Supposed distigmai-obeli in 03 |
I heard Payne give this argument in 2015 and wrote about my reaction to it back then. Here I thought I would add just a little to that.
The crux of the argument is that certain NT paragraphoi are longer than others and occur alongside distigmai (double dots). These “distigme-obelos” marks are then said to mark textual additions which Payne identifies, as before, using the NA apparatus. Payne does concede this time around that seven of his eight “distigme-obelos” symbols might also mark paragraph breaks, but he is quite confident that they are more than that. The upshot in all this for Payne is that 1 Cor 14.34–35 is not Pauline and the apparent inconsistency with 1 Cor 11 is thereby removed.
As before, I find the manuscript argument nearly impossible to believe.
All of Payne’s supposed obeli happen at natural breaks in the text, none of them are actual obeli like we find in the OT portion of Vaticanus (or in other MSS), and all look just like the other paragraphoi to the naked eye. The fact that Payne has to measure them in millimeters to show their distinctiveness only proves the point, in my mind. How could any ancient reader be expected to identify them without measurement?
Add to this the fact that seven of his eight “obeli-distigmai” texts do not actually include the supposed “added text” and the problems are just too much. The ancient reader would have no idea what text these marks were actually marking. The whole point of using an obelos, of course, was that you leave the questionable text in but mark it so that the reader knows what it is. In Payne’s system, the reader is left with no way to know what the symbols mark in 87% (seven of eight) of their cases. Payne’s system, as he envisions it, would never work in practice.
With scholars like Niccum and Miller, then, we should conclude that there is no semantic connection between these distigmai and paragraphoi.
Peter,
ReplyDeleteI think you are right here. I read most of the article he wrote, and read many of the reactions on Twitter and such from those who read the headline. It seems to me that the conclusion fits a modern narrative that garners it more support than the actual argument. If simplicity is to be considered as a criterium for veracity, surely the function of marking a section/paragraph break is to be much more favored. But, as Ockham taught us, simplicity isn’t enough if all factors are not explained.
So my question for you is this: If his explanation is to be rejected, what are the alternative explanations for the dots in these 8 instances?
Here they are probably the same thing they are elsewhere. And on that, I think with Pete Head that they are probably later. As a system of marking variants, they would be pretty worthless to almost any ancient reader.
DeleteFor Head's argument, see here and here.
DeleteCan you provide a simple words explanation for "they are probably the same thing they are elsewhere"? I am having trouble disciphering the rushed recaps in the links you provided below.
DeleteBest post title yet.
ReplyDeleteI appreciate that :)
DeleteI was rather surprised it passed peer review for NTS.
ReplyDeleteThis comment has been removed by the author.
DeleteYou’re in a good company
ReplyDelete...and picked up by the press:
ReplyDeletehttp://nationalpost.com/news/world/bible-was-doctored-to-read-no-women-priests-claims-scholar
The Telegraph too. And Larry Hurtado was impressed by the article.
DeletePeter Gurry writes, “The crux of the argument is that certain NT paragraphoi are longer than others and occur alongside distigmai (double dots).”
ReplyDeleteAs the author of the NTS article, I know that this is not the crux of my argument. See NTS p. 610, “Greater extension into the margin is their primary graphic distinction, but they also average 4.7 mm long compared to the remaining twenty bars’ 3.6 mm average length.” Seven key facts support the conclusion that all eight bars with characteristic features adjacent to a two-dot symbol mark the location of a multi-word textual addition:
1. The Nestle-Aland Novum Testamentum Graece 28th edition identifies a multi-word textual variant somewhere in the line following each characteristic bar. If all eight were simply paragraph marks, this conjunction would have to be mere coincidence. The probability that multi-word variants identified by Nestle-Aland would occur somewhere in each of eight randomly-selected Vaticanus lines is less than one in a trillion (=UK “billion,” namely one million million). Gurry’s assertion in his ETS post, “these are just dots that happen to line up with the paragraphoi” not only gives no explanation for this, the simple mathematical odds of all eight of them “just happening” to coincide with multi-word NA28 textual variants associated with these lines of text in a randon distribution are less than one in a trillion.
2. Even more astounding, Scribe B left a gap at the exact letter where a widely acknowledged, multi-word textual addition begins following every characteristic bar except one that was evidently added by a different hand.
3. None of the other twenty bars adjacent to a two-dot symbol combines as much extension into the margin and total length as any of the eight characteristic bars. All eight extend, on average, almost twice as far into the margin as the other twenty and are, on average, almost one third longer than the other twenty. Gurry’s assertion that these characteristic bars “all look just like the other paragraphoi to the naked eye” is not true, as all three examples showing both distigme-obelos and paragraphos symbols on his post demonstrate. Their differences distinguish them graphically from paragraph bars randomly occurring after dots. Their extension into the margin associates them with their adjacent two-dot symbols, whose purpose was to mark the location of textual variants. The characteristic bars specify what category of variant—multi-word additions—occurs there.
4. Scribe B undoubtedly used horizontal-bars in the Vaticanus Prophets to mark the locations of blocks of added text since one is in the middle of text and since explanations that these bars mark added text display original Vaticanus ink.
5. All eight characteristic bars adjacent to a two-dot symbol resemble the shape and length of each bar marking added text in the Vaticanus Prophets, also penned by scribe B.
6. A horizontal bar was the standard Greek symbol for marking the location of added text.
7. Bars also mark blocks of added text in other manuscripts at John 7:53–8:11 and Mark’s longer ending. Apparently every manuscript with a bar introducing Mark’s longer ending notes that this ending is not “in some of the copies.”
All this supports the conclusion that two-dot+characteristic bar symbols mark the location of multi-word blocks of added text.
to be continued
Payne response Part 2:
ReplyDeleteGurry asserts, “The ancient reader would have no idea what text these marks were actually marking.” This is disproved by the one distigme-obelos with distinctive downward strokes from both dots and the bar, showing that a different scribe added this eighth symbol also at the point of a multi-word block of later added text. So at least that one scribe understood the function of the distigme + long bar extending unusually far into the margin as marking the location where a block of text was added. Nearly 800 distigmai in Vaticanus mark the location of textual differences between Vaticanus and other manuscripts. They apparently provided a guide to where other manuscripts accessible to the scriptorium had a different text than Vaticanus. As I argued in Le manuscrit B de la Bible (downloadable from www.pbpayne.com under Publications: Articles), the mirror-impression distigmai in offset ink demonstrate that Vaticanus was compared to multiple manuscripts. Places where the manuscripts had differing text were systematically marked at that location in Vaticanus, in each case without writing the alternate text.
There is now a general scholarly consensus that the purpose of distigmai was to mark locations where other manuscripts have a different text than Vaticanus. Paul Canart has confirmed (see “Distigmai” in Le manuscrit B de la Bible) that 51 distigmai match the original Vaticanus apricot color ink, some distigmai re-inked in dark-chocolate color ink have apricot color ink protruding from under them, and one has one re-inked dot and the other dot in apricot color ink. One can safely assume, therefore, that scribe B was aware of the pattern of use of distigmai to mark in Vaticanus the locations where other manuscripts have a different text, especially since distigme in one of the eight distigme-obelos symbols matches the original ink color of Vaticanus. It would be inconsistent with this pattern for the obelos to be added only where the Vaticanus text included the addition. The occurrence of a multi-word block of added text at the location of each characteristic bar following a distigme is reasonably explained only if the characteristic bar was used to mark the location of a block of added text.
It is precisely the absence of any of the five distigme-obelos marked blocks of added text in the text of the Gospels in Vaticanus that confirms that the Gospels text in Vaticanus is extraordinarily early. This confirms evidence from the primitive form of the Vaticanus Gospels, having virtually no high stop periods at the ends of sentences–in contrast to its epistles also written by scribe B, which have high stop periods throughout each epistle. Since there are so many similarities between Vaticanus and P75, many scholars have argued that the text of Vaticanus goes back as least to the date P75 was written. Since P75 has high stop periods at the end of sentences throughout, NTS pp. 621-623 argues that the text of the Gospels in Vaticanus is even older than P75, going back to a period before high stops were added at the end of sentences. The fact that seven of the eight texts marked by a distigme-obelos symbol do not actually include the added text is not a “problem” as Gurry imagines. It is merely consistent with the use of distigmai throughout Vaticanus. More important, it provides extraordinary confirmation of the antiquity of the Vaticanus Gospels.
to be continued
Payne response Part 3:
ReplyDeleteGurry’s review of my ETS paper states, “I asked Payne if the fact that the ‘added text’ marked by these symbols wasn’t actually in Vaticanus in his first seven examples means that the symbols would have been completely useless to readers of Vaticanus itself. He admitted that this is true.” I did not state or imply that this is true. My affirmation of instances where the added text marked by the distigme-obelos was not in the text of Vaticanus in no way implied that “the symbols would have been completely useless to readers of Vaticanus.” That insinuation is simply not true. As an aside, one should not, as Gurry does here, use the name “obeli-distigmai” since it inconsistently combines a plural Latin form (obeli) with a plural Greek form (distigmai) where both should be Greek and singular. The correct name in the plural is “distigme-obelos symbols.”
Gurry asserts, “In Payne’s system, the reader is left with no way to know what the symbols mark in 87% (seven of eight) of their cases. Payne’s system, as he envisions it, would never work in practice.” To the contrary, any reader familiar with the use of obeloi to mark the location of added text could deduce that these symbols mark the location of added text, and they would be given a further indication of this if they recognized a distigme marks the location of textual variants, as it also does in LXX G.
Gurry’s same objection that the reader is left with no way to know what the symbols mark could also be made regarding distigmai. But every scholar Gurry cites acknowledges that distigmai mark the location of textual variants. Any user of the scriptorium could discover what variant the distigme and distigme-obelos symbols mark simply by comparing that specific passage with other manuscripts containing that passage in the scriptorium. Even after Vaticanus had left the scriptorium, the distigmai and distigme-obelos symbols would be useful to anyone who realized they identify, respectively, the location of textual variants and the location of blocks of added text. A comparison with other manuscripts anywhere would likely reveal what some of those variants were.
In trying to dismiss the significance of the fact that the gap following every distigme+characteristic bar is at the exact location of a multi-word, widely-acknowledged textual variant, Gurry asserts, “All of Payne’s supposed obeli happen at natural breaks in the text.” As tabulated on NTS p. 619, however, only four of the seven gaps coincide with a paragraph break in NA28, and only three with a paragraph break in the UBS5. Furthermore, as noted on NTS p. 612, a gap occurs in only twelve of the twenty lines following a distigme adjacent to an undisputed paragraphos. So Gurry’s apparent underlying assumption that there is always a gap in the line following a paragraphos is simply not correct. See NTS p. 619, “scribe B may have positioned the first three characteristic bars in the normal paragraphos location because they are all at natural paragraph breaks, then retained the same position for the others to keep them consistent.”
to be continued
Payne response Part 4:
ReplyDeleteGurry writes, “With scholars like Niccum and Miller, then, we should conclude that there is no semantic connection between these distigmai and paragraphoi.” Gurry assumes that these scholars, who to my knowledge have not read my new article, will not reconsider their judgments in light of the new data in my new NTS article. Even without all the new evidence, D. Parker JSNT (2003) 408 rejected C. Niccum’s arguments that it is “likely” the distigmai “originated with Sepulveda … Payne successfully vindicated his case [against Niccum’s critique]. See NTS p. 607.
Gurry writes, “I think with Pete Head that they are probably later.” Distigmai in the fourth- or fifth century LXX G marking the location of Greek textual variants prove that distigmai are not a unified development in the 16th century as Head claimed. See NTS pp. 607-608 with a photograph of one of them. LXX G’s distigmai confirm E. Gravely’s case against “the most recent (and only current) arguments for a late date for all the Vaticanus umlauts” by P. Head. See NTS p. 607.
For a summary of the six groundbreaking discoveries in my new NTS article, see www.pbpayne.com.
I suppose it should be left to the readership to assess whether these 'discoveries' are indeed groundbreaking or not. So far, most MS scholars I've interacted with have seen little ground broken.
DeletePhilip, are you claiming that your eight examples were randomly chosen?
ReplyDeleteNo. The eight instances described and photographed in NTS pp. 610-618 are not randomly chosen. They are treated as a group because they are the ONLY instances of a horizontal bar immediately following a distigme in Vaticanus that extend noticeably farther into the margin (on average, almost twice as far into the margin) as the other twenty bars following a distigme and are also noticeably longer (on average, almost one third longer) than the other twenty bars following a distigme.
DeleteIt is these characteristic graphic features that identify this group of eight. In addition to these characteristic graphic features, each occurs at the location of a widely recognised, multi-word addition to the text. Furthermore, all but one, with downward dipping strokes from both dots and the bar, indicating
a different hand have a gap in the following line of text at the precise location of the widely recognized, multi-word block of added text.
Hi Philip. A quick question. This article https://www.cambridge.org/core/journals/new-testament-studies/article/are-there-distigmeobelos-symbols-in-vaticanus/F2507362D31D370A04A20583EE0E1575/share/837346b6a807c7522a44150fa4400204ff29c9d6 re-examined your measurements and found that the eight you selected over-emphasized the "extend noticeably farther" which was a criteria of your selection of the 8, as you noted above.
DeleteCan you please provide 1) the criteria by which you measured (including if this was a blind measurement) and 2) what number of the non-8 horizontal bar immediately following a distigme mark multi-word textual additions, and which do not?
Philip, the scribe of B left a lot of gaps. It is the absolutely core large level punctuation mechanism that the scribe actually uses. You've got a major selection bias in focusing in on these eight.
ReplyDeleteNone of these eight were identified as in this group because of any relationship to gaps in the Vaticanus text. There is no selection bias because this group includes ALL instances of a horizontal bar immediately following a distigme in Vaticanus that extend noticeably farther into the margin (on average, almost twice as far into the margin) as the other twenty bars following a distigme and are also noticeably longer (on average, almost one third longer) than the other twenty bars following a distigme.
DeletePhilip, the horizontal paragraphoi are not original to the work of scribe B, they are subsequent to the original scribal activity.
ReplyDeleteThe validity of the conclusions drawn from the invariable association of distigme + characteristic bar + widely recognized, multi-word block of added text that begins at the exact point of a gap in the following line of text does not depend on the dating of paragraphoi.
DeletePeter, the copy you sent me of your SBL paper on the Vaticaus Marginalia refers twice to "the original paragrahoi." What causes you now to write "the horizontal paragraphoi are not original to the work of scribe B, they are subsequent to the original scribal activity." Has this been argued in a publication? If so, what are its bibliographic details?
DeletePayne writes, "Gurry’s same objection that the reader is left with no way to know what the symbols mark could also be made regarding distigmai."
ReplyDeleteTo which I can only agree and add that this is why I find Head's explanation for them far more probable. He gives a context in which their form actually makes sense of their purpose.
Under Head’s explanation, just like my explanation, distigmai are not accompanied by the variant reading, so how does Head’s explanation avoid the same criticism you direct at me, that the reader is left with no way to know what the symbols mark?
ReplyDeleteDistigmai in the fourth- or fifth century LXX G marking the location of Greek textual variants prove that distigmai are not a unified development in the 16th century as Head claimed. See NTS pp. 607-608 with a photograph of one of them. LXX G’s distigmai confirm E. Gravely’s case against “the most recent (and only current) arguments for a late date for all the Vaticanus umlauts” by P. Head. See NTS p. 607. For my detailed critique Head's 16th century dating of all distigmai, see https://www.pbpayne.com/wp-content/uploads/2010/03/Critique-of-Vaticanus-Marginalia-15Apr2010.pdf
Payne: “so how does Head’s explanation avoid the same criticism you direct at me, that the reader is left with no way to know what the symbols mark?”
DeleteAssuming I understand Pete’s view correctly, it avoids it precisely by not envisioning a reader as the intended audience. If these are marks made by Sepulveda for use in corresponding with Erasmus, then they are his own private marks that need not be understood by anyone else. On your view, the dots are completely meaningless without insider information and so we are either left to assume that they were useless to most users of 03 or that this knowledge was passed on by means to which we no longer have access.
Your few lines of response to Head in this latest article don’t really refute his view at all. Citing someone else’s opinion is hardly a refutation, of course. And while the dots in LXX G may be significant, there is a leap in the logic and a danger of circularity. We need to know (a) that LXX G really does use the dots to mark variants and (b) that the connection between the dots there and here in 03 is real rather than alleged. I’m not yet convinced on either point. For all I know, the dots in LXX G may be later too or may not mark variants at all.
More seriously, I have problems with the whole method of finding variants in NA and then assuming that the dots must therefore mark variants. If we take large enough collations, there will hardly be a line of Vaticanus that does not have some variation in a modern collation. But that does not tell us anything about Vaticanus; it tells us something about the textual tradition as a whole as we see it from our vantage. We ought to be more careful about reading our knowledge back into Vaticanus and its scribe(s). In short, there is no way to falsify your method of tying the dots to variants since we do not have the exemplar of Vaticanus. Given a large enough apparatus, I could just as easily claim that the dots mark spelling differences only and you would have no way to refute that view. Doubtless such an explanation would not grab headlines the way yours does, but that tells us nothing about whether or not it is true.
Gurry writes, “If these are marks made by Sepulveda for use in corresponding with Erasmus, then they are his own private marks that need not be understood by anyone else. On your view, the dots are completely meaningless without insider information”
DeleteDo you really think that Sepulveda would write in ink “his own private marks” in what is probably the oldest large vellum manuscript in existence and the glory of the great Vatican library, and would do so in nearly 800 places not only in the NT but also sporadically throughout the LXX and flip the pages before the ink was dry leaving mirror impression offset ink on the opposite page in many places? The link I provided exposes many other problems with this view.
On my view, the distigmai were intended to mark the location of textual variants in manuscripts available to the Scriptorium in which they were made. They were not intended for a general readership. My view accounts for the 51 distigmai whose ink matches the apricot color of original Vaticanus ink on the same page, the re-inked distigmai with apricot color ink protruding from under the dark chocolate re-inking, and one distigme with one dark chocolate dot and one apricot color dot. The Sepulveda theory provides no explanation for any of this.
Gurry writes, “Your few lines of response to Head in this latest article don’t really refute his view at all.” Didn’t you notice that my latest NTS article pp. 607-608 n. 13 has the same link to my 35 page rebuttal of Head’s thesis that I provided in my former response? It is at https://www.pbpayne.com/wp-content/uploads/2010/03/Critique-of-Vaticanus-Marginalia-15Apr2010.pdf. In addition, that note cites E. Gravely’s Ph.D. dissertation’s rebuttal of that view, “The Text Critical Sigla in Codex Vaticanus.”
Gurry writes, “We need to know (a) that LXX G really does use the dots to mark variants.” NTS, p. 607 identifies the Greek variant at the exact location of the distigme, includes a photograph, and note 11 identifies another such instance.
Gurry writes, “We need to know ... (b) that the connection between the dots there and here in 03 is real rather than alleged.” The LXX G distigmai (4th to 5th century) disprove Head’s contention that all distigmai are from the 16th century and are not an early symbol. LXX G’s distigmai confirm that D. Parker was correct to reject C. Niccum’s arguments that it is ‘likely’ the distigmai ‘originated with Sepulveda … Payne successfully vindicated his case’. D. Parker, ‘Through a Screen Darkly: Digital Texts and the New Testament’, JSNT (2003) 395–411, at 408 n. 17.
to be continued
Payne response part 2:
ReplyDeleteGurry writes, “I have problems with the whole method of finding variants in NA and then assuming that the dots must therefore mark variants.”
NTS pp. 610-612 explains why NA28’s apparatus is an appropriate basis for assessing whether distigme-obelos symbols are text-critical symbols. NA28 identifies “variants of text-historical relevance.” Multi-word additions have important text-historical relevance. Scribe B marked added text in the LXX prophets with obeloi and scribe B added marginal notes explaining that the obelos signifies that text was added. This proves scribe B regarded blocks of added text as significant. B. Ehrman, as others (e.g. E. Colwell, and B. Aland, see NTS, p. 612, n. 31) argues that “even readings that are not attested in the fragmentary remains of the ante-Nicene age ... are by and large best understood as deriving from documents of the first three centuries ... The vast majority of all textual variants originated during ... the second and third centuries.” This highlights
the value of the NA28 apparatus for identifying early textual variants.
Gurry writes, “there is no way to falsify your method of tying the dots to variants since we do not have the exemplar of Vaticanus.” The exemplar of Vaticanus has nothing to do with the significance of the distigme-obelos symbols in Vaticanus. The pattern could be falsified by showing that there is not a significant correlation between distigme + characteristic bars and widely-acknowledged multi-word blocks of added text. The meaning of symbols is confirmed by establishing a consistent pattern in their usage. I observed a completely consistent pattern in each of the eight distigme + characteristic bars occurring at the locations where multi-word textual variants listed in NA28 occur in Vaticanus. The simple odds of eight randomly-occurring lines in Vaticanus being at the point of a multi-word textual variant is less than one in a trillion (million million). In other words if you were to randomly select eight lines anywhere in the NT of Vaticanus, the odds of your getting all eight to coincide with a line marking the location of a multi-word textual variant of any sort listed in the NA28 is less than one in a trillion.
Gurry writes, “Given a large enough apparatus, I could just as easily claim that the dots mark spelling differences only and you would have no way to refute that view.” I could easily refute that view by noting the frequency of spelling differences in that huge apparatus, determining how many spelling variations occur on average on each line in Vaticanus, and comparing those results to the number lines marked with the distigme-obelos symbols that coincide with a different spelling in that apparatus. If, on average there are two spelling variations on every line of Vaticanus, then getting eight hits in a row wouldn’t even give minimal support to that proposed meaning of the symbol. Indeed, it could be argued that if the distigme + characteristic bar signified a different spelling, surely this symbol would have occurred far more than just eight times.
If, however, an apparatus that fairly represents spelling variations that were probably available to scribe B of Vaticanus actually showed on average only one spelling variation every 31.8 lines of Vaticanus text, and every one of eight symbols all coincided with one of those rare spelling variations, I would conclude that those eight symbols are indeed marks of spelling variations. But since the NA28 does not show any spelling variations on any of those eight lines, I would not be justified in claiming that these symbols signify spelling variations.
Fascinating discussion.
ReplyDeletePhilip, in section 3 of your paper you mention that 8 bars differ in two respects from the Vaticanus paragraphoi: they extend more into the margin and they are longer. Am I right in thinking that they are longer only because they extend more into the margin, and that we therefore have only one phenomenon rather than two? In other words, don't the 8 bars and the undisputed paragraphoi extend into the text by the same amount?
Peters, what is your explanation for the correlation between distance that bars extend into the margin and location of multi-word variants? Philip gives a number of more than 10,000 to 1 and his mathematics is sound. Do you question the data that he uses? Do you propose that Sepulveda was aware of multi-word variants and marked their locations? It would help non-specialists like myself if you could address the statistical argument, which is quite separate from the question of how visually obvious the extra extension into the margin is.
Hi Richard, I hope you are well.
DeleteI don't think the length of the bars or their distance from anything is of any significance whatsoever. These are written by hand! I think the bars and the gaps are two systems marking the same thing - a pause in the thought of the text. In the case of the gaps these are the product of the scribe (in an unknown relation to his textual exemplar and interpretive tradition). In the case of the bars these are a later addition to the text not the part of the initial production (evidence for this is that occasionally they are indepedent of the gaps).
The double dots mark textual variation in the 16th century. Textual variation is not random so occasionally long additions occur at the end of sentences/paragraphs. No big deal, no intrinsic relation between the double dots and the paragraphoi and the spaces - three systems of textual analysis occasionally interacting with each other by co-incidence.
Thank you for your excellent question, Richard. It caused me to examine other graphic aspects about these contrasting bars. The examination revealed an even greater contrast than I had realized.
DeleteBut first, to answer your question, in order to express the full range of differences between the undisputed paragraphoi immediately following distigmai and the characteristic bars immediately following distigmai, both sets of figures are of value since both highlight a different noticeable difference between the 20 paragraphoi and the 8 characteristic bars.
Your question caused me to calculate not just the differing extension into the margin, but also the differing extension into the body text of those paragraphoi and the characteristic bars. This led to the discovery of two additional graphic distinctions between those paragraphoi and characteristic bars. The third is not dramatic, but the fourth provides an even greater contrast than the average extension of characteristic bars almost twice as far into the margin as the average extension into the margin of the undisputed paragraphoi following distigmai.
Here are the results: those 20 paragraphoi extend into the margin on average 1.6 mm but into the body text on average 2.0 mm. In other words, on average they extend farther into the body text than they do into the margin.
In contrast, the characteristic bars extend into the margin on average 3.0 mm but into the body text on average 1.7 mm. So not only do they extend into the margin on average almost twice as far as the undisputed paragraphoi, the characteristic bars extend almost twice as far into the margin as they do into the body text.
The third difference is the length of extension of these two categories of bars into the body text. These 20 paragraphoi extend into the body text on average 2.0 mm, but the 8 characteristic bars extend into the body text on average 1.7 mm. If it were just individual bars that were being compared, that would not draw attention. This is, however, a surprising difference in the average length into the body text of a relatively large number of instances. It is not, however, nearly as striking as the fourth difference.
The fourth difference is in the ratio of the average length these two categories of bars extend into the margin to the average length they extend into the body text. This can be calculated in two different ways, either by focusing on how much of the bar’s length extends into the margin or by focusing on how much of the bar’s length extends into the body text. Either focus results in virtually the same radically contrasting ratio.
Characteristic bars have a ratio showing the bar is predominantly in the margin that is over twice as large as the corresponding ratio for undisputed paragraphoi. This is a dramatic difference given that it is the average of two relatively large groups, 20 undisputed paragraphoi and 8 characteristic bars.
to be continued
Payne reply to Richard, Part 2:
DeleteThe ratio of the average length these undisputed paragraphoi extend into the margin to the average length these paragrahoi extend into the body text is 1.6 mm into the margin to 2.0 mm into the body text. This gives a ratio of 0.8. Since this ratio is less than 1.0 it shows that less of the paragraphoi’s length is in the margin than is in the body text.
The ratio of the average length the characteristic bars extend into the margin to the average length the characteristic bars extend into the body text is 3.0 mm into the margin to 1.7 mm into the body text. This gives a ratio of 1.765 showing almost twice as much of the characteristic bars’ length is in the margin than is in the body text.
A comparison of these two sets of differing ratios is 1.765/0.8 = 2.2. This means that characteristic bars have a ratio of length extending into the margin in contrast to length extending into the body text 2.2 times greater than that of undisputed paragraphoi.
Virtually the same ratio (2.2) results from focusing on the proportion of these bars that extend into the body text rather than the proportion that extends into the margin. For these paragraphoi the ratio is 1.25 (2.0 mm into the body text to 1.6 mm into the margin). It is greater than 1.0 since more of the paragraphoi’s length is in the body text than is in the margin. For characteristic bars the ratio is 0.567 (1.7 mm into the body text to 3.0 mm into the margin). This ratio is far less than 1.0 since most of the length of characteristic bars is not in the body text.1.25 divided by 0.567 = 2.2. This means that undisputed paragraphoi have a ratio of length extending into the body text in contrast to length extending into the margin 2.2 times greater than that of the characteristic bars.
I had never thought of it before you asked your question, but I believe it is this predominant proportion of the characteristic bars being in the margin vs. the predominant portion of the undisputed paragraphoi being in the body text that more than any other factor makes them so visually distinct. Of course, their obviously greater extension into the margin on average, their overall greater length on average, and their smaller extension into the body text on average all contribute to their visual distinction.
Thank you, Richard, for your excellent question that sparked this new discovery. I’m just sorry I had not thought of this before the article was published. But at least readers of this reply will benefit.
In my view this analysis is of the same type as Bible Code analysis of Hebrew texts and the chap who finds Aramaic in Codex Washington. It is inappropriate analysis of the actual simple data - hand written short strokes of the pen.
DeletePhilip, could you provide a table of the data of the 28 cases, showing biblical reference, Vaticanus page, extension into the margin, extension into the body, and the number of words in the textual variant? This would help us with statistical analyses.
DeleteThe number of words in each of the blocks of added text marked by the distigme-obelos symbols are shown with their full text in NTS pp. 612-615. Philip B. Payne, “FULDENSIS, SIGLA FOR VARIANTS IN VATICANUS, AND 1 COR 14.34–5” New Testament Studies 41 (1995), pp. 240–262 is available for free download from https://www.pbpayne.com/wp-admin/Payne1995NTSFuldensis.pdf
DeleteIn contains on p. 253 a table of 27 of the 28 occurrences of a distigme followed by a bar. It includes for each: the biblical reference, Vaticanus page, the nature of the textual variant listed in the NA associated with that line, and a list of the following 20 lines of text in Vaticanus noting whether or not there is a textual variant in NA on each line. It tabulates the total number of occurrences of a NA textual variant on the distigme-obelos lines and on each of the first line following the distigme-obelos symbol (10), the second line after it (11), the third line (9), etc. until the 20th line (7).
I omitted one instance from that 1995 NTS article, Luke 14:24-25 on 1332 C because its distigme was red in my color facsimile, Novum Testamentum E Codice Vaticano Graeco 1209 (Codex B) 1968, so I mistakenly thought it was a later rubrication. Paul Canart determined that this distigme matches the apricot color of the original in of Vaticanus. It was added in the list on pp. 204-209 of the 51 original ink distigmai in Philip B. Payne and Paul Canart “Distigmai Matching the Original Ink of Codex Vaticanus: Do They Mark the Location of Textual Variants?” pages 199-226 in Patrick Andrist, ed., Le manuscrit B de la Bible (Vaticanus graecus 1209): Introduction au fac-similé, Actes du Colloque de Genève (11 juin 2001), Contributions supplémentaires. Lausanne, Switzerland: Éditions du Zèbre, 2009. You may download it free from https://www.pbpayne.com/wp-content/uploads/2010/07/HTB07_199_226.pdf
I have all the measurements hand written on a print out of that 1995 table with Luke 14:24/25 (1332 C) added by hand. I have checked the measurements and confirmed the accuracy of their totals and averages about ten times prior to publication, but I do not have them in an electronic form.
Le manuscrit B de la Bible can be purchased at the discounted price of $49.95 from the secure on-line order form at https://www.linguistsoftware.com/orders/orders.htm.
I am sad to report that Paul Canart passed away September 14, 2017, the very day Cambridge University Press published my article in NTS with so many quotations from him from our collaborations. He wrote appreciatively of this article and was eager that it be published, so I like to think that his voice continues to speak through it.
Thanks, Pete.
ReplyDeleteWhat you say makes good sense, but does not answer my question. Are you conceding that you do indeed need to appeal to a 1 in 10,000 coincidence?
Looking at Philip's photographs of his 8, it seems that the letter Tau is over-represented at the beginning of the lines adjacent to the bars. There are 5 such Taus, where statistically we might expect about 2. We can suspect that the horizontal part of the Tau, which extends into the margin, has encouraged the scribe to make the bars further to the left than he/she would otherwise do. Perhaps someone would like to look into this, unless it has been done already. If we do indeed need to correct for the presence of Tau, then Philip's statistics would change a little. It would be useful if someone could plot the 28 cases on a scatter plot with the x-axis representing the number of millimetres that the bar extends into the margin, and the y-axis representing the number of words in the textual variant. If the cases with Tau are then colour-coded we would have all the information. This approach avoids the need to make arbitrary decisions such as the choice of the number 8 and the 3-word cut-off.
GIGO
DeleteI'd need to see more of the raw data. However, the alleged 1/10,000 probability refers to the two dots, not the supposedly longer lines.
ReplyDeleteI don't find averages in the following data very helpful, especially as the variations in length are so small. Is a difference in one millimetre really so significant?
"The characteristic bar adjacent to a distigme at the interface of 1 Cor 14.33 and 34 extends 3 mm into the margin. In contrast, the seventy-five other bars in 1 Corinthians extend, on average, 2 mm into the margin, and only one of these seventy-five extends 3 mm into the margin (1475 B29). Greater extension into the margin is their primary graphic distinction, but they also average 4.7 mm long compared to the remaining twenty bars’ 3.6 mm average length." (Payne, p. 610).
Peter Williams writes, “the alleged 1/10,000 probability refers to the two dots, not the supposedly longer lines.”
DeleteThis is not correct. Both the simple mathematical probability (less than one in a trillion US = billion UK) and the chi-square probability data in my article, refer to the eight combinations of distigmai with a characteristic bar. Nor is it the “alleged 1/10,000 probability” but the calculated 1/10,000 chi-square probability with the calculations provided and the basis for each figure in the calculation clearly explained. Nor are they “supposedly longer lines” as you can confirm for yourself. Just look at the three samples Gurry posted contrasting long characteristic bars and shorter following paragraphoi.
NTS p. 620 states, “What is the statistical probability that in a random distribution all eight characteristic bars following distigmai would be at the location of a textual variant at least three words long in manuscripts cited in NA28’s apparatus? Using Matthew as a conservative baseline, the probability that all eight Vaticanus lines would coincide with
the location of a multi-word variant listed in NA28 is1 in 31.88 = 1,045,723,722,517. See note 60 below.
Since distigmai mark the location of textual variants, however, lines following distigmai in Vaticanus are more likely to coincide with textual variants, including multi-word variants, than random lines are. So, this author compared the occurrence
of multi-word variants at these eight distigmai adjacent to characteristic bars (eight of eight) to the twenty distigmai adjacent to undisputed paragraphoi (two of twenty).61 The standard probability test shows that the likelihood of such a stark difference occurring at random is far less than one in 10,000. See note 62 below.
This is over 500 times greater than the threshold needed to reject the null hypothesis.
In this case, the null hypothesis is that characteristic bars adjacent to distigmai do not correlate with multi-word textual variants. This test result justifies distinguishing the eight characteristic bars from paragraphoi.
60 By the author’s count NA28’s apparatus contains only 168 multi-word variants in Matthew. Compared to the 5,343 Vaticanus lines in Matthew, this is fewer than one in 31.8 Vaticanus lines. Matthew is probably at the high end of how frequently multi-word variants occur because NA28, 792–9, lists more papyri of Matthew (twenty-four) than of any other NT book except John (thirty). Furthermore, variant readings due to harmonisation, which are often multi-word, are more frequent in the synoptic Gospels than any other part of the NT. Accordingly, five of the eight multi-word variants marked by distigme-obelos symbols are in the synoptic Gospels, two are from Matthew, and three are inter-synoptic harmonisations.
61 The two are Mark 14.70 (1301 B) and Acts 14.18 (1403 B). Both bars are short, only about 3 mm
long, and neither extends much into the margin.
62 For the details of this chi-square test, see www.pbpayne.com/wp-content/uploads///Vaticanus-distigme-obelos-chi-square.pdf.
to be continued
Payne reply, part 2:
ReplyDeletePeter Williams writes, “I don’t find averages in the following data very helpful, especially as the variations in length are so small. Is a difference in one millimetre really so significant?”
In order to establish differences between two groups of data, in this case comparisons between the 20 undisputed paragraphoi immediately following a distigme and the 8 characteristic bars immediately following a distigme, it is necessary to show not just individual differences but differences that characterize each group. Showing the contrast in average extension into the margin of all representatives of each category effectively highlights this graphic difference. It shows that, on average, the characteristic bars extend almost twice as far into the margin as the undisputed paragraphoi. My two-part reply to Richard Fellows above details two new graphic distinctives of the characteristic bars I had not previously calculated. They make the case for the differences between the characteristic bars and the paragraphoi following distigmai even stronger. The portion of the bars in the margin contrasted to the portion of the bars in the body text is over 2.2 times greater on average for the 8 characteristic bars than the 20 undisputed paragraphoi immediately following distigmai.
In typography one millimetre is huge. In this case the characteristic bars extend into the margin on average 1.4 millimetres farther on average than the undisputed paragraphoi. That is the difference in the width between an 8 point and an 18 point Times New Roman lower case x. Print them out next to each other and see if you do not agree that the difference is noticeable.
I apologize that part of the link to the chi-square data did not display properly. It is www.pbpayne.com/wp-content/uploads/2017/01/Vaticanus-distigme-obelos-chi-square.pdf.
ReplyDeleteIn order to replicate the calculation, we need all the measurements. Also, I think the lack of prespecification is problematic.The probability of the particular 10 books on my desk being the 10 books on my desk is vanishingly small. However, you would presumably have considered the difference in horizontal lines significant whether they had extended more into the text but not the margin, or more into the margin and not the text, or if they'd been slightly oblique, or had any other distinguishing feature.
ReplyDeleteI looked at online photographs of all 28 occurrences of the combination of dots and bar, identified by Philip. I measured bar lengths and extensions into the margin. To avoid bias I did this before I know which are associated (by Philip) with multi-word variants. Others would be able to take more accurate measurements than me, but here are some observations.
ReplyDeleteFirstly, the bars did not fall neatly into two categories in terms of length or intrusion into the margin. In fact, I got the same intrusion into the margin for Luke 1:2f (1305A), Acts 2:47f (1385B), Mark3:5f (1280C), Col 2:15f (1504B), and Rom 16:5 (1460B), but Philip puts only the first two of these cases in his category of the 8 that have the most intrusion.
More importantly, perhaps, Matt 18:10, 12 (1259A) should not belong with the 8, according to my measurements. As shown in the photograph in Philip's recent paper, there is a dot just to the left of the bar. I did not consider that dot to be part of the bar. What do people think about whether the dot should be used as the left-hand extremity of the bar? Did you include the dot, Philip?
Also, Matt 24:6f (1268A), which is not identified by Philip as having a multi-word textual variant, had the second longest intrusion into the margin, according to my measurements, and would therefore be included in the 8.
Also, one could argue that the statistical test should not include 1 Cor 14:33 (1474A) itself, since this is the text in question.
All in all, the measurements did not confirm Philips "1 in 10,000" result. While there is clearly a positive correlation between intrusion into the margin and the presence of a multi-word variant, it did not appear to be particularly statistically significant. Philip, could you provide us with your measurements?
Thanks, Richard. Can you share all your raw data? I'd like to see Philip's too. As I understand it, that dismantles the centrepiece of the article.
ReplyDeleteHere are my measurements in millimetres. I hope the table formats OK for you. The cases are sorted in ascending order of intrusion into the margin. The columns are as follows:
ReplyDeleteFirst: Vatican's page
Second: Text
Third: Intrusion into margin
Fourth: Intrusion into body
Fifth: Length
Sixth: Whether the text here in NA28 has a multi-word textual variant at least 3 words long, according to Philip. It would be interesting to see how the pattern changes if we change this criterion.
1403B Acts 14:18 1.3 2.3 3.6 Yes
1428C Jas 4:4 1.3 2.2 3.4
1361A John 7:39f 1.4 2.7 4.1
1365A John 9:41f 1.5 1.8 3.3
1403A Acts 14:13f 1.7 1.8 3.4
1345B Luke 22:58 1.8 1.7 3.4
1500C Phil 2:24 1.9 2.2 4.1
1301B Mark 14:70F 2.0 1.5 3.6 Yes
1470A 1 Cor 10:24f 2.0 2.0 4.1
1442C 2 Joh 8f 2.0 1.5 3.6
1342C Luke 21:19 2.2 1.4 3.6
1262C Matt 21:3 2.3 2.2 4.5
1505B Col 3:18f 2.4 1.7 4.1
1505B Col 3:20 2.4 1.1 3.6
1259A Matt 18:10, 12 2.5 1.4 3.9 Yes
1237B Matt 3:9f 2.7 1.0 3.7
1280C Mark 3:5f 2.8 1.0 3.8
1305A Luke 1:28f 2.8 1.3 4.1 Yes
1385B Acts 2:47f 2.8 1.7 4.5 Yes
1401B Acts 13:16f 2.8 1.4 4.2
1460B Rom 16:5 2.8 1.5 4.3
1504B Col 2:15f 2.8 1.0 3.8
1474A 1 Cor 14:33 2.9 1.5 4.5 ?
1253B Matt 13:50f 3.1 1.7 4.7 Yes
1390A Acts 6:10 3.1 1.7 4.7 Yes
1332C Luke 14:24-5 3.2 1.9 5.1 Yes
1268A Matt 24.6f 3.3 0.4 3.7
1284C Mark 5:40 3.4 0.8 4.2 Yes
As you can see "Yes" in the final column skews towards the bottom of the table but not surprisingly so. Phillip's table would have a "Yes" in the bottom 8 rows. We can also compare average lengths:
Philip's measurements Richard's measurements
Philip's 8 4.7 4.45
Philip's 20 3.6 3.78
Thank you for this work, Richard.
ReplyDeleteAs some have requested, following is a chart of the measurements I used in writing my article in NTS for the 8 characteristic bars and the 20 undisputed paragraphoi following distigmai.
ReplyDeleteCharacteristic Bars’ extension into the margin, total length, # words in added text
NB: NTS p. 610 n. 27 gives the standard definition of the margin, including which letters conventionally straddle the margin. This is the definition used for the measurements below.
Vaticanus p.+col., ref., mm extension into the margin, mm total length, # of added words
1253B, Matt 13.50f, 2.9 4.8 4
1259A, Matt 18.10,12, 3.2 5.0 10
1284C, Mark 5.40, 3.1 4.4 3
1305A, Luke 1.28f, 2.7 4.4 4
1332C, Luke 14.24/25, 3.2 5.0 7
1385B, Acts 2.47f, 2.9 4.2 8
1390A, Acts 6.10, 3.0 5.0 23
1474A, 1 Cor 14.33end, 3.0 5.0 35
8 characteristic bars total 24.0 mm 37.8 mm
24.0 ÷ 8 = 3.0 mm average extension into margin, exactly as stated on NTS p. 610.
37.8 mm ÷ 8 = 4.725 mm average length. NTS p. 610 rounded this down to 4.7 in order not to give an unrealistic impression of this much precision in measurement. even though 4.725 mm would have shown an even sharper contrast between characteristic bars and paragraphoi.
NB: It is common for there to be a gap following the beginning of a bar and its extension to the right, not just in Vaticanus, but also in LXX G, and probably virtually all manuscripts on vellum of this period. Ink tends to dissipate more in between the two ends of horizontal bars. For this reason and because it is aligned with the rest of the bar, I did include the beginning dot of the characteristic bar at 1259A, Matt 18.10, 12.
Undisputed Paragraphoi extension into margin, total length, # words in added text
Vaticanus p.+col., ref., mm extension into the margin, mm total length, # of added words
1237B, Matt 3.9f, 1.7 3.5 1
1262C, Matt 21.3, 1.9 4.2 a different form of the same word
1268A, Matt 24.6f, 2.6 3.6 0
1280C, Mark 3.5f, 2.5 3.6 0
1301B, Mark 14.70f, 1.5 3.4 5
1342C, Luke 21.19, 1.6 3.5 a different form of the same word
1345B, Luke 22.58, 1.3 3.0 a different form of the same word
1361A, John 7.39f, 0.5 3.9 2
1365A, John 9.41f, 0.7 3.0 1
1401B, Acts 13.16f, 2.0 4.0 0
1403A, Acts 14.13f, 1.1 3.1 different forms of the same 2 words
1403B, Acts 14.18, 0.8 3.0 6
1460B, Rom 16.5, 2.0 4.0 2 word substitutions
1470A, 1 Cor 10.24f, 1.2 4.0 1
1500C, Phil 2.24, 1.4 4.0 2
1504B, Col 2.15f, 2.4 3.8 0
1505B, Col 3.18f, 2.0 4.0 0
1505B, Col 3.20, 1.8 3.5 substitute article
1428C, Jas 4.4, 0.9 3.0 0
1442C, 2 John 8f, 1.8 3.4 a different form of the same word
20 undisputed paragraphoi 31.7 sum 71.5 sum
31.7 ÷ 20 = 1.585 mm average extension into margin. NTS p. 610 rounds this up to 16.0 in order not to give an unrealistic impression of this much precision in measurement, even though 1.585 mm would have shown an even sharper contrast between characteristic bars and paragraphoi.
71.5 ÷ 20 = 3.575 mm average length. NTS p. 610 rounded this up to 3.6 mm in order not to give an unrealistic impression of this much precision in measurement, even though 3.575 mm would have shown an even sharper contrast between characteristic bars and paragraphoi.
I found that none of the 20 undisputed paragraphoi extends into the margin as far as any of the 8 characteristic bars. This is the primary graphic characteristic of the bars.
I found only one of the 20 undisputed paragraphoi that is as long as the shortest of the characteristic bars.
Even if one were to judge some of the measurements differently, there is clearly not a single instance where any undisputed paragraphoi combines both as great extension into the margin as any of the characteristic bars and also as great a bar length as any of the characteristic bars.
to be continued
Philip Payne’s measurements of the extension into the margin and length of bars following distigmai comments, part 2:
ReplyDeleteEach of the eight distigme + characteristic bar combinations coincides with the location of a multiword textual variant of three words or more in NA28 apparatus.
In contrast, only two of the twenty undisputed paragraphoi adjacent to a distigme coincide with the location of a multiword textual variant in NA28 apparatus. The two are Mark 14.70 (1301 B) and Acts 14.18 (1403 B). In both instances, the bar is short, only
about 3 mm long, and neither extends much into the margin, only about 1mm. So neither displays the graphic characteristics of the eight distigme-obelos symbols.
Since the distigme by itself indicates the location of a textual variant, it is natural to expect a higher frequency of textual variants, including multi-word textual variants, in lines adjacent to distigmai that in random lines in Vaticanus. This explains why the chi-square test that compares the coincidence of multi-word NA 28 textual variants following the 20 undisputed paragraphoi to those following these 8 characteristic bars shows the probability of this strong a contrast happening is far less than one in 10,000, but the simple mathematical odds of eight lines all coinciding with a multi-word textual variant in the NA28 apparatus is less than one in a trillion (UK = billion).
The random occurrence of the same ten books on Peter Williams’ desk is not comparable because is does not describe a stable subject with the identical controlled variables applying equally to the two sets being compared.
Phil Payne’s critique of Richard Fellows’ measurements:
DeleteRichard writes, “As you can see “Yes” in the final column skews towards the bottom of the table but not surprisingly so. Phillip’s [sic] table would have a “Yes” in the bottom 8 rows.”
The only reason all eight at the bottom do not have a “Yes” is that Richard has inserted eight of the undisputed paragraphoi into the bottom of the chart containing the eight characteristic bars I identify. If the two categories are kept together, the difference is dramatic.
As my earlier reply with all my measurements shows, each of the eight distigme + characteristic bar combinations coincides with the location of a multiword textual variant of three words or more in NA28 apparatus.
In contrast, only two of the twenty undisputed paragraphoi adjacent to a distigme coincide with the location of a 3 word or more textual variant in NA28 apparatus. The two are Mark 14.70 (1301 B) and Acts 14.18 (1403 B). In both instances, the bar is short, only about 3 mm long (only 3.4 mm and 3.0 mm by Richard’s measurements), and neither extends much into the margin, only about 1 mm (2.0 mm and 1.3 mm by Richard’s measurements, still short). So neither is even close to displaying the graphic characteristics of the eight distigme-obelos symbols.
It is crucial when measuring distance into the margin that you measure from the exact margin. The Vaticanus NT (and probably the LXX, though I haven’t examined it as closely; much of it has two columns) was scored for the margins and each line. That is why its text is so even and why its margins are so even. The “margin” is at the far left edge of letters on that margin, excluding the four letters identified in my NTS article, p. 610. n. 27, and the chi, which straddles the margin.
I think accurate measurements can only be done with confidence using the IPZS 1999 color facsimile since it requires moving a perfectly straight edge (I use a blank sheet of paper) gradually left until it covers the ink of letters up to the left margin of that column of text. The 1999 facsimile is a much more precise copy of the original leaves than the 1968 Vatican color facsimile of just the NT of Vaticanus, where double images are common. None of the 51 original ink distigmai in my 1968 edition displays their true apricot color, but all 51 in my copy of the IPZS color facsimile display the original apricot color I remember so vividly from my examination of Vaticanus with Paul Canart at the Vatican Library.
Richard’s images were based on the on-line digital images. The on-line digital images are not nearly as precise as the analogue IPZS photographs in the 1999 color printing of the entire text of Vaticanus. The digital images necessitate pixel shedding or addition at every point where the original analogue image does not exactly fit the digital parameters. IPZS is world famous for its precise analogue photography. They published facsimiles of the Leonardo Da Vinci folios as well.
to be continued
Phil Payne’s critique of Richard Fellows’ measurements, part 2:
ReplyDeleteThere may be small differences in measurement depending on whether the measurement is made with all the ink of letters (besides the five listed in n. 27) covered by paper with a straight edge or if only the continuous strokes along the margin are covered by the blank paper.
There also may be small differences in measurement depending on whether one judges to be extending into the margin the heavier ink at the top and bottom of vertical strokes that protrudes farther left than the trunks of all the vertical strokes in the column.
My NTS article p. 610 states, "Only these eight combine noticeably farther extension into the margin and noticeably greater length." It argues throughout that it is this combination of graphic features that sets the characteristic bars apart.
Your own measurement (3.6 mm) agrees that the paragraphos at Matt 24:6 is not as long as your measurement (4.5 mm) of the characteristic bar at Acts 2:47, or for that matter, any of the eight characteristic bars. It is not even as long as your measurement of the length of the bar at Matt 18: 10, 12 to be 3.9 mm (instead of 5.0 with the initial dot), where you did not include the initial dot even though it is aligned with the (rest of) the bar and even though there are many gaps in bars throughout Vaticanus due to faded ink.
Even if one were to judge some of the measurements differently, there is not a single clear instance where any undisputed paragraphoi combines noticeably farther extension into the margin and noticeably greater length.
Even given your measurements, there are only two undisputed paragraphoi that come anywhere close to combining both of these features, and neither of them combines noticeably farther extension into the margin and noticeably greater length. Furthermore, I strongly dispute your measurements for both of these, especially your measurements for their extension into the margin.
You list for the paragraphos at Acts 13.16f an extension into the margin of 2.8 mm. I remeasured it in the IPZS volume as 2.0 mm, and even adjusting the margin so parts of the characters are in the margin only gave a measurement of 2.4 mm. I again found its total length to be 4.0 mm as opposed to your 4.2 mm measurement.
You list for Rom 16:5 the extension into the margin as 2.8 mm. I remeasured it in the IPZS volume as 2.0 mm, and even adjusting the margin so parts of the characters are in the margin only gave a measurement of 2.2 mm. I again found its total length to be 4.0 mm as opposed to your 4.3 mm measurement.
With these adjustments, neither of these two (Acts 13.16f, Rom 16:5), and according to your own measurements, none of the other 18 undisputed paragraphoi, combine noticeably farther extension into the margin and noticeably greater length than any of the characteristic bars, even with your exclusion of the dot from the bar in Matt 18.10, 12 in both your measurements for that bar's extension into the margin and its total length.
Thus, the result of the standard probability test, the chi-square test, rests on secure data. The chi-square result is over 500 times greater than the threshold needed to exclude the null hypothesis. The null hypothesis is that characteristic bars adjacent to distigmai do not correlate with multi-word textual variants at least three words long listed in the NA28 apparatus. This gives extraordinarily strong support to my thesis that the characteristic bars are obeloi that identify the nature of the textual variant marked by the adjacent distigme as a multi-word textual variant. Since scribe B identifies the obelos as marking added text and since a widely recognized, multi-word textual variant occurs on each of these lines, and since all seven by scribe B have a gap in the text (which only the original scribe could leave) at the exact location of the multi-word addition, these eight should be recognized as distigme-obelos symbols marking the location of added text.
Typically if an argument needs this level of hyper-intricate justification (with few people being convinced) upon receiving straightforward feedback, there are some deep problems with it.
ReplyDeletePhilip, none of the points that you raise explains the large discrepancies in our measurements. To find out which measurements are right, let us look at the bars at Matt 24:6 (1268A) and Acts 2:47 (1385B). According to my measurements the extension into the margin of the bar at Matt 24:6 is great than that of the bar at Acts 2:47 (by about 18%). According to your measurements it is the Matt 24:6 bar that extends more (by about 11%). So which bar extends more?
ReplyDeleteIt is a simple matter to find the texts here: https://digi.vatlib.it/view/MSS_Vat.gr.1209
and place them next to each other, for example by taking screen shots and pasting them into the same document. I did this and, Philip, I will email the resulting image to you. Could someone else please also do this comparison and tell us what you find.
Richard judges, as I do, that the characteristic bar at Acts 2:47 is a single bar in spite of the fading of the ink near the right end of the top bar. I believe Richard is correct to include that entire bar in his measurement even though what appear to be two adjacent dots are slightly separated form the rest of the bar and are not as well aligned with the bar as the dot at the left end of the characteristic bar at Matt 18.10, 12.
ReplyDeleteRichard did not include the initial dot as part of the bar at Mat 18:10, 12 even though it is aligned with the rest of the bar and even though there are many gaps in bars throughout Vaticanus due to faded ink, so he lists its length as 3.9 mm instead of 5.0. The bar at Matt18: 10, 12 probably should be understood as a single bar for these reasons and since it makes sense as part of the bar, but makes no sense interpreted as unrelated to the bar.
Richard’s own measurement (3.6 mm) and his email to me agrees that the paragraphos at Matt 24:6 is not as long as his measurement (4.5 mm) of the characteristic bar at Acts 2:47. It is not even as long as his measurement of the length of the characteristic bar at Matt 18:10, 12.
Richard lists for the paragraphos at Acts 13:16 an extension into the margin of 2.8 mm. I remeasured it in the IPZS volume as 2.0 mm, and even adjusting the margin so parts of the characters are in the margin only gave a measurement of 2.4 mm. I again found its total length to be 4.0 mm as opposed to your 4.2 mm measurement.
Richard lists for Rom 16:5 the extension into the margin as 2.8 mm. I remeasured it in the IPZS volume as 2.0 mm, and even adjusting the margin so parts of the characters are in the margin only gave a measurement of 2.2 mm. I again found its total length to be 4.0 mm as opposed to his 4.3 mm measurement.
With these adjustments, even if one merely adjusts Richard’s measurement of the extension into the margin of Acts 13:16f. and Rom 16:5 from 2.8 mm to 2.7 mm, neither of these two, and according to Richard’s own measurements, none of the other 18 undisputed paragraphoi combine noticeably farther extension into the margin and noticeably greater length than any of the characteristic bars, even with his exclusion of the dot from the bar in Matt 18:10, 12 in both Richard’s measurements for that bar's extension into the margin and its total length.
My NTS article p. 610 states, "Only these eight combine noticeably farther extension into the margin and noticeably greater length." It argues throughout that it is this combination of graphic features that sets the characteristic bars apart. I contend that argument still stands and with it all my arguments from statistical probability that the characteristic bars have recognizeably different form and function, and their form and function is best described by the name “distigme-obelos symbols.
Since 46 of Richard’s 56 measurements of the extension of bars into the margin and their total length are longer than mine, I wonder if the primary reason for this is that Richard was measuring images on a computer screen, whereas I was measuring the extraordinarily precise reproduction of Codex Vaticanus B make by IPZS in 1999. Because the original images were analogue and I have been told the computer images were derived from the analogue images, and since computer images are inherently subject to the limitations of screen display, a fair comparison to my measurements could only be done from the IPZS volume.
I leave in few hours for a trip and will not be back in my office until October 11. So I apologize in advance that I will not be able to reply to further comments and questions until I catch up on my work after returning from this trip.
On my blog I posted a discussion, including plots that compare Philip's measurements of the bars with mine. See here. I conclude that he has made systematic errors in his measurements and that his statistical case evaporates when these are corrected.
ReplyDeleteI, Philip B. Payne, want to thank Richard Fellows for drawing to my attention one additional characteristic bar that I had missed on line 12 of 1285B, Mark 6:11. This bar extends exceptionally far into the margin toward the adjacent distigme and is long. There is a large gap in middle of the text at the exact letter that a 15-word addition is entered at the end of Mark 6:11, apparently added from the parallel account in Matt 10:15. This is important for three key reasons:
Delete1. It significantly increases the statistical case that the characteristic bars are distigme-obelos symbols marking the location of widely acknowledged, multi-word additions to the text. The simple mathematical odds that nine randomly chosen lines would all coincide with a multi-word textual variant in the NA28 text is less than one in 31 trillion.
2. It adds to evidence that the text of the Gospels in Vaticanus is very early since now all six (not just all five) of the blocks of text added to the Gospels marked by the Vaticanus distigme-obelos symbols are not in the text of Vaticanus. Thus, the text of Vaticanus comes from such an early part of the stream of the text that it had not been contaminated by any of these six additions.
3. It indicates that the original scribe of Vaticanus marked the location of not just seven, but eight three-words-or-more blocks of added text by a distigme + characteristic bar (obelos) + a gap at the exact letter where the multi-word block of text was added.
It is simply not true that Richard Fellows’ measurements cause my statistical case to evaporate.
In each of the citations of Richard Fellows’ measurements below, I am using the figures he kindly sent to me that he says average his three measurements and adjust them all by an undisclosed amount to keep letters that should be within the margin within the margin. He emailed to me that the figures he posted on Evangelical Textual Criticism were only his first measurement. In each case, I use his measurements of both the extension into the margin and the total length of the characteristic bar at Matt 18:10, 12 even though they exclude from that bar the dot that I judged to be part of the bar. This results in Matt 18:10, 12 having, by Richard Fellow’s measurements, the least extension into the margin of any of the characteristic bars and the least total length of any of the characteristic bars. As I show next, even using Richard Fellow’s measurements, the eight characteristic bars all still fall into a separate category than the twenty undisputed paragraphoi.
Even using Richard Fellows’ measurements, the 8 characteristic bars adjacent to a distigme extend into the margin on average 2.68875 mm. In contrast, the 20 undisputed paragraphoi adjacent to a distigme extend into the margin on average 1.802 mm. Thus, the characteristic bars extend into the margin approximately 50% farther than the undisputed paragraphoi do. It would be more than 50% if the 9th characteristic bar were included because it extends exceptionally far into the margin.
Similarly, even on Richard Fellow’s measurements, the average total length of the characteristic bars (4.4175 mm) is significantly longer than the average total length of undisputed paragraphoi (3.762 mm).
Even on Richard Fellow’s measurements, there is not a single undisputed paragraphos that combines greater extension into the margin than even the shortest characteristic bar’s extension into the margin and also a total length as great as the average total length of the characteristic bars.
Furthermore, even on Richard Fellow’s measurements, there is not a single undisputed paragraphos that combines greater total length than even the shortest characteristic bar’s total length and also extension into the margin as great as the average extension into the margin of the characteristic bars.
In other words, even just considering the combination of extension into the margin and total length, none of the undisputed paragraphoi qualify as characteristic bars.
To be continued
Payne reply to Richard Fellows, part 2:
ReplyDeleteRemember that my NTS article identifies the following characteristics of “characteristic bars” on pp. 620–621:
1. Each occurs immediately after a distigme.
2. Each extends noticeably further into the margin than most bars adjacent to distigmai.
3. Each is noticeably longer than most bars adjacent to distigmai.
4. Each occurs at the location of a widely recognised, multi-word addition.
5. A gap at the precise location of this addition follows all seven apparently original characteristic bars.
Even with Richard Fellows’ measurements, none of the other twenty bars adjacent to a distigme, which fulfills the first characteristic, shares more than two of the remaining four characteristics. Since my statistical probability analysis was based on comparing “characteristic bars” according to this definition, Richard Fellows’ measurements have no effect on my statistical probability analysis.
It should also be remembered that Richard Fellow’s measurements were not made from the most accurate analogue images of Codex Vaticanus B, the 1999 highest resolution color images, but from measurements he told me he made on screen. Consequently, they were limited by both the lower resolution on-screen digital images and by the limitations of his screen’s resolution.
I do not plan to add comments to Richard Fellow’s blog post because he refused to make many corrections I identified, such as attributing to me data I did not give. Furthermore, some of the measurements on his charts are nowhere close to the measurements he provided to me. The chart of number of words in the variant would have dramatically demonstrated my case if Richard Fellows had plotted the chart as I requested, according to the number of words added to the text as I provided them to him rather than his own count, attributed to me.
Philip, I checked the data that I sent you, and it is exactly what I plotted.
ReplyDeleteYou are right that none of the 20 "undisputed paragraphoi" qualify as "characteristic bars" under your new criteria. The problem is that most of your "characteristic" bars do not qualify as "characteristic bars" either, using the same criteria.
I have to move on to other work, but you may want to plot bar length against number of added words for ALL the bars (not just the 28 with dots). You would then be able to see whether this larger data set shows any statistically significant correlation. Do the measurements without reading the text, to avoid bias.
Richard Fellows writes, “You are right that none of the 20 ‘undisputed paragraphoi’ qualify as “characteristic bars” under your new criteria. The problem is that most of your ‘characteristic’ bars do not qualify as ‘characteristic bars’ either, using the same criteria.”
ReplyDeleteIt is simply false that “most of your ‘characteristic’ bars do not qualify as ‘characteristic bars’ either, using the same criteria.” As documented below, all eight characteristic bars qualify as “characteristic bars.”
In the context of comparing the average extension into the margin and length of the eight characteristic bars to the twenty undisputed paragraphoi my article states, “Of the twenty-eight bars following a distigme, only these eight combine noticeably further extension into the margin with noticeably greater length.”
By Richard Fellows’ own measurements, each of the eight characteristic bars not only extends farther into the margin than the average of the twenty undisputed paragraphoi, each is also longer than the average length of the twenty undisputed paragraphoi. Consequently, even just considering the combination of extension into the margin and total length, all of the eight occurrences identified in my NTS article qualify as characteristic bars, whereas none of the undisputed paragaraphoi qualify as characteristic bars.
By Richard Fellows’ own measurements, by far the least extension into the margin of any of the eight characteristic bars is 2.24 mm, and that is much lower than all the others since he did not include the dot at the beginning of the bar at Matt 18:10, 12. Even with this reduction in length, this lowest extension into the margin of all eight characteristic bars is noticeably longer than the average extension into the margin of the twenty undisputed paragraphoi by the average of his measurements, 1.802 mm.
By Richard Fellows’ own measurements, by far the shortest of any of the eight characteristic bars is 3.88 mm, and that is much lower than all the others since he did not include the dot at the beginning of the bar at Matt 18:10, 12. Even with this reduction in length, this bar is noticeably longer than the average length of the twenty undisputed paragraphos bars by the average of his own measurements, 3.762 mm.
Therefore, all eight characteristic bars qualify as “characteristic bars.” And this does not even include the other two characteristics of the characteristic bars: that each occurs at the location of a widely recognised, multi-word addition and that a gap occurs at the precise location of this addition following all seven (now eight) apparently original characteristic bars.
Philip, you are now not judging your 8 characteristic bars by the same criteria as your paragraphoi. For example, you are judging your undisputed paragraphoi against "the average total length of the characteristic bars", but you are judging your characteristic bars against "the average length of the twenty undisputed paragraphos bars". The undisputed paragraphos at Acts 13.16f, for example, has an extension into the margin of 2.60mm and a length of 4.13mm. Both these measurements are greater than those of the characteristic bar at Luke 1:28 (2.39mm and 4.00mm). However creative you are, there is just no way to find criteria that enable you to disqualify Acts 13.16f but allow Luke 1:28, for example (with my impartial measurements). And Dirk is right, of course.
DeleteUpon returning to my office today, I checked the color of the ink of both the dot and the bar that appears to be an extension of it at Matt 18:10, 12 in the most faithful reproduction of the actual ink colors of Vaticanus, the 1999 Codex Vaticanus B facsimile. In that facsimile, both the dot and the bar extending to its right have the same color. Since it is common for there to be gaps without visible ink immediately to the right of the beginning of bars and since this dot is aligned properly for this bar to be an extension of it, it probably should be regarded as part of this bar. See the first photograph on p. 613 in my NTS article, and compare the three similarly interrupted bars at 1505B lines 26, 29, and 36. When a dot makes sense as part of the bar, but makes no sense interpreted as unrelated to the bar, as this one does, this favors the judgment that it is part of the bar. Since Richard Fellows left it open that the dot may be part of the bar, it is not fair of him to cite this as an illegitimate characteristic bar.
DeleteIncluding the dot, there is not a single undisputed paragraphos that has anywhere near as great extension into the margin or total length as the characteristic bar at Matt 18:10, 12. The same is true of the ninth characteristic bar Richard Fellows helpfully pointed out at Mark 6:11 (1285 B). It has a gap in the text at the exact point of a 15 word later addition to the text that is not the Vaticanus text and so further confirms that the text of the Gospels in Vaticanus comes from a stream of the text so early that none of the six additions marked by a distigme + characteristic bar had yet entered that text.
Richard Fellows writes, “The undisputed paragraphos at Acts 13:16, for example, has an extension into the margin of 2.60mm and a length of 4.13mm. Both these measurements are greater than those of the characteristic bar at Luke 1:28 (2.39mm and 4.00mm). However creative you are, there is just no way to find criteria that enable you to disqualify Acts 13:16 but allow Luke 1:28, for example.”
I note that Richard Fellows reduced the total length of the characteristic bar at Luke 1:28 from the length he emailed to me as being his average of three measurements from 4.01 mm to 4.00 mm. This is not much, but contrasted to his measurement of 4.13 mm for Acts 13:16, it makes a slim difference even slimmer.
Richard Fellows’ twice repeated “for example” gives the impression that there are many such cases. Indeed, his previous post asserted that “most of your ‘characteristic’ bars do not qualify as ‘characteristic bars’.” Yet, excluding Richard’s truncated measurements the bar without the dot at Matt18:10, 12, Luke 1:28 is the only one of the nine characteristic bars for which by his own measurements there is any undisputed paragraphos with both a greater extension into the margin and greater length.
If one considers as part of the length of the bar at Luke 1:28 the full length of its downward angling right extension measured along this angle, as I do, it is doubtful that even Luke 1:28 is an exception.
In any event, my article repeatedly argues that there are five features of the characteristic bars, and of the four besides being adjacent to a distigme, none of the undisputed paragraphoi displays more than two of those four features.
to be continued
Payne responds to Richard Fellows, part 2:
DeleteAs I look again at Richard Fellows’ graphs, I realize that although in the previous graphs he marked with a red x each of Payne’s “characteristic bars,” in the final graph plotting each bar’s extension into the margin against its length, he has included in “Payne’s ‘multi-word variants’” marked with a red x both instances of multi-word variants in undisputed paragraphoi. The chart was introduced with “Philip tells me that it is a combination of features that define ‘characteristic bars’, so here is the plot of extension into the margin against length.” Consequently, I expected that the chart would contrast the instances I had identified with this combination of features, namely the “characteristic bars” to the “undisputed paragraphoi, and I assumed that as in the previous graphs each red x represents one of the characteristic bars with this combination of features. This expectation was reinforced by the sentence below the graph, “I suppose we could define a "characteristic bar", according to the line shown, … and there would still be 4 cases of "multi-word variants" that do not make the cut.” The “cut” here is clearly the “cut” to be included in characteristic bars. But neither of the two with a red x that are lowest and farthest left have I ever identified as “characteristic bars” or as “Payne’s multi-word variants” marked by characteristic bars. Neither of the two red x entries on the far left of the chart are any where near close to meeting the greater extension into the margin criterion or the greater length criterion. It was because I wrongly assumed that the red x marks represented “characteristic bars” that I responded, “some of the measurements on his charts are nowhere close to the measurements he provided to me.” I apologize for this false statement.
Since distigmai mark the location of textual variants, it is simply to be expected that out of twenty one or more would be multi-word additions. So that by itself could not make them characteristic bars. Furthermore, the next lowest and farthest left is only there because Richard Fellows did not include the dot with the bar at Matt 18:10, 12. If this chart had highlighted in red only the characteristic bars, it would have illustrated their sharp contrast with the undisputed paragraphoi, especially if it had included the dot with the bar at Matt 18:10, 12.
"this bar is noticeably longer than the average"
ReplyDeleteSo less than 2/10 mm in length above an average is noticeable? [Double emphasis on 'above an average' and on 'noticeable'].
You've lost me.
Remember that I used for these figures the lowest measurements Richard provides, namely his measurements for the bar at Matt 18:10, 12 without the dot. I was showing that even with his measurements, the characteristic bars can be distinguished in light of their five features.
DeleteUpon return to my office today, I checked the color of the ink of both the dot and the bar that appears to be an extension of it at Matt 18:10, 12 in the most faithful reproduction of the actual ink colors of Vaticanus, the 1999 Codex Vaticanus B facsimile. In that facsimile, both the dot and the bar extending to its right have the same color. Since it is common for there to be gaps without visible ink immediately to the right of the beginning of bars and since this dot is aligned properly for this bar to be an extension of it, it probably should be regarded as part of this bar. See the first photograph on p. 613 in my NTS article, and compare the three similarly interrupted bars at 1505B lines 26, 29, and 36. When a dot makes sense as part of the bar, but makes no sense interpreted as unrelated to the bar, as this one does, this favors the judgment that it is part of the bar. Since Richard left it open that the dot may be part of the bar, it is not fair of him to cite this as an illegitimate characteristic bar.
Including the dot, there is not a single undisputed paragraphos that has anywhere near as great extension into the margin or total length as the characteristic bar at Matt 18:10, 12. The same is true of the ninth characteristic bar Richard Fellows helpfully pointed out at Mark 6:11 (1285 B). It has a gap in the text at the exact point of a 15 word later addition to the text that is not the Vaticanus text and so further confirms that the text of the Gospels in Vaticanus comes from a stream of the text so early that none of the six additions marked by a distigme + characteristic bar had yet entered that text.
Remember, too, that I repeatedly emphasize in my article that the small number you highlight is the one of the four graphic characteristics that is the least striking. Even by Richard’s measurements the nine characteristic bars have on average over 50% greater extension into the margin that the undisputed paragraphoi.
In any event, my article repeatedly argues that there are five features of the characteristic bars, and of the four besides being adjacent to a distigme, none of the undisputed paragraphoi displays more than two of those four features. This sharply contrasts with all nine characteristic bars.
On 9/27/2017 Peter Head commented, “I think the bars and the gaps are two systems marking the same thing - a pause in the thought of the text. In the case of the gaps these are the product of the scribe …. In the case of the bars these are a later addition to the text not the part of the initial production (evidence for this is that occasionally they are independent of the gaps).”
ReplyDeleteThanks for this explanation, Peter. Has any publication ever defended your new view that the paragraphoi in Vaticanus are not part of its original production? If so, what are its bibliographic details? The copy you sent me of your SBL paper on the Vaticaus Marginalia refers twice to “the original paragraphoi.”
If Peter Head is correct that the paragraphoi are later than the work of scribe B, even if there is an undisputed paragraphos that both extends into the margin slightly more and is slightly longer than the shortest characteristic bar, it would provide no evidence against the distigme + characteristic bar + gap in the text being a mark of the exact point of a multi-word addition to the text. In that case, at the time scribe B left the original gap at the exact point of a multi-word addition and, according to my argument, signaled this in the margin with a distigme-obelos symbol, there would have been no paragraphoi in Vaticanus with which the obelos portion of the distigme-obelos symbol could be confused.
Philip, as you see it, what would falsify your view?
DeleteSince there are several inter-related aspects of my view, there are various ways it might be falsified. My research began in the 1990s trying to identify the purpose of distigmai in Vaticanus. What I found was a pattern of a far higher incidence of NA apparatus textual variants in lines marked with distigmai than in the following twenty lines of text. A chi-square analysis of the results showed that the likelihood of this strong a contrast happening at random is far less that one in ten thousand. As far as a know, every subsequent analysis of the distigmai, including at least one Ph.D. dissertation, has confirmed that distigmai mark the location of textual variants.
ReplyDeleteHow could that have been falsified? Theoretically, other broader studies could have found that there is no significant correlation between distigmai and the location of textual variants in other parts of Vaticanus. Or theoretically one might have shown that the textual variants I claimed to occur at these locations do not in fact occur there. No one has ever alleged this, and in fact, since I identified each location precisely and the nature of the textual variant, anyone can confirm that they do exist.
Similarly, in the distigme-obelos current study, if it were shown that the textual variants I cite on pages 612–615 do not exist, this would falsify my view for that instance or those instances. In the distigme-obelos investigation as in my distigme investigation, chi-square analysis of the results showed that the likelihood of this strong a contrast happening at random is far less that one in ten thousand. Now that Richard Fellows has identified a ninth case that perfectly fits the pattern of distigme-obelos symbols marking the location of a widely acknowledged, multi-word block of added text beginning at the exact location of a gap in the following line of text, the chi-square results are so high they are off every chi-square chart I found on the Internet.
Although agreeing that distigmai mark the location of textual variants, Curt Niccum, proposed that all the distigmai were added to Codex Vaticanus by Sepulveda in the sixteenth century. Paul Canart, however, confirmed that 51 of these distigmai match the apricot color of the original ink on the same page of Vaticanus and that others have apricot color ink protruding from under the dark chocolate color of the reinking of the text about AD 1000, demonstrating that at least these original-ink distigmai date to the original publication of Vaticanus. David Parker rejected C. Niccum’s arguments that it is ‘likely’ the distigmai ‘originated with Sepulveda … Payne successfully vindicated his case [against Niccum’s critique]’ in D. Parker, ‘Through a Screen Darkly: Digital Texts and the New Testament’, JSNT (2003) 395–411, at 408 n. 17. See the scholars in my NTS article, pages 605–606 and n. 7.
Peter Head alleged that distigmai are the result of a single process in the sixteenth century. The distigmai in LXX G at the exact location of known Greek textual variants, however, confirm that distigmai were in use in the fourth/fifth century. See the photograph and description on p. 607 and n. 11 in my current NTS article. The original ink distigmai identified by Paul Canart, the distigmai in LXX G, and the one distigme with one apricot color dot and one dark chocolate color dot accompanying my article on the Vaticanus distigmai in plate 8 of Le manuscrit B de la Bible falsify Peter Head’s view that the distigmai are all the result of a single process in the sixteenth century. Now that distigmai have been confirmed to exist in the fourth to fifth century LXX G, and the most renowned scholar of Codex Vaticanus has confirmed that 51 of the ones in Vaticanus match the apricot color of the original ink of Vaticanus on the same page and that some apricot color ink protruded from under some reinked distigmai, I cannot imagine how the dating of at least these to the original production of Vaticanus could be credibly falsified.
to be continued
Payne response to Peter Gurry, part 2:
ReplyDeleteSince one of the 51 distimgai that Canart confirmed matches the original ink color of Vaticanus is part of a distigme-obelos symbol and since eight (including the distigme-obelos symbol at Mark 6:11, 1285B line 12, that Richard Fellows kindly brought to my attention) are followed by a gap (that only the original scribe could put in the text) at the exact letter where the widely acknowledged, multi-word textual variant occurs, it would be correspondingly difficult to falsify the claim that these go back to the original scribe.
Theoretically, one might falsify my view by demonstrating that scribe B was not aware that obeloi mark added text. My NTS article, however, on pages 608–609 shows that it cannot be credibly denied that scribe B penned obeloi, including one in the middle of text (1016 C12 at Isa 14.23) or that scribe B did not understand that obeloi mark added text. To the contrary, scribe B repeatedly added notes explaining that: “the [lines] marked with an obelos contain [text] not in [the] Hebrew [text].”
My observation of the paucity of high stops in the Gospels could be falsified by demonstrating that high stops are abundant in the Gospels, or similarly abundant to the high stops in the epistles. Having gone through the Vaticanus text repeatedly, I have no fear of this being falsified.
I suppose theoretically one might be able to prove that scribe B intended to deceive readers by making the text of Vaticanus look primitive by removing high stops. The many findings I delineate in my NTS article give evidence of the exact opposite, that scribe B was extraordinarily faithful to preserve the text of the Vaticanus exemplars, even where scribe B disagreed with that text, as shown by the obeloi in the Vaticanus prophetic books and the one distigme-obelos in the epistles. Not only is the form of the text of the Vaticanus Gospels primitive with virtually no high stops, none of the six blocks (including 1285 B) of added text marked by distigme-obelos symbols are in the text of the Vaticanus Gospels. In contrast, the one block of added text a distigme-obelos marks in the Vaticanus epistles is in the Vaticanus text. These distigme-obelos symbols can be ascribed to scribe B since only scribe B could leave the gaps in the text associated with these distigme-obelos symbols. The lack of these six additions and of high stops shows that the text of Vaticanus comes from a portion of the text stream that had hardly any punctuation marks and had added none of these six blocks of text known to scribe B. Furthermore, it is now the consensus of text critical scholars that Vaticanus is a remarkably faithful guide to a very early form of the text, especially in the Gospels.
to be continued
Payne response to Peter Gurry, part 3:
ReplyDeletePerhaps a more important questions is, how could one confirm my view that there is pattern of distigmai with characteristic bars followed by a gap in the text at the exact location of a widely acknowledged, multi-word addition to the text? The best confirmation would be if someone were to identify another distigme adjacent to a characteristic bar followed by a gap in the text at the exact location of a widely acknowledged, multi-word addition to the text. I use “characteristic bar” as I do repeatedly in the article as sharing five characteristic traits: a bar (1) adjacent to distigmai (2) that extends noticeably farther into the margin than most undisputed paragraphoi adjacent to distigmai, (3) is longer than the average length of undisputed paragraphoi adjacent to distigmai, and is (4) followed by a gap (5) at the exact location of a widely acknowledged, multi-word addition to the text.
If another distigme adjacent to a bar extending unusually far into the margin could be found, the likelihood of it being followed by a gap at the exact point of a multi-word addition in a random distribution is small, unless my view is correct. The likelihood of a multi-word textual variant of any sort listed in the NA28 is less than one in 31.8 lines, and this would have to be a particular kind of multi-word variant, namely a multi-word addition to the text. Furthermore, the block of added text would have to occur at the exact location of a gap in the text. Most lines do not have gaps. And since there are sixteen letters to a line on average in Vaticanus, it would have to happen at the exact letter the multi-word addition begins.
Thanks to Richard Fellows, this confirmation has become an actuality. The characteristic bar immediately to the right and below the distigme at Mark 6:11, 1285B line 12, extends into the margin farther and is longer than any of the twenty undisputed paragraphoi that happen to occur adjacent to a distigme. It is followed by a gap in the text at the exact point the NA28 lists many manuscripts adding a 15-word block of text. Does anyone really think all eight of these cases share all five of these characteristic traits by coincidence? If so, I challenge that person to find a comparable case of a symbol that occurs at least eight times in a manuscript, all of which coincide with a specific kind of textual variant that occurs on average less than once in 30 lines of text, but where this is mere coincidence and the symbol did not in fact identify that specific kind of variant.
I have faith that we are a long way from the record here but personally I don't ever remember this many comments where no one has admitted they were wrong about something.
ReplyDeleteI'm wrong.
ReplyDeleteSomebody else has probably said this, but surely there would have to be some sort of marker for the end of the added section.
ReplyDeleteJan Krans and I have each published short articles in the current issue of NTS, rebutting Philip's 2017 NTS article. Jan and I worked independently, but our articles seem to compliment each other quite nicely, and they come to the same conclusion. You can access my article for fee by using the link on my blog here.
ReplyDelete(Should read 'free'!) Good work, on which it is hard to comment without getting a ruler out. But it strikes me as hardly conceivable that, if somebody were designing a system of signification for the margins of a manuscript, they would make a horizontal line mean two different things depending on its length. There are so many other possible marks that would be much clearer and less dependent on accuracy in a) making and b) comprehending them.
DeleteTo save the less expert of us a little time, 1 Corinthians 14.28 ff is here: https://digi.vatlib.it/view/MSS_Vat.gr.1209/1478 (page 1474). Surely if the scribe wanted to mark added text they would mark the end as well as the beginning of the added section? But there are no marks at the end of verse 35.
DeleteNot only Payne argues that there's a significance in interpretation of a symbol (in this case a horizontal line) based on a difference in its length, but what goes unmentioned in many articles, is that that difference is only about 1 mm! Given that all lengths vary between 3 and 5 mm, or a spread of 2mm, how is a 1 mm even within the physical control of a scribe, let alone significant enough to mark such an important thing as a textual variant?
DeleteAnonymous, your point is valid, but has been addressed in the comments, and in my NTS article. Payne's arguments about dots and dashes are not valid, but he is nevertheless correct that 1 Cor 14:34-35 was an interpolation.
DeleteMy uncle, Phil Payne, has posted a counter-rebuttal to the Fellows and Krans critiques on his website, https://www.pbpayne.com/. Enjoy.
ReplyDelete