Resource bug: Transliterated headwords are defined as 'English'

Several dictionaries use transliterated headwords. As a minimum:

Eerdmans Bible Dictionary
Nelson's New Christian Dictionary
New Strong's Dictionary of Greek and Hebrew words.
Vine's Complete Expository Dictionary

Unfortunately, in every occasion these transliterated headwords are tagged with the language 'english', rather than the language 'translit'. As a result, none of these resource return lookups from transliterated text.

To reproduce:

search for the word 'Christos' in your library, in its transliterated form, using the syntax translit:Christos. You may need to turn match all forms off.
Open a result, then double-click on the word 'Christos' in the resource.
The power lookup tool will open, and no dictionaries will return results.

Now:

Search for english:Christos to find results that aren't marked as transliterated. The results normally display faulty tagging (we'll overlook that [:)]), or a personal name.
Again, open a result, then double-click on the word 'Christos'.
This time a dictionary will open to the entry on 'Christ' or Χριστός.

Expected result:

That the dictionary/power lookup works when the text is marked as 'translit'.

<edit>Following discussion below, it seems that it might be preferable for the headwords to be marked as both 'translit' and 'english', because around 20% of occurrences in the text seem be incorrectly marked as English. I'll leave others to discuss whether two wrongs make a right!</edit>

Find more posts tagged with

Comments

Mike Binks

Thank you Mark - well (and understandably) put.

NB.Mick

Several dictionaries use transliterated headwords. As a minimum:

Eerdmans Bible Dictionary

Nelson's New Christian Dictionary

New Strong's Dictionary of Greek and Hebrew words.

Vine's Complete Expository Dictionary

Unfortunately, in every occasion these transliterated headwords are tagged with the language 'english', rather than the language 'translit'. As a result, none of these resource return lookups from transliterated text.

To reproduce:

search for the word 'Christos' in your library, in its transliterated form, using the syntax translit:Christos. You may need to turn match all forms off.

Open a result, then double-click on the word 'Christos' in the resource.

The power lookup tool will open, and no dictionaries will return results.

Now:

Search for english:Christos to find results that aren't marked as transliterated. The results normally display faulty tagging (we'll overlook that ), or a personal name.

Again, open a result, then double-click on the word 'Christos'.

This time a dictionary will open to the entry on 'Christ' or Χριστός.

Expected result:

That the dictionary/power lookup works when the text is marked as 'translit'.

Mark: I fully agree with the issue - however I might think that the solution in re-tagging the English theological dictionaries might be shortsighted. As you know there's transliterated text everywhere. It is found in Vyrso books and PBs (in both it will not be able to be tagged as 'Translit' language but be either English or Greek) and it is found for all inflected forms of all Greek words in the NT, e.g. in the UBS Handbooks, thus the handful of lemmas in those dictionaries won't do. Transliterated Greek text is not another language - it is still Greek, just written with other characters. The Greek lexicons would take it up much better than the theological dictionaries (and there is already an Analytical lexicon of NT Greek - actually I think there are two, one from Lexham and one from the Fribergs). It might be better to re-tag such text as Greek, or even English, and have all versions picked up by the Analytical Lexicon.

MJ. Smith

https://community.logos.com/discussion/comment/783264#Comment_783264

A related aside: In the early testing of the concordance feature I reported hundreds of systematic errors in the language coding.

Mark Barnes

https://community.logos.com/discussion/comment/783264#Comment_783264

It might be better to re-tag such text as Greek, or even English, and have all versions picked up by the Analytical Lexicon.

I've said elsewhere that it would be helpful to have an analytical lexicon that works for transliteration.

But I don't understand why you think it's better to retag all transliterated text as 'Greek' or English', rather than simply to add transliterated headwords into the lexicon. (Apart from anything else, there's no way Logos is going to retag thousands of books.)

MJ. Smith

https://community.logos.com/discussion/comment/783270#Comment_783270

why you think it's better to retag all transliterated text as 'Greek' or English'

Actually I find it annoying that I can't find Sanskrit, Pali, Avestan, Old Persian ... and other interesting languages because they are hidden under the tag "transliteration". I personally don't care whether it is in some other script or the Roman alphabet ... Pali doesn't even have a standard script. Why I care is what language it is and whether or not I can read or muddle through it... The current position is untenable for multi-faith commentaries such as John P. Keenan's The Wisdom of James.

Mark Barnes

https://community.logos.com/discussion/comment/783273#Comment_783273

Actually I find it annoying that I can't find Sanskrit, Pali, Avestan, Old Persian ... and other interesting languages because they are hidden under the tag "transliteration". I personally don't care whether it is in some other scripture or the Roman alphabet ... Pali doesn't even have a standard script. Why I care is what language it is and whether or not I can read or muddle through it...

Understood. It would have been better still if Logos had designed the transliteration system to be able to specify the original language.

But retagging all transliteration with the original language will take tens of thousands of hours. I wasn't asking why that would be nice to have. I was really asking why this was felt to be a productive use of time, particularly when there's an alternative solution that achieves most of that and would take 'only' a few hundred hours.

MJ. Smith

https://community.logos.com/discussion/comment/783274#Comment_783274

But with community tagging and sermon labels Faithlife is beginning to allow us to fix such things in the resources that matter to us. If they improved their tagging moving forward and recoded a few key resources, they could fix rather than perpetuate the problem.

BTW have you found anything with Personal Letter labels?

NB.Mick

https://community.logos.com/discussion/comment/783270#Comment_783270

It might be better to re-tag such text as Greek, or even English, and have all versions picked up by the Analytical Lexicon.

I've said elsewhere that it would be helpful to have an analytical lexicon that works for transliteration.

On this we are fully in sync - and I think that a revised edition of https://www.logos.com/product/4580/lexham-analytical-lexicon-to-the-greek-new-testament and https://www.logos.com/product/27295/lexham-analytical-lexicon-to-the-septuagint-hb-swete-edition could easily have the words tagged as Greek, English and 'Translit' headwords and then we're done for the Greek side of it.

But I don't understand why you think it's better to retag all transliterated text as 'Greek' or English', rather than simply to add transliterated headwords into the lexicon. (Apart from anything else, there's no way Logos is going to retag thousands of books.)

I'm not even sure Logos would retag the handful of theological dictionaries that carry transliterated headwords that aren't tagged as such. I have several points why I think this is not right (just voicing opinion, and it might well be that Logos follows your suggestion nevertheless):

"Translit" is not a language. We have transliterated text from all over the globe - wherever a non-Latin script is used. In fact this text is Greek, or Hebrew, or Chinese or whatever - just written in Latin characters. So - regardless of the implementation - from a purist's standpoint the tagging is wrong.
The wrong dictionaries will be picked. Logos should pick the right lexicons - which are those for the original language - for the right words. As much as I deem the theological dictionaries helpful, the software should (maybe via the interim step of an analytical lexicon) pick the user's highest prioritized lexicon of the respective language.
If the tagging was Greek or English, from what some users tell, it might possibly already work today - at least for lemmas.
I don't see Logos tagging thousands of books where the transliterated text is not tagged in that weird 'Translit' language, either. It definitely won't happen for Vyrso books, and afaik there's no way for us users to tag our PBs as such in MS Word. So whatever the solution is, it should work for all three language tags that are existing for Greek transliterated text. Retagging some (arbitrary) dictionary headwords to 'Translit' would make the situation only marginally better. I rather thought that Logos might change a line in the program code to understand "Translit' as English since that would make it work against the dictionaries that have entries for transliterated words tagged as English.

Mark Barnes

https://community.logos.com/discussion/comment/783285#Comment_783285

I'm not even sure Logos would retag the handful of theological dictionaries that carry transliterated headwords that aren't tagged as such.

I'm not suggesting that.

Retagging some (arbitrary) dictionary headwords to 'Translit' would make the situation only marginally better.

Or that.

I'm not suggesting Logos add transliterated headwords to any resource. I'm only asking them to correct the language errors in transliterated headwords that already exist — or at least in one or two resources (Strong's and Vine's).

NB.Mick

https://community.logos.com/discussion/comment/783287#Comment_783287

I think we should clarify that.

We both see the best solution for this in analytical lexicons that work for all transliterated original language words, however tagged.

Currently it seems that there are dictionaries that have transliterated headwords tagged as English - subsequently they work for resources where these transliterated words are contained and tagged as English, but not when thy are tagged as 'Translit'.

Your suggestion would be to retag those transliterated headwords into language Translit and thus have an index of Translit Headwords - then the respective dictionaries will work for resources where transliterated words (actually lemmas) are used and will cease working for those, where the transliterated text is tagged with English language (which includes all Vyrso and PB resources plus a good number in the Logos editions). I personally don't see this as language errors since Translit is not a language. Whether as a stopgap measure this would help Mike's issue from the other thread? Might be. It looks like trying to close a gap by opening another to me.

Mark Barnes

https://community.logos.com/discussion/comment/783294#Comment_783294

Your suggestion would be to retag those transliterated headwords into language Translit and thus have an index of Translit Headwords - then the respective dictionaries will work for resources where transliterated words (actually lemmas) are used and will cease working for those, where the transliterated text is tagged with English language (which includes all Vyrso and PB resources plus a good number in the Logos editions).

I agree with this analysis.

FWIW, a simple search for hesed and theos in my library reveales 4,001 results in English, and 27,131 results in transliteration. So there would be a very definite improvement if the headword language was also 'translit'. (And several of the 4,001 results are actually correct because they refer to 'Hesed' as a proper noun.)

I personally don't see this as language errors since Translit is not a language.

In Logos-speak, translit is a language. That's the way they've designed the system.

NB.Mick

https://community.logos.com/discussion/comment/783303#Comment_783303

FWIW, a simple search for hesed and theos in my library reveales 4,001 results in English, and 27,131 results in transliteration. So there would be a very definite improvement if the headword language was also 'translit'. (And several of the 4,001 results are actually correct because they refer to 'Hesed' as a proper noun.)

Interesting. The results are quite different for me. And if I look at books instead of individual search hits, I don't a large difference percentagewise between Translit and English (and sometimes there are more on the English side, sometimes more on the Translit. I found Greek tagging of theos or christos only in one PB each, thus this doesn't complicsate the issue)

(sorry, screenshot uses German UI)

EDIT: the thousands of hits for translit theos come from the Lexham interlinears (NT, LXX etc) that use this language coding - seems the other interlinears and RIs are not tagged at all for the transliterated line - thus books is the right comparison metric)

MJ. Smith

https://community.logos.com/discussion/comment/783532#Comment_783532

Mark Barnes

https://community.logos.com/discussion/comment/783532#Comment_783532

EDIT: the thousands of hits for translit theos come from the Lexham interlinears (NT, LXX etc) that use this language coding - seems the other interlinears and RIs are not tagged at all for the transliterated line - thus books is the right comparison metric)

I re-ran my search excluding Bibles, and returned the following:

English: 3,888 in 1,222 resources
Translit: 15,079 in 1,781 resources

I don't think resources is the best metric (once Bibles are excluded). By the way, I tranlisterated as חֶסֶד as hesed, not chesed. The former is much more common (9,400 vs 801 in my library).

NB.Mick

https://community.logos.com/discussion/comment/783543#Comment_783543

Mark,

I'm not disputing the commonality of the transliteration, I only remembered 'chesed' as an alternative transliteration (which should be included in a future solution that provides for Hebrew - even if it was only for the well-known lemmas) and included it for completeness. Overall, I guess your library contains some high-value commentary sets and/or journals which are tagged with this language.

(as an aside: we discuss this in two threads without crosslinking them - https://community.logos.com/forums/t/119096.aspx was what triggered this one)

NB.Mick

<edit>Following discussion below, it seems that it might be preferable for the headwords to be marked as both 'translit' and 'english', because around 20% of occurrences in the text seem be incorrectly marked as English. I'll leave others to discuss whether two wrongs make a right!</edit>

Just saw your edit in the OP (I highlighted the most relevant part) and I think you are right, this would be a good short-term solution for the lemmas included in dictionaries - while still looking forward to a complete analytical lexicon of transliterated bible text.

Eli Evans (Logos)

https://community.logos.com/discussion/comment/783622#Comment_783622

Skimming: I'll alert Kyle of this thread to address the book issues.