Tuesday, January 26, 2016

What Are Text-Types For?

Martin L. West
In his book Textual Criticism and Editorial Technique (1973), Martin West raises an important issue for the use of text-types in NT textual criticism. 

Here is West:
When the critic has established that no stemma can be constructed, how is he to proceed? He must, of course, see what groupings are apparent among the manuscripts, and whether the individual groups can be analysed stemmatically… (p. 42).
Here we should pause and note that, as Colwell noted in his essay on the genealogical method, this is exactly what Westcott and Hort did in rejecting the “Syrian” text. They applied stemmatic principles, not to individual manuscripts, but to groups. Having done this, they were able to exclude the Syrian text from consideration on the principle that it was purely derivative. We might call this principle eliminatio textuum descriptorum.

Westoctt and Hort’s stemma has since been modified and the results have not usually been treated with such stemmatic rigor. But West goes on to explain how such groupings can still be useful to the critic:
…even if they [groups] cannot [be analysed stemmatically], he can treat them as units in his further cogitations, provided that they have a sharply defined identity. Thus he reduces his problem to its basic terms.
This reduction of the problem is a major reason why text-types have been so valuable in NT textual studies. Where you have four manuscripts in a tradition, you don’t need to reduce the material. But for the NT such reduction is a huge benefit, even a necessity. No one can keep dozens let alone hundreds of manuscript relations in their head and then apply them to specific variations. But three or four relationships is no trouble at all. Hence the value and appeal of text-types. They “reduce” the problem.

But note the key qualification in West’s sentence. The critic can treat groups as units in his further cogitations, provided that they have a sharply defined identity.

If this is true, then it would seem that the use of text-types in our text critical “cogitations” is in trouble since no such definition exists. Even Eldon Epp in his excellent essay on “textual clusters” says that “the tricky issue, of course, is determining, in percentage terms [West’s ‘sharply defined identity’], what extent of agreement in readings joins members into a group, and what degree of separation in agreements determines the existence of a separate group” (“Textual Clusters: Their Past and Future,” [2013], p. 571). 

Unfortunately Epp doesn’t have an answer to this “tricky issue” which makes me wonder if our failing effort to define text-types is an indication that we’re trying to solve the wrong problem. Maybe trying to reduce the problem is our problem and we should start looking for ways to use more manuscripts (not less) in studying the history of the text.


  1. And there are various levels of groupings as well. Taking Schmid's work on the textual history of Revelation as an example, there's a big difference between certain, relatively uniform, groups within the Andreas tradition or even the Koine stream on the one hand and the "P47 ℵ Origen" or A C Oecumenius" on the other. (Note that I didn't refer to the Andreas text-type as it may not be as coherent a unit as we used to think.)

  2. Epp: "“the tricky issue, of course, is determining, **in percentage terms**, what extent of agreement in readings joins members into a group, and **what degree of separation** in agreements determines the existence of a separate group”

    The fallacy here lies in the attempt to make Colwell's definition of a texttype into a standard measuring rod, which is quite likely an exercise in futility.

    I would suggest that neither a specific percentage of agreement (Colwell's 70%) nor any specific degree of separation (Colwell's 10%) will actually reflect a valid demarcation of "texttypes" per se. Rather, one should seek "near-neighbor clusters" of manuscripts (John G. Griffith's term) that happen to associate together among themselves within the manuscript base but without any specific percentages being imposed in order to establish the existence and validity of such group clusters (which then can be termed "texttypes" in their own right).

    1. Dr. R.,
      How would such "near neighbor clusters" be identified? Would a method like Tim Finney's "Multivariate analysis (MVA)" work? This method seems to group clusters without the percentage category.


    2. The article by John G. Griffith (e.g. "Numerical Taxonomy and Some Primary Manuscripts of the Gospels, JTS 20 [1969] 389-406) would seem to offer a reasonable pattern for manuscript grouping that would provide some sort of texttype-oriented arrangement.

      Multivariate analysis seems to provide something similar, but given the programming and mathematics involved, I remain reluctant to declare anything beyond my known limits of incompetence.

    3. Dr. R.,
      Yeah, it, MVA, is way beyond my competence levels as well!😀 I will have to read the article by Griffith.


    4. Maurice, I think the questions remains as to what value such groupings would have. To put it another way, what degree of demarcation is necessary for what types of use? Also, any quantitative analysis involves counting. It can't be avoided.

    5. Peter,

      If the primary issue is related association within a clustering type of concept, quantity of support -- even while obviously present -- becomes less important than the fact that MSS x y z tend to associate with each other far more than they do with MSS a b c d e or with MSS f g h i etc., even allowing for some minor degree of overlap due to association based on shared readings.

      Recognition of distinctions such as these should then allow for some reasonable group approximations, I would think.

    6. Dr. R.,
      Textual clusters appear to have greater meaning to those who put emphasis on documentary evidence. I would imagine, subject to correction, that for the eclectic of varying stripes, their reliance on internal evidence makes textual clusters irrelevant.


  3. The other question is the fallacy of (presumed) equality. Since that phrase is used more in social settings, we can talk of what in logical terms is the fallacy of equivalence, or the fallacy of false equivalence.

    Here it means that those items that are placed in comparable sounding categories or groups would start off with a presumption of having essentially equal significance. After all, it has been simplified to a few groups.

    Thus a category of Greek manuscripts with thousands of items placed in the category (e.g. Byzantine) is therefore equivalent in substance to a Greek ms. category that is a whiff in the wind (e.g. Western) or a modest number (Alexandrian).

    Ironically, in practice, still using byproducts of the defunct hortian approaches, the Byzantine is the one considered irrelevant.

    Steven Avery

  4. Agree with Dr. Robinson,

    Furthermore, both Wisse/McReynolds rightly saw both terms i.e. group/cluster as more appropriate rather than the whole termed in hortian fashion as a "type". This is correct.

    However, the Byzantine groups/clusters have a "core" which rarely diverges among all the groups/clusters identified among the Byzantines hence text is only appropriate taken as a "whole". If a more keen understanding and/or interest of the Byzantines was practiced in NT-TC this common misunderstanding would fade away. But, this is part of of our own research and work.

    Paul Anderson

  5. May I point out that because Martin Litchfield West (1937-2015) was more of a classical scholar studying with an interest in "textual criticism", without a particularly Bible text viewpoint, without the common indoctrinations, his statements tend to be far more sensible than New Testament modern scientific textual critics.


    Textual Criticism and Editorial Technique (1973)
    Martin Litchfield West

    "This is not to say that the age of a manuscript is necessarily a guide to its quality. Recentiores, non deteriores: that is the famous heading of a chapter in which Pasquali protested against the tendency to equate the two terms, and showed that true readings are sometimes preserved only among the latest manuscripts. ... very old copies such as papyri sometimes disappoint expectation by giving a worse text than the medieval tradition instead of a better one. The quality of a manuscript can only be established by reading it. And when an opinion has been formed on the quality of a manuscript, it can be used as a criterion only when other criteria give no clear answer. The absurdity of following whatever is regarded as the best manuscript so long as its readings are not impossible (continues with example)... " p. 50

    When we choose the 'more difficult' reading, however, we must be sure that it is in itself a plausible reading. The principle should not be used in support of dubious syntax, or phrasing that it would not have been natural for the author to use. There is an important difference between a more difficult reading and a more unlikely reading. p. 51


    Steven Avery

  6. Is there a resource available that provides an authoritative up-to-date listing of NT text groupings? It seems to me that serious textual critics have abandoned the model of the three or four major text-types. But those are still the default that non-TC specialists learn in most NT classes.

    //This reduction of the problem is a major reason why text-types have been so valuable in NT textual studies. Where you have four manuscripts in a tradition, you don’t need to reduce the material. But for the NT such reduction is a huge benefit, even a necessity. No one can keep dozens let alone hundreds of manuscript relations in their head and then apply them to specific variations. But three or four relationships is no trouble at all. Hence the value and appeal of text-types. They “reduce” the problem.//

    This is true, but only inasmuch as the groupings are reliable. It's probably better not to reduce the material if that reduction is by way of imposing a poor model on the data.

  7. Eric, no authoritative source to my knowledge. But you might check out Tim Finney's website. He is doing some interesting things with textual groupings.

  8. Eric, read Eldon Epp's essay in The Text of the NT in Contemporary Research. I would not go so far to say that the concept of text types has been abandoned (although certainly their geographical connections). However, the aim of the CBGM is to provide "higher resolution" so to speak.

  9. @Tommy, "been abandoned (although certainly their geographical connections)", but by whom? As I don't think you intend to say that there isn't anyone who is still using this geographical terminology (since we both know there are), you must mean something like, 'no one who I consider to be a serious text-critic', right? And probably only for Greek NT manuscripts and not for the Latin tradition?

  10. Leaving aside the "geographical connections" (assuming one seriously doesn't consider the Western text predominantly Western or the Alexandrian text predominantly Egyptian, etc.), there is no reason why one who prefers to maintain the concept could not relate to the various texttypes under Kenyon's clear and unassuming general designations α β γ δ (even if such happens not to be my preference).

  11. Maybe another way to get at the issue would be to ask what would falsify one of the particular text-types? Maybe debates about the Cesarean text-type prove the point. How can it be so hard to know whether an entire text-type exists or not?

  12. Peter,

    The issue should not be considered in absolute terms, since MSS that pertain to some general pattern of readings (a key element that would characterize texttype affinities) will not concur with such a pattern at all points. Rather, individual MSS that otherwise cluster together in relation to shared readings still will demonstrate differing degrees of deviation from any overall pattern, even while not aligning with patterns that deviate significantly from that found in an initial cluster. The issue remains a general clustering of MSS in relation to their approximation to a shared pattern of readings rather than any particular fixed entity more sharply defined in texttypical terms.

    In other words, I don't have any major problem in distinguishing a predominantly Alexandrian reading from one that is predominantly Byzantine, even if the MSS supporting either side are not always identical in relation to each variant unit.

  13. Text-types were used, with a lot of finagling and logic deficiencies galore, to achieve the Alexandrian Vaticanus-primacy Critical Text, staring in 1871-1881.

    Later, Aland gave some lip-service to putting aside text-types, and then accomplished essentially the same results by simply assigning categories, where even a late minuscule like the forgery 2427, perhaps by Simonides, was assigned a Category One assignment .. simply due to general Vaticanus affinity. Ergo, back-door Vaticanus primacy.

    Now, if someone abandons or dismisses text-types and categories, and still keeps the edifice of the same text without a new and now somehow sensible conceptual base ---

    ... is their resulting textual structure anything other than a Potemkin Village?

    Steven Avery