Wednesday, December 12, 2018

Hand-to-Hand Combat: P38 vs. GA 2401 (Round 1)

            “The older the manuscript, the better the text” . . . right?  It seems perfectly reasonable to expect the text in manuscripts closer to the original documents to be more accurate than the text in medieval manuscripts.  But at the same time, it’s also perfectly reasonable to reckon that a text that has passed through ten generations of careful copying will be more accurate than a text that has passed through five generations of careless copying.  Today, let’s compare an early copy – Papyrus 38, a fragment produced in the 200s, containing text from Acts 18:27-19:6 and 19:12-16 – to the medieval manuscript GA 2401, which was produced in the 1100s. 
            Henry Sanders described P38 in 1927 in an article that appeared in Harvard Theological ReviewA Papyrus Fragment of Acts in the Michigan Collection – and his data was further refined in Chapter XXIII of The Beginnings of Christianity – Part One, The Acts of the Apostles, Vol. 5; that chapter, The Michigan Papyrus Fragment 1271 was written by Silva New (1933) and includes an uncial transcription.
         GA 2401, meanwhile, is a Praxapostolos manuscript (containing Acts, the Pauline Epistles, and the General Epistles, with book-summaries and some other supplemental compositions).  It is part of the Goodspeed Manuscript Collection at the University of Chicago.
         Contrary to the claim of James White that “Every one of the papyrus manuscripts we have discovered has been a representative of the Alexandrian text-type,” (See The King James Only Controversy, p. 195, 2009 ed.) it is well-established that the text of Papyrus 38 is not Alexandrian.  It is very far from Alexandrian, as we shall see. 
         This hand-to-hand contest will take two rounds; one side of P38 will be considered in each round.   Let’s review the ground-rules:  contractions of sacred names are not counted as variants; transpositions are mentioned but not counted; NA27 is used as the standard of comparison (i.e., for the purposes of this contest, NA27 is the proxy for the original text), and bracketed words in NA27 are counted as text.  In addition, because 2401 contains some secondary corrections, I will make two calculations of 2401’s closeness to NA27:  one with the corrections taken into consideration, and one without the corrections in the equation.  Also, although it would be possible to reconstruct non-extant readings in P38, I will only consider extant readings throughout P38.
         (I thought about trimming away parts of 2401’s text along the contours of P38, so that 2401 would not be at a disadvantage, but after seeing initial results of the comparison, such a step seemed unnecessary.)
         Now let’s get ready to rumble!

Here’s the text of Acts 18:27-19:6 (with corrections) in GA 2401, compared to the text of Nestle-Aland Novum Testamentum Graece:

Acts 18:27-19:6:  2401 Compared to NA27

18:27 – 2401c reads εις τὴν Ἀχαϊαν after παραγενόμενος (+12, -1)
18:28 – 2401 reads διακατηλεγχε instead of διακατηλέγχετο (+0, -2)
18:28 – 2401c reads διαλεγόμενος και after δημοσια (+15, -0)
19:1 – 2401 reads ευρων instead of ευρειν (+1, -2)
19:2 – 2401 reads ειπε instead of ειπεν τε (+0, -3)
19:2 – 2401 reads ειπον after οι δε (+5, -0)
19:2 – 2401 reads ουδε instead of ουδ’ (+1, -0)
19:3 – 2401 reads ειπε δε instead of ειπεν τε (+1, -2)
19:3 – 2401 reads ειπον instead of ειπαν (+1, -1)
19:4 – 2401 reads ειπε instead of ειπεν (+0, -1)
19:4 – 2401 reads μεν before εβαπτισεν (+3, -0)
19:4 – 2401 reads εβαπτισε instead of εβαπτισεν (+0, -1)
19:4 – 2401 reads πιστευσωσι instead of πιστευσωσιν (+0, -1)
19:4 – 2401 reads Χν Ιν instead of Ιησουν (+2, -0)  
19:5 – no variations
19:6 – 2401 reads προεφήτευον instead of επροφήτευον (+1, -1)

            Thus, the text of Acts 18:27-19:6 in 2401, including corrections, has 42 non-original letters, and is missing 15 original letters, for a total of 57 letters’ worth of scribal corruption in this passage.  If we undo the effects of the corrections in 2401 (generally detectable due to the darker ink used by the corrector), then 2401 has 15 non-original letters, and is missing 14 original letters, for a total of 29 letters’ worth of scribal corruption in this passage.

Now let’s consider the text of Acts 18:27-19:6 in P38.  Letters which were only tentatively identified by those who studied the manuscript are shown in red, and are not included in the calculations.

Acts 18:27-19:6:  P38 Compared to NA27

18:27 – P38 reads –ς τὴν Ἀχαϊα after παραγενόμενος (+9, -0)
18:27 – P38 transposes to read πολυ συνε
18:28 – P38 reads δια[λεγόμεν]ος after δημοσια (+2, -0)       
18:28 – P38 reads θελονι-[ος] after Ιην (+6, -0)
19:1 – P38 reads [Π]αυλου κατα τη[ν] (+11, -0)
19:1 – P38 reads [βου]λη[ν] (+2, -1)
19:1 – P38 does not include εγενετο δε εν τω τον Ἀπολλω (+0, -22)
19:1 – P38 reads –ι εις Ιεροσόλυμα (+14, -0)
19:1 – P38 reads το (+2, -0)
19:1 – P38 reads –εφειν εις τ- (+7, -0)
19:1 – P38 reads –ρχετα- instead of κατελθειν (+4, -9) 
19:1 – P38 reads μαθηταις instead of μαθητας (+1, -0)
19:2 – P38 does not include –πεν τε προς αυτους (+0, -15)
19:2 – P38 reads δ’ instead of δε (+0, -1)
19:2 – P38 reads απεκρειναντο (+7, -0)
19:2 – P38 reads λαμβαν[ουσιν τι]νες (+8, -0)
19:3 – P38 reads ο δε Παυλος προς αυ[του]ς instead of ειπεν τε (+16, -7)
19:3 – P38 reads ελεγον instead of ειπαν (+5, -4)
19:4 – no variations
19:5 – P38 reads -φεσιν αμαρτιων (+4, -0)
19:5 – P38 reads επε[πεσεν] instead of ηλθε (+2, -4)

            Thus, in the extant text of Acts 18:27-19:6 in P38, there are 90 non-original letters, and 62 original letters are missing, for a total of 152 letters’ worth of corruption.  (It should be emphasized that this only takes the extant text into consideration.)    

            We have a clear winner in Round One, ladies and gentlemen.  Although the seasoned veteran P38 entered the ring with the advantage of not having as much extant text as GA 2401, this advantage was not nearly enough.  The text of GA 2401 is far, far more accurate than the text in P38.      
            These results have some interesting implications regarding the transmission-streams that produced these two manuscripts.  In the transmission-stream of 2401 (prior to its “correction”), it took scribes about a thousand years to introduce 29 letters’ worth of corruption in this passage (and six of those letters constitute trivial orthographic variations).  Meanwhile in Egypt (if P38 was produced in the same vicinity where it was excavated), it took scribes less than 300 years to introduce 152 letters’ worth of corruption in this passage.

Readers are invited to double-check the data in this post, especially Joey McCollum.


Joey McCollum said...

Me? :)

I checked the microfilms for the relevant pages of 2401 and the transcription and images for P38 on the NTVMR. My collation follows:

GA 2401

18:27 - The entire phrase ος παραγενομενος εις την αχαιαν is in darker ink, so it's possible that the original reading of 2401 was either ος παραγενομενος or possibly ος επιδημησας. Given that no manuscript is known to have the reading ος παραγενομενος without the addition of εις την αχαιαν, and given that the latter reading seems to be a D-P38 reading, the first option seems more likely; your assessment of the addition is probably right. (Translatable)
18:27 - 2401 substitutes πεπιστευκοσι for πεπιστευκοσιν. (Movable nu)
18:28 - 2401 substitutes διακατηλεγχε for διακατηλεγχετο. (Conjugation)
18:28 - 2401c adds διαλεγομενος και after δημοσια. (Translatable)
19:1 - 2401 substitutes ελθειν for κατελθειν. (Preposition)
19:1 - 2401 substitutes ευρων for ευρειν. (Conjugation)
19:2 - 2401 substitutes ειπε for ειπεν. (Movable nu)
19:2 - 2401 omits τε after ειπε. (Conjunction)

It's worth noting that this variant is probably dependent on the substitution of ευρων at the end of the previous verse, and the combination of both variants renders a translatable difference that has some bearing on how we might divide the verses. In NA27, we have a text that could be translated, "Paul went through the upper region to come to Ephesus and to find some disciples. Then he said to them..." In 2401, the text could be rendered, "Paul went through the upper region to come to Ephesus. Finding some disciples, he said to them..."

19:2 - 2401 adds ειπον after οι δε. (Translatable)
19:2 - 2401 substitutes ουδε for ουδ'. (Vowel elision)
19:3 - 2401 substitutes ειπε for ειπεν. (Movable nu)
19:3 - 2401 substitutes δε for τε. (Conjunction)

This last variant may also be related to the ευρων / ευρειν variant in v 1 and the omission of τε in v 2; I haven't looked at Luke's usage throughout Acts, but I think the presence of τε in v 2 would warrant a second usage here, resulting in the sequence τε-τε-δε found in NA27. Without the initial τε, the sequence omit-δε-δε seems to me like it would be more natural.

19:3 - 2401 adds προς αυτους after ειπε δε. (Translatable)
19:3 - 2401 substitutes ειπον for ειπαν. (Orthographic)
19:4 - 2401 substitutes ειπε for ειπεν. (Movable nu)
19:4 - 2401 adds μεν after Ιωαννης. (Conjunction)
19:4 - 2401 substitutes εβαπτισε for εβαπτισεν. (Movable nu)
19:4 - 2401 substitutes πιστευσωσι for πιστευσωσιν. (Movable nu)
19:4 - 2401 adds χριστον before Ιησουν. (Nomen sacrum expansion / contraction)
19:5 - I didn't see any variation, either.
19:6 - 2401 substitutes προεφητευον for επροφητευον. (Orthographic)

In summary, I count 20 variants at this coarse level of detail (6 additions, 1 omission, 13 substitutions), 4 of which are translatable and not easily explained by a common mechanism (although the addition of ειπον after οι δε and the addition of προς αυτους after ειπε δε are not too hard to imagine a scribe supplying absentmindedly).

Joey McCollum said...

GA P38

Given its close relationship to D, P38 probably substitutes ος επιδημησας for ος παραγενομενος in 18:27, but since only the final sigma is extant, we can't say for sure; we'll assume no difference to be conservative here.

18:27 - P38 adds τη αχαια after ος παραγενομενος. (Translatable)
18:28 - P38 adds δια̣[λεγομεν]ος̣ after δημοσια. (Translatable)
18:28 - P38 omits τον before χριστον. (Article)
19:1 - P38 and D change so much material in this verse that I couldn't find a good way to divide up the points of variation. Much of the content from Εγενετο δε to εις Εφεσον is rearranged through transposition, conjugation, and substitution of vocabulary. Near the end of the verse, the phrase και ευρειν τινας μαθητας, dependent on the preceding material, is dropped in P38. This is a very significant difference. (Translatable)
19:2 - P38 substitutes κ̣α̣ι̣ [ειπεν τοι]ς μαθηταις for ειπεν τε προς αυτους. (Conjunction, clarification of referent)

This would more properly be grouped with the previous variation unit, as the substitution for the pronoun αυτους is likely occasioned by the absence of the phrase τινας μαθητας in the previous verse.

19:2 - P38 adds απεκρεινα[ντο] after οι δε. (Translatable)
19:2 - P38 substitutes λαμβαν̣[ουσιν τι]ν̣ες̣ for εστιν after πνευμα αγι̣ον̣. (Translatable)
19:3 - P38 substitutes ο δε Παυλος προς αυτους for ειπεν τε. (Translatable)

Again, this may be considered related to the variant at the beginning of v 2; since there was no τε there, a second one here wouldn't fit as well. Moreover, one clarification of referent was made, and the clarifying phrase προς αυτους was supplied.

19:3 - P38 substitutes ελεγον for ειπαν. (Conjugation)
19:4 - I didn't see any variation here, either.

In 19:5, there's enough space after κυριον in P38 to fit the word ημων, which would result in the expanded nomen sacrum του κ̅[υ̅] [ημων] [ι̅η̅υ̅] [χ̅ρ̅υ̅]; unfortunately, the text is not extant, so we can't say much else.

19:5 - P38 adds [εις αφεσιν αμαρ]τ̣ι̣ων at the end of the verse. (Translatable)

19:6 - P38 adds ε̣υ̣θ̣[εως] after χειρας. (Translatable) I got this one from the transcription, but I must admit, I'm having a hard time seeing it myself.

The total here is 11 coarse differences (5 additions, 1 omission, 5 substitutions), but this doesn't account for the weight of the change made to 19:1 in its entirety. Even if we only count that as a single point of variation, 8 of the 11 differences between P38 and NA27 are translatable and likely genealogically significant. So while our collation data may differ in some respects, I would agree with your conclusions on the accuracy of the scribes of 2401 and P38 relative to the NA27 base text.

James Snapp said...

Thanks Joey McCollum; that was terrific.

Daniel Buck said...
This comment has been removed by the author.