The Role of Tradition in New Testament Textual Criticism

            Yesterday, Joshua Gibbs of the Talking Christianity podcast hosted a round-table discussion on the subject of the role of tradition in New Testament textual criticism, with guests Jeff Riddle (representing a Confessional Bibliologist approach), Peter Gurry (representing a Reasoned Eclectic app, and myself (representing an Equitable Eclectic approach).  A looong discussion commenced.  Here it is in two parts.  (At one point my internet connection died, but then it got better.)

Part One:

And Part Two: 

Sit back, grab some popcorn, and enjoy!

Here is the text of my closing statement:

          In closing, I’d like to briefly consider three different approaches to the role of tradition in the compilation of the text of the New Testament. 

          One view is to look at how much agreement there is, along all transmission-lines, and conclude, “Textual variants do not matter.  Everyone agrees that you’ve got the basics of the gospel if you use the Textus Receptus.  So let’s use that, for the sake of stability.” 

          Another approach is to focus on the disagreements.  A person might say, “This is very complicated, and we just can’t tell what the original readings are.  The safest course of action is to just go on using what the church has traditionally used.” 

          If one defines “the church” in terms of what emerged from the Reformation, that approach will provoke the adoption of the Textus Receptus.  If one defines “the church” in a wider sense, the traditional Greek text is the Byzantine Text.  Ecclesiastical approval is on its side.

          But both of those approaches are basically appeals to authority:  authority in the form of tradition.  And an appeal to authority is not the same as an appeal to evidence.  A reading is authoritative because it is original – not simply because it is thought to be original.  Except for scribal blunders, practically all major readings were thought to be original by somebody; that is why they are in the manuscripts.  It takes more than being accepted by someone to vindicate a reading.

          When dogmatic statements are used instead of arguments from evidence, it’s like saying, “We have been using mumpsimus, so mumpsimus is what should be said.”  But readings do not become authoritative by being used.  An original reading is authoritative at the point of its inspired creation.   And scribal corruptions are never authoritative, because they are not inspired – no matter how many people like them.       

          There is also a third view, in which someone says, “If ecclesiastical usage is what endows a reading with authority, then all readings are valid, because they all have at least a little bit of ecclesiastical usage in their favor,” and this provokes a temptation to clutter the margins with a multitude of textual variants.  Not only does this render the text more unstable than ever, inviting readers to pick and choose, but it is the exact opposite of what textual critics are supposed to do, which is, make decisions about textual contests. 

          I suggest that tradition does have a valid role, though, in certain cases:

          ● if two competing textual variants both have strong external support, and

          ● they convey two different messages, and

          ● neither is shown by internal considerations to be non-original, and

          ● one or both readings says something that is not confirmed in other passages,

that is a situation that merits a footnote. 

          But which reading goes into the text, and which one goes into the margin?  After those qualifications are met, there is something to be said for the principle that possession is nine-tenth of the law.   If one reading consistently dominates the other, in terms of widespread and longstanding use, then, instead of having a relatively brief Council of Bishops to break the tie, we have a very long Council of Use.  This approach might not resolve every case, but it will help keep textual instability to a minimum, without giving tradition the right to veto the original text.

Talking Christianity Apologetics Podcast (Part 1)

           Yesterday I was interviewed by Joshua Gibbs at the Talking-Christianity Apologetics podcast, for a friendly discussion about New Testament textual criticism, the early history of the New Testament text, Equitable Eclecticism, the nature of the Byzantine Text, and some questions involving Mark 16:9-20 and John 7:53-8:11.  Despite a few gaffes on my part – at one point, I repeatedly called Eusebius “Erasmus,” and said “Mark” a couple times when I meant to say “John,” and momentarily forgot where the Pentecost-lection begins, and somehow put Irenaeus in the 200s instead of the 100s – I am happy with the overall result.
            Here’s the video of the podcast, which lasts a little more than an hour and 41 minutes.  Hopefully Part 2 will commence early next year.

Equitable Eclecticism - Part 2

(Continuing the presentation of a slightly updated version of my 2010 essay Equitable Eclecticism:  The Future of New Testament Textual Criticism)

Competing Analytical Approaches

In the Byzantine Priority view, Greek manuscripts which display the Byzantine Text are considered superior witnesses on the grounds that their text has a plausible transmission-history.  Pick any series of readings in the Byzantine Text, and it can be shown to have considerable manuscript support.  The Nestle-Aland compilation, meanwhile, is considered a “test-tube text,” because it often combines readings in a series that is unattested in any Greek manuscript.  And although it has been argued that this is unavoidably what one gets when selecting variants from among different text-types, the point remains that a heavy burden of proof should be upon the compiler whose work implies a transmission-history in which no copyists have preserved the original combination of readings in hundreds of passages.
On the other end of the spectrum, the approach used by Hort may seem like something very different from Byzantine Priority, but in terms of methodology the two approaches are similar:  Hort regarded a specific set of manuscripts as superior to all others (in this case, Codex Vaticanus and whatever allies Hort could find for it), and he built a transmission-model that vindicated its readings.  Having established Vaticanus as the best overall witness in a relatively small series of contests, Hort gave it enormous weight, with the result that its text just kept getting better and better, as more and more contests were decided by “the weight of the witnesses” – to the point that long segments of Hort’s compilation resemble transcripts of Codex Vaticanus.      

Two other approaches were developed by textual critics in the 1900’s by scholars aspiring to produce an eclectic text (that is, a text obtained via the utilization of a variety of sources).

Thoroughgoing Eclecticism (also known as Rigorous Eclecticism) values the relative intrinsic qualities of rival variants as the best means to determine their relationships, effectively rejecting Hort’s axiom.  In this approach, even if a reading appears exclusively in late witnesses, if its intrinsic qualities are judged to be better than its rivals, it is adopted, on the premise that its young supporters echo an older text – the autograph – at that point. 
Building on the theory that text-types did not stabilize until the 200’s or later, thoroughgoing eclectics resort to the only sort of reconstruction which can be undertaken without appealing to the relationships of text-types:  the relationships of rival variants.  Advocates of this approach tend to be more willing to introduce conjectural emendations, if an emendation possesses superior intrinsic qualities to its rival extant variants. 

Reasoned Eclecticism (also known as Rational Eclecticism), in theory, considers the relative intrinsic qualities of rival variants, but also considers the quality of each variant’s sources, their date, and their scope.  The text of the United Bible Societies’ Greek New Testament was compiled using a form of reasoned eclecticism.  However, in its companion-volume, A Textual Commentary on the Greek New Testament, Bruce Metzger’s comments show that the quality of sources tended to be measured according to Hort’s model of transmission-history.  In The Text of the New Testament, Metzger wrote, “Theoretically it is possible that the Koine text” – that is, the Byzantine Text – “may preserve an early reading which was lost from the other types of text, but such instances are extremely rare.”  This anti-Byzantine bias is pervasive.  It is no surprise, therefore, that the UBS text varies only slightly from Hort’s text, even though more evidence in favor of Byzantine readings is available to researchers than ever before.  (For more on this subject seem my four-part essay, The Text of Reasoned Eclecticism:  Is it Reasonable and Eclectic?)

  

An alternative is Equitable Eclecticism, in which the relative intrinsic qualities of rival variants are considered, and each variant’s sources, their date, and their scope are also considered.  Equitable Eclecticism begins by developing a generalized model of transmission-history, and estimates of the relative values of the readings of groups, through a five-step process:

            ● First, the witnesses are organized into groups which share distinctive variants.

            ● Second, variant-units involving variants distinct to each group are analyzed according to text-critical principles, or canons.   

            ● Third, a tentative model of transmission-history is developed, cumulatively explaining the relationships of the competing groups to one another by explaining the relationships of their component-parts where distinctive variants are involved.  This model of transmission-history utilizes the premise that the earliest stratum of the Byzantine Text of the Gospels (echoed by Family Π, the Peshitta, Codex A, part of Codex W, the Gothic version, and the Purple Codices N-O-Σ-Φ) arose without the involvement of witnesses that contained the Alexandrian, Western, or Caesarean texts.  Even readings supported by a higher stratum of the Byzantine Text and not by the lowest one are not rejected automatically. 

            ● Fourth, values are assigned to groups rather than to individual witnesses.  Less dependence by one group upon another group, as implied cumulatively by the relationship of its variants to the rival variants in other groups, yields a higher assigned value.

            ● Fifth, all reasonably significant variant-units (those which make a translatable difference) are analyzed according to text-critical canons, using all potentially helpful materials, including readings that are not characteristic of groups.  When internal considerations are finely balanced and a decision is difficult, special consideration is given to readings attested by whatever group appears to be the least dependent upon the others in the proximity of the difficult variant-unit.       

This will yield the archetype of all groups, albeit with some points of instability (at especially difficult variant-units) and with a degree of instability in regard to orthography.

 

Additional Principles

Equitable Eclecticism, besides rejecting the theory that the Byzantine Text was formed entirely via a consultation of manuscripts containing Alexandrian and Western readings, utilizes some additional principles which set it apart from the kinds of textual criticism which produced the revised text and its modern-day representatives:

1.  Textual criticism is a science, not an art.  It is an enterprise of reconstruction, not creation.

2.  The text of the New Testament should be reconstructed in its component-parts:  the Gospels, and Acts, and the General Epistles, and the Pauline Epistles, and Revelation. 

3.  Relationships shown by patterns of readings in one part of the New Testament should not be assumed to exist in the others.

4.  The genealogical descent of a group of manuscripts from an ancestor-manuscript other than the autograph is not assumed without actual evidence that establishes links among specific manuscripts (such as shared formats, shared marginalia, shared miniatures, or readings which conclusively show a historical connection).

5.  Variants involving nomina sacra are placed in a special class, and receive special attention.

6.  The assumption of preference for the shorter reading is rejected.

7.  If a variant has very sporadic support from witnesses greatly separated by age and textual character, this may indicate that the variant was liable to be spontaneously created by copyists, rather than that it was transmitted by distant transmission-streams.

8.  Exceptional intrinsic merit is required for the adoption of variants attested exclusively or nearly exclusively by bilingual manuscripts in which a Greek variant may have originated via retro-translation.

9.  Conjectural emendations are not to be placed in the text. 

           

Equitable Eclecticism also utilizes principles shared by other approaches.  These principles are all superseded by Principle Zero:  no principle should be applied mechanically.

1.  A variant which explains its rivals with greater elegance and force than it is explained by any of them is more likely to be original.
2.  A variant supported by witnesses representing two or more locales of early Christendom is more likely to be original than a variant supported by witnesses that represent only one locale.
3.  A variant which can be shown to have had, in the course of the transmission of the text, the appearance of difficulty (either real or imagined), and which is rivaled by variants without such difficulty, is more likely than its rivals to be original.
4.  A variant supported by early attestation is more likely to be original than a rival variant supported exclusively by late attestation.
5.  A variant which conforms a statement to the form of a similar statement in a similar document, or in the same document, is less likely to be original than a rival variant that does not exhibit conformity.
6.  A variant which involves a rare, obscure, or ambiguous term or expression is more likely to be original than a rival variant which involves an ordinary or specific term or expression.
7.  A variant which is consistent with the author's discernible style and vocabulary is more likely to be original than a rival variant which deviates from the author's usual style and vocabulary and the vocabulary which he may naturally be expected to have been capable of using.
8.  A variant which is fully explained as a liturgical adjustment is less likely to be original than a rival variant which cannot be thus explained.

9.  A variant which is capable of expressing anti-Judaic sentiment is less likely to be original than a rival variant which is less capable of such expression.

10.  A variant which can be explained as an easy transcriptional error is less likely to be original than a rival variant which cannot be explained as an easy transcriptional error or as one which would be less easily made.     

11.  A variant which can be explained as a deliberate alteration is less likely to be original than a rival variant which is less capable of originating in the same way.

12.  Ceteris paribus, in the Synoptic Gospels, a variant which does not result in a Minor Agreement is more likely to be original than a rival variant which results in a Minor Agreement.

Closing Thoughts

Christian readers may feel intimidated or exasperated at the realization that the original text of the New Testament can only be fully reconstructed by a careful analysis of the witnesses – a massive and intricate task which currently involves no less than 135 papyri, about 320 uncials, about 2,900 minuscules, and about 2,450 lectionaries, plus versional and patristic materials.  The feeling may be increased when one also realizes that even the most erudite textual critics have reached divergent conclusions, and that their conclusions must be subject to the implications of future discoveries.

This may lead some readers to decline to investigate the text, deciding instead to hopefully adhere to whatever text (or texts) they already use.  Such an expedient response is understandable, especially in light of the often-repeated (but false) claim that textual variants have no significant doctrinal impact.  Nevertheless, for those few who are not content to place their confidence in textual critics, or to posit providential favor upon a particular set of variants on account of its popularity or for other reasons, the best option is to become textual critics.
Becoming acquainted with the contents of the manuscripts and other witnesses gives additional responsibility, but also additional confidence, somewhat like the confidence of a traveler who knows his maps, as opposed to one who does not and must trust his guides.  

Knowing the message of the map that we have – and being aware of which parts are still questioned, and why, concerning how closely their form corresponds to the form of the original – makes one a confident traveler where one should be confident, and cautious where one should be cautious.  But after we have done our best to conduct research with scientific detachment, it will do us little good if we only possess the map.  Let us walk in the path that the Holy Spirit reveals to us through the Word.  With that thought I leave the reader to consider the words of J. A. Bengel, one of the pioneers of New Testament textual criticism:

   

Te totum applica ad textum:

rem totam applica ad te.

Apply all of yourself to the text,

Apply it all to yourself.

Equitable Eclecticism - Part 1

            In 2010, I wrote an essay called Equitable Eclecticism:  The Future of New Testament Textual Criticism, and included it in the Kindle e-book Assorted Essays on New Testament Textual Criticism.  Since then, I have frequently been asked about how my text-critical approach differs from the Byzantine Priority approach and “Reasoned Eclecticism.”  So, as a convenient reference, here is the essay on Equitable Eclecticism, presented in two parts (with some improvements).

Equitable Eclecticism: 

The Future of New Testament Textual Criticism

Textual Criticism:  Its Goals and Risks 

The textual criticism of the Gospels is a scientific task which has two goals.  The primary goal is the reconstruction of the text of each Gospel in its original form, that is, the form in which it was initially received by the church.  The secondary goal is the reconstruction of the transmission-history of the text.  This involves both the evaluation of rival readings in specific variant-units, and the evaluation of the documents in which the readings are found.  Hort, in his 1881 Introduction, argued that if superior readings are consistently found in a particular document or set of documents, in cases that seem easy to decide, then the character of the documents should be a factor when considering harder cases. 
            Hort expressed this principle as an axiom:  “Knowledge of documents should precede final judgment upon readings.”  The consideration of individual variant-units should never be completely detached from the question of the relative quality of the witnesses, or from the question of how groups of variants became characteristic readings of text-types.  Accurate text-critical judgments will assist in the estimation of the relative values of witnesses, and in the reconstruction of the text’s transmission-history; simultaneously, accurate assignments of relative value to the witnesses, combined with accurate reconstructions of the text’s transmission-history, will assist specific text-critical decisions.     

The textual critic who engages this method should vigilantly avoid circularity; the adoption of a reading because “the best manuscripts” support it ought to be a last resort.  After observing, on analytical grounds, that certain witnesses seem to consistently contain the best readings, a textual critic might then be tempted to abandon the initial approach which led to that premise, and proceed to use the premise itself to justify a tendency to adopt the readings of those witnesses.  Similarly, a textual critic who notices that a group of witnesses tends to contain the worst readings might be tempted to reject the remainder of the testimony of that group of witnesses.  If a textual critic proceeds to build on both such premises, the premises will virtually determine the results of the rest of the analysis. The “best manuscripts” will seemingly get better and better.  

Competing Models of Transmission-History

The model of transmission-history adopted by a textual critic has a strong effect upon the values which a textual critic assigns to the testimony of groups of witnesses, and therefore also upon the final evaluation of variants.  In this respect, the approach which I advocate – Equitable Eclecticism – resembles the approach used by Hort.  However, Equitable Eclecticism yields an archetype which is significantly different from the Revised Text produced by Westcott & Hort, and from the modern descendants of the Revised Text (such as the 28th edition of the Nestle-Aland Novum Testamentum Graece).  This is because research and discoveries subsequent to Westcott & Hort have required the adoption of a transmission-model significantly different from the one used by Hort. 

Hort, building on premises developed by previous investigators, reasoned that the Byzantine Text was essentially the result of a recension that consisted of readings drawn from manuscripts with Alexandrian or Western readings; Byzantine variants were derived from the Alexandrian Text, or the Western Text, or both, or, in some cases, came into being during the recension.  Hort therefore rejected all distinctive Byzantine variants.  After dismissing the Western Text as the result of scribal creativity, embellishment, and a general lack of discipline (with the exception of a smattering of readings), Hort declared the Alexandrian Text (which he called the “Neutral” text) the only text-type which could possibly be regarded as the depository of the original text of the Gospels. 

Hort’s endorsement of the Alexandrian Text was not absolute, but it was so strong that he openly stated that variants shared by the Alexandrian Text’s two flagship codices (B and À) “should be accepted as the true readings until strong internal evidence is found to the contrary,” and “No readings of ÀB can safely be rejected absolutely,” while “All distinctively Syrian” – that is, Byzantine – “readings must be at once rejected.”

Thus, in the approach used by Hort, the degree of favor that was given to the Alexandrian Text was matched only by the degree of disregard that was given to the Byzantine Text.  The categorical rejection of Byzantine readings was a natural implications of Hort’s model of transmission-history in which the Western Text was derived from the Alexandrian Text, and the Byzantine Text was derived from both the Alexandrian Text and the Western Text. 

However, Hort acknowledged that such a clear-cut genealogical model would be out of place if a transmission-model persistently involved readings which all had some clearly ancient attestation.  [See Hort’s Introduction, page 286, § 373.]

This very thing, or something very close to it, was subsequently proposed by textual critics in the 1900s.  Eminent scholars such as E. C. Colwell, G. D. Kilpatrick, and Kurt and Barbara Aland maintained, respectively, that “The overwhelming majority of readings,” “almost all variants,” and “practically all the substantive variants in the text of the New Testament” existed before the year 200.  Nevertheless the Hortian text has not been overthrown.  Only slightly changed, it has become entrenched in NA-28 and UBS-5 as the primary, and nearly exclusive, Greek New Testament used in seminaries. 

With the discovery and publication of Egyptian New Testament papyri in the 1900s – beginning with Grenfell and Hunt’s work at Oxyrhynchus – Hort’s  claim that the Alexandrian readings have a demonstrably greater antiquity than their rivals has eroded.  Harry A. Sturz collected and categorized dozens of distinctive Byzantine variants which were supported by at least one early papyrus.  Sturz’s data does not vindicate the entire Byzantine Text (and we should not expect it to do so).  What it does do is demonstrate that Hort’s main reason for rejecting distinctive Byzantine readings was unsound.  According to Hort’s transmission-model, none of the early distinctive Byzantine readings listed by Sturz should exist.  The fact that they obviously did exist, even in papyri found in Egypt, demonstrates that the Byzantine Text may, at any given point, attest to an ancient distinctive reading.  Hort’s theory of the origin of distinct Byzantine readings was wrong.

In addition, discoveries about the texts in the papyri, in early versions, and in early parchment codices have contributed to the erosion of one of the building-blocks of Hort’s approach:  the proposal that conflations in the Byzantine Text demonstrate that it is later than the Alexandrian Text and the Western Text.  As Edward Miller objected in 1897, eight conflations cannot justify the rejection of the entire Byzantine Text.  They may be comparable to recently minted coins dropped in an ancient well. 

Dr. Wilbur Pickering, in Appendix D of his book The Identity of the New Testament Text, showed that an apparent conflation exists in Codex Sinaiticus at John 13:24 (where the Alexandrian Text has και λεγει αυτω ειπε τις εστιν, the Byzantine Text has πυθεσθαι τις αν ειη, and Sinaiticus reads πυθεσθαι τις αν ειη περι ου ελεγεν, και λεγει αυτω ειπε τις εστιν).  A conflation appears to occur in B at Ephesians 2:5 and at Colossians 1:12 (where the Western Text has καλεσαντι, the Byzantine Text has ικανωσαντι, and B has καλεσαντι και ικανωσαντι).  In D, a conflation appears to occur at Acts 10:48 and John 5:37 (where the Alexandrian Text – supported by P75 – has εκεινος μεμαρτυρηκεν, the Byzantine Text – supported by P66 - has αυτος μεμαρτυρηκεν, and D has εκεινος αυτος μεμαρτυρηκεν).                

In the world according to Hort,
this should not happen.

The papyri have supplied direct evidence against Hort’s belief that apparent conflations imply that the text in which they are found must be late.  In P53, the text of Matthew 26:36 seems to read ου αν, where the Byzantine text has ου and the Alexandrian Text and Western Text have αν.  Papyrus 66 reads σχισμα ουν παλιν at John 10:19 (agreeing with the Byzantine Text), where the Alexandrian Text has σχισμα παλιν and the Western Text has σχισμα ουν.  Similarly, P66 reads εβαστασαν ουν παλιν at John 10:31 (again agreeing with the Byzantine Text), where the Alexandrian Text has εβαστασαν παλιν and the Western Text has εβαστασαν ουν.  

The appearance of such readings in very early manuscripts forces the concession that they do not imply that the text in which they appear is late.  Instead, they prove that an early text can appear to include conflations.  Nevertheless some modern-day textual critics still appeal to Hort’s list of eight Byzantine conflations as if it demonstrated that the entire Byzantine Text was secondary. [See for example Dan Wallace’s treatment of the data in his online essay The Conspiracy Behind the New Bible Translations.]    

Ironically, as the papyri-discoveries took away the pedestal upon which Hort’s transmission-model had stood, they also tended to exonerate Hort’s favored text of the Gospels, the Alexandrian Text, by demonstrating the high antiquity of the Alexandrian text of Luke and John.  Papyrus 75, in particular, possesses a remarkably high rate of agreement with B.  This shows that the Alexandrian Text of Luke and John was carefully preserved in the 200s, and this has tended to alleviate the suspicions of some earlier scholars that the Alexandrian Text was the result of editorial activity in the 200s.

The correspondence between Papyrus 75 and Codex B was interpreted by some textual critics as a demonstration of the antiquity and superiority of the entire Alexandrian Text.  Kurt Aland compared the situation to sampling a jar of jelly or jam:  a mere spoonful is enough to show what is in the rest of the jar.  However, although the agreement between P75 and B proves that the Alexandrian Text of Luke and John is not the result of scribal editing conducted in the 200s, it did not prove that Alexandrian readings are not results of earlier scribal editing.  

Theoretically, if the Western Text could develop in the period prior to the production of P75, so could the Alexandrian Text.  Papyrus 75 proved that the Alexandrian Text of Luke and John is very early; it did not prove that Alexandrian readings are not the result of very early editorial activity.  (As late as 1992, Bruce Metzger maintained that most scholars “are still inclined to regard the Alexandrian text as on the whole the best ancient recension,” on page 216 of The Text of the New Testament, third edition (1992), emphasis added.)

  

Nor did Papyrus 75 prove that the Byzantine Text is less ancient than the Alexandrian Text.  It shows what kind of Gospels-text (or at least, major parts of the Gospels-text) was in use in Upper Egypt in the early 200s.  It does not constitute evidence about what form of text was used, or was not used, in other places.  

The most significant evidence for the absence of the Byzantine Text prior to the 300s is the lack of patristic testimony for its use, but this is largely an argument from silence.  The natural destructive effects of humidity upon papyrus-material, allied with Roman persecutors who sought to destroy Christian literature, silenced a large proportion of the Christian communities of the first three centuries of Christendom.  According to Hort’s theories, when these communities adopted the Byzantine Text in the 300s and 400s, they embraced a new, imported text of the Gospels, setting aside whatever they had used previously.  A more plausible alternative is that they simply continued to use their own local texts which consisted primarily of Byzantine readings.  (For additional thoughts on this subject see my post Byzantine Manuscripts:  Where Were They Before the 300s?.)         

The discovery of the papyri led some textual critics to advocate an undue emphasis upon the ages of witnesses, resulting in a lack of equity toward variants with no support in Egypt.  Because the Egyptian climate allowed the preservation of papyrus, the oldest copies will almost always be copies from Egypt.  To favor the variant with the oldest attestation is to tilt the playing-field, so to speak, in favor of whatever readings are found in whatever manuscripts were stored in the gentlest climate.  But this is no more reasonable than favoring the variants of a manuscript because it was found closer to the equator than other manuscripts.  Certainly when two rival variants are evaluated, and the first is uniformly attested in early witnesses, while the second is found exclusively in very late witnesses, the case for the first one is enhanced.  But to assign values to witnesses according to their ages without considering factors such as climate is to introduce a lack of equity into one’s analysis.  

The papyri-discoveries elicited another interesting development.  Before Hort, pioneering scholars such as Griesbach had organized witnesses into three main groups – Western, Byzantine, and Alexandrian.  Each group, characterized by consistent patterns of readings, was considered a text-type, and manuscripts sharing those special patterns of readings were viewed as relatives of one another.  Hort then divided the Alexandrian group into two text-types, calling its earlier stratum the “Neutral” text, supported by Sinaiticus and Vaticanus.  Then, following analysis by Kirsopp Lake, the Caesarean text of the Gospels was added.  But the evidence from the papyri indicates that even in a single locale (Egypt), the text existed in forms other than those four. 

Consider Papyrus 45, a fragmentary copy of the Gospels and Acts from the early 200s (or slightly earlier).  In Mark 7:25-37, when P45 disagrees with either B or the Byzantine Text or both, P45 agrees with B 22% of the time, it agrees with the Byzantine Text 30% of the time, and 48% of the time it disagrees with them both.  Such departures from the usual profiles of text-types has led some textual critics to reconsider the existence of early text-types, arguing instead that the text in the 100s and 200s was in a state of fluctuation.  A plausible alternative is that some of the papyri attest to the existence of localized text-forms which became extinct, without implying that the Western, Byzantine, and Caesarean forms did not exist prior to the 300s.

Competing Greek New Testaments

In the late 1800s, Westcott & Hort’s Greek text of the New Testament faced several obstacles.  First was the popularity of the Textus Receptus, which, as the base-text of the King James Version, had the status of an ancient landmark in English-speaking countries, regardless of how carefully attempts were made to demonstrate that its Reformation-era compilers, or some stealthy editors in ancient times, were the real landmark-movers.   

In 1898, the Würrtemburg Bible Society published the first edition of Novum Testamentum Graece, an inexpensive Greek New Testament which closely resembled the Westcott-Hort compilation, and which was designed to compete with the edition of the Textus Receptus which was being widely disseminated by the British and Foreign Bible Society.  (The leaders of BFBS apparently had not found Hort’s 1881 case for his compilation irresistible.)  

Eberhard Nestle wrote an enthusiastic recommendation of this handy Greek New Testament; his brief review appeared in the Expository Times in June of 1898.  He pointed out how “disgraceful” it would be to continue to circulate Erasmus’ errors in Rev. 17:8 and Rev. 22:19-21.  He invited the British and Foreign Bible Society to begin to circulate Novum Testamentum Graece instead of the Textus Receptus.  In 1904 the British and Foreign Bible Society began circulating the fourth edition of Novum Testamentum Graece.  Its editor:  Eberhard Nestle. 

While that was happening, a scholar named Hermann von Soden was in the process of compiling an edition of the Greek New Testament which textual scholars expected to become definitive, superseding all previous editions.  But when von Soden’s Greek New Testament was released in 1902-1911, it was found to be extremely cumbersome, and it was flawed in various ways.  Nestle’s Novum Testamentum Graece was on hand to meet the need of seminarians and other textual researchers, and it has done so ever since – and it eventually was adopted, in later editions, as the primary base-text for new translations.

But should that be the case?  According to Kurt and Barbara Aland, the 27th edition of NTG differs from the text compiled by Eberhard Nestle “in merely 700 passages.”  Considering the high number of variant-units involved, this implies that the text of the Gospels in NA-27 and UBS-4 is essentially the same text that was published by Eberhard Nestle in the early 1900s.  (See page 20 of The Text of the New Testament:  “In its 657 printed pages the early Nestle differs from the new text in merely seven hundred passages.”  Consider that in the Gospels alone, the 25th and 27th editions of NTG disagree at over 400 places.)

It is as if the papyri (and the research into early versions, and the revisions of patristic writings, and other significant discoveries and research undertaken in the 1900s) have scarcely had an impact, whereas in reality they cracked the transmission-model that was a large part of the foundation of the Westcott-Hort compilation.

The marketplace for Greek New Testaments in the early 1900s rapidly became crowded:  Bernard Weiss, Alexander Souter, and J. M. S. Baljon made compilations which rivaled Nestle’s.  F. H. A. Scrivener’s editions of the Textus Receptus remained in circulation. Thomas Newberry’s 1870 Englishman’s Greek New Testament – an interlinear edition of the Textus Receptus which featured a presentation of variants adopted by textual critics prior to Westcott & Hort (Griesbach, Lachmann, Tregelles, Tischendorf, Alford, and Wordsworth) – also remained in print.  The public generally had to choose between either a Greek text similar to the 1881 revision of Westcott & Hort, or the Textus Receptus.  

That changed in 1982, when Zane Hodges and Arthur Farstad published a compilation called The Greek New Testament According to the Majority Text.  As its name implies, this text was intended to consist of the readings shared by the majority of Greek manuscripts.  Hodges and Farstad proposed that the Alexandrian Text is a heavily edited, pruned form of the text, and that the Majority Text is much better, inasmuch as “In any tradition where there are not major disruptions in the transmissional history, the individual reading which has the earliest beginning is the one most likely to survive in a majority of documents.”  The work of Hodges and Farstad was the basis for many text-critical footnotes in the New Testament in the New King James Version, which was published around the same time under Dr. Farstad’s supervision. 

A similar work was released in 1991 by Maurice Robinson and William Pierpont, called The New Testament in the Original Greek According to the Byzantine/Majority Textform.  A second edition was published in 2005.  Rejecting any notion of defending the Textus Receptus (which differs from the Byzantine Text at over 1,800 points, about 1,000 of which are translatable), Robinson and Pierpont regarded the Byzantine Text as virtually congruent to the original text.  The Byzantine Textform consists of a series of majority readings, wherever majority readings clearly exist.  Outside the pericope adulterae (John 7:53-8:11) and the book of Revelation, almost no analytical attempts to reconstruct the relationships of variants within the Byzantine tradition seems evident, since the question is usually settled by a numerical count (or, by a consultation of representative manuscripts, using data from von Soden’s work).

In some respects, Hodges & Farstad and Robinson & Pierpont have paved a trail that was blazed in the 1800s by John Burgon, who opposed the theories of Westcott & Hort.  Burgon’s aggressive writing-style sometimes overshadowed his argumentation; nevertheless some of his views have been vindicated by subsequent research.  

For example, Hort asserted that “even among the numerous unquestionably spurious readings of the New Testament there are no signs of deliberate falsification of the text for dogmatic purposes,” but Burgon insisted that the opposite was true.  Burgon’s posthumously published Causes of Corruption (1896) even included a sub-chapter titled “Corruption by the Orthodox.”  Almost a century later in 1993, a variation on Burgon’s theme was upheld by Bart Ehrman in the similarly titled book The Orthodox Corruption of Scripture.  As a result, although Ehrman exaggerated his case in many respects, no textual critics now consider Hort’s assertion to be correct. 

Many scholars and interested bystanders, noticing that the weaknesses of several of Hort’s key premises and assertions have been exposed, have been willing to consider the model of transmission-history proposed by the supporters of the Byzantine Textform.  Others have irresponsibly attempted to associate it with the fundamentalist doctrine of King James Onlyism.  

Others have rejected it because, despite detailed lists of principles of internal and external evidence in Dr. Robinson’s essay The Case for Byzantine Priority, the factor that usually determines the adoption of a variant in the approach advocated by Robinson is its attestation in over 80% of the Greek manuscripts.  Patristic evidence and the testimony of early versions are not included in the equation of what constitutes the majority reading.  Distinctive Alexandrian variants, Western variants, Caesarean variants, and even minority readings attested by the oldest Byzantine witnesses (such as parts of Codices A and W) have no chance of being adopted; generally, if a variant is supported by over 80% of the Greek manuscripts, it is adopted.  

The validity of such an approach depends upon the validity of the premise that the transmission of the text of the Gospels was free from “major disruptions.”  However, major disruptions have had enormous impacts upon the transmission of the text.  Roman persecutions – followed by Roman sponsorship – wartime and peacetime, dark ages and golden ages – all these things, plus innovations and inventions related to the copying of manuscripts, drastically changed the circumstances in which the text was transmitted, and while all text-types were affected by them, they were not all affected to the same extent.  It is no more scientifically valid to adopt a reading because it was favored in Byzantine scriptoriums than it is to adopt a reading because the manuscripts that support it were kept in an area with low humidity (namely Egypt) and thus lasted longer than the manuscripts in other places.        

[Continued in Part 2]

James White and Hebrews 1 - A Response

The first page of Hebrews
in Papyrus 46.
(Yellow circle = verse 3.)

            Recently, my post Fool and Knave! Hebrews 1 in Codex Vaticanus was the subject of part of a video broadcast by apologist James White (whom I shall call Dr. White in this post, since he accorded me the same title in his broadcast).  His main point seemed to be that it is possible for other people – particularly competitors or opponents – to misinterpret, reframe, or selectively edit one’s statements so as to misrepresent their intention. 

         

            Of course I agree with Dr. White about that.  Jesus Himself had His statements misused by those who wished to oppose Him.  However, along the way, some things were said which call for clarification, so I have cobbled together this response.            

● 25:50 – Just a little quibble:  Dr. White called the note in Codex Vaticanus alongside Hebrews 1:3 a colophon; however, a colophon is technically something else – more like a note by a copyist (rather than a later corrector) about when and where and by whom the manuscript was produced.  The note in Codex B alongside Hebrews 1:3 is just a note. 

● 27:30 – In the course of offering some thoughts about a variant-unit in Hebrews 1:8, Dr. White described me as a “Byzantine Priority-type textual critical scholar;” however, the term “Byzantine Priority” refers to a text-critical approach which favors Byzantine readings all the time, and that is not my view.  (I adopt και κριτὴς in James 4:12, for example.)  My approach is Equitable Eclecticism, in which readings with very strong intrinsic evidence in their favor are capable of being adopted instead of rival readings with much more abundant attestation.

            In his discussion about the presence or absence of “and ever” in Hebrews 1:8, Dr. White stated, “Other people would respond to that by saying, well, while that’s possible, and certainly there is something to be said about Alexandrian scribes desiring a concise, tight type text, it’s also just as possible and probable that there is the opposite tendency amongst the Byzantines to have a more flowing text, and  to have fuller titles, and especially in worship contexts, liturgical contexts, things like that; it could go the opposite direction.  So both of those have to be balanced.”

            I would be interested to know the identity of these “other people” who reject “and ever” in Hebrews 1:8.  For it is not just Byzantine Prioritists who adopt the longer reading there:  it is adopted in the Nestle-Aland compilation, and the UBS Greek New Testament accepts the longer reading without drawing attention to the shorter reading’s existence.  Michael Holmes (editor of the SBL-GNT) rejected Codex Vaticanus’ reading here too.  As far as I can tell, one has to reach back over a century, to Hort (of course) to find any editor willing to even bracket του αἰωνος; the inclusion of the words is reflected not only in the KJV and NKJV but also in the NIV, ESV, CSB, NLT, NRSV, and NASB. 

            Via his insistence that “Both of these have to be balanced” – both a reading that has massive and ancient and widespread support, and a reading with extremely limited support – Dr. White has illustrated a problem with the approach that has dominated the field of New Testament textual criticism for far too long:  it is often possible, if one is sufficiently creative, to imagine reasons to prefer readings with minimal external support.  To restate:  reasoned eclecticism, as it is currently practiced, opens a wide door to the acceptance of quirk-readings. 

            That is why the NIV currently has a reading in Mark 1:41 that is supported by only one Greek manuscript.  This is why the TNIV removed “Son of God” from Mark 1:1.  That is why Bart Ehrman argues for the reading “apart from God” in Hebrews 2:9.  That is why there is no reference to fasting in the text of Mark 9:29 in the ESV.  That is why, as Dr. White himself has observed, although the reading “not yet” in John 7:8 is supported by “an awesome array of witnesses” (including not only the Byzantine Text but also early papyri), the Nestle-Aland/UBS compilation reads, instead, “not.”  And that is why the editors of the latest edition of the Nestle-Aland compilation took the next step in Second Peter 3:10:  they adopted a reading for which the editors created what was, to them, a plausible case, even though it is found in no Greek manuscripts.    

    

● 30:00 – Dr. White stated that he has said that “If you apply the same standards of hermeneutics and exegesis to the most Alexandrian text, and the most Byzantine text, you’re not going to come up with a different Christian faith.”  However, were Dr. White to undertake that task, I suspect that he would end up with two different Bibliologies – one in which the text of the New Testament is errant, and another in which the text of the New Testament is inerrant.      

            Inerrancy is a doctrine which cannot be proven; however, a demonstrable error in the text would show that the text is errant, Q.E.D.  Briefly leaving Dr. White’s video-broadcast, let’s explore the approach he used in his book The King James Only Controversy, when discussing a textual contest in Mark 1:2, and test its effects when applied to another variant.  (Speaking of The King James Only Controversy:  when will Dr. White acknowledge that his account of Tischendorf’s first encounter with Codex Sinaiticus in that book is incorrect??)

            Dr. White argued (I mean “argued” in the technical sense; he reasoned) that between the reading “in the prophets” (read by the vast majority of Greek manuscripts, and supported in a second-century composition by Irenaeus) and the reading “in Isaiah the prophet,” “it is much easier to understand why a scribe would try to “help Mark out,” so to speak, and correct what seems to be an errant citation than to figure out why someone would change it to “Isaiah the prophet.””

            Dr. White proceeded to propose that what appeared to scribes to be an error is actually a conventional way of referring to a conflated citation.  Dr. White and others have made similar arguments regarding not only Mark 1:2, but also Matthew 27:9, where two prophecies with similar themes are referenced using the name of only one.  However, does such reasoning work at Matthew 13:35? 

            Codex Sinaiticus – one of the two fourth-century manuscripts upon which the Nestle-Aland compilers lean most heavily and most often – identifies the prophet cited in Matthew 13:35 as Isaiah.  Problem:  the passage quoted there is Psalm 78:2, which was written by Asaph, not by Isaiah.  Eberhard Nestle (the first editor of what later became the Nestle-Aland compilation) considered Sinaiticus’ reading “certainly genuine.”  And if we were to apply the reasoning that Dr. White used to justify adopting “Isaiah the prophet” at Mark 1:2, how could we say otherwise?  It is much easier (using Dr. White’s reasoning) to understand why a scribe would try to help Matthew out, by removing the error, than to figure out why someone would put Isaiah’s name where it clearly does not belong. 

            And thus a case is made for the quirk-reading – in this case, a quirk-reading that, if adopted, would impact the doctrine of inerrancy, barring some desperate verbal acrobatics.  (The question of whether Biblical inerrancy is a cardinal Christian doctrine seems to be in dispute among some textual critics, by the way, even evangelical ones at Dallas Theological Seminary.)    

            Perhaps another example will be instructive.  The Alexandrian text, in Matthew 27:49, includes a statement that Jesus was struck with a spear before He died, which flatly contradicts the testimony in John 19:30-35.  (Does Dr. White think that this reading, too, has to be balanced?)  An argument could be made that this must be the original reading, on the grounds that it is the more difficult reading.  Without addressing the merits or defects of such a case, can it be denied that if one were to conclude that the errant Alexandrian reading in Matthew 27:49 were original, it would yield a different Bibliology than the non-problematic Byzantine reading? 

            More frequently, what is at stake is not a Christian doctrine, but the ability of a particular passage to teach what it originally taught.  And although, for the most part, Dr. White and I are saying the same thing about the ability of both the Nestle-Aland compilation and the Byzantine text (and some independent compilations) to adequately express Christian doctrine, we advocate two different approaches at this point.  

            Dr. White seems content, when encountering two rival variants, to say, “Both of those have to be balanced,” even when one has extremely meager external support, as long as a scenario can be imagined in which scribes created the reading that has massive and ancient support – whereas I would argue that quirk-readings should be treated as quirk-readings, and be rejected, rather than share the page with the inspired text. 

            Two examples of quirk-readings have already been provided, both from Hebrews 1:3:  it is not difficult to see how the words δι’ αυτου (or δι’ ἑαυτου) – “by Himself ” – could be lost via scribal negligence, when a copyist’s line of sight drifted from the final letters of δυνάμεως αὐτου to the letters του in the next word, thus skipping the letters in between.  The same kind of mistake explains the loss of the word ἡμων after ἀμαρτιων.  (In text-critical jargon, these are cases of parablepsis elicited by homoeoteleuton, and it is a very ordinary kind of scribal mistake.)  

Let’s look at three more examples of quirk-readings, all from the Gospel of Mark. 

            ● In Mark 10:50, in the account about the healing of blind Bartimaeus, Bartimaeus is said to throw aside his cloak as he jumps up, obeying Christ’s call.  In a respectably old Syriac manuscript, though, the text says that Bartimaeus, rather than putting down his cloak, took up his cloak.  A case can be imagined for the minority reading, the idea being that the story was tweaked by scribes who desired to turn Bartimaeus’ experience into an allegorical picture of repentance, in which the garments stained with sin are laid aside.  Should both of these readings therefore be perpetually retained, one in the text, and one in the margin, in our Greek compilations and in English versions, assuring that nobody will be confident about what Bartimaeus did with his cloak?       

            ● In Mark 10:24, in the vast majority of the Greek manuscripts of Mark (and not just the medieval copies), Jesus tells His disciples, “Children, how hard it is for those who trust in riches to enter the kingdom of God.”  This reading is also supported by the Gothic version (made in the mid-300’s) and by a quotation made by Ephrem Syrus (in the mid-300’s) from Tatian’s Diatessaron (which was made around 170).  The flagship manuscripts of the Alexandrian text, however, do not include the phrase “for those who trust in riches,” thus yielding the sentence, “Children, how hard it is to enter the kingdom of God.” 

            Dr. White favored the Alexandrian reading of Mark 10:24 when he wrote The King James Only Controversy.  He reasoned that “It is easier to understand how the phrase could be added than to understand why it would be deleted.”  Is that so?  When I look at the Greek text –  Τέκνα, πως δύσκολόν ἐστιν τοὺς πεποιθότας ἐπὶ χρήμασιν εἰς τὴν βασιλείαν του Θεου εἰσελθειν – it is very easy to see how the phrase would be deleted:  by simple scribal negligence, when a copyist’s line of sight drifted from the letters –ιν at the end of the word ἐστιν to the same letters at the end of the word χρήμασιν, thus accidentally skipping the words in between.  The longer reading, in this case, accounts for the origin of the shorter reading, and does so rather elegantly.  

            Dr. White acknowledged in his book that such a simple scribal error “is, of course, a possibility.”  But he did not conclude that the Alexandrian reading was incorrect.  Instead, he recommended putting a footnote at this passage:  “The reading should be noted if it is not contained in the text; or, if it is contained in the text, its absence in ﬡ or B should be noted as well.  In either case, the reader should be given all the information available.”

            However, this is both wishful thinking – for there is not an English translation in existence which gives readers all the information available – and wrong.  Such generosity toward minority-readings inevitably adulterates the force of the original reading, so as to cause readers to constantly wonder, “Where is the Word of God:  in the text, or in the footnote?”  Dr. White seems perfectly willing to reply that it doesn’t matter, since the same teachings emanate from the New Testament as a whole no matter which reading is adopted in that particular passage.  However, such an approach compels the original text in all those passages to share its authority with a scribal corruption.            

            ● A third example:  in Mark 10:19, the words “Do not defraud” are not in Codex Vaticanus, and some other uncial manuscripts concur.  This is not hard to explain:  the same mechanism that caused the loss of four words in Mark 10:24 caused the loss of two words here, when a copyist’s line of sight drifted from the letters –ρήσης at the end of μὴ ψευδομαρτυρήσης to the same letters at the end of μὴ ἀποστερήσης.     

            Is there really a need to inform Bible-readers about every passage where the copyists of Sinaiticus and Vaticanus made mistakes?  Should we clamor for more trivia in Bible-footnotes, so as to allow – nay, to cause – more glitch-readings to adulterate the authority of the original text?  Granting that we could afford to do so, I submit that the responsibility of textual critics is, rather, in the opposite direction:  the effects of scribal corruptions should be removed from the base-text, and expressions of indecision should be resorted to only in cases where the evidence is extremely closely contested.  I hope that Dr. White, upon further reflection, will agree.    

(Postscript:  in the forty-fourth minute of his video-lecture, as Dr. White read the textual apparatus for Hebrews 1:3, he misinterpreted the letter D as “Bezae Cantabrigiensis.”  Actually D in this passage stands for a different manuscript, Codex Claromontanus.)