Elijah Hixson - Myths and Mistakes on Palaeographic Dating

Steven Avery

Administrator
(look for more SInaiticus)
On the problems with paleographic dating due to the dearth of securely dated manuscripts, see Don Barker, “How Long and Old Is the Codex of Which P.Oxy 1353 Is a Leaf?,” in Jewish and Christian Scripture as Artifact and Canon, ed. Craig A. Evans and H. Daniel Zacharias, SSEJC 13/LSTS 70 (New York: T&T Clark, 2009), 194-96. 19 See especially Guglielmo Cavallo, Ricerche sulla maiuscola biblica, Studi e testi di papirologia editi dall’Istituto Papirologico «G. Vitelli» di Firenze 2 (Florence: Le Monnier, 1967). See also Guglielmo Cavallo and Herwig Maehler, Greek Bookhands of the Early Byzantine Period A.D. 300–800 (London: Institute of Classical Studies, 1987); Pasquale Orsini, Manoscritti in maiuscola biblica: Materiali per un aggiornamento, SAAFLS 7 (Cassino: Edizioni dell’Università degli studi di Cassino, 2005). 20 Turner, Greek Manuscripts of the Ancient World, 20. 21 See, for example, Brent Nongbri, “The Limits of Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P. Bodmer II (P66),” MH 71, no. 1 (2014): 1-35; Nongbri, “Use and Abuse of P52”; Pasquale Orsini and Willy Clarysse, “Early New Testament Manuscripts and Their Dates: A Critique of Theological Palaeography,” ETL 88, no. 4 (2012): 443-74. See also Christian Askeland, “Dating Early Greek and Coptic Literary Hands,” in The Nag Hammadi Codices and Late Antique Egypt, ed. Hugo Lundhaug and Lance Jennott, STAC 110 (Tübingen: Mohr Siebeck, 2018), 457-89.

Dating Myths, Part One

does not allow such precision for undated texts.” Turner warns that for “literary” hands, “a period of 50 years is the least acceptable spread of time.”22 A comparison of P.Ryl. 16 and scribe A of Codex Sinaiticus illustrates the problem that resulted in Turner’s hesitancy to assign any date ranges shorter than fifty years on the basis of paleography alone.23 These two manuscripts are written in the same general style, called “biblical majuscule,” but they were written forty to two hundred years apart from each other. The comparison aims to show why one should not be too eager to accept the earliest possible or even the latest possible dates. P.Ryl. 16 is a fragment of a comedy that was reused as a letter. The letter on the back can be dated to AD 256, giving this papyrus its terminus ante quem, but we do not know how long before AD 256 it was written. In 1967, Guglielmo Cavallo paleographically dated the manuscript to the narrow window of circa 220–225 (!).24 The manuscript is part of an archive, however, which can shed some additional light on its date. In a study of reused papyri with dates on both sides, Turner suggested that P.Ryl. 16 might be as early as AD 150 because other reused papyri in the archive were about one hundred years old before reuse.25 22

97

Figure 5.2. P.Ryl. 16 is a fragment of a comedy with a dated letter on the backside

Foster, “Bold Claims, Wishful Thinking,” 199; Turner, Greek Manuscripts of the Ancient World, 20. Thanks are due to Brent Nongbri, who inspired me to compare P.Ryl. 16 with Codex Sinaiticus. 24 Cavallo, Ricerche sulla maiuscola biblica, 45-46. 25 Turner, “Recto and Verso,” 106. However, there is no guarantee that P.Ryl. 16 was written as early as AD 150. At least one papyrus in this archive was reused after only nine months, and several others after fifteen to thirty years. 23

98

Elijah Hixson

The other hand of our comparison is scribe A of Codex Sinaiticus, which has a terminus post quem because it was produced with the Eusebian apparatus in the Gospels. It obviously could not have been made before Eusebius had developed this cross-reference system, which probably happened during the window of AD 290–340. Codex Sinaiticus could be later but not earlier. On the basis of the handwriting, Cavallo dated Codex Sinaiticus to circa 360 “or a few years later” (!).26 Cavallo described the hand he calls “biblical majuscule” as emerging from a “sober and undecorated script.” “True biblical majuscule” has a “visible contrast between thin horizontal strokes and fatter vertical ones (particularly gamma, pi, tau), while oblique [i.e., diagonal] strokes appear in between (alpha, delta, lambda).” Cavallo continued, “Among late examples . . . the script shows a stronger contrast between fat and thin strokes and decorative buttons at the extremities of the latter, in particular on the horizontal strokes of gamma, delta, epsilon, pi and tau.”27 In other words, the hand that we are comparing should have thick vertical strokes and thin horizontal strokes. The later the hand is, the more the contrast there will be. Later examples of the hands tend to have decorative serifs on horizontal strokes, but earlier examples tend not to have them. In the above comparison, we see handwriting from two manuscripts that Cavallo dated 135 to 140 years apart, and to be fair, P.Ryl. 16 is earlier than Codex Sinaiticus. There are differences between the two hands, admittedly. The epsilon (ε) and sigma (ϲ) are not the same—they are slightly more decorative in Codex Sinaiticus. The phi (φ) in Codex Sinaiticus is more angled than round. Codex Sinaiticus has generally more difference between the thick vertical strokes of nu (Ν) and the thinner diagonal stroke (but not always; see the ν in μενειν [menein]).28 Still, the two hands bear a remarkable similarity. Additionally, tau (Τ) in P.Ryl. 16 has more prominent serifs, or “hooks,” on the ends of the horizontal stroke than tau in Codex Sinaiticus—a feature Cavallo considered to be more characteristic of later examples of “biblical majuscule.” Cavallo himself dated the manuscripts 135 years apart, but this comparison is even more problematic if P.Ryl. 16 is a few decades earlier, as Turner argued. 26
Cavallo, Ricerche sulla maiuscola biblica, 58. Guglielmo Cavallo, “Greek and Latin Writing in the Papyri,” in The Oxford Handbook of Papyrology, ed. Roger S. Bagnall (Oxford: Oxford University Press, 2009), 128-29. 28 It is also problematic to look merely at “test letters,” so I have reproduced the samples in groups of the same words. 27

99

Dating Myths, Part One

P.Ryl. 16 (AD 220–225, according to Cavallo)

Codex Sinaiticus, Scribe A (ca. AD 360, according to Cavallo)

Figure 5.3. A comparison of P.Ryl. 16 (left) and scribe A of Codex Sinaiticus (right), showing their similar hands 

The point of the comparison is to illustrate how difficult it can be to assign a date paleographically. Yes, something is different between the two hands, but can we accurately measure it to be 135 years’ worth of development? Furthermore, “development” assumes that the styles changed in linear fashion, as Turner noted. But, he said, “If [a particular style] was written in several centres it is likely that cross-influences will have affected this style, as they did other styles.”29 To state it alternatively, do we really see 135 years’ worth of linear development between the hands of P.Ryl. 16 and scribe A of Codex Sinaiticus, or are the two similar because the scribes had similar influences? If the latter, we cannot really be confident enough to assign such narrow dates as Cavallo does. Additionally, it is possible for scribes contemporary with each other to have conflicting features with regard to date. P. J. Parsons mentioned features of Codices Sinaiticus, Vaticanus (03), and Alexandrinus (02), in which a scribe or scribes exhibit later features of handwriting than other scribes working at the same time on the same manuscript.30 Even a near-perfect 29

Turner, Greek Manuscripts of the Ancient World, 22. P. J. Parsons, “Review of Ricerche sulla maiuscola biblica by Guglielmo Cavallo,” Gn 42, no. 4 (1970): 380.

30

100

Elijah Hixson

match with a securely dated manuscript is not enough to assign a secure date paleographically.31 Other methods. Other aspects might occasionally provide more information on the date of a manuscript. For example, Turner suggests that the format and dimensions of a codex might shed light on its date.32 Illuminated manuscripts have an additional layer of available information in their art. If a manuscript has been found with an intact cover, the cover might provide a terminus post quem. A deed dated October 7, AD 348, for example, was reused in the cover of Nag Hammadi Codex VII, indicating that the codex had to have been made after that date and suggesting a similar date range for the other Nag Hammadi codices.33 Scientific testing could be of some use as well. The York Gospels (York, Minster Library, Add Manuscript 1) have a terminus ante quem of around AD 1020 because the manuscript came to York through Wulfstan, the archbishop of York, who died in AD 1023, but it was unclear how long before AD 1020 it was written. DNA analysis revealed that an unusually high percentage of the parchment sampled came from female calves. One would expect male calves to be used for parchment because males are less valuable for maintaining and growing a herd. Cattle numbers in the British Isles suffered in the late tenth century because of a cattle plague that swept through between AD 986 and 988, and female cattle would have been especially valuable to repopulate herds after the plague. The best explanation for the unusually high presence of parchment made from female calves in the York Gospels is that its pages were made from the skins of female calves that died in that plague, which, if correct, suggests a strong possibility for a composition date of the York Gospels around the year AD 990.34 31

Brent Nongbri writes of P52, “I have not provided any third-century documentary papyri that are absolute ‘dead ringers’ for the handwriting of P52, and even if I had done so, that would not force us to date P52 at some exact point in the third century. Paleographic evidence does not work that way” (“Use and Abuse of P52,” 46). 32 E. G. Turner, The Typology of the Early Codex (Philadelphia: University of Pennsylvania Press, 1977), 7. 33 J. W. B. Barns, G. M. Browne, and J. C. Shelton, eds., Nag Hammadi Codices: Greek and Coptic Papyri from the Cartonnage of the Covers, NHS 16 (Leiden: Brill, 1981), 57. Stephen Emmel rightly questions how much later they were made, however, in “The Coptic Gnostic Texts as Witnesses to the Production and Transmission of Gnostic (and Other) Traditions,” in Das Thomasevangelium: Entstehung-Rezeption-Theologie, ed. Jörg Frey, Enno Edzard Popkes, and Jens Schröter, BZNW 157 (Berlin: de Gruyter, 2008), 38. 34 Matthew D. Teasdale et al., “The York Gospels: A 1000-Year Biological Palimpsest,” Royal Society Open Science 4, no. 10 (2017): 7.

Dating Myths, Part One

101

Radiocarbon dating might be a possibility, though it requires the destruction of a small piece of the item dated, and most conservators would probably forbid such destruction. Additionally, scientific testing is often not able to settle historians’ questions. R. E. Taylor and Ofer Bar-Yosef write, “Radiocarbon ‘warps’ create periods throughout the 14C time scale where there are inherent systemic limitations in the precision with which a 14Cbased time segment can be expressed.”35 Like paleography, radiocarbon dating can give only a range of most probable dates, but because of these “warps,” there are periods within the overall timeline where radiocarbon dating results are less useful than others. For example, as Josephine K. Dru shows, “14C science can often distinguish the 130s from the 40s CE. But it cannot distinguish the 130s from the 220s CE—though the time difference is the same!”36 These specifics mean that, if P52 were to be radiocarbon dated, the best we could hope for (if the tests achieved state-of-the-art accuracy and precision) is a better sense of whether its papyrus (the material on which the text was written, not the writing itself) originated before 130, between circa 135–225, or later. Carbon-14 analysis could not show whether 135, 150, 175, 200, or 225 is more probable. Is that worth the cost of a slightly smaller P52? In summary, the process of dating a manuscript often involves a combination of methods and/or sources, and even still, assigned dates must be taken with caution. Turner writes, “Palaeography is neither a science nor an art, but works through a continual interaction of the methods appropriate to both approaches. And in the last resort a judgment has to be made—and judgment [sic] is fallible.”37 P52: DOES IT DATE TO AROUND 125?

C. H. Roberts and the earliest manuscript. In light of the difficulties of paleographic dating and the fallibility of judgments, we turn to the fragment that holds the distinction of being the most commonly cited early manuscript: P52 (P.Ryl. 457). It is a fragment of John’s Gospel about three-and-a-half by 35

R. E. Taylor and Ofer Bar-Yosef, Radiocarbon Dating: An Archaeological Perspective, 2nd ed. (Walnut Creek, CA: Left Coast Press, 2014), 59-60, but see a fuller discussion of evaluating radiocarbon dates at 130-71. 36 Josephine K. Dru, “Radiocarbon Dating for Manuscripts on Papyrus or Parchment: Improving Interpretation Through Interdisciplinary Dialogue,” poster presented at ManuSciences, Villa Clythia, Fréjus, France, September 2017 (emphasis original). 37 Turner, Greek Manuscripts of the Ancient World, 20.

102

Elijah Hixson

two-and-a-half inches in size to which academics and apologists alike frequently appeal as the oldest extant manuscript of the New Testament. When C. H. Roberts first published P52 in 1935, he assigned it a date on the basis of its handwriting: “On the whole we may accept with some confidence the first half of the second century as the period in which P. Ryl. Gr. 457 was most probably written.”38 Roberts arrived at that date by comparing the handwriting of P52 to that of other known papyri.39 It is important to note that Roberts did not say that P52 was written between AD 100 and 150 but that it was most probably written then. A few considerations provide reason for revising Roberts’s early date for P52 (“most probably” AD 100–150). First, Roberts’s two closest matches to the hand of P52 were not themselves securely dated. Second, the securely dated specimens in general were not close matches. Third, there are now many more published manuscripts with which to compare P52 than when Roberts first published it in 1935, such that consensus regarding the paleographic dates can change. In the case of one of the two “close matches”—P.Egerton 2—it did. Roberts compared P52 to an early dated manuscript that is no longer considered to be so early.40 A recent redating of P.Egerton 2 concluded that it dates to circa AD 150–250 and that “it is not impossible that [P.Egerton 2] was produced sometime at the turn of the third century.”41 In summary, Roberts assigned the early date to P52 primarily on the basis of problematic comparisons or manuscripts without secure dates, one of which has recently been dated much later. In light of the many papyri published since 1935, we must now consider whether P52 really is as old as is commonly claimed. Recent rejection of an early date. In 1975, Eric Turner accepted Roberts’s date of circa AD 100–150, but with reservation.42 Turner mentioned P.Amh. 38

C. H. Roberts, An Unpublished Fragment of the Fourth Gospel in the John Rylands Library (Manchester: Manchester University Press, 1935), 16. 39 Roberts, Unpublished Fragment of the Fourth Gospel, 14-16. Brent Nongbri helpfully provides images of each of these with discussion of their relative value for dating P52 in Nongbri, “Use and Abuse of P52.” 40 Nongbri writes, “The problematic nature of paleographically dating thee papyri comes into even sharper relief when we notice that the principle comparanda for dating Egerton Papyrus 2 are for the most part the same as those later used by Roberts to date P52. The independent value of Egerton Papyrus 2 for dating P52 is thus minimal,” in “Use and Abuse of P52,” 34 (emphasis original). 41 Peter Malik and Lorne R. Zelyck, “Reconsidering the Date(s) of the Egerton Gospel,” ZPE 204 (2017): 63. 42 Turner, Typology of the Early Codex, 100.

Dating Myths, Part One

103

II 78 as a manuscript with “similarities”—it is a petition firmly dated to AD 184. In 1989, Andreas Schmidt identified two other manuscripts—still only paleographically dated—that exhibit close similarities to P52. Schmidt then suggested that P52 dates to circa AD 170, plus or minus twenty-five years.43 Sixteen years after Schmidt’s study, Brent Nongbri made an even stronger case for adopting a wider date range for P52. Nongbri began his important article thus: “The thesis of this paper is simple: we as critical readers of the New Testament often use John Rylands Greek Papyrus 3.457, also known as P52, in inappropriate ways, and we should stop doing so.” To be perfectly honest, he isn’t completely wrong. His work demonstrated that most of Roberts’s comparanda (manuscripts with comparable handwriting) are not close matches to P52, and he also introduced several additional securely dated comparanda. Nongbri concluded, “Any serious consideration of the window of possible dates for P52 must include dates in the later second and early third centuries.”44 It is important to note that Nongbri is not suggesting that we abandon the possibility of an early date of P52. Rather, he is arguing that the early date range for P52 should be extended to include later dates.45 More recently, other manuscript specialists have rejected the AD 100–150 date for P52. Don Barker, a papyrologist at Macquarie University, writes, “It is difficult to place [P52] in a very narrow time period,” and he assigns P52 to anywhere in the second or third centuries.46 Barker continues, “This may be unsatisfactory for those who would like to locate [P52] in a narrower time frame but the palaeographical evidence will not allow it.” 47 Christian Askeland cites Nongbri and Barker’s work with approval, condemning “indefensible arguments for ridiculously early dates of various New Testament papyri” in his own article on the difficulties of paleographic dating. 48 43

Andreas Schmidt, “Zwei Anmerkungen zu P. Ryl. III 457,” APF 35 (1989): 11-12. Nongbri, “Use and Abuse of P52,” 23, 46. 45 For example, Nongbri writes, “I think it is safe to say that although a few Christian books may be as old as the second century, none of them must be that old, not even the celebrated fragment of the fourth gospel in the John Rylands Library [i.e., P52],” in God’s Library: The Archaeology of the Earliest Christian Manuscripts (New Haven, CT: Yale University Press, 2018), 269 (emphasis original). God’s Library is an excellent introduction to several important issues regarding early manuscripts. 46 Don Barker, “The Dating of New Testament Papyri,” NTS 57, no. 4 (2011): 574-75. In addition to the studies discussed here, see also Foster, “Bold Claims, Wishful Thinking.” 47 Barker, “Dating of New Testament Papyri,” 575. 48 Askeland, “Dating Early Greek and Coptic Literary Hands,” specifically, 466. 44

104

Elijah Hixson

Whether or not Nongbri and Barker are correct that P52 could be as late as the third century, they are absolutely correct that dates of circa AD 125 or AD 100–150 are too early.49 How useful is P52 to apologetics? First, P52 does not and cannot offer definitive proof that John’s Gospel is a first-century composition by an eyewitness. Even if P52 were written in the afternoon of April 26, AD 125 (it wasn’t), it would prove only that sections from John 18 were in Egypt by AD 125. Technically, such a date does not prove that John’s Gospel was in its “final” (canonical) form by then, nor does it prove that the text it contains is any more than a few months old. An early date of P52 might render these possibilities unlikely—even extremely unlikely—but it cannot disprove them. Two examples from redaction-critical commentaries demonstrate this point. First, Rudolf Bultmann accepted a date of P52 in the period of AD 100–150 and still argued that as much as forty years could have passed between the original writing of John’s Gospel and a final redaction that left it in the canonical form we have today. Second, Walter Schmithals was well aware of the existence of P52, but he still dated a final redaction of John’s Gospel to around AD 160–180.50 Given the uncertain nature of paleographical dating and the fact that P52 has not deterred source-critical scholars from adopting second-century dates of a final redaction to John’s Gospel, we quote again Paul Foster’s remarks about the usefulness of P52: “Was John’s Gospel written before the end of the first century? Yes, probably. Does P52 prove this to be the case? No, probably not.”51 Second, the occasional absence of P52 in critical editions of the Greek New Testament shows why it is not necessary for reconstructing the original text. The UBS3 (published in 1975) makes no mention of P52. It is included in the index of manuscripts to the UBS4 and UBS5, but neither edition cites P52 for or against any variant in the text. The NA28 contains eleven variation units for John 18:31-33, 37-38 (the verses for which P52 is extant for at least part), but because the papyrus is so fragmentary, the NA28 cites P52 only 49

Pasquale Orsini and Willy Clarysse assign P52 a date of circa 125–175 in their important essay “Early New Testament Manuscripts and Their Dates,” 470. 50 Rudolf Bultmann, The Gospel of John: A Commentary, trans. G. R. Beasley-Murray, R. W. N. Hoare, and J. K. Riches (Philadelphia: Westminster, 1971), 12; Walter Schmithals, Johannesevangelium und Johannesbriefe: Forschungsgeschichte und Analyse, BZNW 64 (Berlin: de Gruyter, 1992), 242. See also his discussion of the date of P52 on pp. 9-11. 51 Foster, “Bold Claims, Wishful Thinking,” 204.

Dating Myths, Part One

105

once. The one instance is a variant more or less with respect to word order at John 18:33. The reading of P52 was adopted as the text of John’s Gospel against the reading of the majority of manuscripts, but Westcott and Hort had already adopted this reading as the text of John decades before P52 was published. Likewise, P52 makes no appearance in the apparatus of the recent Greek New Testament Produced at Tyndale House Cambridge, an edition that emphasizes its dependence and use of early manuscripts and scribal habits not only for the text but even the spelling and paragraphing of the New Testament.52 Third, some scholars who are neither trained papyrologists nor paleographers have proposed unusually early or narrow dates for P52, and these dates should not be accepted. Karl Jaroš (AD 80–125), Philip Comfort (AD 110–125), and Carsten Peter Thiede (AD 80–130) are each controversial for their early dates, which have failed to gain scholarly acceptance.53 Pasquale Orsini and Willy Clarysse criticize Jaroš and Comfort (specifically, the edition of New Testament papyri he cowrote with David Barrett) for their early dates, and they dismiss Thiede altogether.54 With respect to the date of P52, Orsini and Clarysse note that the comparanda used by Comfort-Barrett and Jaroš are inappropriate because they are not even the same “style” of handwriting as P52. They show that the early dates proposed by Jaroš and Comfort-Barrett are methodologically unsound, concluding, “Biblical scholars should realise that some of the dates proposed by some of their colleagues are not acceptable to Greek palaeographers and papyrologists.”55 52

Dirk Jongkind et al., eds., The Greek New Testament, Produced at Tyndale House, Cambridge (Wheaton, IL: Crossway, 2017), 505-23. 53 Karl Jaroš, Die ältesten griechischen Handschriften des Neuen Testaments (Cologne: Böhlau, 2014), 71; Comfort, Commentary on the Manuscripts and Text, 65-66; Carsten Peter Thiede, The Earliest Gospel Manuscript? The Qumran Fragment 7Q5 and Its Significance for New Testament Studies (Carlisle, UK: Paternoster, 1992), 21-22. 54 Orsini and Clarysse, “Early New Testament Manuscripts and Their Dates.” On Thiede’s claim that P64+67 could date to the late first century, “perhaps (though not necessarily) predating A.D. 70” (Carsten Peter Thiede, “Papyrus Magdalen Greek 17 [Gregory-Aland P64]: A Reappraisal,” TynBul 46, no. 1 [1995]: 38; Carsten Peter Thiede and Matthew D’Ancona, The Jesus Papyrus [London: Weidenfeld & Nicolson, 1996]), Klaus Wachtel writes, “Thiedes Argumentation für eine Datierung des P64/67 ins 1. Jahrhundert ist jedoch als methodisch unzulänglich und sachlich falsch zurückzuweisen” (“Thiede’s argument for a dating of P64/67 in the first century, however, is rejected as methodologically insufficient and factually wrong”), in “P64/67: Fragmente des Matthäusevangeliums aus dem 1. Jahrhundert?,” ZPE 107 (1995): 80. For Comfort and Barrett’s edition of New Testament papyri, see Philip W. Comfort and David Barrett, The Text of the Earliest New Testament Greek Manuscripts, corrected, enl. ed. (Wheaton, IL: Tyndale House, 2001). 55 Orsini and Clarysse, “Early New Testament Manuscripts and Their Dates,” 462, 466.

106

Elijah Hixson

Christian Askeland rightly rejects the conclusions of Jaroš, Comfort, and Thiede, describing their respective works under the heading “Paleography Gone Wrong.”56 Craig A. Evans describes “assertions of very early dates for some papyri” by a few scholars including Thiede and Comfort as “especially problematic.”57 Many recent scholars who mention the date of P52 give it a mid-second century or late-second century date.58 As it happens, P52 is not the only New Testament manuscript to receive such controversial treatment from these authors and others. In his Commentary on the Text and Manuscripts of the New Testament, Comfort cites “manuscripts with certain dating” in defense of his circa 200 date for P45, but none of the manuscripts he cites are dated securely. Two are reused rolls with dates on one side (but not the side relevant to P45), and the other is a paleographic date of a cursive hand—admittedly easier to date than a literary hand but by no means certain. Comfort also mentions E. C. Colwell’s famous study of scribal habits, including Colwell’s conclusion that the scribe of P75 copied “letter by letter,” but he seems unaware of Klaus Junack’s study from 1981 criticizing Colwell’s conclusions about “letter by letter” copying or Dirk Jongkind’s more recent article that touches on the same issue.59 Comfort cites 56

Askeland, “Dating Early Greek and Coptic Literary Hands,” 464-65. Craig A. Evans, “Christian Demographics and the Dates of Early New Testament Papyri,” in The Language and Literature of the New Testament: Essays in Honor of Stanley E. Porter’s 60th Birthday, ed. Lois K. Fuller Dow, Craig A. Evans, and Andrew W. Pitts, Biblical Interpretation Series 150 (Leiden: Brill, 2016), 201n3. 58 Mid-second century: David C. Parker, An Introduction to the New Testament Manuscripts and Their Texts (Cambridge: Cambridge University Press, 2008), 234; Juan Chapa, “The Early Text of John,” in Hill and Kruger, Early Text of the New Testament, 141; Orsini and Clarysse, “Early New Testament Manuscripts and Their Dates,” 460; Alan Mugridge, Copying Early Christian Texts: A Study of Scribal Practice, WUNT 362 (Tübingen: Mohr Siebeck, 2016), 250-51. Late second century: Charles E. Hill, “Did the Scribe of P52 Use the Nomina Sacra? Another Look,” NTS 48, no. 4 (2002): 592; Roger S. Bagnall, Early Christian Books in Egypt (Princeton, NJ: Princeton University Press, 2009), 89. Lonnie D. Bell allows any date in the second century, in The Early Textual Transmission of John: Stability and Fluidity in Its Second and Third Century Greek Manuscripts, NTTSD 54 (Leiden: Brill 2018), 38. Barker allows a date for P52 anywhere in the second or third centuries, in “Dating of New Testament Papyri,” 575. 59 Comfort, Commentary on the Manuscripts and Text, 60-61. See also E. C. Colwell, “Method in Evaluating Scribal Habits: A Study of P45, P66, P25,” in Studies in Methodology in Textual Criticism of the New Testament, ed. Bruce M. Metzger, NTTS 9 (Grand Rapids: Eerdmans, 1969), 106-24; Klaus Junack, “Abschreibpraktiken und Schreibergewohnheiten in ihrer Auswirkung auf die Textüberlieferung,” in New Testament Textual Criticism: Its Significance for Exegesis; Essays in Honor of Bruce M. Metzger, ed. Eldon J. Epp and Gordon D. Fee (Oxford: Clarendon, 1981), 277-95; Dirk Jongkind, “Singular Readings in Sinaiticus: The Possible, the Impossible, and the Nature of Copying,” in Textual Variation: Theological and Social Tendencies? Papers from the Fifth Birmingham Colloquium on the Textual Criticism of the New 57

Dating Myths, Part One

107

H. J. M. Milne and T. C. Skeat for having “demonstrated that Scribe A of Codex Vaticanus was likely the same scribe as Scribe D of Codex Sinaiticus,” but this statement is simply untrue.60 Milne and Skeat came to the opposite conclusion. After a discussion of the similarities between the hands, they write, “It would be hazardous to argue identity of the two hands.”61 Comfort even appeals to Guglielmo Cavallo for a mid-second-century date of P66.62 Cavallo did accept that date in 1967, but he changed his mind in favor of a later date by 1975, so Comfort’s appeal to Cavallo in the present tense is misleading.63 In general, one gets the impression that Comfort gives emphasis to references that could be used to support his controversial conclusions, and even then he does not always represent them accurately. CONCLUDING SUGGESTIONS

We do have early manuscripts of the New Testament, and apologists are right to appeal to them. Even if our extant witnesses are not quite as early as we once thought, the number of early manuscripts of Christian Scriptures is a testimony to their importance to early Christians. That we can even identify tiny fragments as New Testament manuscripts by the text they contain is a testimony to the macrostability of the New Testament text. Below are a few practical suggestions for identifying and reporting the dates of early New Testament manuscripts responsibly. • Always use a full date range for a manuscript rather than the midpoint or early end of a date range. A date of “circa AD 200” might imply either a range of 175–225 or a range of 150–250, and it does not convey the fact that a manuscript dated to 150–250 is just as likely to be written Testament, ed. H. A. G. Houghton and D. C. Parker, TS, Third Series 6 (Piscataway, NJ: Gorgias, 2008), 35-54. 60 Comfort, Commentary on the Manuscripts and Text, 94. 61 H. J. M. Milne and T. C. Skeat, Scribes and Correctors of the Codex Sinaiticus (London: British Museum Press, 1938), 90. 62 Comfort, Commentary on the Manuscripts and Text, 70. Michael J. Kruger also accepts a secondcentury date for P66, in Canon Revisited: Establishing the Origins and Authority of the New Testament Books (Wheaton, IL: Crossway, 2012), 235. Kruger appeals to Herbert Hunger’s controversial dating, though more recent scholarship has rejected Hunger’s second-century proposal. Herbert Hunger, “Zur Datierung des Papyrus Bodmer II (P66),” AÖAW 4 (1960): 12-33. See, however, Turner, Greek Manuscripts of the Ancient World, 108; Nongbri, “Limits of Palaeographic Dating of Literary Papyri,” 9-13; Orsini and Clarysse, “Early New Testament Manuscripts and Their Dates,” 470. 63 On Cavallo’s change of mind, see Nongbri, “Limits of Palaeographic Dating of Literary Papyri,” 13.

108

Elijah Hixson

in 248 as it is likely to be written in 156, or anywhere in between. We simply cannot be more precise.64 • Avoid sensational dates and excessively early or narrow date ranges. If it sounds too good to be true, assume that it is. Remember Eric Turner’s assessment that for literary hands (which applies to almost all early New Testament manuscripts), fifty years is the smallest acceptable window for a date range assigned on paleographic grounds. A date in the first century for any New Testament manuscript is an immediate red flag—especially for an unpublished manuscript. • Accept the full range of the date given by INTF. This information can be found in the back of NA28 (pp. 792-819) or via the electronic version of their official register of Greek New Testament manuscripts.65 The electronic availability of the Liste ensures that their information is freely accessible with nothing more than an internet connection, and as they are the keepers of the official register of New Testament manuscripts, no one can fault a nonspecialist for trusting their judgment. INTF dates P52 in the second century. The online Liste assigns P52 to the mid-second century (125–175), but NA28 gives the full second century as the possible date. I suggest adopting the broader range.66 For P66 and P75, the online Liste has dated them to the early third century, with dates of AD 200–225. NA28 reports “c. 200” for each manuscript. Thus, in these instances, I suggest accepting the broader range of AD 200–225. Turner’s objections to a range of fewer than fifty years notwithstanding, again—no one can fault a nonspecialist for accepting the date range adopted by INTF.67 64

At this point I must confess my own sins. In 2015, I gave the dates of P46 and P75 as “Around A.D. 200” in a chapter I coauthored with Timothy Paul Jones, “How Was the New Testament Copied?,” in How We Got the Bible, by Timothy Paul Jones (Torrance, CA: Rose, 2015), 120. If I could do it over again, I would give the date ranges of AD 200–300 for P75 and AD 175–250 for P46. 65 “Liste,” Institut für Neutestamentliche Textforschung, http://ntvmr.uni-muenster.de/liste (accessed January 12, 2018). In the “Manuscript Num.” box with the dropdown menu on “GA,” type the Gregory-Aland number for a manuscript (e.g., P52, 023, 1739, or l1747) to search for it. If the dropdown menu is set to “ID,” manuscripts are identified slightly differently. Each manuscript has a five-digit code beginning with 1, 2, 3, or 4, for papyri, majuscules, minuscules, and lectionaries, respectively. The initial number is followed by the numerical portion of the GA number, preceded by zeros if necessary to fill out five digits. Thus, for the examples given above, the IDs would be 10052, 20023, 31739, and 41747. 66 As of January 12, 2018, for P52 as well as P66 and P75 below. 67 Brent Nongbri has suggested that P66 and P75 could possibly be as late as the fourth century.

109

Dating Myths, Part One

• Rather than appealing to P52 to rule out a second-century composition of John, appeal to other factors such as its reception in the second century to argue for its authenticity. This approach is admittedly more complex, but Charles E. Hill has written a thorough discussion of its second-century reception that would be helpful in formulating such an argument.68 Finally, remember that Christianity is a faith that stood unshaken for centuries while the earliest copies of its texts lay quietly buried in the sands of Egypt. People became Christians long before P52 was discovered, and they will continue to do so even if it is not quite as ancient as was previously thought.69

Key Takeaways uu Often a manuscript can be dated only by paleography, which is a difficult and imprecise way of assigning the date by means of an assessment of the handwriting. uu It is almost always unwise to assign a date range of fewer than fifty years on the basis of paleography; a range of seventy-five to one hundred years is typically more preferable. uu The middle year of a date range is no more likely (but also no less likely) to be the “actual date” of a manuscript than any other date in that range. Always try to give the full range; do not assume the earliest date is the right date. uu The responsible date range of P52, probably our earliest New Testament manuscript, is AD 100–200, and a few scholars even extend this range into the 200s.

See Nongbri, “Limits of Palaeographic Dating of Literary Papyri”; Nongbri, “Reconsidering the Place of Papyrus Bodmer XIV-XV (P75) in the Textual Criticism of the New Testament,” JBL 135, no. 2 (2016): 405-37. If researchers at the INTF are persuaded by his arguments, perhaps they will change their entry for these manuscripts. 68 Charles E. Hill, The Johannine Corpus in the Early Church (Oxford: Oxford University Press, 2004). 69 Several individuals read early drafts of this chapter, and I am especially thankful to Brent Nongbri, Josephine Dru, and Grant Edwards for their helpful feedback. Remaining errors are, of course, my own.

CHAPTER SIX
 
Last edited:
Top