Tuesday, February 06, 2018

Palaeographic Method, Comparison and Dating: Considerations for an Updated Discussion (Guest post by Pasquale Orsini)

As some of you may remember, Brent Nongbri and I have recently had a nice back and forth about palaeographical method. It started with my post on the Egerton Gospel, where I mentioned some hesitations about Brent’s recent suggestions concerning P66 and P75, to which Brent responded here and then, followed by my comment, here). Now, I’m very pleased to report that Pasquale Orsini, one of the great palaeographers of the recent generation, decided to chip into our discussion and I’m very honoured to include his guest post on our blog. [Update: I should also like to bring to your attention Pasquale's forthcoming monograph Studies on Greek and Coptic Majuscule Scripts and Books (Studies in Manuscript Cultures 15; Berlin/New York: De Gruyter, 2018).]

Rome, 6 February 2018

I want to contribute briefly to the debate between Peter Malik and Brent Nongbri on some principles of the paleographic method, without repeating opinions and concepts already written.

First of all, there is no ‘Italian’ method of paleography, but a paleographic method. This method is based on comparison (a paleographer once said that paleography is the ‘science of comparison’) of the graphic structure of scripts before that of single letters. However, in the last decades some paleographers (Guglielmo Cavallo ‘in primis’, Edoardo Crisci and myself) have questioned some principles of the paleographic method (for example, the concept of ‘canon’, its standard distribution in development / perfection / decline, etc.), recognizing the need for a reflection that is historically closer to the available data. These studies further develop the paleographic method of the 60s of the last century. And these studies should be read, even if they have been written in Italian language (with all the difficulties of this language).

Stylistic problems and ‘appropriate’ comparisons

P66 (P. Bodmer II)

The article of Don Barker (‘The Dating of New Testament Papyri’, NTS 57 (2011), 578–582) attributes the script of this manuscript to the graphic stream with angular formation of some letters (delta, ypsilon, phi) and serifs, and not to the ‘Alexandrian stylistic class’. According to this assignment, he proposes a dating between the middle of the second and the middle of the third century. As comparisons he proposes P. Oxy. 1622 (a fragment of a roll of Thucydides, assigned by Grenfell and Hunt to the ‘early second century” because the reverse was used for a contract dated 148 AD) and P. Oxy. 3030 (an official letter dated 207 AD).

I mentioned Barker’s article just to explain that in P. Bodmer II there is the roundness of the letters and not the angularity; there is no emphasis on the upper notional line; serifs or blobs on the ends of the strokes are characteristic of the ‘Alexandrian stylistic class” and then of the ‘Alexandrian majuscule”.

Brent Nongbri, in a more articulated work (‘The Limits of Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P. Bodmer II (P66)’, Museum Helveticum 71 (2014), 1–35), reviews the various proposals of dating, analyzing the comparisons with manuscripts dated on basis of paleographic method. He does not find these comparisons satisfactory and suggests three new manuscripts: P. Bodmer XX, P. Cairo Isid. 2 and P. Lond. 1920.

P. Bodmer XX—part of the famous Miscellaneous Codex, which includes P. Bodmer V, X, XI, VII, XIII, XII, XX, IX, VIII—contains the ‘Apology of Phileas”: Phileas, the bishop of Thmouis in the Delta was martyred in the year 305, which gives a clear terminus post quem for the manuscript. P. Bodmer XX is therefore dated around the middle or in second half of the fourth century, when his apology had developed into a literary work. Nongbri says: ‘P. Bodmer 2 and P. Bodmer 20 show a number of compelling similarities in spacing, letter forms and overall appearance”, and soon after: “the two hands are noticeably similar in a number of ways.’ For Nongbri the chief (but insignificant) difference between these two hands is ‘the presence of serifs, or blobs, at the end of certain strokes’ in P. Bodmer II.

These two Bodmer papyri belong, however, to two different graphic typologies: the scribe of P. Bodmer XX used a ‘mixed style”, with elements of Biblical majuscule (alpha and ypsilon; contrast between fine and thick strokes) and of Alexandrian stylistic class (delta, epsilon, kappa, lambda, omega); the scribe of P. Bodmer II used an Alexandrian stylistic class (see particularly alpha, my, ypsilon). These different scripts cannot be compared for dating, even if they are contemporary.

More ‘appropriate’ is the comparison with P. Cairo Isid. 2, a letter from the archive of Aurelius Isidorus written in AD 298, and with P. Lond. 1920, a letter from a Greco-Coptic dossier of the monastery of Phathor dated about 330–340. Both scripts belong to the Alexandrian stylistic class and have elements in common with P. Bodmer II. These comparisons are convincing for me, and for this reason I accept a date between third and fourth centuries (expanding the chronological terms of previous article written together with Willy Clarysse, ‘Early New Testament Manuscripts and Their Dates. A Critique of Theological Palaeography’, ETL 88/4 [2012], pp. 443–474: 465, 470; for P. Bodmer XX and P. Bodmer II see also P. Orsini, ‘I papiri Bodmer: scritture e libri’, Adamantius 21 [2015], 60–78: 61 n. 5, 63–64, 77 [Tab. 2]).

P75 (P. Bodmer XIV-XV)

P75 was written by a single scribe in ‘severe style’. The dates proposed for this codex vary from the late second to the second half of the third century (see B. Nongbri, ‘Reconsidering the Place of Papyrus Bodmer XIV-XV (75) in the Textual Criticism of the New Testament’, JBL 135 (2016), 405–437). Though comparisons with the manuscripts of the third century can indeed be made (P. Oxy. 1012 [205–250 AD; on the recto there is a document, P. Oxy. 1045, ca. 205 AD]; P. Oxy. 1016 [235–299 AD; on the recto there is a document, P. Oxy. 1044, ca. 234–235 AD]), a persuasive comparison, proposed by Nongbri, is also possible with P. Herm. 4 (about 317–323 AD) and P. Herm. 5 (about 325 AD), from the archive of Theophanes. For this reason I agree to extend the chronological terms, including the early fourth century (see Orsini, ‘I papiri Bodmer’, 77 [Tab. 2]).

The basic misunderstanding is to consider the script of the two Herm. papyri as belonging to ‘sloping pointed majuscule’: instead it belongs to the ‘severe style’. I shall not repeat here the characteristics of the ‘sloping pointed majuscule’ and its chronological problems (see my article: ‘La maiuscola ogivale inclinata. Contributo preliminare’, Scripta 9 [2016] 89–116), but the ‘sloping pointed majuscule’ develops through a transformation of the ‘severe style’ (Grenfell and Hunt, Schubart, Lameere and finally Cavallo have stated this), between the end of the IV and the beginning of the V century. The two Herm. papyri belong to a process of initial transformation of the ‘severe style’, but they do not yet present all the characteristics of the ‘sloping pointed majuscule’.

Egerton Gospel (P. Egerton 2 + P. Köln VI 255)

I agree with the dating proposed by Malik (150–250 AD; see P. Malik and L.E. Zelyck, ‘Reconsidering the Date(s) of the Egerton Gospel’, ZPE 204 [2017], 55–71: 65) for this ‘Alexandrian stylistic class’. Furthermore, I add—as ‘impression d’ensemble’—to the comparisons proposed by him P. Oxy. 4625 (AD 200–299) and P. Oxy. 3313 (AD 100–199), although these have some differences in detail.


  1. Thank you very much Paquale! May I ask you to explain the reason for the very narrow time frame for the date of P4+P64+P67 (in Paris, Oxford and Montserrat), which you and Claryssee date to 175-200 CE in the 2012 article you mentioned (Nestle-Aland 28 indicates 200-250 CE). What is the relevant comparanda?

    1. This script (Biblical majuscule) belongs to the initial phase of this "normative majuscule", and I have proposed (P. Orsini, "Manoscritti in maiuscola biblica", Cassino 2005, 85-86) as comparisons P. Vindob. G 29768 and P. Vindob. G 29784, both generally attributed to the end of the II century. "AD 175-200" is the translation (certainly, too precise) in numbers of the expression "end of the II century".

    2. Thank you, so if you were to rewrite the 2012 article, what interval would you give? 150-200 CE or something?

    3. For now, I maintain the formula "AD 175-200", which best expresses the concept of the final part of the II century, even if it may seem too "narrow" chronologically.

  2. Pasquale Orsini, thank you for your helpful blogpost. I especially appreciate your comment “that paleography is the ‘science of comparison’” as I have received the impression at SBL that there is a growing skepticism towards palaeography as a truly objective discipline. In this I wanted to ask the relevance of the comparison of the hands of P.Bodmer II with that of P. Cairo Isid. 2, and P. Lond. 1920. In your article, “Early New Testament Manuscripts and Their Dates” page 448, note 23, you and Clarysse define a “stylistic class”;

    “A “stylistic class is a set of writings sharing a general framework, form and structure (in number, sequence and direction of strokes) of some (but not necessarily all) letters; moreover, they may contain graphic variants of the same letter. The term “stylistic class” attempts to recognize a distinctive writing with no rigid and fixed rules.”

    I understand that slight differences in letter formation are normal within a stylistic class. However, at what point do differences in letter formation become too great for adequate comparison? For example, the deltas of P.Lond. 1920 are angular with the right and left obliques meeting at the top and crossing each other to form an X at the top. This is quite different than the deltas of P.Bodmer II which are much more looped and flattened, with only the right hand oblique extending to the left. There are marked differences in the other letter forms as well. For example, in P. Lond. 1920 there are “cross” shaped psi’s while the arms of the psi’s in P. Bodmer II curve upward. There are differences in the alphas as well. With P. Cairo. Isid. 2., the loop of the alpha is more squashed with an angular tip of the loop when compared to the alphas of P. Bodmer II and the phi of P. Cairo. Isid. 2. is slightly diamond shaped and more flattened. Then there is the near absence of serifs in both P. Lond. 1920 and P. Cairo. Isid. 2. When compared to P. Bodmer II.
    I understand that a blogpost is not the ideal place for a discussion such as this, so feel free not to respond.
    Thanks again for your excellent post!

    1. In a "stylistic class" the differences in structure of some letters are common, because the graphic phenomenon has not been "normalized" according to a scheme (as it happens in a "canon" or "normative majuscule"). The differences observed by you are part of this variability of a "stylistic class". The elements of the style (for example round shapes with loops, oblique and horizontal strokes prolonged, some letters written in a single sequence) prevail over the structure of the individual letters.
      This is a topic too complex to be treated in a short answer on a blog.
      This problem is at the base of the "comparison" process. When someone says that scripts belonging to different styles can be compared (for the purposes of dating), I think it is a mistake: comparison is possible if there are elements of style, some structures and forms in common. Otherwise what do we compare?

    2. This is very insightful, Pasquale. Thanks! Will you address these issues more fully in your forthcoming monograph?

    3. Yes. Although my book is the re-publication of previous works, I wanted to write an introduction to discuss a series of open problems, like this one.

    4. Thank you very much for your response, it is helpful. I understand, once again that a blog post is not the best place for a discussion of this detail, so I appreciate your reply. Blessings to you in your work.

  3. Pasquale Orsini,
    Does this mean that for you 175-200 is more the centre of a date range than the full possible date range itself?

    1. For this ancient phase of Biblical majuscule we have a witness with a chronological element: P. Ryl. I 16, terminus ante quem AD 253 or 256 (on the verso there is a letter from Siro to Heronino, P. Ryl. II 236, TM 12980). The text on the recto was written before the letter on the verso: I think that, for caution, we can assume the first half of the third century as a possible date of P. Ryl. I 16. Well, the script of P4+P64+P67 seems to me to be earlier than that of P. Ryl. I 16. An element seems to me an index of anteriority: the shading (result of a particular writing angle) in P4 + P64 + P67 has slight contrast and is not regular (it is regular in Biblical majuscule when it is approx. 75°). The main reason for my dating (end of the second century) is this. Therefore, the two chronological extremes (175-200) are obviously not impassable borders, but new paleographic elements must appear to re-discuss the history of the biblical majuscule.

  4. The curators of the Voynich Manuscript surrendered four slivers of marginal parchment to Carbon-14 spectrometric dating, and they yielded a very precise date in the middle of the first quarter of the 15th century, thus putting to rest several theories as to its authorship.
    There is no scientific reason why any number of ancient biblical manuscripts could not be subjected to the same analysis, thus putting to rest numerous disputes over dating. Why, even the Modern Sinaiticus Theory could possbly be put to rest.

    1. C14 wouldn't necessarily give us a tighter or more accurate date range.

    2. Meaning, professional paleographers whose reputations are at stake would refuse to accept any results outside a predetermined acceptable range?

    3. Many of them would not; rather they would definitely appreciate greater precision (even if that meant their proposed dates were proven wrong). Most of us want to know the truth, at the end of the day. But C14 involves limitations and flexibilities of interpretation as well and it doesn't necessarily follow that C14 will disprove or replace palaeographical method.

    4. Daniel, there is no reason to get insulting. If my brief comments above were offensive, I apologize for the clipped tone, I was writing quickly I did not mean to come across insulting.
      I am definitely willing to be wrong. I will admit being wrong right now on this blog. I will admit that I forgot that P.Lond.1920 is written in Coptic not Greek. The differences in the "deltas" I was describing are actually two different letters. So that explains some of the differences in structure. There, I admitted I was wrong, that wasn't so hard!
      There are three problems with C14 dating (when compared with palaeography) that make it nearly useless in dating when a manuscript was written.
      1) C14 can and often gives a date range spanning 100-200 years, sometimes even greater than this. Palaeography on the other hand can be as precise as 100 years or so depending on the script, etc.
      2) C14 destroys the piece that it is testing. For large manuscripts such as P. Bodmer II this may not be a that big of a problem. But for something like P104 or P52, C14 is out of the question because most of the manuscript would be destroyed.
      3) Finally, in my opinion the biggest problem with C14 is that it would reveal only the age of the papyrus or parchment and would tell almost nothing as to when the piece was actually written upon. Papyrus and parchment can last for hundreds of years and be reused several times over. The "Gospel of Jesus' Wife" fragment is a good example of this. But there are many examples from antiquity of a manuscript being reused several generations after they were first written upon.
      I hope that better explains my brief comments above.
      Blessings to you.

    5. Thank you Timothy, for explaining that so graciously. I think the size problem has been taken care of now; I mean, look at GOJW--the size of a business card, with no margins anywhere, seemingly purposely so to prevent carbon-dating, and it was still carbon-dated to medieval times (i.e. too late to fit a scenario of being authentic). Granted, the first test was way off because the sample was too small to allow a full testing regimen--but that is an extreme case, and the second lab was able to work around that limitation, even with the first sample having been taken. Your other reasons, however, do make your point. I apologize for the cynicism in my response: Carbon-14 dating seems to be capable of a lot more precision than those whose theories such precision may overturn are willing to give it credit for, and as a scientist that frustrates me. But as a gentleman I shouldn't have let that frustration color my response.

    6. No worries Daniel, thanks for your comments. I appreciate your reiterating the C14 of the GJW fragment. My understanding of C14 is obviously dated (see what I did there? ;-) ) because I thought larger portions of material was required)

  5. If anyone is intetested, I put some images of the letters from the dicferent papyri here;