Tuesday, September 26, 2017

Trovato on why we need to face the ‘awkward problem’ of conjectural emendation

The following post is from Paolo Trovato who has been a welcome contributor to the comments the last few months. It can be good to have an outside perspective sometimes and that is what Paolo gives us. For those who need an introduction to his work, see my interview here and here.

Dear all,

Even though I don’t belong to your community, I am really interested in your posts and comments. One of the aspects that strikes me is the following position, which seems to be a very popular position in Greek NT studies: We have so many MSS that at least one of them must preserve the original reading of any passage of NT.

I only quote two different but clear examples of similar statements
PG: “it is assumed that manuscripts to which we don’t have access contained original readings. But if we don’t have access to them, how do we know they had original readings? From the argument as stated, we don’t. We need other evidence.”

MAR: “there is a reason to postulate a greater likelihood of the original reading being present in our existing MSS, and that indeed is the quantity of extant data that has been preserved — particularly in contrast to the extremely limited number of works of Latin antiquity (or even Greek classical antiquity) that virtually force conjecture by their paucity.

To put it statistically: even assuming, say, that 60% of all Greek NT MSS that ever existed have been lost, the odds remain extremely favorable that the remaining MSS would preserve in relative proportion the same readings that might have appeared in those now-lost documents; if so, then there is no good reason to postulate much if any supposed “loss” of original readings as matters now stand.”
(Both quotations are drawn from a very interesting post of last January.)

I have probably already bored some of you with my partial attempts of applying to the textual criticism of Greek NT some standard approaches of genealogical textual scholars of Classical and Romance literature. For those who are still interested in a comparison with different textual schools and methodologies I will try, very briefly, to explain why, in my opinion, these positions don’t seem very likely.
  1. The message of the Gospels was so new, moving and fascinating that it created very soon a strong interest (i.e. a strong demand of copies), but also persecution, theological opposition among the faithful, different degrees of standardization etc.
  2. There is a gap of centuries between the lost original(s) and the lost first copies of the Gospels and the few oldest extant witnesses
  3. Every act of copying produces fresh innovations., mostly but not always insignificant. No witness is a Xerox copy of its exemplar. In the domain of classic and Romance transmissions, four or five major, significant innovations every 10 pages seem a realistic rate for the progressive entropy of the textual transmission of any not too short text. So, a copy of 4th rank should feature 4 or 5 × 4 (= 20 or 25) significant innovations every 10 pages.
  4. In the first decades of printed books (1470–1530), every time we happen to know the runs of a book and we look for its extant copies, we find a loss rate of about 85% or 90%. I suspect that, the value T(ime) being higher for Greek NT MSS, their loss rate must be higher than that of the printed books of the Renaissance.
  5. Let’s suppose that after the original(s) of NT, there existed 3 copies of 1st rank, 9 copies of 2nd rank, 27 copies of 3rd rank, 81 copies of 4th, 243 copies of 5th rank, 729 copies of 6th rank, that is, a complete tree of 1,092 copies. (I will stop here for the sake of simplicity, even if I imagine that the complete tree, that is, ALL the MSS of Greek NT that ever existed were probably no less than 20,000). If we assume an optimistic random loss rate of 90%, the extant MSS should be 98 out of 1,092. 
Even without introducing other factors (the obsolescence of materials, the old fashion layout etc.), mere math suggests that most of the extant MSS must belong to ranks 5 and 6 (972 out of 1,092 MSS). It is probable that only one or no MS of ranks 1 and 2 (12 MSS) could survive to a rate loss of 90%. But even if a couple from these ranks was preserved nowadays, they should present 4 to 10 major innovations every 10 pages. No need to say that most of the surviving MSS must belong to rank 6 and that, as a rule, they should be marred by a good number of innovations.

Therefore, I would say that, at least in a few passages, not even my colleagues devoted to textual criticism of the Greek NT, can feel completely freed from the awkward problem of conjectural emendation.

(By the way, Dante’s Commedia, which I have been studying for 15 years, is transmitted by 600 extant MSS plus 200 fragments. Notwithstanding the fact that the oldest extant witnesses are only 15 years later than the completion of the work, a few passages do exist in which emendation is really necessary).


  1. As a supplement to the above consider not just The Difficult Reading Principle but The Extremely Difficult Reading Principle which says there can be an extreme inverse relationship between degree of difficulty and Manuscript support. Using GMark as example which was the problem child of the early Gospel family that Patristics did not like to talk about, the most difficult reading may be "angry" of 1:41. Here Jesus is described with a strong negative emotion for no apparent good reason. ETC generally does not think "angry" is original but I think the majority of ETC consider it a serious candidate? "Angry" only has one direct Greek Manuscript support (Bezae = an oldie and a goodie and probably about as good as any Manuscript for Markan difficult readings). It does also have a few (again quality) Latin Manuscripts too. So this is getting close to no Manuscript support.

    The above suggests that if a reading was sufficiently difficult it's possible that there might not be any extant Manuscript support. This uncertainty needs to be combined with the general uncertainty regarding anything that was written 2,000 years ago.

    1. A similar variant, and one that is even more firmly within the territory of conjectural emendation, is that found in 2 Pet 3:10. There is no known Greek MS support for the reading οὐχ εὑρεθήσεται, but, like Bezae's reading in Mk 1:41, equivalents for it have been preserved in versional evidence (syᵖ, syʰ ᵐˢˢ, sa, and possibly cv, according to NA28).

      The conjectural emendation of 2 Pet 3:10, however, seems to reflect the view that the difficult reading principle has its limits. The reading εὑρεθήσεται, found in a handful of Greek MSS, is more difficult, but it is so difficult that the ECM / NA28 editors could not consider it original. The conjectured reading, therefore, is actually much easier than the nearest surviving reading.

      All this makes me wonder about the rules for conjectural emendation. Must conjectural emendations by their nature be easier than some existing readings (as with 2 Pet 3:10)? And if not, then how would a textual critic know what to conjecture in the other cases (e.g., lacking any MS evidence for the "angry" reading in Mk 1:41, what would lead a critic to propose such a reading)?

  2. 'Angry' may be the exegetically harder reading, but it is the linguistically easier one. Outside Christian texts the verb for 'compassion' is extremely rare and there are good grounds for suggesting that a highly literate Greek speaker who had not previously encountered specifically Christian or Jewish vocabulary would have found it extremely difficult.

    1. I could concede the point that a highly literate Greek speaker may have found "compassion" difficult, but I don't yet think I've read a satisfactory theory for why such a speaker would choose - out of all the other words a highly literate person would have to choose from - a word like "angry" , especially when that word is, as you note, exegetically harder.

      To put it another way, perhaps one that returns a bit to the OP, if a highly literate reader was simply trying to fix a linguistic difficulty, it defies probability that he would just happen to select a term that, exegetically, conveys the exact opposite meaning.

      Bottom line, I think your linguistic point is true but moot, as the variation was most likely exegetically motivated, and thus must be evaluated on exegetical grounds.

  3. I'm not thinking of anyone deliberately changing compassion into anger. For words ending in -ισθεις, οργισθεις is the most common in Greek by some margin. From a transcriptional perspective it is not only easy, it is the easiest.

    1. Sure, absolutely. I don't think we can look only at the ending though; that big honking splank on the front just, to my mine, makes the letter sets far too different.

    2. Paolo Trovato9/27/2017 6:20 am

      Gmark 1.41. Even If I don’t remember a discussion on this aspect in handbooks or papers I read, I would say that in the case of the reading “angry” the very fact that the final part of the variant “angry” is paleographically and morphologically close to the variant “filled with compassion” (PW “For words ending in -ισθεις, οργισθεις is the most common in Greek”) weakens the interest of the possible lectio difficilior, which could depends on the bad condition of the exemplar in that passage. Secondly, Metzger (A Textual Commentary) remembers that maybe the exchange depended on “confusion” (of the scribe!) “between similar words in Aramaic (e.g., in Italian “h” is not pronunced and at school I repeatedly confused “angry” and “hungry”). Last, but not least: why Gesus should have been “angry”, when a sick man asked him to be freed by his illness. Apart my general lack of expertise a/o ignorance in this field of textual criticism, for which I can only apologize, there is a general point that I want to underline. The criterion of the “harder reading” requires a refined application, it cannot be appplied as a rule, mechanistically, if the textual critic think that the hard variant isn’t fit to the context .

    3. Paolo Trovato9/27/2017 7:46 am

      Errata corrige:
      Gmark > GMark
      could depends > could depend
      “between similar words in Aramaic > “between similar words in Aramaic"
      pronunced > pronounced
      textual critic think > ...thinks


  4. Since one's of Trovato's main points is that conjectural emendation may be unavoidable "at least in a few passages," I thought it would help to provide a few palpable examples. I'd be interested in hearing others' opinions on some of better-known and more recent conjectures that have been proposed for the NT. I'll list the few that I can think of off the top of my head:

    1. The Lost Ending of Mark: Not sure if any ETC regulars hold to a lost original ending of Mark, but for anyone who's thought about it, what would the extent of a conjectural emendation here be? It seems like the best one can do is to propose that such an ending existed, having nothing to say of its content.
    2. The Epilogue of John: The omission of Jn 21 is the most extensive case of a "negative" conjecture I can think of. Unlike the previous example, this type of conjecture has the convenience of certainty regarding content (the longer reading is already known, and the conjecture simply omits it). I'm not familiar with the arguments for this conjecture, but I'd be interested to hear them. Given that the surviving MS tradition for Jn 21 is not noticeably fractured, it seems odd to me to view it as a secondary development.
    3. Gal 4:25: In his dissertation (The Text of Galatians and Its History, pp. 222-226), Carlson follows earlier textual critics in proposing that the phrase τὸ γὰρ Σινᾶ ἐστὶν ἐν τῇ Ἀραβία, found in some form or another in all MSS, was originally a gloss to the text, though Paul himself may have been responsible for it. Carlson's arguments are compelling, but I don't think I've seen any comment on this proposal in ETC.
    4. Jud 1:5: Wasserman's critical text of Jude has the reading εἰδότας ὑμᾶς ἅπαξ πάντα, ὅτι κύριος here, which is not found in any Greek MS. But whether or not this counts as a conjectural emendation is debatable; as Wasserman's breakdown of the variant in his commentary shows, the individual components of this reading all have some MS support, so the conjectural nature of the reading depends on where we draw the lines for a "reading".

    I'm sure there are plenty of other examples I'm missing, so feel free to add to these. I'm curious to see which passages have the "most unavoidable" need for conjectural emendation.

    1. Paolo Trovato9/26/2017 7:42 pm

      I'm curious too, of course ;)

    2. Dr. Trovato,
      Is your experience with CE, is it used when all manuscripts read the same text, like in the example of 1 Corinthians 14:34-35? If so, who or what is the basis of such? What would the basis be for a 'negative' CE as described above for John 21?
      Finally, if you will bear with my ignorance, if manuscript evidence exists for a reading, even in another language, would you consider that a CE?


    3. Paolo Trovato9/26/2017 9:40 pm

      Dear Tim (if I may)
      Forgive very probable mistakes, but I am a bit tired and my English is far from perfection.

      I start from the last and more simple question. I wouldn’t call CE a reading which exist in the MS tradition, even if it is found only within the so-called indirect tradition (translations, commentaries etc.).

      I made some examples of CE in Latin, Italian, Spanish texts (and of the arguments which are necessary to prove it) in my handbook (a chapter is entitled “The ineluctability of critical judgment (choice out of variants, conjecture)”).

      I present here a fresh example from Dante’s “Commedia”, which I draw from a recent paper of Riccardo Tesi (unfortunately I have a very scarce expertise of Greek NT’s criticism). In Paradiso XXVI, lines , 133-136, Adam informs Dante on the variation of the name of God in the old language: “Ere I descended to the grieving place / below, the Highest Good, from whom proceeds / the joy
      which swathes me, was on earth called “I”; / “EL” was it called thereafter…”

      All the MSS (and all the scholarly editions so far) bear the reading “I” or “EI” or “un”, but all the scholarly literature of Dante’s time (saint Albert, saint Thomas etc.), which Dante masters, maintain that the old name of God was “Ia”. In my edition I will read without any doubt “Ia”, because I consider “I” an innovation which occurred at the level of the archetype (the copyist of the archetype wasn’t so acquainted as Dante with contemporary philosophers and theologians).

      Of course, in this case the reason is the scarce likeliness that Dante, who chose to address that problem, read only one philosofical source on this topic and his copy of that source was uncorrect. But the need to correct the tradition can depend on the context, on linguistic inconsistency and on a number of different reasons.

    4. I'm not sure how well-known or recent this is, but if I were more inclined to accept CEs in the NT, I'd take a hard look at "God" for "jealousy" in Jm 4:5 (see the database mentioned in a post below).

    5. According to the database, this one's been proposed twice in different forms (τον θεον in one case, as recorded by Wettstein in 1730, and θεον by Könnecke in 1896). If I'm reading it correctly, it changes the sense of the scriptural citation to "The Spirit...desires God."

      I'd like to hear more! What do you think makes this a more compelling conjecture? Does it fit the context of Jas 4 especially well? Or does it conform more to the scripture being cited? (And on that note, what exactly is the scripture being cited?)

    6. not that I'd necessarily endorse his views or anything, but I believe Wettlaufer did a chapter on this one in his brill volume...

    7. Thanks for that one Ryan, I enjoyed it.

    8. As I understand it, the idea is that with the CE the reference would be to Ps 42:1 (LXX). In that case there would be a specific OT passage being cited, and arguably Jm 4:5 itself would become more intelligible. As indicated elsewhere in this thread, I'm not personally in favor of CEs for the NT, but I thought I'd bring this one up in the interest of fairness.

  5. Note that a database of numerous CEs is available to consult: See http://evangelicaltextualcriticism.blogspot.com/2016/09/over-on-jan-krans-announces-public.html.

  6. While it must be conceded that relatively little survives from the first few centuries and that the large majority of MSS once to have existed no longer exist, I would say that the diversity of origins for our witnesses argues strongly against an original reading not being found anywhere in the extant MS tradition. It's not as though we have only, say, a few dozen MSS from only one or two monasteries or printing presses with historical interconnections. Moreover, to glance at Church history, the theological as well as geographical divides that an error would have to leap to stamp out the original reading altogether would have been very great.

  7. Paolo Trovato9/27/2017 6:30 pm

    I would be deeply surprised if I could persuade all of you that one of the more pacifically accepted notions of your discipline must be dismissed at once.

    What I tried to do is simply to explain the reasons why I don't think that the chance that all the readings of the original(s) are preserved in the extant tradition seems rather unlikely.

    Of course my model of transmission is just a model (a rose is a rose is a rose...). You could choice 2 or 2,5 instead of 3 as a multiplier, but the outcome is neither particularly different, nor too far removed from the real situation of extant Greek NT MSS: you will have a few old copies and hundreds and hundreds of low rank ones. And, if we compare the experience of classic and Romance philology with your transmissional data, it is difficult to imagine that low rank copies, even if from differents areas, are so useful in preserving the good readings etc.

    1. Paolo Trovato9/27/2017 6:35 pm

      Sorry: I mean, ...to explain the reasons why I think

    2. We certainly do appreciate your thoughts, and some in the NT text-critical discipline would agree with you as regards a need for CE. I certainly would allow for it in many other cases than the NT, but for my part I think that there is enough non-low grade material to be quite skeptical of the feasibility of a CE for the NT.

  8. Paolo Trovato9/28/2017 9:20 am

    The rich comment of Joey McCollum (9/26/2017) on 2 Pet 3:10 allows me to make a step forward in the comparation between methods of NT TC and Neo-Lachmannian approach.

    JMC “he conjectural emendation of 2 Pet 3:10, however, seems to reflect the view that the difficult reading principle has its limits. The reading εὑρεθήσεται, found in a handful of Greek MSS, is more difficult, but it is so difficult that the ECM / NA28 editors could not consider it original. The conjectured reading, therefore, is actually much easier than the nearest surviving reading.
    All this makes me wonder about the rules for conjectural emendation. Must conjectural emendations by their nature be easier than some existing readings (as with 2 Pet 3:10)”?

    I remember that we are discussing here of the substantive variants (that is, 5 to 10% of any text with a long MS transmission) and I limit my comment to 2 main points.

    1. Your general diffidence against genealogical reconstruction (the refusal of terms as “groups”, “families”, the difficult use of the ambiguous “text-type”) means that you must use a set of rules (and a collection of readings) without a hierarchy. “These rules cannot rigidly be applied in every situation. Intuition and common sense must guide the process” etc. (Wegner, TC of the Bible, ed. 2006, 1.1). To some extent, this is true also for Neo-lachmannians scholars, but (without a temptative classification of the witnesses and a provisional genealogical tree) I imagine that often the textual critic finds himself or herself like a child in a supermarket (shall I buy this or this or this?).

    After 15 year of classification of the extant copies of Dante’s text I have a sort of transmissional map. According to the map, the highest nodes in the transmission are 4 Northern Italian families (60 MSS out of 600) linked in 2 subarchetypes (2 families versus the other 2). When I look at the general genealogical tree, I immediately see that a reading that is in a witness of a subsubsubsubfamily of the Florentine vulgate tradition, crowded with innovations and deeply contaminated (i.e, the other 540 complete MSS) doesn’t have a chance of “preserving” a original reading (it is absent in his ancestors!). So in the best hypothesis the reading is a good CE and (if it is really a good one) I can register in the apparatus that the copyist of MS *** suggest the fascinating reading “platypus” (or “octopus”) instead of “byke” ,which is in ALL the MSS in the top of the tree.

    By the way, most of the members of my 4 “good” families show too often inconsistent behaviours (i.e. contamination or too many errores singulares), so, in order to fix the text, I chose 2 representatives only for family Beta, 3 for each of the other 3 northern families, that is, I use as representatives of the 4 groups the more consistent and conservative witnesses.

    Of course, in 15% / 20” of the cases stemmatics doesn’t work, e.g. 2 of my “good” families read “grapes” and 2 read “bananas” (Pasquali introduced for this situation the term “open recensio”). In those cases, if it is possible, I have to use the afore-mentioned set of rules. But (if I worked hard and consistently during the classification) I am happily out of the supermarket. As a rule, the discussion doesn’t regard any more hundreds and hundres of readings that generated only in the low ranks of the tradition, but only a more little series of readings which could have been in the original. (follows).

    1. The first paragraph of your point 1 relates to an important issue that has been discussed on ETC before. NT textual critics recognize that there is a problem of circularity in the evaluation of internal evidence (pertaining to readings) and external evidence (pertaining to manuscripts): some readings are good because they are preserved by good manuscripts, and some manuscripts are good because they preserve good readings. To use your illustration, we're in the supermarket, and we don't know where to begin.

      In order to have a useful hierarchy, how do we get to a provisional stemma? We can know a handful of things about manuscripts without appealing to the undecided priority of important readings, such as their age (roughly), their provenance (sometimes), and the care with which their scribes copied them (on the basis of obvious errors and orthographic preferences). Similarly, we can know some things about readings without appealing to the number of undecided weight of the manuscripts that bear them, using factors like historical context, authorial style, scribal habits, and probable motivations for intentional scribal changes.

      If I understand your approach (and please correct me if I don't), in order to produce a provisional stemma that we could use to weigh manuscripts, we'd have to do so using textual variants that (1) have resolutions so obvious that we wouldn't need to appeal to the weight of manuscript support; and (2) represent monogenetic changes to the text (i.e., ones that are very unlikely to have arisen more than once independently), to ensure that they induce a meaningful partition of the manuscript data. Scribal errors often have quality (1) but not (2). Variants that divide the majority induce desirable partitions, but they often are polygenetic and less significant (e.g., transposition of word order), besides being difficult to resolve. I imagine that variants with qualities (1) and (2) will be rare. Of course, if a provisional stemma is expected to be more like the "spine" of a fuller stemma, then I suppose the sparsity of its branches shouldn't be a problem.

    2. Paolo Trovato9/28/2017 7:05 pm

      You perfectly understood the essential problems. Indeed variants with qualities (1) and (2) are rare. But if we have got a few of them, also confirmatory readings can help to verify that the classification works. I don’t want to bore too much the readers of this post on this point; so, if you are interested to more in depth indications, please refer to my book Everything You Always etc. etc. :)

  9. Paolo Trovato9/28/2017 9:21 am

    2. Textual criticism is a difficult “sport”. Of all the rules we can use, the Harder Reading Principle is the most difficult. The most difficult reading must not be a absurd or inconsistent reading.

    “Easy and difficult are not absolute terms, and what is difficult, that is, unusual, for us may have been easy for people in other ages”.
    Pasquali, 122

    “Lectio difficilior is […] a criterion that needs to be employed with great care”
    Contini, 111 [= id., Frammenti, II, 998]

    To my knowledge (my “source” is again Timpanaro, 38 with note [= id. (Most ), 68 with note]), the concept of lectio diffi cilior [literally: ‘the more diffi cult variant’] was already correctly theorized in Jean Le Clerc’s Ars critica in 1696 (“Si una ex iis [sc. lectionibus] obscurior sit, ceterae clariores, tum vero credibile est obscuriorem esse veram, alias glossemata”). I will attempt an up-to-date translation of this passage (Timpanaro had already objected to the reduction of all competing readings to glosses, that is, explanations):

    “If (weight in the stemma being equal) in a variation place one reading is more diffi cult and the others easier, it is more likely that the lectio difficilior is the original one, and any attempts to make the passage more easily understandable, whether intentional ( glosses) or not (unconscious banalizations), are secondary readings.”

    This theory needs to be clarified with a couple of examples. The difficulty can be of various kinds, e.g., lexical, syntactical, or conceptual. As late as the fifteenth century, some Italian writers still refer to lovers by the learned compound word “filocapti,” from the Greek phìlos [friend] and the Latin capere [to capture]. The word was frequently used in medieval Latin (Latin examples can be found in Ducange’s dictionary or online s.vv. philocaptus or filocaptus; some Italian examples are listed in GDLI and LIZ, s.v. filocatto). If in the course of the transmission of the work a copyist were to introduce the phrase “preso d’amore” [captured by love] or such, usually this would be assumed to be a typical lexical banalization, and the Greek-Latin compound would be the lectio difficilior […].

    (I am a lazy man, so I must confess that point 2 is exactly the beginning of a paragraph of my handbook: 3.2. From lectio difficilior to diffraction).

  10. Paolo T.,

    The key phrase in all this being, "Let’s suppose."

    Btw, regarding that CE in Dante, if "I" is the extant reading, why not conjecture that was was meant was not "Ia," but "Iehovah," and that a period has been lost from the abbreviation?

  11. Paolo Trovato9/29/2017 7:27 pm

    Dear Peter,
    1.You are right.The key phase in good textual criticism is “Let’s suppose”.

    Indeed, «The Faculty, which God has given Man to supply the want of clear certain Knowledge in Cases where that cannot be had, is Judgement» (Locke, Essay, IV, 14, 3, 653).

    «The grounds of it [scil. Probability] are, in short, these two following: First, The Conformity of any thing with our own Knowledge, Observation, and Experience. Secondly, The Testimony of others, vouching their Observation and Experience» (Locke Essay, IV, 15, 4, 656).

    «Upon these grounds depends the Probability of any Proposition: And as the conformity of our Knowledge, as the certainty of Observations, as the frequency and constancy of Experience, and the number and credibility of Testimonies, do more or less agree, or disagree with it, so is any Proposition in it self, more or less probable» (Locke Essay, IV, 15, 6, 657).

    2.As you know, Dante’s Commedia is a poem and it’s written in hendecasyllables (high quality hendecasyllables, stressed in the canonical syllables).

    The line of the CE is:
    “I (corrige: IA) s’appellava in terra il Sommo Bene”
    (syllable 1 IA / 2 s’ap- / 3 pel-/ 4 la- / 5 va in / 6 ter- / 7 ra il / 8 Som- / 9 mo / 10 Be- / 11 ne).

    Unfortunately there isn’t enough room for a trisyllabic word like “Iehovah”.

    1. Paolo Trovato9/29/2017 7:32 pm

      phase > phrase (But also with "phase" we get a good suggestion)

    2. Paolo Trovato9/29/2017 7:37 pm

      Peter > James (I apologize, but as a textual critic I am rather absent-minded)

    3. Paolo: does Riccardo Tesi's discovery mean that we now recognize two archetypal errors preserved in all the extant witnesses, the other being rocco for crocco? (I was going to ask you privately, but perhaps it will be useful for other people to have your answer available online.)

    4. Paolo Trovato10/01/2017 11:39 am

      There are more than two archetypal errors shared by the extant MSS of Dante's Commedia. Whoever you are, you can write me.


    5. Paolo,
      << You are right.The key phase in good textual criticism is “Let’s suppose” >>

      Let me try to cut through what appears to me to be a sort of sophistry. Here is what I mean: it is easy to say that *if* the text has been transmitted a certain way, then we may expect a certain amount of readings to be lost. *If,* for example, the text of a particular composition were transmitted in four distinct transmission-lines, and each transmission-line retained 10 original readings not transmitted at all by the others, and *if* one transmission-line was destroyed, *then* we could conclude than 10 original readings were lost, and could only be recovered via conjectural emendation.

      But that assumption is an assumption. One must justify the premise, first, before one can scientifically build on it. Otherwise one is eating the breakfast of the man who said, "If we had some ham, we could have some ham and eggs, if we had eggs."

      Or to put it another way: imagining a need for emendation is one thing; demonstrating a need for emendation is another. There just aren't very many passages in the New Testament where a strong case that there is a need for conjectural emendation has been made. Meanwhile, as Bowyer showed long ago (in https://archive.org/details/criticalconject00bowygoog ), allowing textual emendations into the text would adulteration more than it would purify.

    6. Paolo Trovato10/02/2017 12:57 pm

      Dear James
      You last comment puts a clear question and I think that any clear question requires, if it is possible, a clear answer.

      A) I start from the last point. I don’t say nor I think that NT TC or any other field of TC must be based on conjecture. CE is our last weapon: it must be used only in the few cases in which a thorough analysis of the variants of a passages shows that all them, for different reasons, are absurd or unacceptable.

      (I also gave a look to the first samples of the collection of comments included in William Bowyer in 1812!!!! Then TC, in all fields, was still in is infancy and editors didn’t limit themselves to correct scribal errors. Very often they arbitrarily corrected the style of the very authors according to their subjective taste. It is much more instructive to read high quality works of the last 50 years such as Timpanaro, The Genesis of Lachmann’s Method, or Kenney, The Classical Text. Aspects of Editing in the Age of the Printed Book, or Reeve, Manuscripts and Methods, or better all three of them).

      B) I come to the transmission. For the sake of semplicity, I leave on one side the issue of different collections of Gospels, Act etc. and I only consider a single Gospel. We have two possibilities.
      1. That there is one original (whatever it may be, and whoever is the author).
      2. That we have more originals. In this case we should make as many editions as many originals we can trace.

      I don’t address the also theological problem which position (2) implies (I am absolutely incompetent) and I go back to (1), which is probably more interesting for you too. The original of an extant work, which in the case of a Gospel has an almost higly probable datation, can be copied 1 /2 / 3/ … / n times, giving birth to 1 /2 / … lines of transmission. Each copy of 1st rank can be copied 1 / 2/ …/ n times. And so on.

      Mostly for reason of datation no extant copy of the Gosples seems to be of 1st 2nd or 3rd rank. Studies and direct experiences about many MSS traditions, experiments with the computer (Weitzmann 1987) and calculations of a nuclear phisicist on the complete genealogical tree of Renaissance printed books (Guidi-Trovato 2004) agree that, when facing a realistic loss rate of 80% / 90% (again, a figure agreed by many scholars), the most thin lines of transmission of a genealogical tree disappear without leaving a trace. On the contrary huge families (vulgate traditions) survive in many copies. The problem of loss of exemplars is deeply linked to the notion of archetype (I refer here to my book Everything etc., ad indicem).

      Thus, if you find the original of NT you only need to reproduce the original, if you don’t have it, you must use criticism.

      I find unlikely that only NT transmission doesn’t follow, to a certain extent, the general trend, which isn’t an arbitrary construction of my own, but appears rather probable for the afore-mentioned criteria of Locke: “as the conformity of our Knowledge, as the certainty of Observations, as the frequency and constancy of Experience, and the number and credibility of Testimonies, do more or less agree, or disagree with it, so is any Proposition in it self, more or less probable” (Locke, Essay, IV, 15, 6, 657).

    7. Paolo Trovato10/02/2017 12:59 pm

      Of course, this was a rough synthesis of a very complex problem, but I hope I clarified my position.


  12. Paolo,
    PT: "I find unlikely that only NT transmission doesn’t follow, to a certain extent, the general trend."

    Why? That sounds like more like an assumption than a finding.

    1. Paolo Trovato10/02/2017 8:08 pm

      Dear James
      In fourty years of work I found tens and tens of cases in which we have a few old and testually genuine MSS against many recent ones of bad textual quality (and few exceptions as well: “recentiores non deteriores”, as Pasquali puts it), I studied the simulation of Weitzmann and the calculations of Guidi. I cannot spend the rest of my life looking for new confirmations of such an obvious pattern.

      In my opinion it’s you that should find at least a sound example of the kind of transmission that you imagine: that is, if I understand well, a lot of late MSS all of the same textual quality of the oldest ones. And this notwithstanding the fact that at least since the time of Westcott and Hort the Byz seems to offer many double lessons, that is, a typical proof of contamination.

  13. " 4 to 10 major innovations every 10 pages."

    Scripture has an inherent counterweight to "major innovations" when copied within a church and believer environment. Tertullian talks of the discipline of a scribe who innovated.

    Take the Hebrew Bible, at least from 800 AD to printing. Safe to say, 0 innovations, at least in the Pentateuch.

    This type of analysis is so loosey-goosey, trying to apply flying leaps of analysis (Dante, medieval books, scripture) that I find it a bit humorous.

    Now, from a humanistic standpoint, I could agree that there could be missing variants from the original. Hort speculated this for his primitive corruptions, but no real evidence was ever found for his speculations. From a humanistic standpoint.. the Bible is just another ho-hum book.


  14. Paolo Trovato10/15/2017 4:21 am

    Dear Steven,
    my point is very simple. I think that, to a certain extent, textual scholars use the same set of rules in any field of its application and that the fact of addressing slightly different problems can favour the finding of interesting solutions. For this reason I use to spend some of my time looking, in a rather amateurish way, at my neighbours’ gardens, where I always happen to learn something new and useful. I consider this a very healthy practice and omho super-specialisation doesn’t make any good to the disciplines. (Within a discipline most scholars repeat for many years the same basic things and almost nobody checks if they are true. I would say that, on the long run, it works like, in the old times, a chain of marriages between relatives.)

    Just reading some comments here I understood that some problematic places of variations do exist also in Greek NT and that at least a minority of NT scholars was interested in my point of view. If you like, you can also dramatically reduce my figures (e. g., "1 to 3 major innovations every 25 pages"), but the problem I tried to pose remains the same: unless we don’t think that the Holy Spirit decided to help Greek NT scribes, but did not help at all, e.g., the printers of the Vinegar Bible.

    P.S. Just about the Hebrew Bible, I read Ronald Hendel, Steps to a New Edition of the Hebrew Bible (Atlanta, SBL Press, 2016),.which the author was so kind to send me. In this book Professor Hendel seems rather interested to the position of Classical and Romance textual scholars.