by Maurice A. Robinson
Principles of External Evidence
The Byzantine-priority method looks at external evidence as a primary consideration within a transmissional-historical framework. The key issue in any unit of variation is not mere number, but how each reading may have arisen and developed in the course of transmission to reflect whatever quantitative alignments and textual groupings might exist. To this end a careful consideration and application of various external principles must be applied to each reading within a variant unit. Certain of these criteria are shared among various eclectic methodologies, but none demonstrate a clear linkage to transmissional-historical factors.
The quantity of preserved evidence for the text of the NT precludes conjectural emendation. The NT text has been preserved to an extent far exceeding that of any other hand-transmitted literature of antiquity. Thus, the likelihood that conjectural emendation might restore the original form of the text is virtually nil. While other critics do not exclude conjectural emendation as a possibility, conjecture does not gain a serious foothold in contemporary praxis, nor is there any pressing need for such.69 Conjecture argues a historical model requiring an unparalleled transmissional catastrophe in which all known witnesses–manuscript, versional, and patristic–failed to preserve the original text at a given point.
Given the quantity of NT evidence, such becomes doubtful in the extreme, and if otherwise valid would call into question every word found in any extant witness.
Readings which appear sporadically within transmissional history are suspect. Assuming the general normality of manuscript transmission, the original text should leave a significant imprint over the range of transmissional history. Optimally, an original reading should demonstrate a continuity of perpetuation from the autograph to the invention of printing. Readings which fit this criterion have an initial presumptive authenticity that cannot easily be overturned. Certain corollaries follow:
- A reading preserved in only a single MS, version or father is suspect. As with conjecture, it remains transmissionally unlikely that all MSS, versions, and fathers save one should have strayed from the original reading. Even if some witnesses are considered “best” within a given portion of text, it remains unlikely that any such witness standing alone would have preserved the original text against all other witnesses. So too the next corollary:
- Readings preserved in a small group of witnesses are suspect. Just as with single testimony, readings preserved in but two witnesses are unlikely to have preserved the original reading against all remaining testimony. This principle can be extended to other small groups, whether three or four MSS, or even more, so long as such groups remain smaller than a larger texttype (which is treated under other principles). Such cases reflect only sporadic or limited transmission.
- Variety of testimony is highly regarded. This principle addresses two areas, neither sufficient to establish the text, but either of which lends support to a given reading.
- A reading supported by various versions and fathers demonstrates a wider variety of support than a reading lacking such. The greater the variety of support, the more weight is lent to a reading. However, if a reading possesses only versional or patristic support without being evidenced in the Greek manuscript tradition, such a reading is secondary. Isolated patristic or versional testimony is not sufficient to overturn the reading most strongly supported among the Greek MS base.
- Among Greek MSS, a reading shared among differing texttypes is more strongly supported than that which is localized to a single texttype or family group. Diversity of support for a reading is far stronger than the testimony of any single manuscript or small group of MSS. Overlooked by many is the fact that the Byzantine Textform is the most frequent beneficiary of such diverse support: there are far more instances wherein an Alexandrian-Byzantine or Western-Byzantine alignment exists than an Alexandrian-Western alignment wherein the Byzantine stands wholly apart. Indeed, were all Alexandrian-Byzantine or Western-Byzantine readings in the MSS, fathers, and versions considered as primarily representing the Byzantine Textform (in accord with the present hypothesis), all witnesses would appear far more “Byzantine” than by methods which exclude such co-alignments from consideration as Byzantine. Specific texttype alignments in either case naturally remain distinct on the basis of quantitative analysis.
Wherever possible, the raw number of MSS should be intelligently reduced. “Genealogical method” is accepted whenever such can be firmly established. “Family” groups such as f1 and f13 have long been cited under one siglum, and a few MSS are known copies of earlier extant witnesses. In many other cases a close genealogical connection can be established and thus mere numbers can be reduced in a proper manner. At times a group of MSS can be shown to stem from a single scribe with one exemplar (e.g., the eight MSS copied by George Hermonymus or the seven copied by Theodore Hagiopetrites); other MSS stem from a single recension (e.g., the ca. 124 MSS of Theophylact’s commentary on John, which differ so little from one another that Theophylact’s Johannine archetype readily can be reconstructed). Such numerical reductions restore the source text of the descendants and prevent a multiplication of totals for the sake of mere number. Such also includes grouping the various Byzantine subtypes (K1 Ka Ki Kr etc.) according to their hypothetical archetypes; these then become single secondary-level sources within the Byzantine Textform. The Kr subtype in particular is known to be late and secondary, having been produced out of the Kx type with lectionary and liturgical interests in mind. The MSS of that subtype resemble each other far more than they do the dominant Kx type. When recognizable genealogical ties can be established, MSS can be grouped under their reconstructed archetype and reduced to a common siglum, wherein number carries no more weight than its archetype.
What is not legitimate is to force the genealogical method to do more than it can, and to impose a genealogy which treats a texttype as a single witness. Less legitimate is to claim a given texttype or texttypes as the assumed parent(s) of other texttypes without demonstrable transmissional evidence. Such was the essence of Westcott and Hort’s hypothetical stemma and subsequent claims made with the sole intent of discrediting the Byzantine Textform. On the basis of transmissional considerations, the Byzantine-priority hypothesis would claim that the original form of the NT text would be more likely to manifest itself within whatever texttype might be overwhelmingly attested within the manuscript tradition, to the exclusion of all others.
Such appeals to “normality,” and is far more plausible than a piecemeal eclectic reassemblage of a hypothetical “original” which finds no representative among the extant witnesses. The texttype which on the basis of transmissional factors would appear to possess the strongest claim to reflect the original text can be termed the “Textform” from which all other texttypes and subtypes necessarily derive. The present theory asserts that the Byzantine best fulfills this demand, thus the designation “Byzantine Textform.” All competing forms of the text reflect “texttypes,” “subtypes,” or “families,” each of which developed transmissionally out of that original Textform.
Manuscripts still need to be weighed and not merely counted. This principle encompassed the intelligent reduction of witnesses based upon proven genealogical ties. Yet all MSS still need to be categorized regarding their text-critical value and “weight.” A basic component of “weight” is the transcriptional reliability of a MS. A later MS may preserve an earlier form of text; a well-copied MS may preserve an inferior form of text; a poorly-copied MS may preserve an otherwise superior form of text.
The effects upon transmission caused by individual scribal practice need to be taken into consideration when assigning a particular “weight” to a given MS at any point of variation. Thus, a determination of individual scribal habits becomes of prime importance. A MS whose scribe had a penchant for haplography or changes in word order will be of less significance when evaluating variant readings which parallel those types of error.
A scribe whose problems involved dittography or frequent substitutions of synonyms will be of less weight regarding readings reflecting those types of variation. The study of scribal habits of individual MSS has not taken place on a wide scale, despite the oft-repeated claim that “weight” prevails over mere “number” (one suspects the slogan is used more as a catch-phrase to discredit the Byzantine numerical majority rather than a call for establishing on solid grounds the true text-critical “weight” of individual MSS). Much more needs to be done in this regard, since the studies which so far have appeared have only scratched the surface. An evaluation of individual scribal habits would allow a better perception of the significance of individual MSS as they support or oppose given variants.
It is important to seek out readings with demonstrable antiquity. While the age of a MS is not as significant as the text it contains (which text is earlier than that particular MS), it is important to determine the earliest known attestation for a variant reading amid the extant evidence. A reading which lacks even a modicum of early support may be suspect. This is particularly so when the earliest testimony for a reading occurs quite late in the transmissional process. One problem is determining “late” versus “early.” While readings found in sources of a given date are at least as old as the witnesses involved, silence in the earliest period (due to a paucity of evidence) does not require rejection of readings solely because they lack early attestation. When extant testimony decreases, some loss of attestation is to be expected, and readings lacking attestation in the early period cannot be summarily dismissed. Methodological failure on this point neutralizes Westcott and Hort, since subsequent discoveries established the early existence of many readings which they had considered late and secondary.
Had such information been available to them, those readings could not have been as easily dismissed. Indeed, if most sensible readings were in existence by AD 200, caution should be applied when establishing the antiquity of a reading based solely on extant representatives. Chronologically “late” MSS are known to preserve earlier non-Byzantine texts well into the minuscule era; there is no reason to assume that minuscules preserving a Byzantine type of text fail to reflect a similar “early” character. Where, indeed, might one make a demarcation? While some may prefer a fourth-century boundary, there is no compelling reason to disqualify the fifth or sixth century, or even the ninth or tenth century. The real issue appears to be an opposition to any authoritative inroad for the Byzantine Textform.
There are valid reasons for considering all MSS extending into the late tenth or early eleventh century as “early” in regard to their texts. An explanation is in order: Apart from colophon information which would date the time of writing and the age of the exemplar, one cannot establish the actual antiquity of the text in any given MS. Since colophons of such detail do not exist, other means of assessing textual antiquity must be considered. Pertinent to this point are two major disruptions within transmissional history: “copying revolutions,” wherein numerous ancient MSS were subjected to massive recopying efforts, replacing their previous exemplars en masse.
The first “copying revolution” occurred when Christianity was legitimized under Constantine. The church of the early fourth century moved from a persecuted minority to an approved entity with governmental sponsorship. It is no coincidence that a change in writing material (from cheap and fragile papyrus to costly and durable vellum) occurred at this time. The earliest extant vellum MSS (i.e., the fourth- and fifth-century uncials , A, B, C, D, and W) and many later uncials would have been copied directly from papyrus exemplars. This is demonstrated by the lack of stemmatic or genealogical ties among the early vellum and papyrus witnesses. The common archetypes of closely-related uncials such as EFGH or SUVW as well as those of the relatively “independent” uncials up through the ninth century all are likely to have been early papyrus exemplars. This principle would not have been missed had the later uncials not been Byzantine in character. If correct, then all vellum uncials should be utilized when attempting to restore the original text of the NT. Most extant vellum uncials or (at least) their immediate archetypes would have been copied from papyrus exemplars, many of which would have preceded the change of writing material engendered by the altered political status of the previously persecuted church.
The second “copying revolution” occurred in the ninth century when handwriting switched rapidly from uncial to minuscule script. This change likely was initiated by Theodore of Studium and was swiftly accepted throughout the Greek-speaking world as a replacement for the more ponderous uncial script. Within a century and a half uncial script had ceased to exist among continuous-text NT MSS and soon after that disappeared even from the more traditional and conservative lectionaries. The upshot of this copying revolution was similar to what transpired following the papyrus-to-vellum conversion of the fourth century: uncial MSS of far earlier date were recopied in great quantity into the new and popular minuscule script and then destroyed. As Streeter noted,
In the ninth century there was a notable revival of learning in the Byzantine Empire. A natural result of this would be to cause Christian scholars to seek a better text of the Gospels by going back from current texts to more ancient MSS … An analogy may be found in the effect of the revival of learning under Charlemagne on the text of the Latin classics. MSS of the seventh and eighth centuries … are full of corruptions which do not occur in MSS of the subsequent period.
Part Seven will be published on July 15th.
This excellent article is reprinted with permission of TC: A Journal of Biblical Textual Criticism .
© TC: A Journal of Biblical Textual Criticism, 2001.