Wednesday, March 01, 2006

Manuscript transmission experiment

Rick Brannan has set up a manuscript transmission experiment, involving modern copying of NT manuscripts. Here are the first and second reports on the project. This is something I've thought about myself, in consultation with Peter Head. I had conceived of the experiment somewhat differently from Rick. Start with one text, copy it twice, copy each copy twice, etc. Devise a simple unique label for every ms involving a number for its 'generation' and then A or B for whether it is the first or second copy, plus further designators as necessary. I'd hope that at least 1054 copies could be produced that way. We could then look at the copying errors that occurred and do various tests to see whether it was possible to restore the original on the basis of partial evidence. Ideally the experiment would be carried out on a mass volunteer basis, with certain controls, e.g. that people do not copy a copy that they themselves have produced. Volunteers would submit their copies in scanned form.

The experiment could work in Greek or in English. An English version would open up the possibility of more volunteers and possibly many thousands of mss.

My favourite copying experiment actually would not be to produce manuscripts. I'd start with a bit of the KJV in original spelling on computer and a user interface inviting people to copy it by typing. Perhaps it should all be block capitals. Once two copies have been produced, these then form the foundation of the four copies of generation 2, etc. Since all data would be entered electronically it would be very simple to analyse it. Obviously certain mistakes would occur on qwerty or azerty keyboards that would not occur in manual copying. However, it might still be possible to get some useful observations on various forms of parablepsis.

Why KJV in original spelling? This would be to mimic simultaneously the familiarity and linguistic strangeness that many a Greek scribe must have felt when copying the NT.

Aim: 100,000 copies of 1 chapter of the Bible. Who's going to set this up?


  1. Hey, thanks for the mention. In the shorter term, I hope to post images of each page of each submitted MS, collate them, and generate transcriptions from the collations. Once that is done, I'd like to consider selecting a MS or two and using those to create further generations of the text. Then collate, etc. Lather, rinse, repeat.

    It might be easier to roust volunteers with English, but one of my secondary goals was to become more familiar with the Greek of 2 Timothy -- and what better way to do that than to collate a bunch of different MSS?

    Rick Brannan

  2. If you want to find out what it was like to write in the past, you are going to have to throw away your modern felt tipped pens and get either metal pens or reeds to write with and dip then in the ink like all the rest of us. Once you realize that by dipping the pen in the ink and then writing, the time delay of using the pen will affect the accuracy of the writing.

    Writing the Bible is not that difficult, the errors can be seen easily enough as you have a bit more time writing the page than you would think. The errors actually come about when I am writing chapters that are similar to each other, that start with the same line. That is when the errors creep into the manuscript.
    IN John chapters 18 and 19 I wrote an interlinear greek latin and english using the complutensian Polyglot. As it was the only manuscript I had on hand to use. I do not read Greek or latin and I use the handwriting of 15th century to write the Greek. I used a dictionary to look up the words to place them in line with the KJV. It took a month for each book. But then again, I learned quite a bit and have come up with a great concordance.
    If you use modern pens expect your results to be not accurate.
    The United Methodist News Service had an article on my bible a few weeks ago and is going to have Video TV production out next week on my work on the web site. You can find out about my hand made illumianted mansucript of the bible at my web site google "The Pepper Bible."

    If you are looking for the original KJV, are you talking about the 1611 or 1612 or the 1715?
    And be careful, the American KJV are missing words, punctuation, and books for that matter!

    James G. Pepper
    Antiquarius Domini

  3. "If you want to find out what it was like to write in the past, you are going to have to throw away your modern felt tipped pens..."

    I copied the Apocalypse of John and the Gospel of John using a gold plated Caran d'Ache ball point pen which my father gave me 35 years ago. He was given the pen as a gratuity after doing some consulting for Brown Boveri in London. You cannot make mistakes with a Caran d'Ache. I wrote all my exams in seminary with it. Did wonders for my grades. :-)


  4. I think there is room for several experiments that test different things. Writing with reeds is a great idea, but will probably not be able to generate many thousands of manuscripts.

    As far as KJV's go, I'd go for what is probably the first printing of 1611 (see David Norton, A Textual History of the King James Bible). There's a cheap reprint of this by Hendrickson. The art work of the first edition is essential for authenticity :-). Since acquiring a copy of the editio princeps I rarely refer to other editions of the KJV.

  5. James Pepper said: 'Once you realize that by dipping the pen in the ink and then writing, the time delay of using the pen will affect the accuracy of the writing.'

    Excellent! For a stimulating discussion of this phenomenon see:

    P.M. Head & M. Warren, 'Re-inking the Pen: Evidence from P. Oxy. 657 (P13) concerning unintentional scribal errors' New Testament Studies 43 (1997) 466-473.

    This is on-line at

  6. PJW: "Start with one text, copy it twice, copy each copy twice, etc. [...] I'd hope that at least 1054 copies could be produced that way. We could then look at the copying errors that occurred and do various tests to see whether it was possible to restore the original on the basis of partial evidence.

    Such an experiment could be helpful when attempting to establish and categorize the various types and relative frequencies of scribal error.

    I doubt whether the experiment would offer any real help toward restoration of the precise form of the original NT text, given its parameters.

    The system proposed is closed, and thus excludes mixture from readings found in other exemplars, as well as independent scribal alteration (whether an attempt to "repair" a difficulty or error, or where a prior familiarity with the passage might lead to correction of what appears in the exemplar).

    At best, this experiment represents a "worst-case scenario", with each stage of such a closed transmission resulting in a text less accurate than its immediate predecessor. Had this actually occurred in NT transmissional history, the case made by Ehrman in Misquoting Jesus would have strong merit.

    NT Greek manuscript transmission, however, is characterized by mixture. W-H long ago recognized that mixture utterly skews genealogical descent and textual restoration based on classical stemmatic principles. They therefore attempted to force genealogy and stemmatics into a quite new and different mould in order to retain restoration of the original as a valid goal.

    While the closed system experiment (which perhaps should be conducted) might permit a true stemma to be constructed and the original text of that experiment to be restored by careful analysis, such would say very little regarding the actual Greek manuscript transmission of the NT. It would, however, assist in the recognition and categorization of scribal habits.

  7. The scribe will copy the best example they have, and if they can get a copy of the original work, they will do so.The idea that they are copying one manuscript is wrong, they will use as many authentic copies as possible. I have always beleived, given my experience in writing 1500 finished pages of the Bible as well as all of the pages I had to stop for errors, that the differences in the manuscripts was a matter of translation, not writing. Espcially if you consider the diffeences in the latin vulgate. How many revisions took place in the early church due to problems of translation and the multiple sources. How many of the grek texts were the result of translation problems from multiple sources?

    I catch most of the errors when I write the pages, and then I proofread it over and over and over again. And given that these manuscripts would be in the area where they were made and people would be looking at them, any other errors would be noticed and corrected within a relatively short time frmae from their being made. People are looking for errors all the time in my work, mainly to try to "catch" me. It is very strange. No these differences in these manuscripts have another reason. To blame it all on scribal error, and yes there are scribal errors, is an easy answer made by people who have not the experience of actually writing the text.

    If you use the 1611 will you correct the he/she error?

    James G. Pepper

  8. I've done this a number of times with both lay classes and grad text crit classes. Though I never tried to produces thousands of copies, I have run over 50, including misture (by giving a student 2 or more copies from which to produce the next one), and have actually produced a textual aparratus of the results in NA-style. I've only used it to illustrate, however, not to assess actual scribal habits. Interesting and helpful to the students--in their opinion.

  9. The results of such an experiment would be compromised by the difficulty of finding in the 21st century a group of professional scribes with the same cultural background and training as a first century professional scribe.

    According to the normal canons of social research a 21st century control group doesn't tell you much about what would happen in antiquity.


  10. Prof. Robinson, I agree with everything you say. Any transmission experiment would have rather limited aims and would not be attempting to replicate many significant aspects of the NT transmission process.

  11. I'm all for it. Take away experimentation from science and all you have left is dogma.

    I'm somewhat of a newcomer to NT TC, having entered the field as an outgrowth of my interest in the Textual Criticism of the King James Version. My editio princeps is of 1980's vintage; a Thomas Nelson reprint of the Cambridge Press 300th anniversary commemerative reprint of 1911. I recently got the chance to see the 1911 edition, and was surprised to find it something like 5 volumes in single columns. Computerised typesetting is indeed a boon!

    There is no current critical edition of the KJV. I believe producing one would be a helpful experiment in TC, as well as a valid attempt to restore the text of the original ms, which was lost, apparently, in the London Fire of 1666.

    Although Scrivener collated all the editions to his day, subsequent comparisons of his collation with the Nelson edition have turned up many additional variants.

    I don't know how thoroughly the Nelson edition was proofread against the original printing(s), but I have been impressed by several instances of an inverted u or n, obviously a painstaking reproduction of the orinal hand-set type.

    I've spent several years trying to read entirely through my editio princeps (I should be done within the month), and have frequently been sidetracked by what I initially thought was a printing error, which turned out after much study to just be a little bit of Middle English carried over from a "former translation diligently compared" but not "revised."

    Example: 'or ever' as a synonym for 'before' in Daniel 6:24. I thought it must be 'ere ever', but sure enough, it was no misprint. Shakespeare used the term in Hamlet, and also as 'or ere' in King John.

  12. If you use the 1611 will you correct the he/she error?
    --James G. Pepper

    It can't be corrected with any certainty. Each reading is found in one of the original printings (Oxford/Cambridge), and the Hebrew is no help at reconstructing the original, as the Masoretic Text also contains both variants.

  13. I must say that I think that textual criticism on the KJV is on the whole not so interesting. I enjoyed reading Norton's Textual History, but as Norton said, full collation of the variants within the KJV tradition is not possible. There are simply too many printings, each with their minor variations. Moreover, early on even one printing could have variations since pages were printed individually and pages from different print runs might be mixed in binding. Even KJVs from the same 1611 edition are not identical in letter sequences. Before about 1730, when stereotyping (a metal cast for each page) was introduced, moveable type also allowed the alteration of individual letters during a single 'printing'.

    My suggestion of use of the KJV was only that it might serve as a basis for an experiment to find out about human errors in copying. For this purpose, the Geneva Bible or Bishops' Bible would be equally good. All we need is a version in a form of English that is rather different from that of the copyists so that we can see whether linguistic updating is common and how much copyists conform to the spelling of their own day.

    Another variant on the experiment involves the creation of an artificial new texts with 'traps' (e.g. sentences with similar endings) to try to see the extent to which these increase the likelihood of error. For this purpose something bible-like could be used.

    The advantage of using a form of the Bible as the base text is that you get interference from people's own familiarity with the Bible.

    Part of the initial questionnaire for anyone volunteering to copy would involve them disclosing their educational level and an assessment of their familiarity with the Bible.

  14. Then why not use Tyndales New Testament, the one he printed after the frist one that is in modern english; I think its the 1537 edition. It is not in the Old English typeface and its in modern english but its variation from the KJV would be enough that people might write what they remember as the KJV instead of the actual text. I used it in a Psalter writing it and the KJV side by side. It is almost the same as the KJV. Handel used it for the Messiah.
    You could try Matthews Bible or the Great Bible but I think this is just the right balance to make the effort interesting.

  15. "Not in the Old English Typeface"

    Why not? How better to simulate the transmission of the text from uncial to minuscule?

  16. Would you randomly destroy parts of the earliest copies to simulate papyrological evidence?

  17. Why don't you orgainze the experiment where people write down verbatim "quotations" of leactures they heard twenty or forty years ago?

  18. Christian, I don't think we'd destroy anything, though it would be interesting to 'hide' copies, and see how accurately we could reconstruct what happened on the basis of partial evidence.

    Anon, I'm not sure what the experiment about lectures is meant to be testing. People nowadays will not be able to give verbatim quotations from lectures they heard yesterday unless the lecturer used a particularly memorable turn of phrase or if they wrote down a quotation at the time, or sought to memorise it at the time.

  19. Sounds like a little literary criticism on the origin of the verbiage of the Gospels.

  20. Did this experiment happen? If so, what were the conclusions?