Tagging problem in GNT, UBS, NA, LXX tagged texts

[Rod Decker]

I just noticed a problem in the tagging systems used by the various Greek texts. I was curious how often ὅτι occurred as a conjunction as opposed to it being the neut sing nom/acc indefinite relative pronoun. Most current texts print the pronoun as ὅ τι, but some have used ὅτι. I was initially a bit frustrated when a search for

 

[PRONOUN relative indefinite neuter sing]

 

produced no hits in some texts. I finally figured out that some of these texts (e.g., UBS4-T) tag the two pieces separately as a rel pron and an indef pron, but other texts (e.g., NA27-T) tag the two items together even though they are printed as two separate words. (The space between is not a regular space apparently since both parts hi-lite together.) The second would be, I think, highly preferable as a tag. The problem is that if you select "ὅ τι" in NA27-T and do a right click, Search for... inflected, you get an error beep and no search is performed. I think that's indicative of a problem either in the tagging or the way that Accord handles the inflected search with this sort of tagging.

 

I know this is in part related to the fact that diff texts have been tagged by different teams of scholars and the tagging schemes are not consistent. It would be a tremendous help if all tagged Greek texts used a consistent set of tags, e.g., all used 2aor tags (highly preferred) or none did. The biggest diffs are btwn NT texts and LXX, but there are diffs within the NT tagged texts as well.

[Rick Bennett]

I just noticed a problem in the tagging systems used by the various Greek texts. I was curious how often ὅτι occurred as a conjunction as opposed to it being the neut sing nom/acc indefinite relative pronoun. Most current texts print the pronoun as ὅ τι, but some have used ὅτι. I was initially a bit frustrated when a search for

 

[PRONOUN relative indefinite neuter sing]

 

produced no hits in some texts. I finally figured out that some of these texts (e.g., UBS4-T) tag the two pieces separately as a rel pron and an indef pron, but other texts (e.g., NA27-T) tag the two items together even though they are printed as two separate words. (The space between is not a regular space apparently since both parts hi-lite together.) The second would be, I think, highly preferable as a tag. The problem is that if you select "ὅ τι" in NA27-T and do a right click, Search for... inflected, you get an error beep and no search is performed. I think that's indicative of a problem either in the tagging or the way that Accord handles the inflected search with this sort of tagging.

 

I know this is in part related to the fact that diff texts have been tagged by different teams of scholars and the tagging schemes are not consistent. It would be a tremendous help if all tagged Greek texts used a consistent set of tags, e.g., all used 2aor tags (highly preferred) or none did. The biggest diffs are btwn NT texts and LXX, but there are diffs within the NT tagged texts as well.

 

We're utilizing the same morph db for GNT-T, UBS4-T and NA27-T so theoretically the results should be the same. We can pass this on for review to find out why they are not.

 

For the LXX, as you probably know, this is maintained by someone else and their methodology does differ at times. We are going to be working on an update of that for the Fall and if you have any specific items, we can potentially discuss them and recommend changes; but it's not up to us (as far as I know) whether they are adopted or not.

 

Thanks for the feedback and report.