Textual Evidence

Robinson Interview
on Byzantine Text

Excerpt from: Interview with Maurice Robinson, D.A. Black Interview, (2006)

Page Index

Interview: - With Maurice Robinson:
    Introduction - and selected questions
    4. Dating Readings: Byzantine Text-type
    5. Criteria : Choosing the best Byzantine reading
    6. Internal Evidence: value and place in methodology
    7. Eclecticism: versus External MSS evidence
    8.The Original Text: do we have it?
    9.The Alexandrian Text: why is it inferior?
    10.Byzantine Text: possibility of errors
    11.KJV/NKJV: How good are modern translations?

Return to Index

D.A. Black


Taken from:

Headings have been added for clarity and navigation purposes.

Interview with
Maurice Robinson (2006)

Byzantine Text-Type

Dave Black:
I recently asked the students in my textual criticism class to write out some questions they wanted to ask Dr. Maurice Robinson, my colleague at Southeastern Seminary, having already read his Case for Byzantine Priority. Dr. Robinson graciously agreed to answer them. I publish his written response, verbatim; [excerpted here]

4. What are the criteria for determining the date of a reading of the Byzantine text?

Let me return the question: what are the criteria for determining the date of any reading (as opposed to the date of an existing witness)? The only certainty is that a reading appearing in any ms is either older than that ms or created by the scribe of that ms. A reading found in a 2nd-century ms is certainly at least that old, but not necessarily any older nor necessarily “original”. Equally, a reading in a 12th-century ms might be “old” and possibly original, according to some textual critics. Most critics acknowledge that most readings found in mss of any era — regardless of texttype — have demonstrable existence in mss, versions, or fathers dating from the second and third centuries. Colwell and others rightly have postulated that, except for late singular or near-singular readings, all meaningful variants are “old” and originate before ad 200. Readings dated solely on the basis of ms age do not carry very far when attempting to determine the autograph or “original” reading.

The real issue involves which reading in any particular variant unit is the oldest, and therefore the original or autograph reading. Such cannot be determined solely or even primarily on the basis of the date of whatever ms or mss contain a given reading. Many factors — external evidence, internal evidence, intrinsic probability, transcriptional probability, transmissional probability —must be considered in order to determine the most likely original reading in any variant unit. Such factors weigh differently according to various text-critical theories and methodologies, resulting in conclusions that vary according to the theory espoused.

I suggest that Occam’s Razor applies, and thus prefer the theory and method that, when consistently applied, appears to possess the least number of problems or complications. Assuming such to have occurred, that theory and method should be followed steadfastly, brooking no exceptions so long as the evidence and conclusions remain commensurate with the principles that characterize that theory (cases remain where any theory will retain some problematic variant units; I do not speak here regarding those). As a result of the consistent application of these principles, I draw a solid conclusion in favor of the Byzantine-priority hypothesis.

5. When the Byzantine mss disagree among themselves, what criteria do you use to determine the best Byzantine reading?

Were the Byzantine Textform monolithic (as some seem to suggest), no division would exist among its member mss; only a shared unity reflective of the original Byzantine archetype. Such is not the case, since the Byzantine archetype was transmitted amid many transmissional lines that at times differ among themselves. At any given point of variation some deviation is expected, although usually to a minor degree. When the division among the Byzantine mss approaches 50% in any variant unit, however, more care must be taken in order to determine the Byzantine archetype reading..

Our position as stated is to consider support greater than 70% among the Byzantine mss as reflecting the archetype of that Textform (Hodges and Farstad in their “Majority Text” edition suggest 80%, but only when including mss from non-Byzantine witnesses, such as the Western or Caesarean mss). Within our system, as the level of support drops from 70% down to 50%, external reliability steadily decreases. When the support ranges from 45%-55%, we consider the external data too closely divided, and non-determinative. At such points, the establishment of the Byzantine archetype requires a judicious application of various forms of internal evidence, as noted previously: intrinsic probability, transcriptional probability, transmissional probability, consideration of known scribal practices and proclivities, and the like. The primary reading suggested by each category carefully must be weighed in conjunction with all other categories; the best attested reading overall gaining the ascendancy. Where external testimony is closely divided, we place the alternatives in the side margin. As the division lessens from 50%-50% and approaches 70%-30%, alternate readings continue to be evaluated on the basis of internal evidence, but with a steadily increasing inclusion of external support as a determining factor. Our theory and method should not be described as “Byzantine eclecticism”: the primary establishment of the text remains externally based. Internal evidence becomes a confirming and balancing factor, but is not determinative, except where a variant unit externally is closely divided. This discussion is continued under the next question.

6. What place should be given to the internal evidence?

The primary determination of the text within Byzantine-priority is based on external criteria. So long as these criteria are nearly unanimous within the Byzantine tradition, internal evidence does not play a determining role in the establishment of the Byzantine archetype. This does not mean that internal evidence plays no role, but that its role is secondary and confirmatory: internal evidence is used to explain the rise of non-Byzantine and sub-Byzantine variants in relation to the archetypal reading established on external grounds. The use of internal evidence remains necessary. We believe that internal criteria generally will confirm the originality of the primary Byzantine reading and demonstrate the secondary nature of its various competitors — not always with the same degree of confidence nor always using the same internal criteria in any given case.

Where Byzantine readings are closely divided, the external criteria do not define a clear Byzantine archetype. In such cases, the various forms of internal criteria take a leading role. Closely divided external evidence can only eliminate readings that are weakly supported or non-Byzantine. The remaining closely divided readings need to be evaluated on the basis of internal criteria (transmissional, transcriptional/scribal, and intrinsic probabilities). One must remember, however, that divided Byzantine testimony in a given variant unit is relatively rare; for nearly all of the text, the Byzantine archetype reading is established directly on the basis of external attestation, with internal criteria confirming that result.

7. How does your approach differ from eclecticism (of any kind – radical, reasoned, Sturzian)?

“Sturzian” must be a textual variant of some sort. By way of definition, radical or rigorous eclecticism is based almost exclusively on internal evidence, while reasoned eclecticism is based on varying combinations of external and internal evidence.

Other theories are primarily externally based. These include the preference for a favorite ms (Tischendorf), small group of mss (Westcott-Hort), or a particular texttype (Amphoux/Heimerdinger for the Western; Robinson-Pierpont for the Byzantine). The Sturzian view also is primarily external, but follows the reading found in a majority of geographically diverse texttypes.

The simplest answer has been noted above. Eclecticism (any type) treats the text piecemeal, on a variant-by-variant basis. Byzantine-priority does not function on a variant-by-variant basis, but considers the entirety of the text as primary. Only after the text is established on the wider scale are the competing internal claims evaluated within the sequential variant units. In no case does Byzantine-priority adopt as primary a reading that lacks significant support. This principle clearly distinguishes Byzantine-priority from all forms of eclectic methodology. Internal criteria are utilized to confirm text-critically those readings established by external criteria and to demonstrate the secondary nature of the various competing alternatives.

The Byzantine Textform has been criticized because its specific pattern of readings is not found among surviving pre-fourth century textual witnesses. Yet the text produced by the various forms of modern eclecticism creates a sequence of claimed “original” readings whose pattern cannot be demonstrated ever to have existed within transmissional history. The Byzantine Textform, on the other hand, does in its aggregate reflect a legitimate and historically valid form of text that at any point retains a basic consensus of support among its related witnesses. Even where the Byzantine mss are divided, a transmissional consistency permeates a substantial portion of Byzantine witnesses. The Byzantine Textform thus differs significantly from the results produced by various eclectic schools.

8. How would you defend your view that we now have the New Testament in the original Greek against the agnostic position that we cannot get back to the original?

I dealt with this issue in my ETS 2005 paper, “The Integrity of the Early New Testament Text: A Collation-based Comparison”. In general, any claim that suggests absence of the physical autograph equals absence of textual reliability or biblical authority is bogus. The manuscript copies we possess remain substantially identical to the autographs. As demonstrated in my paper, the earliest extant (non-Byzantine) papyri compared against the text of Byzantine minuscule mss copied a thousand years later share a verbal identity approximating 92% — including orthographic and non-translatable differences. With such a large percentage of common text, even over more than a millennium of transmission, it is clear that the autograph text substantially has been preserved, even among disparate copies representing quite different textual traditions. On the same principle, dispute hardly should arise as to whether the autograph text similarly was preserved during the much shorter period between autograph composition and the earliest extant mss. Transmissional observations suggest an equally reliable transmissional history during the short period from which no evidence exists. In addition, all doctrinal essentials are clearly present within the ca. 92% average base text; no doctrine is established or negated within the remaining ca. 8% where differences occur. Also, most variants are quite minor and generally stylistic in nature. If the orthographic, non-translatable, and minor stylistic variants are excluded, the overall agreement among the earliest and latest mss rises substantially. The existing documents accurately represent the autographs in all essential points. The text we now possess is sufficient and substantial for establishing and maintaining all doctrinal positions held within orthodox Christianity, skeptics and postmodernists such as Ehrman, Epp, Parker, or the media to the contrary.

9. What is the leading argument in your mind for the inferiority of the Alexandrian text type?

“Reasoned transmissionalism”! Had any texttype other than the Byzantine more closely represented the autograph form of the text in any nt book, that texttype should have thoroughly permeated the primary Greek-speaking region of the Empire beyond the first few centuries. Any later-developing “new” texttype would fail to dominate against a presumed liturgically entrenched and widely disseminated “original” Textform. One need only consider in this regard the failure of the “Western” text to gain a substantial hold within the Greek ms tradition; similarly, one can consider the limited and apparently “localized” nature of the Alexandrian texttype.

Westcott and Hort acknowledged that the Byzantine Textform dominated the Greek-speaking Eastern Empire from the mid-4th century onward. They also noted that such dominance could have occurred only in two ways: either (1) the Byzantine Textform was the product of a formal, ecclesiastically sanctioned revision, promulgated with full ecclesiastical authority behind it (the Alands’ “Byzantine Imperial Text”); or (2) the Byzantine Textform reflects the autograph form of the text, which — under a normal process of transmission — would be expected to produce an overwhelming number of descendants “at each stage of transmission” (W-H, Introduction, 45). W‑H argued the first alternative, without which their preferred B-Í type of text could not be maintained.

The W-H “revision” hypothesis generally has been discarded, due to lack of historical corroborating evidence. A “process” view is now instituted in its place, suggesting that, over a lengthy period of time, the Byzantine Textform slowly evolved into what finally becomes a relatively fixed form during the post-ninth century minuscule era. But, as Zane Hodges long ago pointed out:

“No one has yet explained how a long, slow process spread out over many centuries as well as over a wide geographical area, and involving a multitude of copyists, who often knew nothing of the state of the text outside of their own monasteries or scriptoria, could achieve this widespread uniformity out of the diversity presented by the earlier forms of text. Even an official edition of the New Testament — promoted with ecclesiastical sanction throughout the known world — would have had great difficulty achieving this result as the history of Jerome's Vulgate amply demonstrates. But an unguided process achieving relative stability and uniformity in the diversified textual, historical, and cultural circumstances in which the New Testament was copied, imposes impossible strains on our imagination”

(Hodges, Appendix C, in Wilbur N. Pickering, The Identity of the New Testament Text, 166).

Other claims, such as the influence of Chrysostom, the Constantinopolitan Church, or the supposed destruction of Alexandrian mss due to the Islamic conquest, are discussed in my full-length essay, “The Case for Byzantine Priority”, available on the internet and as an appendix to the R-P Byzantine Greek nt. I might observe that, if the Alexandrian text could have been wiped out by the Islamic conquest, that predominantly Egyptian text was not widespread, but reflected only a more localized tradition; also, for either Chrysostom or Constantinople to effect such a significant change in the Church’s base text, full ecclesiastical authority and proclamation would have been necessary in order to accomplish its general acceptance throughout the Eastern Empire. No such proclamation or imposition of ecclesiastical authority seems ever to have occurred. The implication returns to Byzantine originality as the more probable cause of that Textform’s dominance within the transmissional history of the nt.

To return to the original question, I do not reject the Alexandrian texttype primarily on the basis of subjective judgments regarding a presumed inferior quality of its readings; nor do I approve the Byzantine Textform because of a supposed superior quality regarding its readings. Textual establishment is primarily a matter of evaluating the external evidence in order to determine on the basis of transmissional considerations whatever sequentially connected set of readings appears most likely to have been original. Hort correctly stated the leading principles in this regard: “Knowledge of documents should precede final judgment upon readings” (W-H, Introduction 30), and “All trustworthy restoration of corrupted texts is founded on the study of their history” (Ibid., 40; small caps original in both cases).

10. Do you allow any possibility that the Byzantine text preserved the wrong reading?

Michael Holmes asked that question during our 2000 on-campus text-critical conference. My answer was definite: NO — at least in regard to the 99% bulk of the Byzantine Textform where significant division does not appear among its mss. This again distinguishes our position from any form of “Byzantine eclecticism”. Had we treated the variant units merely on a case-by-case basis without following our principles consistently, the level of external support would fluctuate, and necessarily would produce a text that was no longer “authentically Byzantine”. As my text-critical mentor, Kenneth W. Clark, long ago suggested, on transmissional grounds, it is more likely that the original text is preserved in a single texttype, rather than “original readings” having been scattered among diverse textual witnesses that reflect widely divergent streams of transmission. Many good reasons can be cited in support of Clark’s view, one of which is the lack of support within eclectic editions for the pattern of readings claimed to be “original” (cf. my forthcoming essay regarding 105 whole verses of na27/ubs4 which, when tabulated as whole‑verse units, lack support from any ms, version, or father within transmissional history).

Modern eclecticism presupposes a highly skewed transmission of the text, sarcastically described by Calvin Porter:

“Very early the original text was rent piecemeal and so carried to the ends of the earth where the textual critic, like lamenting Isis, must seek it by his skill”

(Calvin L. Porter, “A Textual Analysis of the Earliest Manuscripts of the Gospel of John”, PhD Diss., Duke University, 1961, 12).

In contrast, a relatively “normal” process of transmission maintains the greater likelihood of autograph preservation within a single dominant branch of the transmissional tradition, absent any evidence demonstrating a major upheaval in the transmissional process. This argument particularly is cogent when transmissional dominance is accompanied by a high level of internal plausibility for the readings that appear within that dominant Textform.

11. What is your viewpoint on the reliability of the KJV/NKJV versions?

Text-critically, the KJV and NKJV are translated from a text commonly termed TR (Textus Receptus, or Received Text), with a smattering of readings from the Latin Vulgate, various retranslations from Latin sources, or editorial conjectures. All other nt translations tend to reflect a general Alexandrian type of text. In both cases, I have reservations, since neither agrees totally with the Byzantine tradition.

Yet the NKJV includes footnotes regarding significant variant readings, and clearly indicates whether a variant reading is found in the Nestle-ubs predominantly Alexandrian tradition (“NU-text”) or the Byzantine tradition (the “M-text”, being the Hodges-Farstad “majority” text). Footnotes in most English translations lack texttype-specific identification, leaving their readers in the dark regarding the textual nature of any variant. In this sense, there is a real benefit in the NKJV footnotes, unmatched in any other English translation. I would prefer to see a good, formal-equivalence English translation based on the Byzantine Textform, with footnotes indicating translatable variants from the alternative traditions.

Should one inquire about the translational quality of the KJV or NKJV, that is a separate matter. I strongly prefer formal equivalence; the KJV and NKJV qualify in that regard. However, I do not use the KJV for teaching or study purposes due to its archaic language and an inconsistency in rendering Greek words and phrases, particularly in parallel passages. The NKJV is superior in this regard, but remains inconsistent in rendering some words and phrases. Both translations are generally accurate and reliable: they will not mislead a reader in doctrinal matters. Other modern formal equivalence translations (ESV, NASV) are superior in translational quality, and would be strongly recommended had they been based on a Byzantine text or had they identified a sufficient number of Byzantine variants in their footnotes.

Return to Index