Friday, September 08, 2006

Textual circularity: unavoidable?

30
In the comment section to the previous post, criticism is leveled against Alands’ classification of manuscripts, including a reference to Bart D. Ehrman, “A Problem of Textual Circularity: The Alands on the Classification of New Testament Manuscripts,” Biblica 70 (1989): 377-88. (Read more about Aland’s classification at http://www.skypoint.com/~waltzmn/MSCategories.html.)

Now my question is: how can we avoid textual circularity when considering external criteria?

When text-critics speak of “external evidence” they usually consider the textual character of a witness among other things (date, geographical distribution, etc). The evaluation of the textual character (among other things) involves a process of weighing internal evidence at as many points of variation as possible, in order to find out about the overall quality of a witness, i.e. the external evidence. The external evidence is then applied again to individual problems. It seems impossible to avoid a circular reasoning, but the question is how to best control the procedure and make it as objective as possible.

Gerd Mink describes the method currently being developed and implemented at the INTF in Münster in his article “Problems of a Highly Contaminated Tradition: The New Testament. Stemmata of Variants As a Source of a Genealogy for Witnesses,” in Studies in Stemmatology II (ed. Pieter van Reenen, August den Hollander and Margot van Mulken; Amsterdam: John Benjamins, 2004), 13-85.

Mink states:

“An overall hypothesis can therefore only be acquired through approximation and iterative revision of all the intermediate results. I am referring to the circular reasoning which cannot be entirely avoided in textual criticism, but has to be controlled” (p. 46).

“In other words: witnesses are good because of their good variants, variants are good because of their good witnesses. This circle cannot be avoided, but it has to be controlled. We need a method, therefore, which can provide an overall view of the consequences of all the decisions we take, so that also the overall plausibility of what we are doing can be examined. In the present method [Coherence-Based Genealogical Method] this is done through an iterative process, especially designed to perform this examination” (p. 25).

30 comments

  1. My take on this can be found in the online commentary at:

    http://www-user.uni-bremen.de/~wie/TCG/Introduction-2-MSS.pdf
    Please read pages 2-3. Sorry, but I cannot express this in a few words only.
    Comments?

    PS: How is the syntax for a web link in the Comments?

    ReplyDelete
  2. Manuscripts can be grouped into families according to their affinity with one another. Or, when their level of diversity defies grouping, they can be plotted as points on a coordinate plane in such a way as to graphically portray their levels of nearness to one another (unfortunately, this method would involve mathematics, which some text critics are not open to using). This is a non-circular way of categorizing mss.

    After these objective groupings of mss are recognized, then a scholar can appraise each group's value according to his reconstruction of the text's transmission. This ranking must be honestly portrayed for what it is, a secondary category that is dependent on the hypothesized transmission.

    The Alands' textual categories reflect the latter approach absent the necessary explication of the former. There is some overlap between their rankings and the usual recognized text-types, but not a consistent one, as that is not the object of their system. When I first read their book as a beginning Greek student this confused me. I thought that in an introductory text on the NT text, I would learn the basics about text-types. For that I had to read Metzger.

    ReplyDelete
  3. 'When text-critics speak of "external evidence" they usually consider the textual character of a witness among other things (date, geographical distribution, etc). The evaluation of the textual character (among other things) involves a process of weighing internal evidence at as many points of variation as possible, in order to find out about the overall quality of a witness, i.e. the external evidence....It seems impossible to avoid a circular reasoning...'

    The large amount of circularity in the discipline of textual criticism indicates that we have too little hard data to work with. Thus we have to fill in the gaps with speculation - and lots of it. (I wonder in what other fields of study is speculation termed "evidence"?). Many of the conclusions are very suspect, and don't deserve the level of certainty often exhibited.

    Casey

    ReplyDelete
  4. ER: "When I first read their book as a beginning Greek student this confused me. I thought that in an introductory text on the NT text, I would learn the basics about text-types. For that I had to read Metzger."

    Indeed you had to read Metzger, because the Alands never embraced the text-type paradigm in the way Metzger did.

    Anyhow, I do agree that the old category system of the Alands was rather arbitrary and actually it overlapped with the Text und Textwert evaluation which was a bit different. But much has happened since the time the Alands wrote their introduction. Now the institute is implementing Gerd Mink's Coherence-Based Genealogical Method (CBGM). It will take to long time to describe the method in detail, and I am not the right person to do it (read Mink's own articles). However, there is an nice little introduction to what the Münster institute hopes to be a paradigm shift, away from the old "text-type" paradigm at the blog Ekthesis, with a summary of a paper that the new director at the INTF in Münster, Holger Struthwolf read:
    http://ekthesis.blogspot.com/2006/05/day-
    conference-on-tc-and-na-text-2-of.html

    Also, Eric, I think that the attempt to graphically portray the manuscript relationships is indeed what the CBGM does, perhaps not exactly in the way you describe it, but it involves statistics coupled with classical philological methods.

    I am very familiar with the type of manuscript grouping you describe, and it has a great value. However, one of the problems with the plotting you describe is that we cannot tell the stemmatic relationship of the text-type members. How do we know who is the "purest" Alexandrian witness in a certain section of the NT? Which MSS would you consider as the best representative of the Alexandrian text-type, and on what grounds? I think you will end up in a circular reasoning.

    Thus, we can see that manuscripts are related with a method similar to Aland's Teststellen method (but of course we prefer full collations without selection of test passages) and that is important, but the goal of the CBGM is to take further steps, and attempt to draw up as many probable stemmas as possible, on the local level, then on the global, and then back on the local level again in an iterative process.

    During my own work in the Catholic Epistles, I must admit that I have become more and more sceptical to the text-type paradigm.

    ReplyDelete
  5. I am certainly not prepared to attach any non-byzantine mss to any text-types, Alexandrian or otherwise, being far less familiar with the material than many members of this blog. I hesitate to define a text-type by 2 mss, such as B and P75; even if I did I would hesitate to label it with any geographic term, such as "Alexandrian;" and even if I did that I would hesitate to attempt any stemma for the small family. But, should anybody want to do that, they first need to recognize where in the mss pool any consanguinity exists. And this first step does involve grouping mss in a non-circular way. It may or may not end up being the case that any text groups can really be defined by geographic terms, or that stemmata can be identified in the groups. It may even turn out that the only groups of mss that deserve the label "text-type" are groups within the larger byzantine family. Nevertheless, if the question is only whether mss can be grouped in a non-circular way (contra the Alands) the answer must be yes.

    ReplyDelete
  6. Indeed mathematics are indespensible to a coherent comparison of manuscripts.

    Notice how many mathematicians have become textual critics.

    Andrew Wilson has written an entire chapter in his book critiquing the Alands' approach, going into much more detail than has been attained here. I quote just a brief excerpt:

    "Take, for example, the way that Reuben Swanson in his extremely useful series of collations of Greek Manuscripts in Horizontal Line format (about which the present author cannot pay enough compliments) describes why he has used Codex Vaticanus (B) as his exemplar:

    'Vaticanus, widely considered to be superior to all other witnesses, is chosen as the primary witness, for, as Sir Frederick Kenyon wrote, ‘Codex Vaticanus [is] the most valuable of all the manuscripts of the Bible’. Professor Kurt Aland, noted textual critic, likewise evaluated Vaticanus most highly, writing, ‘Among the uncials, B has a position of undisputed precedence in the gospels’.

    There are a number of issues we need to focus on from this brief excerpt. The first thing we need to notice is that Swanson assures us of Vaticanus’ great value by quoting certain scholars who consider it to be of great value. He says that Vaticanus is ‘widely considered to be superior to all other witnesses’. Swanson’s basic proof of the value of Vaticanus is an appeal to the collective authority of the scholarly community considered to have the authority to value manuscripts, rather than an appeal to any intrinsic features of the manuscript. However, since Swanson himself is among those same scholars who value Vaticanus highly, we could state the matter in rather more circular terms: most scholars consider Vaticanus to be of great value because most scholars consider Vaticanus to be of great value.

    Swanson's appeal to authority is a circular argument which really proves little at all. Of course, Swanson had to use some MS as the base for his collation, so his valuation of Vaticanus does no one any harm and we all praise Swanson for an incredible achievement and trust God will enable him to complete the entire New Testament in his series. But the fact remains that Swanson struggles to provide any better reason for the primacy of Vaticanus than an appeal to authority. This is in fact the premier argument used today in valuing external evidence."

    ReplyDelete
  7. Daniel Buck wrote:
    "But the fact remains that Swanson struggles to provide any better reason for the primacy of Vaticanus than an appeal to authority. This is in fact the premier argument used today in valuing external evidence.""

    This is nonsense.
    The superiority of Vaticanus is based on the superiority of its readings. This has been established again and again. Recently by Text&Textwert and also (if I am allowed to be so bold) in the online commentary (read the passage I mentioned above).

    ReplyDelete
  8. Wieland:

    Anyone can write a textual commentary (with a pre-conceived view of which manuscripts are best) and then use internal arguments to 'prove' that these manuscripts repeatedly have the 'best' readings.

    As Colwell wrote (forgive my paraphrase, I don't have the exact quote to hand): 'the more learned the textual critic, the easier it is to provide a justification of any reading'.

    This is precisely what Hort did, but as Salmon (the Chancellor? of Trinity College, Dublin, who was not at all anti-Hort), wrote 'That which gained Hort so many adherents had some adverse influence with myself - I mean his extreme cleverness as an advocate; for I have felt as if there were no reading so improbable that he could not give good reasons for thinking it to be the only genuine'.

    Someone writing from a Byzantine or Western textual viewpoint can also play the same game and delude themselves that they have proven that their favourite manuscripts repeatedly give the best readings and therefore reinforce their initial prejudices. It all depends on how much you believe your own propoganda.

    Tommy:
    My question would be: why is it necessary to have some iterative process in the first place? If someone is committed to considering individual variant readings on their own merits (i.e. an eclectic approach) then why does the consideration of a previous and unrelated variant need to be borne in mind when considering the merits of the variant reading under consideration?

    ReplyDelete
  9. I give in.
    A. Wilson wrote:
    "Anyone can write a textual commentary (with a pre-conceived view of which manuscripts are best)""

    If you would have read the above mentioned passage you would know that what I did was exactly the opposite.

    As they say, an ignorant can question more things in half an hour than all wise man in the world can answer in their lifetimes.
    If you are unwilling to read and understand the basics of textual criticism, then I see no basis for discussion.

    ReplyDelete
  10. AW:
    "My question would be: why is it necessary to have some iterative process in the first place? If someone is committed to considering individual variant readings on their own merits (i.e. an eclectic approach) then why does the consideration of a previous and unrelated variant need to be borne in mind when considering the merits of the variant reading under consideration?"

    Andrew, if we compare successively the readings of two documents in all their variations, we have ample materials for ascertaining the leading merits and defects of each. Readings authenticated by the coincidence of strong Intrinsic and strong Transcriptional Probability, or it may be by one alone of these. Probabilities in exceptional strength and clearness and uncontradicted by the other, are almost always to be found sufficiently numerous to supply a solid basis for inferencee. Moreover they can safely be supplemented by provisional judgements on similar evidence in the more numerous varations where a critic cannot but form a strong impression as to the probabilities of reading, though he dare not trust it absolutely. Where then one of the documents is found habitually to contain these morally certain or at least strongly preferred readings, and the other habitually to contain their rejected rivals, we can have no doubt, first, that the text of the first has been transmitted in comparative purity, and that the text of the second has suffered comparatively large corruption; and next, that the superiority of the first must be as great in the variations in which Internal Evidence of Readings has furnished no decisive criterion as in those which have enabled us to form a comparative appreciation of the two texts. By this cautious advance from the known to the unknown we are enabled to deal confidently with a great mass of those remaining variations, open variations, so to speak, the confidence being materially increased when, as usually happens, the document thus found to have the better text is also the older. Inference from the ascertained character of other readings whithin the identical text, transmitted, it is to be assumed, throughout under identical conditions, must have a higher order of certinty than the inferences dependent on general probabilities which in most cases make up Internal Evidence of Readings.

    Sorry, I couldn't resist it. Ok, add Westcott and Hort to the introductions to textual criticism that students should read. (Citation above, after "Andrew" from WH, Introduction, pp. 32-33).

    ReplyDelete
  11. There's no real circularity, but more of a question of working from the known to the unknown.

    Some readings are more clearly right or wrong on internal grounds than others. The idea is to use our knowledge about the variation units where the application of internal criteria is clear to identify which witnesses to use in conjunction with the external evidence for when the internal evidence is not so clear.

    This is more like a bootstrapping process than a circularity.

    Stephen

    ReplyDelete
  12. This comment has been removed by a blog administrator.

    ReplyDelete
  13. The syntax for weblinks is [a href="http://www.address.com"]Name[/a], but where [ stands for an angle bracket.

    Yes, all arguments ultimately are circular, but this is where one needs to take seriously Hort's comments about TC being negative. It more easily establishes that a text (e.g. D) has suffered corruption than that a text (e.g. B) is relatively free from it (though Hort argued that this could be established as plausible for B).

    Ehrman's own methods are not free from circularity, since readings become primary evidence for the existence of people who change texts for doctrinal reasons and alleged people who change texts for doctrinal reasons are used as evidence for which readings should be judged prior.

    ReplyDelete
  14. Wieland:

    I have read those pages of your commentary before, and for todays penance have now even read them again.

    You define the best manuscripts as those closest to the Originals, but then proceed to rate manuscripts on how consistently they follow YOUR chosen readings. You have put a lot of work into your commentary and there is a lot of good information in it, but I find myself often disagreeing with your conclusions, your preferred readings - even the ones you consider relatively certain.

    There's the problem. Your rating system does not show which are the BEST manuscripts, it simply shows which are YOUR preferred manuscripts.

    I know ... you will reply by telling me that many other textual scholars often agree with your preferred readings. But this is nothing more than the appeal to authority that you previously condemned as nonsense. I agree that such an appeal to authority is nonsensical, but despite the fact that appeals to authority are usually classed among the logical fallacies, some textual critics can find no better argument, bless them.

    I too have written a textual commentary on Romans and Galatians (forgive my self-referencing, it wouldn't be allowed on Wikipedia) and when I did my 'checksumming' at the end to see which manuscripts were the best, I found to my delight that my 'original' text aligned marvellously well with certain manuscripts that, if I had been honest with myself, I preferred from the start. I defy even J. K. Elliot to honestly declare himself entirely free from preferences for manuscripts - I suspect it is humanly impossible.

    Wieland, your defence seems to be that you did your 'checksumming' at the end of the work on the commentary. We already knew that. But you seem blind to the fact that what you get out at the end is simply what you put in right from the beginning. This is the circularity problem: garbage in, garbage out, and there is accounting for bad taste. Just as in hermeneutics, there is no one without their presuppositions.

    Hort's 'method' is simply the same self-deception described more elegantly. By his 'morally certain' readings we might substitute Burgon's 'ring of truth'. By Hort's 'strong' probabilities he simply means his own strong preferences - and Hort had stronger manuscript preferences than virtually any other textual critic dead or alive. Why should we appoint a judge whose blatant prejudices have been seen in previous cases to decide the most delicate and difficult cases?

    ReplyDelete
  15. I do not think that Wieland should feel obliged to answer this question. It would be better asked privately. Contributors are merely asked to respect the ethos of the blog.

    ReplyDelete
  16. Daniel Buck wrote:
    "But the fact remains that Swanson struggles . . .

    Wieland, I only point out that the quotation regarding Swanson is none of mine, but quoted from Andrew Wilson's book.

    I would add that when reading the Online Textual Commentary of the Gospels, I often run across such a valuation as "this mss is best in ch. 5-10" or such the like. I always interpret this to mean that 'good' or 'best' refers strictly to how closely it reads to NA27; not a particularly useful definition for me, especially given that errors in NA27 are still being discovered and published on this blog.

    ReplyDelete
  17. DB:
    "... not a particularly useful definition for me, especially given that errors in NA27 are still being discovered and published on this blog."

    I do not think that the errors that are "being discovered" in the apparatus of NA27 is of much relevance since their correction will hardly affect the reconstructed text at any point. I have not used the NA27 in my work on Jude, since it is superceded by the ECM, but I think 1505 is at some point wrongly cited somewhere. This will have absolutely no effect on the reconstructed text of Jude. On the other hand, the implementation of the CBGM has resulted in three changes in the reconstructed text of Jude, two of which involved words in square brackets. The third change in Jude 5 will be discussed at the SBL annual meeting in a paper read by Klaus Wachtel.

    In any case it is apparent that we end up with different reconstructions and we should have respect for each others arguments as long as they are coherent in relation to the overall view. Personally I am not surprised by examples where B has an inferior reading, but the question for me is the overall quality of the manuscript (cf. citation of WH above). Perhaps we can agree that this is a relevant question even if we favor different manuscripts. In my own experience B offers a very good text and I have worked mostly in the Catholic Epistles. I will save the details for my dissertation...

    ReplyDelete
  18. T. Wasserman wrote:
    " The third change in Jude 5 will be discussed at the SBL annual meeting in a paper read by Klaus Wachtel."

    I am looking forward to this. But I think it is not surprising that the CBGM cannot reconstruct the NA reading, because there is no MS support for exactly this reading.
    :-)

    http://www-user.uni-bremen.de/~wie/ECM/ECM-Variants-23JohnJude.html

    ReplyDelete
  19. WW: "I am looking forward to this. But I think it is not surprising that the CBGM cannot reconstruct the NA reading, because there is no MS support for exactly this reading"

    Well, as M.A. Robinson has pointed out there are over 100 verses in the NA27 without MS support. However, it all depends on how you define the variation-unit. When Wacthel treated this verse in his dissertation (Der Byzantinische Text; in the textual commentary on the Teststellen) he treated "the reading" in four subsets of variants, i.e. four unrelated problems, which I think was sound. For example, the subject of APWLESEN is a separate problem from the problem with APAX inside or outside the OTI-clause. Thus, it is not surprising that in this verse (one of the most complex in the New Testament) there is no support for the whole variant reading, but indeed for each subset. The interesting thing is what goes together.

    The significance of a sequence of "original" readings and subvariants preserved or not preserved in manuscripts is a related and interesting discussion (see e.g. M. A. Robinson's and M. Silva's respective contribution in D. A. Black, ed., Rethinking Textual Criticism).

    ReplyDelete
  20. Inasmuch as Maurice Robinson is never far from any discussion on this blog, he is eminently qualified to quickly clear up a small matter of dispute between Wieland Willker and I:

    Wieland's online commentary mentions 1350 Byzantine mss of John as containing the Pericope Adulterae. I suspected this figure as being inflated with fragmentary mss of John that don't actually retain that passage. Wieland's number is based on T und T, but I know that Dr. Robinson has independently collated this passage.

    So what is the actual number?

    ReplyDelete
  21. Daniel,
    Dr. Robinson's paper says, "In John 1665 collations of the PA area were performed among the continous-text MSS...All other continuous-text MSS (totaling around 1350) include the PA."

    His full paper is avaiable here:
    http://www.tren.com/e-docs/search.cfm

    ReplyDelete
  22. I am often out of the loop; but never too far from the blog....

    To help resolve all "dispute":

    Whatever numbers were previously published are clearly too low.

    The number cited in my "Preliminary Observations" article represented collation data as of 1998 -- before I had collated the bulk of lectionary MSS containing the passage. A number of continuous-text MSS and lectionaries also have been added due to subsequent collation opportunities (this includes additional MSS obtained by the INTF since 1998).

    As to the issues raised:

    Figures are not inflated, nor do fragmentary MSS affect the totals (a MS either contains the PA or not). Most MSS containing the passage tend to be complete or nearly so.

    Additional note: MSS not containing the PA were tabulated separately, even while collated over the portion 7:51b-8:13a.

    >Wieland's number is based on T und T

    True; however, Klaus Wachtel, Ivo Tamm and I jointly had gone over the TuT listing during its pre-publication state, in order to weed out errors and to make certain that we were on the same track.

    >So what is the actual number?

    Hmm...wish I knew, but I haven't yet updated the overall total (the raw-count data are on computer, but I have not run the totals).

    I can say that 43 additional MSS containing the PA were collated at the INTF during the Spring of 2005. If these are added to my lectionary log total, the number of MSS + lectionaries that contain the PA is at least 1350+43+470 = 1863 total MSS (there are somewhat more than 280 continuous-text MSS that do not include the PA (excluding lectionaries, where the PA only appears sporadically, when certain specified saints happen to be honored therein).

    ReplyDelete
  23. This is just to note that, with regret, for the second time in the blog's history I have used the facility to delete a comment (anonymous). As some may remember, it used abusive language.

    All civil, intelligent and relevant comments continue to be welcome.

    ReplyDelete
  24. MAR:
    "the number of MSS + lectionaries that contain the PA is at least 1350+43+470 = 1863 total MSS"

    Excellent. I'm sure Wieland Willker and I can both use this information.

    Thanks Dr. Robinson!

    ReplyDelete
  25. 100 verse in NA27 without MSS support? Really?

    Does that include stuff below the line too?

    A hard reference for this would be handy.

    Regards,

    David.

    ReplyDelete
  26. David,

    I think the issue has been mentioned on Wieland's discussion list (search the archives). The "hard reference" I am afraid is to a paper not yet published. The paper was read at a colloquium in Canada. I had the chance to read and discuss the interesting paper when I met the author just a week before in Münster. The paper will be published in a conference volume scheduled for publicatin in 2007:

    Maurice A.Robinson, “Rule 9, Isolated Variants, and the ‘Test-Tube’ Nature of the NA27 Text,” in From Text to Translation: The Proceedings of the Bingham Colloquium held at McMaster Divinity College, Hamilton, Ontario, 26-28 May 2005. Edited by Stanley E. Porter and Mark J. Boda. Grand Rapids: Eerdmans, forthcoming 2007.

    ReplyDelete
  27. I hope I'm not posting this too long after the original post here for my question to be addressed.
    But Tommy, the exthesis link you gave is good, but not detailed enough for me to get a grasp of the topic. Can you recommend any other specific English language treatments of the text-type question and providing an understanding of the newer understanding that the Institute is employing. I found an English article by Mink dating to 1985 on computer assisted textual research. But I gather that it won't address this topic in full.

    This issue seems like a major topic in NT TC. There are alot of correlary topics that arise. I'm surprised I haven't seen more about it. Does Ehrman's update of Metzger's Intro address this?

    ReplyDelete
  28. ER:

    You can try to follow the link on this blog to the INTF homepage, and then look at their publications "Veröffentlichungen" and check Mink's list of publications.

    I received an off-print from Mink of this one, which is probably the best treatment thus far:

    Gerd Mink, “Problems of a Highly Contaminated Tradition: The New Testament. Stemmata of Variants As a Source of a Genealogy for Witnesses,” in Studies in Stemmatology II (ed. Pieter van Reenen, August den Hollander and Margot van Mulken. Amsterdam: John Benjamins, 2004), 13-85.

    In fact this item is available on-line (you might have to install a special reader plug-in to your browser):

    http://site.ebrary.com/pub/benjamins/Doc?
    isbn=1588115356

    I cannot tell you if Ehrman brings up the subject in his introduction, since I haven't had chance to read it yet.

    You can expect that the Editio Critica Maior, in their very next publication which will be the fifth installment of the Catholic Epistles (the supplementary studies to these epistles), will describe the method and supply examples, along with a textual commentary, at least at those points where there are bold dots involved.

    You can also search the archives of Wieland's list to see if the issue has been discussed. Unfortunately, Mink's method is rather complicated so not many scholars seems to have a grip of what exactly he is doing.

    Another stemmatic approach is applied by Stephen Carlson, who supplies an introduction on his homepage:

    http://www.mindspring.com/~scarlson/tc/

    A more general description:

    http://www.skypoint.com/~waltzmn/Stemma.html

    Good luck!

    ReplyDelete
  29. Thanks, TW. I hope other readers don't miss that great link you just gave for the free online Studies in Stemmatology. I intend to read the Mink chapter.
    I also took a look at his publications at the INTF and, unfortunately for me, my german isn't good enough for most of them I think.
    But I was really as interested in reading more on the status quaedtionis of text-types as I was in stemmatology, which was brought up in that exthesis link. Any recommendations for that?

    ReplyDelete
  30. ER: But I was really as interested in reading more on the status quaedtionis of text-types ... Any recommendations for that?

    Eric,

    I can highly recommend that you read Wachtel's thesis, Der Byzantinische Text ... (which you will find listed under his name at the INTF homepage, where you found Mink's works), where he develops his perspective of "Traditionssträngen" and the "Entwicklungsprozess" of the streams of tradition. You can start reading reviews of his dissertation. Elliott had one in Novum Testamentum 38,4 (1996), 409-11.

    Good luck!

    ReplyDelete