The Tors-Costa Debate, Part 2

THE DEBATERS RESPOND

            At the outset of the second part of the Tors-Costa debate about how the original text of the New Testament should be reconstructed, technical difficulties caused Tony Costa’s voice to be briefly inaudible.  When the sound resumed, he was addressing Tors’ claim that the Alexandrian Text has an error in Mark 1:2, where the main part of the quotation is from Malachi.  (Tors has argued in an online essay that although evangelical apologists argue that it was not unusual to blend separate quotations and attribute them to a single source, the evidence for the existence of such a custom is extremely elusive.)  Let’s resume there.  Once again, I will summarize the debaters’ statements and offer comment in italics. 

            

            Consider – Costa reasoned – Matthew 27:9, where words that are found in Zechariah 11:12 are attributed by Matthew to Jeremiah.  Some copyists saw the difficulty and therefore altered the text.  Similarly in Acts 20:28, where the Textus Receptus has a difficult reading that seems to imply that God has blood, scribes altered the text.   (It may have been ill-advised spin to say that “a  Majority Text manuscript” changed the difficult reading to an easier one in Matthew 27:9, for this only makes the point stand out more clearly that the vast majority of manuscripts do not avoid the more difficult reading in that case; most scribes thus acted contrary to the premise that Costa is defending.) 

            Byzantine manuscripts, Costa asserted, changed the text in Matthew 27:9.  He provides their identities:  Codex 22, which has Zechariah’s name in the text, and “Family 33, which is also Majority Text, omits the prophet’s name altogether.”  

            (There are just two problems with that.  First, minuscule 22’s text frequently diverges from the Byzantine norm; it is not a typical Byzantine manuscript.  Second, Costa misinterpreted a printed reference to “F 33” as if the F is an abbreviation for  “family” but that is not what it means (there is no “family 33”); it represents Codex Boreelianus, and 33 refers to minuscule 33, which is, according to Metzger, “an excellent representative of the Alexandrian type of text.” Thus Costa has misdirected his accusation of scribal unreliability in two ways.  First, he is illustrating the dangers of relying on small minorities of manuscripts; that is where the non-original reading is found in this case.  The approach that he is arguing against is precisely the approach that avoids the adoption of such errors.  Second, he has unintentionally exposed the unreliability of a chief member of the group of manuscripts that he is trying to convince his listeners is the most reliable.)  

Tony Costa, demonstrating
a rationalistic approach.

            Next, Costa addressed Tors’ statement about Byzantine readings in the papyri, responding (as expected) that none of the papyri exhibit strong and sustained patterns of agreement with the Byzantine Text.  (This was not a very effective response, since Tors had only claimed that the papyri contained mixed texts; the question being how Byzantine readings got into the mix if the Byzantine Text didn’t yet exist.  It’s like finding a small mound containing salt, pepper, and paprika in a kitchen where there is not supposed to be any paprika.  Simply saying that a container full of paprika was not found does not change the implication.)

            Costa then said, “John said that there are mistakes in the Majority Text,” and as an example of a type of mistake, he mentioned harmonizations – specifically, harmonizations in the Lord’s Prayer in Luke (i.e., Luke 11:2-4), where, in the Majority Text, “The Lord’s Prayer in Luke 11 has expanded to match the Lord’s Prayer in the Gospel of Matthew.”  Costa put things more forcefully:  “In the later manuscripts, you’ll notice that the Lord’s Prayer in the Gospel of Luke is exactly the same as the Gospel of Matthew, with the doxology added to the end, ‘For Thine is the kingdom, the glory, and the power, and so forth.’”

            (Which is forceful – but wrong.  In the Byzantine Text, where Matthew 6:11 has δὸς, Luke 11:3 has δίδου.  Where Matthew’s text has σήμερον, Luke’s text has τὸ καθ’ ἡμέραν.  Where Matthew’s text has τὰ ὀφειλήματα, Luke’s text has τὰς ἁμαρτίας.  Where Matthew’s text has ὡς και, Luke’s text has και γὰρ.  And – passing by some other differences – the Lord’s Prayer in Luke in the Byzantine Text does NOT have the doxology that appears in Matthew 6:13 in the Byzantine Text.  

            (Not only is Costa’s claim false, but it can be reversed, and the absence of “but deliver us from evil” in the Alexandrian Text of Luke 11:4 can be used as an example of Alexandrian omission, and the absence of the doxology in the Alexandrian Text of Matthew can be used as an example of Alexandrian harmonization – in this case, a conformation of Matthew’s text to the parallel in Luke.  Fortunately for Costa, his friendly opponent did not pursue these points.)

            (A better example of a rationalistic approach in action – in which someone paints a reasonable-sounding picture of how scribes made harmonizations, while failing to carefully examine the evidence – could scarcely be hoped for.)

            (Also, the order in the doxology is kingdom, then power, then glory.)

            Then, in response to Tors’ observation about the plethora of scribal errors in Papyrus 66, Costa said that scribes made so many mistakes because they were working under stressful conditions, specifically, Roman persecution.  The Alexandrian scribes, Costa proposed, were “on the run.” (I suspect/hope that this idea came to Costa on the spot, without a test of its plausibility.)  

            The last section of Costa’s response to Tors’ opening statement was a disjointed collection of miscellaneous and tangential points: (One of these points ended up being revisited later in the debate.)     

            ● God had used the Vulgate and a variety of versions based on different texts (Tors’ focus had been altogether elsewhere; perhaps this observation was made as if to suggest that God does not share the “very adversarial approach” that Costa attributed to Tors.  An adversarial approach in a debate; what a concept.) 

            ● A “majority rules” approach does not work with the Old Testament text.  (Costa thus opposed a view that Tors never advocated or even mentioned in his opening statement.)   

            ● The Majority text does not have “daily” in Luke 9:23; isn’t that a scribal harmonization?  (Tors never answered this question, but if he had, the answer would probably be that the minority reading with “daily” (καθ’ ἡμέραν), familiar due to its inclusion in the Textus Receptus, is an Alexandrian harmonization from First Corinthians 15:31.)

            ● A note (in Vaticanus at Hebrews 1:3) rebuking a copyist who tried to correct the text shows that Alexandrian copyists faithfully transmitted the text.  (Costa seemed unaware that the old reading which the note-writer is zealous to preserve (φανερον) is a mistake; the reading there in Nestle-Aland and in the Byzantine Text is φέρων.)

            ● Textual criticism is very complicated and people should read Metzger & Ehrman’s The Text of the New Testament.  (Because there’s nothing like recommending a book by an ecumenical Bible-condenser and an atheist to convince people at church that your position is faith-friendly.)

            ● Modern translations such as the ESV (which Costa called the Evangelical Standard Version; this was a joke, I think) do not really deny any important doctrines. 

Then Tors responded. (It was very advantageous to go second in this part of the debate, because Tors could thus respond not only to Costa’s opening statement but also to Costa’s response.)  He used the words that Costa had just delivered as part of the framework for his response.    Tors made eight corrective points before addressing Costa’s earlier remarks:

            ● Nobody is saying that all the Byzantine manuscripts agree perfectly; of course the Byzantine manuscripts are not 100% uniform.  That does not say anything about the validity of the Majority Reading approach. 

            ● About the note in Vaticanus:  this supports exactly what Tors has been saying:  although some scribes, on rare occasions, tried to alter the text, their efforts were opposed.

            ● Regarding Matthew 27:9:  we see some scribes alter the text, but how many?  A very small number, nowhere remotely close to a majority.  This once again supports the view that the vast majority of scribes did not consciously alter their texts, and most people did not accept such rare alterations.

            ● About the rival readings in Mark 1:2:  is the Byzantine reading correct?  Yes; even if one posits a thematic connection to a passage in Exodus (I think that must have been mentioned in the part of Costa’s statement where there is no audio), Moses was a prophet, so it’s not a problem.  But is the Alexandrian reading correct?  No.  The apologists try to say, “There was this scribal practice of ascribing several quotes from several sources to one source” but (Tors says) when he checked it out, he found no substantial basis for that assertion.  (This is not an approach that I would take; rather, granting that a writer could validly cite more than one thematically related passages from the Old Testament but only name one of them specifically – I would point to the strong scribal tendency to identify unnamed prophets and other unnamed individuals in the Gospels as evidence that in cases such as Mark 1:2, the idea that the less specific reading is to be preferred should be brought to bear.)

            ● What about Matthew 27:9?  It’s not a problem because the text refers to what was spoken by Jeremiah.  Nothing precludes the idea that Jeremiah spoke a prophecy that was also written by Zechariah.  (It might have been helpful to point out the parallels between parts of Jeremiah 49, and parts of the text of Obadiah.)  Similarly in Matthew 1:23, Matthew refers to a prophecy that was spoken but which is not found in written form anywhere in the books of the prophets. There simply was not a custom of saying that what had been written by one man was written by someone else.  (In my view, a case can be made that Matthew – writing to a readership well-acquainted with the Old Testament writings – felt that it would not be problematic to draw a thematic link between a passage in Jeremiah, and a passage in Zechariah without naming both sources.  That does not mean that Mark was in a similar situation; for further analysis regarding Mark 1:2 see my online essay about that variant-unit.) 

            ● It is not an ad hominem argument to call Griesbach a “rationalist.”  That is an accurate assessment of his philosophy and of the basis for his text-critical assumptions:  he valued what seemed reasonable, rather than what was tested and experienced. 

            ● Regarding Papyrus 66:  Roman persecution came in waves; the scribes simply were not under the kind of constant stress that Costa has described.  (Costa’s description is speculation through and through.  Tors surely pulled his punches here.)

            ● Is the majority always right?  Not always; after all, Tors is advocating a view held by a minority of scholars.  But showing that some majorities are wrong does nothing to show that the Majority Text approach is wrong.

            Tors then presented a map of the Byzantine Empire as it existed in the year 600.  Revisiting Costa’s claims about why the Alexandrian Text disappeared, Tors points out that the Byzantine Empire included a huge swath of territory; it was not a localized corner of text-production.  Against Costa’s claim that the Byzantine Text did not become the majority until the ninth century, when we look at the use of the Byzantine Text in this huge territory, we have to ask, if the Alexandrian Text was in the majority, why did the people in this area stop using it?  Roman persecution ended in the 300’s; Christianity was the official religion of the Roman Empire by the 400’s – and we see the Byzantine Text used in the 400’s and 500’s.  Why not the Alexandrian Text, if that was the majority? 

             Also, we keep being told by advocates of the Alexandrian Text that the early church fathers in the first 300 years quoted the Alexandrian Text.  But when we examine their quotations, that’s arguable.  (Tors built the case gradually:  arguable, then questionable, then . . . )  It is mistaken.  The early patristic quotations of the New Testament are a textual mishmash; they do not support one form or another – but they side with the Byzantine Text more than with anything else.  (It would have been helpful to provide a few concrete examples, such as Clement’s text of Matthew.)  In addition, the pro-Alexandrian case benefits from an absence of evidence where early patristic evidence is concerned:  there are no early patristic writings from Antioch and the surrounding area that are substantial enough to analyze.  (Tors seemed to have more to say about this, but he apparently had some technical difficulty with his digital slides, and moved on.)

John Tors, commenting about
intentional textual changes.

            Tors then addressed Griesbach’s fundamental premise:  did scribes intentionally change the text? (Costa had mentioned a couple of passages that he considered alterations.  Tors could have easily gotten distracted and attempted to focus on those specific passages in detail – but instead he kept his focus.)  Without granting that any reading in the Majority Text is not original, Tors simply noted that he had never said that deliberate alterations never happened; he only insisted that they were rare.  And Costa only gave a few examples – because they are rare.  (Tors could have really hammered this point by pointing out that even the most prominent examples of harmonization that Costa gave – the exact harmonization of the Lord’s Prayer, and the inclusion of the doxology in the Lord’s Prayer in Luke 11 – are non-existent.)    

            As Tors wrapped up this part of the debate, he once again stated that although the reasoned eclectic approach yields a text that contains errors, the Majority Reading approach yields a text that is inerrant.  And it is the original inerrant text the textual critics should aim to reconstruct, even though God has shown that He is able to use imperfect compilations and even flawed translations such as the Vulgate.

            In this part of the debate, Costa presented practically no point for which Tors (even without taking advantage of Costa’s gaffes) did not have an effective answer.  Tors did not, however, spend much time addressing Costa’s claim that it would be “dangerous” to apply the Majority Reading approach to the Old Testament text, probably because Tors never suggested doing so, and because the announced subject of the debate was not the Old Testament text.  Somehow this soon drifted into the main current of the debate. 

            Coming soon:  Part 3: The Debaters Cross-Examine Each Other.    

The Tors-Costa Debate, Part 1

The Tors-Costa Debate
can be viewed at YouTube.

            Last week in Toronto, John Tors, an advocate for the Majority Text, won a debate against Tony Costa, who attempted to defend the Nestle-Aland compilation.  The stated purpose of the debate was to examine the Greek texts behind popular translations of the New Testament.  Today (and in the next two posts, God willing) I will summarize their debate, and offer comments in italics at various points.

            The moderator, Johnny Yao-Chung Chao, welcomed guests to the Toronto Free Presbyterian Church, introduced the debaters, and offered an opening prayer.  The debate was designed to consist of opening statements, followed by responses and a cross-examination period, followed by a time of spontaneous questions from the listeners. 

            Dr. Costa began with a standard summary of the purpose and materials of New Testament textual criticism, noting that the manuscripts contain many variants – mostly trivial, but not all – and he asked, in light of those variants, “How can we be sure and confident that we still have the New Testament today?” – “How do we get back to the original?”  Arguing for the Alexandrian Text, Costa stated that “During the first millennium, the Alexandrian manuscripts were actually in the majority.  The Byzantine manuscripts did not become the majority until the ninth century A.D.”

            At this point, Costa used a graphic that resembles the graph that one can consult on page 153 of James White’s book The King James Only Controversy (first edition).  (Costa’s claim, similarly, is found on the opposite page; White wrote that the Byzantine text “does not become the “majority” until the ninth century.”  That is not a realistic appraisal of the implications of the evidence; perhaps I will write more about this another day.  White’s/Costa’s chart shows how many manuscripts have survived from each century.  That’s all.  One would think that those who object to the idea of “normal” transmission would also object to the idea of “normal” survival, but apparently not when attempting to bolster their case.)

            Costa proposed that use of Alexandrian Greek manuscripts declines because of three factors:  (1)  the destruction of manuscripts by Roman persecutors, (2) the shift from Greek to Latin, and (3) the expansion of Islam.  Costa proposed that in countries governed by shariah-law, it was not safe to have scriptoriums.  (However, many monasteries in Egypt and other countries continued to produce New Testament manuscripts long after the territory in which they were situated was under Islamic rule.)  

            Costa then asserted, “For the first 300 years of the history of the church, all of the church fathers quoted from the Alexandrian text-type manuscripts, not from the Byzantine text-type manuscripts.  At least not till about 350, with John Chrysostom.”  Costa argued that the majority can change, and that the majority is not always right – just look at the majority of Germans who supported the Nazis, for example.  

            As he concluded his opening remarks, Costa made a theological point and a criticism of the Byzantine Priority view:  “God usually works with the small remnants sometimes.”  (Gaffe:  which is it:  usually, or sometimes?)  And, “The Majority Text does not approach a uniform text.  Maurice Robinson openly admits this.  The Majority Text suffers textual corruption as well.” 

            Tors, in his opening statement, restated the basic question on which the debate was intended to center:  Which  method of textual reconstruction should be used:  “reasoned eclecticism” or the Majority Reading Approach?  He began by addressing the much-repeated claim that it does not matter which method is used:  claims that the differences are trivial, and that no doctrine is affected by textual variants.  And then the gauntlet was hurled down:  “But that is not true.” 

            The textual variant in First Timothy 3:16, Tors insisted, has an impact on doctrine.  And the variant in John 7:8, where P66 and P75 and Codex Vaticanus disagree with the Nestle-Aland compilation, also has an impact.  There are many, many more examples.  (One could wish that he had specified a few more, such as Matthew 27:49 and Mark 6:22.) The survival of at least one doctrine of the faith – inerrancy – depends on which approach is used.

            Tors then reviewed some text-critical guidelines, or canons, utilized by supporters of the Alexandrian Text.  He emphasized the premises behind them, such as: 

            (1)  Scribes were more prone to add than to omit. 

            (2)  Scribes were prone to correct errors. 

            (3) Scribes were prone to harmonize. 

            Tors also pointed out that the critical text depends very heavily on Vaticanus and Sinaiticus, even though they disagree, on average, four times in every five verses.  (Gaffe:  Vaticanus was not discovered in the 1800’s; its New Testament text was reliably edited and made available to researchers at that time; its existence had been known for centuries.)

            In addition, Tors protested that at the root of the critical text is the genealogical method proposed by Hort back in 1881 – a theoretical transmission-history that Hort never bothered to prove. (As Colwell put it:  “That Westcott and Hort did not apply this method to the manuscripts of the New Testament is obvious.  Where are the charts which start with the majority of late manuscripts and climb back through diminishing generations of ancestors to the Neutral and Western Texts?  The answer is that they are nowhere.”) 

            Tors also mentioned that the advocates of the pro-Alexandrian school also cite the discovery of Egyptian papyri as a basis for their position – but before investigating that further, he turned to the just-listed canons, and asked the audience to see if they could detect the basic idea that they express.  That idea, he said, is the fundamental foundation of the reasoned eclectic case:  the idea that scribes altered the text on purpose.  Griesbach – the scholar in the late 1700’s and early 1800’s who developed these canons – believed this, not because he had conducted thorough analytical research about scribal tendencies, but because he embraced a rationalistic philosophy, and these ideas simply seemed to make sense.  His assumptions were accepted by textual critics for 200 years.  However, Tors continued, those assumptions do not align with most of the evidence.  Patristic writers consistently denounce those who altered the text of Scripture.  And one research-study after another – such as James Royse’s – show that the most common scribal error was omission.      

            Tors then turned his attention to the early papyri.  He pointed out that contrary to what James White has claimed, the papyri do not all support the Alexandrian Text.  There is considerable mixture in the text of some of the early papyri.  Tors then showed a graphic, summarizing data from Pickering, who in turn had extracted data from research by Klijn:

            P66:  agreed with Aleph 14 times, agreed with B 29 times, and agreed with TR 33 times.

            P75:  agreed with Aleph 9 times, agreed with B 33 times, and agreed with TR 29 times.

(In the debate, it did not seem very clear at all what Tors’ comparison was comparing, so I will explain:  those numbers are not derived from a consideration of the entire text of P66 and P75; they are part of an analysis (conducted by Klijn, and used by Pickering) of just the parts of John in which P45, P66, and P75 are all extant.  These statistics shouldn’t be relied upon as anything but a demonstration that P66’s text is significantly different from P75 – which is still a significant point, since if P66 and P75, with this much variation, are both Alexandrian, the term doesn’t mean much.  But this evidence-bundle looks carefully picked.  Nevertheless I don’t think anyone will contest Tors’ basic point that some of the papyri have texts that fail to display the patterns of readings displayed in the flagship-manuscripts of any text-type.  Aland acknowledged this, as Tors mentioned.)     

            After mentioning that the early papyri (better:  some of the early papyri) are not strongly aligned with the Alexandrian Text, Tors pointed out that 150 distinctly Byzantine readings have been found in the papyri.  (This finding by Harry Sturz is often belittled by defenders of the Alexandrian Text; a typical response is that early Byzantine readings do not show the existence of an early Byzantine Text.  And that was Costa’s response, almost verbatim.  But such a reaction overlooks the chief implication of Sturz’s research, which is that wherever these readings came from, it was not from a simple amalgamation of readings drawn from Western and Alexandrian exemplars, at least not as we know them.  If there was any such amalgamation-work, it must have involved a third source of readings – in which case, Hort’s main reason to categorically reject distinct Byzantine readings falls to pieces.) 

            Approaching his conclusion, Tors stated that in view of all this – the research that has undermined the “prefer the shorter reading” canon, the analyses that have shown that most scribes simply aspired to accurately reproduce their exemplars’ contents, and the discoveries of early distinctly Byzantine readings that weigh in against Hort’s theory of the origin of the Byzantine Text – “Nestle-Aland is dead.  They don’t admit it, but it is.”

            Tors’ final argument for the Majority Reading Approach in his opening statement consisted of an appeal to statistical analysis.  Using a mathematical model, Tors showed that if a manuscript were copied five times, then, all things equal, any error would have to be copied three times to be the majority reading.  And the number of times an error would have to be reproduced in each subsequent copying-generation would necessarily increase.  Thus the probability of any error being the majority reading is staggeringly low.  Tors said, “This is based on real-life numbers.”  

            (However, what are the numbers based on?  They are a hypothetical mathematical construct, not a reflection of historically verified circumstances – a grid, not a map.  Of course one can imagine a tree that grows 10 branches, with 10 twigs on each branch, and 10 fruits on each twig, but one can also walk outside and observe trees with branches hacked away, twigs broken off, widely different numbers of twigs on different branches, fruit plucked by birds and squirrels, and so forth.  If a reading’s status as part of a majority infallibly implied what Tors says it implies, the Eusebian Sections and chapter-divisions would also be part of the original text.  Still, aside from this, Tors presented some sound reasons why the Majority Reading Approach is more trustworthy than the pro-Alexandrian approach.)

Thus ended the opening statements.  Next:  Part 2:  The Debaters Respond.