More On Dating New Testament Manuscripts (and the Rylands Fragment P52 again)

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

In my previous post I addressed the question of the famous P52 manuscript. But the article by Pasquale Orsini and Willy Clarysse is more generally a critique of “theological palaeography” and I highlight here some of their other more points about the principles involved with the dating of manuscripts.

The page references are from Pasquale Orsini & Willy Clarysse, “Early New Testament Manuscripts and Their Dates: A Critique of Theological Palaeography,” Ephemerides Theologicae Lovanienses 88 (2012): 443-74. (In the extracts I am responsible for all bolding of text and formatting that goes beyond normal paragraphing.)

1. The Problem of Dating Literary Papyri

Only a few literary papyri can be dated thanks to

  • circumstantial evidence — i.e. their archaeological or historical context
  • or when they belong to a dated archive
  • or when they are written on the back or front of documentary text (this can give a terminus post — if written on the back; or an ante quem — if a dated document is written on their verso/back)

Other manuscripts (the majority) are thus dated by comparing their handwriting to datable scripts. This gives a relative, not absolute, date for most.

2. New Testament Texts and their Dates

New Testament manuscripts are more problematic than other literary texts since they are nearly always written as part of a codex. This means that the script is the same on both sides of each page and neither side can be used to establish a terminus ante or post quem.

Gradually, however, an uneasy consensus has been reached among papyrologists, and the result of this is found in the dates put forward by Nestle-Aland.

NESTLE–ALAND, 1994 = K. ALAND, Kurzgefasste Liste der griechischen Hand-schriften des Neuen Testaments. Zweite, neugearbeitete und ergänzte Auflage, bearbeitet von K. ALAND, in Verbindung mit M. WELTE, B. KÖSTER und K. JUNACK (Arbeiten zur Neutestamentlichen Textforschung 1), Berlin – New York, 1994;

see updates in: http://intf.uni-muenster.de/vmr/NTVMR/ListeHandschriften.php

As I cited in my previous post, no NT manuscripts are dated to the first century and “only very few to the second century.”

Recently even these early dates have been called into question by R.S. Bagnall [see R.S. BAGNALL, Early Christian Books in Egypt, Princeton, NJ – Oxford, 2009, pp. 11-18.]

Stepping outside of the Orsini-Clarysse article for a moment, here are three online reviews of Bagnall’s book:

1. By Paul Foster, University of Edinburgh [PDF file]:

In the first chapter ‘the Dating of the Earliest Christian Books in Egypt’, Bagnall advances the following proposition: ‘the narrowness of much [scholarship] has permitted its practitioners to reach conclusions that I believe are profoundly at odds with fundamental social realities of the ancient world and with basic probability’ (p. 1). . . .

More on the early dating of P52


Ultimately Bagnall makes two important claims in this book that require consideration. First, there has been a tendency to date Christian papyri too early. And, secondly, there is little evidence for the emergence or visibility of Christianity in Egypt before the end of the second century. Both of these points deserve close reflection. It is surprising that Bagnall did not attack the early dating of P52 more severely, especially given the fact that it was dated in relation to P.Egerton 2.
The publication of P.Köln 255, which is a continuation of one of the pages of P.Egerton 2, has resulted in a revision of the dating of P.Egerton 2. This is due to P.Köln 255 containing a punctuation mark not evidenced prior to the end of the second century. Hence, instead of the initial proposal of dating P.Egerton 2 to around 130, it is now seen as being written sometime closer to 200. However, despite the fact that the hands of P52 and P.Egerton 2 were seen as similar, and hence P52 was dated to 125 on this basis, there has been no attempt to move the dating of P52 to around the end of the second century. Such an insight would have strengthened Bagnall’s case of the ideologically driven nature of the dating of many New Testament papyri. The case for the spread of Christianity in Egypt beyond the confines of Alexandria as occurring in the third rather than the second century may also be a corrective to prevailing theories. . . .

2. By Larry Hurtado, University of Edinburgh [PDF file].

See comment by mcduff on previous post.

3. By Benjamin Garstad, Grant MacEwan University

Roger Bagnall is one of the foremost authorities on the written remains of Roman Egypt and the evidence they offer us for a lost world . . . .

More on the claims that the Codex signalled something special about Christianity

English: Folio 339 verso and 340 recto of the ...
Codex (Photo credit: Wikipedia)

The fourth and final chapter offers an accessible survey of the evidence for the replacement of the book roll with the codex in Egypt. Scholars and interested laymen alike have traditionally held the assumption that the codex form was especially popular amongst Christians and that the eventual predominance of the Church was somehow linked with the eventual prevalence of the codex. Bagnall shows that this cannot really be demonstrated from a statistical analysis of the data. The use of the codex by Christians seems to be proportionally consistent with its use over time in other circles, with the exception of devotees of classical Greek literature. Exploiting the evidence of the so-called Theban Magical Library he suggests, moreover, that the influence of Christianity cannot account for the rise of the codex. What is remarkable is the regular preference of Christians for the codex for the copying of scriptural texts. With very few exceptions the books of the Bible, as well as extra-canonical books which Bagnall describes as “also rans”, are preserved in codices. Bagnall dismisses some of the current explanations of this phenomenon in favour of what seems a compelling solution to the problem. The codex probably comes from Rome, originating with a number of tablets strung together. Bagnall suggests that we consider the spread of the codex an example of Romanization in the eastern Mediterranean world. The resistance to this thesis, he proposes, stems from a desire to see Christianity as a counter-cultural resistance movement, deeply antipathetic to Roman imperial power. The idea of the Christians exploiting a distinctly Roman artefact to preserve their Scriptures does not sit well with such a presumption.

And on (what should be) the obvious

And here we return to one of the themes that unites Bagnall’s remarkably wide-ranging little book. To what extent have researchers’ often unstated assumptions, or perhaps aspirations, their deeply held sympathies and loyalties determined the course and the findings of the study of ancient Christian books? This recurring question and Bagnall’s astute answers make these essays more than a survey or a ‘state of the scholarship’. They offer a healthy corrective to the potential contamination of research by preconceived ideas and ideological commitments, and not only in the author’s chosen subject, but in our field generally.

The Leuven Database of Ancient Books (LDAB) was used by Bagnall. The current entry for P52 in the LDAB assigns the following date range:


Grenfell and Hunt have been criticized for dating the codices too late, but the criticisms against their late dating have been “refuted by B. Nongbri:

B. NONGBRI, Grenfell and Hunt on the Dates of Early Christian Codices: Setting the Record Straight, in Bulletin of the American Society of Papyrologists 48 (2011) 149-162.

Enter the New Testament scholars

Recently this consensus has been put to the test by some New Testament scholars, who argue for an earlier date for most of our oldest New Testament papyri and propose a first century date for several of these. The first to do so was C.P. Thiede in a series of articles and books dealing with P64+67+4, of which fragments are kept in Magdalene College Oxford, Paris and Montserrat. J. O’Callaghan and Thiede even found Gospel fragments among the Greek papyri from Qumran (LDAB 6820 and 7341), which were certainly written before the suppression of the Jewish revolt in AD 69. Papyrologists have generally reacted negatively to these identifications and these scraps are not taken into account of here. (p. 444)

The more one reads of scholarly works the more one learns not to be overwhelmed by pages of footnotes and detail. Details, however copious, still need to be evaluated. Scholarly publications that appear to be so thorough and authoritative need to be evaluated on the basis of the evidence actually supplied for their arguments. (This lesson ought to be applied also to several publications on the archaeology of Nazareth.) So here is how the attempt to date New Testament manuscripts early (against the consensus of other papyrologists) made its entrance:

Two publications, offering a well-documented survey of the oldest New Testament Greek manuscripts (attributed to the first three centuries of Christianity), have recently broadened the attack on the traditional dating to all early New Testament manuscripts. Comfort–Barrett, in 1999 and in the enlarged edition of this work published in 2001 (cf. COMFORT– BARRETT, 1999, 2001), offer a full Greek text with (in most cases) photographs and a discussion of 64 New Testament manuscripts, for which they consider a date before AD 313 (or AD 320 according to JAROŠ, 2006, pp. 10-12) either likely or possible. In a similar work, published on a CD-ROM of no less than 5163 pages, Jaroš (cf. JAROŠ, 2006) offers the full Greek text, including a diplomatic transcript, and a translation of 92 manuscripts (with photographs).

Impressive. But here is Clarysse’s cautionary comment:

In his review of Comfort–Barrett’s edition, for the Journal of Biblical Textual Criticism, D.C. Parker noted in 1999 that “some attempts at early dating owe more to apologetics than to palaeography”, but his main criticism went to the transcriptions. The second edition of Comfort–Barrett, with some corrections, was given a reasonably positive review in 2001 by M.A. Robinson in the same Journal of Biblical Textual Criticism; the reviewer concluded that this was “a useful primer for the student of New Testament papyri”, though he also stressed that “Comfort–Barrett appear to apply their own palaeographical criteria in dating various manuscripts and tend to claim an earlier date for many manuscripts included in their volume than might be allowed by other palaeographers“. (p. 445)

3. Comparison Between the Theological and the Papyrological Approach

Clarysse publishes a table comparing the datings found in Nestle-Aland (the consensus of papyrologists) and Jaroš.

Nestle-Aland allow a much wider range of possible dates — “a full hundred years is the norm” — than found in Jaroš. Jaroš is much more “precise”, using a range of 50 years, “and often less”.

After 300 CE in particular, Jaroš assigns the latest possible date for manuscripts as older than the earliest possible date accepted by Nestle-Aland (the papyrologists).

This observation is important for the argumentation because the date of one text often depends on that of others. If some of the dates in Jaroš can be shown to be too early, this will have implications also for the rest, since his datings are nearly always based on comparisons within the same corpus. Besides the narrow range of parallel texts (mostly biblical papyri), the dating method of Jaroš is also heavily dependent on that of Comfort–Barrett, which is regularly represented by him as forming a new consensus, though this is clearly not the case. (pp. 446-447)

Thus Jaroš apparently tends to rely on comparisons a much narrower corpus of material (biblical texts alone?) than papyrologists would consider best practice.

4. The Comparative Method in Palaeography

Comfort-Barrett and Jaroš explain their criteria for dating of NT manuscripts:

  • historical and archaeological data
  • re-use of an earlier manuscript in a more recent one
  • recto and verso criterion
  • cursive and documentary hands used in marginalia
  • and finally — but only if all these more “objective” criteria are absent — palaeographical comparison“.

Orsini (I use his name alone since he wrote this portion of the article, though both authors accept responsibility for the whole) here skips past the complex problems involved even with the above “objective” criteria and chooses to focus on the last mentioned: “the method of comparative palaeography adopted by Comfort-Barrett and Jaroš.

What is this method? In the absence of clear chronological elements, bookhands are dated by comparing them with other scripts which are dated or datable, and similar from a general point of view, in their style, their way of execution and their characteristic shapes. From a methodological point of view it is unsound to compare scripts belonging to different graphic categories, simply on the basis of a superficial impression21. In order to be “appropriate”, comparisons should be similia cum similibus. (pp. 447-448)

Orsini explains this problem within footnote 21:

21. For example, the script of P49 is compared with that of P53 (cf. COMFORT–BARRETT, 2001, pp. 357, 369), and the script of P104 with that of PSI XI 1213 [NORSA, 1929-1946, pl. 9a; LDAB 886] and of P. Oxy. LXII 4301 [P. Oxy. LXII, pl. I; LDAB 888] (cf. COMFORT–BARRETT, 2001, pp. 643-644), on the basis of the similarity of single letters, without considering the general graphic aspect.

How to get a chronological result

Palaeographical comparison may lead to chronological results when an undated manuscript is compared to an explicitly dated or to a datable one (i.e., a manuscript that contains no explicit date but objective chronological data, such as references to known people, places or events). Such parallels may lead to different results. They may:

1. connect an undated script with the same general graphic background to one or more dated and/or datable examples;

2. bring an undated manuscript into the context of a “stylistic class”*, whose chronological range can be reconstructed thanks to various dated manuscripts;

3. link an undated script to a “style”*, whose history and main distinctive aspects can be reconstructed thanks to dated and undated manuscripts;

4. connect an undated script with a “canonical”* or “normative script”* for which a system of internal rules and a history can be reconstructed;

5. attribute an undated manuscript to the hand of a scribe, known by other manuscripts, dated or undated. (pp. 448-449)


* Stylistic Class: a set of writings sharing a general framework, form and structure (in the 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 recognise a distinctive writing with no rigid and fixed rules . . .
* Individual “styles” are formed within a stylistic class, when the most frequent and most typical characteristics are selected and organized in a graphic structure with well-defined and homogeneous features
* Texts constituting a “canon” exhibit the repetition of a style, extended in time, i.e., canonical styles have lost their original spontaneity and repeat themselves nearly unchanged over a period of several centuries, for extra-graphic, historical and cultural reasons. Even if a text within a canon has a unitary and closed graphic structure, it may also have an internal dynamic, with chronological and geographical differences.

* In a recent publication also CAVALLO, 2008, p. 15, has questioned the very concept of canons, since such terminology is “too rigid for graphic forms which lack any theoretical basis of fixed rules to be followed, and which also are not required models but simply represent one possible choice among others”; for this reason he now prefers the term “normative scripts” . . .

So what does all this amount to?

In practice, not a single surviving literary manuscript between the first and sixth century is actually dated, and only few are datable. Therefore, for book hands within the chronological span considered by Comfort–Barrett and Jaroš (first to fourth centuries) only a comparison with datable manuscripts can be used, whereas dated manuscripts are only available for documentary hands, i.e., when the script used for the literary text is closely akin to a bureaucratic or administrative hand. (p. 449)


5. Palaeography as a Historical Science

The comparative method in palaeography is not only based on the shape of single letters. Indeed, writing is not a simple sequence of signs, to be considered as static entities. Moreover, palaeography, as a historical study of shapes (not only as a technique for dating and localising manuscripts), has in the last decades aimed at a comprehensive view of graphic phenomena.

In the works of Comfort–Barrett from 1999-2001 and Jaroš in 200629, however, there is no historical approach to the graphic evidence; their interest is simply in chronology. Their palaeographical analysis and their comparative method are based on impressionistic suggestions and on the shape of single and separate signs. As a result, in Comfort–Barrett the general graphic framework, reconstructed from the manuscripts that they take into account, is reduced to just a few types. In the work coordinated by Jaroš there is not even a hint at a general vision of graphic trends attested in his material, but palaeographical discussion is reduced to some notes added to the general description of the manuscripts presented. (pp. 449-450)

At this point the authors enter into detailed examinations of the reconstructions of Comfort-Barrett. They then place the manuscripts within their comprehensive and historical framework. I cited a few snippets from this lengthy discussion — those that related specifically to P52 — in my previous post.

But I post the above in order as a more informative background to the previous article. Well, at least it’s helped me understand a little better what’s involved with dating manuscripts.

The following two tabs change content below.

Neil Godfrey

Neil is the author of this post. To read more about Neil, see our About page.

Latest posts by Neil Godfrey (see all)

If you enjoyed this post, please consider donating to Vridar. Thanks!

11 thoughts on “More On Dating New Testament Manuscripts (and the Rylands Fragment P52 again)”

  1. There are also problems with palaeographical dates for Hebrew scripts found at Qumran (i.e. the Dead Sea Scrolls). The problem there is that so-called Herodian scripts were dated to the first CE by FM Cross in the early 1960s based on faulty archaeological data, when the proper dating of the archaeological context for 2 of 3 of his data points would have moved the dates earlier into the Hasmonean Era. Due to Cross’s authority as a scholar, his system of dating Qumran scripts was perceived as authoritative among scholars who didn’t have the expertise to chutzpah to challenge him. Now, 50+ years later, it is finally becoming normal to express doubts as to the accuracy of his palaeographical dates, but there has not yet been a systematic revisiting of the primary evidence to construct a satisfactory new system.

    BTW no scrolls scholar that I am aware of accepts the proposal that the Greek fragments include excerpts from the gospels. I have read the articles in which this was proposed and found them unconvincing.

    1. There never was a basis for reading James and Paul (Robert Eisenman) or Jesus (Barbara Thiering) into the scrolls. For one thing, radiocarbon dating indicates the scrolls are older than required for either theory. Eisenman’s methods for finding evidence for his theories can be described as follows. (1) Any parallel between Pauline controversies and the scrolls counts as a parallel. (2) Any contradiction between Pauline controversies and the scrolls counts as a parallel, only “inverted”!! By such logic, you could prove the scrolls belonged to any historical context desired. At any rate, neither theory has gained any supporter outside the author him or herself, and neither deserve serious consideration. Not that a lack of evidence has ever discouraged either author.

  2. A non-destructive physical form of dating has been available for a decade or more: Activated Neutron Analysis.

    A minute scrap of papyrus can be dated within a decade, if it has ink on it, the ink can be dated too.

    At the very least, it will give an “earliest possible” dating.

    Why has not this been proposed?

    Any country with a research synchrotron, such as Australia, can perform this for a small charge.

    Such information would be invaluable, if not embarrassing.

    1. I thought Activated Neutron Analysis was useful for determining chemical composition (that is, the constituent elements or atoms), like spectrography. How does this translate to dating? Any references to literature or news articles on dating of ink, papyrus or leather?

      I can see how analysis of an ink might show “modern” ingredients not used in ancient times, and would therefore be useful in identifying forgeries, but I’m curious on how the process results in absolute dates, like radiocarbon dating.

      1. And the material so analysed has to be treated as if low to medium radioactive waste for decades afterwards. That would be a little limiting I think!

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from Vridar

Subscribe now to keep reading and get access to the full archive.

Continue reading