BUG: Search results for loved in Romans

L4 4.5 RC-1 Conducted Bible Search for "loved" in Romans (all forms)

Analysis View, noticed 7 search results listed lemma as ὁ. Have never seen the article translated as love before.

See this thread in the L4 Mac Beta forum http://community.logos.com/forums/p/43973/327251.aspx#327251 as this is not a Mac specific problem, but a gross tagging error in all Reverse Interlinears.

Though I did not try this with any other word, I am pretty sure it will exist for every English word rendered from more than a single word in the original.

Find more posts tagged with

Comments

Dave Hooton

This is typical of the reverse interlinears, Jack - not just the word "loved". What is more, there are gross inconsistencies between different RI's for the same passage.

Jack Caviness

https://community.logos.com/discussion/comment/327329#Comment_327329

This is typical of the reverse interlinears, Jack - not just the word "loved". What is more, there are gross inconsistencies between different RI's for the same passage.

Understood. But it makes the Analysis feature of Bible Search of questionable value.

Dave Hooton

https://community.logos.com/discussion/comment/327331#Comment_327331

Understood. But it makes the Analysis feature of Bible Search of questionable value.

It reports all words in the one cell as the original for the translated word.

Eph 5.33 in ESV "houtos agapato" is translated simply as "love". Look at the way NIV re-arranges the Greek words and positively indicates that houtos (word 11) is not translated!

Jack Caviness

https://community.logos.com/discussion/comment/327342#Comment_327342

Look at the way NIV re-arranges the Greek words and positively indicates that houtos (word 11) is not translated!

See that, and also see that the ESV RI gives the impression that houtos is the word translated as "love". Since I like Interlinears almost as much George does, I had not noticed that tagging was done so poorly. Logos really needs to fix this.

Rick Brannan

https://community.logos.com/discussion/comment/327342#Comment_327342

Hi Dave.

Eph 5.33 in ESV "houtos agapato" is translated simply as "love". Look at the way NIV re-arranges the Greek words and positively indicates that houtos (word 11) is not translated!

(Responding to the analysis issue and not the indexing issue)

I'm not sure exactly why it is said the NIV "re-arranges" the Greek words. The NIV is English; the Greek is lined up as best as it can be given the translation. There are places where decisions have to be made, and in those places one makes them and moves on.

My thought (as best as I can recall) on ουτως in this verse is that in situations like this it can be translated "in the same way"; there is some equivalence going on in the comparison between loving one's wife and loving one's self. That fits best right after "must love his wife" (and, following "his wife", is actually in Greek order), as in: "must love his wife [in the same way] as he loves himself". I suppose I could have linked both words (ουτως ως) to "as", but that has other issues. Another alternate would be to conceive of ουτως as "so" (probably the most popular way it is translated in the NIV) and do something like "must [so] love his wife as he loves himself" but that doesn't seem too NIV-like.

LEB takes a different tack and is a bit more stilted but at least gives English "thus" for ουτως, "must thus love his own wife," making the alignment easy.

The ESV NT alignment, under the supervision of the ESV editorial board, tended to want to include as much Greek as possible. That is, they didn't want to leave any Greek words untranslated (bulleted) unless they couldn't get away from it. So they put it where they put it.

These sorts of things happen when one wants to completely account for all words in both texts, which is what we try to do with our reverse interlinears. Particles, articles, adverbs, conjunctions and interjections are important. But language is messy, and trying to align two of them together is messier still.

Dave Hooton

https://community.logos.com/discussion/comment/327520#Comment_327520

I'm not sure exactly why it is said the NIV "re-arranges" the Greek words. The NIV is English; the Greek is lined up as best as it can be given the translation.

Rick, sorry about the way it was expressed but I was primarily concerned about the different ways of indicating that "houtos" was not translated.and its effect on a Search.

There are places where decisions have to be made, and in those places one makes them and moves on.

My sentiment goes with you but my desire for consistency remains e.g. there are passages where the primary English word is not aligned with the Greek but other RI's make the alignment. I haven't written these down so I can't provide an example at this time.

Jack Caviness

https://community.logos.com/discussion/comment/327520#Comment_327520

Rick

I understand your problem, but ουτως is still not the word translated "love" in this verse, nor does the article mean "love" as the initial post in the referenced thread discovered.

At the very least, align the English words with the correct Greek/Hebrew word. Current alignment makes results from the Analysis feature of Bible Search suspect.

Rick Brannan

https://community.logos.com/discussion/comment/327551#Comment_327551

Jack & Dave --

Multiple issues I'll try to respond to here.

First, my response was only regarding ουτως in Eph 5.33. Our ESV NT alignment was coordinated with, reviewed and approved by the ESV folks. So one can debate it, but that's where they wanted to put ουτως. We have similar constraints with our NLT NT alignment. In my previous response I was trying to explain that we have some constraints in these types of things. I was also trying to explain that placement of bullets in English is typically driven by where the aligner thinks the term would/could be translated. This is one of the reasons why an untranslated post-positive δε is almost always represented with a bullet at the beginning of the English clause (where the translation would go) instead of one word/phrase into the clause (where it is in the Greek order).

Regarding the include-the-article issue. This is a known issue. In our raw alignment data, we have distinctions between 'primary' and 'secondary' alignments. The articles, particularly in the cases like Jack cites here, are distinguished as 'secondary' alignments with the primary alignment being between love and αγαπη.

On consistency: it simply isn't that simple, particularly when translations themselves are not consistent internally in the way they translate (note: I'm fine with that, I don't think translations should be concordant because languages don't work that way).

One example would be the (contrived in my head just now) phrase αιμα του Χριστου. You might translate this as "blood of Christ", "Christ's blood", "blood of the Christ", or render Χριστου as "Messiah", or even some way I haven't just now thought of. If the translation was "Christ's blood", we'd align του Χριστου -> "Christ's" (with του as secondary) and αιμα -> blood. Does this mean that if another translation has "blood of Christ" that we should mark 'of' as untranslated to be 'consistent'? Or insert a bullet in the English for the seemingly missing 'the', considering the article to be untranslated?

I agree that we need to do some overall analysis of alignments to make similar things similar; that's one reason why it is already on our tasklist. (Yes, it really is.) But this is a long term project, not an immediate one. One step will (hopefully) be to come up with some way to elegantly and simply respect the primary/secondary linkage data in our raw alignment data; that (again, hopefully) will make most of the problems like those Jack brings up disappear (note, however, complex idioms would continue to be a sticking point, they will require a different solution). I've long advocated for something like this internally, but also know enough about the problem on the indexing side (remember, I'm a data guy, not an application programmer guy) to know it is difficult.

Apologies if I come across as overly defensive, but these particular issues are complex and seem to get more complex the deeper one digs into them. They are also near and dear to me as I have a lot of personal time and effort invested in making these alignments for Greek materials. I (and we) do appreciate the feedback. Know that some sort of consistency analysis and revision is on our radar, but it is not something that will be implemented in the near-term.

Keep Smiling 4 Jesus :)

https://community.logos.com/discussion/comment/327618#Comment_327618

I agree that we need to do some overall analysis of alignments to make similar things similar; that's one reason why it is already on our tasklist. (Yes, it really is.) But this is a long term project, not an immediate one. One step will (hopefully) be to come up with some way to elegantly and simply respect the primary/secondary linkage data in our raw alignment data; that (again, hopefully) will make most of the problems like those Jack brings up disappear (note, however, complex idioms would continue to be a sticking point, they will require a different solution). I've long advocated for something like this internally, but also know enough about the problem on the indexing side (remember, I'm a data guy, not an application programmer guy) to know it is difficult.

Wonder about adding translation linkage data to word tagging ? that would allow Analysis to have a Linkage column, which could be used to group search results. Thus, articles could be grouped as secondary.

If translation linkage is added to word tagging, would also appreciate tagging words added by translators for smooth English rendering (possibly in conjunction with some Bibles that italicize added words). Likewise Greek words that are not translated, could have linkage tagging of untranslated (to join dot used in Reverse Interlinear display).

Note: the translation linkage tagging could be a one character code (possibly added with @L for use in Morph Search).

Keep Smiling [:)]

Dave Hooton

https://community.logos.com/discussion/comment/327618#Comment_327618

I agree that we need to do some overall analysis of alignments to make similar things similar; that's one reason why it is already on our tasklist.

"making similar things similar" is what I mean by "being consistent"; the contrived example is not that.

Know that some sort of consistency analysis and revision is on our radar, but it is not something that will be implemented in the near-term.

And I'm grateful for that.

Now, if we can only get the metadata fixed near-term[:D]

Jack Caviness

https://community.logos.com/discussion/comment/327618#Comment_327618

Regarding the include-the-article issue. This is a known issue. In our raw alignment data, we have distinctions between 'primary' and 'secondary' alignments. The articles, particularly in the cases like Jack cites here, are distinguished as 'secondary' alignments with the primary alignment being between love and αγαπη.

Understand the complexities of coding this, but half the results of the search that started this discussion are incorrect and therefore worthless. It's not they type of search I would normally do, but Mo Proctor described it in Camp Logos 1, and so I gave it a try. When the results were inconsistent with reality, I created this thread, and the one in the Mac Beta Forum.