetai to be original over against the Byzantine katakahsetai, but obviously the critical text editors do so presume). As I recently commented on the Evangelical Textual Criticism blog,

The problem I have with conjectural readings is not restricted to a priori concerns related to a Byzantine Priority or majority text position, but rather as ultimately involving transmissional considerations; i.e., any conjectured reading — assuming such supposedly to be more reasonable than what appears among the existing witnesses — would have to explain transmissionally how and why such would utterly disappear from our known transmissional history. Were such conjectures actually superior to all extant alternatives, I would consider their lack of perpetuation to be inexplicable.

Q:  Could you briefly explain how the NA28 has many conjectural emendations if one considers short series of variant-units instead of just single variant-units?

Robinson:  I have already written extensively on the so-called “zero-support” verses in the Nestle-Aland editions, both in a published essay in Translating the New Testament: Text, Translation, Theology and in a subsequent ETS presentation in 2012. 
          To summarize: if individual variant support in NA27 is considered in a linear manner (i.e., the stated documentary support for an entire verse containing at least two variant units, when reduced to its combined joint agreement), at least 105 whole verses exist in the critical text that apparently lack any actual existence in the form published, whether from any Greek manuscript, ancient version, or patristic source. In those instances, the result obviously becomes de facto conjecture.
          In my follow-up paper, I examined two-verse segments in NA27 using the same criteria, and found an additional 210 similar portions of text that again as published lack attestation from any known witness (suffice it to say that among the manuscripts comprising the Byzantine Textform, such never occurs in relation to passages of similar length). Note that the same findings apply to the NA28 edition as well, since its main text and apparatus support basically remain the same.

The Greek New Testament
for Beginning Readers
Q:  Could you explain again why Revelation is so different in the Hodges-Farstad compilation?

Robinson:  Although more differences appear between H-F and RP in Revelation than elsewhere in the New Testament, they are not that extensive as when either text is compared against the Old Uncial form of the critical editions. Rather, the primary differences between the Byzantine form of text in H-F and RP mainly involve H-F utilizing a particular stemmatic approach that prefers a minority Byzantine subgroup that they considered original – a group that at times represents less than 30% of the Byzantine manuscripts of that book. In contrast, RP2005 presents a non-stemmatic model representing a general consensus among the two primary Byzantine groups within that book.
          Where these two groups divide, RP generally follow the Majority-Koine group except where a significant number of its manuscripts align with the Majority-Andreas group – this because the Majority-Andreas group appears to reflect a single archetype derived from the Andreas commentary that usually accompanies those particular manuscripts.

Q:  It was recently acknowledged by a textual critic from Dallas Theological Seminary that “many” people subscribe to the Byzantine Priority school.  Besides you, who are these people?

Robinson:  What Dallas critic might have suggested such (I speak as a fool)?  I would prefer to say that many have a preference for a text similar to the Byzantine who might claim to be majority or Byzantine supporters, but who speedily dissent from such whenever the Byzantine Textform departs from their favored Textus Receptus/KJV type of reading.
            Beyond Pierpont and myself, among those who are not TR/KJV partisans but who favor some form of the Byzantine text (not necessarily agreeing with our specific theory or methodology nor resultant form of the text) would include Hodges and Farstad, John Wenham, Jakob van Bruggen, Peter Johnston, Harry Sturz, Wilbur Pickering, Paul Anderson, Thomas Edgar, James Davis, Donald Brake, Timothy Friberg (not all still living) and others, including several more Europeans along with many of my own students.  Not all of these have published in relation to textual matters, and thus some names may be unfamiliar; yet in general they remain pro-Byzantine to some degree.  There also are numerous laypeople that have communicated with me or these others over the years who hold to some sort of Byzantine or majority text position, but I only mention here a few who have published within academia.

Q:  Finally:  in NA28, in Second Peter 2:18, the editors rejected oligws and adopted ontws, even though the adoption of oligws had previously been given an “A” ranking (as if the editors were certain that it was correct).  Any idea how that happened?

Robinson:  The answer apparently is the “wag the dog” influence of CBGM and little else; this particularly in view of Metzger’s previous strong defense of oligws in his Textual Commentary. I also note that in the process UBS5 lowered the rating from “A” to “C” – again without providing any particular reason or justification for such.  Nice that they here adopted a Byzantine reading, but clearly not for the same reason that I would do so.

Thank you for sharing your thoughts.

  end of interview  

Links:  Part One.  Part Two.

Wednesday, September 7, 2016

Interview with Maurice Robinson, Part 2

          Today as we continue the interview with Dr. Maurice Robinson, we investigate how the Byzantine Priority position deals with one specific textual variant-unit.

Q: Could you provide an example of your text-critical approach by walking me through why you favor the Byzantine reading of Luke 6:1, instead of the Alexandrian reading?

Robinson:  I addressed this particular unit in a 1993 article (in Faith and Mission 11), and somewhat repeat myself here.
          I regard the shorter Alexandrian reading as possibly due to homoioteleuton in the Alexandrian archetype, even while suggesting other proximate causes of variation also to exist.  A primary principle of internal evidence should be to presume accidental error as more likely than intentional alteration, especially where omission might result from transcriptional failure and yet still produce a “sensible” reading — one that then would be less likely to receive correction in subsequent copies within its particular transmissional line.
          As for a short walk-through, the external evidence is basically clear:  the dominant united Byzantine Textform reads en sabbatw deuteroprwtw diaporeuesqai, while a small minority made up of predominantly Alexandrian-type manuscripts, along with several versional witnesses, omits deuteroprwtw.  Assuming the originality of the more dominant Byzantine reading, the presence of deuteroprwtw appears best to explain the alternative than vice versa, and that on two grounds:

1. Accidental omission.  The transcriptional factor is fairly obvious: the shorter reading lacks deuteroprwtw; this easily could have occurred — particularly in a small number of manuscripts — due to haplography, skipping from -tw d- to -tw d- in the phrase sabbatw deuteroprwtw diap-.  Such readily could have been the proximate cause of the Alexandrian archetypal reading underlying its omitting manuscripts at this point, and also for those otherwise unrelated minuscules that likely reflect independent error.  But this is not the only possible factor, since intentional removal of a difficulty cannot be ruled out.

2. Intentional omission.  Due to the problematic nature of the longer reading (i.e., what does deuteroprwtw actually mean?), intentional excision of a questionable word equally may have occurred in the Alexandrian archetype or otherwise unrelated witnesses.  Certainly, whatever significance deuteroprwtw may have had to first-century readers remains unclear (I suggest “second chief sabbath” — but even this is open to interpretation).

          Interestingly, similar problematic designations appear in the Septuagint: Psalm 23’s title speaks of “the first of the sabbaths” (ths mias sabbatwn).  Psalm 47’s title mentions “the second sabbath” (deutera sabbatou); and Psalm 93’s title reads “the fourth of the sabbaths” (tetradi sabbatwn) — even though today we have no knowledge of what these designations might mean, despite their unquestioned presence in the Septuagint.  Also, Meyer (In Vol. 2, pages 47-50 of his Handbook to the Gospels of Mark and Luke) notes the similar deuteroescatos (penultimate, next to last) and deuterodekath (“the second tenth”; Jerome, ad Ez. 45).  There thus is no reason why sabbatw deuteroprwtw may not have been understood by the original first-century recipients of the Lukan narrative, even if we today are uncertain.
          In contrast, Metzger (in his Textual Commentary) claims to defend the Alexandrian omission of deuteroprwtw by suggesting “transcriptional blunder” to have caused the rise of the Byzantine deuteroprwtw reading. Yet Metzger’s convoluted explanation of such (taken from Westcott & Hort and Meyer) is highly problematic:

● Step 1: “Perhaps some copyist introduced prwtw as a correlative of en eterw sabbatw in ver. 6, and”
● Step 2:  “a second copyist, in view of [Lk] 4.31, wrote deuterw, deleting prwtw by using dots over the letters — which was the customary way of cancelling a word.”
● Step 3:  “A subsequent transcriber, not noticing the dots, mistakenly combined the two words into one, which he [then] introduced into the text.”
● Step 4 (required, but not mentioned by Metzger): further adjusting the orthography from the two-word form deuterw prwtw into the single-word form deuteroprwtw
● Step 5 (also required but not mentioned by Metzger):  this highly problematic reading then would not only be accepted as valid, but also perpetuated without removal or correction among nearly all remaining manuscripts over the centuries of manual transmission.

In addition to the above, the second edition of Metzger’s Textual Commentary offers an alternative but even more convoluted explanation, “as Skeat has suggested” in his “Final Solution” article:

● Step 1:  “by dittography the letters batw were added to sabbatw.”
● Step 2:  “A later copyist interpreted the b as deuterw and the a as prwtw, and”
● Step 3:  “took tw as an indication that the adjective was to agree with sabbatw.”
Step 4 (not stated, but obviously required): In a resultant copy the presumably abbreviated b and a would have to be written in full form as deuterw and prwtw;
● Step 5: (also not stated but required): the extraneous tw would have to be removed;
● Step 6: (similarly): the two words then would be combined into one, with the orthography corrected (as noted above in Metzger’s Step 4); and then
● Step 7 (Metzger’s Step 5): almost all remaining manuscripts then would accept this highly problematic reading and perpetuate it without removal or correction over the centuries of manual transmission].

(Notice that Skeat’s third step simply makes no sense, since the result would have been en sabbatw deuterw prwtw tw diaporeuesqai auton — a construction occurring nowhere else in the New Testament or Septuagint (the usual expression requires en tw [or 2x in the LXX ama tw] + infinitive + accusative subject of the infinitive).  Yet Skeat boldly claims that “Once it has been accepted that deuteroprwtw is a feasible expansion of BATW” everything else will follow without complication.”)

          Metzger’s explanation represents a noble attempt to explain the rise of the more difficult deuteroprwtw by means of compounded transcriptional errors — yet according to his or even Skeat’s hypotheses all initial steps somehow would have to occur in a single exemplar manuscript, with a resultant copy then containing a “clean” form of that “more difficult” reading. That manuscript then becomes the progenitor of that which appears among almost all other manuscripts throughout copying history without subsequent correction.  
          As I stated in 1993, “the logic simply does not follow.”  Unless there were some sort of formal revision process based upon that one particular resultant manuscript, the odds against the massive perpetuation of presumably a “nonsense” reading would be utterly overwhelming — especially since under normal transmissional circumstances such a supposedly egregious error should have been systematically and speedily corrected by the usual scribal diligence, and thus lack perpetuation among most subsequent copies.
          In essence, the Metzger/Hort/Meyer explanations offer a perfect example for Occam’s Razor:  is it easier to accept a highly convoluted 5- or 7-step process that requires a complex and problematic transmissional history?  Or is an assumption of either accidental haplography (skipping from -tw d- to -tw d-) or deliberate removal of a single problematic word the simpler solution?  The answer would seem obvious to anyone not overly enthralled with the Alexandrian text-type.

(More of this interview coming soon!)