Reliability of Morph searches in Logos vs Bibleworks vs Accordance

Page 1 of 4 (75 items) 1 2 3 4 Next >
This post has 74 Replies | 6 Followers

Posts 3646
Francis | Forum Activity | Posted: Wed, Oct 22 2014 2:16 AM

Let's state from the onset here that I am not trying to advertise Bibleworks or Accordance or enter into a debate about which software is best. Please do not highjack this thread with this conversation.

AS A LOGOS USER, here is the problem:

1. Doing PhD work, we search a root in Hebrew, we get a count. Then we read on the root and find a different count in a lexical resource. After toying with different type of searches, we finally find the missing entries (as I recall, it was cognate participles).

2. The word is out in the academic community that Accordance or Bibleworks are more reliable when it comes to language work. As I heard it from an eminent professor, different Logos teams sometimes use different criteria to tag language resources.

What I am looking for: 

Feedback from users who are REALLY COGNIZANT of these issues. Are you aware of some of the discrepancies when comparing what these different programs produce? Do you know -- for a fact (please no guesswork) -- what account for those discrepancies (or possibly, simply different approaches)? What does it mean for scholarly work reliance on Logos language resources (esp., morphologically tagged primary texts and associated functionalities)?

The purpose: if there is a problem with Logos tagging, what specific feedback can we give so that it can offer a better product for languages in particular.

What would be counter-productive: wasting everyone's time by going on a "defend Logos" campaign. This is not about criticizing Logos or coming to its defense. It's about asking whether there is a liability that can certainly be improved to make it a better academic product.

Posts 3646
Francis | Forum Activity | Replied: Wed, Oct 22 2014 6:47 AM

BTW, compare the following results:

(same count for all English Bibles with Logos morphology).

When one does a BWS and click on the root, all of the above results (that is, a number of English Bibles and 1 Hebrew text) will appear. Clearly the results are not consistent though I am sure there is a reason for it. The point however is this: if an English Bible is searchable by root, should it not have the same count as a Hebrew Bible?

Posts 7943
LogosEmployee

Francis:
BTW, compare the following results:

Thanks for posting a specific example.

In this case, it appears that LEB is missing some roots. I've opened an inquiry about this.

Posts 3646
Francis | Forum Activity | Replied: Wed, Oct 22 2014 7:56 AM

Thanks, Bradley. Actually ALL English Bibles with Logos Morphology (NASB, ESV, NKJV, etc.) have the same, lower count, as compared to LHB.

Posts 2181
Forum MVP
John Fidel | Forum Activity | Replied: Wed, Oct 22 2014 7:58 AM

I think to be comparable, you need to run searches on the original language texts and compare. You cannot even run a comparable original language search using an English text with the other two software programs. You can run a Strong's search.

With that said, I do appreciate you bringing these differences to Logos attention as the original language searches should be scholar grade as well.

I am not proficient in Hebrew, but would suspect that those would be the texts that show the most differences.

Posts 49
LogosEmployee
Chris Robbers | Forum Activity | Replied: Wed, Oct 22 2014 9:49 AM

Bradley Grainger (Faithlife):
In this case, it appears that LEB is missing some roots. I've opened an inquiry about this.

This will be fixed in an upcoming update to the LEB reverse interlinear resource. Thanks for the report!

Posts 1502
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Wed, Oct 22 2014 10:15 AM

As Francis mentioned, it's not just the LEB. It's every single Bible that's not named LHB, whether it is in English or not, including, for example, LHI, Logos LXX, LXXALT, NRSV, etc., etc.

Posts 433
Vincent Setterholm | Forum Activity | Replied: Wed, Oct 22 2014 10:17 AM

Chris Robbers (Faithlife):

Bradley Grainger (Faithlife):
In this case, it appears that LEB is missing some roots. I've opened an inquiry about this.

This will be fixed in an upcoming update to the LEB reverse interlinear resource. Thanks for the report!

We actually have a significant update to LHB and all the resources that get their morph data from that source - LHI, the Reverse Interlinears, etc. - sitting in the wings. The last obstacle to getting it out the door is waiting for a new release of the 5.3 beta, as these resources won't be recognized by beta 3. At this point I don't know if we'll ship these resources immediately when the beta is released or if we'll just roll it into the planned 'massive update' scheduled for next week.

Speaking to the broader question of data accuracy: I'm proud of this pending update, but I'm sure perfection still eludes us. The single best way to report issues with the LHB morph/lemma/root data is to use the report typo feature directly on the LHB text so that the feedback ends up in one place, instead of scattered across all the RIs. I check the reporting on that resource regularly and appreciate feedback. Thanks for reporting this issue!

Posts 3646
Francis | Forum Activity | Replied: Wed, Oct 22 2014 10:37 AM

Thank you for the response. I don't know what Roza has also noted in Septuagintal resources. Perhaps, much in the same way as there is a beta for features, there could be user-based testing of the language resources? It may be scary to do so, but a general invitation to scholars and qualified users to point out gaps and problems could help bring morphological materials in Logos to Cadillac level? I think that Logos is doing great in serving the large public but seems to be behind in terms of reputation among scholars. 

Posts 433
Vincent Setterholm | Forum Activity | Replied: Wed, Oct 22 2014 10:45 AM

David Knoll:

David, we updated to the latest Westminster data almost a year ago (version 4.18, which you probably got for free if you already had the old 4.2) and teased apocopation out into a separate field in the structure of our morph codes so that there wouldn't be problems finding hits for words that were both apocopated and had another ending tag, like paragogic nun.

Posts 1502
Forum MVP
Fr Devin Roza | Forum Activity | Replied: Wed, Oct 22 2014 10:59 AM

Francis:

Thank you for the response. I don't know what Roza has also noted in Septuagintal resources. Perhaps, much in the same way as there is a beta for features, there could be user-based testing of the language resources? It may be scary to do so, but a general invitation to scholars and qualified users to point out gaps and problems could help bring morphological materials in Logos to Cadillac level? I think that Logos is doing great in serving the large public but seems to be behind in terms of reputation among scholars. 

I simply ran your search on the Morph tab, and switched over to the Bible tab. In the Bible tab, run the search against ALL Bibles. Then switch to Grid view.

In the Grid view, what you'll see is that the exact same hits that are in the LHB but are not in the LEB are also not in the LXX or any other reverse interlinear. So, that's what I was referring to.

I can say in general that I am quite happy with the quality of the data in Logos, and in their responsiveness in fixing issues when they are identified (as this thread testifies). As well, another nice thing about Logos (at least compared to BibleWorks) is that they have so many different Morph databases. So, for serious research you can actually run the search in different databases and compare if you have any reason to doubt. 

FYI, I am constantly doing advanced searches at the Pontifical Biblical Institute and comparing to data from professors or from Biblework to verify the quality of Logos's data, and Logos has always come out just fine. Where I worry more about the quality of the data is in the areas where Logos is really breaking ground, such as the tagging of people, places, things, clause searching, etc. But that is more because these areas are so new, and are just being implemented, so they are naturally more prone to errors, or to not having been fully implemented, not because I think they aren't doing it well or haven't seen solid methodology, etc. 

As best I can tell, Logos's lack of penetration in some scholarly circles has more to do with the way Logos was 5-10 years ago than how the program is today.

Posts 911
David Knoll | Forum Activity | Replied: Wed, Oct 22 2014 7:25 PM

Vincent Setterholm:

David Knoll:

David, we updated to the latest Westminster data almost a year ago (version 4.18, which you probably got for free if you already had the old 4.2) and teased apocopation out into a separate field in the structure of our morph codes so that there wouldn't be problems finding hits for words that were both apocopated and had another ending tag, like paragogic nun.

Hi Vincent,

The poster asked how Logos earned that reputation, and I gave him an example. Other software did not have such problems before 4.18.  At the time one was unable to search the Qumran sectarian database for the verb כתב. For me, this is how Logos lost its credibility as a scholarly tool. I bought something else and mostly use Logos as an e-reader (which is why I didn't know you fixed that).  I am not sure this is part of the past. I have repeatedly asked here what is going on with the Wolfgang Richter morphology. This was the main reason I invested in Logos in the past, but I got no reply. Logos promised the Richter Morphology before Logos 5 was out, when it was out Logos promised again, two years have passed and no word. Logos used to allow the scholar to make syntax searches in Eep Telstra's syntax module, then one day it was gone (Logos 4). How can a scholar rely on Logos?  I am not saying this to cause damage to anyone. I am just saying that sadly the OP has a point, and you should listen. 

Posts 2465
Lee | Forum Activity | Replied: Wed, Oct 22 2014 11:21 PM

Yes I just feel a need to show my support for the above post. Not in the specifics maybe, but in the generalities for sure.

Posts 26104
Forum MVP
MJ. Smith | Forum Activity | Replied: Wed, Oct 22 2014 11:43 PM

David, if I may be so bold as to make a suggestion, scholars need to press Logos hard for the tools they need. Logos does want to reach the academic community but Bob is not himself an academic. He has some solid scholars working for him, but they are often working with in-house tools to create tools already envisioned. Yes, I at times have been a real pest, but I've also seen several things I championed become a reality. Where I have backed down, Logos convinced me that there was a good reason for not doing it my way because of the resources required to implement my vision. So all scholars should become pests for the tools they need to accomplish their goals ... that is how Logos will know there is a demand, that the market is greater than 1, that they'll enhance their reputation among academics if they provide the requested feature. Yell. complain, instigate a insurrection and otherwise help all of us get a better product.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 2465
Lee | Forum Activity | Replied: Thu, Oct 23 2014 12:42 AM

The posts so far provide some clues.

  • no time to waste
  • alternatives are out there
  • perceived snubs from fanboys
Posts 911
David Knoll | Forum Activity | Replied: Thu, Oct 23 2014 12:58 AM

MJ I am not even speaking of revolutionary features: It is so cumbersome to search for three Hebrew words in a row where the middle one is optional,  or two words which come one after the other and a third one which occurs anywhere in the verse. I am trying to search for the lemma שאל and the lemma שלום in the same verse and it takes me ages to figure that out (where exactly is שאל when you type sal or shal? Why can't we input with Israeli keyboard layout?). This needs to be intuitive simple and quick. The results should be reliable. The main differences between morphologies need to be described. Search examples need to be provided. You cannot expect the scholar to ask Dave Hooton (if you are reading this, thank you ever so much for your help) on the forums how to perform a syntax search. It needs to be intuitive and there needs to be a manual with examples. 

 If Vincent or Bob need to learn how scholars work with bible software I invite them to Jerusalem to see how we work in HUBP with a competing software. I challenge them to present an intuitive fast and efficient Logos solution for every task we perform, with minimum mouse clicks, minimum typing, and simple and intuitive syntax. 

Posts 26104
Forum MVP
MJ. Smith | Forum Activity | Replied: Thu, Oct 23 2014 1:13 AM

Offer Bob some great Jerusalem food and you might have a deal - he seems to love travelingSmile

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 911
David Knoll | Forum Activity | Replied: Thu, Oct 23 2014 1:20 AM

I would expect, that since I am offering to help HIM, he should offer me food...

Logos isn't doing the scholars any favours. They are the customers!

Posts 26104
Forum MVP
MJ. Smith | Forum Activity | Replied: Thu, Oct 23 2014 1:23 AM

Okay, I'll admit that the last time I visited Verbum they bought lunch. Wink

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Page 1 of 4 (75 items) 1 2 3 4 Next > | RSS