Monday, April 06, 2009

Comparing GNT Texts (and Manuscripts)

Over at Biblical Studies and Technological Tools, Mark Hoffman has posted an evaluation of how some of the available software tools can handle comparisons of Greek New Testament texts (editions). The fuller version of the comparison is available here). He has used Accordance8, BibleWorks8, Logos3, and the online Manuscript Comparator (by Weston Ruter).

Hoffman concludes:

Each program has its strengths and weaknesses. Manuscript Comparator does the best job of displaying differences, but it lacks the NA27, and results cannot be saved. Accordance does a good job of display and creates useful lists of differences, but only two texts at a time can be compared. BibleWorks has the most versatility, but it is difficult to save results. Logos has the most texts available for comparison and results export easily, but one must get accustomed to the way it displays differences

I think many people do these type of comparisons. It would be interesting to hear what our readers think about the methodology and results?

One important thing is to get the software to disregard some features that are insignificant from the genealogical perspective (e.g., editorial features such as moveable nu, etc).

In the comment section I pointed Hoffman to the software Collate and Anastasia which are more suitable for advanced manuscript comparisons and the production of critical apparatuses. These tools are used for the NT transcript prototype (INTF, Münster) and the forthcoming digital NA28. I used Collate 2.0 for my comparison of 560 MSS in Jude. Read more about those tools here. BTW, Anastasia and Collate have their own blog nowadays (but it seems not to be updated very often).

In this connection I should also mention the on-line software LaParola which allows for comparisons and includes a vast amount of editions and manuscripts. It is also possible to insert the whole on-line site into BibleWorks.


  1. For those interested in the total number of differences between a MT and a CT, here is an email my friend sent me. He got it from LaParola, I think. (BTW, LaParola actually provided a complete printout of each 'variant.' This LaParola person is good!!)

    I can't easily do an exact count, but approximately I think it is about 6,980. However, the difference in numbers (from Wallace at 6,577) is due not only to the different texts, but the fact that it depends on the way you define a difference or a variant. For example, if one text has the words A B C and another D B E, is that one difference or two? Even A B against C D could be one or two differences, depending upon how closely connected the words are. For a concrete example, the first variant listed is in Matthew 1:5, where NA has BOES twice and Byz has BOOZ twice. I counted that as one variant, although some might count it as two (because there are two differences).

    Probably some "fundie" who removed the BOOZ :o )

  2. Question on variants.... Doesn't the newer term "variant unit" try to reduce (make more accurate) the number of variants counted, such as ABC compared to DCB, since Greek word order may vary with no intent of varying the contents of the proposition (maybe just the emphasis).

  3. Mark has done a good job in representing the data here, but a few notes are in order.

    In regards to Accordance, as was mentioned in the comments, he is running it on a PC under emulation. While this is all fine and well, the speed at which it runs is considerably less than on a Mac (n.b. Decker's comments on the speed to generate differences in Mark and the entire NT). Further, he (and others) have commented that it would be nice to ignore 'editorial differences' in the text (i.e. movable nu's). This can be done in Accordance by comparing 'Lemmas,' which would automatically ignore any differences that do not affect the lexical form. He also does not explain why Accordance only compares two texts at a time, and how it is very easy to swap back and forth between the two texts compared (see David's explanation in the comments previously linked to).

    Regarding more text-critical matters, he consistently misspelled Comfort & Barrett's work (spelling it Comfort & Barrow!). Also, while listing the availability of codices א, B, D and W in Accordance, he does not mention the possibility to compare these texts as well (likely because they aren't included in the Scholar's Standard level which he has). The compare texts feature works very nice here with the 'lemma' option, and will ignore editorial differences such as final nu's indicated by an overbar. This can also be done with the papyri in the Comfort and Barrett module since the Accordance version treats them as a separate text module both in canonical and mss order, with full morphological tagging.

    I am very interested to see how the BW projected underway will be able to compare and work with mss. Kent Clarke presented his work in San Diego, and looks to be very promising in their scope and attention to detail (with several people checking and re-checking the work). It was unfortunate that his presentation was scheduled at the same time as Parker's session on his book.