Monday, June 24, 2019

Palaeography of an invoice

One thing I’ve already seen since Holmes’ email last night (see some discussion by Brent Nongbri here and here) is a questioning of the authenticity of the First-century Gospels invoice. Some have cautioned against taking it at face value. I even had one conservative Christian ask me directly this morning if I think it is real.

I don’t have any reason to doubt that it’s real, but I also think we can quantify that a little bit. Are we not text critics? Is it not part of our job to analyze handwriting on handwritten documents? I don’t mean to make light of a very serious situation, but I do think it could be helpful to post an analysis like this. I freely admit that I am not trained in contemporary forensic handwriting analysis, so my thoughts here should not be taken as definitive. I am only analysing the letters as I would give an informal analysis on the fly if a friend asked me to describe the letters in a Greek NT manuscript. I’m pretty sure the date of the Mark fragment itself is proof that an opinion can change when something is studied in greater detail.

With that in mind, I offer this assessment of the handwriting of two samples. The first is the handwriting on the invoice where it is signed Dirk Obbink, which I designate INV in the discussions that follow. I have not used the signature line in the comparison, because it is qualitatively a different style—it is a ‘signature’ style, rather than a ‘print’ hand used elsewhere.

The other is the handwriting of the paper note, which I designate PAP.
Descriptions for each letter are below the screenshots of them.

1, 3

For the numeral 1, INV uses both hooked and non-hooked forms. PAP uses non-hooked forms. The numeral 3 in both INV and PAP has a lower curve that is larger than the upper curve, but in both hands, the difference is not always strongly pronounced.


For the numeral 2, both INV and PAP lack a loop at the lower left part of the numeral, and the horizontal is fairly straight. I am not sure what we can tell from the correction, but I give it here anyway. Both instances of 2 in INV have a ‘low’ and broad curve. One instance of the PAP 2 has the same, but the other two are more rounded.

 In the number 5, the top horizontal is straight, it connects to a short vertical at almost a right angle, and the curve is not too rounded. One instance of a PAP 5 is almost identical to the INV 5. None of the instances of 5 has a curve that turns and come back upward.


 The numeral 6 in both INV and PAP begin with a near-vertical downstroke that ends in a counter-clockwise loop. The size of the loop varies, but the formation is the same in both hands.


 7 seems reasonably distinctive. I’m not sure how many people add the horizontal line in the middle. I do, but that is because of my science background and the need to distinguish 7s and 1s on paperwork. 7 in both INV and PAP lack a small hook at the top left that is occasionally present in handwritten 7s by others, but they both include the horizontal line in the middle. The angle at which the two main strokes of 7 connect is similar, though there is variation in PAP.


 The best I can tell from the low resolution is that the 9 might be a two-stroke character. The loop is not perfectly round but wider than it is tall, and the vertical is nearly completely straight.


 H is made of three separate but straight strokes. In INV, the horizontal can extend to the left beyond the first vertical. Distinctively, in every instance, the horizontal is near the bottom of the letter. The two halves are not symmetrical; the top half is noticeably larger than the bottom half.

R, K

 R and K appear in the same sequence in both INV and PAP. In most instances of R and K, the vertical is slightly slanted, leaning left. The lower oblique extends out at the same angle. R is made in two movements: a vertical, then the round and oblique in one movement. K is similarly made in two movements beginning with the vertical. The two obliques of K always connect to the vertical, but the curve and oblique of R occasionally do not, with a non-connecting instance in both INV and PAP. Unusually, both INV and PAP have an instance of R in which the loop is much larger than the area required for the oblique below it.

Internal U
 Although we have only one instance of internal U in each sample, they are identical. The U of UK in INV (visible above in the images for R, K) begins higher, but that is probably due to its position beginning an abbreviation. Even the Initial U has a similar letter formation in that it comes back down in a final vertical without lifting the writing instrument. The two verticals in internal U are uneven, with the right one longer and higher. The bowl is shallow, and the right vertical extends below the bowl.


 E is formed of straight strokes, and the center horizontal extends beyond the vertical. The vertical and lowest horizontal are probably formed in a single two-stroke motion without lifting the writing instrument.

upper-lower asymmetry

 In characters that could normally be divided symmetrically between upper and lower halves that contain curved strokes (e.g. 3, B, 8), there is a noticeable tendency to use asymmetrical forms. INV consistently uses forms with larger lower halves and smaller upper halves. This tendency is there but not as pronounced or consistent in PAP. 

Ends of letters

 INV displays a noticeable tendency to end some letters in horizontals that extend beyond the normal space of the letter into (or nearly into) the following letter. There appears to be at least one instance of this in PAP (the A in Mark), and a possible second instance (E in Matthew).

In conclusion:

I think there is enough consistency across both hands to conclude that it is reasonable to assume that the hand that signed the invoice as Dirk Obbink was the same hand that wrote paper note identifying New Testament text. Higher-resolution photos would of course be ideal, but what we have shows a consistency across both hands with some identical or very similar letters and letter formation patterns.  In addition to higher-resolution images, one source of information could be metadata with the original photograph of the note. Often you can tell things like what camera was used to take a photograph, whose computer was used to create the file, etc. from that metadata. Also helpful would be to have a handwriting sample that is indisputably that of Obbink. What I have done here is not exactly the same as giving a palaeographical assessment of an ancient fragment—I know my own handwriting has a lot of internal inconsistency, but still, it looks to me from what is extant here that these two samples were written by the same person.


  1. There is analysis, over-analysis, and pointless over-analysis. I suggest this post is one of them.

    1. I'm happy to delete the post (or allow a blog editor to do so) if you think it should be deleted. I thought something like this—asking if the two writing samples could be the same hand [without passing judgment on *who's* hand it is] on the basis of comparing the letters like we do with papyri—could be a little bit of a change of pace from the sensational journalistic tone this issue has had for several years now (to which I admit contributing my share of cat gifs).

    2. This is gold. I favour keeping it online.

    3. I'm not proposing deleting it. I just can't quite believe you wrote it.

    4. This comment has been removed by the author.

  2. Now that the cat is out of the bag, it seems inevitable that there will be legal proceedings that will settle any question about whether or not this is authentic. Either Obbink engaged in and can be proven guilty of very serious fraud, or either Dr. Holmes or someone else engaged in very serious libel against him.

    For my part I am completely confident that the invoice is genuine and that Obbink engaged in fraud.

  3. Why would the real Dr Obbink write Christ Church as "Christchurch"?

    1. The second "C" looks capitalized to me.

    2. I mean, all the letters are caps. But it looks larger.

  4. You guys never cease to amaze.

  5. If I were being picky, I would note that only page 4 of the invoice document is damning; it is not in dispute that Dirk Obbink sold papyri to the Greens, and pages 2 and 3 could easily have originally have referred to those. Equally it is not in dispute that Dirk Obbink discussed loaning a set of EES papyrus fragments to the Green collection for inclusion in their 'Passages' exhibition in the Vatican for Easter 2012; and was authorised to do so. So the photographed list (which is nowhere referenced in the invoice) could have an innocent explanation. It might have been possible to substitute the current page 4 for one relating to other fragments entirely.

    On the other hand there must have been earlier written communications from the prospective vendor of FCM to the Green collection before February 2013, as Dan Wallace reports being made to sign a non-disclosure agreement by Jerry Pattengale (as a formal requirement of the unknown vendor)'later in 2012'. So, if Mike Holmes has found the February invoice in the files of the Green collection, he could well have corresponding earlier correspondence setting out the terms under which images of the FCM papyrus could be shared.

  6. It might be well if EEF convey an image of the 5345 record card.
    And I hope, if he is interested, that Revel A. Coles would comment.
    Minor point: The Matthew parchment fragment is described as on skin of a cow, "ranch raised." How might one know the latter? Maybe merely a Texas/Oklahoma aside meaning domestic.