Text Comparison - Majority Text Hodges/Farstad overlap

I assume by Majority, you mean Hodges. 

Years back, one of the curious-ities of neural net analysis, was that Robinson/Pierpont was much closer to 'author-ial' than the NA series. Meaning, for example, a 'Jude' pattern matched back against Jude, was far more consistent in the Byzantine, than the NA group (any author style will only partially match, since humans vary, as they serially write (or dictate).  Paul was more Paul. Luke more Luke. And so on. This suggested either the amalgam process in the Byzantine was more attuned to author, or the underlying mss's being used were more attuned.

At the same time, it accidentally became clear the neural nets were most interested in conjunctions, particles, prepositions .... glue words, not vocabulary or morphing. If you simply stripped the glue out, author matching was difficult to impossible. If you smoothed the other data, the neurals could still locate author.

A book in Logos sort of discusses the problem and is well recommended:

https://www.logos.com/product/9413/sentence-conjunctions-in-the-gospel-of-matthew-kai-de-tote-gar-sun-and-asyndeton-in-narrative-discourse 

It's a more traditional look at the problem (statistical) but discusses similarly ... authors, at least in the koine, seem to distinguish at the discourse level, and slightly below.

Returning to Hodges, it'd be interesting, but you'd need someone pretty disciplined. Translation is often a tribal affair ... absent clear differences, look like the tribe's wording.

A good example of this is the Source translation (panned a month back). Some of the translation looks as if written by an early church father (vs modern 1900s translations, as most now tie to about the 1960s).

Dave Hooton said:

Give it an abbreviated title

True, but poor coding. Columns should never overlap.

Sounds like we’re back to Lotus!