This past semester, for the first time in a long while, I taught a few sessions on New Testament textual criticism. I tried to refresh myself on some of the changes in the field in the last couple decades. I’m well aware that I don’t have an especially strong grasp of the details of the Coherence-Based Genealogical Method (CBGM) that is being used by the Institute for New Testament Textual Research to produce the Editio Critica Maior of the books of the New Testament. So, I wanted to take the opportunity to brush up on this somewhat technical methodology.
The predominant textual flow diagram for non-fragmentary witnesses in the Catholic Epistles; image source: Peter J. Gurry, A Critical Examination of the Coherence-Based Genealogical Method in New Testament Textual Criticism (Brill, 2017), pp. 60-61
First off, let me offer a hearty endorsement of Tommy Wasserman and Peter Gurry, A New Approach to Textual Criticism: An Introduction to the Coherence-Based Genealogical Method (SBL Press, 2017). Concise, clear, and really, really useful for those of us struggling to get our heads around this approach. I still have some questions, but I feel a lot better having read this book.
One of the confusing things about the CBGM is the definition of terms. A major point on which I still lack clarity relates to the Ausgangstext, the “initial text” that is the outcome of the application of the CBGM. In Wasserman and Gurry’s helpful glossary, the “initial text” is defined as “the hypothetical witness from which all the extant witnesses derive.” To me this sounded quite similar to what we would call the “archetype” in classical textual criticism (although in the framework of the CBGM what is at issue are immaterial witnesses rather than material manuscripts). But I see now that Klaus Wachtel draws a sharp line between the “initial text” and the “archetype” in the following way:
“…the definition of the term ‘initial text’ must be carefully distinguished from the archetype of the tradition, on the one hand, and from the original text of the author, on the other. The archetype of the tradition was a real manuscript, the copy by which the transmission started that put forth the manuscripts we have―and many more that are lost. The original text of the author predates the manuscripts we have by more than a century in most cases. The initial text is the hypothetical reconstruction of the text as it was before the archetype of the tradition emerged. The initial text is the result of methodical efforts to approximate most closely the lost text of the author based on all relevant evidence, not excluding any trace of transmission predating the archetype.” (“Conclusion” to The Textual History of the Greek New Testament: Changing Views in Contemporary Research, SBL Press, 2011)
I think I understand what is being argued here, but I am not quite clear on–I’m not sure exactly how to say this–how far behind(?) the archetype one can go with “traces of transmission that predate the archetype.” I suppose it would depend upon the quality of the “traces”–in this case references to NT passages in the texts of early Christian authors (which, themselves have material manuscript histories with which we must deal). If I understand correctly, then, the “date” of the initial text will vary from book to book (or even from passage to passage?) in the New Testament, depending on the quality of (the texts of) the available manuscripts and “traces of transmission preceding the archetype.”
Gurry has a very informative dissection of Wachtel’s chapter and other contributions to the Textual History volume in his more detailed monograph, A Critical Examination of the Coherence-Based Genealogical Method in New Testament Textual Criticism (Brill, 2017, see especially the third chapter,”Recovering the Initial Text”). And he comes close to addressing the issue that is nagging me:
“…the solution to the misunderstanding about the use of the term ‘initial text’ is to distinguish the meaning from the referent. The term means ‘that text from which the extant tradition descends,’ a definition which allows it to refer to any number of historical entities including the author’s original text, the archetype, or some editorialized text subsequent to both. The benefit of this understanding is that it may allow those with differing opinions about the referent to nevertheless agree on the editorial text itself.”
I think I like this distinction, but I miss a definition of “tradition” in the sense that Gurry uses it here [[see Update below]], which seems different from the use of Wachtel, who refers to “the archetype of the tradition” in direct contrast to “the initial text.” For Gurry, it would seem that “tradition” also includes “traces of transmission predating the archetype.” But maybe I am missing something?
I wonder if it would be helpful to think in concrete terms: 2 Timothy. I think the earliest surviving material manuscript of 2 Timothy would be Sinaiticus (generally assigned to the fourth century), although I guess the earliest surviving immaterial text might be considered 1739, perhaps pushing things back to the time of Origen. Would the initial text of 2 Timothy that we could produce through the CBGM then be considered a text supposedly extant in the middle of the third century, with the exception of those passages cited in Clement of Alexandria, for which we could then say our initial text dates to closer to the end of the second or the beginning of the third century (on the assumption that our critical text of Clement is sound)?
Or am I misunderstanding something in the method?
Update 10 June 2020: Gurry elaborates via e-mail:
Variant Readings 
There’s a lot going on, and the INTF isn’t entirely clear. Let me make a few points:
1. “Initial text” or Ausgangstext seems to have (at least) two senses in the ECM. One is a particular goal of their method, a “hypothetical witness from which all extant witnesses derive.” Another sense is the working text of the CBGM algorithm, designated by “A” in the textual flow diagrams and iteratively refined. At the end of this refinement, it is a hypothesis that the two senses are the same.
2. Because the ECM establishes the text with the aid of conjecture, the goal of their text is by definition prior to the archetype, as the archetypal text does not depend on conjecture.
3. An archetypal text partially corrected by conjecture cannot be equated to an original text either. For one, there may well be necessary emendations we just cannot discover.
4. Some scholars have rightly problematized the notion of a singular original, but I would go further and hold that, even with due recognition of this plurality, we nonetheless lack the evidence and methods to pick out and distinguish any particular authorial draft or scribal copy prior to the archetype.
Thanks, Stephen. I appreciate especially your point 2 on conjectural emendation, which I think is quite helpful in seeing the difference between the CBGM’s “initial text” and the archetype (Gurry also mentions conjectures in his book).
Pingback: A Question About the “Initial Text” of New Testament Documents | Variant Readings | Talmidimblogging
From Mike Holmes:
Re: “initial text” and “archetype”: Yes, the initial text is, as you rightly note, “quite similar to what we would call the ‘archetype’ in classical textual criticism,” and the differences are subtle. For CBGM, the archetype is the text as established by the evidence of the surviving witnesses (in classical terms, the result of recensio). But it is possible that the evidence of the surviving witnesses is flawed—that all extant MSS preserve at one or more points a faulty text, as at, e.g., 1 Cor 6.5, where διακρῖναι ἀνὰ μέσον τοῦ ἀδελφοῦ αὐτοῦ is almost certainly not what Paul dictated (see Zuntz, Text of the Epistles, 15, for the argument). Yet that is the reading of the archetype. In such cases, the classical remedy is emendatio (which the ECM has utilized in Acts and 2 Peter). The result of this stage is the “initial text,” a state of text in which errors of omission or confusion in the archetype have been (at least in theory) repaired. Now the next question is: what is the relationship between the initial text and the authorial text (a problematic term whose discussion I’ll leave aside for now). It is at least theoretically possible that the initial text (and the archetype after it) may transmit errors of addition (i.e., interpolation); to give one possible example from 1 Corinthians, what about μὴ ὑπὲρ ἃ γέγραπται in 4.6? Here examinatio comes into play, as one assesses the probability that the transmitted text is the authorial text (this step led Westcott & Hort to their category of “primitive errors”). Or, for those who think 2 Corinthians is a composite document: every extant witness presents the same 13-chapter form of the letter. If the composite theory is right, then some sort of editorial intervention occurred between the “authorial” and the “initial text” stages. Pragmatically speaking, for most of the NT the archetype = initial text = authorial text. But rather than assume this, as NT scholarship has almost universally done, the CBGM turns this traditional “taken for granted” assumption into a hypothesis to be tested against the evidence—a useful heuristic move.
This response comes with a warning that perhaps I don’t understand the CBGM! However, last year I had to prepare a talk for a local theological group (most of whom did not know Greek) on textual criticism and the CBGM.
In answer to the question at the end of your post, I think you may be misunderstanding the CBGM – certainly on the basis of your concrete example. For the CBGM, the Initial Text is a hypothetical text which coherently “explains” all subsequent variants, ie that produces a genealogical coherent text flow with a single point of origin for each variant found in the various texts. In CBGM terms, there will be an optimal local substemma for each variant and the combination of decisions in these local substemma will produce the Initial Text. Different practitioners define the Initial Text in different ways, so some regard the IT as an hypothetical version of the autograph, and some would say it is the text in the form in which it began to be copied and distributed, ie the origin of all the textual tradition which has survived. But as Gurry suggests in the passage you quoted, this makes no practical difference to how the CBGM is employed.
In your concrete example, it does not matter at what point in history we can identify a particular version of the text, whether it be the 2nd, 3rd, or 4th century or later. What the CBGM is doing is taking all the surviving versions of the text, either complete or fragmentary and producing a hypothetical text (the Initial Text) which is genealogical coherent, ie “explains” coherently the point of origin of all the variants which have arisen.
I would like to express my appreciation of your blog, “God’s Library”, your articles and your kindness in sharing pdfs of the latter. Sorry if I have wrongly “corrected” you over the CBGM!
Thanks for the input, Graham. I take your (and Gurry’s) point that the initial text could in theory be taken to refer to an author’s text or the origin of the surviving textual tradition. I also appreciate the distinction that the CBGM makes between manuscripts and texts. I think where I am getting confused is the question of how the product of the CBGM is to be related to historical time. This is something that Gurry and Wasserman mention on several occasions, as in the following from p. 108: “the global stemma can discard the date of the ink and papyrus (or parchment or paper) when relating texts in a way that a stemma codicum cannot. More importantly, this means that the global stemma cannot be read as telling us which manuscript was copied from which; it only tells us about textual influence. That means we cannot directly use the dates of manuscripts to date sections of the global stemma.” I take this to be similar or identical to the point you are making in your second to last paragraph. So, I guess what I am struggling with is how to characterize the text resulting from the CBGM. At an intuitive level, I somehow want the age of the manuscripts (or even the age of the text, as in the case of 1739) to have some kind of role in the process. Wachtel has touched on this point: “The global stemma is (or rather, will be) a hypothesis about the relative chronology of the development of the text as preserved in the manuscripts. All we know about the history of the manuscripts can be incorporated in the CBGM procedures of constructing the local stemmata of variants and the optimal sub-stemmata of witnesses. The relative chronology of the development of text as shown by the global stemma can and will be put into relation with known historical data like the dates of the manuscripts carrying the textual witnesses, the dates of authors citing from the respective writings, and the dates of translations” (http://purl.org/TC/v20/TC-2015-CBGM-history.pdf). It’s that last step I’m having trouble seeing (the 2 Timothy example was my effort at thinking through that).
As far as I can tell, they haven’t actually produced a global stemma for the Catholics or Acts, and so that last step has never been done.
Stephen, as I understand it, work is still going on to produce a global stemma for the Catholic Letters and I presume they are also working on Acts but I have not come across specific reference to this. Seeking to combine and reduce the optimal substemma to a single global stemma is a huge task. Whereas I see the merits of the CBGM as a valuable text critical tool, as I indicated in a previous comment, I struggle to see what exactly the global stemma will tell us when it is finally available.
Thank you for this, Brent. I think you may be highlighting some distinction between the thinking of Wachtel and Gerd Mink. I read Mink as very much wanting to distance himself from the concept that the global stemma is in any sense a history of the text – which is why the CBGM speaks of text flow. Since all mss suffer from some level of contamination (which indeed the CBGM is seeking to address) then the place in the textual flow will differ for each variant in each ms. In the circumstances, it is difficult therefore to see how the global stemma could be tied to the point in time of particular text represented in a particular ms since the text of a particular ms will have potentially a different genealogical relatuonship within each variant. The optimal substemma reduces that down to show flow but does not identify in time how it relates to an exemplar which first contained that variant..
If I am understanding Mink’s approach correctly, then one of my problems with the CBGM is that I don’t really see what the function of the global stemma actually is if it doesn’t do the sort of thing you are wanting. However, given how the CBGM works, I don’t see how it can. At the level of producing local stemma and optimal substemma I see entirely how the method is helpful but the logic of combining the substemma to reach a global stemma seems to me to have no purpose. Maybe I’m just restating your problem in a different way?
The term Ausgangstext (abbrev: AT) comes from Translation Theory = English “source text”, French texte de départ.”
Yes, but it seems to have taken on a life of its own within the framework of the CBGM.
Thanks Brent, I will give a brief response on the ETC blog at some point … but I am so busy with other deadlines and stuff before the holiday that this may take a while. I have too little time for this fun stuff. 😦