Inconsistencies in Andersen-Forbes

I am exploring Andersen-Forbes text type tagging.
I found it very interesting the tips on the wiki page: Extended Tips for Highlighting and Visual Filters (http://wiki.logos.com/Extended_Tips_for_Highlighting_and_Visual_Filters). So, after exploring a little bit A-F I found some tagging not very much appropriated.
Well, let's go to an example:
In the Hebrew Bible: Andersen-Forbes Analyzed Text, in Gen 3:14-19, it is God speaking to the serpent, Eve and Adam. See that God's words to the serpent and to Adam are tagged as "Divine Blessing," while for Eve the words are tagged as "Divine Judgment." Why is that?
Someone could say that there is a difference between the words to the serpent and Adam and the words to Eve. For the serpent and Adam God says "Because you...cursed is (are)..." as for Eve, this sequence does no exists. However, there is no doubt that the words to Eve is a curse as much as it is for the serpent and Adam.

Let's assume that their criteria for tagging the text is based mostly on its syntactical structure, and because of that the words to Eve is different because it does not have the "construction of a Blessings/curse formula." Well, still I see inconsistencies by this criteria because they tag differently the words to the serpent and to Adam as well.

See that for the serpent, they interpret that the "Divine Blessing" (as they tag), begins with:

“Because you have done this,

cursed are you above all livestock..." (The Holy Bible : English Standard Version. (Wheaton: Standard Bible Society, 2001), Ge 3:14.)

As for Adam, the "Because you..." is tagged as "Divine Judgment," and only when God says "course is the ground because of your..." they tag it as "Divine Blessing."

So, do you guys agree with my analysis. Does anyone see they rational on a way that it all could make sense?

Find more posts tagged with

Comments

MJ. Smith

Forms of Old Testament Literature, my "go to" source for literary forms shows:

I am surprised that the AF Glossary doesn't seem to include the form tags - either that or I'm not seeing it because my first cup of coffee is unfinished. But my AF seems to label 3:16-17 as "Divine instruction" - neither blessing nor curse. The serpent verse is simply tagged "narrative"

As for Tooman which I know we have no definitions for:

Dave Hooton

https://community.logos.com/discussion/comment/245887#Comment_245887

But my AF seems to label 3:16-17 as "Divine instruction" - neither blessing nor curse. The serpent verse is simply tagged "narrative"

Without my first cup, I see different words appropriately labelled "narrative" "instruction" then the rest is "blessing" or "judgement"; but the "blessing" is not appropriate (similarly for Gen 4:11-12).

Dave Hooton

Does anyone see they rational on a way that it all could make sense?

It's not comforting to think that it was mechanically tagged and it is positively disturbing to think that it was not checked.

MJ. Smith

https://community.logos.com/discussion/comment/245920#Comment_245920

Without my first cup, I see different words appropriately labelled "narrative" "instruction" then the rest is "blessing" or "judgement"; but the "blessing" is not appropriate (similarly for Gen 4:11-12).

Don't do that to me - it means my filter isn't working correctly. [:'(]

Edit: It's even worse - I had typed in the wrong reference. As my Mother says "Once I thought I had made a mistake but found I was mistaken". The coding should be reported as a typo.

Dave Hooton

https://community.logos.com/discussion/comment/245937#Comment_245937

The coding should be reported as a typo.

Now i've had my coffee, I found some energy to send an email to syntax@logos.com!

Rick Brannan

https://community.logos.com/discussion/comment/245933#Comment_245933

Hi Dave

It's not comforting to think that it was mechanically tagged and it is positively disturbing to think that it was not checked.

I'm not nearly as familiar with the Hebrew side of syntax resources, but my understanding is that Andersen-Forbes was iteratively produced. That is, there wasn't/isn't a parser per se that produces it, but it was built up iteratively over time, a product of both automated processes and personal review cycles. It took 30 years, and they're still working on it.

Also, are you aware of http://andersen-forbes.org/ ? You might be able to offer feedback through their web site.

[Note: This really is Rick; I've just recently gone through the pain of separating my personal and employee Logos accounts (I think this is my personal account), so my post count reverts back to 0 (now 1, I think)]

Dave Hooton

https://community.logos.com/discussion/comment/245965#Comment_245965

Also, are you aware of http://andersen-forbes.org/ ? You might be able to offer feedback through their web site.

I didn't observe any way to offer feedback, but we've been down this track before with OpenText ...at least Andersen-Forbes' offerings are more current, so I have confidence that Logos will be able to offer feedback!

Philip E. Sumpter

https://community.logos.com/discussion/comment/245971#Comment_245971

A more grave example of inconsistencies in my book is their tagging of the superscription in the psalms. Note the different ways they tag the same phrase מזמר לדוד.

Nielsen Tomazini

https://community.logos.com/discussion/comment/245976#Comment_245976

A more grave example of inconsistencies in my book is their tagging of the superscription in the psalms. Note the different ways they tag the same phrase מזמר לדוד.

Philip, could you give me a couple of Psalms where you found inconsistencies in superscriptions in the psalms? I look a series of Psalms of David but I did not find differences. However, I did not look at them all.

It is really unfortunate that A-F has got these inconsistencies. I wish I could use this syntax tagged resource as a reference to some Hebrew studies, but as for now it is impossible. A-F, the way it is now is not reliable.

This is quite a critical problem. I have participated of consultations with scholars, debating the use of Bible Softwares for academic research. When we see this kind of inconsistency it pushes back some scholars who were willing to try to use a Bible Software. In the seminary that I work and study, one faculty, who is not in favor of Bible Software for teaching and researching, asked one guy to perform a search for him in a Bible Software. I did the search as well, and we got inconsistencies, we could not find what that faculty member wanted. Then, he said: "I told you!, These tools are not reliable..." Sometimes it happens because the Bible Software is not capable of performing the search, other times it is due to inconsistencies.

George Somsel

https://community.logos.com/discussion/comment/245993#Comment_245993

A more grave example of inconsistencies in my book is their tagging of the superscription in the psalms. Note the different ways they tag the same phrase מזמר לדוד.

Philip, could you give me a couple of Psalms where you found inconsistencies in superscriptions in the psalms? I look a series of Psalms of David but I did not find differences. However, I did not look at them all.

It is really unfortunate that A-F has got these inconsistencies. I wish I could use this syntax tagged resource as a reference to some Hebrew studies, but as for now it is impossible. A-F, the way it is now is not reliable.

This is quite a critical problem. I have participated of consultations with scholars, debating the use of Bible Softwares for academic research. When we see this kind of inconsistency it pushes back some scholars who were willing to try to use a Bible Software. In the seminary that I work and study, one faculty, who is not in favor of Bible Software for teaching and researching, asked one guy to perform a search for him in a Bible Software. I did the search as well, and we got inconsistencies, we could not find what that faculty member wanted. Then, he said: "I told you!, These tools are not reliable..." Sometimes it happens because the Bible Software is not capable of performing the search, other times it is due to inconsistencies.

I'm with MJ. I generally use FOTL if I wish to check my analysis of a passage, but then the editor (and author of the Numbers volume) was my mentor in grad school so I'm a bit prejudiced. The main uses I make of A-F is (1) posting Hebrew text where I don't want to include cantillation (and don't wish to go through the work of removing it) and (2) the Eissfeldt analysis of the threads in the Pentateuch (though this is better in L3 since the colors are more discernable).

David Knoll

https://community.logos.com/discussion/comment/245993#Comment_245993

This is quite a critical problem. I have participated of consultations with scholars, debating the use of Bible Softwares for academic research. When we see this kind of inconsistency it pushes back some scholars who were willing to try to use a Bible Software. In the seminary that I work and study, one faculty, who is not in favor of Bible Software for teaching and researching, asked one guy to perform a search for him in a Bible Software. I did the search as well, and we got inconsistencies, we could not find what that faculty member wanted. Then, he said: "I told you!, These tools are not reliable..." Sometimes it happens because the Bible Software is not capable of performing the search, other times it is due to inconsistencies.

Well he was right. But he should try the new syntax database in Accordance. Searches are simple and the results are accurate (But the linguistic theory is bizarre).

Don't rely on AFAT. First of all nobody knows how to search it effectively. It is too complicated. It is also as you quite rightly put it, inconsistent.

Logos is not good at Hebrew searches. They don't plan or think what the user wants to achieve by using a Hebrew resource before they publish it. The GUI is not user friendly. Searches are too complicated to construct and too often they are not even possible.

The ideal syntax database would begin by having a look at a syntax text book such as Davidson. The phenomena described in the textbook should be categorised and analysed. Then the GUI queries should be developed and only then the scholar should start developing the tagging scheme. The database should be user-centric and not scholar/tagger-centric.

In the realm of Hebrew Syntax databases the scholar would have to think how to present parallel pairs because many people are interested in searching for word pairs and hemistich structures.

I tried using AFAT. I looked for documentation (none can be found) played with it for sometime but I was never able to get the right results. For a month I have been using the database prepared by Accordance. It is a great leap forward (It only covers Genesis and the Psalms at the moment) but there is still room for improvement. Logos can do better. For instance in how they deal with the verbless clause and in their treatment (or rather lack of it) of the relationship between parallel hemistichs.

At the moment, once again, for those who want a professional tool for Hebrew searches, Logos is not the answer. It is a nice tool for those who don't know Hebrew and want to show off in a cocktail party some niceties that they have figured out using a reverse interlinear or the Bible Word Study Tool. On the other hand, for those who are well versed in the Hebrew Language and seek to ask questions about the text and get accurate results, Logos is the worst option.

I am sorry to say this. I am heavily invested in Logos, and I was rather hoping to see improvement over the last year but it seems the scholarly community isn't listed high in Logos' list of priorities. Logos is still a very elegant reading platform and they sell wonderful content. I use it a lot for commentaries, lexicons, and grammars.

MJ. Smith

https://community.logos.com/discussion/comment/246101#Comment_246101

The ideal syntax database would begin by having a look at a syntax text book such as Davidson. The phenomena described in the textbook should be categorised and analysed. Then the GUI queries should be developed and only then the scholar should start developing the tagging scheme. The database should be user-centric and not scholar/tagger-centric.

Sounds suspiciously like some of my complaints on other topics. [:)]

David Knoll

https://community.logos.com/discussion/comment/246131#Comment_246131

Sounds suspiciously like some of my complaints on other topics.

Great minds think alike. Especially the modest ones. [:)]

Nielsen Tomazini

https://community.logos.com/discussion/comment/246101#Comment_246101

Thank you David for such good insights about Hebrew in Logos.

Thank you for your good sense of humor despite our lack of a good Hebrew syntax tool in Logos 4.

The more I search in Logos 4 (for Hebrew) the more I see that I need another tool. In this case, I need Accordance. Maybe in a couple of month I will be able to buy it.

In Logos, I believe, the best resource for syntax search is the SESB, it is a serious work which is not difficult to work with. The BHS/WIVU Constituency Tree is not difficult to perform syntax searches. However, there are two problems with it:

1 - The SESB syntax search is available only for Libronix 3 . It is up to the German Bible Society to release a Logos 4 compatible version;

2 - It still lacks the 'root' search.

So, at this moment, the solution to my Hebrew syntax queries is about $300.00 away from me.

David Knoll

https://community.logos.com/discussion/comment/246155#Comment_246155

In this case, I need Accordance.

Nielsen pay attention to the fact that while Accordance runs on a PC through an emulator you are unable to print or export your results to the PC. So if you don't have a mac Accordance is of limited value for you. Send me an e-mail to my firstname.my lastname at huji.co.il

Dean Forbes

Nielsen--

I am sorry to see that you have been frustrated as you tried to use the AF text-type assignments. As will become clear once our book on grammar is published ("Biblical Hebrew Grammar Visualized"), our primary focus is on syntax. (To date, Eisenbrauns has delayed publication by over a year leading to enormous frustration for us and for (would-be) users of our data.)

The text type assignments were made about 25 years ago, were rarely used, and are quite unsatisfactory both in their underlying theory and implimentation. We should have withheld them, but we naively did what we did.

Combining blessings and curses under one code has proven to be less self-explanatory than we expected. We thought it would be clear to users which was which...Big mistake!

Just last month I wrote a white paper proposing an approach to making the assignments both taxonomically solid and self consistent. But, improvements in the phrase marker representation take precedence over "enhancements." So carrying out the text type improvements will have to wait.

I hope that you will find the syntactic database that is perfect for your needs.

Sincerely,

Dean Forbes

Dave Hooton

https://community.logos.com/discussion/comment/246292#Comment_246292

Dean,

Thank you for responding on behalf of yourself and Frank Andersen.

The text type assignments were made about 25 years ago, were rarely used, and are quite unsatisfactory both in their underlying theory and implimentation. We should have withheld them, but we naively did what we did.

Combining blessings and curses under one code has proven to be less self-explanatory than we expected. We thought it would be clear to users which was which...Big mistake!

When the (Logos) Text Type is labelled as "Blessing" we don't expect "Cursing" and would certainly prefer to search for these separately, but that could easily be clarified in the current (Logos) implementation.

Nielsen Tomazini

https://community.logos.com/discussion/comment/246160#Comment_246160

Nielsen pay attention to the fact that while Accordance runs on a PC through an emulator you are unable to print or export your results to the PC. So if you don't have a mac Accordance is of limited value for you. Send me an e-mail to my firstname.my lastname at huji.co.il

Dear David,

Thank you for allowing me contact you via e-mail. Unfortunately I could not send you a message. I have tried the address you gave me, I even tried different forms of your e-mail but all my attempts apparently were returned. Could you confirm your e-mail? I understood that it is: yourfirstname dot yourlastname at huji.co.il

Shalom

David Knoll

https://community.logos.com/discussion/comment/246326#Comment_246326

Oh that should read huji.ac.il

Pardon

Dean Forbes

https://community.logos.com/discussion/comment/246303#Comment_246303

Dean,

When the (Logos) Text Type is labelled as "Blessing" we don't expect "Cursing" and would certainly prefer to search for these separately, but that could easily be clarified in the current (Logos) implementation.

You are, of course, quite correct. We have tagged 206 blessings/curses. I'll look into dividing them up for our next release, not too big a job.

Dean

David Knoll

https://community.logos.com/discussion/comment/246335#Comment_246335

Dear Mr. Forbes,

I wonder if you could explain why several predicates are marked as "joined perfect" (Deut 1:5, Josh 3:16 Job 20:19) while many other clauses that contain two verbs are split into two predicates (Daniel 9:15, Hos 5:11 and many many more). What is the formal difference between these cases?

many thanks,

David Knoll

Nielsen Tomazini

https://community.logos.com/discussion/comment/246292#Comment_246292

Dear Mr. Forbes,

First of all, thank you very much for your explanations and for taking your time to interact with us here.

As I wrote in the original post, I was really enjoying using A-F and exploring the text-type tags. Actually, the problem is not to see a text about blessing or curse under the same category. Certainly, by reading the text we can figure out if it is a blessing or a curse. But this kind of resource (with tagging) in a Bible Software is really valuable when we are able to search through its data. And that is where the inconsistencies do not allow me to have confidence in the results. And as such, I can only rely on the searches if I read every single result. Even doing that I would be suspicious if any occurrence was not omitted.

For instance, if I want to find all the occurrences of Divine Blessings (does not matter if blessing or curse), which begins with KY ("because, for, that"), I would get Gen 3:14 but not Gen 3:17.

The idea and work of tagging the Hebrew text the way you and Mr. Andersen did is awesome. When I saw the tagging I saw the potential for researching a data that we did not have before. Your explanation, that the primary focus is on syntax, clarifies to me why the divine words towards Eve was not tagged as Divine Blessing as well as the other words (to the serpent and Adam). So, this case is not a matter of inconsistency but of methodology, I understand that, and maybe with the publication of your work, many features in A-F will be clarified.

It is sad that the improvements in the text-type database will take time to be done. But I don't thing there is any resource with text type tagging. So, even if it takes time, please do it!! The Old Testament studies can only gain with more data to be researched.

I need to say that I am quite impressed with your curriculum.

Physics in Harvard;
Working with Hewlett-Packard in 1971;
Eight patents in cardiac electrophysiology;
Visiting scholar at Stanford;

Wow, really impressive!

Blessings,

Nielsen Tomazini

https://community.logos.com/discussion/comment/246327#Comment_246327

David,

I am still not getting it. Could you send me an e-mail? Mine is nielsen tomazini at gmail dot com. No space or dot between my names.
Thank you,

Philip E. Sumpter

https://community.logos.com/discussion/comment/245993#Comment_245993

Nielsen,

I've just looked at the ones from Ps 3-26 and couldn't find any. I was writing from memory of looking at the superscriptions in general. I'm guessing I made a mistake or I've confused the problem with something else ... sorry for writing without double checking!

Dean Forbes,

I'm glad to hear to book is still scheduled to come out, I've been waiting for it for quite some time now! This database is the reason I bought logos.

Nielsen Tomazini

https://community.logos.com/discussion/comment/246383#Comment_246383

Nielsen,

I've just looked at the ones from Ps 3-26 and couldn't find any. I was writing from memory of looking at the superscriptions in general. I'm guessing I made a mistake or I've confused the problem with something else ... sorry for writing without double checking!

That's fine Philip, thank for checking it out.

Blessings,

Dean Forbes

https://community.logos.com/discussion/comment/246337#Comment_246337

The sort of formation of a compound predicator that you refer to is an instance of what I call "Free Convention" inconsistency in my paper on consistency issues.The human over-reader (Professor Andersen) resorted to it 168 times (119x as a continuous join, 49x as a discontinuous join). It will certainly be cleaned up as we move into discourse-level analysis. I have put it on my list of interim irritants that we might set right sooner.

Five kinds of inconsistency are discussed, with examples, in:

Forbes, A. D., "The Challenge of Consistency." In Luis Vegas Montaner et al. (eds.), Computer Assisted Research in the 21st Century, (Piscataway, NJ: Gorgias, 2010), 111-126.

This paper also provides statistical estimates of the incidence of inconsistency in AF v0.9 (overall about 0.1%) based on my concept of surrogate texts.

Dean

P.S.: We read this forum when it touches on our work. We may not respond to comments very often, but they do affect our scheduling of tasks. We would ask that contributors make sure that their comments are based on our latest release and that they do their best to insure that their assertions are valid. We want to be responsive to user needs, but can ill afford to waste our time dealing with false alarms.

Dave Hooton

In the Hebrew Bible: Andersen-Forbes Analyzed Text, in Gen 3:14-19, it is God speaking to the serpent, Eve and Adam. See that God's words to the serpent and to Adam are tagged as "Divine Blessing," while for Eve the words are tagged as "Divine Judgment." Why is that?

These anomalies appear to be resolved in the recent update of the Andersen-Forbes resources.