‘For a book is never simply a remarkable object. Like every other technology, it is invariably the product of human agency in complex and highly volatile contexts which a responsible scholarship must seek to recover if we are to understand better the creation and communication of meaning as the defining characteristic of human societies’.1
1. Introduction
What is the earliest surviving specimen of the codex format? One candidate for that honor is a scrap of a leaf from a parchment codex found at the Egyptian city of Oxyrhynchus bearing a historical text in Latin, P.Oxy. I 30. The exact date this codex was produced is unknown, but multiple experts have found it credible that this book was produced in the late first century ce.2 A recent study pushes that date back even further to the first half of the first century ce.3 Whatever the date of this particular piece, there is no doubt that the site of Oxyrhynchus accounts for a considerable portion of our earliest surviving samples of the codex format. As of July 2024, the Leuven Database of Ancient Books (LDAB) contained a total of 857 codices produced through the end of the fourth century.4 Of these, 283 have a provenance listed as Oxyrhynchus.5 The site is thus thought to account for almost a third of all published samples of early codices. For that reason, Oxyrhynchus is an ideal laboratory for thinking about the development and spread of codex technology.
In what follows, we present a brief bird’s-eye view of codicological data from Oxyrhynchus, focusing particularly on the Christian codices. We then zoom in on two codex fragments, P.Oxy. VII 1010 (LDAB 3181, TM 62022), a leaf from a parchment codex containing verses of Ezra in Greek and P.Oxy. I 2 (LDAB 2940, TM 61787), a fragmentary bifolium from a papyrus codex containing the Gospel According to Matthew in Greek. These two pieces illustrate some of the benefits and drawbacks of working with this particular corpus. On the one hand, the highly fragmentary nature of this material recovered from a trash context can present us with certain codicological puzzles that challenge us in ways that differ from the problems presented by more fully preserved specimens. On the other hand, our knowledge that these manuscripts come from a common source provides us with an unusually rich context for thinking about both the production and the use of early codices.
2. Oxyrhynchus and its Papyri
The story of the discovery of the ‘Oxyrhynchus papyri’ is most closely associated with the names of Bernard P. Grenfell and Arthur S. Hunt, two late-nineteenth century graduates of Oxford University who headed a decade of excavations specifically focused on the recovery of the manuscripts of Greek and Roman Egypt.6 With the support of the Egypt Exploration Fund (now the Egypt Exploration Society), Grenfell and Hunt began work at Oxyrhynchus in the winter of 1896–1897. Over the next decade, they spent five more winters in Oxyrhynchus uncovering a massive number of papyrus and parchment fragments from the city’s trash heaps. The bulk of these manuscripts were documents: the letters, receipts, lists, tax records, magical spells, and other ephemera of daily life in an ancient city. But they also found the remains of literary texts, including lost classical texts and early Christian works. While a small selection of these pieces remained in Cairo, the majority of the scraps were set between sheets of The Oxford Gazette in tin boxes that were packed into wooden crates and shipped to Oxford.7 Editing and publishing the material they excavated would keep Grenfell and Hunt occupied for the remainder of their lives (Grenfell died in 1926 and Hunt in 1934). Meanwhile, other excavators continued to work at Oxyrhynchus, including teams led by Giulio Farina and Ermenegildo Pistelli (1910–1914), Hilda and William Matthew Flinders Petrie (1922), Evaristo Breccia (1927–1934), the Egyptian Antiquities Organization (1982–1992), Géza Fehérvári (1985–1987), and the University of Barcelona commenced in 1992 and continuing until the present.8 Of these later expeditions, Petrie’s project and the Italian missions of 1910–1914 and 1927–1934 were responsible for the most significant addition to the total number of manuscripts excavated from Oxyrhynchus. There also seems to have been a substantial amount of illicit digging at Oxyrhynchus that put manuscripts into the hands of antiquities dealers and eventually into various collections around the world.9 The exact number of fragments recovered from Oxyrhynchus is not known. The Italian excavations recovered 1229 papyri, with just over half published as of 2016.10 Those with knowledge of the unpublished Oxyrhynchus material at Oxford put the total number of fragments recovered by Grenfell and Hunt anywhere from 500,000 to over a million.11 The lower number seems more credible, but much depends on how one counts ‘fragments’.
2.1 An Overview of the Corpus of Christian Codices at Oxyrhynchus
From among these hundreds of thousands of pieces collected by Grenfell, Hunt and other excavators, roughly six thousand have now been published, mostly in The Oxyrhynchus Papyri series but also in many other outlets. As of 2024, the LDAB records that 259 Christian books from Oxyrhynchus have been published.12 Because so much of the material at Oxford remains unpublished, it has been difficult to ascertain the statistical significance of this published portion. In March of 2019, however, the Egypt Exploration Society revealed the number of Christian literary manuscripts that have been identified but still await publication:
Some twenty New Testament inedita have been identified, none of them apparently earlier than the late 2nd to early 3rd century AD. They have all been assigned to editors, and will be published in The Oxyrhynchus Papyri series as the editors complete their work over the next few years. There may be more small fragments still unidentified because, like the Mark fragment recently published (LXXXIII 5345), their identity only emerges from much more detailed study than is feasible when cataloguing. We note that Grenfell and Hunt were particularly keen to find New Testament texts, and so sorted out possible cases as they processed their finds in Egypt and back at Oxford, and published many of them. Some ten patristic texts have also been identified and assigned to editors, and over eighty Septuagint and related texts are currently known to us and will gradually be assigned and published.13
Thus, at least 110 additional Christian pieces will be published in due course from the materials at Oxford (some of the Septuagint pieces may of course have been produced by Jews rather than Christians). Additional items will no doubt turn up in other collections. Nevertheless, it is probably safe to estimate that more than half of the Christian papyri collected from Oxyrhynchus have been published.14
Of these 259 published pieces, 164 are classified as codices by the LDAB.15 Most of the ‘codices’ from Oxyrhynchus are fragments of a single leaf. Precisely how many actual individual codices are represented by these fragmentary remains is difficult to determine. For instance, the Septuagint papyri published in The Oxyrhynchus Papyri LXXXIV consist of ten fragmentary leaves representing five discrete texts (Numbers, Deuteronomy, Joshua, 1 Kings, and Jeremiah). They are numbered 5404–5408. Yet, the editor notes that the script and the format of the pages appears to be identical (18.7 cm wide and 35 cm high) for all the leaves, and despite the presence of some page numbers and quire signatures, ‘there is no way of telling how many codices are represented’.16 Keeping such cases in mind, we are probably dealing with a corpus of at least 150 to 160 Christian codices from Oxyrhynchus.
The main characteristic of the entire assemblage is variety. Codices from Oxyrhynchus employed different materials. Papyrus was by far the better represented medium (126 papyrus examples and 38 parchment examples). Both single-quire and multi-quire codices are attested. They vary in size from miniature formats (dimensions under 10 cm wide and 10 cm high) to larger format codices with heights of up to an estimated 35 cm (the recently published Septuagint leaves mentioned above are among the largest surviving papyrus codex leaves). Overall, production value ranges from the humble to the luxurious. We find a full spectrum of types of handwriting, everything from competent cursives and capitals to the struggling hands of inexperienced writers.17 And it is the handwriting of the pieces that has been the main means by which scholars have assigned dates to the fragments. These palaeographic judgements, made with varying degrees of confidence, have been subject to greater scrutiny in recent years, but a reasonable estimate is that the codex fragments recovered from Oxyrhynchus range in date from the second or third century ce to perhaps the seventh or eighth century ce.18 Within that range, the majority of samples (about 80%) fall into the period from the third century to the fifth century.19 From among these codices, we have chosen two examples for a closer examination, P.Oxy. I 2 and P.Oxy. VII 1010.
3.1 P.Oxy. I 2: A Papyrus Codex Containing the Gospel According to Matthew
We begin then with one of the first pieces unearthed by Grenfell and Hunt from the trash heaps in their first season at Oxyrhynchus, P.Oxy. I 2, a fragmentary bifolium from a papyrus codex, one leaf of which contains the opening verses of the Gospel According to Matthew in Greek.20 This fragment presents several codicological challenges and illustrates some of the problems of working with the fragmentary material from Oxyrhynchus. Grenfell and Hunt assigned the fragment to the third century, and this date is still generally accepted.21
Verses 1:1–1:12 of Matthew are written against the fibers and labelled as page ⲁ (see Fig. 1). Verses 1:14–1:20 are written along the fibers and labelled as page ⲃ (see Fig. 2).
Much of the conjoint leaf is lost. The small surviving portion is blank on the side with vertical fibers. On the horizontal fibers, there are the beginnings of three lines written in a larger hand apparently distinct from that of the copyist of the gospel text. The upper margin of the conjoint leaf is 2.5 cm, considerably wider than the upper margin of the leaf containing the text of Matthew (0.7–0.9 cm). Grenfell and Hunt read the text as:
ⲉⲅⲉⲛ̣[
ⲡⲁⲣ[
ⲙⲏⲧ̣[
The fragment thus presents a couple of codicological puzzles. The first concerns the original dimensions of the codex. The surviving portions of the leaves can give the impression of a roughly square format, but this is deceptive. Indeed, some confusion has surrounded the measurements of the bifolium.
As it is now framed, the bifolium is in four parts. A large upper section (numbered ‘2’ in red ink on the vertical fibers), a large lower fragment (also numbered ‘2’ in red ink on the vertical fibers), a smaller detached fragment of about 2.5 cm wide placed below the larger pieces, and a very small fragment that has come loose and is out of place below the 2.5 cm fragment. Grenfell and Hunt provided only the maximum dimensions of the bifolium as represented by the two large fragments (14.7 cm wide and 15.0 cm high). Kurt Aland reported both the dimensions of the combined larger fragments and the dimensions of the smaller detached fragment placed below the larger pieces: ‘Zwei Fragmente eines Doppelblattes aus einem Kodex (14,7 x 15 und 2.5 x 3)’.22 José O’Callaghan subsequently cited Aland’s quotation but misunderstood the measurement of the smaller detached fragment as a reference to the smaller of the two preserved folia. He thus took the two folia to be separate, which resulted in an impossible codicological reconstruction and the mistaken suggestion that the remains of three lines preserved on the smaller folium were part of Matt. 2:14 written against the fibers.23 Consequently, the contents of the bifolium are misreported in several sources.24
3.2 The Dimensions of P.Oxy. I 2
A second problem concerns the dimensions of the individual leaf of the codex. Grenfell and Hunt made no explicit statement about this measurement. The width of the page containing the text of Matthew is fully preserved, but reported measurements differ considerably. Kurt Aland reported a width of 13 cm, while Eric Turner reported a width of 12 cm.25 The discrepancy seems to be the result of differing identifications of the central fold of the bifolium and probably also measurements taken from different areas of the papyrus, as illustrated in Fig. 3.
Turner seems closer to the mark. It appears to us that a fold at about 12 cm from the outer edge of the page can be seen faintly on both sides of the bifolium, and the mirrored pattern of holes at some points on either side of that vertical line also supports its identification as the central fold of the bifolium. Furthermore, along this fold, there appear to be small narrow slits that could well be the openings through which a binding thread or tacket passed. If the fold instead fell at the 13 cm mark as Aland proposed, the stray three lines on the smaller folium would have a left margin of just 0.5 to 1.0 cm, quite disproportionate to the right margin of the conjoint leaf, which would be about 2 cm. On Turner’s measurement, those numbers would be reversed, which would make more sense, if, as has been proposed, the remains of the three lines in a different hand represent a paratext of some sort, since paratexts such as titles or colophons are frequently indented rather than aligned flush with the left margin.26
The vertical dimensions of the leaf must be estimated using the number of missing lines of text between the lowest preserved portion of page ⲁ and the top of page ⲃ. Aland estimated a height of 25 cm, while Turner estimated a height of 24.7 cm, a distinction without a difference when speculating about restored dimensions. Turner’s proposed dimensions of 12 cm wide and 24.7 cm high placed this codex into his Group 8 (height roughly equal to twice the width). This will be our working assumption for the dimensions of this codex. So, we must imagine a taller and somewhat narrow folium, as illustrated in Fig 4.
3.3 The Place of P.Oxy I 2 within its Quire
With a working hypothesis for measurements, we can now turn to the more complicated issue of the position of this bifolium within its quire. Grenfell and Hunt summed up the problem succinctly in 1897: ‘As the arrangement in the quire of the two leaves forming the sheet is wholly uncertain, the question what relation, if any, the beginnings of the three lines on the other leaf have to the St. Matthew fragment cannot be determined. The difference in the handwriting and the greater margin above the three broken lines distinguish them from the text of St. Matthew, though they may have formed a title of some kind’.27 A few other explanations have been raised over the years, but none has proven persuasive.28 More recent scholarship has tended to follow the suggestion of Grenfell and Hunt that the letters may be a title related somehow to the gospel text, even though the surviving letters do not correspond to any known title of the gospel in any manuscript.29 In any event, the presumption would be that the bifolium was folded with the horizontal fibers on the outside and the vertical fibers on the inside. The ‘title’ page (→) with the three stray lines would then be followed by an apparently blank page (↓), and then page ⲁ (↓), followed by page ⲃ (→). The unstated assumption, then, is that this surviving bifolium was either the center of a quire (and thus preceded in the codex by an unknown number of leaves) or a quire consisting of just a single bifolium.30 The other possibility with this folding pattern would be that the bifolium was not the center of the quire, meaning that an unknown number of leaves intervened between the ‘title’ page and page ⲁ of Matthew, decreasing the likelihood that these three lines have any relationship at all with the gospel text.
An alternative possibility, apparently so far unexplored, is that the bifolium was folded with the vertical fibers on the outside and the horizontal fibers on the inside.31 This folding pattern would be more suggestive of a single quire construction. The stray letters on the fragmentary leaf could be thought of as part of a concluding title or colophon for the work, with a complete copy of the Gospel According to Matthew in intervening leaves. This option has the virtue of making the page labelled ⲁ actually be the first page of this codicological unit. If this bifolium was part of a single quire codex, the number of pages required to contain the whole of the Gospel According to Matthew in something like the form it appears in modern editions would be roughly ninety. Because the leaf conjoint with the leaf ⲁ-ⲃ seems to have contained little or no text from the gospel, we could imagine a quire composed of ninety-two pages, that is, twenty-three bifolia.32 Examples of single-quire books of this format (Turner’s Group 8) that are more completely preserved show that such a thickness is relatively common. Nag Hammadi Codex V (LDAB/TM 107745) especially resembles this imaginary single quire P.Oxy. I 2. So also does another single-quire gospel codex from Oxyrhynchus, P.Oxy. II 208 + XV 1781 (LDAB 2780, TM 61630), a copy of the Gospel According to John in Greek.33 The so-called Qau Codex (LDAB, TM 107755) containing a copy of the Gospel According to John in Lycopolitan Coptic is also quite similar.34 Tab. 1 organizes the data. Estimated quantities are in [brackets].
| Width of external folia | Height of folia | Thickness (number of bifolia) | |
|---|---|---|---|
| Nag Hammadi Codex V | 13.6 cm | 24.3 cm | 23 ½ bifolia |
| Qau Codex | 11.6 cm | 24.6 cm | [25 bifolia] |
| P.Oxy. II 208 + XV 1781 | [12.5 cm] | 25.0 cm | [25 bifolia] |
| P.Oxy. I 2 as hypothetical single quire codex | 12.0 cm | [24.7 cm] | [23 bifolia] |
The seemingly small distance between the right margin of page ⲃ and the central fold of the bifolium (about 1 cm) is also paralleled by both P.Oxy. II 208 and Nag Hammadi Codex V, in which the inner margin is about 1 cm ( a width of 2 cm separates the two columns of writing on the facing pages on the outermost bifolia). The outer leaves of the Qau codex have a slightly larger distance between the columns of writing (2.3–2.6 cm).35 On the whole, then, it would seem that the physical features of P.Oxy. I 2 are not incompatible with it being the outermost bifolium of a single quire codex (perhaps also the most likely bifolium to break off from the quire and be sent to the trash heap).36 At the end of the day, however, we do not have enough data to decide definitively among these various possibilities. None of them can be categorically ruled out. Such are the joys and frustrations of working with trash.
4.1 A Miniature Codex with 6 Ezra (P.Oxy. VII 1010)
Our second case study involves a tiny vellum page containing a few verses of 6 Ezra, part of what is also known as 2 Esdras (P.Oxy. VII 1010).37 From the more complete Latin translation, we can see that the short passage our parchment leaf preserves, 6 Ezra (2 Esdras) 15:57–59, is addressed to the province of Asia, personified as a woman.38 The message is rather depressing:
(Your children) shall die of hunger, and you shall fall by the sword; and your cities shall be destroyed, and all your people that are in the plains shall fall by the sword, and they that are on the mountains and highlands shall die of hunger and shall eat their own flesh and drink their own blood in hunger for bread and thirst for water. And first you have come to be wretched and again a second time (you shall receive woe).39
The palaeography and codicology of this page open up larger questions regarding the work, its circulation, and readers. In fact, this small page has outsized significance in the transmission history of 6 Ezra. According to Theodore Bergren, its publication in 1910 was ‘the single most important event for the textual history of 6 Ezra’.40 This page constitutes the only Greek manuscript of 6 Ezra. As such, the Oxyrhynchus fragment of 6 Ezra proves that the Latin version of the text was translated from Greek, and also testifies to the work’s circulation in the fourth century.41
While this mini-page confirms the circulation of a Greek text of 6 Ezra in the fourth century generally, it also situates this work more up close in the possession of inhabitants of Oxyrhynchus together with the numerous other literary and documentary texts and other trash from that city.42 Multiple other Greek works previously known only in translation or even previously entirely unknown texts have turned up from the Oxyrhynchite trash heaps.43 Together with fragments of texts now considered canonical (Septuagint and New Testament), this great variety of sources from Oxyrhynchus makes the site of the utmost importance for the study of early Christianity.44 These different texts also raise the question of who owned and read them.45 In addition to its textual significance, this manuscript as a physical object contributes to the history of the work and to the sociology of reading at Oxyrhynchus. Who read this manuscript? What does the materiality of this page reveal about its production and owner? In order to glean information about the sociology of reading of this book, we will first examine book format, handwriting and pagination, and then inquire about readers and owners, including a possible connection with a documentary letter (P.Oxy. LXIII 4365).
4.2 The Format of P.Oxy. VII 1010
Measuring only 5.6 cm wide and 8.4 cm tall, this page formed part of a strikingly tiny book (see Fig 5).46
To give an indication of just how small this book is: the page matches almost exactly the size of a modern credit card.47 When read from, such a miniature codex fits comfortably into the hands of a reader, creating an intimate reading experience. Its small size also makes it highly portable.48 This small book was made of parchment.49 While the cost of animal skin made the production of parchment manuscripts more expensive than plant-based papyrus manuscripts,50 miniature books could be made from leftover scraps of parchment or from the skin of just one sheep or goat, rendering them more affordable.51
In the history of the codex, small books are attested from the beginning.52 Miniature codices were not uncommon in antiquity; the LDAB lists 243.53 The majority of these (191 or 79%) are, just as P.Oxy. VII 1010, written on parchment. 204 of these miniature codices are Christian.54 Locally, at Oxyrhynchus, 37 miniature codices are currently known; 28 of these are classified as Christian.55 Gamble interpreted the small format as evidence for private reading, comparing it to modern pocket books.56 As we will see below, several Christian authors associate small books especially with women.
Page 40 is a codicological recto (i.e., the righthand side of an opening of the codex). The leaf broke off in the middle at the folding line. Damage at two places on what was the fold line of the bifolium of P.Oxy. VII 1010 indicates that the leaf came from a bound codex.57 The sewing holes have torn horizontally into the written surface of the vellum, perhaps an indication of use. No traces of the binding material survive, nor is there any indication whether this was a single or multiple quire codex.
Regarding the mise-en-page: The lower margin is wider than the upper margin, as is common in ancient manuscripts.58 Furthermore, the page has a relatively wide outer margin, and a much narrower inner margin; in other words, the text is not centered on the page, but is closer to the center of the bifolium. Especially on the hair side, the text is penned very close to the page fold. According to Turner, it remains to be discovered whether there is any established relationship between the written space and the complete page (that is, upper, lower, and side margins around the written space)’.59 Nevertheless, the practice of placing the text closer to the inner margin can be seen in contemporary manuscripts also, for instance, PSI I 5 (LDAB 2771, TM 61621), an Oxyrhynchite Epistle of James miniature codex of roughly similar format as ours,60 P.Bodm. II (LDAB 2777, TM 61627), the Gospel of John,61 the Freer Gospels (LDAB 2985, TM 61831), the Medinet Madi Kephalaia codex (LDAB, TM 108111),62 and the Gospel of the Lots of Mary (LDAB, TM 382619).63
Why did the ancients prefer this layout? Is this an aesthetic decision to have wider empty spaces on the outside?64 Or does it have a more practical function? For one, the margin is the place where books are handled. The Gospel of the Lots of Mary codex exemplifies that margins become dirty and are vulnerable to destruction, what Turner describes as ‘the liability of edges to damage’.65 The Ezra page, P.Oxy. VII 1010, has a tear in the bottom corner at the spot where one turns the page.66 Moreover, readers sometimes wrote corrections and comments—marginalia—in these outer margins, though that is not the case with this page. Whether this page layout with the wider outer margins and narrower inner margins is a normal practice in our earliest codices we would often not know because the outer margin of pages is often lost or shows repairs. For instance, P.Bodm. 19 (LDAB, TM 107759) has very narrow inner margins and large outer margins and often shows damage to the outer edges of the leaves.67 All in all, the page layout of this small codex with 6 Ezra falls fully in place with contemporary ancient manuscripts. The person who inscribed it followed a common bookish practice.
This type of mise-en-page has implications for the binding and consequently the reading experience. Book binding is a special skill in and of itself.68 The binder needed to make sure the codex could open flat enough to give access to the entire text block, and so the binding must have some space. If bound too tight, the book will get damaged from opening it.69 In the case of this tiny Ezra page, the relatively deep horizontal tears at the sewing holes signal tension in the binding. On the plus side, this kind of damage allows us to imagine readers of this text, opening the codex and turning its tiny pages.
4.3 The Palaeography of P.Oxy. VII 1010
The text on the flesh side (recto) is well preserved, but the hair side has suffered damage, making it more difficult to decipher.70 The handwriting is an outgrowth of the Severe style and dates probably to the fourth century.71 Alpha and delta are pointed; rho and phi have long extenders; upsilon sports a long, at times squiggly tail, with high divider. But mu and omega have rounded shapes. There appears to have been some effort for a contrast between thick vertical strokes and thin horizontal strokes (note the horizontal strokes of delta and tau, for instance).
Despite the small size of the letters (between 0.2 and 0.3 cm), the copyist wrote legibly and strove to maintain left and right margins.72 Each side has 12 lines of text, but on the recto, the letter ωροις are added in a smaller script below line 12 on the bottom of the page as a correction; presumably the copyist had omitted it when turning the page and added it later. The length of the lines varies from 8 to 13 letters; most of them have 10 or 11 letters. Other, common, scribal features are: suspended nu at the end of line 5 and at the end of line 23; and diairesis over upsilon in line 21. The Greek is good; apart from the iotacism in forms of the word λιμός (λειμῷ for λιμῷ in line 13 and, presumably λειμοῦ for λιμοῦ in line 19; this is thus also how the first word on the recto/flesh side has to be reconstructed).
While the script gives the impression of an experienced writer, this is not a professional bookhand.73 Forms of individual letters vary substantially, and the copyist does not maintain a consistent size or spacing. The script is practiced and competent, but not particularly elegant.74 Eric Turner phrased it well, discussing handwriting more broadly in second and third century papyrus codices: ‘Now it may be readily admitted that it is not easy to find examples of calligraphy among papyrus codices of the second or third centuries. Their handwriting is in fact often of an informal and workaday type, fairly quickly written, serviceable rather than beautiful, of value to a man interested in the content of what he is reading rather than its presentation’.75 A reasonably close parallel for the hand of 1010 is P.Oxy. XI 1358 (LDAB 1241, TM 60127, which LDAB dates 250–350 ce), fragments of Hesiod penned on the back side of ‘third-century official accounts.’ 1358 is much more professionally executed (and in much larger script), but some of the letter forms are similar, and if 1358 had been written on parchment, we might see some of the same shading effects with thick/thin variation that 1010 has.76 Other comparanda, with slightly better penmanship, might include P.Oxy. VIII 1080 (LDAB 2793, TM 61643) a copy of Revelation usually assigned to the fourth century ce77 and PSI I 5 (LDAB 2771, TM 61621) a leaf of James usually assigned to the fourth or fifth century ce.78 Both these items are also parchment miniature codices. Although these latter two examples lack firm dates, it seems fair to conclude that these miniature books were en vogue in roughly the fourth or fifth century in Oxyrhynchus. The handwriting and format of P.Oxy. VII 1010 together suggest that this book belonged to a private person, instead of a church, where it would more likely be intended for reading out loud to a group. This evidently puts its owner(s) among the small part of the population that was literate.79
4.4 The Implications of Page Number(s) of P.Oxy. VII 1010
A mu with supralinear stroke in the center top margin of the flesh side, penned by the same hand as the main text, marks this as ‘40.’ The placement of page numbers in Greek papyri has as far as we know never received a full treatment.80 According to Turner, ‘the center of the upper margin’ is ‘the favorite place’ for a page number, although he admits there is no consistency.81 Furthermore, while in modern practice, right side pages (the codicological recto) have uneven numbers, in antiquity there is no fixed system of pagination yet. In this case, probably the first page (or pages) of the quire was left empty, for instance to be fastened to the cover, and the text began on the left hand page so that page 40 falls on the right.82 The hair side lacks a page number.83 This seemingly small omission presents an important codicological issue regarding the length of the work: is this page 40 (folium 20) or folium 40 (page 80)?84 In other words: did 39 pages precede this one, or 79?
This is not just a matter of the practice of numbering pages. What is at stake here is whether what we call 6 Ezra circulated as an independent text or whether it was at this point in its circulation already attached to 4 Ezra, the Jewish apocalypse (a.k.a. 2 Esdras chapters 3–14). The single small Greek character of the page number thus has big implications for the textual transmission of 6 Ezra. Since this is the only preserved Greek part of the text, we must turn to the Latin translation to estimate the amount of text preceding these verses. Μiniature codices with hundreds of pages are attested. For instance, the Coptic Psalter in the Freer Collection (LDAB, TM 107936), a manuscript of about the same page size as the Ezra codex, may have contained more than 700 pages.85 However, the number 40 makes it unlikely that 14 chapters of 5 and 4 Ezra (2 Esdras 1–14, or even just the 4 Ezra part minus 5 Ezra, i.e., 2 Edras 3–14) preceded this section; even 79 miniature pages would not suffice for that amount of text.86
The Greek of P.Oxy. VII 1010 has 47 (28 + 19) words in this section, for an average of 23.5 words per page. This matches almost exactly the words in the Latin translation.87 Now the Latin of 6 Ezra (2 Esdras 15–16) has 56 verses with 888 words before the section preserved on our page. Spread over 39 miniature pages, that yields an average of almost 23 (22.77) words per page; or spread over 79 pages, roughly 11 (11.24) words per page. This calculation makes it likely that the mu indicates page 40, and that the codex began with what we call 6 Ezra.88 The seemingly insignificant detail of the page number thus indicates that when this codex was produced, 6 Ezra likely circulated by itself, not attached to 4 Ezra.89
The Latin of 6 Ezra continues for another 1154 words to the end of the work.90 Assuming a continued close correlation between the Greek and Latin and approximately 23.5 words per page, another 49 pages completed the codex for a book of roughly 90 pages, or 23 bifolia (if this was a multi quire codex, 24 bifolia would make 6 quires of 4 bifolia each).
Much remains unknown, of course, such as: Did this work already circulate under the name of Ezra? Did the codex contain only 6 Ezra, or did another work come after it? And: Whose hands held these small pages to read these apocalyptic words?
4.5 A Possible Owner for P.Oxy. VII 1010?
Until recently, the scholarly consensus was that 6 Ezra was a Christian appendix to Jewish 4 Ezra.91 John Marshall, however, has argued that 6 Ezra is itself a Jewish composition: ‘6 Ezra makes no mention of Jesus, or a Lamb, gives no indication of churches, claims no distinction from Jews or Judaism, makes no clear allusions to Christian documents or Christian rituals’.92 Yet, just as with so much other Jewish literature, 6 Ezra has been transmitted by Christians.93 The question of whether 6 Ezra is a Jewish or Christian text also matters here regarding the readers: Were the readers of this small codex Oxyrhynchite Jews or Christians? Jews figure in multiple literary and documentary Oxyrhynchus papyri. Although the revolt under emperor Trajan (115–117) had impacted local Jews seriously, there is clear evidence of a Jewish community again towards the end of the third century (the time of Diocletian) and afterwards.94
Typically, features such as nomina sacra (scribal contractions of the words God, Lord, Jesus, and Christ, among others) and the codex format are regarded as reasons to believe that a manuscript originated in an environment that was Christian rather than Jewish.95 But this folium does not contain any of the words usually contracted as nomina sacra, and it would also be too simplistic to imagine that Jews read exclusively from rolls and Christians from codices. The miniature format has been especially closely associated with Christian readers, to the point that Colin Roberts once characterized them as ‘a Christian invention.’96 Yet, examples of non-Christian literature in this format (especially Isocrates) are not uncommon.97 And Jews read and used texts in miniature formats as well.98
The larger question that this folium raises is: What does it mean for a ‘Jewish’ text to become ‘Christian’? Meron Piotrkowski has argued that the Christian transmission of ‘Jewish’ literature at Oxyrhynchus suggests that Oxyrhynchite Christians at one point had obtained these works from Jews, which in turn indicates friendly contacts among certain Jews and Christians, and also a continued interest from Jews in these works.99 Nevertheless, if Christians at Oxyrhynchus transmitted literature that clearly emerged in a Jewish milieu (such as Septuagint pieces or so-called pseudepigrapha), this does not necessarily indicate these texts that emerged in a specifically Oxyrhynchite Jewish milieu. Such texts could also have entered Oxyrhynchite Christian communities as ‘Christian texts’ already; the point of transition or contact with an originating Jewish community could have occurred at a much earlier time than the date of our surviving manuscripts.100 At some point in the manuscript transmission process, Greek Torah texts did become ‘Old Testament’ texts. But the exact mechanics of this process, and certainly the exact moment(s) when this happened, are lost to us. All things considered, it is probably more likely that P.Oxy. VII 1010 was produced and used in a Christian rather than a Jewish setting, but the issue is perhaps more ambiguous than it is often presented.101
Thomas Kraus concludes his codicological and paleographical discussion of P.Oxy. VII 1010 by writing that the small codex ‘opens a window to the past and its people’.102 But he does not point out the possible link with a documentary papyrus from Oxyrhynchus that mentions a book of Ezra, a short letter (P.Oxy. LXIII 4365).103 The reason for Kraus’s reticence may be that other scholars have firmly rejected any connection between this document and the leaf from our codex of Ezra. Eldon Epp for instance, has argued that ‘only the wildest speculation would identify that [leaf, P.Oxy. VII 1010] with the ‘Ezra’ of our letter [P.Oxy. LXIII 4365]’.104 But just how ‘wild’ would such speculation actually be?
The letter (P. Oxy. LXIII 4365) was penned on the back side of a petition dated to the second half of the third century (published as P.Oxy. LXIII 4364).105 The reuse happened probably in the fourth century.106 The writing of the letter is reasonably well preserved (see Fig. 6).
The note writer seems impatient to receive the copy of Ezra as part of an exchange of books in which another text had already passed hands: ‘To my dearest lady sister in the Lord, greetings. Lend the Ezra, since I lent you the Little Genesis. Farewell from us in God’.107
We encounter an anonymous woman (‘my dearest lady sister’) who owns an Ezra manuscript and an equally anonymous owner, whether female or male,108 of ‘the little Genesis’, probably a reference to the book of Jubilees.109 Both parties likely had additional manuscripts in their possession.110 All indications are that the works mentioned in the note are private copies, in collections owned by individuals.111 Could P.Oxy. VII 1010 be the ‘Ezra’ of P.Oxy. LXIII 4365? As we have seen, the handwriting of P.Oxy. VII 1010 is not incompatible with a fourth century date. Both the handwriting style and the miniature format of P.Oxy. VII 1010 are consistent with what we expect of a book in a private collection. It is perhaps also worth noting that multiple Christian writers associate miniature books with women.112 A connection between these two manuscripts is indeed speculative, but the speculation is perhaps not quite so wild as some would have it.
5. Conclusions
From just these two examples, it should be clear that the large repository of codex remains from Oxyrhynchus, despite their frequently poor state of preservation, offer us rich possibilities for thinking about different aspects of the early codex, from the details of their construction to the multiple possible lived contexts in which they were used before being consigned to the trash heaps. Indeed, the case studies we present here just begin to scratch the surface of what the papyrological remains from Oxyrhynchus can teach us about the codex. We have focused almost exclusively on Christian evidence, but the non-Christian use of the codex is quite well attested at Oxyrhynchus. Furthermore, the early pieces published by Grenfell and Hunt are in need of reassessment in light of the vastly increased body of comparative evidence, and some of the successors of Grenfell and Hunt (notably Edgar Lobel) rarely focused on issues of codicology, preferring to concentrate solely on text critical issues.113 A fuller study of the codex at Oxyrhynchus remains a desideratum.114
Bibliography
Aland, Kurt (1967), Studien zur Überlieferung des neuen Testaments und seines Textes, Berlin: de Gruyter.
Aland, Kurt (1976), Repertorium der griechischen christlichen Papyri I: Biblische Papyri, Berlin: de Gruyter.
Askeland, Christian (2018), ‘Dating Early Greek and Coptic Literary Hands’, in Hugo Lundhaug and Lance Jenott (eds), The Nag Hammadi Codices and Late Antique Egypt, Tübingen: Mohr Siebeck, 457–489.
Bagnall, Roger S. (2009), Early Christian Books in Egypt, Princeton: Princeton University.
Bagnall, Roger S. (2011a), Everyday Writing in the Graeco-Roman East, Berkeley: University of California.
Bagnall, Roger S. (2011b), ‘The Readers of Christian Books: Further Speculations’, in Guido Bastianini and Angelo Casanova (eds), I papiri letterari cristiani: atti del Convegno internazionale di studi in memoria di Mario Naldini, Firenze, 10-11 giugno 2010, Firenze: Istituto papirologico G. Vitelli, 23–30.
Bagnall, Roger S. and Raffaella Cribiore (2008), Women’s Letters from Ancient Egypt, 300 BC-AD 800. ACLS Humanities E-Book. New York: ACLS Humanities E-Book.
Bergren, Theodore A. (1998), Sixth Ezra: The Text and Origin, Oxford University.
Bergren, Theodore A. (2013), ‘Sixth Ezra’ in Richard Bauckham, James R. Davila, and Alexander Panayotov (eds), Old Testament Pseudepigrapha: More Noncanonical Scriptures, Grand Rapids: Eerdmans, 483–97.
Blumell, Lincoln H. (2012), Lettered Christians: Christians, Letters, and Late Antique Oxyrhynchus (New Testament Tools, Studies, and Documents, 39), Leiden, Boston: Brill.
Blumell, Lincoln H. and Thomas A. Wayment (eds) (2015), Christian Oxyrhynchus: Texts, Documents, and Sources, Waco: Baylor University.
Bowman, Alan K., Revel A. Coles, Nikolaos Gonis, Dirk Obbink, and Peter J. Parsons (eds) (2007), Oxyrhynchus: A City and Its Texts, London: Egypt Exploration Society.
Boudalis, Georgios (2018), The Codex and Crafts in Late Antiquity, New York: Bard Graduate Center.
Bromer, Anne and Julian I. Edison (2007), Miniature Books: 4,000 Years of Tiny Treasures, New York: Harry N. Abrams.
Charlesworth, Scott D. (2016), Early Christian Gospels: Their Production and Transmission, Florence: Gonnelli.
Claremont College Digital Library (no date), Nag Hammadi Archive, <https://ccdl.claremont.edu/digital/collection/nha/id/2046/rec/23> (last accessed 30 December 2020).
Clarysse, Willy (1983), ‘Literary Papyri in Documentary Archives’, in Edmond Van ’t Dack, Peter Van Dessel, and Wilfried Van Gucht (eds), Egypt and the Hellenistic World. Proceedings of the International Colloquium Leuven, 24-26 May 1982 (Studia Hellenistica, 27), Leuven: >Nationaal Fonds voor Wetenschappelijk Onderzoek, 43–61.
Cockle, Walter E. H. and W. Benjamin Henry (2019), ‘P.Oxy. LXXXIV 5404–8 Septuagint Codices,’ in Amin Benaissa, Nikolaos Gonis, W. Benjamin Henry, and Micaela Langellotti (eds), The Oxyrhynchus Papyri, Part LXXXIV, London: Egypt Exploration Society, 1–20.
Del Corso, Lucio (2006), ‘Lo “stile severo” nei P.Oxy.: una lista’, in Aegyptus, 86: 81–106.
Depauw, Mark and Tom Gheldof (2014), ‘Trismegistos: An Interdisciplinary Platform for Ancient World Texts and Related Information’, in Łukasz Bolikowski, Vittore Casarosa, Paula Goodale, Nikos Houssos, Paolo Manghi, and Jochen Schirrwagen (eds), Theory and Practice of Digital Libraries - TPDL 2013 Selected Workshops (Communications in Computer and Information Science, 416), Cham: Springer, 40–52.
Egypt Exploration Society (2019), Unpublished EES Biblical Papyri. <https://www.ees.ac.uk/news/unpublished-ees-biblical-papyri> (last accessed 30/12/2020).
Epp, Eldon Jay (1997), ‘The New Testament Papyri at Oxyrhynchus in their Social and Intellectual Context,’ in William L. Petersen, Johan S. Vos, and Henk J. de Jonge (eds), Sayings of Jesus: Canonical and Non-canonical: Essays in Honor of Tjitze Baarda (Supplements to Novum Testamentum, 89), Leiden: Brill, 47–68.
Epp, Eldon Jay (2004), ‘The Oxyrhynchus New Testament Papyri: “Not Without Honor Except in Their Hometown”?’, in Journal of Biblical Literature, 123/1: 5–55.
Epp, Eldon Jay (2006), ‘The Jews and the Jewish Community in Oxyrhynchus: Socio-Religious Context for the New Testament Papyri’, in Thomas J. Kraus and Tobias Nicklas (eds), New Testament Manuscripts: Their Texts and Their World (Texts and Editions for New Testament Study, 2), Leiden, Boston: Brill, 13–52.
Epp, Eldon Jay (2007), ‘The New Testament Papyri at Oxyrhynchus: Their Significance for Understanding the Transmission of the Early New Testament Text’, in Alan K. Bowman, Revel A. Coles, Nikolaos Gonis, Dirk Obbink, and Peter J. Parsons (eds), Oxyrhynchus: A City and Its Texts, London: Egypt Exploration Society, 315–31.
Feldman, Ariel (2022), Tefillin and Mezuzot from Qumran: New Readings and Interpretations (Beihefte zur Zeitschrift für die alttestamentliche Wissenschaft, 538), Berlin: De Gruyter.
Fournet, Jean-Luc (2010), ‘Femmes et culture dans l’Égypte byzantine (IVe-VIIe s.)’, in Béatrice Caseau (ed.), Les Réseaux Familiaux: Antiquité Tardive et Moyen-Âge. In Memoriam A. Laiou et É. Patlagean (Centre de Recherche d’Histoire et Civilisation de Byzance – Monographie 37), Paris: ACHCByz, 135–45.
Gamble, Harry Y. (1995), Books and Readers in the Early Church: A History of Early Christian Texts, New Haven: Yale University.
Gonis, Nikolaos (1997), ‘Notes on Two Epistolary Conventions’, in Zeitschrift für Papyrologie und Epigraphik, 119: 148–154.
Grenfell, Bernard P. (1897), ‘The Oldest Record of Christ’s Life’, in McClure’s Magazine, 9: 1022–1030.
Grenfell, Bernard P. and Arthur S. Hunt (1898), The Oxyrhynchus Papyri, Part I, London: Egypt Exploration Fund.
Grenfell, Bernard P. and Arthur S. Hunt (1899), The Oxyrhynchus Papyri, Part II, London: Egypt Exploration Fund.
Hubai, Peter and Angelika Balog (2009), Koptische Apokryphen aus Nubien: der Kasr el-Wizz Kodex (Texte und Untersuchungen zur Geschichte der altchristlichen Literatur, 163), Berlin: de Gruyter.
Hunt, Arthur S. (1910), The Oxyrhynchus Papyri, Part VII, London: Egypt Exploration Fund.
Hurtado, Larry W. (2006), The Earliest Christian Artifacts: Manuscripts and Christian Origins, Grand Rapids, Mich.: Eerdmans.
Ibscher, Hugo (1940), ‘Die Handschrift’ in Hans Jakob Polotsky and Alexander Böhlig (eds), Manichäische Handschriften der Staatlichen Museen Berlin: Kephalaia, 1. Hälfte (Lieferung 1–10), Stuttgart: Kohlhammer, v–xiv.
International Organization for Standardization (2019), Identification Cards—Physical Characteristics. <https://www.iso.org/standard/70483.html> (last accessed 30 December 2020).
Johnson, William Allen (2004), Bookrolls and Scribes in Oxyrhynchus, Toronto: University of Toronto.
Kasher, Aryeh (1981), ‘The Jewish Community of Oxyrhynchus in the Roman Period’, in Journal of Jewish Studies, 32/2: 151–157.
Kasser, Rodolphe (1962), Évangile de Matthieu XIV, 28-XXVIII, 20: Épître aux Romains I, 1–II, 3 En Sahidique, Cologny-Geneva: Bibliotheca Bodmeriana.
Koskenniemi, Heikki (1956), Studien zur Idee und Phraseologie des griechischen Briefes bis 400 n. Chr. (Suomalaisen Tiedeakatemian toimituksia Sarja B = Series B, 102/2), Helsinki: Akateeminen Kirjakauppa.
Kraus, Thomas J. (2001), ‘Bücherleihe im 4. Jh. n. Chr. P. Oxy. LXIII 4365—ein Brief auf Papyrus und die gegenseitige Leihe von apokryph gewordener Literatur’, in Biblos, 50/2: 285–296.
Kraus, Thomas J. (2010), ‘Die Welt der Miniaturbücher in der Antike und Spätantike. Prolegomena und erste methodische Annäherungen für eine Datensammlung,’ in Studien zum Neuen Testament und seiner Umwelt,35: 79–110.
Kraus, Thomas J. (2016), ‘Miniature Codices in Late Antiquity: Preliminary Remarks and Tendencies about a Specific Book Format’, in Early Christianity, 7/1: 134–152.
Lied, Liv Ingeborg and Hugo Lundhaug (eds) (2017), Snapshots of Evolving Traditions: Jewish and Christian Manuscript Culture, Textual Fluidity, and New Philology (Texte und Untersuchungen zur Geschichte der altchristlichen Literatur, 175), Berlin: de Gruyter.
Luijendijk, AnneMarie (2008), Greetings in the Lord: Early Christians and the Oxyrhynchus Papyri (Harvard Theological Studies, 60), Cambridge, Mass: Harvard University.
Luijendijk, AnneMarie (2010a), ‘Sacred Scriptures as Trash: Biblical Papyri from Oxyrhynchus’, in Vigiliae Christianae, 64: 217–254.
Luijendijk, AnneMarie (2010b), ‘A New Testament Papyrus and its Documentary Context: An Early Christian Writing Exercise from the Archive of Leonides (P. Oxy. II 209/P10)’, in Journal of Biblical Literature, 129: 575–596.
Luijendijk, AnneMarie (2011), ‘Reading the Gospel of Thomas in the Third Century: Three Oxyrhynchus Papyri and Origen’s Homilies’, in Claire Clivaz and Jean Zumstein (eds), Reading New Testament Papyri in Context / Lire des papyrus du Nouveau Testament dans leur context (Bibliotheca Ephemeridum Theologicarum Lovaniensium, 242), Leuven: Peeters, 241–267.
Luijendijk, AnneMarie (2014), Forbidden Oracles? The Gospel of the Lots of Mary (Studien und Texte zu Antike und Christentum, 89), Tübingen: Mohr Siebeck.
Luijendijk, AnneMarie (2018), ‘Books and Private Readers in Early Christian Oxyrhynchus: “A Spiritual Meadow and a Garden of Delight”’, in Karl Shuve (ed.), Books and Readers in the Pre-Modern World: Essays in Honor of Harry Gamble (Writings from the Greco-Roman World Supplement Series, 12), Atlanta: Society of Biblical Literature, 101–135.
Malik, Peter (2017), ‘P.Oxy. VIII 1080 A Fresh Edition and Textual Notes on a Miniature Codex of the Apocalypse’, in Archiv für Papyrusforschung, 63/2: 310–320
Mallon, Jean (1949), ‘Quel est le plus ancien exemple connu d’un manuscrit latin en forme de codex?’, in Emerita, 17: 1–8.
Mango, Marlia Mundell (1982), ‘Patrons and Scribes Indicated in Syriac Manuscripts, 411 to 800 AD’, in Jahrbuch der Österreichischen Byzantinistik, 32: 3–12.
Marshall, John W. (2013), ‘6 Ezra and Apocalyptic Judaism in Asia Minor’, in Eduard Iricinschi, Lance Jenott, Nicola Denzey Lewis, and Philippa Townsend (eds), Beyond the Gnostic Gospels: Studies Building on the Work of Elaine Pagels (Studien und Texte zu Antike und Christentum, 82), Tübingen: Mohr Siebeck, 427–445.
Mathieson, Erica A. (2014), Christian Women in the Greek Papyri of Egypt to 400 ce (Studia Antiqua Australiensia, 6), Turnhout: Brepols.
Mazza, Roberta (2022), ‘Narratives of Discovery: Petrie, Grenfell and Hunt, and the First Finding of the Oxyrhynchus Papyri’, Bulletin of the American Society of Papyrologists, 59: 221–258.
McKenzie, Donald F. (1999), Bibliography and the Sociology of Texts, Cambridge: Cambridge University.
Mugridge, Alan (2016), Copying Early Christian Texts: A Study of Scribal Practice (Wissenschaftliche Untersuchungen zum Neuen Testament, 362), Tübingen: Mohr Siebeck.
Nongbri, Brent (2013), ‘Losing a Curious Christian Scroll but Gaining a Curious Christian Codex: An Oxyrhynchus Papyrus of Exodus and Revelation’, in Novum Testamentum, 55: 77–88.
Nongbri, Brent (2014), ‘The Limits of Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P.Bodmer II (P66)’, in Museum Helveticum, 71: 1–35.
Nongbri, Brent (2018a), God’s Library: The Archaeology of the Earliest Christian Manuscripts, New Haven: Yale University.
Nongbri, Brent (2018b), ‘Excavating the Oxyrhynchus Philo Codex’, in Variant Readings <https://brentnongbri.com/2018/05/14/excavating-the-oxyrhynchus-philo-codex/> (last accessed 30 December 2020).
Nongbri, Brent (2019a), ‘Palaeographic Analysis of Codices from the Early Christian Period: A Point of Method’, in Journal for the Study of the New Testament, 42: 84–97.
Nongbri, Brent (2019b), ‘How Much Material from Oxyrhynchus Remains Unpublished?’, in Variant Readings <https://brentnongbri.com/2019/10/20/how-much-material-from-oxyrhynchus-remains-unpublished/> (last accessed 30 December 2020).
O’Callaghan, José (1971), ’¿Mt 2,14 en el fragment adéspota de P1?’, in Studia Papyrologica, 10: 87–92.
O’Hogan, Cillian (2021), ‘The Date of the « Oldest Codex » Reconsidered’, Scriptorium, 75: 101–107.
Orsini, Pasquale and Willy Clarysse (2012), ‘Early New Testament Manuscripts and their Dates: A Critique of Theological Palaeography’, Ephemerides Theologicae Lovanienses, 88: 443–474.
Parsons, Peter J. (2007), City of the Sharp Nosed Fish: Greek Lives in Roman Egypt, London: Weidenfeld and Nicolson.
Piotrkowski, Meron M. (2018), ‘“Literary Jews”: The Jewish Community of Oxyrhynchus in Light of Non-Documentary Texts on Papyrus’, in Meron M. Piotrkowski, Saskia Dönitz, and Geoffrey Herman (eds), Sources and Interpretation in Ancient Judaism. Studies for Tal Ilan at Sixty (Ancient Judaism and Christianity, 104), Leiden, Boston: Brill, 143–173.
Rapp, Claudia (2007), ‘Holy Texts, Holy Men and Holy Scribes: Aspects of Scriptural Holiness in Late Antiquity’, in William E. Klingshirn and Linda Safran (eds), The Early Christian Book, Washington, D.C: Catholic University of America, 194–222.
Rea, John Rowland (1996), The Oxyrhynchus Papyri, Volume LXIII, London: Egypt Exploration Society.
Roberts, Colin H. (1979), Manuscript, Society, and Belief in Early Christian Egypt (Schweich Lectures 1977), London: Oxford University.
Rudy, Kathryn Margaret (2010), ‘Dirty Books: Quantifying Patterns of Use in Medieval Manuscripts Using a Densitometer’ in Journal of Historians of Netherlandish Art, 2: 1–44.
Schofield, Ellwood M. (1936), The Papyrus Fragments of the Greek New Testament, PhD dissertation, Southern Baptist Theological Seminary.
Smith, Geoffrey S. (2023), ‘The Archaeology of Two Early Gospels: P.Oxy. 1 and 2 and the Trash Mounds of Oxyrhynchus’, in Jaimie Gunderson, Tony Keddie, and Douglas Boin (eds.), The Social Worlds of Ancient Jews and Christians: Essays in Honor of L. Michael White, Leiden: Brill, 297–316.
Turner, Eric G. (1977), The Typology of the Early Codex (The Eighteenth Publication in the Haney Foundation Series), Philadelphia: University of Pennsylvania.
Turner, Eric G. (1987), Greek Manuscripts of the Ancient World, 2nd edn, London: Institute of Classical Studies.
VanderKam, James C. (2018), Jubilees: A Commentary on the Book of Jubilees, 2 vols, Minneapolis: Fortress.
van Haelst, Joseph (1976), Catalogue des papyrus littéraires juifs et chrétiens, Paris: Publications de la Sorbonne.
Van Minnen, Peter (1994), ‘House-to-house enquiries: an interdisciplinary approach to Roman Karanis’, in Zeitschrift für Papyrologie und Epigraphik, 100: 227–251.
Wessely, Charles (1908), ‘Les plus anciens monuments du christianisme écrits sur papyrus’, in Patrologia orientalis, 4: 97–210.
Worrell, William Hoyt (1916), The Coptic Psalter in the Freer Collection, New York: Macmillan.
Worrell, William Hoyt (1923), The Coptic Manuscripts in the Freer Collection, New York: Macmillan.
Notes
1 McKenzie 1999, 4.
2 Throughout, papyrological abbreviations follow the conventions of papyri.info. For the possibility that P.Oxy. I 30 (LDAB 4472, TM 63267) was produced in the first century ce, see Mallon 1949.
3 See O’Hogan 2021.
4 A good deal of our research relies on the papyrological databases papyri.info and Trismegistos (hereafter TM), and we are greatly indebted to the scholars who compile and maintain them. On TM, see Depauw and Gheldof 2014.
5 A search of the Leuven Database of Ancient Books (hereafter LADB, https://www.trismegistos.org/ldab/) was carried out in July 2024 for the criteria ‘Bookform = codex’ and ‘Date = 1 ce to 400 ce)’. A second search added the criterion ‘Provenance = Oxyrhynchus’.
6 For an overview of the history of Oxyrhynchus and its papyri, see Parsons 2007.
7 For the excavations of Grenfell and Hunt, see their seasonal excavation reports collected in Bowman et al. 2007, 345–368. For a more critical take on the work of Grenfell and Hunt, see Mazza 2022.
8 An overview of the excavations at Oxyrhynchus after Grenfell and Hunt current up to about the turn of the millennium can be found in Bowman et al. 2007, 50–138.
9 One often finds pieces described as ‘Provenance: Oxyrhynchus (?)’ and there are surely many with this designation that come from sites other than Oxyrhynchus. But at least some pieces that travelled the markets can be connected to pieces that definitely came from Oxyrhynchus. This is the case, for instance, with P.Haun. I 8, a fragment of a bifolium from a papyrus codex containing works by Philo. Other pieces from this codex (LDAB 3540, TM 62375) were excavated on site by the teams of Grenfell and Hunt (P.Oxy. IX 1173, P.Oxy. XI 1356, P.Oxy. XVIII 2158, and P.Oxy. LXXXII 5291, all probably excavated in 1905) and later of Breccia (PSI XI 1207 excavated in 1932). See Nongbri 2018b.
10 See Nongbri 2018a, 341, note 35.
11 See the evidence gathered in Nongbri 2019b.
12 LDAB searches were carried out in July 2024. In this instance, we searched with the criteria ‘Religion = Christian’ and ‘Provenance = Oxyrhynchus’. We use these numbers with due caution, given the difficulty of determining in some cases what counts as a ‘Christian’ manuscript (rather than, say, a Jewish manuscript); see the discussion of P.Oxy. VII 1010 below. Also, we often lack secure knowledge about those pieces that were purchased on the antiquities market but attributed to Oxyrhynchus. For searches like this, the LDAB numbers are perhaps overly inclusive. For a relatively recent assessment that wrestles with some of this ambiguity, see Nongbri 2018a, 228–234.
13 Egypt Exploration Society 2019.
14 Editorial selection over the last century has likely disproportionately favored the publication of Christian literary material from Oxyrhynchus. On this phenomenon, see Bagnall 2011a,70–78.
15 We added the criterion ‘Bookform = codex’ to the earlier search (‘Religion = Christian’ and ‘Provenance = Oxyrhynchus’). Again, due caution must be exercised. It is not always easy to identify a fragment as a codex rather than a roll. See, for example, Nongbri 2013, 77–88.
16 Cockle and Henry (2019), 1. The number of the tin box that held most of these fragments indicate that they were likely found close together during Grenfell and Hunt’s first excavation season (1897).
17 For a more precise breakdown of the numbers of published manuscripts current through 2016, see Nongbri 2018a, 216–246 and 273–280.
18 For a general overview of recent developments in palaeographic dating of Greek and Coptic manuscripts of the early Christian era, see Askeland 2018 and Nongbri 2018a, 47–82. For the special set of problems associated with the palaeographic dating of early Greek and Coptic codices, see Nongbri 2019a. For a chronological overview of the Christian literary material from Oxyrhynchus, see Nongbri 2018, 273–280.
19 A search on the LDAB for ‘Religion = Christian,’ ‘Provenance = Oxyrhynchus’, and ‘Bookform = codex’ was divided into centuries. For those cases in which the LDAB assigns date ranges that cross centuries (second to third, third to fourth, etc.), we split the numbers of these groups, assigning half to the earlier century and half to the later century. When the total was odd, we added the extra to the earlier century. The resulting numbers of Christian codices from Oxyrhynchus from each century were: first century: 0, second century: 8, third century: 47, fourth century: 50, fifth century: 36, sixth century: 17, seventh century: 3, eighth century: 3.
20 P.Oxy. I 2 is generally said to have been found in the trash mound Grenfell and hunt designated K 1, but a recent study has challenged this identification and suggested that the findspot was instead K 18-19 (Smith 2023).
21 Their more exact statement of the date was as follows: ‘Though the writing is somewhat later in style than that of the ‘Logia’ [P.Oxy. I 1, LDAB 4028, TM 62838], there is no likelihood of its being subsequent to the beginning of the fourth century, and it may with greater probability be assigned to the third’ (Grenfell and Hunt 1898, 4). Grenfell and Hunt appear to have used some contextual data in addition to handwriting to arrive at this date, but this information is vague: ‘The evidence of both the handwriting and of the dated papyri with which they were found makes it certain that both the ‘Logia’ and the St. Matthew fragment were written not later than the third century’ (Grenfell 1897, 1027–1028). For a recent reevaluation of the script that reaches a similar conclusion about the date, see Orsini and Clarysse 2012, 469. For a revised view of the archaeological context, see Smith 2023.
22 Aland 1967, 107.
23 O’Callaghan 1971, 88: ‘Ahora bien, aunque los dos fragmentos se encontraron juntos, no se sigue que ambos deban pertenecer al mismo pliego. Mi opinión es que son parte de dos pliegos consecutivas’. But the two folia are conjoint, and the three lines are written along the fibers not against them, so there are multiple reasons that O’Callaghan’s proposal cannot be correct.
24 Sources that report Matt. 2:14 among the contents of the papyrus include van Haelst 1976, 124, item 332; Blumell and Wayment 2015, 31, item 5; and Mugridge 2016, 227, item 153.
25 See Turner 1977, 145 and Aland 1976, 215. Scott Charlesworth follows Aland in reporting the width as 13 cm, but he estimates a height of 24–25.5 cm. See Charlesworth 2016, 80.
26 See, for example, the end titles of the gospels in the Freer Codex (LDAB 2985, TM 61831).
27 Grenfell and Hunt 1898, 4.
28 As noted above, O’Callaghan mistook the letters as part of Matt. 2:14. Carl Wessely toyed with the possibility that another gospel had preceded Matthew, and that these stray three lines are all that remains, but he ultimately rejected the idea. See Wessely 1908, 142.
29 See, for instance, Blumell and Wayment 2015, 32: ‘It may be that this particular codex bore a title that referred to Jesus’ mother, as εγεν[ ] παρ[ ] μητ[ ](“virgin mother begot”) suggest. This must, however, remain a conjecture’. If this suggested theme is correct, it may be that the words are more likely to be a closing colophon rather than an opening title.
30 Most authors seem to assume that the piece is a quire consisting of just a single bifolium. This assumption is generally left unstated, although it is occasionally made explicit. See, for instance, Schofield 1936, 87: ‘If the order suggested is correct, then the codex consisted of quires of a single-folded sheet with recto side [that is, horizontal fibers] out’. Early papyrus codices composed of single-sheet quires are not unknown, but they seem to be very rare. For an example, see Chester Beatty Biblical Codex I (LDAB 2980, TM 61826) and the discussion in Nongbri 2018a, 138–140.
31 It is true that the horizontal fibers of this bifolium are much more abraded than the vertical fibers, which could be taken to indicate that the horizontal fibers were subject to more wear and tear during the useful life of the book and hence more likely to be the outer side of the folded bifolium. Yet, the pattern of abrasion on the horizontal fibers is quite unusual and suggestive of damage that occurred after the bifolium was crumpled and thrown out (see Fig. 2).
32 We arrive at this number by dividing the number of words in Matthew in a modern edition of the Greek New Testament (about 18346) by the number of words and restored words on the first page of P.Oxy. I 2 (about 204). Calculations like this can never be precise because of the number of unknowns: Was the text of this manuscript actuallyso similar to the modern reconstructed text? This assumption seems safe based on the surviving text, which is quite close to the text of most modern editions. The width of leaves in single-quire codices becomes smaller as one approaches the central leaves of the quire. The central leaves thus have a smaller space for writing and fewer words. Should we thus increase the number of sheets we expect? Perhaps, but copyists of single-quire codices are known to ‘cram’ near the end when they recognize they are running out of space (Turner 1977, 74). Were all of the sheets complete bifolia or were some of them one folium plus a stub, as is common in single-quire codices? We cannot know. Despite these variables, 23 bifolia seems a reasonable estimate, although it may well be that a slightly higher number (perhaps 25) could be an equally defensible estimate.
33 A fragmentary bifolium of this codex (P.Oxy. II 208) was found in Grenfell and Hunt’s first season. A portion of another leaf was published over twenty years later (P.Oxy. XV 1781). It is not clear during which season this latter fragment was excavated.
34 Turner reports dimensions for the Qau codex as 12.5 cm wide and 25.0 cm high. For the sources of our measurements, see the next note.
35 For the measurement of P.Oxy. II 208, we rely on the statement of Grenfell and Hunt 1899, 1: ‘The margin between the two columns of writing in the flattened sheet is only about 2 cm. wide’. For the measurement of Nag Hammadi Codex V, we rely on the scale in the image of pages 2+85 available at the Claremont College Digital Library, Nag Hammadi Archive. For the Qau codex, we are grateful to the Cambridge University Library for providing digital images of pages 16 and 85 with a scale for measurement.
36 It should be noted, however, that books at Oxyrhynchus seem often to have been discarded whole. See Luijendijk 2010a.
37 6 Ezra (also known as 2 Esdras, chapters 15–16) is commonly understood as aChristian work appended to the Jewish apocalypse 4 Ezra (2 Esdras, chapter 3–14). For an extensive overview of books named after Ezra, including the canonical Ezra-Nehemiah, and 1–6 Ezra, seeMarshall 2013, 427–45, esp. ‘Appendix’, 443–45. On 6 Ezra, see Bergren1998and Bergren2013, 483–97.
38 Both van Haelst and Mugridge give the content as chapter 16; it is chapter 15 (van Haelst 1976, 201, no. 574; Mugridge 2016, 284, no. 267).
39 [καὶ τὰ παιδία σου λει]μῷ διαφθα|ρήσεται καὶ | σὺ ἐν ῥομφαί|ᾳ πέσῃ καὶ αἱ | πόλεις σου συ(ν)|τριβήσονται | καὶ πάντες σου | οἱ ἐν τοῖς πεδί|οις πεσοῦν|ται ἐν μαχαί|ρῃ καὶ οἱ ἐν τοῖς | ὄρεσι καὶ μετε|ώροις || ἐν λειμῷ δια|φθαρήσονται | καὶ ἔδονται τὰ[ς] | σάρκας αὐτῶν | καὶ τὸ αἷμα αὐ|τῶν πίονται | ἀπὸ λει̣μ̣ο̣ῦ̣ ἄ̣ρ̣|του καὶ δίψης̣ | ὕδατος πρῶτα̣| μὲν̣ ἥ̣κ̣εις τά|λα[ι]να κα[ὶ π]άλι(ν)| ἐκ δευτέρου (13 λειμῷ: l. λιμῷ; 19 λειμοῦ: l. λιμοῦ).Text and translation adapted from Hunt 1910, 14. The parchment is now catalogued as Oxford, Bodleian Library MS. Gr. bibl. g. 3 (P). See also Kraus 2016, 134–36 and Kraus 2010, 105.
40 Bergren 1998, 12. See also his description and discussion of the piece; Bergren 1998, 30–31.
41 Bergren 1998, 12. Also: ‘The fragment of a Greek text of 6 Ezra 15:57–59 found at Oxyrhynchus strongly suggests that the surviving Latin version was based on a Greek Vorlage. That is, there is no evidence, nor any reasons to speculate, that the Greek version preserved at Oxyrhynchus was translated from the Latin’(Bergren 1998, 17). Hunt had already concluded that the Greek formed the Vorlage for the Latin (Hunt 1910, 12; also van Haelst 1976, 201, no. 574).
42 More generally about the situatedness of literary and documentary papyri, see Clarysse1983; Van Minnen 1994; and Piotrkowski 2018.
43 Another Greek text found at Oxyrhynchus previously known only in translation is, for instance, P.Oxy. III 403 (LDAB 3471, TM 62308), a copy of the Apocalypse of Baruch. Previously unknown Christian texts attested at Oxyrhynchus include, for example, the Gospel of Thomas in three copies: P.Oxy. I 1 (LDAB 4028, TM 62838), P.Oxy. IV 654 (LDAB 4030, TM62840) and P.Oxy. IV 655 (LDAB 4029, TM 62839); the Gospel of Mary in two copies:P.Oxy. L 3525 (LDAB 5406, TM 64187) and P.Ryl. III 463 (LDAB 4190, TM 62998); and the Martyrdom of Pamoun (P.Oxy. LXX 4759, LDAB 10692, TM 92139).
44 See Blumell and Wayment 2015; Blumell 2012; Epp 1997; Epp 2004; Epp 2007; and Luijendijk 2008.
45 For other case studies into readers of Christian texts from Oxyrhynchus, see Luijendijk 2010b; Luijendijk 2011; and Luijendijk 2018.
46 According to Thomas Kraus, it measures 5.2 cm wide and 8.3 cm high. See Kraus 2010, 105; Kraus 2016, 135. Scholars vary in the specific measurements for what they deem miniature manuscripts: Eric Turner classifies as miniatures those ‘codices which are less than 10 cm. broad, using that breadth as a conventional upper limit’ (Turner 1977, 25). Thomas Kraus (with reference to the Library of Congress) defines miniature books as measuring at most 101.6 mm or 4 inches (Kraus 2010, 80–81). Anne Bromer includes books measuring 76 mm (3 inches) or less in height and width (Bromer and Edison 2007, 11).
47 Or, as Mary Jane Cuyler pointed out to us, a credit card has the size of this book, an indication that this is a good size for a hand held object. Credit cards measure 8.56 cm wide and 5.398 cm high (3.37 inches wide and 2.125 inches high), as specified by the International Organization for Standardizationand the International Electrotechnical Commission (International Organization for Standardization 2019).
48 Bromer refers to such books as ‘one–hand books’(Bromer and Edison 2007,11). See also Luijendijk 2014, 51–56 and Rapp 2007.
49 In modern times, the name Oxyrhynchus has become almost synonymous with papyrus fragments, yet as noted above, the textual remains from the city also comprise multiple fragments of text written on animal skin, even though these are but a fraction of the overall find. A search on the LDAB for ‘Material = parchment’ and ‘Provenance = Oxyrhynchus’ gives 98 results. In contrast, a search for ‘Material = papyrus’ and ‘Provenance = Oxyrhynchus’ gives 3540 results (31 July 2024).
50 On the cost of books in antiquity, including differences in cost for papyrus and parchment codices, see Bagnall 2009, Chapter 3: ‘The Economics of Book Production’, 50–69, table on 57. In her important article on Syriac manuscripts, Marlia Mundell Mango concluded that some early (pre-640 ce), undecorated Syriac Christian manuscripts were ‘relatively inexpensive, were produced in a variety of places, and bought or owned by a middle stratum of society, that is village churches, rural monasteries and some individuals in these milieux’ (Mango 1982, 8). Thanks to Dina Boero for drawing our attention to this article.
51 See Hubai and Balog 2009, 22; Luijendijk 2014, 56. On the use of parchment trimmings, see Worrell, 1923, xi and Luijendijk 2008, 147–149.
52 As Gamble noted, ‘The first evidence of the development [from roll to codex] is found in the Epigrams of the Roman poet Martial in 84 to 86 C.E. Martial recommends that readers who wish to carry his poems on journeys should purchase “those that parchment confines in small pages” (Epigr. I. 2: quos artat brevibus membrana tabellis) so that they can be held in one hand’ (Gamble 1995, 52). Further speculation about the physical reality of the items Martial describes can be found in Nongbri 2018a, 36.
53 LDAB search for ‘Bookform = miniature and codex’; 31 July 2024. 107 (or 44%) are written in Greek; 87 (or 35.8%) in Coptic.
54 LDAB search for ‘Religion = Christian’, ‘Bookform = miniature and codex’; 31 July 2024. 27 or 73% are written on parchment; 10 or 27% on papyrus. 35 or 94.55% are in Greek; one in Coptic, one in Latin.
55 LDAB search for ‘Place = Oxyrhynchus’, ‘Bookform = miniature and codex’, with added search for ‘Religion = Christian’; 31 July 2024.
56 ‘There is a special category of ancient manuscripts that consists of miniature codices, roughly analogous to modern pocket books, and clearly produced for private reading’ (Gamble 1995, 235). Gamble also notes ‘the preponderance of Christian writings found in these small codices are apocryphal … This underscores the popular nature of the apocryphal literature by showing its use for edifying private reading, and it also shows that official efforts to control what was read privately, whether by drawing up lists or formulating a general principle, were responses to the currency, especially in private hands, of apocryphal books’ (Gamble 1995, 236).
57 In 1916, William Worrell suggested that miniature codices were left unbound because of the perceived difficulty in binding such small books (Worrell 1916, xi). However, the Kölner Mani codex (P.Köln inv. 4780; LDAB 5804, TM 64574, with pages measuring 3.5 wide and 4.5 cm high) demonstrates that even the smallest of these books could be (and almost certainly were) bound.
58 Turner 1977, 8. This was already the case in bookrolls (Johnson 2004, chapter 3, ‘Formal Characteristics of the Bookroll’, especially 130–41). The lower part of a roll or codex could get damaged from rubbing against clothes or other reading surfaces.
59 Turner 1977, 8.
60 For comparison, in P.Oxy. VIII 1080 (LDAB 2793, TM 61643), a miniature codex of Revelation, the difference in width of inner and outer margins is less pronounced.
61 On this codex, see Nongbri 2014.
62 The layout of the text on the page of the Kephalaia codex is nicely illustrated in Ibscher 1940, xiv.
63 This type of mise-en-page continues into later manuscripts; see for instance images in Rudy 2010.
64 On the aesthetics of bookrolls, see Johnson 2004, 85–160.
65 Turner 1977, 8. Turner notes that ‘it is usually the margins that have suffered most in our fragments’ (Turner 1977, 25).
66 It is also damaged on the inward facing margin, where the binding holes have torn into the page, presumably also from turning pages or opening the codex. Furthermore, a small section from the inner bottom corner has broken off entirely. The Oxyrhynchite miniature Epistle of James (PSI I 5) is overall rather battered, but has a particularly badly damaged outer margin in the midsection of the page.
67 Kasser 1962.
68 See, for instance, Boudalis 2018.
69 Kate Rudy determined how Medieval people held their books open through scientific analysis of manuscripts using a densitometer (Rudy 2010).
70 This was already the case when Hunt edited the fragment, but, as he notes, ‘only in one place (ll. 21–2) is there a real doubt about the reading’ (Hunt 1910, 11).
71 For a detailed description of the handwriting and codicology, see also Kraus 2016, 134–36. On the Severe style in manuscripts from Oxyrhynchus with mainly classical Greek works, see Del Corso 2006.
72 The letters are actually large relative to the size of the page. For a scaled comparison between the writing of P.Oxy. VII 1010 and the minute script of another codex of roughly the same size, P.Oxy. V 840 (LBAD 5831, TM 64601), see Nongbri 2018a, 241, figure 6.10.
73 Contra Mugridge, who classifies it as ‘calligraphic’, a category 1– in his scale, in which ‘Categories 1 and 2 are professional or scribal hands: Category 1 is “calligraphic” and Category 2 “secretarial” or “plain”’ (Mugridge 2016, 22 and 284–85, no. 267).
74 On the writing of educated bureaucrats, see Bagnall 2011b, 23–30. Those, however, would use more ligatures than in this page.
75 Turner 1977, 37.
76 P.Oxy. XI 1358, 44; image: Plate II.
77 On this piece, see Malik 2017.
78 PSI I 5, p. 8. See also Blumell and Wayment 2015, 179–81, no. 45. Measuring 5.9 cm wide and 7.2 cm high, this manuscript is similar in size and also in page layout, yet it is not the same manuscript.
79 Scholars have proposed diverging estimates on levels of literacy in Antiquity. Roger Bagnall argues for a kind of ‘everyday’literacy (Bagnall 2011a).
80 Eric Turner treats page numbers briefly in Turner 1977 (‘Chapter 6: The Codex and the Scribe’, especially 74–79).
81 Turner 1977, 76.
82 In the case of the Gospel of the Lots of Mary, the pagination appears also on the right hand page (in the upper right corner), but with uneven numbers; the first inscribed left hand page was not counted; see Luijendijk 2014, 43.
83 There may never have been a page number, or it may have been abraded (Hunt1910, 13). The latter seems less likely given the overall preservation of the page, where ink traces are visible throughout (this is based on examination of a digital image).
84 Hunt already wrestled with this, asking: ‘Does this refer to the fortieth leaf or the fortieth page?’ (Hunt 1910, 12–13).
85 ‘Manuscript No. 1 must have been, when complete, a volume of seven hundred pages of about 2 ¾ x 3 ⅛ inches’ (Worrell 1923, xi). The highest preserved page number in the manuscript is 162 (Worrell 1916, 67). Further pages survive (though the numbering does not) up to a presumed page 258 (Worrell 1916, 106).
86 See also Bergren: ‘Of greater moment is the implication of this pagination for the form in which 6 Ezra circulated in this manuscript. It is impossible, given the pagination of the fragment, that the far longer 4 Ezra preceded 6 Ezra in this manuscript … Thus it is almost certain that, in the form of the Greek text of 6 Ezra preserved at Oxyrhynchus, 4 Ezra did not precede 6 Ezra. This contravenes the theory, generally accepted in Hunt’s time and promoted even by many modern scholars, that 6 Ezra was written as an appendix to the Greek version of 4 Ezra’(Bergren 1998, 31).
87 The Latin translation has 43 words in this section; however, as Bergren also notes, at two places there are doublets in the Greek that are absent in the Latin (Bergren 1998, 58). If we translate the Latin to Greek with the doublets, the number of words is identical in both languages in this small section. Overall, however, one should take into account the fluidity of works like these in a Late Antique context (Lied and Lundhaug 2017).
88 Without the benefit of electronic editions through which word counts are easy, Hunt used the number of lines in the Latin translation to calculate whether mu indicated page 40 or leaf 40, as follows: The two verses of Greek of P.Oxy. VII 1010 equal four lines in the Latin; the 56 preceding verses take up 96 lines in the Latin. If calculating by leaves, Hunt counted 24 preceding leaves (96:4), if by pages, 48 preceding pages (Hunt 1910, 13). The results from this method appear less precise, even though either way of estimating the size of the codex gives just an approximation, as Hunt also noted.
89 So already Hunt 1910, 13, Bergren 1998, 31.
90 Based on the text of the Vulgate available on the LLT-A (Brepols); last consulted on December 22, 2020.
91 See, for instance, the publications by Bergren mentioned above, and the discussion of previous scholarship by Marshall 2013, 431–436.
92 Marshall 2013, 442. Also: ‘the reasons for regarding [6 Ezra] as a Christian composition are inadequate. Rather, understood as a Jewish text, 6 Ezra offers insight into the variety of Judaism in Asia Minor and the reception history of the Book of Revelation in that context’ (Marshall 2013, 427).
93 ‘We receive 6 Ezra as preserved, copied, translated, and published by Christians and used by scholars as evidence for ancient Christianity’ (Marshall 2013, 429).
94 Kasher 1981; Epp 2006; and Piotrkowski, 2018.
95 On nomina sacra, see, for instance, Hurtado 2006, 93–134; Luijendijk 2008, 57–78; and Blumell 2012, 49–51.
96 Roberts 1979, 10–12, quotation at 12.
97 See, for example, the following pieces: P.Ant. II 84 (Isocrates, [8.8 x10 cm], parchment); P.Oxy. VIII 1096 (Isocrates, [9] x 10 cm, parchment); P.Oxy. LXIX 4717 (Isocrates, 7.5 x 10.2 cm, parchment); P.Oxy. LXXVIII 5133 (Isocrates, 7.5 cm w and 9.8 cm high, parchment); P.Oxy. LXXXIV 5420 (Apollonius Rhodius, about 8 cm w, height of written column 8.1 cm, parchment).
98 See Feldman 2022.
99 ‘Thus, although the Sitz im Leben of our papyri listed above is probably best sought with Oxyrhynchan Christians, it remains notoriously difficult to conclude with utmost certainty that all of these papyri, indeed, have a Christian background … These texts, after all, are genuine Jewish compositions and Oxyrhynchan Christians must have obtained them from somewhere. The fact that they were copied “mainly” by Christians, so it seems, points to cordial relations between Jews and at least some Christians in Oxyrhynchus. But it also shows that these texts were read and probably also continued to be read by local Jews’ (Piotrkowski 2018, 153).
100 Or, it is also possible that some manuscripts, and hence the texts they carried, became Christian along with Jewish owners who became Christian.
101 For an ambiguous case of another sort at Oxyrhynchus, see P.Oxy. XLI 2944 (LDAB 4502, TM 63296), ‘Peri apophaseon,’ a story of Solomon’s judgement, in the form of a miniature roll. It is difficult to know if it was produced and used by Jews or a Jewish story that had been taken over into a non-Jewish environment.
102 Kraus 2016, 136.
103 Kraus did mention the Ezra miniature codex in his earlier article on the letter: ‘Wegen eines ebenso in Oxyrhynchos zu Tage geförderten Pergamentblattes eines Miniaturcodex … (P.Oxy. VII 1010) mit griechischem Rest von IV Esra und der Verwendung ebendieser Bezeichnung in einer Bücherliste aus dem siebten oder achten Jahrhundert … mag der Schluß naheliegen, es handele sich in Z. 3 des Briefes um IV Esra’ (Kraus 2001, 287).
104 Epp 2004, 34. As Blumell and Wayment comment: The presence of P.Oxy. VII 1010, the Ezra page, ‘should at the very least serve to increase the probability that the Ezra under question was a copy of one of the many pseudepigraphical works circulating under his name and not the book of Ezra from the OT’ (Blumell and Wayment 2015, 511 n. 3). More open to a connection between the two manuscripts is Luijendijk 2018, 109–10.
105 Rea dates the petition to the third or fourth century, more specifically to the second half of the third century (Rea 1996, 43, note to line 7).
106 ‘The date range [of the petition on the recto] raises the possibility that the letter on the back dates from before about 325’ (Rea 1996, 43, note to line 7). Nick Gonis has shown that the vocabulary in line 1 of the letter, P.Oxy. LXIII 4365, suggests a fourth-century date, for ‘in no other private letter from the first three centuries of Roman rule in Egypt does φίλτατος qualify ἀδελφός’ (Gonis, 1997, 150). According to Gonis, the construction in this note ‘would have seemed intolerable in earlier times’ (Gonis 1997, 148). When Koskenniemi published his book on Greek epistolography in 1956, he could state that the adjective was used predominantly for men in correspondence: ‘Eine nähere Untersuchung zeigt, dass es fast ausschliesslich im Verkehr zwischen Männern gebraucht wird’ (Koskenniemi 1956, 97).
107 τῇ κυρίᾳ μου φιλτάτῃ ἀδελφῇ ἐν κ(υρί)ῳ χαίρειν. χ̣ρ̣ῆσον τὸν̣ Ἔσδραν, ἐ̣π̣εὶ ἔχρη̣σά σοι τὴν λεπτὴν Γένεσιν. ἔρρωσο ἡμεῖν ἐν θ(ε)ῷ. For an overview of scholarship on this papyrus, see Blumell and Wayment 2015, 509–12, no. 141. In addition, see Mathieson 2014, 219–20; and Luijendijk 2018, 107–12.
108 Bagnall and Cribiore rejected the identity of the petitioner’s hand with that of the letter writer’s (“We do not believe that the hands [of P.Oxy. LXIII 4364 and 4365] are the same, but it remains perfectly possible that 4365 was written by a woman”); Bagnall and Cribiore 2008, chapter 6, n. 1. See also Luijendijk 2008, 73 n. 56; Luijendijk 2018, 108 n. 20. In the editio princeps, John Rea argued that the letter’s handwriting was the same as that of the document’s subscription on the recto (P.Oxy. LXVIII 4364), by a woman called Aurelia Soteira alias Hesychium (Rea 1996, 44). So also Fournet 2012, 142. Consequently, Epp proposed that these correspondents are two female leaders in the Oxyrhynchite church (Epp, 2004, 31–35). According to Blumell, women more frequently correspond among themselves (Blumell 2012, 262–63 and 262–63 n. 114).
109 Currently, no Greek manuscripts of the book of Jubilees exist, making this note written in Greek an important, even the earliest, reference to this lost version; see VanderKam 2018, vol. 1.10–1.11.
110 See Luijendijk 2018, 109.
111 The text is phrased in the first person singular with a plural ‘farewell from us in God’ at the end presumably referring to the sender’s household. See also Luijendijk 2018, 108 n. 22.
112 E.g., Hermas, Vision 2,5; John Chrysostom, Stat. 19.14; Hom. Matt. 72; Isidore of Pelusium, Epistula 150.2 (ad Epimachum).
113 From today’s standpoint, it seems incredible that Lobel dedicated so little discussion to the remarkable P.Oxy. XX 2258 (LDAB 523, TM 59424), a codex containing works of Callimachus and commentary copied on leaves that Turner estimated to be 37 cm wide and 28 cm high (Turner 1977, 14). The codex is discussed further in Turner 1987 (item 47).
114 We are grateful to the organizers of the 2016-17 Mellon Sawyer Seminar, Cultural and Textual Exchanges: The Manuscript Across Premodern Eurasia hosted at the University of Iowa. The participants offered useful discussion and feedback. Thanks also to the sharp-eyed peer reviewer at ISAW Papers who saved us from numerous slips and offered several helpful suggestions. Nongbri’s research was funded in part by the Norwegian Research Council, project number 314240, The Early History of the Codex: A New Methodology and Ethics for Manuscript Studies (EthiCodex). This essay was written in 2020; the statistics were updated in July 2024, and the essay was lightly revised at that time.