Help me figure out why this search isn't working

[mgvh]

I'm still comparing what I do in BibleWorks with what Accordance can do. I am only using Accordance Lite, but I do have the tagged GNT text. (I.e., there is probably a more elegant way to do the search using the visual search, but I don't have that.)

 

What I'm trying to find are all the instances in the GNT of the grammatical construction εν τῳ plus an infinitive. Mindful that a δε might sometimes sneak in after the preposition, I'm running this search from top line

WORDS: εν <FOLLOWED BY> <WITHIN 2 Words> [ARTICLE neuter singular dative] <FOLLOWED BY> <WITHIN 1 Words> [VERB infinitive]

 

I get hits for results in 45 verses.

 

When I run the same search in BibleWorks, I get hits in 48 verses. What's going on? It appears that Accordance is missing:

• Luke 8.5
• Luke 9.34
• Gal 4.18

There is nothing tricky about those verses, and there isn't a text variant issue.

Should I be using a different search? Or is there some other explanation?

 

(BTW, this is a search I have my students run across both the LXX and the GNT to show how common it is in the LXX as a rendering of the Hebrew, and that it's overwhelming occurence in Luke is his way of making his Greek more biblisch, i.e., sounding like good old LXX Scripture!)

 

Thanks for your help!

[mgvh]

Hmmm.... I wondered if the "FOLLOWED BY" was throwing things off, so I just tried:
WORDS: εν <WITHIN 2 Words> [ARTICLE neuter singular dative] <WITHIN 1 Words> [VERB infinitive]

When I do that, it picks up the 3 texts I missed as noted above, but it returns 2 extra ones:

• Phil 1.13 though it doesn't have τῳ in it
• 1Thess 4.6 which is a correct result except that it doesn't have the order I want and why I used the FOLLOWED BY

[Ken Simpson]

Hi mgvh,

I can’t explain why that search doesn’t work. It seems properly formed syntactically, but I do remember some time ago Helen saying that the search bar doesn’t cope with complex Followed By and Within strings terribly well. Perhaps Joel will have a better answer.

 

However, when you use a construct search, it performs perfectly finding all the examples you cite that the other programme finds.

 

Sorry I can’t be more helpful than that.

 

[mbcvida]

"εν τῷ" [VERB infinitive] <OR> "εν δε τῷ" [verb infinitive] gave me 48 hits - is this what you were looking at?

[Luke Bert]

Take out the Neut for the article. Accordance tags the article as Masc at least in Luke 8:2 most likely agreeing with αυτος. Also in general I suggest the construct search for things like this it is usually easier and more effective. 

[Joel Brown]

Generally when you are combining multiple terms in a row it is simpler to use the construct.  "x followed by y followed by z" can get complex and unexpected logically, so the construct is great for full clarity.

[mgvh]

Thank you, all!

@Dick Roberts: Your search returns the results that I anticipated. I had tried avoiding the OR by allowing for any word to show up in there. (E.g., sometimes a μεν might also be inserted.)

@Ken Simpson; @Joel Brown: It does appear to be the case that the search line is not reliable for searches with FOLLOWED or WITHIN

"Construct Search": I'm assuming by this that you are referring to the visually oriented search tool in Accorcdance (which is not available in Accordance Lite).

 

Thanks again!

[Joel Brown]

To add a bit more clarity.  With "x followed by within 1 y followed by within 1 z", this is ambiguous.  Do you mean the group "x and y" are followed by within 1 of z?  The problem is, at this point, y is within 1 of z, but not x. This is why the luke 8:5 hit is missing.  If you modify your search as so, it helps clarify how you want the WITHIN x connected:

 

εν ([ARTICLE neuter singular dative] [VERB infinitive])

 

This seems to give the results you want.

 

But, this may not be intuitive to everyone, which is why in such ambiguous cases we encourage use of the construct, as shown by Ken.

[mgvh]

@Joel Brown: Perfect! That explains the logic part I was overlooking. The search string you provide works exactly as it should. Thanks!

[Λύχνις Δαν]

This is the best explanation I have seen to this issue yet. Thank you Joel. I have long wondered why Accordance did not support a parenthesis syntax to specify precedence but apparently it does. I confess to not having tested this and simply resorted to constructs but honestly a proper bracketing syntax that respects user desired precedence is easier to work with sometimes. Presumably (εν <FOLLOWED BY> <WITHIN 2 Words> [ARTICLE neuter singular dative]) <FOLLOWED BY> <WITHIN 1 Words> [VERB infinitive] would also work just as well in this case.

 

Previously I've heard comments made that such searches as this are unreliable and that construct is the way to go. But is it really just a case that the default precedence rules sometimes result in unexpected results rather than there actually being a limitation in the command line search support ? This is a pretty important point for anyone wanting to write scripted queries in Acc because you cannot use constructs.

 

Additionally, can one carry this sort of parenthetical nesting to arbitrary depth ?

 

thx

D

[Joel Brown]

Your string:

 

(εν [ARTICLE neuter singular dative]) [VERB infinitive]

 

 

Is not the same, because here it specifies that terms x and y must both be within 1 word of the verb, which is obviously impossible.  We describe the searches as unreliable because both A: They often give different results from what the user expects, and B: they have not been as exhaustively tested when getting to the nth degree of complexity.

 

Edit: You'll see the results of that search string are actually the same as without any parentheses at all, and thus is the default interpretation.  I honestly don't know why there are any results in this case, so again, please use the construct

[Λύχνις Δαν]

So FOLLOWED BY ... WITHIN is not behaving as a binary operator ?

I realise this is getting a little arcane and curiosity will be my death but ...

 

Thx

D

[mgvh]

I don't mean to overwork this, since it's clear that using the Construct Search is the preferred method... but I tried:

(εν <FOLLOWED BY> <WITHIN 2 Words> [ARTICLE neuter singular dative]) <FOLLOWED BY> <WITHIN 2 Words> [VERB infinitive]

This gives me 52 hits.

 

This returned some results that aren't the prepositional infinitive clauses that I want but do adhere to the logic of the string. (E.g., 2 Pet 1.13: ἐν τούτῳ τῷ σκηνώματι, διεγείρειν)

It did turn up one other hit that I had not previously found anywhere in Heb 2:8: ἐν τῷ γὰρ ὑποτάξαι - I forgot that the γαρ might show up after the article (though it only happens In Exod 9.14 and Heb 2.8)

But I still don't understand why it fails to return Luke 2:27 (καὶ ἐν τῷ εἰσαγαγεῖν ) or Gal 4.18 (καὶ μὴ μόνον ἐν τῷ παρεῖναί με)

Edited February 24, 2017 by mgvh

[Mark Allison]

Try the following search:

 

ἐν <FOLLOWED BY> <WITHIN 2 Words> [ARTICLE neuter singular dative] [VERB infinitive]

 

You don't need either the "Within" or the "Followed by" commands between the article and verb if there are no intervening words.

Edited February 24, 2017 by Mark Allison

[mgvh]

@Mark Allison: Well, that does the trick and returns the correct 48 hits I was expecting.

Now try:

ἐν <WITHIN 3 Words> [ARTICLE neuter singular dative] [VERB infinitive]

 

Since you point out that the FOLLOWED BY is not needed if there are no intervening words, this search works the same way and returns all 48 hits. (I needed to make it WITHIN 3 WORDS so that it included the infinitive.)

 

Thanks again to all!

[Mark Allison]

 

ἐν <WITHIN 3 Words> [ARTICLE neuter singular dative] [VERB infinitive]

 

Excellent!