Problem searching for only 1 instances of a search term

I want to find verses where a particular word only occurs once. In this example, I am looking for a single occurrence of the Greek word δέ. I did a morphological search on the NA27 as follows:

lemma:δέ NOT BEFORE lemma:δέ

This produces erratic search results. In addition to legitimate examples of a single occurrence of δέ (e.g. Matt 3:11; 5:19, 33), I get the following types of erroneous matches:

Some verses where δέ occurs more than once, such as Matt 10:28, 13:11; 21:29, 32; 23:8. Not all verse with more than instance of δέ are returned (such as Matt 1:2, 3, 4, 5, 6, etc.). I cannot figure out why it returns the ones it does.
Some verses where δέ occurs with the contracted form δʼ. For example, in Matt 5:22 (δέ appears once, δʼ twice). These should be excluded since I am searching on a lemma and the lemma of δʼ is correctly tagged as δέ.
If δέ occurs in brackets the verse appears, but [δέ] is highlighted (e.g. Matt 6:1).

The search does not return all verses with δέ, so it is doing some type of exclusion. I cannot figure out what it is really doing.

The other quirk is that the search results are highlighting both δέ and another word. The word changes in each verse, but it tends to be a conjunction, participle or adverb such as oti, μὲν, μὴ, ἐὰν, or οὐ).

Here is a screen shot showing some of the search results.

Any wisdom about why this search is failing? How can I search for verses with only 1 instance of a particular word?

Find more posts tagged with

Comments

Rosie Perera

You could search for all verses where it occurs at all (δέ) and save the results to a Passage List. Then do another search for where it occurs at least 2x (δέ BEFORE δέ) -- I don't know Greek or the Morphology search feature very well, but I don't think you need to use lemma: for this search. Save those results to another Passage List. Now combine the two Passage Lists using Merge - Difference:

That will exclude all the verses where it occurs 2 or more times from the initial list, and you'll be left with just the verses where it occurs once and only once.

Harry Hahne

I think I understand why this did not work. If I search for lemma:δέ NOT BEFORE lemma:δέ, it treats the NOT as a search term, rather than a negation of the BEFORE operator. That is why it was highlighting certain words, including δέ when it means "but" or "not this . . . but that".

So is there a way to say one word must not occur before another?

The reason I have to do this work around is that you cannot say lemma:δέ NOT lemma:δέ. With some Bible programs I can do this and I will find verses where δέ only occurs once.

So I am still trying to figure out how to find verses with only one occurrence of a certain word.

Graham Criddle

lemma:δέ NOT BEFORE lemma:δέ

"NOT" isn't a valid search operator ("ANDNOT" is) so it is being treated as a search term

So your search is being treated as:

"lemma:δέ" AND "not BEFORE lemma:δέ" (and the "NOT" is being matched against the English gloss associated with the text)

I may not have explained that very well but hopefully you get the sense[:)]

I don't believe you can construct the search you are looking for so Rosie's suggestion about Passage Lists is a good one to go for

EDIT: Sorry - I see we were typing at the same time.

Harry Hahne

https://community.logos.com/discussion/comment/469960#Comment_469960

Thanks, Graham. After my original post, I figured this out.

So here is my current work around for the lack of a NOTBEFORE operator in Logos:

lemma:δέ ANDNOT (lemma:δέ BEFORE lemma:δέ)

This appears to find verses with only one occurrence of the word. I don't know if there are any unforeseen side effects of this.

The reason I am nervous about it, is that I know that this syntax is not completely reliable. I have tried previously to find anarthrous nouns and this search fails (see the discussion at http://community.logos.com/forums/t/49655.aspx):

@N ANDNOT (@D BEFORE 1-2 WORDS @N)

This will find some anarthrous nouns, but not all of them. It excludes verses where there are any articles (e.g. John 1:1). I have never figured out a way to reliably say something must not occur before a search term.

Graham Criddle

https://community.logos.com/discussion/comment/469961#Comment_469961

Hi Harry

So here is my current work around for the lack of a NOTBEFORE operator in Logos:

lemma:δέ ANDNOT (lemma:δέ BEFORE lemma:δέ)

That's nice[Y]

I don't know if there are any unforeseen side effects of this.

I can't see why there would be but I can see why you would be concerned.

Graham

Harry Hahne

https://community.logos.com/discussion/comment/469967#Comment_469967

I don't know if there are any unforeseen side effects of this.

I can't see why there would be but I can see why you would be concerned.

The reason for my concern is that I can verify that this syntax does not always work. Here is the same syntax that does not work as expected:

@N ANDNOT (@D BEFORE 1-2 WORDS @N)

In this case, Logos seems to only find verses that have no article (@D) at all. So it fails to find verses in which there are any articlesat all.

Dave Hooton

https://community.logos.com/discussion/comment/470066#Comment_470066

Here is the same syntax that does not work as expected:

@N ANDNOT (@D BEFORE 1-2 WORDS @N)

Once the condition (@D BEFORE 1-2 WORDS @N) is satisfied the verse is thrown out. In the case of nouns this is a notoriously unreliable query. It gives roughly 3260 results but a (more accurate) Syntax Search will give over 10100 results (it picks up results like "the Isaac Isaac", but the second Isaac is in a different Clause).

We need the Libronix Graphical Query!

Harry Hahne

https://community.logos.com/discussion/comment/470131#Comment_470131

We need the Libronix Graphical Query!

Amen! [Y]

This could also be solved with exclusion operators, like NOTBEFORE. But the best way is the Libronix Graphical Query, which allows proper exclusions as well as the morphological agreement which is required for this search.