ETC Interview with Maurice Robinson: Part 1

If you’re still wondering, the answer to my quiz of last week is none other than that Byzantine Beatle, Maurice A. Robinson. Maurice also happens to be the first participant in what I hope will be an ongoing series of interviews with text critics. In the past, we have interviewed Bart Ehrman, Dan Wallace, and Stanley Porter and these were well received. So I thought we should continue the tradition. I don’t have any detailed criteria by which to pick our interviewees (suggestions welcome), but I can say I am quite pleased with those who have already agreed to be interviewed. There are many familiar names on the list, but also some lesser-known or younger scholars that I am excited to introduce to our readers. So without further ado, I present our first interview.

---

As a regular commentator and sometime contributor at the ETC blog, Maurice Robinson is no stranger to regular readers. But despite the blog’s great fame, he is most well-known for his work editing and defending the Byzantine textform. He teaches at Southeastern Baptist Theological Seminary in Wake Forest, NC where he was recently named research professor of New Testament studies. He’s been interviewed a number of times before, but I thought there were a few things those interviews didn’t cover, especially the final question of part 2. Enjoy!

Peter Gurry: Many readers might be surprised to learn that you worked with Kenneth W. Clark during your master’s work. Can you tell us how that relationship has (or hasn’t) influenced your own view of textual criticism?

[Maurice A. Robinson] I began studying with Clark (1898–1979) in 1971 during my MDiv program at Southeastern Baptist Theological Seminary (where I currently serve as Research Professor); this was arranged by the then text-critical professor here, since he said I already knew more about the subject from previous self-study than did he. Clark at that time was already emeritus from Duke, having retired from teaching in 1967, but he genuinely was excited about my interest in the field, since at that time very few students anywhere were becoming interested or involved in the subject. As a result, Clark and I began and maintained a very good relationship from 1971–1977 (when I moved to Texas for my PhD studies), despite our evangelical versus liberal theological differences.

My position at that time was one of reasoned eclecticism, basically following the Metzger-style theory and praxis; Clark, however, in various of his publications had already raised serious questions as to whether that or any type of eclectic method really represented a solution rather than a symptom (a theme later discussed by Epp in 1976). Clark therefore strongly encouraged me to study, heavily read, and critically examine various alternative views, including those favoring a primarily external and transmissional approach to the text as opposed to those theories that placed a more subjective emphasis on internal criteria (including both thoroughgoing and reasoned eclecticism). In essence, what Clark strongly suggested was a return back to primarily external principles such as espoused by Westcott and Hort, but without their unsupported speculative historical baggage regarding a “Syrian recension” being the creative cause of the Byzantine Textform.

One of Kenneth Clark’s
most often cited essays.

Clark thus was advocating a return to Hort’s “theoretical presumption” (Introduction, 45) that a majority of existing documents should reflect on the basis of transmissional principles a majority of predecessor documents at each stage of transmission, moving back toward the autograph — such in essence was already a “Byzantine priority” position, even though neither he nor I had developed a term for such at that point. As Clark clearly expressed back then, were he 40 years younger, he would have begun research and writing toward a new paradigm that would more strongly favor the Byzantine text as the most likely transmissional scenario leading to the establishment of the original text. Since he was already in his mid-70s, however, he strongly encouraged me to move in that direction, and that of course is exactly what I did; this occurred slowly, however, since even in my 1975 ThM thesis (which Clark guest-supervised) my position was still in process and necessarily hesitant. Only after I became associated with William Pierpont in 1976 (an event that for two years overlapped with my continued studies with Clark) did my position actually crystallize into a Byzantine-priority position based upon an accompanying transmissional theory.

[PG] I’m always interested in how editors of Greek New Testaments produce their editions. Can you tell us a bit about the mechanics involved in producing the Robinson-Pierpont GNT?

[MAR] Since my last sentence conveniently segued to the point of my association with Pierpont, the mechanics actually began with him. Pierpont originally had followed the Westcott-Hort position that he had been taught in the 1930s but in the mid-1960s began to question such. As part of his research from 1965–1975 he had already accurately tabulated most of the primary Byzantine readings within the NT, using von Soden’s bold K apparatus designations as the starting point, also taking note of the various Byzantine subgroups cited in that apparatus. He further compared these data with other resources available at that time (including other critical edition apparatuses, published collation data, and the various IGNTP volumes for Mt Mk and Lk); these allowed him to correct several errors of von Soden at various points. So basically, by the time Pierpont contacted me in 1976, the primary legwork toward establishing a basic Byzantine text had already been accomplished. From that point on, Pierpont and I worked together (mostly by postal mail correspondence) to fine-tune, tweak, and adjust his basic data by further examination of existing materials, as well as by exploring internal arguments raised both in support of and opposition to any particular Byzantine reading. The basic principle for establishing the text, however, remained externally based, and did not deteriorate into a mere form of eclecticism having a general Byzantine preference.

Pierpont’s original data had been typed in corrective note format (in that era on a manual Greek typewriter), and after three years of our joint checking and rechecking the data, that typescript was published in 1979 as an appendix to Green’s Interlinear Bible NT volume (for many years now that appendix has no longer appeared in various reprints of that work). This note-form document listed whatever changes to the base TBS TR text of that edition that would alter that text to reflect Byzantine readings; this initial “edition”, if it can be called such, predated by three years the appearance of the 1982 Hodges-Farstad Majority Text edition, on which I served as a proofreader.

Obviously, matters did not stand still from that time onward, but further checking and rechecking as newer materials became available, as well as revision of various points of internal consideration regarding individual variant units continued apace for another decade. With the advent of the personal computer era after 1980 (more ancient history!) I was able to begin digitizing the Greek NT in 1985, manually typing in the entire Stephens 1550 TR base text from Berry’s Interlinear; this was followed by a second adjustment to reproduce the TBS Scrivener TR, and finally the revised Pierpont notes were utilized to create the complete sequential Byzantine text in electronic form. These texts all became available in the DOS-based Online Bible software program in late 1986 or early 1987, and thus constituted the first actual “publication” of the RP Byzantine Textform in continuous (and searchable) form; later electronic developments added parsing/declension and lemmatization data, but those remain external to the textual base issues. At some point, the electronic form of the text caught the eye of Original Word Publishers, who produced the first print version in 1991 (limited to 1000 copies). That first published form of the text had certain limitations, leaving the text in a somewhat less adequate state, particularly in regard to display of divided Byzantine readings. Still, our work continued, and with further format and revision refinements (mostly aesthetic and stylistic) occurring into the late 1990s, at which point Chilton Book Publishing picked up the work, prepared a formal page format for the work, and sent reading proofs to us, all before Pierpont’s death from cancer in 2003. The actual publication occurred in November 2005, with announcement and presentation at the Valley Forge ETS meeting.

Since that time, Jeffrey Dodson and I have collaborated to produce the 2010 Edition for Beginning Readers, incorporating a near-final correction of various accent, breathing, punctuation, and capitalization faults that had appeared in the 2005 edition. That Reader’s Edition includes verb parsing data and definitions for all words occurring less than 50 times, with an appendix covering the data for all words and parsing forms occurring more than 50 times (presenting more information than other Reader’s Editions). So at this point, this is where matters stand; the RP main text remains identical in the 2005 and 2010 editions, although differing at points from the 1991 or earlier forms.

[PG] You’re currently working on a large project on the Pericope Adulterae (PA)? Can you tell us a bit about its aims and scope?

[MAR] Around 1995 I began to collate the PA from existing Greek MSS, starting at Duke University with the Kenneth W. Clark collection. A decade earlier I had attempted to obtain grant funding for travel and research to Münster to be able to pursue the project from their extensive collection of microfilms, but grants were not forthcoming (one reviewer actually asked “Why would anyone want to do that?”; another claimed the information obtained would be “useless for any text-critical purpose”; another said it would offer “no information that would be helpful in determining textual transmission or interrelationships among the manuscripts”). So much for grant agencies. As a result, only when my first sabbatical arrived in 1997–1998 did I in fact travel to Münster to formally begin the project; this was funded only by my continued salary and by a generous contribution from William Pierpont. Also, thanks to the courtesy of Ulrich Schmid (who that year was a research fellow at NIAS in Holland), we were able to sublease his apartment at a major discount, without which we still would have had financial difficulty in completing the post-doctoral study). I thus began collating the PA and its border regions (John 7.51b–52, 8.12–13a) from the microfilms in Münster in September 1997, beginning with the papyri and uncials, followed by the minuscules according to century. At that time I really did not think I would get beyond the 12th century MSS in a 10-month session.

To my surprise, I not only collated the PA (in MSS possessing such) or its border regions (in MSS not including the PA) in all available papyri, uncials, and minuscules by the end of April 1998, but still had time remaining. I then continued the project with the largely unexplored region of the Lectionaries, finishing the sabbatical term with over 100 lectionaries also completed (these data became the basis for my Filología Neotestmentaria article on the PA). Since I had begun the collation work in the lectionaries (and found it fascinating, James Leonard to the contrary), I determined to continue the research, and so again with no further grant funding (why bother to apply again?), I returned to Münster several times for 3-month stints (Summer 1999, Summer 2000, Spring 2005) in order to complete collation of the lectionary PA portions. Beyond that point I have spent several short 2-week sessions in Münster adding data from various new films obtained that include the PA region. Beyond all this, I have further supplemented the data through various online sites, including Dan Wallace’s digital photographs (and I eagerly await more from his current expedition!), also photographs Jeff Cate kindly obtained for me of MSS in California. I hope for even more material from Athonite monasteries (assuming Paul Anderson ever can get permission from them to allow serious researchers actually to work with the photographs).

Having said all that by way of background, the “scope” of the research is basically covered. The “aim” of the project, however, is to provide a comprehensive database citing all variation found in the MSS containing the PA or its surrounding region, including misspellings, itacisms and scribal blunders. Not cited in the electronic database, but included on the original collation sheets are matters regarding line division, nomina sacra, and movable Nu. The nomina sacra words appearing in full form, however, are cited in the database. The comprehensive data should not only allow for examination of the various lines of transmission (involving even more than von Soden’s seven), but also to permit skilled researchers or computer programs to postulate the likely lines of descent of the MSS that comprise each line of transmission.

The form of presentation in the PA database is not as complex as that of ECM, nor does it appear in the same format as Text und Textwert; rather, as Barbara Aland had suggested to me, it is a straightforward collation against the Oxford 1873 TR (noting lacunae where necessary). Thus, at any given instance all MSS that are neither lacunose nor specifically cited necessarily concur with the base collation text. At this moment, the database collation data involves 1494 continuous-text MSS containing the PA or some portion thereof, 268 continuous-text MSS that do not contain the PA but have all or part of the border regions, 495 lectionary MSS that contain the PA, and (with electronic data entry still in progress) over 650 lectionary MSS that do not contain the PA, but do have the border regions (being for them the end of the Pentecost lection). Almost certainly these numbers will increase as new documents come to light from the various photographic research projects currently underway.

[PG] The current editors of the Nestle text have argued that our NT witnesses cannot be meaningfully grouped into the traditional text-types (Western, Alexandrian, etc). Assuming you disagree with this assessment, how do you distinguish one text-type as over against another?

[MAR] Contrary to the prevailing concepts currently expressed, I see no reason to abandon the basic texttype designations as providing a convenient grouping terminology for MSS that associate more closely with each other than with those that associate within other grouping situations. This does not mean that I view the concept as representing “recensions” per se — a term that implies a formal revision process at some particular place and time; rather, I view texttypes in John G. Griffith’s sense of “near-neighbor clusters,” that is, MSS that by their agreement in a reasonable pattern of readings are recognized as somewhat but not entirely related to one another. Nor in this would I suggest that the traditional texttype names necessarily restrict their presence to any particular time or locale, even if those existing MS clusters might tend to fall primarily within certain regional or temporal boundaries. As Wachtel has noted, given the overall high percentage of agreement regarding the bulk of the NT text, all our MSS basically represent a single primary text from which they have descended; the clustering that occurs at places of disagreement (variant units) is what differentiates one group of MSS from another — these divisions reasonably can be called texttypes in their own right.

As to distinguishing one texttype from another, the principle is simple: start with the overall Byzantine consensus text, noting wherever a significant representative sample of any particular group of MSS tends to differ on repeated occasions. Such MSS then would represent some sort of “texttype” or “subtype,” even if all such MSS only mostly but not always concur on every reading. Given these parameters, I have no real difficulty in recognizing as texttypes either an “Alexandrian” (or “Egyptian”, or whatever term might be preferred), a “Western” (even with limited Greek support, but bolstered by various Old Latin and Syriac sources), and even a “Caesarean” type of text, based upon a particular pattern of readings reflected in its respective clusterings. What should not be done with the texttype concept, however, is to utilize a “majority of texttypes” approach as some solution presumed to determine the autograph (cf. Sturz in this regard) — such becomes merely another form of “nose-counting” and should be rejected solely on those grounds; more seriously, it actually places the relatively unstable and questionable “Western” text in the position of ultimate arbiter of originality (e.g., Alex+West and Byz+West combinations would tend to dominate over the less-frequent Alex+Byz combination, the Caesarean generally being disregarded in such considerations).

[PG] One of your arguments against reasoned eclecticism is that it produces a “test tube” text, one that does not exist in extenso in any given manuscript. But it seems to me that the logic of this criticism would require you to adopt the text of a single witness in its entirety. In other words, how does your criticism of the Nestle text not apply to your own GNT taken as whole?

[MAR] Perhaps the essay should be carefully re-read in view of what it says as opposed to what people seem to presume it says. The point of that essay never implied that the so-called “test-tube text” concept (Kurt Aland’s term) would have to exist “in extenso in any given manuscript.” We all know that in general no two MSS — even of the same texttype — tend to agree in all particulars throughout their entire content or even within a single book (Pickering, however, seems to have demonstrated that within the tightly controlled fam. 35 group of MSS some of them in fact are totally identical within shorter individual NT books, but this really does not say much in the overall sense).

The actual issue raised in my essay was that in the current critical editions (NA^27/28/UBS^4/5) Aland’s “test-tube text” concept does exist within extremely short segments of text (where for convenience I used single verse units as the demarcation to establish the point). The basis for testing thus became a single verse unit with at least two variant units present. From these, the NA²⁷ apparatus was utilized directly to demonstrate in 105 separate instances that within the particular identified single verses the MSS that support one main text reading were not the MSS supporting the remaining variant(s) within the same verse. Thus, each of those 105 short single verses, published as the main text of the NA/UBS edition, when considered as a unit, presents for each of those verses a form of the text that demonstrably has zero support from any existing MS (and also neither from versions or patristic quotations).

That, and that alone represents the purpose and scope of the “test-tube text” essay, with the stated contrast from my theoretical perspective that in any similar single-verse situation one would never find the Byzantine text reflecting a similar scenario. Instead, within any single-verse segment — even when the Byzantine MSS are sharply divided — a significant proportion of its MSS still would support the primary Byzantine reading: quite the opposite of what has transpired in the case of NA/UBS. Thus, when put in the proper perspective — contrary to various misassumptions and distortions that have been expressed — the issue never involved the entire NT text, nor the entire text of any NT book or chapter. The only issue was whether the NA/UBS text by its own eclectic praxis has itself departed from any real MS continuity over relatively short segments of text; quite demonstrably, so it has done, nothing more, nothing less.

To further expand the point, I continued the study (presented in an ETS paper), extending the scope to two-verse segments within NA/UBS (none of these encompassing any of the 105 single-verse segments tabulated in the original essay). That second experiment resulted in a further doubling of the extent of the NA/UBS “test-tube text,” with an additional 210 two-verse segments that as printed demonstrably have no actual existence in any extant MS, version, or father. Once more, in those two-verse segments, the Byzantine text again can be demonstrated to maintain a viable form of text among a significant number of its MSS, even while NA/UBS reach zero support in the same single- and two-verse segments (now totaling 315 whole verses). The question thus becomes one of de facto conjecture in regard to the short verse units described, even when each separate variant unit within those verses might have one or more MSS in support. As noted, such de facto conjecture does not apply within the same parameters to the MSS comprising the Byzantine Textform. I hope this now clarifies the point to everyone’s satisfaction.

---

Part 2 is posted here.