Friday, January 30, 2015

The Text of Reasoned Eclecticism: Is It Reasonable and Eclectic? Part Two of a Four-Part Response to Dan Wallace

External Evidence

            Wallace began his consideration of what external evidence implies about the Byzantine Text with a simple misrepresentation.  He stated that the primary premise in the Byzantine Priority view is, “Any reading overwhelmingly attested by the manuscript tradition is more likely to be original than its rivals,” he added, “In other words when the majority of manuscripts agree, that is the original.”  But somewhere in those other words, the word “overwhelmingly” was murdered – sacrificed for the sake of caricature-drawing.  A majority might be 50.1%.  The overwhelming support for Byzantine readings that Wallace routinely vetoes in favor of Alexandrian readings include majorities of 85%, 98% and 99% and higher proportions of the manuscript-evidence. 
            Wallace says, “In historical investigation, presumption is only presumption.”  But when facing textual variants in which 99% of the Greek manuscripts disagree with the Alexandrian reading, we do not face mere presumptions; we face implications:  either a corruption permeated 99% of all Greek manuscripts, or else a corruption was adopted in 1% of them.  In general, which is more likely:  that many copyists in many places created or adopted a non-original reading, or that few copyists in few places did so?  The original text is the text with the head start, so to speak, and if one is to posit that a non-original reading overtook it so as to become more popular, one must explain how that happened – and this is not easy to do without making some assumptions, or presumptions.  So let the axiom, “Presumption is only presumption,” be aimed at all presumptions, not just those that maintain Byzantine Priority. 
            Moving along.  Wallace claimed that if the Byzantine Text were the original text, then one would expect to find it “in the earliest Greek manuscripts, in the earliest versions, and in the earliest church fathers,” and “One would expect it to be in a majority of manuscripts, versions, and fathers.”
            “But,” he continues, “that is not what is found.”  However, by definition, the majority text is what is found in the majority of manuscripts (at least in passages where a majority exists, rather than a split-decision among three or more rival variants).  Wallace attempts to circumvent this obvious fact by redefining the majority as the majority of long-lived manuscripts.  “As far as the extant witnesses reveal,” he claims, “the majority text did not exist in the first four centuries.” 
            Considering that it comes from someone attempting to avoid presumptions, that is an extremely presumptive claim.  Other than the papyri from Egypt, there is not much New Testament manuscript-evidence to indicate what texts were being used throughout the Roman Empire before the year 400.  The available manuscript-evidence is not remotely close to being extensive enough to justify statements about what the majority of manuscripts read in the second or third centuries, at points where the testimony of the extant evidence is diverse.  To presume that the manuscript-evidence from Egypt depicts the text that was used in other locales is a huge presumption.  By the year 235 or so, Origen stated that the manuscripts were in disagreement with each other.13  That is difficult to reconcile with the idea that a uniform Alexandrian Text, or any text-type, was an established standard text at that time in a multitude of non-Egyptian locales. 
             When Wallace appeals to the papyri as vindication for his idea that the Byzantine Text did not exist in the first three centuries of Christendom, he states, “More than fifty of these came from before the middle of the fourth century. Yet not one belongs to the majority text.”  That is not quite true.  Papyrus 104, which currently contends with Papyrus 52 for the claim of earliest-known-New-Testament-manuscript, is a fragment of text from Matthew 21 that agrees with the Byzantine Text as much as it agrees with the text of Codex Vaticanus.14  In addition, large portions of Codex W (from the late 300’s or early 400’s) display the Byzantine Text.15  
            Papyrus 98, a fragment from the late 100’s or early 200’s which contains text from Revelation 1:13-2:1, disagrees twice with the Byzantine Text (though in one of these two instances, the Byzantine Text is divided), and disagrees once with the Nestle-Aland text – so it is rather presumptive to say that the fragment favors one text-type significantly more than the other, especially since P98 disagrees with Codex Sinaiticus five times.  Papyrus 16, a fragment from the 300’s with text from Philippians, diverges from the Byzantine Text eleven times, but it also disagrees with Codex Vaticanus nine times.  The much-mutilated Papyrus 45, which is currently the earliest known manuscript of the Gospel of Mark, from the early 200’s, agrees much more closely with the text of Mark in Codex W than with the text of Mark in Codex Vaticanus, and in Mark 7, P45 agrees repeatedly with the Byzantine Text.  Papyrus 46 also frequently disagrees with the Nestle-Aland text.16
            Wallace can’t have it both ways:  if the mere existence of a non-Byzantine local text displayed in the early papyri constitutes strong evidence that the Byzantine Text did not exist, then the existence of a non-Alexandrian text in the early papyri (such as what is seen in P45) constitutes strong evidence that the Alexandrian Text didn’t exist.  Obviously this sort of reasoning is an overextrapolation, since witnesses such as P75 show that the parts of the Alexandrian Text that they contain did exist in the 200’s.  The non-existence of Byzantine papyri in Egypt does not imply the non-existence of Byzantine papyri in other locales, just as non-Alexandrian readings in some Egyptian papyri do not imply the non-existence of Alexandrian readings in other Egyptian papyri.
            Wallace stated, “Many hypotheses can be put forth as to why there are no early Byzantine manuscripts.  But once again an ounce of evidence is worth a pound of presumption.”  As if there is some mystery about it!  Until the 300’s, New Testament manuscripts were made of papyrus, which decomposes in virtually every climate except for the low-humidity climate of Egypt.  This is not a presumption; it is a scientific fact.  High humidity was even more systematic and thorough than the Roman persecutors who destroyed Christian manuscripts during the Diocletian persecution. 
            This is a perfectly reasonable explanation why we have New Testament manuscripts – and fragments of the works of Homer, and Greek poetry, and tax-receipts, etc. – from Egypt, and not from very many other locales, from the 200’s and 300’s.  It’s not as if Christians in other locales were not reading the Gospels, or reading the Iliad, or writing receipts and letters.  The unique climate of Egypt is the factor that has resulted in the preservation of papyrus documents there.  And they provide a fairly good sample of the texts that were used in Egypt (especially at Oxyrhynchus).  But they cannot do the impossible; they do not tell us what sort of New Testament text was in use elsewhere. 
            Wallace moves on to consider the early versions.  He correctly points out that the Old Latin evidence is consistently Western (where its witnesses are not barnacled by Vulgate readings, at least).  However, some Old Latin witnesses habitually collide with other Old Latin witnesses; there was not one monolithic Old Latin tradition; there were, instead, numerous independent Latin versions, as Jerome indicated in his preface to the Vulgate Gospels.  The extant Old Latin manuscripts are samples from that collection.  And while the Old Latin texts are not Byzantine, they agree with the Byzantine Text much more than they agree with the Alexandrian Text.  This is no more proof of the non-existence of the Byzantine Text than it is proof of the non-existence of the Alexandrian Text. 
            The Coptic version, Wallace states, “goes back to an early date, probably the second century.”  There was not just one Coptic version.  What we have are New Testaments (or at least portions of the New Testament) in several Egyptian dialects, displaying several different forms of the text from different areas and different eras:  Sahidic, Bohairic, Achmimic, Sub-Achmimic, Middle Egyptian, and Fayyumic.  Part of the Sahidic version is strongly Alexandrian, but the collective testimony of the Coptic versions is very far from a uniform endorsement of the Alexandrian Text; the Coptic Glazier Codex (CopG67), for example, displays a thoroughly Western text of Acts.  In addition, one should consider that the second-century origin of the Sahidic version is a calculated guess.  Due to the uniformity of Coptic lettering across centuries, the production-dates of Coptic manuscripts are notoriously difficult to specify on a paleographical basis.    
            Wallace next turns his attention to the Gothic version, which he, in agreement with Metzger, affirms to be the earliest representative of the Byzantine Text.  However, he blurs its production-date, stating that it was produced “at the end of the fourth century,” i.e., the late 300’s.  The Gothic version was produced by Wulfilas, who was appointed to be a bishop in 341; he undertook his translation-work shortly after that, in the mid-300’s, that is, at about the same time Codex Sinaiticus was produced.
            What does the existence of these early versions imply about the Byzantine Text?  Wallace proposes two implications.  First, he proposes that “If the majority text view is right, then each one of these versions was based on polluted Greek manuscripts.”  As far as the Old Latin versions are concerned, that sword cuts both ways:  advocates of the Alexandrian Text consider the Old Latin versions’ texts to be thoroughly corrupt.  Wallace has no right to treat this as a problem, since he believes it too.    
            Second, he proposes that the early versions represent the texts used in a wide variety of locales:  “the Coptic, Ethiopic, Latin, and Syriac versions came from all over the Mediterranean region. In none of these locales was the Byzantine text apparently used.”  (Did you see how he just threw the Ethiopic version in there, as if it existed before the 300’s?)  I challenge Wallace to name a single Byzantine reading anywhere in the Gospels that does not have some support from one or more of these versions.  The Syriac Peshitta version, in particular, exhibits strong alignment with the Byzantine Text.  And neither the Sinaitic Syriac nor the Curetonian Syriac displays an Alexandrian Text; put either one alongside the text of Vaticanus or Sinaiticus and you will observe a plethora of disagreements.  So when Wallace’s sentence is filtered by reality, and only versions from before the 300’s are in view, this is what survives:  “The Sahidic, Latin, and Syriac versions came from all over the Mediterranean region.  Only in the Sahidic version was the Alexandrian Text apparently used.”   
            The early Sahidic version did not come from all over the Mediterranean region.  It was a local text.    
            Rather than constituting “strong evidence that the Byzantine text simply did not exist in the first three centuries—anywhere,” the Old Latin evidence that Wallace has called to the stand testifies that diverse forms of the Western Text were used as the basis for Latin translations.  They are no more anti-Byzantine than they are anti-Alexandrian.  And, I note in passing, that in many cases, the Old Latin aligns with the Byzantine Text and not with the Alexandrian Text.  (This alignment provides pro-Alexandrian critics with an excuse to see only a few early Byzantine readings:  when a Byzantine reading agrees with a Western witness, the reading is categorized as Western.) 
            Again:  the only early version with a strongly Alexandrian Text is the Sahidic made-in-Egypt version.  Thus, what Wallace has in the early versions – even when the Gothic version and the Peshitta are set aside – is not evidence that the Byzantine Text did not exist anywhere.  The evidence does not come remotely close to warranting such a sweeping conclusion.  It implies, rather, that by the time the Old Latin versions were made, the Western Text had already developed, and that by the time the earliest strata of the Sahidic version was made, the Alexandrian Text had developed in Egypt.  It does not, and cannot, inform us about the text that was being used in other locales.

            What about the early patristic writers?  Wallace affirms that “Many of them lived much earlier than the date of any Greek manuscripts now extant for a particular book.”  His readers could easily get the impression that many patristic authors before the year 300 wrote so extensively that researchers can confidently observe what text-type they used in their utilizations of the New Testament.  However, in 1881, Hort wrote,

            “The only extant patristic writings which to any considerable extent support Pre-Syrian readings at variance with Western readings are connected with Alexandria, that is, the remains of Clement and Origen, as mentioned above (§ 159), together with the fragments of Dionysius and Peter of Alexandria from the second half of the third century, and in a certain measure the works of Eusebius of Caesarea, who was deeply versed in the theological literature of Alexandria.”17

            In other words, if one searches through all the known patristic literature produced before 300, the only places one finds substantial agreement with non-Byzantine, non-Western forms of the New Testament are in the writings of a few individuals who were linked to Egypt either geographically (Clement of Alexandria lived in Alexandria, of course, and Origen worked there prior to moving to Caesarea around 230) or in terms of training (Eusebius of Caesarea was a fan and defender of Origen).  
            Even in some of the writings of individuals who were either in, or from, Egypt – where one would naturally expect the local Alexandrian Text to exert the most influence – there is as much evidence for the Byzantine Text as for the Alexandrian Text.  In Carl Cosaert’s analysis of the Gospels-text used by Clement of Alexandria, Cosaert listed 125 utilizations of the text of the Gospel of Luke in which Clement’s text agrees with either one or two members of a group consisting of Codex Vaticanus, Codex Bezae, and the Textus Receptus.  Clement’s text, according to Cosaert, agrees with D 62 times (49.6%), with B 65 times (52%), and with the Byzantine Text 68 times (54.4%).18  This is not evidence that Clement used a manuscript of Luke that closely resembled the Byzantine Text – but it is evidence (contrary to what Wallace is trying to show) that Clement’s support for the idea that the Byzantine Text of Luke was non-existent when he wrote is not any greater than Clement’s support for the idea that the Alexandrian Text of Luke was non-existent when he wrote.
            None of the patristic research conducted in the last 80 years has shown that any writers outside Egypt and Caesarea prior to the year 300 used the Alexandrian Text.  There simply are not “many” patristic writers before 300 who wrote enough, and cited the New Testament enough, to clearly show that they favored the Alexandrian Text.     
            Wallace’s claim that “The early fathers had a text that keeps looking more like modern critical editions” does not accurately describe a single early patristic writer outside the borders of Egypt and Caesarea.  It does not even accurately describe the text used by Clement of Alexandria.  In some cases, the texts used by early patristic writers look more Byzantine than Alexandrian.  For example, the Alexandrian Text does not contain Mark 16:9-20, but utilizations of Mark 16:9-20 are found in the writings of Justin Martyr (160), Tatian (172), Irenaeus (180’s), Hippolytus (220’s), and the pagan author Hierocles (305), who was very likely recycling material composed by Porphyry in the 270’s.  Matthew 17:21 is another example of a non-Alexandrian reading supported by a patristic author who is supposed to have an Alexandrian text:  this verse was cited by Origen, but the entire verse is absent in the Alexandrian Text.19 
            Wallace presents the Greek manuscripts, the early versions, and early patristic quotations as “a threefold cord” of testimony.  But in reality those three things do not come together:  the early Egyptian papyri display the texts used in Egypt.  The texts in the Old Latin versions and the Sinaitic Syriac and Curetonian Syriac are not Byzantine, but they are certainly not Alexandrian either.  Early patristic evidence from outside Egypt-and-Caesarea does not support the Alexandrian Text, and even some writers in Egypt and Caesarea support readings that are in the Byzantine Text, against the Alexandrian Text. 
            Everything we see in these three forms of evidence – manuscripts, versions, and patristic utilizations of the New Testament – indicates that the New Testament text was disseminated in localized forms.  And a local text, while capable of showing us what text was used in a specific locale, does not show us what text was being used in another locale hundreds of miles away.  We don’t look at an early Old Latin manuscript such as Codex Vercellensis and conclude that it reveals the local text of Alexandria.  Nor do we look at the Sinaitic Syriac and conclude that it reveals the local text of southern Italy.  But Wallace apparently wants us to look at a local Greek text of Egypt, and very different local Latin texts from who-knows-where – possibly also from Egypt, in some cases20 – and conclude that they reveal the local texts of Antioch, Asia, Cyprus, Edessa, Nicomedia, and the cities of Greece
            What text of the New Testament was being used by Christians in that vast territory before the year 300?  We do not know:  manuscripts, versions, and substantial patristic writings from that area, in the ante-Nicene era, are not extant.  But in the 400’s, the Byzantine Text was the Greek text that was in use in the Greek-speaking churches in these areas, and the Peshitta was the Syriac text that was in use in the Syriac-speaking churches.  In addition, we observe that
            ● the Gospels-text in Codex Alexandrinus (from around 400) is mainly Byzantine,
            ● portions of the Gospels in Codex W (from the late 300’s or 400’s) are Byzantine,
            ● Basil of Caesarea (330-379) used a text of Matthew that was primarily Byzantine,21
            ● The texts of John and the Pauline Epistles used by Gregory of Nyssa (335-395) agree more with the Byzantine Text than with any other text-type.22

            Before moving on to address Wallace’s comments about internal evidence, there is one more of his claims about external evidence that invites a response:  he stated that some patristic statements show that what is the majority-reading now was not the majority-reading when those statements were made.  Wallace provided only two specific examples of this:  (1)  Jerome’s statement (in Ad Hedibiam) that Mark 16:9-20 “is met with in only a few copies of the Gospel – almost all the codices of Greece being without this passage,” and (2) Jerome’s statement that at Matthew 5:22 “most of the ancient copies” do not contain εικη.   Regarding the first example, I believe that anyone who takes the time to compare Jerome’s comments in Ad Hedibiam to Eusebius’ comments in Ad Marinum will conclude that the part of Jerome’s composition in which this statement is found is essentially a loose recycling of Eusebius’ material; in the course of answering Hedibia’s broad question about how to reconcile the Gospels’ accounts of events after Christ’s resurrection, Jerome utilized three of Marinus’ specific questions on the subject, as well as three of Eusebius’ answers, in the same order in which they appear in Ad Marinum
            This should provide some instruction about the high degree of caution that should accompany patristic references to quantities of manuscripts.  In some cases, such as we see in Ad Hedibiam, the claim may have been borrowed second-hand from a source who was describing manuscripts in a different time and place.  In other cases, it may indeed reflect what the author has encountered, but it would be quite a leap to conclude that what the author encountered is what one would encounter when surveying all manuscripts everywhere that were contemporary to him.  In other words, there is no justification for the assumption that a reading found in the majority of manuscripts known to a specific author would also be found in the majority of manuscript that were not known to that author.  When we approach a statement about manuscripts that an author knows about, that is what we should understand it to be – not a statement about manuscripts about which the author knows nothing.  There is no necessary correlation between the contents of majorities of manuscripts known to Origen, or to Eusebius, or to Jerome, and the contents of actual majorities of manuscripts at the time of Origen, or Eusebius, or Jerome.  This point seems to have completely eluded Wallace.

- Continued in Part Three - 



13 – See Metzger’s quotation of Origen on page 88 of New Testament Tools & Studies – Historical and Literary Studies, Vol. 8 (1968), at the outset of his article, Explicit References in the Works of Origen to Variant Readings in New Testament Manuscripts:  “The differences among the manuscripts have become great, either through the negligence of some copyist or through the perverse audacity of others,” etc.
14 – A thorough description of P104 is among the files of the NT Textual Criticism group on Facebook.
15 – Specifically, Codex W is essentially Byzantine in Matthew and in Luke 8:13-24:53.  In Mark 5-12, it seems to be loosely, and uniquely, aligned with the text found in P45.  The text of Luke 1:1-8:12 and John 5:12-21:25 is essentially Alexandrian.  This block-mixture shows that it was possible for rival text-types to exist side-by-side in the same locale.
16 – See the comments by Dennis Kenaga regarding P46 on page of Skeptical Trends in New Testament TextualCriticism:  “The oldest witness, P46, was rejected 303 times, 30% of the time, in 1 Corinthians.”  (The wording of this sentence could be improved, but the basic point is correct.)    
17 – See page 127 of Hort’s Introduction to the New Testament in the Original Greek.
18 – See the data in Carl P. Cosaert’s The Text of the Gospels in Clement of Alexandria, © 2008 by the Society of Biblical Literature.  A preview is at .
19 – It is feasible that Origen was quoting from Mark 9:29, but he tended to quote from Matthew much more frequently than from Mark.  Even if one were to grant that Origen was quoting Mark 9:29, the quotation is clearly not based on the Alexandrian Text, because Origen includes the words “and fasting,” which are not in the Alexandrian Text of Mark 9:29.
20 – See Metzger’s comment on page 37 of The Bible in Translation:  The Coptic versions of the Old Testament frequently show a relationship with the Old Latin version . . . . This is not surprising, because the Old Latin version is regarded as having been of preeminent importance for the African Church.”
21 – See Jean-Francoise Racine’s 2004 book, The Text of Matthew in the Writings of Basil of Caesarea.
22 – See James A. Brooks’ 1991 book, The New Testament Text of Gregory of Nyssa.      

No comments: