Thursday, January 24, 2019

Coherence at 1 John 4.19

29
To follow up from yesterday’s post, here is the textual flow diagram and local stemma for 1 John 4.19/5 in INTF’s CBGM. Reading a = omit; b = αυτον; c = τον θεον; d = τον κυριον

In this diagram, none of the readings have perfect coherence as all show at least one witness needing a source outside its own attestation. In particular, reading a leads to b once and vice versa. But still, not bad coherence as a whole. Reading c has quite bad coherence, developing from b multiple times and maybe from a a few times as well. Reading d develops from b which is no surprise. Witness 6 is noteworthy in that it has close ancestors with reading b and c, suggesting that either could lead to the shorter reading as I said before. But 617 with reading b is closer and so more likely as a whole.

If we keep the same CBGM dataset but set the initial text (A) at this point to reading b instead of a, the coherence gets worse for a and becomes perfect for b. In short, more support for b and less for a.
To make things more interesting, here is the same attestation and local stemma but now from my custom version of the CBGM where the initial text (A) is defined as the Byzantine text across the entire Catholic Letters. Here we get perfect coherence for reading b and see it leading to the other three.
As with other types of evidence, the evidence from coherence is a balance of probabilities, but here it does give more support to reading b as the initial text than I expected it to. That doesn’t necessarily mean it’s enough to overturn other evidence for reading a, but it’s at least enough to make me reconsider it. It would be most ironic if this were a place where the CBGM supports Maurice Robinson’s Byzantine Priority position!

29 comments

  1. I'm not sure we should ever say "the evidence from coherence ...".

    ReplyDelete
    Replies
    1. How about “the evidence from these colorful graphs…”

      Delete
    2. 'The impression given by how we present these colourful graphs ...'

      Delete
    3. “The overall agreement between witnesses combined with the strongest direction of textual influence as determined by our editorial judgments and displayed here in simplified but colorful graphs suggests that…” I don’t know. That just doesn’t have the same ring to it.

      Delete
  2. Thanks for looking into this!

    ReplyDelete
  3. Peter Gurry,
    Not that I think we should remove later witnesses from consideration, but what would happen here if we removed all witnesses later than the 1100s from consideration? (I'm thinkin', "Latin/Greek 629? Really?") Does the flow look the same?

    ReplyDelete
    Replies
    1. Peter Gurry,
      Well, if the method is valid, it should be as true in 1100 as it is in the 1500s, shouldn't it? Would it be very inconvenient to crunch the numbers (or texts) eliminating post-1100 MSS and find out?

      Delete
    2. Yes it would be. And I don’t know what you mean by “just as true in…” Coherence has to do with witness relations so naturally the relations depend on the witnesses included.

      Delete
    3. Peter Gurry,
      Basically, I mean that if the method is sound, then it should point in the same direction when used in 1100. We can't go back in time to 1100, but we can remove all post-1100 MSS from the picture. In which case, what happens, if anything?

      Delete
  4. To return to the original post, I fail to see anything "most ironic" in that CBGM happens to support a Byzantine reading at this point, given that it already has done so many times within the General Epistles and Acts as noted in the ECM editions.

    The mystery is why CBGM and its colorful graphs can be correct only part of the time.

    ReplyDelete
  5. OK, now that we know the coherence, how does it help us with establishing the initial text?

    ReplyDelete
    Replies
    1. It probably doesn’t offer much help here. Readings 1 and 2 have good enough coherence to be the initial text I would think. We probably have to rest the argument on other evidence.

      Delete
    2. Leaving the Byzantine "b" reading aside, I fail to see why reading "a" with its limited and disparate level of support would have more "coherence" than the better supported "c" reading that apparently is not considered a viable candidate.

      Would the CBGM gurus care to explain for those of us incapable of perceiving the inner workings of the Mystery by explaining something that transcends the colorful diagrams? Basically, this is asking who or what actually determines that "Reading c has quite bad coherence, developing from b multiple times and maybe from a a few times as well"? And does it really? Or is this simply something suggested by the arcane computer-based machinations?

      Delete
    3. MAR, I am not getting a strong vibe from your comment that you actually want to understand how the CBGM does what it does. But I'll give you the benefit of the doubt.

      What determines the coherence of reading c is the level of textual agreement between those readings that attest it. It is also affected by the editors' own decisions though this is less determinative (as evidenced by the worse coherence in my CBGM as well as INTF's). So the short answer is, the weaker level of agreement among the witnesses of c primarily determines that reading c has worse coherence than the other readings. Make sense?

      Delete
    4. Even though I am not a CBGM partisan, it would help if the process could actually be understood in plain and clear terms. So yes, I do actually want to understand how CBGM works; but no, the response still does not clarify my understanding.

      With what I see as the bemuddled and ambiguous concept, "the level of textual agreement between those readings that attest it" -- is this saying that the overall percentage of agreement among witnesses throughout an entire book (as in the Text und Textwert data) becomes determinative of how the text supposedly will "flow" in any given individual variant unit (which seems to be an illegitimate totality transfer)? Or is there some other supposed evaluation restricted solely within an individual variant unit that results in the determination of "flow"? Feel free to clarify and help my understanding.

      Delete
    5. Well, that’s disappointing. I noticed that the online tool has a local stemma, favoring reading a over the rest. I can’t tell whether they made this decision before or after looking at coherence. If before, I think it means that they feel the internal evidence clearly favors reading a.

      Delete
    6. SC,
      it has become clear to me, at least, that as much as Peter Gurry and others claim that the editors decisions are not decisive, they actually quite often determine the text that is listed.

      Tim

      Delete
    7. This comment has been removed by the author.

      Delete
    8. Well, I do think that the editor's job is to decide but also to present the evidentiary basis for the decision. My interest with the CBGM I lies I ascertaining how much it is bringing to the table for a particular variant and how much is already baked into the cake with the local genealogy.

      Delete
    9. SC, now you've done it -- I keep hearing the MacArthur Park refrain: "...melting in the dark...Someone left the cake out in the rain; I don't think that I can take it, 'cause it took so long to bake it, and I'll never have that recipe again."

      Delete
    10. SC,
      I’m not opposed to the idea of the editor making a decision because I don’t know who else can. My point is that the CBGM and it’s defenders protest too loudly that the editors decisions do not ‘really’ determine the output!
      Tim

      Delete
    11. My impression is somewhat different, but maybe I haven't been following the CBGM proponents as closely as you have. Do you have an example of this?

      Delete
    12. I believe "presupposition" is the answer to your initial inquiry good Doctor. -M.M.R.

      Delete
    13. The Ill coherence of reading c is very likely due to Aleph. We shall most probably be seeing the Sinai codex given no quarter when it parts ways with B/Vaticanus in future CBGM analysis. Why? It's a lone ranger- And it's coherence with it's very own sister mss. B is even lackluster. -M.M.R

      Delete
  6. I've had a thought about the "my custom version of the CBGM where the initial text (A) is defined as the Byzantine text across the entire Catholic Letters" in connection to your previous last post. If I understand it correctly, the initial initial text (A) is a subset of the variants where the local genealogies are clear. Presumably this means that the 'a' reading should be evident from internal evidence alone. But if the 'a' reading is set to the Byzantine reading in all variation units, then there ought to be cases where the Byzantine reading is only weakly supported by internal evidence (because sometimes it is inconclusive) or counter-indicated by the internal evidence. In other words, I wonder if there is a tension between the local genealogies and the hypothesis that the Byzantine reading is the initial reading?

    ReplyDelete
  7. Peter, this is ever so slightly off-topic for this post, but I'm going to ask nonetheless. In regards to the CBGM, could you perhaps make a post at some point where you give a simple example demonstrating how pregenealogical coherence acts as a check and balance against any subjectivity introduced during the editorial process at the level of local stemmata?

    In other words, if you start with a given local stemmata, which incorporates editorial decisions that are subjective, show how considering the pregenealogical coherence highlights mistakes that might have been made in forming the local stemmata, and thus how to correct it.

    A real world example which you have encountered would be ideal.

    Perhaps there's already an example of this out there somewhere?

    ReplyDelete
    Replies
    1. Thank you for the post Dr. Gurry. I noticed that Metzger gives no heed to the possibility of a simple scribal slip (ht./sound) in his textual commentary. I was wondering if you considered this in your evaluation? ....ENAUTONOTIAUTOC or
      ....ev autov oti autos. -M.M.R.

      Delete