More on Rethinking the Textual Transmission of the Gospels
In earlier postings, I’ve questioned the paradigm/model of an early “wild” period of textual transmission of the Gospels and a subsequent/later period of textual stability (here, here, and here). The model may seem intuitively credible, but the manuscript evidence doesn’t seem to support it. To cite one thing, the manuscript evidence suggests that significant textual variants continued to appear, or at least first became widely accepted, well into the Byzantine and early Medieval periods. Here are some examples.
Take the sizeable variant known as the pericope of the adulteress.[1] The passage first appears in the extant manuscripts in the fifth century (e.g., Codex Bezae), and not in earlier manuscripts (e.g., P66, P75, Codex Vaticanus, Codex Sinaiticus, as well as a number of somewhat later manuscripts). But thereafter the passage became a regular part of the Gospel of John at 7:53-8:11. So, the “success” of this passage, in terms of widescale incorporation into the Gospel of John, came, not in some early “wild” period, but later, in the period of supposed textual stability. The variant may well have first appeared early in some now-lost early manuscripts, but obviously did not enjoy that “success” till later.
Or consider the “long ending” of the Gospel of Mark (which became 16:9-20 in the Medieval text). Due to the comparatively less frequent usage and copying of Mark in the earliest centuries, the first Greek manuscripts that allow us to check the matter are Codex Vaticanus and Codex Sinaiticus, which don’t have these verses. Again, to find the variant in the manuscript tradition we have to go later, to the fifth/sixth century, in Codex Alexandrinus, Codex Bezae and others. And once more, this sizeable variant became thereafter solidly a part of the Medieval text of Mark. So, another sizeable variant that may have featured in some early manuscripts but only became a “success” later.
Still more striking, consider the variant at 1 John 5:7-8. The earliest Greek manuscripts with these words (or variations on them) are a small number from the Medieval period. Yet the variant became part of the “Textus receptus” (and so part of the KJV) thereafter.[2]
Now if we combine these facts with studies of the textual character of our earliest Greek manuscripts (papyri commonly dated to the third century or earlier), there is further reason to question the model of early/wild and later/stable textual transmission. Studies by Barbara Aland and Min on the papyri for the Gospel of Matthew, and the more recent and more sophisticated study of early papyri of the Gospel of John by Lonnie Bell, all show that these texts were transmitted basically with care—no “wild” variants in evidence.[3]
The major points to make are these: (1) The process of producing and incorporating significant and intentional textual variants continued well into the supposed period of textual stability. Indeed, it appears that factors (yet to be identified) may have made for the “success” of such variants in the later period, in contrast with the earlier centuries. (2) The earliest extant Greek evidence does not support the notion that in the first centuries the text of the Gospels was transmitted in some free-wheeling “wild” manner in which variants were introduced freely, producing some highly fluid state of affairs. The earliest witnesses do show variants, to be sure, although almost entirely those relatively small ones incurred in the process of copying, not the larger and intentional variants, and clearly some copyists were more skilfull than others, and followed somewhat distinguishable practices.[4]
The larger point is that there is much to be explored about the textual transmission of the Gospels all across the first several centuries down through the Byzantine and early Medieval periods. What factors made for the “success” of major variants only in these later times? What factors made for a relatively cohesive transmission of Gospel texts reflected in our earliest Greek evidence? Time for a more data-based model/paradigm!
[1] Chris Keith, “Recent and Previous Research on the Pericope Adulterae (John 7.53–8.11),” Currents in Biblical Research 6 (2008): 377-404; David Alan Black and Jacob N. Cerone, eds., The Pericope of the Adulteress in Contemporary Research (London: Bloomsbury T&T Clark, 2016).
[2] On this and the other variants mentioned in this blog, see, e.g., Bruce M. Metzger, A Textual Commentary on the Greek New Testament, Second Edition, 2nd ed. (Stuttgart: Deutsche Bibelgesellschaft/United Bible Societies, 1994), and the references cited at the respective places.
[3] Barbara Aland, “Das Zeugnis der frühen Papyri für den Text der Evangelien: Diskutiert am Matthäusevangelium,,” in The Four Gospels 1992, ed. F. Van Segbroeck et al. (Leuven: Leuven University Press, 1992), 325-35; Kyoung Shik Min, Die früheste Überlieferung des Matthäusevangeliums (bis zum 3./4. Jh): Edition und Untersuchung, ANTF 34 (Berlin: Walter de Gruyter, 2005); Lonnie D. Bell, The Early Textual Transmission of John: Stability and Fluidity in Its Second and Third Century Greek Manuscripts, NTTSD 54 (Leiden: Brill, 2018).
[4] E.g., James R. Royse, Scribal Habits in Early Greek New Testament Papyri, NTTS 36 (Leiden: Brill, 2007).
Comments are closed.
Larry Hurtado's Blog 
Sorry, I didn’t realise that a long version of Mark’s 16th chapter was dated so early. I misunderstood the thrust of the post. I thought it was saying that ‘look, later it had occasional “wildness” too’, and thankfully we can track that, account for it and revert away from it.
I’m not sure if my other comments have been dealt with but if you want to leave it there then ok. Fascinating reading as ever, and appreciation for all the interaction this post has clearly generated.
Larry, it doesn’t occur to you that the obvious answer as to why there are few changes to the NT is because they were created that way in the first place. This means that the original manuscripts written (created) during the second/third centuries, as Brent Nongbri implies.
No, Geoff. Neither I nor Nongbri posits the Gospels created in the 2nd century! For heaven’s sake, do read more carefully! Also, there are certainly many variants in the early manuscripts, but of the sort that arise from hand copying. There were likely also others that don’t survive in the extant evidence, and so must not have been all that popular. But I’d guess that no amount of reason or evidence would move you from your peculiar pet theory. So, just drop the conversation as there’s nothing to be gained for you or others.
Thank you Larry, that clarifies things a lot; and I find your terminology very helpful.
What seems to have been established in empirical studies of late 2nd and 3rd century papyri, is that the normal process of copying in Egypt in this period was able to sustain a ‘strict’ text alongside various (and fewer) ‘wilder’ texts. Do I have this right? From this I understand you as proposing to generalise that copying in the earlier period (for which we have no ‘dated’ papyri); and copying in other major centres of Christian book production (Rome?, Antioch?, Caesarea?) will similarly have sustained a ‘strict’ text. If so, I would agree. What cannot be established is how far the strict text copied in Caesarea may have differed from the strict text copied in Alexandria; and whether any variation between these strict texts was ‘major’ or ‘minor’. So it is not impossible that the strict text copied in Lyon contained the longer ending of Mark.
Which, to my mind, established two questions for further research.
– where, how, and at what stage did the ‘major’ variants originate; if we no longer postulate a period of ‘wild’ copying?
– where, how, and at what state did so many ‘major’ variants not found in the successful text shared by Vaticanus and Sinaiticus become present in the successful text a hundred years later in Alexandrinus?
I take it as read that, given the investment in material and scholarship to create a pandect bible, the text of a pandect will always be ‘successful’ for its time (or otherwise maybe officially intended to be successful for the future). Further I understand that , although Vaticanus and Sinaiticus differ a great deal in minor variants in the text (the sort that could result from slips in copying); they differ very little in ‘significant and intentional’ variants. Again am I right in this?
But, in the gospels, we then find many differences in these ‘significant and intentional’ variants when we come to Alexandrinus. So one possibility is that in a hundred years the ‘successful’ text has changed. Another being that Alexandrinus stemmed from the ‘strict’ text of a different area.
May I propose as one possible evidential source for exploring these questions could be the Ethiopic Version. In 356 Constantius II attempted to bully the Christianised Ethiopic Axumite kingdom into returning their new bishop to Alexandria to be re-educated in Arian theology. The Axumite kings did not comply, but appear rather to have cut off all official and ecclesiastical contact; and to have ordered a complete translation of the whole Greek Bible into Ethiopic (Ge’ez); the gospel translations being complete by the last quarter of the fourth century. Given that the translatops had a rather limited command of Koine Greek, it is rarely possible to retrovert from the Ethiopic back to a particular Greek text, but on ‘significant and intentional’ variants this is less of a problem. And we do have the gospels in Ethiopic in two complete manuscripts of 6th century date; so we can be confident they witness a late 4th century text created outside of any central Byzantine ecclesiastical authority; translating texts collected in Egypt around 350.
Overall, in the synoptic Gospels this text tends to agree to a considerable degree with Alexandrinus, in John more evenly between Alexandrinus and Sinaiticus, though not with Vaticanus where this differs. But for ‘significant and intentional’ variants the picture is more mixed.
A few examples:
– the Ethiopic supports the longer ending of Mark, against Vaticanus and with Alexandrinus
– the Ethiopic supports Matthew 16 2-3 (red skies as signs of the times), against Vaticanus, but missing in Alexandrinus
– the Ethiopic supports Matthew 24:36 (nor the Son), with Vatricanus, missing in Alexandrinus
– the Ethiopic supports Matthew 27:49 (water and blood) with Vaticanus against Alexandrinus
– the Ethiopic supports the first phrase of John 5:4 (the angel at the pool) against Vaticanus and with Alexandrinus; but omits the second phrase.
– the Ethiopic omits the pericope of the adulteress at John 7:53, with all three pandect witnesses.
Since the Greek texts underlying the Ethiopic must have been near contemporaries with Vaticanus and Sinaiticus, that suggests to me that we can take it that gospel texts of ‘Byzantine’ character were circulating in Egypt at that time, reproducing many ‘significant and intentional’ variants. But that this ‘Early Byzantine’ gospel text should not be identified – particularly in respect of such variants, with the later ‘Majority’ text; nor indeed is it exactly the same as the successful gospel text to be found in Alexandrinus.
Tom: One clarification in order. The significant and intentional variants such as those I cited may well have originated earlier than their first attestation in the manuscript tradition, but it appears that it was only in the fifth century and later that they became firmly established as part of the text, “successful” in that sense.
Larry; thanks for the clarification.
If significant and intentional variants indeed originated earlier than their first attestation; do you have a view as to how they may have originated?
If such variants only became ‘successful’ in the fifth century; so you have a view as to how they may have become successful at this time – when they had not been so successful before?
Tom: As per previous postings from me, I think that most intentional variants arose in the process of reading/studying the text, not in copying it. Individuals made annotations, e.g., in the margins or interlinear spaces. Then when that copy was copied (often much later), the material was incorporated into the text, the copyist assuming that it was some sort of correction.
As to why some major variants appear to have “succeeded” only in the fifth century and later, I have no specific wisdom. The sense I get is that the emerging “Byzantine/medieval text” was perhaps shaped by a desire for fullness, giving a textual “berth” to variants deemed positively.
Another point may be worth making. If something like John 21 is added late then it is less likely to displace the original version of the text. In other words, the original, shorter version of the text will survive alongside the longer version. On the other hand, if the change is made *very* early then it might completely displace the original. But if the change *was* made very early, then it is more likely to have been made by someone who knew what he was talking about, someone who is still in touch with reliable tradition.
Could it be a case of “heads we win, tails we win”? Just a thought.
Not sure David.
All the Gospel manuscripts that survive are copies; most likely copies of copies of copies in the best case. So, for a reading or text not to survive, it only needs not to have been copied. We do not need to postulate a ‘purge’ of non-standard readings, only a preference in copying for readings that conform to the commonly recognised text for a particular copying centre.
Tom: The notion of “copying centres” sounds a bit like the old and discredited “local texts” theory. The early evidence that we have comes from various locations (albeit mainly in Egypt), and various copyists, and both public and private uses. That none of the major textual variants appears suggests, not that they weren’t already around, but that they hadn’t yet become as widely accepted as they became later.
Tom, the fact remains that the easiest way to corrupt a text is to have the copying done in one place, and the hardest way to do it is to introduce a change in one manuscript tradition which then needs to displace the original form in numerous other manuscript traditions.
Thanks for the clarification on ‘local texts’
I suspect I have been focusing on the slightly later period – specifically the first half of the fourth century. We do have two complete pandect witnesses from this period to the text of the Gospels, in Vaticanus ans Sinaiticus; but we also have the evidence of three versions most likely translated from Greek manuscripts copied in this period – the Ethiopic, Gothic and Peshitta Syriac. In a number of major textual variants (e.g, Matthew 16: 2-3) the readings found in these versions (where it survives) consistently differs from that in the two contemporary pandect witnesses. The Ethiopic version is particularly valuable in this respect as it can be dated to a fairly narrow window.
Following on from David Madison’s comment, I’d like us to consider a more neutral position here. If we follow through on Dr. Hurtado’s point, which seeks to correct misinformation about the weighting of scribal insertions, we have to think a bit more about methodology. If the strategy is to point out that some significant insertions happened in the so-called stable period and that a diminuendo is thus proven false, we are left with a problem. In fact, this post, to my mind is unfinished (I suppose that comes part and parcel with the medium, blog posts are essentially conversation, which is why they are so precious). Where it leads me is to ask: do we want to imply that *therefore* significant alterations happened to a low degree throughout Christian manuscript history? That should be considered a pretty problematic angle for Christianity. David, you seem to be wanting to say “no” to this option (back to that in a second).
The only alternative angle I can see that Dr. Hurtado’s post gives us, and I don’t for a moment think you are saying this Dr. Hurtado, is that alterations of this kind occurring in a later period (e.g. Byzantine) somehow imply that they weren’t also occurring much earlier (1st-3rd C.).
Put this plainly, I find the post both reassuring and disturbing. I want to believe (as a Christian, no shame there) that the critical text is super-solid (and I do believe it is very solid!). At the same time, I find no basis to think that earlier copyists and scribes (yes, I am sorry I don’t see quite such a universal separation there, even if some distinction in places, times and function is helpful) would have avoided such alterations *more* than in later times.
So to return to your point, David, if I have heard you right, you want to argue against early corruption as *wholesale* corruption. *Wholesale* isn’t worrying for studies of early Christianity, nor is *wild*, and scholars have reacted well to correct that. What should be of concern to us is quite simply to concede that *some* corruption may have always been integral to copying. We can more easily affirm this for later surviving manuscripts than the earliest, perished ones. The perished variety are by their very nature, forever unverifiable. I have to concede that this is what we would probably expect in terms of the evidence.
I close with a reminder that “corruption” is not meant in the negative, common sense, even if some have equivocated here. For readers who have taken the time to read Ehrman’s Orthodox Corruption of Scripture, combined with the ideas of Hermeneutic Circle dynamics such as discussed by French philosopher, Paul Ricoeur, I believe they will feel quite unperturbed by it. Christians have always believed in truth, I think, but mostly one that transcends and runs deeper than individual words and sentences, which cannot exist in a vacuum and ultimately existed in human form. Ehrman’s point is that when the custodians of the text were confronted with various heresies, the proto-orthodox church felt *obliged* – in their deep commitment to that truth – to act in order to prevent it being misconstrued by their opponents. Such action included “textual action” and probably always has.
JB,The existence of variants might seem to be a problem, but, as I understand it, there are reliable methods for reconstructing the original text. As long as the original form of a verse exists somewhere among the manuscripts and we are able to identify it, we don’t have to worry. But what if the original forms *don’t* exist anywhere in existing manuscripts? That would be a problem. But how likely is that? Christian texts were widely distributed at an early stage and copies were being made in numerous locations. This would lead to textual variants but the original form should be preserved somewhere.
As I understand it, David; the most widely held scholarly opinion on the Gospel of Mark, is that the original ending was lost; and, as yet, has not been found.
Tom (and David): I think that the majority view nowadays is that GMark originally ended at 16:8. See, e.g., Adela Yarbro Collins, , Mark: A Commentary, Hermeneia (Minneapolis: Fortress Press, 2007), 125.
The problem we face (together), is that we have no way of knowing if we are reconstructing *throughout* a set of original texts. This is not, as I see it, an either/or. It is not either we have a reliable critical text of the original texts or we don’t. The question your post raises, whether we like it or not (and that is why I said I found its implications to be both reassuring and disturbing as a Christian) is whether we have a solid reconstruction throughout. It would only take a few “scribal” insertions or alterations of super early manuscripts lost to us to maintain the balance of later centuries (on the assumption that there is an equal balance). This could and I am willing to grant *would* grant a text that is in the high 90s of reliability with respect to lost originals.
The debate about early wildness doesn’t stand. However, what we would love to claim but can’t, is that there are likely no (ie ZERO) significant alterations prior to the current state of traceability, ie manuscript evidence and textual-critical work in 2018.
John: I’m not making the silly claim that you pose. There is good evidence that the larger variants (e.g., endings of Mark) were known as early as the second century. But it appears that they weren’t as widely accepted or copied as they became in the fifth century and later.
John, you talk about the possibility of insertions in “super early manuscripts” but in that case the sceptic is already on the back foot. It isn’t easy to think of ways in which the texts could be corrupted without leaving any trace. Of course, we can’t completely rule out the possibility of corruption but, as you indicate yourself, there are good reasons to be confident overall.
Thanks David.
Yes, I am confident in the strong reliability of the critical text as established now. Unfortunately and fortunately, I feel confident also that the quality of our critical text is still set to improve in small ways, maybe very small ways, but improve nonetheless. In addition to the point I made earlier (based on logical argument), this improvement process points to the impossibility of knowing what the text *will* say in 2100 and what it (or rather they) said exactly in 100. I disagree with you that say a first generation of copy error (and remember that all copies contain changes) particularly to a more robust and popular copy would necessarily be traceable, although am glad of LH’s historical point of lifespans of manuscripts exceeding this first “generation”). Ie the orthographs may have held corrective power over some of the corruptions/errors inevitably made in those first copies. All?!
Thanks for this, Larry. I agree about the importance of looking into later transmission history to think about the success of some of these larger variation units (and I’m looking forward to reading the new book on the woman caught in adultery by Knust and Wasserman). Just one note on the “longer ending” of Mark: I know you’re talking here about our surviving manuscripts, but to say that it “may have featured in some early manuscripts” seems a little overly cautious. Irenaeus (Against Heresies 3.10.6) cites from the longer ending and attributes it explicitly to Mark, which indicates its presence in at least one manuscript already in the late second century.
Brent: Sorry if my “may have featured” seems “overly cautious.” I didn’t intend to deny it, only that the “long ending” didn’t “succeed” apparently till later.
You seem to be arguing that textual fluidity in later periods means that the early period was not as fluid as supposed. But that doesn’t follow. Evidence concerning later transmission has no bearing on the earliest period, unless I have missed something?
As for the manuscript evidence not favouring early wild variation, well there is little or no manuscript evidence for the very earliest period. So there is little or no direct manuscript evidence one way or another. What there is are internal, contextual and social reasons for supposing the earliest period of textual transmission relatively fluid.
No, Donald. I’m not arguing what you wrongly infer from my posting. What I mean is what I said: Major variants (which may well have originated early, who knows) succeeded more evidently in the period when things are often thought to have been stablized and fixed textually, and the extant evidence of earliest mss doesn’t actually reflect a wildly fluid textual transmission. If there had been such a status at an earlier point, then surely there would be some trace of it in earliest mss. Your unspecified “internal, contextual and social reasons” for “supposing” otherwise, are only suppositions. In fact, as I and others have shown, there are actually forces that would have made for limits to the textual fluidity in texts treated as scripture in the second century. See, e.g., L.W. Hurtado, “The New Testament Text in the Second Century: Text, Collections, and Canon,” Transmission and Reception: New Testament Text-Critical and Exegetical Studies, eds. J.W. Childers & D. C. Parker (Piscataway, NJ: Gorgias Press, 2006), 3-27. The pre-publication version appears on this site under the “Essays” tab.
“If there had been such a status at an earlier point, then surely there would be some trace of it in earliest mss.”
How can there be MS evidence for wild variation in the second century when there are hardly any MSS from the second century to begin with? It’s like saying that no photograph supports the idea that Napoleon was short, when there are no photographs of Napoleon at all because he was a few years too early for photographs.
What we do know, by reading the documents themselves, is that 2 Corinthians, John 21, and other parts of the New Testament indicate significant revision in the early life of the text.
Donald: Gee whiz! Do try to follow the reasoning better. My point should be clear. “Earliest mss” = not only the very few 2nd century ones but also 3rd century ones. There is no trace of some wild process of copying in these mss. On your assumptions (and they’re only that), all traces of a supposedly wild state of textual affairs were somehow lost, eliminated. But in a time when there was no ecclesiastical control over the copying process, that’s actually very hard to imagine.
Your final sentence, yet once again, confuses suggestions about an early *redaction/editing* of certain texts *before the copying process* and as distinguished from that copying process. That 2 Cor, for example, may have been put into its familiar form at some early redactional stage says nothing about how well it was copied thereafter. I do wish that you would drop your hostile attitude, and try to follow the discussion more carefully. Deal?
Professor,
Your studies in this area have helped me in several ways. I have been one of the “wild” advocates and can see your point clearly.
However, why is 1 John 5:7-8 (5:7b-8a) applicable? as this passage never made it into the mainstream of even the Byz. text. Is not it found in only 1% of the extant Gk. mss. of this passage?
MANUSCRIPTS WITH 1 JOHN 5:7-8 IN THE TEXT:
ms. # 629 (Codex Ottobonianus) – fourteenth century
ms. # 61 (Codex Montfortianus) – sixteenth century
ms. # 918 (in Spain) – sixteenth century
ms. # 2318 (in Romania) – eighteenth century
MANUSCRIPTS WITH 1 JOHN 5:7-8 LATER ADDED TO THE MARGIN:
ms. # 221 (in Oxford) – tenth century
ms. # 88 (in Italy) – twelfth century (sixteenth century hand)
ms. # 636 (in Italy) – fifteenth century
ms. # 429 (in Wolfenbuttel) – fifteenth century
Corrections and oversights are welcome.
Ron: I offered 1 John 5:7-8 as an example of how a variant that appeared only later was “successful”, in that it became part of the “textus receptus” and so the translations based on it, such as the KJV.
A reader might wish to change or “improve” a text but there is an obstacle to anyone who sets out to do this, and that is the continued existence of the original form of the text. If you want your version of the text to prevail, then you need all copies of the original to disappear. And the more copies there are of a text, the less likely that is to happen. If there is a theory of the early production of Christian texts which would allow wholesale corruption to occur while the originals conveniently disappear, I would like to know what it is.