Yesterday at church, I happened to be reading Eph 5.22 and thinking again about the relationship between Eph 5.21 and Eph 5.22. My NA26, which I had with me at the time, notes the possibility of punctuating v. 21 with v. 22 or separately from it. The NA28 punctuates it with v. 22 and the paragraphing follows suit. One of the reasons for doing this is because v. 22 doesn’t have a main verb but one that is implied from the participle ὑποτασσόμενοι (“being in submission to”) from v. 21. In this, v. 21 is set apart from the other similar participles in 5.16–20 that unpack what it means to “walk worthy” (5.15).
However, the apparatus of NA also notes that most manuscripts have an explicit imperative in 5.22 which would make 5.22 line up naturally with the other second person imperatives in the rest of Eph 5.22–6.20 (see 5.25, 6.1, 6.5, 6.10). Each of these starts its own paragraph in NA.
There are two alternate readings in 5.22. The first is ὑποτάσσεσθε (“y’all be in submission to”) found in K L 630 Byz syr. The other is ὑποτασσέσθωσαν (“they should be in submission”) found in 01 A I P 0278 6 33 1505 1739 lat syh co etc. What I realized yesterday in thinking about this is that the second reading has a really good claim to originality; in fact, I now think it may have the best claim to that.
Not only is it attested early and well, but it can easily explain both the alternate readings while alternatively not being well explained by either. It explains the shorter reading (found only in P46 B Clem Hier) by simple homoioteleuton, the word being omitted because of the repeated ν on ἀνδράσιν just before it. On the other hand, it explains the Byzantine reading which is the obvious way to assimilate this verse to the rest of this section, the other imperatives being 2nd person rather than 3rd person as we have with ὑποτασσέσθωσαν. I think this latter point is also good grounds against preferring the shorter reading; if a scribe were going to add a verb here (as, I readily admit, would be natural), it would most likely have taken the form of the 2nd person imperative to fit with the others. In other words, it would take the form of the dominant Byzantine reading.
There is, then, a strong case to be made for ὑποτασσέσθωσαν as the original reading. And, if so, then v. 21 should be read more with what precedes and v. 22, more with what follows. The paragraph break thus belongs after v. 21 not before it.
The question I had at church was whether or not anyone else had taken this view. Sure enough, Tregelles prefers ὑποτασσέσθωσαν and he is followed by the editors of the new THGNT. The latter also has a paragraph break at v. 22. [Update: Lachmann has it and WH give ὑποτασσέσθωσαν as a marginal reading.] So, I am at least in good company.
Whatever your view, this is certainly not an insignificant decision. Given the debates about vv. 21 and 22, the choice of variant and its effect on where to break the text, affects how you read and apply the text. Let no one say that textual criticism doesn’t influence interpretation and application. The other lesson from this example is: always take your GNT to church! You never know what you’ll discover.
However, the apparatus of NA also notes that most manuscripts have an explicit imperative in 5.22 which would make 5.22 line up naturally with the other second person imperatives in the rest of Eph 5.22–6.20 (see 5.25, 6.1, 6.5, 6.10). Each of these starts its own paragraph in NA.
There are two alternate readings in 5.22. The first is ὑποτάσσεσθε (“y’all be in submission to”) found in K L 630 Byz syr. The other is ὑποτασσέσθωσαν (“they should be in submission”) found in 01 A I P 0278 6 33 1505 1739 lat syh co etc. What I realized yesterday in thinking about this is that the second reading has a really good claim to originality; in fact, I now think it may have the best claim to that.
 |
| Variant and paragraph break at Eph 5.22 in 01 |
There is, then, a strong case to be made for ὑποτασσέσθωσαν as the original reading. And, if so, then v. 21 should be read more with what precedes and v. 22, more with what follows. The paragraph break thus belongs after v. 21 not before it.
The question I had at church was whether or not anyone else had taken this view. Sure enough, Tregelles prefers ὑποτασσέσθωσαν and he is followed by the editors of the new THGNT. The latter also has a paragraph break at v. 22. [Update: Lachmann has it and WH give ὑποτασσέσθωσαν as a marginal reading.] So, I am at least in good company.
Whatever your view, this is certainly not an insignificant decision. Given the debates about vv. 21 and 22, the choice of variant and its effect on where to break the text, affects how you read and apply the text. Let no one say that textual criticism doesn’t influence interpretation and application. The other lesson from this example is: always take your GNT to church! You never know what you’ll discover.
You would have been saved from all this trouble had you chosen to take your THGNT copy, rather than the NA, to church.
ReplyDelete“As every conjecture provokes refutation, this at all events advantages our understanding of the passage, and only the best conjectures will win acceptance” etc. (Maas [Flower], § 18).
DeleteAs I understand it, one of the salient features of the household codes genre is the direct attention each party receives. If so, it seems unlikely that Paul would vacilate between second and third person address. The second reading here could be due to faulty memory in the copying process, to the visual similarity of the third person ending to the suffix of the previous word and to the next word, or to an attempt to minimize the moral and social autonomy assigned to women (c.f. the minority readings in 1 Tim 5:16 and Jm 4:4, which might reflect the same tendency) by not appealing directly to them. If assimilation was a major factor, it would seem that more witnesses would shift the imperative to directly after the noun of direct address.
ReplyDeleteAre you suggesting that assimilation is not a factor in any of the variants in play here?
DeleteI'm suggesting that most witnesses to the second person form are not reproducing an assimilation but rather transmitting the authorial text.
ReplyDeleteAh. So how were the other two readings created on your view?
DeleteHomoeoteleuton in the case of the short reading. The third person form is harder to account for, but I just supplied my thoughts, such as they are.
DeleteTwo clarifications might not be out of place. First, Byzantine priority theorists might not all agree on how most plausibly to explain various minority readings. I do not know how MAR, Jonathan Borland, or others approach this.
DeleteSecond, I do see assimilation as a major factor in biblical textual transmission. Yet the practice is sporadic, not systematic. Thus for assimilation to have occurred dozens or hundreds of times in this place seems unlikely.
Logically, it only needs to have happened once. Historically, it may have only happened a few dozen times. Not hard to imagine, I’d say. But, a study of coherence (once the data are available) will shed light on that.
DeleteOnly once? You mean, if the Byzantine textype originates in a single exemplar?
DeleteAs to coherence, that of course depends on the validity of the CBGM and the uses made if it. Maybe when I understand these things, I'll change my time. I haven't finished reading your and Tommy Wasserman's book yet. :)
Well, then, quit reading this and get back to that excellent book! ;)
DeleteAs for Byz, as a text, it has to originate from somewhere. Byz prioritists think it always originates from the single exemplar that is the autograph. So, why not?
In the process of transcribing manuscripts of Ephesians for the IGNTP, I've had a lot of time to think about this variant myself. I agree with some of your points in favor of υποτασσεσθωσαν, but there are other intrinsic and transcriptional factors to consider.
ReplyDeleteIn defense of the shorter reading, homoioteleuton on the basis of a single letter is less plausible than it would have been if larger common ending were involved. That doesn't mean that it's impossible here, but it seems a bit strained, especially given the length of the word being omitted. I agree that on intrinsic grounds, either longer reading seems more original than the shorter reading, but on transcriptional grounds, we could view the two competing longer readings as independent attempts to complete an incomplete sentence after the paragraph breaks had been introduced.
A better argument against the shorter reading, in my opinion, would be that the omission was early and deliberate, intended to remove redundancy in what was perceived to be a single long sentence.
That leaves us with the choice between third-person and second-person in the longer reading. I agree that assimilation might well have changed an original υποτασσεσθωσαν to υποτασσεσθε, but I feel there's a better transcriptional argument in the opposite direction. A scribe copying this section for the first time might not know that it was part of a series of commands addressed to specific groups. Without the later context in mind, such a scribe might mistake the vocative αι γυναικες for the nominative (which looks exactly the same) and thus adopt the third-person subjunctive. Once the pattern became clear with the address to husbands, the scribe would understand that second-person verbs were appropriate.
Thanks, Joey. Are you suggesting that the longer readings were simultaneously added to complete an incomplete sentence and removed early on to remove redundancy? That seems conflicted to me.
DeleteAs for your preferred explanation, it seems that on your scenario, the scribe was quite cognizant of the change from 3rd to 2nd person; wouldn't he then go back and fix his mistake then? In any case, I would think the vocative would be obvious enough on the first read (assuming a scribe copying a text without any initial familiarity with it).
Re: homoioteleuton, how many letters does it take to enable such a mistake? It seems to me that one is quite enough. But even if you think it less likely (not the same as unlikely), that would well explain why we find this shorter reading in so few witnesses. Most witnesses eventually came to have υποτασσεσθε, however, so there was less and less opportunity over time for that particular mistake.
I agree with Joey on caution about a single letter justification for a homoioteleuton leap.
DeletePeter,
DeleteI wanted to offer different transcriptional scenarios that could explain either the longer or shorter reading for the sake of completeness, but my point was not that all of those scenarios happened together. It's a possibility that such a thing could have happened, but I agree with you that it is quite unlikely.
Regarding my explanation for the rise of υποτασσεσθωσαν, I agree that the scribe responsible would have to be making a conscious change, but if he was copying by ear or working in a hurry, then he could either have forgotten about the change or decided not to concern himself with it by the time he got to the section on husbands. I'm sure we could think of other cases where scribes of the NT messed up an early part of a sequence, then got the later parts right, but didn't fix their original mistake. Perhaps some scribes even developed some convictions from Luke 9:62 about their work? :)
I feel like the immediate context makes the vocative less obvious, especially since no other second-person verb or adjective occurs in the verse or in Paul's explanation. More importantly, your intrinsic argument in favor of the third person assumes that Paul himself would not have considered the vocative obvious. Moreover, as Stephen pointed out, it's hard for me to see why Paul would have phrased the commands inconsistently from one audience to the next; even when he addresses the children in Eph 6:1, he uses second-person.
As for homoioteleuton, I can't say I have a formula to determine the likelihood with which it would occur in a given instance, but if I did have one, I imagine that it would involve a ratio between the length of the common ending and the length of the text being omitted. A similar ending consisting of a single letter is more plausible if a very short word is being omitted; the omission of ως after δουλευοντες in Eph 6:7 by many witnesses could be such an example. For a longer word, homoioteleuton on a common word ending would be more plausible; some good nearby examples would be the reduction by a few witnesses of αδοντες και ψαλλοντες to αδοντες in Eph 5:19 (with a moderate-length common ending -οντες) or the change from τοις ιδιοις ανδρασιν to τοις ανδρασιν among some witnesses in Eph 5:24 (with a shorter common ending -οις). For an entire verse or part of a verse, we might expect an entire word or phrase to be the occasion for homoioteleuton (e.g., Eph 5:30, which you've discussed on the blog before).
I think the strongest argument for υποτασσεσθωσαν is its early and diverse external support. If it could be established that this reading gave rise to the shorter one by homoioteleuton, then that would give it an even earlier place in the textual history, but I'm still not convinced that homoioteleuton is more likely than a deliberate stylistic omission (which could have given rise to the shorter reading from either of the longer ones). Nevertheless, a big obstacle to my argument for υποτασσεσθε is the diversity of the witnesses in favor of υποτασσεσθωσαν. It might be reasonable to suppose that one scribe changed the verb from second- to third-person by mistake, but how reasonable would it be to say that the textual groups in support of υποτασσεσθωσαν (witnesses from different Alexandrian strata, Family 1739, the Harklean group, and the Latin witnesses) all got this reading from one common ancestor? It would have to have been a very early mistake.
Regarding the last point, if the short reading does derive from the third person form, the alliance of Aleph 1739 etc here is not too surprising. It's just another example of interconnected witnesses known to be such from elsewhere (e.g. Eph 1:1). (In this scenario, this unit would be a case of B not always preserving the original Alexandrian reading.) Yes, the third person reading must be early, but of course it is generally agreed that most major variants arise early (perhaps due in part to hasty copying work in the early days of the Church?).
DeleteSo, you agree that υποτασσεσθωσαν could have led to the shorter reading, you just think an intentional change is more likely than an unintentional change?
DeleteReturning to homoioteleuton, I note that D shows homoioteleuton in Matt 5.32 with only two letters in common. The same is true of the shorter reading, in my view, at Matt 19.9. To quote Mike Holmes, "the really key factor facilitating homoioteleuton is not the repetition of the entire word (μοιχαται/μοιχευθηναι] but only of the last syllable..."
In any case, looking at how similar -σιν and -σαν are in 01 above, homoioteleuton doesn’t seem hard to imagine here at all.
DeleteI do agree that υποτασσεσθωσαν could have given rise to the shorter reading; I even agree that between the possibilities of homoioteleuton and stylistic excision, it has a slightly higher probability of giving rise to the shorter reading than υποτασσεσθε does. Nevertheless, while I normally adhere to the rule of prioritizing accidental changes over intentional ones, the explanation of homoioteleuton here requires enough qualifications to make it less likely to me. The most likely transcriptional explanation I can think of for the shorter reading is a single early omission made for stylistic reasons, and on this basis, either longer reading could precede the shorter one.
DeleteBut I consider transcriptional probability secondary to intrinsic probability, and on these grounds, υποτασσεσθωσαν has the same problem as the shorter reading: it creates a stylistic inconsistency in a sequence where we would reasonably expect Paul to be consistent.
(Thanks for the link to Holmes's paper, by the way; it's a great resource!)
I note that P46 omits απεκδεχομεθα in Phil 3.20 which Royse says is due to a leap (p. 277). I assume he means a leap involving a single letter alpha.
DeleteFor Joey. "how reasonable would it be to say that the textual groups in support of υποτασσεσθωσαν (witnesses from different Alexandrian strata, Family 1739, the Harklean group, and the Latin witnesses) all got this reading from one common ancestor? It would have to have been a very early mistake."
ReplyDeleteWhen a work is so popular and important, we can not imagine a MS transmission without contamination. Of course, for sttistical reason it is more likely that contamination goes toward the readings of a very popular textual type than toward a very rare, old and out of fashion MS.
Paolo,
DeleteAre you arguing that the omission, which is found in just such a manuscript, B as well as the oldest manuscript P46 is likely the original?
Tim
Paolo,
DeleteBoth good points! It could be that υποτασσεσθωσαν was at an earlier stage a more popular reading and that this led to its early adoption in other textual groups. To demonstrate this, of course, we'd need a fuller picture of the transmissional history of Ephesians to see if such a situation holds true for these groups in other variants.
Thanks, Joey.
DeleteFor Tim. No, Tim. I don't dare to propose solution in a field so far removed from my studies. I wanted only to remember that in overabundant transmissions a high rate of contamination is the rule. The path is described by Joey: "It could be that υποτασσεσθωσαν was at an earlier stage a more popular reading and that this led to its early adoption in other textual groups. To demonstrate this, of course, we'd need a fuller picture of the transmissional history of Ephesians to see if such a situation holds true for these groups in other variants."
Col 3:17 also "drops" an expected imperative that otherwise would repeat the preceding verb. "καὶ πᾶν ὅ τι ἐὰν ποιῆτε ἐν λόγῳ ἢ ἐν ἔργῳ, πάντα ἐν ὀνόματι ⸂κυρίου Ἰησοῦ⸃"
ReplyDeleteAmidst all this theory-spinning (including the validity of possible haplography -- I do not call it homoioteleuton when only a single letter may be a proximate cause, which can and does occur frequently enough among the MSS):
ReplyDeleteApparently no one has noticed the similar situation at 1Cor 14:34. There Byz reads υποτασσεσθαι while NA reads as primary the same υποτασσεσθωσαν as is being discussed here from a secondary standpoint -- more importantly, the NA reading in 1Cor is supported by many of the same witnesses as in the Eph passage, specifically א.A.33.81.365.(1175).1241s.2464.
I would see the 3pl imperative both here and in the 1Cor passage as reflective of attempted stylistic editorial improvement on the part of some archetype underlying the MSS supporting such at those points, while continuing to take the Byz reading as original in both instances (no surprise there, of course).
MAR, no surprise indeed. But what do you mean by stylistic improvements? In 1 Cor 14.34 the 3rd person makes good contextual sense given the other 3rd person verbs there (cf. θελουσιν in v. 35). (Even there, the infinitive is the most natural stylistic change since we have one just prior in λαλειν.)
DeleteHere in Eph 5, there is no real reason to prefer the 3rd person given the train of 2nd person imperatives that follow. There is no omission in 1 cor 14.34 like there is here either.
So, I must demur on your preference for the Byz reading in both cases (little surprise there!).
Actually I see the 3pl imperative as a much easier reading, given the preceding vocative αι γυναικες (which might appear awkward to various scribes when joined with a 2pl imperative, whereas as subject of the 3pl imperative everything flows smoothly). That is one issue.
DeleteAs for the NA omission of the verb entirely, this likely reflects harmonization to a parallel construction, where 5.21-22 is paralleled by the 5.24 follow-up in which υποτασσεται is initially mentioned (in relation to the church), followed by ουτως και οι γυναικες without any form of υποτασσω present. Such attraction to a nearby parallel thus may well have influenced the NA omission of the verb from a Byzantine original presence at 5.22.
Many thanks, Peter, for drawing attention to this variation unit. As it happens, it caught my attention a few months ago just minutes before meeting with a young married couple who disagree with each other on Ephesians 5: 21-22.
ReplyDeleteI was concerned that Westcott and Hort gave the text without an explicit verb form in v. 22 on the basis of no manuscript evidence other than B-03. Aside from the limited Patristic support for their text at this point, the case for the verbless text is not much stronger nowadays. B-03 was joined by P46, while 0278 (dated later, I know) has joined the evidence for the the third-person plural imperative (which I would translate as a middle).
I would add a contextual factor for consideration. When the 'paragraph' is summed up in v. 33 (contrast the second-personal plural imperative in 6:1), the commands are framed in the third person, one third-person imperative, one noteworthy instance of the conjunction hina marking a third-person subjunctive as jussive/imperatival.
That said, v. 25 has a second-person plural imperative...
Thanks, Patrick. So, are you inclined to go against WH then?
Delete