Word Translation Rings Anomaly

Kevin Purcell raised this question, and it seems it should become an issue. In testing the word translated 'ill' in John 11:1 (ESV) using the PG and then clicking on the Greek word to produce a Word Study (it could be done directly as a Word Study, too), the Word ring results are substantially different from 4.0a SR-3 and 4.0b Beta 9. Please refer to this thread to view the particulars. It needs some explanation.

http://community.logos.com/forums/t/11576.aspx

Find more posts tagged with

Comments

George Allakhverdyan

Mark you mentioned in that thread in a later post than this one

"I have gone back to look at the ring in 4.0a SR-3. The small segments to the upper right and upper center are for combinations of words which also include the base Greek word in them (sort of a phrase translation than just focusing on the original Greek word translation). Examples are 'weak person', 'Lazarus was ill', etc. In the Beta ring graph, these seem to have been merged under the translation for the base Greek word, as I think they should (for the above two examples, merged under 'weak' and 'ill'). So I think Logos is cleaning up the underlying source (The ESV RI text) or the way it is handled for this report."

Do you still feel this an issue that needs to be given attention to, is there problems in the beta?

Mark Smith

https://community.logos.com/discussion/comment/90770#Comment_90770

I think it would be good to clarify if what I think is going on is in fact the case. Has the programming been changed in this way? (I don't recall seeing that in all the beta release notes, but I don't read them scrupulously.) Otherwise there could be a bug. Rather than me having to do a lot of detective work, someone at Logos should have a ready answer.

Thanks for asking.

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/90891#Comment_90891

The ring graph was changed (in 4.0b Beta 1) to group solely under the "primary translation" term instead of removing stop words (we, he, you, I, etc.) and grouping by the remaining terms (the algorithm in 4.0a). This removes extraneous segments from many graphs. For example, for ἀσθενέω, "was ill" is under "ill", but "became ill" is its own segment. In particularly bad cases, sometimes all of the translation terms (e.g., both "we" and "ought" for ὀφείλω) were in the stopword list, so there was nothing left to group by.

Mark Smith

https://community.logos.com/discussion/comment/91576#Comment_91576

Thanks for the explanation. The new approach is much better.