Release Announcement: open-cbgm

Bible Study software, Unicode, Fonts, Keyboards, creating Web pages in Greek, and other software issues.
Post Reply
Brent Niedergall
Posts: 25
Joined: August 22nd, 2018, 1:00 pm

Release Announcement: open-cbgm

Post by Brent Niedergall »

It was only a matter of time until someone reverse engineered the Coherence-Based Genealogical Method. The “black box” is now open and freely available thanks to Joey McCollum. His C++ implementation is open-source, compact, fast, and compliant with established standards of the field. It runs on Windows, MacOS, and Linux. And for a first, it can produce a complete global stemma of a text (very quickly at that). You can see Joey has done so for 3 John.

Get it here: https://github.com/jjmccollum/open-cbgm

From the readme:
The Coherence-Based Genealogical Method (CBGM) is a novel approach to textual criticism, popularized by the Institut für Neutestamentliche Textforschung (INTF) for its use in the production of the Editio Critica Maior (ECM) of the New Testament. It is a meta-method, combining methodology-dependent philological decisions from the user with efficient computer-based calculations to highlight genealogical relationships between different stages of the text. To establish genealogical relationships in the presence of contamination (understood to be a problem in the textual tradition of the New Testament), the CBGM makes a number of philosophical and methodological innovations, such as the abstracting of texts away from the manuscripts that preserve them (and the resulting rejection of hypothetical ancestors as used in traditional stemmata), the encoding of the textual critic’s decisions in local stemmata of variants, and the use of coherence in textual flow to evaluate hypotheses about the priority of variant readings.
Brent Niedergall
Stephen Carlson
Posts: 3350
Joined: May 11th, 2011, 10:51 am
Location: Melbourne
Contact:

Re: Release Announcement: open-cbgm

Post by Stephen Carlson »

Brent Niedergall wrote: January 17th, 2020, 4:41 pm It was only a matter of time until someone reverse engineered the Coherence-Based Genealogical Method. The “black box” is now open and freely available thanks to Joey McCollum. His C++ implementation is open-source, compact, fast, and compliant with established standards of the field. It runs on Windows, MacOS, and Linux. And for a first, it can produce a complete global stemma of a text (very quickly at that). You can see Joey has done so for 3 John.

Get it here: https://github.com/jjmccollum/open-cbgm
Thanks for that. I think this is a step forward for greater understanding of the method, but I have a question: why does the "global stemma" have multiple roots unlinked to the initial text 'A'?
Stephen C. Carlson, Ph.D.
Melbourne, Australia
jjmccollum
Posts: 4
Joined: January 17th, 2020, 5:55 pm

Re: Release Announcement: open-cbgm

Post by jjmccollum »

There are two reasons the global stemma is disconnected. First, whenever fragmentary witnesses (e.g., GA 365, which has only 10 extant readings in 3 John) are not excluded, it is possible that they will have equal priority to the A-text at every passage where they are defined. As a result, they have no potential ancestors according to the CBGM and will be isolated in the global stemma. Second, and more importantly, whenever a reading in a local stemma has an unclear relationship to the rest of the stemma, it becomes impossible to explain that reading in the highest-priority witness that supports it. Because an optimized substemma must account for every reading in a witness, no feasible substemma will exist for that witness. This happens for a handful of witnesses in 3 John, including GA 03 (with reading j at 3Jo 1:4/2-10) and 35 (with reading c at 3Jo 1:12/28-30). The problem goes away if the disconnected local stemmata are ignored (which we can do by commenting out the variation units in the input XML file), or, preferably, if every local stemmata accounts for all of its readings.
Stephen Carlson
Posts: 3350
Joined: May 11th, 2011, 10:51 am
Location: Melbourne
Contact:

Re: Release Announcement: open-cbgm

Post by Stephen Carlson »

Thanks for your speedy response.
jjmccollum wrote: January 17th, 2020, 6:11 pm There are two reasons the global stemma is disconnected. First, whenever fragmentary witnesses (e.g., GA 365, which has only 10 extant readings in 3 John) are not excluded, it is possible that they will have equal priority to the A-text at every passage where they are defined. As a result, they have no potential ancestors according to the CBGM and will be isolated in the global stemma.
OK. I believe Mink's provision for cases of equal priority is to link them with a horizontal bar. At any rate, the INTF has a provision for ignoring fragmentary witnesses (not much can be said about them in any case).
jjmccollum wrote: January 17th, 2020, 6:11 pmSecond, and more importantly, whenever a reading in a local stemma has an unclear relationship to the rest of the stemma, it becomes impossible to explain that reading in the highest-priority witness that supports it. Because an optimized substemma must account for every reading in a witness, no feasible substemma will exist for that witness. This happens for a handful of witnesses in 3 John, including GA 03 (with reading j at 3Jo 1:4/2-10) and 35 (with reading c at 3Jo 1:12/28-30). The problem goes away if the disconnected local stemmata are ignored (which we can do by commenting out the variation units in the input XML file), or, preferably, if every local stemmata accounts for all of its readings.
This is a little hard to follow; perhaps a more detailed exposition would be helpful. At any rate, I've got a few questions. Where are you getting the local stemmata from? At both places, the INTF punts on the local stemma, showing a '?' as the origin of all but the 'a' reading. (Is this what you mean by 'disconnected local stemmata'?) Are you using the INTF's local stemmata (so that's the cause of the problem), or do you generate your own?
Stephen C. Carlson, Ph.D.
Melbourne, Australia
jjmccollum
Posts: 4
Joined: January 17th, 2020, 5:55 pm

Re: Release Announcement: open-cbgm

Post by jjmccollum »

Equal priority could be supported, but this would allow the theoretical possibility of two witnesses that connect to each other in the global stemma to the exclusion of potential ancestors with more prior readings. This would also result in a disconnected global stemma. I chose to avoid this possibility by imposing a strict ordering of potential ancestors in terms of priority.

I got my local stemmata from the INTF Genealogical Queries Tool, which is indeed the cause of the problem. The "?" placeholder, as far as I can tell, indicates that some other reading in the local stemma is the source of the readings below it, but the editors did not decide which reading was the source. Functionally, this disconnects the local stemma, as no path exists from the initial reading to the readings under the "?", and therefore, these readings cannot be "explained" by readings elsewhere in the local stemma for the purpose of substemma optimization.

To illustrate what happens without these problems, I've removed the variation units with unclear reading sources from the collation input file and used a minimum extant readings threshold of 11 to filter out the fragmentary witnesses GA P74 and 365. The GitHub page (https://github.com/jjmccollum/open-cbgm) now includes an image of the resulting connected global stemma.
Stephen Carlson
Posts: 3350
Joined: May 11th, 2011, 10:51 am
Location: Melbourne
Contact:

Re: Release Announcement: open-cbgm

Post by Stephen Carlson »

Thanks. The new image is more like what I expected to see.
Stephen C. Carlson, Ph.D.
Melbourne, Australia
Post Reply

Return to “Software”