The Color(s) of Papyrus and Pliny’s Instructions

Posted on December 31, 2023 by Brent Nongbri

A few months ago, I observed that two stalks of papyrus from the same plant could produce quite different colored sheets of papyrus. I’ve also noticed that the same stalk of papyrus can produce strips that, although they look the same when freshly cut, have different colors when dry. I wondered whether this color difference might relate to the area of the stalk from which the strips come (the top end or the root end).

I figured this was easy enough to test using a stalk from one of the papyrus plants I grew from seeds earlier this year. The stalks on my plants are quite skinny, but I’m still able to make small sheets from them. So, I cut the stalk into sections and then made strips from them. When freshly cut and shorn of its green husk, the whole stalk has a translucent grey-ish white color. But after a few minutes in the air, the portion nearer the root began to turn a reddish color. This effect diminished as you moved up the stalk toward the top, where a lighter yellow color developed.

The different colors of a papyrus stalk after a few minutes of exposure to air; lighter near the top, darker near the root

This pattern of coloration became more pronounced as the strips dried. The reddish area at the bottom dried to a fairly dark brown color after 24 hours. Strips from the upper part dried to a much lighter color. The image below shows small sheets made from middle strips on the left and lower strips on the right (the upper-most strips are simply too narrow to use effectively). The color difference is fairly stark:

Papyrus sheet made from strips from the middle of the stalk (left); papyrus sheet made from strips from the lower portion of the stalk (right)

I tried this on three stalks from the same root cluster and had similar results. I also observed a slightly less pronounced version of this phenomenon on strips cut from larger stalks from the local botanical garden.

In general, the upper portions produce strips of a much lighter color, but the strips become very thin due to the narrowing of the stalk near the top. At the lower portion of the stalk, the strips are broad, but the color of the papyrus becomes quite dark when exposed to air. For the purpose of making usable papyrus sheets, it’s really the portion of the stalk in the middle–where the breadth is decent and the color is still light–that is the best part.

This observation potentially sheds some light on a passage in Pliny’s famous description of the manufacture of papyrus (Natural History, book 13). Pliny writes the following (according to the text and translation in Lewis, Papyrus in Classical Antiquity):

Praeparatur ex eo charta diviso acu in praetenues sed quam latissimas philyras; principatus medio, atque inde scissurae ordine.

Paper is made from the papyrus plant by separating it with a needle point into very thin strips as broad as possible. The choice quality comes from the centre, and thence in the order of slicing.

As Lewis notes, there are several puzzles in this short passage, but for now I’m interested in the phrase principatus medio. Lewis understands medio as the center, or core portion, of the pith. But if my observations about the color differences of the pith up and down the stalk are right, then it seems like medium could equally refer to the middle part of the length of the stalk, between top and the root, where the strips are lighter color but still relatively wide. And in fact, already in 1976, Adam Bülow-Jacobsen argued for this interpretation of medio on other grounds:

“I am convinced that medium must be taken quite plainly to mean the part of the plant which is mid-way between the top and the root, and there is good physical reason to understand it thus: All fibers in a papyrus stem run all the way from bottom to top and a cross-section at the top and at the bottom of the same stem will reveal the same number of vascular bundles in both, although the diameter of the stem is much smaller at the top than at the bottom. In terms of strips cut along the fibers this means that a strip from near the bottom of the plant will seem to contain fewer and thicker fibers, which will stand out like wires once the pith between them has been hammered out flat and dried. A strip from near the top of the plant will have great concentration of fibers with little pith between them and will be too narrow for convenience.”1

Bülow-Jacobsen notes that other experimenters came to different conclusions, but the argument from the differences in color in the different levels of the stalk would seem to support his analysis and exegesis of Pliny’s text.

  1. Adam Bülow-Jacobsen, “principatus medio: Pliny, N.H. XIII, 72 sqq.,” Zeitschrift für Papyrologie und Epigraphik 20 (1976) 113-116. ↩︎
Posted in Papyrus Making | Tagged , , , | 7 Comments

The Robinson Papyri, and the Mississippi Papyri, and William H. Willis

Posted on November 22, 2023 by Brent Nongbri

Several years ago, I tried to sort out some of the confusion that surrounds the so-called Robinson Papyri, a collection of papyrus manuscripts accumulated by the archaeologist David M. Robinson (1880-1958) and eventually inherited by the classicist William H. Willis (1916-2000). The confusion results from seemingly conflicting descriptions of the collection published by Willis, and the close but unclear relationship between the Robinson Papyri and two other collections, the Mississippi Papyri and the Deaton Papyri. I had first encountered the Robinson Papyri because of my interest in still another intersecting collection, the Bodmer Papyri. Figuring out the Robinson Papyri became more urgent in 2015, when Dirk Obbink claimed that the Robinson Papyri were the source of the now infamous Sappho papyrus that he published (“P.Sapph.Obbink”).

David M. Robinson and William H. Willis

This series of posts (here, here, and here) raised a number of questions. Many of them have now been answered by an informative and thorough new article:

Daniel B. Sharp, “The Provenance of the Robinson and Mississippi Papyri,” Archiv für Papyrusforschung 69 (2023) 162-192

Sharp’s detailed archival research at the University of Mississippi, Duke University, and elsewhere has uncovered some of the reasons for the confusion within these collections. What all these collections have in common is the figure of Willis, who had physical custody of the Robinson Papyri (which he owned through inheritance), the Mississippi Papyri, the Deaton Papyri, and a few scraps of Bodmer Papyri. Sharp demonstrates quite clearly that Willis allowed these groups to intermingle in different ways. This matters because Willis owned the Robinson Papyri but not the other collections. It is thus surprising to learn that some of the Mississippi Papyri and Deaton Papyri became Robinson Papyri and were donated to Duke as the personal property of Willis. As Sharp notes in the case of P.Deaton 28, which was donated to Duke by John Deaton:

“It is impossible to know with certainty why Willis decided to claim personal ownership of P.Deaton 28 rather than donate it to Duke in Deaton’s name. What can be documented is that Willis had P.Deaton 28 appraised for $12,000 as part of his donation to Duke and that he claimed those donations as tax write-offs.”

Willis’s numerous and important contributions to classical studies and papyrology are well known. Sharp’s research adds a new dimension to Willis’s academic legacy and sheds light on the entanglement of papyrologists and the antiquities trade in the twentieth century.

Posted in Antiquities Dealers and Collectors, Antiquities Market, Dirk Obbink, Duke Papyri, P.Sapph. Obbink, Robinson Papyri | 5 Comments

P.Oxy. 87.5575 and P.Oxy. 60.4009: The Same Copyist

Posted on September 30, 2023 by Brent Nongbri

The editors of P.Oxy. 87.5575, the recently published papyrus fragment with a collection of sayings of Jesus, stated that P.Oxy. 60.4009, another papyrus with material about Jesus, “may well be in the same hand, though the loops in that papyrus are sometimes more pronounced and there is perhaps less lifting of the pen.” In preparation for a meeting of the North American Society for the Study of Christian Apocryphal Literature yesterday, I spent some time with images of the two papyri and managed to convince myself that these pieces were in fact copied by the same person.

P.Oxy. 87.5575 and P.Oxy. 60.4009, at basically the same scale; image sources: Evangelical Textual Criticism and Egypt Exploration Society / Oxford University

I’m often critical of efforts to assign precise dates on the basis of handwriting to undated manuscripts of Greek, Latin, and Coptic writing of the Roman era. The comparisons involved in that process depend on a number of assumptions that I’ve discussed elsewhere on multiple occasions. The comparisons made in palaeographic dating are always a matter of more similar and less similar, and there is always room for debate. When the question has to do with identifying the work of a particular copyist, however, the task is sometimes easier. I say “sometimes” because when it comes to expert execution of highly formal scripts, it really can be difficult to tell the work of one copyist from another (think of the debates around the number of copyists involved in the production of Codex Vaticanus). But when the scripts involved are less formal and the degree of execution more, we might say, relaxed, the process of identifying the work of a single copyist in multiple manuscripts can be more straightforward. It’s a matter of saying, “Look at this combination of idiosyncrasies. These two samples of handwriting were produced by the same person.”

So, here’s a side-by-side comparison of what I would consider a set of idiosyncrasies sufficient to identify the work of a single copyist in P.Oxy. 87.5575 and P.Oxy. 60.4009:

The mu is formed with no lifting of the stylus and with the belly sitting on the lower notional line. The crossbar of the epsilon extends to the top of the first vertical stroke in the nu. The formation of the ξ in both pieces is virtually identical, made without lifting the stylus. The upper right quadrant of the omicron in both pieces is occasionally flattened. The epsilon-iota combination is regularly produced with the iota simply trailing off the middle bar of the epsilon and extending below the lower notional line. The oblique strokes of the kappa in και extend out in an almost horizontal fashion and bend down slightly toward the lower loop of the alpha.

Each of these individual features could be paralleled pretty easily in other papyri, but the combination of all of them in both pieces and the level of graphic similarity on display here make me comfortable saying these two fragments were the work of the same copyist. There are occasionally some slight differences in the formation of some letters, but these all fall within what I would call a normal level of variation for any copyist.

There is at least one other example among the Christian material at Oxyrhynchus in which we find the same copyist responsible for two manuscripts. AnneMarie Luijendijk pointed out to me several years ago that P.Oxy. 8.1078 (Hebrews) and P.Oxy. 6.850 (Acts of John) were likely written by the same person. Arthur Hunt had actually made the identification in his edition of P.Oxy. 8.1078, but I had not noted the reference. It is also worth recalling the recently published set of Septuagint papyri, P.Oxy. 84.5404-5408, which seem to have been copied by the same person but apparently were part of different codices.

I would be even more interested to see if it is possible to identify the work of a single copyist who was responsible for both Christian and non-Christian papyri.

Posted in Oxyrhynchus Papyri, Palaeography | 4 Comments

More Details on the Possible Codex at Graz

Posted on September 12, 2023 by Brent Nongbri

Earlier this year, I noted the news that we may have a new candidate for the earliest surviving portion of a codex, P.Hib. 113, a papyrus excavated from the Egyptian town of Hibeh and now kept at the University of Graz. Theresa Zammit Lupi and the team at Graz have now produced a much more detailed study of the papyrus, which is freely available here.

I have not yet had the opportunity to read the paper carefully and think it through, but a quick scan shows that it is well illustrated and promises to make further study of this artifact much more informed. I look forward to the ongoing discussion.

Posted in Book binding, Codices, Codicology, Mummy cartonnage, P.Hib. 113 | 1 Comment

The Date of the New Oxyrhynchus Sayings of Jesus P.Oxy. 87.5575

Posted on September 9, 2023 by Brent Nongbri

The publication of the latest volume of The Oxyrhynchus Papyri has been in the news. Among the newly published pieces is a small fragment of a leaf of a papyrus codex that contains a previously unknown collection of sayings of Jesus. Candida Moss has a nice summary and analysis in her recent column. I have now read the edition of the papyrus, and I offer a few first impressions. I see that I first encountered this papyrus (part of it, anyway) several years ago when it emerged from Jerry Pattengale’s pocket during a lecture in 2011.

Jerry Pattengale with a stolen Oxyrhynchus papyrus during a lecture in the “Passages” series in 2011

On that occasion, Pattengale described the papyrus as a copy of “the end of Matthew 6” that had been “dated to about 140 to 160.” Pattengale continued, elaborating on how the date was established: “It’s early. And you have in the room a couple people, um, that can do that. And then Dr. Obbink as well.” The reference is to Professor Dirk Obbink, who allegedly stole this and many other papyri from the Oxyrhynchus collection and sold them to the Green Collection and other buyers. So, these stolen papyri that were returned to the Egypt Exploration Society are now being published.

In his lecture, Pattengale stressed the quite early dating of the papyrus in order to emphasize the “reliability” of the New Testament documents:

“My friends, this is 200 years earlier than a lot of the texts that are much in the sensational news today. This is part of that list that supports the canon.”

Not exactly. Now that the papyrus has been more thoroughly studied, it turns out that it is not a copy of the Gospel according to Matthew, but rather a previously unknown collection of sayings of Jesus that has some similarities with material found in Matthew and Luke and the Gospel of Thomas. In other words, it’s a non-canonical text about Jesus, which, if the dating is correct (and that’s a significant “if”), is earlier than almost all surviving copies of anything in the New Testament. In the new publication, the editors assign the copying of the fragment more broadly to the second century.

So, there is quite a bit of excitement around the fragment. But it is challenging to assign a date to a papyrus like this. We have no criteria to judge the date except for the handwriting on the fragment, which can sometimes not be a very reliable guide, as I have discussed on many occasions. Good images of the papyrus as published are not yet available. Peter Gurry at Evangelical Textual Criticism has posted these images of the ex-Green Collection fragment (the published edition contains another small fragment that was identified in the collection at Oxford):

The ex-Green Collection portion of P.Oxy. 87.5575, horizontal fibers (→); image source: Evangelical Textual Criticism
The ex-Green Collection portion of P.Oxy. 87.5575, vertical fibers (↓); image source: Evangelical Textual Criticism

The script falls into Turner’s rather broad “informal round” classification and is “only approximately bilinear,” meaning that the letters do not always stay between upper and lower notional lines. The editors note that the letters show a slight slope to the left (\) and have “several cursive elements.” They offer four securely dated manuscripts that they regard as having a similar script. All the pieces have dates in the second half of the second century:

CPG II. 1 App. 1178 CE
P.CtYBR inv. 685157-160 CE or 180-188 CE
P.Oxy. 36.2761161-169 CE
C. Pap. Gr. II.1 63185 CE

The editors also note the very close similarity of script with another collection of the sayings of Jesus, P.Oxy. 60.4009, even suggesting the possibility that the same copyist was responsible for both manuscripts, stating that the two “may well be in the same hand.”

P.Oxy. 60.4009, horizontal fibers (→); image source: Egypt Exploration Society, University of Oxford
P.Oxy. 60.4009, vertical fibers (↓); image source: Egypt Exploration Society, University of Oxford

The scripts are quite similar. It is interesting, then, that the editors of P.Oxy. 60.4009 offered a different set of dated samples to justify their dating of 4009. These samples cluster in the first half of the second century rather than the second half:

Schubart, Pal. Abb. 8181 CE
Norsa, Scritt. Doc. XVc133-136 CE
Schubart, PGB 22b135 CE (?)
Schubart, PGB 24148 CE

If we agree to the basic assumption of palaeographic dating (similar visual appearance of scripts = similar dates of production), then the evaluation of these claims means having a close look at the suggested comparative evidence and seeing how similar the samples actually are. To facilitate that process, I gather here links to the images of the relevant dated manuscripts that have featured in the discussion so far.

CPG II. 1 App. 1, a report of an accidental death copied in 178 CE:

CPG II. 1 App. 1; image source: Egypt Exploration Society, University of Oxford

P. CtYBR inv. 685, a lease for a house copied in the period 157-160 CE or 180-188 CE

P.CtYBR inv. 685; image source: Yale University Library Digital Collections

P.Oxy. 36.2761, report of a death copied in the period 161-169 CE

P.Oxy. 36.2761; image source: Egypt Exploration Society, University of Oxford

C. Pap. Gr. II.1 63 (=P.Petaus 8=P.Köln inv. 388), a report of a death copied in 185 CE

P.Köln inv. 338; image source: Kölner Papyri

The editors of P.Oxy. 60.4009 suggested the following script samples as useful comparanda:

Schubart, Pal. Abb. 81 (=P.Lond. 1.130), horoscopes copied after 13 September 81 CE

P.Lond. 1.130; image source: The British Library

Norsa, Scritt. Doc. XVc (=PSI 5.446), an edict of Petronius Mamertinus copied during 133-136 CE

PSI 5.446; image source: PSI Online

Schubart, PGB 22b ( = BGU 1.136 = P.Berol. inv. 6855), a copy. ofthe proceedings of a trial that took place in 135 CE

P.Berol. inv. 6855; image source: Berliner Papyrus datenbank

Schubart, PGB 24 (=BGU 1.300=P.Berol. inv. 6849), a transfer of power of attorney copied in 148 CE

P.Berol. inv. 6849; image source: Berliner Papyrusdatenbank

I tend to agree with the editors about the similarity of the scripts of 5575 and 4009, but in my first look at the proposals for dated parallels (for both the pieces), I cannot say that I find any of them especially compelling. This is not to criticize the work of the editors. It is very difficult to find good, securely dated comparanda for scripts like these. A more detailed evaluation will have to wait for another occasion.

Before finishing this post, I should also point out that the editors of P.Oxy. 87.5575 state that “P. Orsini, cited by Trismegistos, has placed 4009 in the first half of the fourth century, but we have not found evidence to support such a dating for either of these papyri.” The reference is to the noted palaeographer Pasquale Orsini, whose assessment appears on the Leuven Database of Ancient Books. When this fourth century date was published on the Database several years ago, I asked Professor Orisini about the evidence for this new assessment, and he provided a list of several other undated literary papyri. So a full and convincing argument in favor a fourth century date has yet to be made.

Posted in Antiquities Market, Dirk Obbink, Green Collection, Oxyrhynchus Papyri, Palaeography, Passages Speakers Series | 9 Comments

Notes on a Forger’s Methods

Posted on August 12, 2023 by Brent Nongbri

John de Monins Johnson (1882-1956) was in some ways a kind of successor to Grenfell and Hunt for a short period in the early twentieth century. Before taking up a position at Oxford University Press, he had training as a papyrologist under Hunt’s supervision. After Grenfell and Hunt stopped excavating for papyri in Egypt, Johnson led expeditions to Greek and Roman sites for the Egypt Exploration Fund for a number of years. In the winter of 1913-1914, he led excavations in Egypt at Antinoopolis, and his reports from the dig make for interesting reading. He was more meticulous about record keeping than Grenfell and Hunt, and he also made some interesting digressions in his reports. My attention was drawn especially to his account of the discovery of large amounts of blank papyrus in the rubbish mounds of Antinoopolis: “[Mound] N was remarkable throughout for what is often a feature of late mounds, the quantity of blank papyrus it provided. In this case sacks might have been filled.”

I was surprised to read Johnson’s description of what they did with all this blank papyrus and why they did it:

“This blank papyrus is carefully torn into small fragments before being given to the winds so that material shall not be added to the store of the papyrus forger. At the risk of a further digression a note may perhaps here be inserted on the methods of the latter. Most papyrologists are familiar with his work, fewer know the forger himself. Hufuta, an uneducated fellah of the Fayum, formerly of Hawara but now living and plying his trade in the Medina, is an amusing and not wholly unattractive character. In stature and appearance he resembles somewhat his notorious countertype, Islam Akhun, the forger of central Asian books, and his work, like Islam Akhun’s, has won the distinction of being published with facsimile in a serious book. His methods are roughly threefold. Blank sheets of papyrus from the mounds, cut into uniform pages and inscribed in red or black ink with a sequence of meaningless signs resembling somewhat pothooks and hangers, are sewn together and bound in thicker sheets of papyrus which are covered with mummy-cloth. The second method is similar but relies on skin in imitation of vellum as its medium, this being bound in skin with strips of mummy cloth or even with an elaborate arrangement of copper corners linked with wire and beads. By his third method, perhaps a more deceptive one, numerous tiny fragments of genuine inscribed papyrus, too small to find a sale, are glued together till they form a sheet, a strange literary mosaic in which the lines are incontinuous and Ptolemaic rubs shoulders with Byzantine, and this is then tightly rolled. The smell of the oil in which his productions are soaked before being buried is often a sufficient test for those who do not happen to be scholars.”

Items like Hufuta’s third type of forgery are pretty common in museums with Egyptological collections. Here is an example from the National Museum of Archaeology in Dublin:

A fake papyrus roll assembled from ancient materials, on display at the National Museum of Archaeology in Dublin; image source: Brent Nongbri 2022

The papyrus fragments in such rolls can be genuinely ancient, but they also sometimes contain ancient papyrus with modern writing. And the “rolls” themselves are modern productions. The World Museum in Liverpool offers another example:

A fake papyrus roll assembled from ancient materials; image source: National Museums Liverpool, World Museum

A similar-looking roll was the vehicle for the published fake that Johnson mentions in the quotation above, P.Stras. 1 39. Apologies for the quality of the scan of the edition:

One of the pieces removed from the roll, P.Stras. 1 39A, is a nice example of the gibberish symbols that characterize fake papyri that have ended up in a number of collections over the years:

For a working list of known fake papyri from Egypt, see the helpful spreadsheet of Macquarie University’s “Forging Antiquity” project: http://www.forgingantiquity.com/forgeries.

Sources:

John de Monins Johnson, “Antinoë and its Papyri: Excavation by the Graeco-Roman Branch, 1913-14,”
Journal of Egyptian Archaeology 1914, 168-181.

Friedrich Preisigke, Griechische Papyrus der Kaiserlichen Universitäts- und Landesbibliothek zu Strassburg (Leipzig: Hinrichs, 1912).

Posted in Antiquities Market, Fakes and Forgeries | 6 Comments

Is “Papyrus (Roll) Maker” (chartopoios) an Ancient Word?

Posted on July 23, 2023 by Brent Nongbri

In literature about the process of papyrus production in antiquity, the word χαρτοποιός (“papyrus maker” or “papyrus roll manufacturer”) comes up with some frequency. What’s odd is that its role in the discussion is out of proportion to its attestation. The term appears in only two search results on papyri.info, and both have some problems. The first instance is P.Tebt. 1 5, a list of edicts of Euergetes II from the year 118 BCE. The relevant section is a list of different types of workers who are exempt from a requirement–textile workers (wool weavers and others), pig herders, goose keepers, makers of oil, beekeepers, and brewers. The relevant lines are here in the form they appear on papyri.info:

…καὶ τανυφά[ντας πάντ]α̣ς καὶ τοὺς ὑοφορβοὺς
καὶ χηνοβο(σκοὺς) κ[αὶ χαρτοποιοὺ]ς(*) καὶ ἐλαιουργοὺς καὶ
κικιουργοὺς καὶ με[λισσουργο]ὺς καὶ ζυτοποιοὺς…

The asterisk indicates that the reading given in this line on papyri.info differs from what the original editors (Grenfell and Hunt) proposed, which was this:

καὶ χηνοβο(σκοὺς) κ[αὶ. . . . . . . . . . ]ς καὶ ἐλαιουργοὺς καὶ

In support of the restoration of [χαρτοποιού]ς, the Berichtigungsliste gives references that lead ultimately to a single Russian work that I have been unable to track down. The reference is given in the following forms:

  • W. G. Borukhovich, Papirusnye svidetelstva ob organizacii proizvodstva i prodaje charty ν Egiptie vremeni Ptolemeev. Problemy socialno-ekonomicheskoy istorii Drevnego mira. Sbornik pamyati akademika A. I. Tiumeneva = Papyrus-testimonies concerning the organization of the production and sale of charta in Egypt in the times of the Ptolemies. The problems of the social and economic history of the Ancient World. A collection of essays in memory of the academician A. I. Tiumenev. Moscow-Leningrad, 1963, pp. 271-287.
  • D.G. Boruknovič, Papyrologische Quellen zur Organisation der Herstellung und des Handels mit Papyrus im ptolem. Ägypten, Gedächtnisschrift für A.J. Tjumenev, 1963, S. 271 ff.

(If anyone who has access to the chapter and reads Russian wants to offer a translation of the relevant bits, that would be great!)

In any event, a quick look at an image of the papyrus suggests that Grenfell and Hunt’s more conservative transcription is probably safer, as there really is just a light trace of (maybe) a final sigma and nothing else. The dotted line is where our missing word would be:

P.Tebt. 1 5, with dotted line showing the space where [χαρτοποιού]ς is suggested to fill the lacuna; image adapted from Berkeley Library Digital Collections

The second occurrence is in another papyrus from Tebtunis, P.Tebt. 1 112 (=P.Tebt. 5 1151), an account of income and expenses. We have a little more to work with here in the relevant lines:

Ἁρφαήσει μαχί(μωι) ὁμοίως τι(μῆς) χαρτῶν εἰς συ(μ)-
πλήρωσι<ν> τῶν διαγεγρ(αμμένων) τῶι χαρτ[ο]πο(ιῶι) Γω

We clearly do have a χαρτ- word here, and the context, in which expenditures for other writing materials is discussed, lends itself to a word meaning “seller of papyrus,” but not so much “maker of papyrus.” And I hate to disagree with Grenfell and Hunt, but I cannot convince myself there is a pi and an omicron in where they see it after the lacuna in χαρτ[ο]πο(ιῶι) Γω:

P.Tebt. 1 112, showing the hole and ink traces where Grenfell and Hunt read χαρτ[ο]πο(ιῶι); image adapted from Berkeley Library Digital Collections

I don’t know of any other proposed restorations here. Ulrich Wilcken (Grundzüge, 1.1.255) resolved the abbreviation differently, proposing χαρτοπό(ληι) = χαρτοπώ(ληι), “papyrus seller,” but this word is just as poorly attested (and doesn’t solve the pi–omicron problem). Another thing that bothers me here is the meaning of διαγεγρ(αμμένων) (I’m not sure why papyri.info has διαγεγρ(αμμαμένων) with an extra syllable here). In his re-edition of the papyrus, Verhoogt translates as follows:

“To Harphaesis, armed guard, ditto, as the price of rolls of papyrus for payment in full of the 3,800 paid to the papyrus-maker,

Verhoogt presumably takes διαγεγρ(αμμένων) as a synonym of διαγρ(αφή) (in this context meaning “payment”), which occurs a few lines later in reference to papyrus rolls. But is that the best way to handle this? As Verhoogt notes, the other papyrus rolls mentioned in this account are specified as ἄγρ(αφος), “blank,” “unused,” or “uninscribed.῾ It stands to reason that in such a context, διαγεγρ(αμμένων) would mean the opposite: “used” or “inscribed” (although I don’t know of any parallel usages of this word in this way). If this is the correct understanding, it would raise the question of whether we should imagine buying used papyrus rolls from a “papyrus maker” or a “papyrus seller.”

In any event, I don’t spend a lot of time reading Ptolemaic documents, so I don’t really have any insight into how that gap might be filled in an intelligible way, but given its lack of existence elsewhere in the papyrological record (so far), and the surviving ink traces here, I’m a little skeptical of χαρτοποιῶι.

In an article in ZPE from 2016, Menico Caroli added two more potential candidates for possible occurrences of χαρτοποιός, P.Wisc. 1 29 and P.Flor. 3 388. P.Wisc. 1 29 is a highly fragmentary list written on the back of another document. The relevant line is quite damaged, and the quality of the published image is not great. The transcription from papyri.info is below, along with the relevant section of the image.

κ̣δ Ἀλλοῦτι χαρ̣τ̣[ο]\π̣/[(ώλῃ)] (ἀρτάβη) α

P.Wisc. 1 29, detail of letters editor restores as χαρ̣τ̣[ο]\π̣/[(ώλῃ)] image source: P.J. Sijpesteijn, The Wisconsin Papyri I (Brill, 1967), plate X

The original editor suggested χαρτοπράτῃ as a possible alternative reading for χαρτοπώλῃ. The word χαρτοπράτης (“papyrus seller”) has the virtue of being clearly attested in unabbreviated form in two papyri, but the drawback is that both these papyri date to the seventh century CE at earliest (BGU 1 319 and P.Berol. inv. 2749). The term also appears in Latin transliteration (chartopratis) in the Justinian Code 11.18. Caroli, however, suggests that we have here another instance of χαρτοποιός, but given what we have seen of the overall insecure attestation of this word, proposing such a reading here just begs the question.

P.Flor. 3 388 is a set of private daily accounts from the late first or early second century CE. The online image of this piece is not especially good, but I supply it below with the text of the relevant lines from papyri.info:

[⁦ -ca.?- ⁩]  ̣[⁦ -ca.?- ⁩]  ̣  ̣σακκω̣(  ) αὐτῶ(ν) (δραχμαὶ) ιβ ὀβ(ολοὶ) κ ἀρ[γυρι]κ(οῦ) ἔσχο(ν) δι(ὰ) Πλουτ(ᾶτος) ὀβ(ολοὺς) ε λοιπ(  )
[⁦ -ca.?- ⁩] γ τιμ(ῆ)ς χαρτοπ(   ) δι(ὰ) Παθώ(του) 𐅵 (δραχμῆς) ε[ἰς τ]ιμὴ(ν) ἰχθύω(ν) (δραχμὴ(?)) α

P.Flor. 3 388, showing the letters read as χαρτοπ in the second line; image adapted from PSI Online

The editor left the abbreviation unresolved. Caroli suggests χαρτοποιός. But again, there is no clear reason to resolve the abbreviation in one way or the other.

The TLG doesn’t help much either, giving just one hit, Constantine Porphyrogenitus (905-959 CE), De administrando imperio 52.11, in which χαρτοποιοί are listed among other professions (sailors and murex fishers) that did not provide horses in a levy in the Peloponnese in the tenth century. There is no indication of the exact meaning of the word. “Papyrus maker” seems unlikely in such a temporal and geographic setting, but if the term refers to another writing surface (parchment or paper), surely it must have had a prehistory of some sort in which papyrus figured?

The Lexikon zur byzantinischen Gräzität, if I understand the entry correctly, adds a reference to a particular manuscript of Theodore of Stoudios (8th-9th century): “A. Dobroklonskij, Prepodobnyj Feodor I (Odessa 1913) 413adn.1.: Theod. Stud., catech. magn. in cod. Patm. 112.” I have not had a chance to track this reference.

So, there are still some references to check, but at this point, I don’t think I’m totally convinced that we have any attestations of this word before the medieval period.

Posted in Papyrus Making | 5 Comments

Snowy White Papyrus

Posted on July 12, 2023 by Brent Nongbri

I’ve been experimenting with making my own papyrus and discovering that there are lots of variables to control in order to make good quality writing material. One thing that especially surprised me has to do with the color of papyrus.

Modern commercially available papyrus produced in Egypt is often treated with an alkaline solution and bleach during production, which gives it a yellowish color (as outlined in this video, at the 1:40 mark). Commercially available papyrus produced in Sicily, at least the samples I have seen, tends to have a more brown tint. I’m not sure if the Italians use any additional processing. In any event, the papyrus sheets that one can buy aren’t entirely helpful when thinking about the color of papyrus in antiquity.

In making my own papyrus, I expected that different plants would yield sheets that looked a little different, but I was surprised to find that two stalks from the same plant could produce sheets of quite different colors. The image below shows two small sheets that I made using two stalks from the same plant.

Papyrus sheets of different color produced from the same plant; image source: Brent Nongbri, July 2023

The two stalks were cut at the same time. The pith of each of the stalks was an identical color (white with a slight grey tinge). I cut the strips in the same way from the same central portion of each stalk. I pressed them under the same conditions. Yet the colors are very different.

I’ve wondered about passages in ancient Roman authors that describe papyrus sheets as white. When Pliny discusses different grades of papyrus, he lists several qualities by which papyrus sheets were judged: fineness, firmness, smoothness, and whiteness (candor). At least one Roman poet describes papyrus rolls as “snowy white” (Tibullus [Lygdamus], Elegiae 3.1.9: lutea sed niveum involvat membrana libellum: “But let yellow parchment wrap the snowy white little book”). It makes me wonder if there was something different about the papyrus plants that grew in Egypt in antiquity, or if there was something in the ancient processing that helped to produce uniformly white sheets (for instance, Pliny mentions that papyrus sheets were dried in the sun–could that have a bleaching effect?). It would be challenging to produce papyrus sheets on an industrial scale if you had to worry about strips from different stalks potentially having quite different colors and ending up on the same sheet. That kind of thing did happen occasionally. Notice the differently colored papyrus strips in this bifolium from the Bodmer Menander Codex:

Bifolium from the Bodmer Menander Codex, showing alternate strips of lighter papyrus; image adapted from the Bodmer Lab

You see this phenomenon from time to time (thought not usually so pronounced). For the most part, however, it seems like ancient bookmakers tried to avoid this kind of thing.

Posted in Papyrus Making, Voluminology | 2 Comments

What Do We Mean By “Codex”?

Posted on July 3, 2023 by Brent Nongbri

P.Hib. 113, the papyrus kept at Graz that has recently been proposed as being the earliest surviving remains of a codex, continues to be in the news. I made a brief post about it some days ago, and in the comments to that post, David Kelsey raised an interesting question:

“If I were to take three sheaths of paper, fold them in the middle and insert inside each other to form a quire, then stitch the sections together, as a finished product I essentially have a booklet. But is this a codex? At what point does folded sheets gathered become a codex? Could this papyrus from Graz University be the front page of a booklet only and can it is so proved to have other associated pages written on both side be an actual codex?”

This is a good observation and really boils down to this: What do we mean by the term “codex”? Is it something more than a series of leaves linked together in some fashion? I tend to think of the codex as an example of a kind of technology, so it is indeed these physical characteristics that are most interesting to me. I would answer David’s question affirmatively. What he describes, a stack of three sheets folded together and stitched through the middle, is a small, single-quire codex.

But other questions remain. What about less flexible materials like wood or ivory–can these form codices? What about, say, a diptych joined by hinges rather than threads or tackets–is that a codex?

For some bookbinding specialists, the method of binding is the key element. J.A. Szirmai, for example, has pointed to the method of joining the “leaves” as a determining factor and a firm point of separation between tablets and codices:

“Hardly any textbook in which the origin of the codex is discussed, fails to assure us that it [was] wooden tablets from which the construction of the codex was derived. Even authorities proclaim this with great certainty, so Roberts and Skeat (The Birth of the Codex, 1983, p. 1): ‘There has never been any doubt about the physical origin of the codex, namely that it was developed from the wooden writing tablet. . . ‘. The certainty with which the validity of this statement is taken for granted is in marked contrast with the lack of any substantial evidence or explanation as to the exact nature of this genetic relationship. Yet the assumption is being repeated again and again without any sign of intellectual discomfort about the weakness of the argument. . . The analogy in the geometrical shape of the composing elements or of their ability to be turned along one side is merely superficial; the primitive methods of connecting the elements of the writing tablets (using hinges, metal rings or lacing) have scarcely anything in common with the codex structure.”

Still other specialists also factor in the textual contents. In his reflections on “Les origines du codex,” Joseph van Haelst defined “codex” as follows: “The term ‘codex’ . . . designates a collection of sheets of papyrus or parchment folded in two, grouped into a quire (or quires), sewn together along the spine, and usually protected by a cover. Its content, unlike that of the so-called documentary codex, is a composition, i.e. a text designed for distribution and preservation. This may be literary (classical works) or professional-technical (biblical, legal, magical, medical, scholarly, etc.).”

I don’t really like this division between the (real) codex and “documentary codex,” but for better or worse, it is a part of the scholarly discussion. So, we really have three intertwined issues–physical form (leaves), method of joining (binding), and contents (documentary vs. literary).

Part of the reason for these overlapping categories has to do with etymology. The Latin caudex originally meant “tree trunk” or “block of wood,” and was later used to refer to bound sets of wooden tablets, which were themselves cut from blocks of wood (Greeks seem to have used the words δέλτος or πίναξ to refer to these tablets). We don’t have surviving examples of this kind of artifact from very early periods, just some iconography, such as the famous Douris School Cup of the early fifth century BCE:

A teacher with a set of writing tablets on the “Douris School Cup” produced in the early fifth century BCE; image source: Staatliche Museen zu Berlin, Antikensammlung /CC BY-SA 4.0

Surviving examples of wooden tablets from later in the Roman era, however, are plentiful. The images below show a well-preserved set of wooden tablets dating from the fourth century CE that were found in the Egyptian city of Kellis. They were excavated intact with the “binding” thread in place (left and center), and they are shown in a stack after the thread was removed (right).

The “Kellis Agricultural Account Book,” a set of wooden tablets found intact at Kellis in Egypt; image source: Roger S. Bagnall et al., The Kellis Agricultural Account Book (Oxbow, 1997), plates

We can get a glimpse into how these tablets were made by looking at another artifact found at Kellis in proximity to woodworking tools, namely a wooden block that appears to be an off-cut from a block that had been sawn into a similar set of thin slabs:

An off-cut from a wooden block that had been sawn into thin slabs for tablets, found at Kellis in Egypt; image source: Roger S. Bagnall et al., The Kellis Agricultural Account Book (Oxbow, 1997), plate

So we can see the etymological connection to the caudex as “block of wood.” But there is also the question of how the slabs are held together after they were cut from the block. We don’t have a full picture of how these tablets were usually bound. The example shown above from Kellis has a pretty simple “binding” of a thread looped through the holes that were drilled in the wood slabs and tied off. But other surviving tablets have “covers” (outer slabs) that show evidence of a somewhat more elaborate lacing of the threads, and a 2018 book by Georgios Boudalis offers some speculative but very intriguing ideas about exactly what such lacing might have looked like, and indeed how closely it might resemble the bindings of multi-quire papyrus and parchment codices.

While these bound groups of wooden slabs are sometimes described as “codices,” some scholars (most scholars?) tend to reserve the term for bound groups of leaves made of more flexible materials like parchment, papyrus, and later paper. As noted earlier, in many studies of codices, we find a kind of evolutionary narrative that is based on contents as much as physical characteristics. These wooden tablets often contained documentary writing. The tablet found at Kellis shown above, for example, contained agricultural accounts. Other tablets contained legal documents, wills, public records, writing exercises, or other sorts of “non-literary” writing. We also have papyrus and parchment books that contain similar non-literary sorts of material. But such manuscripts are often described not as codices proper, but rather as “notebooks.”

A typical example of such a “notebook” (the one used as an illustration by Roberts and Skeat in The Birth of the Codex) is a parchment bifolium in Berlin, P. 7358 + P. 7359 (SB 26 16551) that contains some notes on labor carried out and payments made. Roberts and Skeat assigned the manuscript to the second century CE, but the editor of the papyrus preferred a date in the third century (on p. 21, footnote 2 Roberts and Skeat also assign the manuscript to the third century):

The two sides of a bifolium from a parchment “notebook,” Berlin, P. 7358 + P. 7359; image source: Berliner Papyrusdatenbank

In The Birth of the Codex, Roberts and Skeat present the arrival of the “notebook” as a turning point in the story of the codex, as indicated by the chapter title “From Writing Tablet to Parchment Note-Book.” They use the terminology of a transition and name the agent: “Certainly it was the Romans who took the decisive next step, that of replacing the wooden tablet by a lighter, thinner and more pliable material” (p. 15). Now, I’m not sure at present how much I like this linear, evolutionary way of thinking about wooden tablets and papyrus and parchment codices, but if we momentarily allow this scenario, it is at this point in the narrative that the Graz papyrus might shake things up. If the Graz papyrus really is an example of a papyrus codex/notebook from the third century BCE produced by Greek-speakers in Egypt, then the Roberts-Skeat story about the Romans’ invention would require some significant revision.

Sources cited:

Roger S. Bagnall et al., The Kellis Agricultural Account Book (Oxbow, 1997).

Georgios Boudalis, The Codex and Crafts in Late Antiquity (Bard Graduate Center, 2018).

Günter Poethke, “Ein Berliner ‘Notizbuch’ aus Leder,” in C.-B. Arnst et al., Begegnungen: Antike Kulturen im Niltal (Verlag Helmar Wodtke, 2001), 399-403.

C.H. Roberts and T.C Skeat, The Birth of the Codex (Oxford University Press, 1983).

Joseph van Haelst, “Les origines du codex,” in Alain Blanchard, Les débuts du codex (Brepols 1989), 13-35.

J.A. Szirmai, “Wooden Writing Tablets and the Birth of the Codex” Gazette du livre medieval 17 (1990), 31-32.

Posted in Codices, Codicology, P.Hib. 113, Tablets | 11 Comments

The Oldest Codex?

Posted on June 22, 2023 by Brent Nongbri

News reports coming out of Graz today suggest that a papyrus in the collection of the University of Graz may be the earliest surviving specimen of a bound book with pages, possibly as early as the third century BCE. It is a papyrus that was published way back in 1906, P.Hib. 113:

The earliest codex? Graz, UBG Ms 1946; image source: University of Graz, Universitätsbibliothek

“The papyrus fragment (Graz, UBG Ms 1946), which measures only about 15 x 25 cm, has belonged to the University of Graz since 1904. Discovered during an excavation in the Egyptian necropolis of Hibeh (today El Hiba) south of Fayum (El-Fayoum), the fragment now belongs to a collection of 52 papyrus objects, some of which were used as so-called cartonnage for mummy wrappings in the Ptolemaic period (305-30 BC).”

Theresa Zammit Lupi, the head of conservation of the Special Collections of Graz University Library, noticed the presence of a thread, what appears to be a central fold, and holes along the central fold that could indicate binding. This announcement is potentially very exciting news for those of us interested in the early history of the codex, but the news release itself is a bit jumbled. The claims made about the earliest codices in the report seem somewhat confused. According to the press release,

“The earliest codices known to date with evidence of stitching in book form have been dated to 150-250 AD. Two such examples are located in the British Library (Add MS 34473) and the Chester Beatty Library, Dublin (CBL BP II).”

British Library “Add MS 34473” isn’t a designation of a single manuscript. That inventory number covers several parchment manuscripts, none of which (to the best of my knowledge) has been assigned especially early dates. The possible exception might be parchment fragments of a codex containing works of Demosthenes filed under that inventory number (P.Lond.Lit 127, TM 59549), which are sometimes assigned a date as early as the second century CE. There is, however, a parchment codex fragment in the British Library that is generally assigned an even earlier date, British Library Papyrus 745, (TM 63267), which may be as old as the second or even the first century CE. Chester Beatty BP II is, of course, the single-quire papyrus codex of Paul’s letters better known to New Testament scholars as P46 (TM 61855), which has been assigned to a wide range of dates, from the late second century CE to the late third or fourth century CE. But it is generally thought that the earliest of the Beatty Biblical Papyri is the BP VI (TM 61934), the Numbers-Deuteronomy papyrus codex, which is usually assigned to the second or third century CE.

These small details aside, I am excited to learn more about this papyrus. It was reused for cartonnage, which gives us a reasonable terminus ante quem. But as far as I can see from the online images, there is no writing visible on the “back” side of the papyrus. There is writing on the right side of the “front” of the papyrus, but it is tough to decipher. It would be nice to see the whole thing subjected to multi-spectral imaging to see what is legible on the right side of the papyrus as well as the “back” side. If there is continuous text from the back to the front, that would be a nice confirmation that we do in fact have a very early codex here. I look forward to learning more about this piece!

Posted in Codices, Codicology, P.Hib. 113 | 7 Comments