Skip to content

That Newly-published Gospel of Mark Fragment: Focusing on It

June 30, 2019

Recent days have witnessed a dramatic turn of events in connection with the newly-published fragment of the Gospel of Mark (P. Oxy. 5345, now with the N-A number of P137).  One individual involved in the early days when the fragment was first mooted has written his account of things in an article in Christianity Today (here).  Another individual, named in that Christianity Today article, has then posted a response, accusing the author of the CT article of misrepresentation (here).

The allegations and counter-allegations begin to sound somewhat like a messy divorce, and leave a slightly sour taste in the mouth.  I suppose that there will be things to learn from whatever actually transpired, if/when we’re ever sure of what that was.  In the meantime, while the dramatis personae work things out, the rest of us can perhaps more profitably focus on the item at the centre of the controversy, the early fragment of the Gospel of Mark.  I don’t disparage those of an journalistic investigative bent.  I simply offer here some thoughts dealing with the item itself.

It has been published now, in the latest volume of the Oxyrhynchus Papyri series: Dirk Obbink and D. Colomo, “5345 Mark 1:7-9, 16-18,” in The Oxyrhynchus Papyri, Volume LXXXIII, ed. Peter J. Parsons and Nick Gonis (London: Egypt Exploration Society, 2019), 4-7.  See my notice of the publication in an earlier posting here.

Some things we know and can use in our historical work.  First, it is a fragment of a copy of the Gospel of Mark, and is datable on palaeographic grounds to sometime in the late second or early third century.  This provides us with evidence of a second copy of GMark in the first three centuries (the other one being P45, dated to sometime in the third century).  We have, thus, further confirmation that copies of GMark were being made and were circulating in this early period, although not as often as copies of the other NT Gospels, especially GMatthew and GJohn.

Second, the fragment has a provenance.  It was included in the huge body of papyri unearthed by Grenfell and Hunt in their excavations at Oxyrhynchus, and then sent back to London.  (The overwhelming mass of this material is yet to be published.)  So, this fragment shows more specifically that GMark was being copied and circulated in Egypt at this time.

Third, it is a portion of one leaf of a codex.  We know this because the writing proceeds from one side of the fragment to the other.  This provides us with yet another confirmation of the early Christian preference for the codex bookform, especially for the Gospels and other texts that had begun to be treated as scriptures.

The editors of the item estimate that the codex would have required roughly 20 sheets of papyrus, which would yield approximately 40 leaves or 80 pages, some 78 of which were likely needed for this copy of GMark.   This was probably a single “gathering/quire” codex.  That is, all the papyrus sheets were likely bound together in one bundle.  This is the more common way that earliest codices were constructed.  It is further likely that the codex contained only the GMark, for this also was the dominant way in which the Gospels were transmitted in the earliest period.

The writing of GMark commenced on the side of a papyrus leaf with vertical fibres.  This could suggest that this formed the third page of the codex, the first two pages (or first leaf) left blank and forming a cover.  (This assumes that this single-gathering codex was constructed with the papyrus sheets all laid each on the other with the vertical-fibre side down, which is one known pattern.)

This particular codex was likely about 12.4 x 16.6 cm in page-size.  This isn’t a miniature codex, but it is a smallish one, a little bigger than volumes in the Loeb Classical Library series.  The size of the writing, 0.2-0.3 cm, and the line-spacing of ca. 0.5 cm is rightly described by the editors as giving “a closely-packed appearance.”  These things could suggest a copy made primarily for personal reading, but that is only a suggestion.

The “hand” is, thus, small, but the letters are clearly formed and separated, and the copyist aimed for a “bi-linear” text (i.e., the letters all of the same height and the tops and bottoms of the letters all matching on two imaginary lines above and below the letters).  But the effort wasn’t entirely successful, for both the upsilon and phi violate bilinearity.  Also, the copyist wasn’t entirely consistent in the way he (?) formed certain letters, e.g., the alpha.  So, it looks like a skilled copyist, but not able (or concerned) to produce a truly calligraphic copy.  The result, however, is a clear and fully readable text (although the fragment has suffered some abrasion, especially on the horizontal-fibre side).

Portions of only a few verses of GMark are preserved, 1:7-9, 16-18.  In v. 8, the fragment reads that the coming one will baptize πνευματι αγιω (with Vaticanus and some others) instead of εν πνευματι αγιω (supported by Sinaiticus, Alexandrinus, Bezae and a good many others).  But this is a minor stylistic preference, and both variant-readings = “in/with the holy spirit.”

The more interesting (though also minor) variant is in v. 17.  Here, where most manuscripts read “Jesus said, ‘Come follow me’,” the fragment omits “Jesus”.  This variant isn’t noted in the N-A text, or in the Legg volume on GMark.  But the editors cite manuscripts Φ and 1194, “and a further scatter of minuscules” as having the same variant.  They cogently suggest that the omission could have happened by the copyist’s eyes leaping from the final letters of ειπεν αυτοις to the final letters of the likely nomina sacra (abbreviated) form of Ιησους (Jesus), which would have been ις.  So the manuscript being copied from may have read ειπεναυτοιςις, and the copyist’s eyes fell on the final letters, accidentally omitting the ις form of Jesus’ name.  In any case, the context makes it perfectly clear who is speaking.  If one infers from this extant portion that the rest of GMark was copied as carefully, we should expect further accidental variants, but otherwise a copy of GMark fully recognizable and not varying much from what we know today.

The only nomina sacra form extant on the fragment is in v. 8, where πνευματι (spirit) is written as πνι.  This likely confirms that the copyist was either a Christian or else somehow otherwise acquainted with this distinctive early Christian copyist practice of writing certain words in a curious abbreviated form, with a horizontal pen-stroke over the abbreviated form.  (For a brief online explanation of this practice, see here.  I discuss the matter more fully in my book, The Earliest Christian Artifacts:  Manuscripts and Christian Origins, 95-134.)

In sum, though only a small fragment (4.4 x 4 cm), P137 gives us some interesting data nevertheless.  It confirms that GMark was copied and circulated (apparently on its own), that early Christians preferred the codex especially for the NT Gospels and some other texts, that the copyists in this early period were generally competent and didn’t take wild liberties with what they copied, and (though this is more a suggestion than a firm conclusion from P137) that Christians in the second/third century were making copies of the Gospels both for ecclesial reading/usage and also for personal usage.


From → Uncategorized

  1. Dr. Hurtado: You wrote “The more likely scenario is that once Matt & John and Luke were known, Mark just wasn’t copied as often.” Possibly, but BEFORE the other gospels were written Mark was the only game in town, and so could have been copied just as often during that early period. If so the reason we have fewer copies of Mark is that the bulk of the copies came before the bulk of the copies of Matthew and Luke.

    This is similar to the pattern we see with majuscules and minuscules. The former gave way to the lattter, and hence a smaller percentage of the majuscules survived because they were older and/or were ‘retired’ when replaced by the latter. My point is that the evidence just tells us how many of the early copies survived, not how many were originally written, because we don’t know how or where they were written, used, discarded, or destroyed.

    • But, David, in the period from which we have evidence of extant mss, Mark was less frequently copied. Whatever happened before Matthew and Luke were written isn’t relevant.

      • Dr. Hurtado, of course what “happened before Matthew and Luke were written” is relevant. Assuming Mark was copied frequently “before Matthew and Luke were written” BECAUSE it was the only gospel at that point then it’s perfectly reasonable to suggest that when Matthew/Luke came out their much greater ‘coverage’ of Jesus directly led to Mark’s decline, with copies of Mark also being discarded/destroyed in favor of Matthew/Luke.

      • No, DAvid. It’s not relevant to the question: Was GMark copied less frequently AMONG THE FOUR GOSPELS? It isn’t a matter of GMark mss being destroyed. It was simply not copied as frequently as the other Gospels. I think we’re done on this.

  2. Tom Hennell permalink

    One further point proposed by the editors:

    – that the fragment is from the bottom of a page. So, if we assume that the beginning of Mark was the first page of the codex (or if not, that it started a new page); then the text of Mark preceding that witnessed in the fragment must have been of about the same length as that observed in the Codex Sinaiticus. In the 19th century it was commonly speculated, for instance by Karl Lachmann, that the current text of Mark Chapter 1 included multiple later intrusions (verses 2 and 3 especially). This fragment does not support that speculation.

  3. Dr. Hurtado, you wrote: “We have, thus, further confirmation that copies of GMark were being made and were circulating in this early period, although not as often as copies of the other NT Gospels, especially GMatthew and GJohn.”

    The first part of this sentence is clear, but I do not think the evidence provides “confirmation” of your conclusion in the second part. Just because we have more extant portions of GMatthew and GJohn doesn’t tell us anything about how many early copies of each of the synoptic gospels were created. For example, without knowing where any non-extant copies were made we have no information regarding their rate of production, where they were written, how long each copy lasted, or how many remain to be found.

    With such a small number of early extant portions of Mark I think any suggestion that Mark was copied less frequently than GMatthew and GJohn “in this early period” can only be a hypothesis. If you know of a way to test this hypothesis I would like to know what it is.

    • David: It’s an inference, but a reasonable one I think, that the comparative number of surviving copies of works likely reflects the comparative number original produced. Also, citations of GMark in early church writings are much fewer than for the other Gospels, which further suggests that GMark wasn’t read as often.

      • Dr. Hurtado: Your inference assumes that Mark was treated in a similar way to the other gospels. However, you wrote that “GMark wasn’t read as often,” i.e. that it wasn’t treated in a similar way to the others. My inference would be that once Matthew or Luke came out people no longer either wanted Mark or chose to quote from it, and so a higher percentage of the copies of Mark were simply not preserved.

      • Uh, David. The more likely scenario is that once Matt & John and Luke were known, Mark just wasn’t copied as often. So fewer copies survive. We use the same logic in attempting to estimate the relative frequency of copying classical texts. Homer comes out on top!

  4. Fragment of Mark, Jesus seminar, gospel of Jesus’ Wife, Bart erhman, crossan, ossuaries, etc.
    Biblical scholarship needs to regain the trust of the public.

    • Uh, I think you’re mixing apples & oranges. E.g., Ehrman isn’t untrustworthy, though I may think him wrong on some matters.

  5. S Walch permalink

    As an interesting bit of info, the line-spacing (or leading between lines, to take from Johnson, Bookrolls and Sribes in Oxyrhynchus) of 0.5cm is also that seen in P45, which has letter widths of around 0.2-0.3cm too, though sometimes certain letters can get to 0.4cm in width (β/α/δ being most common for 0.4cm).

    Looks a little less compact on P45 mind, as the writing area itself (w x h: ca. 16cm x 19.5-20cm) is larger than the actual size of the codex that P137 comes from.

    Could do with a database somewhere that records the average size of letters and leading between lines in codices and bookrolls. Just because I’m nerdy like that.

  6. ROY KOTANSKY permalink

    Although you and I have not always seen eye-to-eye in the past (with my apologies), I find your comments and hypotheseis here very acute and highly perceptive. Thank you for sharing them.

Comments are closed.

%d bloggers like this: