Among the Christian papyri from Oxyrhynchus is an interesting piece now in the British Library (Pap. 2053) that contains the final verses of Exodus in Greek with an end title (P.Oxy. 8.1075) and on the reverse, in a different script, the opening verses of Revelation (P.Oxy. 8.1079):

British Library Papyrus 2053, showing P.Oxy. 8.1075 (Exodus) on the horizontal fibers (left) and P.Oxy. 8.1079 (Revelation) on the vertical fibers (right); image source: The British Library

The original editor (Arthur S. Hunt) published the texts separately and believed that the piece was a roll of Exodus that was later (but not much later) reused by copying the text of Revelation on the back of the roll. A few years ago, I suggested that this piece was equally likely (if not more likely) to be a leaf of a codex. The abstract of that article in Novum Testamentum laid out the lines of my argument:

British Library Pap. 2053 is a Greek papyrus fragment from Oxyrhynchus. Its two sides, written in distinctly different hands, were published separately in 1911 as P.Oxy. 8.1075 (Rahlfs 909, the last verses of Exodus) and P.Oxy. 8.1079 (P18, early verses of Revelation). The original editor of the piece, Arthur S. Hunt, described Pap. 2053 as part of a roll rather than a codex. He speculated that at some point after Exodus was copied along the fibers, the reverse was then reused for a copy of Revelation written against the fibers. I argue that, for at least three reasons, it is worth entertaining the possibility that Pap. 2053 is not part of a roll but rather a leaf from a codex. First, the amount of text and the format of the text on the papyrus are appropriate for a codex leaf. Second, we now have good evidence (unavailable to Hunt) for the existence of Christian codices with an eclectic mix of contents copied by different scribes. Third, when the back sides of rolls were reused, they were often rotated 180 degrees such that the texts on the two sides of the papyrus were upside down relative to one another, rather than same-side-up, as the columns of Pap. 2053 are.

The full article can be downloaded here. My conclusion was fairly modest: “In light of the absence of any compelling evidence that Pap. 2053 must be a roll, and given the general preference of the codex format on the part of Christians, we should probably no longer regard Pap. 2053 as a fragment of a roll but rather as a leaf from a codex constructed in a manner not unlike” that of more recently published codices with diverse contents produced by multiple different copyists (discussed below).

In the last couple years, two publications have appeared that are rather critical of this article. For the ca. six people in the world who truly care about this kind of thing, let us see if there is any merit to their criticisms. I’ll begin with an article by Scott Charlesworth that appeared in Buried History: The Journal of the Australian Institute of Archaeology in 2017.

According to the abstract, Charlesworth’s self-described “careful” examination shows that “Nongbri’s argument is found to be flawed in a number of ways.” But the first point that stands out about Charlesworth’s article is that much of it is occupied with material that doesn’t really relate to “Nongbri’s argument” (e.g. “[Nongbri] overlooks the fact that a sizable proportion of canonical gospel codices have text blocks wider than the normative column size of literary book rolls…,” a point that is true but has nothing to do with any of the arguments of my article, which is not about gospels, “canonical” or otherwise). So, setting aside this kind of thing, I will isolate and address the actual criticisms that Charlesworth makes.

To my first point: that the amount and formatting of the text on the papyrus–a single column of writing on each side with estimated page dimensions of about 27 cm high and 11.6 cm wide–is consistent with what we might expect on the leaf of a codex, Charlesworth seems to have no objection at all. Although the tone of his writing is quite antagonistic toward my article, Charlesworth’s claims in this regard simply restate the analysis in my article. After repeating my calculations of the size of the hypothetical leaf and confirming (within a few millimeters) my reconstruction of the dimensions, Charlesworth classifies the hypothetical leaf as belonging to “a sub-category of Turner’s Group 8,” again, simply restating, in a weirdly combative way, a conclusion of my own article, in which I wrote: “If the papyrus is indeed a leaf of a codex, it would, with these reconstructed dimensions (about 27 cm high by 11.6 cm wide), fit comfortably into Turner’s Group 8 (height = about 2x breadth) with its aberrants” (Nongbri 2013, 83, emphasis added–Charlesworth’s careful examination seems to have missed those last three words, since he falsely states that I assigned the leaf to Turner’s Group 8 proper).

To my second point: that the Exodus/Revelation papyrus can be fruitfully compared to codices with diverse contents written by different copyists. I suggested a useful comparandum for the Oxyrhynchus papyrus might be a leaf from a codex containing mixed contents, P.Bodmer 20+9, written by two different copyists. I illustrated the point with the leaf containing the end of P.Bodmer 20 (the Apology of Phileas) on one side and the beginning of Psalm 33 copied in a different hand on the other side:

P.Bodmer 20 (Apology of Phileas), written on the horizontal fibers (left) and P.Bodmer 9 (Psalms) written on the vertical fibers; image source: Fondation Martin Bodmer

I also observed some similarities between the two pieces:

“Note the difference in the breadth of the columns, which causes a discrepancy in horizontal margins of a sort comparable to what we see on the faces of Pap. 2053. It is especially instructive that the scribe of the Psalms wrote lines extending to the very edge of the inner margin, a somewhat unexpected feature that would also be true of the scribe of P.Oxy. 8.1079, if it is indeed a leaf from a codex. Like the hands of Pap. 2053, the hands of the Bodmer leaf, though distinctive, are most likely not far apart chronologically. In fact, Turner would locate all of the hands represented in the Bodmer codex in the fourth century. It does not seem unreasonable to suggest that British Library Pap. 2053 came from a similarly constructed codex” (Nongbri 2013, 84)

In response, Charlesworth objects that this is not the usual practice for copying codices:

“Ordinarily, a scribe copying a codex, even a second scribe as here, would want to maintain the uniform appearance of the codex by producing a leaf with text blocks that were as complimentary as possible.”

Charlesworth cites no data to support this claim, probably because the claim is demonstrably false, at least as regards examples in which “a second scribe” is at work on one side of the leaf. The example of P.Bodmer 20+9 amply proves the point, but Charlesworth objects that the comparison is not valid because in the case of the Bodmer codex, one side of the leaf is prose (Phileas) and the other side contains lines in verse (Psalms). But we see the same phenomenon even when only prose texts are involved. So, just keeping to material in the Bodmer collection, consider the leaf containing the end of P.Bodmer 5 (the Genesis of Mary) on one side and the beginning of P.Bodmer 10 (apocryphal correspondence of Paul and the Corinthians) on the reverse:

P.Bodmer 5 (Genesis of Mary) on the vertical fibers (left) and P.Bodmer 10 (Apocryphal correspondence of Paul and the Corinthians) on the horizontal fibers (right); image source: Fondation Martin Bodmer

Again, different hands on the opposite sides of the same leaf, and again, the sizes of the text blocks do not match, either vertically or horizontally. And again, the second copyist is writing closer to the central fold. Charlesworth is simply making up facts to suit his argument.

My third point was framed in tentative terms: “Finally, I offer one possible argument to cast some further doubt on the likelihood that Pap. 2053 is part of a roll. I have observed (from an admittedly non-comprehensive sample) that when the reverse sides of papyrus rolls were put to reuse, the roll was more often than not rotated 180°, so that the side written with the fibers is upside-down relative to the side written against the fibers.”

In a footnote, I added the following: “Frustratingly, I have found that editors often do not include this detail of orientation, and autopsy inspection is often necessary. The desideratum, a comprehensive catalogue of the orientations of reused rolls, would thus be a daunting task to carry out” (Nongbri 2013, 84).

Bearing in mind the incomplete nature of the data, the article went on to list examples of reused rolls with Christian content, most of which show front-to-back orientation that is upside down rather than same-side-up like the Exodus/Revelation papyrus. I noted one exception to this general pattern: “In fact, the only Christian literary text on the reverse of a reused roll known to me to be written right-side up relative to the ‘recto’ is P.Oxy. 4.654, a fragment of the Gospel of Thomas copied on the reverse of a land register.” Charlesworth objects to this line of argument in the following way:

“There is compelling evidence that BL Pap. 2053 also comes from a reused roll and is yet another ‘exception.'”

Charlesworth’s argument here is helpful insofar as it can serve for students as a nice example of the logical fallacy of petitio principii (“begging the question”), in that it assumes the truth of the very point that is being debated. But aside from its usefulness as cautionary example, this argument has no value.

To conclude his article, Charlesworth makes a final appeal to authority. He cites personal correspondence with Peter van Minnen, who made the following claims:

The ink on the back [of Pap. 2053] is occasionally in ‘crevices’ where the strips of papyrus do not join. This may have been very bad papyrus to begin with, but it is far more likely that the ink was applied only after the front had been used for quite a while, that is after the papyrus had been rolled up hundreds of times. The strips on the back then start to crack, leaving crevices. What I saw is typical of rolls that are reused on the back some time after use. On the back, strips run vertically and the ‘rolling’ pulls them apart. On the front, the strips run horizontally and the ‘rolling’ presses them closer together but does not break them. In codices, front and back are put in place at the same time and , after the quires are put together, the strain on either side of the papyrus is the same, actually minimal (another advantage of the codex form)” (letter from van Minnen to Charlesworth, 15 August 2017).

This argument overlooks a very basic codicological fact, recognized already by Hugo Ibscher and confirmed over the course of the twentieth century, namely that in the vast majority of cases, the papyrus used for the leaves of codices already spent an indeterminate amount of time in roll format before the sheets were cut from the roll for use as codex leaves. As a result, all but the most well prepared papyrus will show some degree of ink bleeding when writing against the fibers. Anyone who has rolled a newly made sheet of papyrus and heard it crackle knows that it does not take “hundreds of times” rolling a papyrus to cause this kind of breakage. All it takes is rolling it up once. Thus, “crevices” are completely common on the faces of codex leaves that have vertical fibers. To pull a random example from a codex I was recently working on: the first leaf of the text of P.Bodmer 13, the paschal sermon of Melito. Here are some close-ups of the writing against the vertical fibers, which show the crevices into which the ink bleeds.

Samples from page 1 of P.Bodmer 13; image source: Fondation Martin Bodmer

While not every papyrus codex shows this characteristic as clearly as P.Bodmer 13, this effect is in fact perfectly common in papyrus codices and has more to do with the quality of the papyrus than with any supposed “reuse.” So van Minnen’s argument is just as ineffective as the rest of Charlesworth’s arguments and carries no weight.

I thus remain convinced that, as I claimed before, “it is worth entertaining the possibility that Pap. 2053 is not part of a roll but rather a leaf from a codex.” I’m open to being persuaded that this piece is more likely a roll than a codex, but not a single one of the arguments attempted by Charlesworth is actually valid. This post is already too long, so I will turn to the arguments of the second article (by Peter Malik) in a separate post.