Syntax Search for Anarthrous Noun

Page 1 of 2 (26 items) 1 2 Next >
This post has 25 Replies | 2 Followers

Posts 6
Ben Martin | Forum Activity | Posted: Wed, Sep 8 2010 3:50 AM

Hi,

I am trying to find all instances of θεος in the NT that occur anarthrously. I have looked around the wiki and found nothing specifically, and have searched the forums to no avail. I don't really understand how to construct a syntax search using the cascadia graphs of the NT but I'd like to learn. So my question is two fold:

1. Is there a resource that goes through how to construct a syntax query for Cascadia?

2. How do I construct a specific search to find θεος every time it occurs anarthrously?

 

Kind regards,

Ben.

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 5:01 AM

Ben,

Welcome!

I know that more knowledgeable users will chime in but here is my answer:

1.) There are a few places that explain syntax searching....this forum is one of them...just do a search to find the example searches...

For myself, it's just a matter of starting out by replicating what's already found in the resources...i.e. go to an anarthrous Theos reference and duplicate it and use that as a starting point...

PS: this also works in a bible search..."θεὸς"  ANDNOT "ὁ θεὸς"

 

If I can...I'll see if I can get something together with the cascadia syntax graphs....

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 5:13 AM

Ben,

I can add more details later but this is doable...

I did this:

I opened a new syntax search...and I chose the "predefined subject" as shown:

 

 

Then i customized it with the article...and set it to be "not present" using the options on the right side of the screen in the search box...

Both elements were set to "matching skips levels" and viola!

 

Just replicate this search to get a feel for it.....

 

 

 

As I mentioned...you're going to get more answers that are better than mine....but this is a start....

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 5337
Kevin Becker | Forum Activity | Replied: Wed, Sep 8 2010 5:17 AM

Ben,

Welcome to the forums!

The first page to visit is http://wiki.logos.com/A_Strategy_for_Syntax_Search and http://wiki.logos.com/Setting_up_a_Syntax_Search

I've spent a while trying to get this syntax search to work. However, it is difficult to search for something that isn't there because, in this case the article can come in differing parts of  the syntax graphs.

This particular search is probably best attempted with a morph search because the article comes at a predictable interval in front of the noun.

  1. Find all instances of θεος: Search lemma:θεός
  2. Export Search results to a Passage List
  3. Find all instances of an θεος that come with the article lemma:θεός AFTER 2 WORD @D (The space of 2 allows for a postpositive to intervene)
  4. Export Search results to a Passage List
  5. Merge the first Passage List with the second using Difference. This will subtract all of the articular occurrences from the total number.

Some concluding thoughts

  • There may be a way to do this in a syntax search but I was unsuccessful in my limited attempts (If you like I can post an explanation on why I was unsuccessful)
  • This is the type of search that Libronix's graphical search excelled at. I hope something like it will eventually make its way into L4
  • Syntax searching can be difficult to learn. These forums are the only reason I figured it out... keep posing questions and we'll do our best to help.
Posts 5337
Kevin Becker | Forum Activity | Replied: Wed, Sep 8 2010 5:22 AM

Robert,

Your examples assume that Theos is in the Nominative (for both your morph and syntax search). You will miss a lot of examples, an anarthrous θεος need not function as a subject only.

However, I am glad you posted what you did because it reminded me of parentheses (duh!) The search I suggested above with passage lists can be done with this

lemma:θεός ANDNOT (lemma:θεός AFTER 2 WORD @D)

 

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 5:31 AM

Well,

In my defense I DID SAY that smarter forumites would answer....

 

And  they have... :)

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 5:35 AM

kevin...your search didn't work for me....

 

did you test it or am I doing something wrong....?

 

Edited to add: this worked:

<Lemma = lbs/el/θεός> ANDNOT (((@D BEFORE 0 WORD <Lemma = lbs/el/θεός>) OR (@D BEFORE 1 WORD <Lemma = lbs/el/θεός>) OR (@D BEFORE 2 WORD <Lemma = lbs/el/θεός>)))

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 5337
Kevin Becker | Forum Activity | Replied: Wed, Sep 8 2010 5:50 AM

Robert Pavich:

kevin...your search didn't work for me....

 

did you test it or am I doing something wrong....?

 

Edited to add: this worked:

<Lemma = lbs/el/θεός> ANDNOT (((@D BEFORE 0 WORD <Lemma = lbs/el/θεός>) OR (@D BEFORE 1 WORD <Lemma = lbs/el/θεός>) OR (@D BEFORE 2 WORD <Lemma = lbs/el/θεός>)))

Yes I tested it; it works in a Morph search (NA27 with Logos Morph). Based on your search terms above you must be performing a Bible search of some kind.

 

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 6:18 AM

Kevin Becker:
Based on your search terms above you must be performing a Bible search of some kind.

 

Yep....I'll quit helping now.... Embarrassed

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 25199
Forum MVP
Dave Hooton | Forum Activity | Replied: Wed, Sep 8 2010 8:32 AM

Kevin Becker:
I've spent a while trying to get this syntax search to work. However, it is difficult to search for something that isn't there because, in this case the article can come in differing parts of  the syntax graphs.

It is particularly hideous in Cascadia because of the many levels of nominal phrases. OpenText is more predictable but still involves at least four different constructs. Since we began to look at example constructs from a text query like lemma:θεός ANDNOT (@D BEFORE 2 words lemma:θεός) it is easy to conclude that the verification of a Syntax query is not worth the effort!

 

Dave
===

Windows 10 & Android 8

Posts 5337
Kevin Becker | Forum Activity | Replied: Wed, Sep 8 2010 8:49 AM

Dave Hooton:
It is particularly hideous in Cascadia because of the many levels of nominal phrases. OpenText is more predictable but still involves at least four different constructs. Since we began to look at example constructs from a text query like lemma:θεός ANDNOT (@D BEFORE 2 words lemma:θεός) it is easy to conclude that the verification of a Syntax query is not worth the effort!

Which just goes to show that it isn't enough to know how to construct a syntax search. One must understand the various strengths and weaknesses of the various morph and syntax databases in Logos so one can select the appropriate database(s) to answer a question.

Posts 4508
Robert Pavich | Forum Activity | Replied: Wed, Sep 8 2010 9:16 AM

Kevin Becker:
Which just goes to show that it isn't enough to know how to construct a syntax search. One must understand the various strengths and weaknesses of the various morph and syntax databases in Logos so one can select the appropriate database(s) to answer a question.

 

Which promps me to observe that I've had a love/hate relationship with the various Syntax searches in L3 and L4....on the one hand, they are unique in finding certain things....on the other hand....they are fraught with pitfalls if you're not up on them and careful about how you go about searches...and even then...sometimes a simple look through a BWS Grammatical Relationships section is more than adequate for what I need to know!

Robert Pavich

For help go to the Wiki: http://wiki.logos.com/Table_of_Contents__

Posts 1308
LogosEmployee
Rick Brannan (Faithlife) | Forum Activity | Replied: Wed, Sep 8 2010 9:38 AM

Dave Hooton:

Kevin Becker:
I've spent a while trying to get this syntax search to work. However, it is difficult to search for something that isn't there because, in this case the article can come in differing parts of  the syntax graphs.

It is particularly hideous in Cascadia because of the many levels of nominal phrases. OpenText is more predictable but still involves at least four different constructs. Since we began to look at example constructs from a text query like lemma:θεός ANDNOT (@D BEFORE 2 words lemma:θεός) it is easy to conclude that the verification of a Syntax query is not worth the effort!

At one point, I worked up the following for an anarthrous noun/adjective search using OpenText.org. Perhaps it helps, perhaps not. I don't recall if there were issues with false positives or false negatives.

Rick Brannan
Data Wrangler, Faithlife
My books in print

Posts 222
LogosEmployee
Mike Aubrey | Forum Activity | Replied: Wed, Sep 8 2010 9:47 AM

Dave Hooton:

It is particularly hideous in Cascadia because of the many levels of nominal phrases. OpenText is more predictable but still involves at least four different constructs. Since we began to look at example constructs from a text query like lemma:θεός ANDNOT (@D BEFORE 2 words lemma:θεός) it is easy to conclude that the verification of a Syntax query is not worth the effort!

Dave, are you sure?

UPDATE removed search image.

I'm going to try to come up with something that does it

Posts 222
LogosEmployee
Mike Aubrey | Forum Activity | Replied: Wed, Sep 8 2010 10:27 AM

Michael Aubrey:
I'm going to try to come up with something that does it

So, it seems that it is possible to create a search that locates places where θεος is the head of a Clause Function (Subject, Object, Indirect Object, Adverbial...) with the following search (which is actually incredibly simple):

(Terminal Node has "Matching Skips Levels" checked)

This returns 111 (accurate) hits.

What becomes problematic with Cascadia is to locate places where θεος is a modifier of some sort, particularly in the genitive. In such cases, the recursive phrases (i.e. the many nominal levels) becomes a challenge. I'm still looking for a solution. If I find something, I'll let you know.

Posts 222
LogosEmployee
Mike Aubrey | Forum Activity | Replied: Wed, Sep 8 2010 10:50 AM

Michael Aubrey:
What becomes problematic with Cascadia is to locate places where θεος is a modifier of some sort, particularly in the genitive. In such cases, the recursive phrases (i.e. the many nominal levels) becomes a challenge. I'm still looking for a solution. If I find something, I'll let you know.

The key to finding modifying anarthrous instances of θεος, it seems, is to constrain the search from higher up in the syntactic structure. The following search finds all the instances where anarthrous θεος functions as a modifier. This search finds 18 hits for a total of 129 when included with my previous search. That's not bad out of 1317 occurrences of θεος in the NT.

As in the previous search: Terminal Node has "Matching Skips Levels" checked.

Other contexts where θεος might be found would be coordinate constructions and appositional constructions. It would be possible to create similar searches constrained like this one, but I would need to find an example of how appositions and coordinate NPs are handled by Cascadia before I could create a search. Coordinate constructions would be a challenge though because of occurrences like τὸν θεὸν Ἀβραὰμ καὶ θεὸν Ἰσαὰκ καὶ θεὸν Ἰακώβ in Luke 20:37, where technically, the second two instances of θεος are not anarthrous because they are governed by the very first article.

Posts 5337
Kevin Becker | Forum Activity | Replied: Wed, Sep 8 2010 6:21 PM

Michael,

Thanks, you've given me some good stuff to chew on. However, lemma:θεός ANDNOT (lemma:θεός AFTER 2 WORD @D) returns 240 hits in 219 verses (which coincidentally does not return Luke 20:37... I would expect it too but it doesn't). So even with your good approaches I still think this type of query is best suited for a morph database and not a syntax one.

Posts 25199
Forum MVP
Dave Hooton | Forum Activity | Replied: Wed, Sep 8 2010 6:22 PM

Rick Brannan:
At one point, I worked up the following for an anarthrous noun/adjective search using OpenText.org. Perhaps it helps, perhaps not.

That was what I had thought as being necessary. It is the only Syntax query I would have attempted to verify!

Dave
===

Windows 10 & Android 8

Posts 25199
Forum MVP
Dave Hooton | Forum Activity | Replied: Wed, Sep 8 2010 7:49 PM

Michael Aubrey:
The key to finding modifying anarthrous instances of θεος, it seems, is to constrain the search from higher up in the syntactic structure. The following search finds all the instances where anarthrous θεος functions as a modifier. This search finds 18 hits for a total of 129 when included with my previous search. That's not bad out of 1317 occurrences of θεος in the NT.

Sorry to be alarmist in using "hideous" but given the time I spent on a Cascadia solution vs OpenText solution I was convinced that the return would not be worth any further investment! Your solution confirms my assessment but provides valuable insight to the construction of Cascadia queries!

The text query omits Lk 20:37 which proves both its fallibility and unintended utility given your comment ie. it omits any verse in which theos has the definite article even if there is an apparently anarthrous instance. However, I believe the text query provides the better, more complete, solution.

 

Dave
===

Windows 10 & Android 8

Posts 222
LogosEmployee
Mike Aubrey | Forum Activity | Replied: Thu, Sep 9 2010 8:43 AM

 

Kevin Becker:
Thanks, you've given me some good stuff to chew on. However, lemma:θεός ANDNOT (lemma:θεός AFTER 2 WORD @D) returns 240 hits in 219 verses (which coincidentally does not return Luke 20:37... I would expect it too but it doesn't). So even with your good approaches I still think this type of query is best suited for a morph database and not a syntax one.

Kevin, that statement was never intended to make an observation or comparison between text queries and syntax queries. I use both for different purposes. Linear strings are generally more convenient with text queries -- especially with the analysis view!

With that said, syntactically speaking, if this particular search wasn't looking for θεος, your string would be inadequate for how Greek can potentially structure noun phrases:

  • Its possible to have multiple modifiers between an article and its noun such as in the construction: Article, Adjective, Possessive Pronoun, Noun
  • Its possible to have a multiple word modifying phrase between an article and its noun such as: Article, Prepositional Phrase, Noun OR Article, [Genitive Article, Genitive Noun] Noun.

None of these constructions appear with θεος in the NT that I know of, but they are possible and grammatical, which would then require you to use: AFTER 3 WORD @D rather than 2 in order to allow for post positives, which would then also allow for many, many more false hits since there's more potential for another noun to sneak in there between them. Lest you despair...these structures are in the clear minority and are rare both in the NT and even the non-biblical texts.

As for Luke 20:37, the reason that it doesn't is that the text query is motivated by verse boundaries not clause types. Thus because there is a single instance of θεος with the article, it rejects the entire verse -- as Dave has stated, that is its fallibility.

Dave Hooton:
Sorry to be alarmist in using "hideous" but given the time I spent on a Cascadia solution vs OpenText solution I was convinced that the return would not be worth any further investment! Your solution confirms my assessment but provides valuable insight to the construction of Cascadia queries!

As I said, I never actually checked the text query, so I didn't have any basis line for comparison. I'm looking at it now. But ... if one is searching for examples of a grammatical construction in order to generalize about grammar, there is little need to have every example, as long as you have a sufficient number of hits to compare--though I'd assume for the initial question about the interest in θεος in particular that the original searcher wasn't interested in only the question of definiteness. If that were the case, one would be better off:

  • Using multiple nouns
  • Working through a text to see how participants, objects, and events are introduced and how the use of the article functions as a result of that.

In any case, I'm glad that I was able to help provide some insight into constructing queries with Cascadia. Minimally, that is a good thing.

Dave Hooton:
The text query omits Lk 20:37 which proves both its fallibility and unintended utility given your comment ie. it omits any verse in which theos has the definite article even if there is an apparently anarthrous instance. However, I believe the text query provides the better, more complete, solution.

I did a passage list merge, which resulted in 24 hits missing from the text search.

Looking at the search differences between them, the reason why my search didn't find all of them was a very simple and easily fixable error that I should have caught from the beginning. It only took deleting one small element, as seen below:

The new result was 258 hits in 235 verses. The text query, as I expected, cut out 42 verses that had both an articular and anarthrous θεος.

There were 30 passages where the text query gets it right and syntax misses. Some of these are easily fixable (e.g. Matt 27:46; Luke 2:52; Acts 14:15). The main problem is that merely blocking the determiner is too constrained of an approach and block all instances where there is an article blocking something else...but its less constrained than the previous iteration where the determiner's sister phrase headed by θεος blocked the majority of instances where there was no second phrase level below the initial θεος phrase (i.e. some of the misses in this search were caught by the previous one). Adding an OR and the previous search with the deleted  phrase brings the total up to 276 hits. 

And then there are quite a few where and then even a few where I don't understand why the text query picked them up when it shouldn't have (Luke 18:29; John 17:3)...or perhaps I missed something in performing the search that caused those...

In any case, this little change, pushed the Cascadia syntax search over the top and made it significantly simpler than Opentext and significantly more accurate than the text query. With that said, it still seems that if one is interested in being comprehensive, both search types are required--which is something I would have said all along any way.

All of this also shows that Cascadia requires significant practice in getting the hold of and perhaps also plenty of time familiarizing oneself with it in order to anticipate constructions that might otherwise be missed in creating a search.

Page 1 of 2 (26 items) 1 2 Next > | RSS