[Count 1-25] finds a word occurring 41 times in BHS

[Martin Z]

in Judges 9:25, ארב is found.

CDCH says it occurs 41 times in the BHS.

However, when I open the Analysis Window, it says it occurs only 23 times. I think that is why it is included.

 

ארב to ambush = 23

I don't know where does 23 come from in Accordance.

I'm not sure if it is the only case or not.

I think I've seen another case before.

Blessings,

Martin

[Natan Rubenstein]

Yes, I've been doing a comprehensive review of all Accordance's search features. My analysis is part of larger project on which I am working. I can list bugs like this consistently. I bought Accordance on the recommendation from colleagues that it hosted powerful search features. That is true. It does, but it gives a lot of dirt in the search results as well.

[Ken Simpson]

Hi Martin (and Natan)

Please tell us about cases of “dirt” so that we can either explain why you are getting what you see, or so that we can fix the issue that is causing the “dirt”. We pride ourselves on the accuracy of our search engine, and while nothing is perfect in the material world, we do want to approach said state.

 

On the OP, if you do a search for ארב‎

 

and then an analysis you find

 

Total number of verses = 44

(total number of verses displayed = 44)

 

(ארב 46 total words)

 

Number of different forms = 5:

 

‏ארב‎ to ambush = 23

‏(ארב) אֹרֵב ambush = 18

‏(ארב) אֹרֶב ambush = 2

‏(ארב) אֶרֶב lair = 2

(ארב) אֲרַב Arab = 1

 

As you can see, the ארב‎ lemma occurs 41 times in the first two items (hence CDCH) picking up the standard verb and the noun form of the lemma.

 

CDCH excludes the lemmas that are not ארב‎ and excludes אֹרֶב and אֹרֶב and the word for Arab, and the different lemma that means lair.

 

if you tell accordance to do a search for ארב but confine it to the exact lemma (ארב=) you will find it matches the CDCH result.

Edited December 4, 2014 by Ken Simpson

[Martin Z]

Hi Ken,

 

Thank you very much for the explanation.

I searched for the hits and highlighted all the words. I was surprised to see that a word I did not expect to be highlighted was actually highlighted.

My confusion was that both CDCH and Mitchel's vocab book say it occurs 41 times. I trusted the Lexicon and Vocab book without realizing that the authors sometimes are not consistent throughout the books. I should have checked for myself.

You explanation makes perfect sense.

 

Blessings,

Martin

[Joel Brown]

I'm sorry you feel there is dirt in your results, but the results are highly consistent for what you search for. When you search for [COUNT 1-25], you are asking to list all lemmas that occur in the text, 25 times or less. As you saw from the results in the analysis window, the lemma occurs 23 times, so the search was correct.

 

Different tools often use different schema to calculate their results, so saying one is right or one is wrong isn't accurate - they are simply representing different bits of data. In the case of CDCH, they chose to combine multiple lemmas that Accordance (or rather, the Westminister morphology) has split apart. If you compare the result with the DCH (non-concise), it lists only 26 occurrences of the word. The 3 extra in this case are due to DCH's combination of the noun and verbal forms of the word.

 

Edit: Oops, Ken beat me!

[Ken Simpson]

No oops there Joel, really helpful extra information!

[Martin Z]

CDCH understands אֹרֵב as participle of ארב.

vb.—Qal 38.4.5 Pf. ‏אָרְבוּ‎; impf. ‏יֶאֱרֹב‎, ‏יֶאֱרֹבוּ‎ (Q ‏יורבו‎), ‏נֶאֶרְבָה‎; + waw ‏וַיֶּאֱרֹב‎ (‏וַיָּרֶב‎), ‏וַיֶּאֶרְבוּ‎; ‏וְאָרַב‎; impv. ‏אֱרֹב‎; ptc. ‏אֹרֵב‎, ‏אֹרְבִים‎; inf. ‏אֱרָב־—1

[Ken Simpson]

Ahhh, these Hebrew scholars. If only they could agree!

[Natan Rubenstein]

My research of Accordance's search features is part of a larger review project. I've only frequented the forums to ensure the problems I've found have not been addressed. The results of my study will be posted on a website in due course. I believe a lot of Accordance search problems are based on some naive definitions linguistics, for example, "word" is especially problematic. In addition, the syntax search functions are particularly buggy. I have many more items I could list.

[Abram K-J]

Well... I'm just going to pop some popcorn and follow this topic.

[Ken Simpson]

Well, I am not sure where we can go from here. If Natan is of the view that words as a formal definition are problematic then he is very welcome to that view of Hebrew linguistics. Of course there is an issue to some degree. However Accordance is built on that definition (up to a point) and all the tagging that we get from Westminster-Groves is built on that (or at least I assume so) so I suspect it would be a mammoth task to change the way we approach the Hebrew text.

 

Alternatively, as in most situations like this, you can choose to work with the paradigm that the software gives you with it’s strengths and weaknesses, and reframe the way you think about it.

 

I suspect for Natan, that may not be what he would prefer, but at least we can say that is what we have.

 

On the issue of the syntax databases, it is incomplete, very much a work in progress, and if you find errors in the searches please report them in the forums so we can improve the search engine.

 

Thanks

Edited December 5, 2014 by Ken Simpson