BUG with Clause Search?

I had the L4 Bible Study Library plus additional resources I had purchased. I upgraded to the L5 Minimum Crossgrade because I was interested in the the Clause Search. However I'm finding myself a little confused about some of the results obtained when playing around with it.

Search Example:

My search clause options were Entire Bible in the The Lexham Greek-English Interlinear New

I did a search on subject:moses in conjunction with various verb options that I "thought" would give the same (or at least roughly similar) results. Instead I got very different results as indicated below. If these are the results more experienced Logos users would anticipate, then I'm thinking there are grammatical issues involved that I don't "see" and perhaps I should just go with the "Core Datasets" upgrade--at least at this stage--and cut my upgrade cost in half. Any feedback would be appreciated.

Results (Verb, Number of hits, Scripture References)

verb:to say 7 Matt 7:24, Mark 7:10, Acts 3:22, Rom 10:19, 15:10, Heb 9:20, 12:21

verb:to say, to tell 2 Acts 7:26, 7:37

verb:to speak 0

verb:to speak, to say 2 Acts 26:22, Heb 7:14

verb:to say, to speak 1 Luke 9:31

verb:to talk, to speak 0

verb:to talk with,
to converse with 1 Luke 9:30

verb:to talk with,
to discuss with 0

verb:to speak,
to converse, to talk 0

verb:to talk together 0

UPDATE: Not sure if it will take, but as suggested, I tried changing the title of this topic to include the word "bug".

Find more posts tagged with

Comments

MJ. Smith

Logos has said they haven't implemented the Boolean operators yet - I think you may have found a related bug. If you edit your post to put BUG at the beginning, Logos should notice and give an answer

(see just above your post on the right side for the edit option).

Rick Ausdahl

https://community.logos.com/discussion/comment/426210#Comment_426210

Thanks, MJ Smith.

Guess I don't know for sure if my clause search results reflect a "bug" or a "feature" [:)], but I changed the topic title per your suggestion.

MJ. Smith

https://community.logos.com/discussion/comment/426230#Comment_426230

Some of you options should have the same results ... however, I found the same Greek word translated by the same "say" that appeared if one asked for a single word but not if you asked for two.

Rick Ausdahl

https://community.logos.com/discussion/comment/426281#Comment_426281

Thanks again MJ. Smith, for looking at this.

I figured my search results had to refelect one of two things--a major bug or a misunderstanding of the feature on my part. Either way, it's of no use to me at the moment.

I'll call Logos on Monday. If they agree it's a bug, I'll wait for a fix provided they expect a resolution in the near future. If it's a misunderstanding on my part, then I'll have to determine (with help from others) if that's due to grammatical issues/rules at a granular level that are beyond my current level of expertise, and whether or not I'm likely to be able to reach the level needed to make use of the feature. If that's in serious doubt, I'll probably ask for a refund and drop down to the Core Datasets upgrade. I really hope that's not the case though. I was more excited about the Clause Search feature than anything else I saw in Logos 5.

Sean Boisen

https://community.logos.com/discussion/comment/426462#Comment_426462

I'm not sure exactly what you're expecting to see, but i'll take some guesses:

searching for "subject:Moses" only includes clauses where Moses is the grammatical subject. If you want all clauses that mention Moses in some way, "person:Moses" will give you more results.
at present, you can only search OT and NT individually, not both at the same time. So even though your range is Entire Bible, you're really only searching the NT.
even though the English labels for verbs overlap ("to say" vs "to say, to tell"), they're not searching the same thing: the underlying Greek verb is different. So you should expect to get different results.

I see 104 results for simply "subject:Moses", so the results you're showing don't seem wildly surprising. If something else is what's puzzling you, please post what you expect to see.

Rick Ausdahl

https://community.logos.com/discussion/comment/426793#Comment_426793

Sean, thanks for taking a look at this.

In regard to your comments...

searching for "subject:Moses" only includes clauses where Moses is the grammatical subject. If you want all clauses that mention Moses in some way, "person:Moses" will give you more results.

No, I didn't want all clauses that mention Moses in some way--just those that were dealing with passages reflecting something Moses had said.

at present, you can only search OT and NT individually, not both at the same time. So even though your range is Entire Bible, you're really only searching the NT.

I was not expecting any scripture with OT references (chapter/verse). I knew it was NT only and was thus using the Lexham Greek-English Interlinear NT

even though the English labels for verbs overlap ("to say" vs "to say, to tell"), they're not searching the same thing: the underlying Greek verb is different. So you should expect to get different results.

Here are a couple of examples along the lines of what I anticipated.

The verb to speak, to say to give the same results as the verb to say, to speak.
The verb to say to be either a superset or a subset of the verb to say, to tell--not a totally different set with no overlap at all between them.
The results for the verb, to talk with, to converse with to be comparable results to those for the verbs to talk with, to discuss with; to speak, to converse, to talk; to talk together.
At a higher level, I expected the same (or at least very similar) results across the first six verbs listed in my original post, and the same (or very similar) results across the last four verbs in that post.

Basically, I wanted to see every occurence of Moses being quoted in the NT. One of the advertised benefits of the clause search (as I understood it), is that you don't need to know Greek/Hebrew--you can use English verbs--so I expected my English search verbs to give all the scripture references that would reflect those English verbs in the English translation(s) I specified for display in the search results, regardless of the nuances of the underlying Greek/Hebrew. I certainly didn't expect the English verb to speak, to say to give different results than the English verb to say, to speak. It seems I may have misunderstood this feature and am in fact faced with needing to know/understand the underllying Greek/Hebrew grammar at a fairly granular level in order to anticipate the search results that will be returned. Unfortunately, I don't know either language at all.

[:(]

MJ. Smith

https://community.logos.com/discussion/comment/427032#Comment_427032

Sean, here's the comparison that left me convinced that something was wrong:

why does the right show fewer and different results for the same verb lemma where the arguments are separated by a comma in the input - as given from the drop down menu? I can come up with a number of hypotheses but no real answer.

MJ. Smith

https://community.logos.com/discussion/comment/427035#Comment_427035

Okay folks I finally figured this out ... see http://community.logos.com/forums/t/60050.aspx Logos just didn't think the user interface through all the way. The clause search actually works as intended ... maybe even correctly. The argument notation however comes from confusion himself.

Dave Hooton

https://community.logos.com/discussion/comment/427077#Comment_427077

The clause search actually works as intended ... maybe even correctly.

That's somewhat confusing given your conclusion in that thread (which I agree with, so don't try to say anything different!)

MJ. Smith

https://community.logos.com/discussion/comment/427096#Comment_427096

That's somewhat confusing given your conclusion

[:D]

Rick Ausdahl

https://community.logos.com/discussion/comment/427077#Comment_427077

MJ. Smith--Thanks for the link http://community.logos.com/forums/t/60050.aspx. I checked it out and am glad to see I'm not the only one more than a little confused. But... if the kind of results I got from the 10 searches in my original post are considered by Logos to be working as intended, I'm very disappointed and I don't think it's working as advertised by the features page for this feature.

For example, on the features page for the clause search it says, FIND SPECIFIC ACTIONS: Type in “subject:Jesus verb:heal” to find every time that Jesus heals someone. I don't know about other folks, but what this tells me is that using this clause search will return all the scripture passages where Jesus heals someone, whether or not a Greek word for "heal" is specifically used in the passages.

Instead, I'm now told that I should expect the verb to say, to tell to give totally different results than the verb to say. Now, I might be able to see the logic of getting fewer results with the verb to say, to tell, IF it was explained that the comma separating to say and to tell is not being used as an OR indicator, but as a LEVEL separator, meaning it only returns a particular form/usage of the verb to say. But in that case, it seems to me it should return a subset of the verb to say, not totally different results. I might also be able to see the logic of getting more results with the verb to say, to tell than with just the verb to say IF it was explained that the comma separating to say, to tell is being used as an OR indicator. But in that case, the results should be a superset of the verb to say. But instead, I get totally different results--absolutely no overlap in the scripture references returned. And I'm really baffled that the verbs to say, to speak and to speak, to say give totally different results, yet are working as expected. What percentage of English speaking people would expect those English verbs to return totally different results?

So I'm left wondering--just how do I use the clause search to find every reference in the NT where Moses is being quoted, regardless of whether the Greek nuances distinguish between saying, speaking, telling, discussing, conversing, etc., etc. In fact, based on the way this feature is now being presented as working, I'm wondering how the Logos 5 example of this feature as noted above (finding every occurence/passage where someone is healed by Jesus) can be achieved if the only verb used in the clause search is heal. Perhaps I'm mistaken, but I'd be surprised if the word "heal" is actually used in every instance where Jesus heals someone. I would check it out, but I don't have either a Bible or Logos available at the moment.

fgh

https://community.logos.com/discussion/comment/427035#Comment_427035

Sean, here's the comparison that left me convinced that something was wrong:

why does the right show fewer and different results for the same verb lemma where the arguments are separated by a comma in the input - as given from the drop down menu? I can come up with a number of hypotheses but no real answer.

Did you notice Ac 3:22 (left) and Ac 7:37 (right)? The grammar's a bit different, but the content of those two verses is pretty much identical: Moses said + Deut 18:15.

Dave Hooton

https://community.logos.com/discussion/comment/427178#Comment_427178

why does the right show fewer and different results for the same verb lemma where the arguments are separated by a comma in the input - as given from the drop down menu? I can come up with a number of hypotheses but no real answer.

I agree that Logos should regard the comma as separating two phrases.and that we should be allowed to OR terms in the query e.g. verb:to say, verb:to tell, verb:to speak will match lexical values "to speak, to say, to tell", "to report, to announce, to tell", "to tell, to order"

MJ. Smith

https://community.logos.com/discussion/comment/427170#Comment_427170

And I'm really baffled that the verbs to say, to speak and to speak, to say give totally different results, yet are working as expected.

I actually think I understand the search if Logos gives me a way to distinguish multiple senses from compound names of senses. I can now correctly explain the results - I just can't do a comprehensive job of predicting them. I don't want to add to the confusion until I get a clear answer from Logos.

fgh

https://community.logos.com/discussion/comment/427288#Comment_427288

why does the right show fewer and different results for the same verb lemma where the arguments are separated by a comma in the input - as given from the drop down menu? I can come up with a number of hypotheses but no real answer.

That quote is from MJ, not me!

Rick Ausdahl

https://community.logos.com/discussion/comment/427170#Comment_427170

I can now say from my own test that the clause search does NOT necessarily perform as advertised on the features page for this search--at least on my install of L5.

Listed below are three excerpts from an earlier post I made today at 6:30 AM.

...if the kind of results I got from the 10 searches in my original post are considered by Logos to be working as intended, I'm very disappointed and I don't think it's working as advertised by the features page for this feature.
For example, on the features page for the clause search it says, FIND SPECIFIC ACTIONS: Type in “subject:Jesus verb:heal” to find every time that Jesus heals someone. I don't know about other folks, but what this tells me is that using this clause search will return all the scripture passages where Jesus heals someone, whether or not a Greek word for "heal" is specifically used in the passages.
...based on the way this feature is now being presented as working, I'm wondering how the Logos 5 example of this feature as noted above (finding every occurence/passage where someone is healed by Jesus) can be achieved if the only verb used in the clause search is heal. Perhaps I'm mistaken, but I'd be surprised if the word "heal" is actually used in every instance where Jesus heals someone. I would check it out, but I don't have either a Bible or Logos available at the moment.

I'm back home now and have been able to try the search regarding "Jesus healings" as given in that example on the features page for the clause search. It does NOT find every time that Jesus heals someone. As just one example of a missing healing, consider the healing of the blind man starting in John 9:1.

I also see a discrepancy/problem with another example of the clause search on the features page, where it says: "Logos 5 finds everything" and then gives an example of finding verses for Jesus AND Bethlehem. But I'll save that for later discussion.

In general, I thought the search clause was going to make it much easier to find things, but it feels like it's going in the opposite direction. [:S]

MJ. Smith

https://community.logos.com/discussion/comment/427454#Comment_427454

In general, I thought the search clause was going to make it much easier to find things, but it feels like it's going in the opposite direction

I've found that verbs and related are the most problematic out of the following list of options:

adverbial: --> English
adverbial-surface: --> original languages
indirect-object: --> English
indirect-object-surface:--> original languages
location:
object: --> English
object-surface:--> original languages
participant:
person:
place:
polarity:positive
polarity:negative
related:
subject: --> English
subject-surface:--> original languages
thing:
verb: --> English
verb-gloss:
verb-lemma:
verb-ln: --> Louw-Nid
verb-morph:
verb-surface:--> original languages

I am convinced that there is a disconnect on verbs that Logos has to clarify. And I think that there are some coding errors - but not a high rate. From earlier testing I know there are a group of oddities such a phrases in apposition that I'm not sure I'm convinced by Logos' grammatical argument but their position is understandable.defensible.

Rick Ausdahl

https://community.logos.com/discussion/comment/427471#Comment_427471

I would very much like to hear from a Logos employee whether or not from the company's perspective this feature is working correctly (as advertised on the feature page) or if they agree that at least in the case of verbs, there are problems.

If Logos agrees there are problems and plans to address them in the near future, I'll probably stay with the minimum crossgrade provided a refund will be available down the road if things aren't resolved in what I consider a reasonable time frame. If they don't feel there are problems, I may go back to L4--or at the very least ask for a partial refund and just go with the core datasets.

I do hope they acknowledge problems and plan to fix them. The clause search was my primary motivation for upgrading. If it turns out not to work as advertised, yet per Logos, is working as intended, I'll be disappointed, but I'll get over it and I'll continue using Logos, whether it's L4, or L5 with just the core datasets.

MJ. Smith

https://community.logos.com/discussion/comment/427623#Comment_427623

I would very much like to hear from a Logos employee whether or not from the company's perspective this feature is working correctly (as advertised on the feature page) or if they agree that at least in the case of verbs, there are problems.

Several of us want to know. I'll admit to checking the forums frequently, hoping for the answer,

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/427623#Comment_427623

I would very much like to hear from a Logos employee whether or not from the company's perspective this feature is working correctly (as advertised on the feature page) or if they agree that at least in the case of verbs, there are problems.

If Logos agrees there are problems and plans to address them in the near future, I'll probably stay with the minimum crossgrade provided a refund will be available down the road if things aren't resolved in what I consider a reasonable time frame. If they don't feel there are problems, I may go back to L4--or at the very least ask for a partial refund and just go with the core datasets.

I do hope they acknowledge problems and plan to fix them. The clause search was my primary motivation for upgrading. If it turns out not to work as advertised, yet per Logos, is working as intended, I'll be disappointed, but I'll get over it and I'll continue using Logos, whether it's L4, or L5 with just the core datasets.

We will be changing this in 5.0b (and will probably keep making improvements based on your feedback).

I don't know all the details, so please keep an eye on the Logos Desktop Beta forum for announcements.

Rick Ausdahl

https://community.logos.com/discussion/comment/428835#Comment_428835

Thanks for the update, Bradley. I'll keep eyes and ears open. [:)]