Ancient Books Website

My thanks to Drew Longacre for drawing my attention to a new web resource on ancient books from William A. Johnson and Nicholas Wagner at Duke. As the website explains:

The Ancient Books Website (ABW) joins a long tradition of open- access tools for papyrological research. The website provides data complementary to those in tools like the Leuven Database of Ancient Books (LDAB), now part of Trismegistos (TM), and the Digital Corpus of Literary Papyri (DCLP). The data captured here are focused on two areas: (1) reconstruction of the physical details of each literary papyrus, and (2) analysis of the scribal features.

Currently, they have two main datasets online that give detail on the physical features and scribal features of bookrolls. Again, from the website:

Physical features. The website provides measurements for width and height (measured or calculated) for features like column, intercolumn, roll, letters, and vertical spacing. 

Scribal features. So far, the website provides synoptic analysis of the punctuation, and of scribal usage for nu-movable and iota adscript. We will be adding to these data over time. 

Datasets for early codices is said to be coming in 2026. 

Solving a Puzzle in P136

The following is a blog post I first wrote while on sabbatical in 2022 in Wake Forest. While there, I spent a nice afternoon with Libbie Schrader Polczer at the wonderful David M. Rubenstein Rare Book & Manuscript Library examining their Greek NT manuscripts. One of those is P136.

---

This last week I had the good fortune of getting to spend some time with Duke University’s sole NT papyrus, P. Duke Inv. 1377 (P136). The GA number is your tipoff that this is a recently published papyrus. Andrew and Valerie Smith published the editio princeps only in 2018. It’s so recent that it didn’t make it into the ECM Acts. (As an aside, one of the great things about textual criticism is that we regularly get new material to work with like this.)

It’s a sixth century fragment of Acts and it has some surprising features. The script is the first surprise. This is not your typical six-century majuscule. The Smiths identify it as chancery style and it does fit Kenyon’s description of 6th/7th century non-literary style for Byzantine papyri (Paleography of Greek Papyri, pp. 49-51). 

The more surprising feature is that P136 has text on both sides but is not a codex. The writing is upside down on the recto relative to the verso. The Smiths suggest it is most likely a scroll, written in rotulus format (meaning the text flows vertically rather than horizontally). This format apparently saw a resurgence of use in the sixth century. If this is the right format, the Smiths’ reconstruction would allow for Acts 1:1-13:43 in the original scroll. That’s a little less than half the book (based roughly on the pages in my NA27). 

One last feature that surprised me: it’s big. Bigger than I expected, at least. It would have been a good-size scroll. Here is Libbie holding it up to show the size.

All that is by way of introduction. My real interest here is in a curious reading in P136. I was tipped off by the Duke website which says, 

Text varies from Nestle-Aland 28th edition Novum Testamentum Graecum in 3 locations, most significantly at 3v (Acts 4:28), which reads η χειρ σου και η δουλη instead of η χειρ σου και η βουλη σου.

That lack of σου is not surprising. It’s missing in over a dozen witnesses according to the ECM, including 02* and 03. What is odd is δουλη for βουλη. It’s only a difference of one letter, and the word δουλοι does occur in the next line. But the result has Peter and John claiming that those who conspired to kill Jesus did “whatever your [God’s] hand and your (female) servant had predestined to take place.” That’s definitely odd. 

The letter in question on line 3 of the ↑ side

To be sure, the letter certainly did look to me like a delta not a beta. Then again, I had no prior experience with this script. I do, however, have enough general experience reading Greek manuscripts to know that two letters that look the same at first may well turn out to be different entirely. To add to my suspicion, I checked the NT.VMR transcription and it has βουλη instead of δουλη. (I do not know the source of this transcription.)

So, is this a delta or a beta? Unfortunately, there are no other betas in P136, but there are multiple deltas. And what we can see is that in each case, the final stroke ends down and does not connect with the next letter. This distinction may seem slight, but it is typical of how letters are distinguished in cursive scripts like minuscule. I’m not too surprised to see it here. It’s a subtle difference to us, but they must have recognized it easily enough.

Deltas (orange dot) vs. beta (blue dot).

We can confirm our suspicion by looking at comparanda. The Smiths point to P. Geneva 210 as “remarkably similar” to P136 and we can see why. The hand is very similar. Lucky for us, it has plenty of both betas and deltas and, sure enough, we find the same distinction between them.

P. Geneva 210 showing deltas (orange dot) and betas (blue dot). Image source

Again, the distinguishing mark is that deltas end on a down stroke whereas betas don’t. Instead, they connect in some way with the next letter.

One more example. P. Berol. 7027 is written in a decidedly more elaborate hand than our previous examples. It’s much harder to read to my eyes and so I haven’t highlighted all the examples I could. Here, the beta is not quite the same as P136 or P. Geneva 210. Instead, it looks much more like the form of we find in later minuscule (it looks like a u to me). The key distinction is still the same, however. The delta ends on a down stroke and the beta does not.

P. Berol. 7027 showing deltas (orange dot) and betas (blue dot). Image source

From these examples—both within P136 and without—I think we can confidently say that P136 reads βουλη not δουλη in Acts 4:28 and is a witness to the text otherwise known in all other manuscripts at this point. Kudos to whoever did the transcription for the VMR (anyone know?) for catching this before me. It certainly does look like a delta, but a closer inspection confirms we have a beta. There is still plenty to puzzle us about P136, but I hope to have shown that this one letter is not one of them.

New Minuscule 2957 and Its Allies – Guest Post by Post

I want to introduce Darrell Post, graduate of Virginia Beach Theological Seminary, whom I invited to do this guestpost on the newly registered minuscule 2957 (David M. Rubenstein Rare Book & Manuscript Library, Duke University, Greek MS 053). Darrell has followed this blog for many years and is very interested in New Testament textual criticism. He has done us all a great service by creating and editing useful articles on Wikipedia.

He created the page on New Testament amulet (here). He thoroughly revised and expanded the page for NT papyri (here), uncials (here), minuscules (here), and lectionaries (here).

These helpful lists contain a lot of bibliography and links to images which is very helpful. The lists of minuscules can also be viewed by hosting institution. Darrell told me that someone else started individual pages for each minuscule, but he has not touched those pages since they are full of errors. He just focused on the lists. Here is his guestpost:

Guestpost by Darrell Post

Duke University owns a manuscript (MS 053) of Theophylactus’ commentary on the Gospel of John dated circa 1540 (see images here). The Scripture text appears to be complete with the words of the gospel written in light brown ink and the commentary text in a darker hue. This manuscript has just been issued a GA number, 2957. Wanting to know what sort of text was represented, I collated John 11:1-57 against the 2005 R-P Majority Text. Not surprisingly, MS 053 was found to mostly agree with the Byzantine text with several alternative readings.

Later while collating the text of GA-318, a damaged manuscript containing only John 7:9-12:8, I found it to be a nearly perfect match to Duke’s MS 053. Both manuscripts are commentaries with Scripture text in different color ink, and in John chapter 11 they share almost the exact same words, 943 out 953. Leaving out scribal corrections and most spelling differences, as allowed by the CBGM, the coherence between the two manuscripts is 950 out of 953. The two manuscripts also share a rare abbreviation (IE) of δεκαπεντε and the rare full spelling of Jesus in 11:33 and 38, along with the full spelling of πατηρ in 11:41. They also include a few somewhat hard to find readings of the NA text where it differs from the MT (αυτω instead of αυτου in 11:12, omission of αυτου in 11:54).

According to the test passages evaluated in the Text und Textwert tool available on the INTF web page, GA-318 (and therefore Duke’s MS 053) may be part of a subset of manuscripts belonging to the cluster Wisse identified as Cluster 2148. Working through this list, I have thus far found seven other manuscripts that match the proposed sub-set of Cluster 2148: 315, 742, 817, 819, 854, 1160 and 2735. Several other manuscripts probably also belong to Cluster 2148 including 833, 855, 857 and 2470.

With a few exceptions, all ten of the manuscripts collated thus far include each of the following unique, identifying readings from John 11: The omission of marian in 11:28, the omission of poiesai in 11:37, and the addition of de after the second legei in 11:39.

The table below shows the percentage of agreements between these ten manuscripts for John 11:1-57 against the 953 words in the R-P Majority Text.

	318	MS053	315	817	742	2735	819	1160	854	2148
318	100%	99.7%	99.3%	99.2%	99.3%	99.1%	98.6%	99.0%	98.8%	98.3%
MS053	99.7%	100%	99.2%	99.1%	99.2%	99.0%	98.5%	98.8%	98.7%	98.2%
315	99.3%	99.2%	100%	98.8%	98.5%	98.3%	98.1%	98.4%	98.3%	98.0%
817	99.2%	99.1%	98.8%	100%	99.5%	99.1%	99.1%	99.2%	98.8%	98.5%
742	99.3%	99.2%	98.5%	99.5%	100%	99.2%	99.0%	99.1%	99.0%	98.4%
2735	99.1%	99.0%	98.3%	99.1%	99.2%	100%	99.4%	99.5%	99.8%	98.4%
819	98.6%	98.5%	98.1%	99.1%	99.0%	99.4%	100%	99.3%	99.2%	98.0%
1160	99.0%	98.8%	98.4%	99.2%	99.1%	99.5%	99.3%	100%	99.3%	98.3%
854	98.8%	98.7%	98.3%	98.8%	99.0%	99.8%	99.2%	99.3%	100%	98.4%
2148	98.3%	98.2%	98.0%	98.5%	98.4%	98.4%	98.0%	98.3%	98.4%	100%

POSTSCRIPT: Greg Paulson has drawn attention to the fact that Maurice Robinson mentioned this manuscript in a comment on James Snapp’s blogpost about Greek MSS in the K. W. Clark collection at Duke University in 2017: “Also to be included but not yet digitized is Duke Gr. 53 (ca. AD 1450), Commentary of Theophylact on the Gospel of John with NT text interspersed. Peculiarly, this MS has no GA number even yet, although other Theophylact commentary MSS have a GA number). In this MS, the PA is not included, as is typical for commentaries.” Well, now by 2020, the manuscript has been digitized and yesterday it got its Gregory-Aland number – 2957.