Skip to content

“God’s Library” (by Brent Nongbri): A Review

November 12, 2018

In an essay published nearly two decades ago, I drew attention to several features of early Christian manuscripts, identifying these features as likely our earliest expressions of an emergent Christian visual and material culture.[1]  A few years later, in a contribution to the Society of Biblical Literature Forum series, I wrote about early Christian manuscripts as artefacts.[2]  In the following years in several further publications I emphasized the relevance and importance of early Christian manuscripts for historical questions about the origins and early development of Christianity.[3]  In particular, I single out my 2006 book urging greater attention to the specific features of early Christian manuscripts.[4]

So it is encouraging to see this emphasis on manuscripts taken up in Brent Nongbri’s newly published book, God’s Library:  The Archaeology of the Earliest Christian Manuscripts (Yale University Press, 2018).  Whereas my studies have focused on what we might call the para-textual features of earliest Christian manuscripts, Nongbri focuses on questions about their provenance, acquisition, and how we go about assigning manuscripts dates. The “archaeology” in the sub-title is mainly what he refers to as “museum archaeology,” the detailed “digging” into the archives of the institutions that hold these manuscripts, to address his questions.  And Nongbri has certainly done this sort of digging!  His book is chock-full of discoveries of answers to these matters, or, in other cases, fresh questions about them.

In a number of essays of his own over recent years, Nongbri has surfaced questions about the acquisition of manuscripts and, more provocatively, querying the heretofore widely-accepted dates of some key manuscripts of the Gospels.[5]  In the book reviewed here, in his own words, “issues of dating have recurred throughout this study” (269).  But his argument about dating isn’t so much a firm defence of the later dates that he prefers, so much as it is a (properly in my view) critical querying of the unwarranted specificity of dates assigned to some NT manuscripts by some scholars.

Nongbri repeatedly notes that dates have often been assigned solely (and, in many cases, unavoidably so) on the basis of palaeography; and, as anyone can verify by examining the relevant publications, even recognized experts in Greek palaeography can differ over the likely date of a given manuscript by as much as a century or more.  His main complaint is that scholars have too often preferred the earlier dates of some experts over the later dates assigned by others, without noting that any date based on palaeography can only be approximate.  So our handling of the dates of early papyri, which forms the real focus in the book and in questions about the early texts of NT writings, should be more modest and careful than has sometimes been the case.

But, whatever your view of the dates of early papyri, Nongbri’s book is a rich harvest of his investigations, especially in the archives of some key collections and holding institutions.  I did some of this sort of work myself in the Freer Gallery of Art archives, in preparation for an edited volume on the Freer biblical manuscripts, and I share Nongbri’s enthusiasm for this “museum archaeology.”[6]

The first chapter, “The Early Christian Book,” is a review of writing material used and the techniques of codex construction.  There is excellent detail in this discussion, with helpful visual illustrations.  My one quibble is over Nongbri’s statement that, with “a few” exceptions, “‘the book’ in the early Christian centuries was almost always the codex” (22).  As I’ve shown, the early Christian preference for the codex appears to have been particularly consistent for those texts that were coming to be treated as scriptures, whereas there seems to have been a somewhat greater readiness to copy other texts on rolls (e.g., theological treatises and certain other texts).  By my count, about one-third of the extant manuscripts of the latter types of texts are rolls.[7]  Indeed, Nongbri’s table of the forms of Christian books at Oxyrhynchus (Table 6.1, p. 237) rather clearly supports my point.  This still obviously represents a Christian preference for the codex overall, but I think that the marked distinction in degrees of preference for various kinds of texts is noteworthy.

Chapter two, “The Dating Game,” reviews the methods used to try to establish the likely date of an undated literary manuscript.  These involve noting their contents, their re-use (“opisthographs,” the writing of a text on the outer surface of a roll that was initially used for some earlier, sometimes dated, text), palaeography, radiocarbon analysis, and analysis of the ink of a manuscript.  Readers will find this a very informative chapter.  For example, Nongbri explains the different handwriting styles referred to by palaeographers, and the somewhat different approaches advocated by them.

As noted earlier, Nongbri offers an extended critical appraisal of the limits of palaeographical dating, observing, for example, that a given type of handwriting may well have been practiced over a century or more.  In this chapter he rightly urges that palaeographical dating must proceed on the basis of securely dated comparisons, whenever possible.  He is equally helpful and candid in noting the limits of radiocarbon analysis, and the importance of conducting this sort of analysis with great care to avoid contamination.  And, as for ink analysis, he notes that this can only tell us the chemical contents of the ink, not its age.

In the following chapters, Nongbri lays out the results of his impressively detailed investigation of questions about the provenance and acquisition of some collections of early Christian books.  Chapter four is devoted to the Chester Beatty biblical papyri (eleven codices, the largest single collection of early biblical papyri).  He discusses both the collection and how it was acquired, and also reviews the individual books in the collection.  The “jewels in the crown” among the Chester Beatty papyri, so far as NT scholars are concerned, are obviously P45 (P.Chester Beatty I, a codex that originally contained the four Gospels and the book of Acts) and P46 (P.Chester Beatty II, a collection of Pauline epistles in one codex), both of them typically assigned to sometime in the third century.

Nongbri expresses some hesitation about the dates of these two codices, however, because “our corpus of datable samples in taken almost entirely from rolls,” a format that was “largely supplanted [in general usage for Christian and non-Christian manuscripts] by the codex by the end of the fourth century” (138).  True, but whether roll or codex “documentary” texts tended to have dates written by the copyist, and so they can still be used to help us make comparisons with undated literary texts such as the NT writings.  Perhaps I’m missing something, therefore, but I don’t find his argument persuasive.  His review of dates for the other Chester Beatty codices (e.g., the Numbers-Deuteronomy codex), however, is informative and corrective (more likely third century than second).

Chapter five is devoted to the Bodmer Papyri, a much more diverse body of early Christian writings, and reflecting a much wider timeframe.  Here, too, Nongbri combines a critical discussion of the various stories about the finding and acquisition of these items with the contents of the books.  In several helpful tables, he lays out the details of the various Bodmer codices, drawing upon the Leuven Database of Ancient Books.  The diversity of the Bodmer collection includes the writing material (both papyrus and parchment), the languages (Greek, Coptic, and Latin), and the construction (both single and multiple quire codices). Nongbri classifies the manuscripts into three groups, based on their format, drawing particular attention to the “considerable group” of “basically square codices” (195).  Contra the judgements of some earlier scholars, Nongbri opines that much of the Bodmer collection can be dated in the fourth century.

This includes both P66 and P75 (the latter was purchased from the Bodmer foundation privately and then donated to the Vatican Library).  Nongbri’s arguments here and in his earlier essays on dates of these items deserve careful consideration, to be sure, but I think that he is more effective in his critique of some overly specific dates than in establishing his own preferred ones.

Chapter six is a highly informative discussion of the excavations conducted at Oxyrhynchus by Grenfell and Hunt and subsequent archaeologists.  This is the site from which comes the largest number of manuscripts of early Christian writings.  In an appendix (273-80), Nongbri lists Christian books from Oxyrhynchus as of 2016, classified by types of literature.[8]  Noting that the Oxyrhynchus papyri were unearthed in trash heaps of the ancient city, and Nongbri poses, but doesn’t try to answer, what that might signify.  The story of the excavations by Grenfell and Hunt will make fascinating reading for those unfamiliar with it.  They employed as many as a couple of hundred local men and boys for the digging, and unearthed many thousands of manuscripts (albeit, many of them only fragments), shipping then back to the UK in metal boxes.  To date well over 5,000 have been published in the Oxyrhynchus Papyri series, but this amounts to only a small percentage of the total, much of which is currently housed in the Sackler Library (an extension of the Ashmolean Museum, Oxford, UK).

Chapter seven gives a particularly interesting piece of detective work by Nongbri on a small set of papyrus fragments of the Gospel of Matthew (P64, P67) and the Gospel of Luke (P4).  Nongbri deftly shows that the oft-repeated story of the fragments coming from the binding of a copy of works by Philo of Alexandria arose from a misunderstanding of the original account.  He further shows that the subsequent proposals that these fragments derive from a four-Gospel codex of the second century CE are highly dubious, and seem to rest more on speculation than secure evidence.  Granting that a second-century date is “not impossible,” Nongbri makes it as “equally possible” that they derive from the third or even fourth century (267).

In his epilogue, Nonbri judges that “although a few Christian books may be as old as the second century, none of them must be that old” (269, emphasis his).  Complaining that there remains a “drive to have older and older Christian manuscripts,” he cites the ill-judged rumors of a first-century fragment of a manuscript of Mark as another example of this.[9]  Nongbri concludes by urging the greater use of radiocarbon analysis, greater interest in “museum archaeology,” and the greater digitization of manuscripts held in various institutions.[10]  All of these suggestions are cogent and to be embraced.

Nongbri has produced a “must read” for all those interested in early Christian manuscripts, and will likely persuade some of those who haven’t shown such interest that they should!  His proposals for revisions of dating of some key manuscripts carry varying force, but deserve a careful and considerate evaluation.  But it has to be said that the dates assigned to early Christian manuscripts have tended to be reached by papyrologists and palaeographers, practicing the same methods by which they date non-Christian manuscripts.  So, for example, if too many Christian manuscripts have been assigned too early, then is the same the case for the many more non-Christian manuscripts dated by the same people and by the same methods?  But, to repeat myself for emphasis, we can all be grateful for Nongbri’s impressively researched book, which I am sure will deservedly generate still greater interest in the study of early Christian manuscripts as artefacts.

[1] Larry W. Hurtado, “The Earliest Evidence of an Emerging Christian Material and Visual Culture:  The Codex, the Nomina Sacra, and the Staurogram,” in Text and Artifact in the Religions of Mediterranean Antiquity:  Essays in Honour of Peter Richardson, edited by Stephen G. Wilson and Michel Desjardins (“Studies in Christianity and Judaism 9; Waterloo:  Wilfrid Laurier University Press, 2000), 271-88.  (American and British spellings of the word differ slightly, “artifact” and “artefact” respectively, hence the variation in the titles of publications, depending on their provenance.)

[2] “The Earliest Christian Artifacts:  Manuscripts and Christian Origins,” SBL Forum.  ( 2004.

[3] Hurtado, “The ‘Meta-Data’ of Earliest Christian Manuscripts,” in Identity and Interaction in the Ancient Mediterranean.  Jews, Christians and Others:  Essays in Honour of Stephen G. Wilson, eds. Zeba A. Crook and Philip A. Harland (Sheffield:  Sheffield Phoenix Press, 2007), 149-63; “Early Christian Manuscripts as Artifacts,” in Jewish and Christian Scripture as Artifact and Canon, ed. Craig A. Evans and Daniel Zacharias (Library of Second Temple Studies,70; Leiden:  Brill, 2009), 66-80; “What do Earliest Christian Manuscripts Tell Us About Their Readers?” in The World of Jesus and the Early Church:  Identity and Interpretation in Early Communities of Faith, ed. C. A. Evans (Peabody, MA:  Hendrickson Publishers, 2011), 179-92; “Manuscripts and the Sociology of Early Christian Reading,” in The Early Text of the New Testament, eds. Charles E. Hill, Michael J. Kruger (Oxford:  Oxford University Press, 2012), 49-62; “P45 as an Early Christian Artefact:  What it Reflects about Early Christianity.” Teologisk Tidsskrift 4 (2016): 291-307.  Some of these essays (along with some others) are republished now in Hurtado, Texts and Artefacts: Selected Essays on Textual Criticism and Early Christian Manuscripts LNTS 584 (London:  Bloomsbury T&T Clark, 2018).

[4] Larry W. Hurtado, The Earliest Christian Artifacts:  Manuscripts and Christian Origins (Grand Rapids:  Eerdmans, 2006).

[5] There are sixteen publications by Nongbri listed in the bibliography to his recent book (including several postings from his blog-site).  With particular reference to the dates of certain manuscripts, Brent Nongbri, “The Use and Abuse of P52:  Papyrological Pitfalls in the Dating of the Fourth Gospel,” Harvard Theological Review 98.1 (2005): 23-48; “Grenfell and Hunt on the Dates of Early Christian Codices:  Setting the Record Straight,” Bulletin of the American Society of Papyrologists 48 (2011): 149-62; “The Limits of Palaeographic Dating of Literary Papyri:  Some Observations on the Date and Provenance of P. Bodmer II (P66),” Museum Helveticum 71 (2014): 1-35; “Reconsidering the Place of Papyrus Bodmer XIV-XV (P75) in the Textual Criticism of the New Testament,” Journal of Biblical Literature 135.2 (2016): 405-37.

[6] Larry W. Hurtado (ed.), The Freer Biblical Manuscripts:  Fresh Studies of an American Treasure Trove (Atlanta:  Society of Biblical Literature, 2006).  See esp. the detailed discussion by Kent D. Clarke, “Paleography and Philanthropy:  Charles Lang Freer and His Acquisition of the ‘Freer Biblical Manuscripts’,” 17-73.

[7] Hurtado, The Earliest Christian Artifacts, 53-61.

[8] See now Lincoln H. Blumell and Thomas A. Wayment, Christian Oxyrhynchus:  Texts, Documents, and Sources (Waco, TX: Baylor University Press, 2015), for the texts of 175 items.

[9] The item in question has now been published in the latest volume of the Oxyrhynchus Papyri series.  See my blog posting on the publication here.

[10] I’m pleased to have had a hand in arranging for the digital photography of the Chester Beatty biblical papyri several years ago, which are now available online at the Center for the Study of New Testament Manuscripts (here).

From → Uncategorized

  1. Larry, you wrote:”They preferred it already by the second century”. Who were “they” specifically?

    • Geoff: I think I made it clear in my posting that I was referring to the early Christian preference for the codex. You know, Christians, who had been around since the early first century.

      • Larry, I expected that you would say that. But who were the real wheeler-dealers who had power in a Roman world? Were “they” the same sort of people we see today with power and authority in the the Christian church? Were “they” able to say, write and do exactly what they wanted? This would narrow down who I am talking about to a limited number of people, not your simplistic generalised Christian.

      • Geoff: Do please take the trouble to learn before you utter your private opinions publicly. In the first two centuries there were no power brokers in Christian circles such as you wrongly posit. The humble nature of the papyri from that period reflects the more “grassroots” nature of the movement. You have a seriously anachronistic grasp of the matter, and your views of matters are, I have to say in all candor, worthless. Do go peddle them somewhere else.

  2. Hi Larry: Thanks for the thorough and thoughtful review! I’ve posted a couple follow-up thoughts here:

    • Brent: Thanks for the posting on your site, which clarifies for me your argument. I take it that you see the “problem” as there being few re-used rolls, by which to date literary hands.

    • Tom Hennell permalink

      Fascinating Brent; and along with Larry’s review, you have sold me your book.

      I realise this may be covered there; but just to clarify matters for the moment. Suppose that a paleographic range of dates is being assessed for a Christian papyrus text, on the basis of counterpart literary hands; which have dated documents on the back. And from a set of closely matched comparators the paleographer assesses that there is a 95% probablity that the date for the papyrus in question falls within the range 150 – 250.

      But, as I understand it, you are proposing that the assessed dating range might then properly be extended further into the third and fourth centuries – to take into account the radically reduced potential number of comparators available in the period when fewer literary texts may have had dated documents on the back of them. But how can this be done? Are you proposing that less closely matched comparators of third and fourth century date – or maybe comparators that closely match some aspects of the papyrus in question, but not others – should be ‘promoted’ in some way in the calculus?

      A potential confounding factor here, is a common popular misunderstanding of how probability ranges work. To take the same example, suppose that instead of assessing the date of the papyrus in question as being in the range 150 – 250, with a 95% probability; we now assess the 95% probability range as being from 150 to 350. But, contrary to how it may often be understood, doubling the length of the range does not halve the likelihood of of any one date within that range being the actual date of the papyrus in question. Although it is now more likely that the papyrus is of third century date, it has not become less likely that it is of later 2nd century date. All dates within any 95% range, whether shorter of longer, have the same probability of corresponding with the actual date.

      Of course you are in no way responsible for popular misinterpretation of your published findings.

      • Yes, Tom, a valid point, that extending later the range of possible dates doesn’t mean that an earlier date is any less probable. It just means that an earlier or later date is about equally cogent.

  3. bryantiii permalink

    Dr. Hurtado,

    Thank you for the review of Nogbri’s book. I hope to save the money necessary to obtain it, but you know how that goes.

    Couple of questions.
    First, in you paragraph above in which you quibbled over the preference for the codex. I wonder if the codex, as you indicated as a preference to be used for scriptural books, became more prevalent toward the end of the Third Century CE and into the Fourth Century CE as Christianity became more established than in the First, Second, or early Rhird Century?
    Second, could this preference for the Codex to be used for Scripture have coincided with the gradual acceptance of which books belonged in the NT Canon?
    Third, would the increased preference of the Codex also coincide with the proclamation of Constantine I to have 50 Bibles (OG/LXX and NT) placed in churches throughout the Roman Empire? This would seemed to be a natural conclusion, but we know what happens with “appearances.”

    • Bryant: The early Christian preference for the codex,especially for texts treated as scripture, is attested in our earliest extant evidence (some of which likely goes back as early as the late second century). Constantine and the “success” of Christianity had nothing to do with it.
      Second, the treatment of texts as scripture is one thing, and the formation of a canon is another. The latter took much longer, and the preference for the codex again is not directly connected with a canon.
      Third, there is no “increased preference for the codex” in Christian circles. They preferred it already by the second century, and rather consistently.

Comments are closed.

%d bloggers like this: