Significant change in search results after update

Page 1 of 1 (8 items)
This post has 7 Replies | 4 Followers

Posts 11
Brad Berglind | Forum Activity | Posted: Wed, Nov 13 2013 6:16 PM

Before the update, I searched for the manuscript form of Yom from Gen 1:5 in the NASB95 Old Testament and got 495 hits.  After the update, the same search got 1,399 hits.  It seems that the search is returning hits like יּ֖וֹם and where י֔וֹם is in a word - likeכָּל־הַיּֽוֹם׃.  

What's up with this??  It seems like the search before the update was more correct than after the update.

Brad Berglind

MacBookPro Retina, OSX 10.8, 16GB RAM,  Solid State Drive

Atlanta Logos User Group

Posts 27251
Forum MVP
Dave Hooton | Forum Activity | Replied: Thu, Nov 14 2013 3:48 AM

Brad Berglind:
What's up with this??  It seems like the search before the update was more correct than after the update.

I'm not a Hebrew expert so I can only refer you to this document for the changes that were introduced in 5.2. The comment to note is "By default, Hebrew is not sensitive to any marks in the query" - so has the default changed or are the results now correct? I can reproduce the 5.1 results by using either:

  • [match dagesh] י֔וֹם; or
  • [match pointed] י֔וֹם

I note the results are unchanged for LHB (Lexham Hebrew Bible)  i.e. 241 hits. So I'm at a loss to explain the differences for the reverse interlinear bibles e.g. NASB95, ESV

Dave
===

Windows 10 & Android 8

Posts 86
Scott Fleischman | Forum Activity | Replied: Thu, Nov 14 2013 10:07 AM

As Dave mentioned, the change in search results is due to Hebrew no longer being sensitive to the dagesh mark by default. Yom is a common word, which is why there is such a significant change in the result count.

The reasoning behind the change of default Hebrew search behavior is that the dagesh mark is not an important mark for distinguishing words. Yom is still the same word whether or not it has a dagesh in it, and it makes sense to have the default search behavior reflect this. We added the new search options for various mark sensitivity if you need to search for specific marks, or the lack of specific marks.

Dave also mentions another point: searching for a Hebrew word in LHB and NASB95 return different results. Thanks for calling this to our attention. We will investigate whether we should change this behavior.

The reason for the difference is that NASB95 is backed by a reverse interlinear, which aligns the English text to the Hebrew text. In order to make the alignment precise, the reverse interlinear breaks up the Hebrew text into segments, so "בְּרֵאשִׁ֖ית" in Gen 1:1 is broken up into "בְּ" and "רֵאשִׁ֖ית". Searching NASB95 for "רֵאשִׁ֖ית" returns a hit for Gen 1:1, but searching LHB does not return a hit for Gen 1:1 because it does not match "בְּרֵאשִׁ֖ית". This difference between searching in Hebrew in NASB95 and LHB is not related to the mark sensitivity change in 5.2.

Posts 11
Brad Berglind | Forum Activity | Replied: Thu, Nov 14 2013 11:04 AM

Thanks so much for the explanations.  Any suggestions on what these marks are and which ones change the meaning of the word?  If I want to find all the places in the OT with the same manuscript yom as Gen 1:5 do I ignore the marks or use some of the matched terms.  The number of verses vary a lot:

Search Term

Number verses

Verses in Base but not in Matched search

י֥וֹם

1,237

Base

[match exact] י֥וֹם

108

Gen 1:14,16,18, Gen 2:2, 3,4

[match dagesh] י֥וֹם

438

Gen 1:14,16,18, Gen 2:2 Gen 5:1

[match accents] י֥וֹם

279

Gen 1:14,16,18 Gen 2:2,3,4,17, Gen 3:5,8

[match massora] י֥וֹם

1,237

 

[match rafe] י֥וֹם

1,237

 

[match critical] י֥וֹם

1,237

 

[match maqqef] י֥וֹם

1,237

 

[match pointed] י֥וֹם

438

Gen 1:14,16,18, Gen 2:2 Gen 5:1

[match cantillated] י֥וֹם

108

Gen 1:14,16,18, Gen 2:2, 3,4

So if I am trying to do an exegesis analysis of Gen 1 and Yom, should I look at 108 verses or 1,237 verses.  If the marks don't change the definition of the word then the answer is 1,237.  If they do, it could be 108 or 279 or 438 verses.  Please advise.

Thanks for your help!

Brad

Brad Berglind

MacBookPro Retina, OSX 10.8, 16GB RAM,  Solid State Drive

Atlanta Logos User Group

Posts 817
LogosEmployee
Eli Evans (Faithlife) | Forum Activity | Replied: Mon, Dec 9 2013 11:50 AM

Not knowing exactly what you hope to find, I guess that [match pointed] is what you want. That will get you surface text ("manuscript") words with the same letters and will ignore the cantillation. Broadly speaking, the cantillation marks mainly concern how the verse is to be recited -- their function lies somewhere between musical notation and our punctuation marks like periods and commas. As such, they don't affect meaning; none of the introductory Hebrew grammars I've read mentions them except in passing. Gesenius has the longest treatment I could find. Far more than you ever wanted to know about the cantillation marks can be found in this exhaustive (and exhauting!) article by Helmut Richter.

But beware: Hebrew orthography is a complex topic. Searching on a single Hebrew surface text spelling is a fragile method that is subject to many hidden complications:

  • Sometimes even the vowel points are changed just a little a bit when a word appears at the end of a verse (pausal forms). The pausal and non-pausal forms of Yom happen to be spelled with the same vowels, but that's just a coincidence. Many words have different pausal forms.
  • Sometimes grammatical/functional categories which can greatly affect meaning are not reflected in how the word is spelled. For example, the construct and absolute forms of Yom happen to be spelled the same way. You have to use other clues from the sentence to tell them apart.
  • Yom ends with a final form of the letter Mem. There are some occurrences of the word which have pronominal suffixes attached at the end of the word which changes the final Mem to its medial form -- which is the same consonant, but encoded with a different Unicode code point, that is, a different consonant as far as Logos Search is concerned. For example, take בְּיוֹמוֹ֩ ("the same day," literally "in his day") in Deuteronomy 24:15. Just guessing, but I bet you want that verse on your list -- it mentions sunset as part of the definition of "the same day" -- but none of the surface text searches you're trying will find it because "Yom" is spelled there with a medial Mem.

A better way to begin would be to search for a combination of lemma and morph. The morphology codes don't document how a word is spelled, but rather how it is formed (which is closely related to, but distinct from, how it is spelled) and how it functions in a given context (I'm painting with a very broad brush here). More the point, searching for a lemma+morph combination helps to lump together insignificant variations in surface spelling while at the same time differentiating between different forms of a word (or even different homographs!) that happen to be spelled the same.

  • For example, searching for lemma:יוֹם.1@NC?SA in LHB (Yom when it is a noun, common, singular, absolute) yields 1,141 results in 1,021 verses. Looking at that list, you'll find the "Yom" for the sixth day in Genesis 1:31 is missing, because it's in the construct, rather than absolute, state.
  • Searching for lemma:יוֹם.1@NC?SC in LHB (yom when it is a noun, common, singular, construct) yields an additional 308 results in 274 verses, including Genesis 1:31.
  • Searching for lemma:יוֹם.1@NC?S (noun common singular, no regard to absolute/construct state) yields the sum of 1,141 + 308, or 1,449 results in 1,248 verses. 

I think that list of 1,449 results in 1,248 verses is a better starting point for studying how Yom (singular) is used throughout scripture than any surface text search would get you. Deuteronomy 24:15 is on that list.

I highly recommend the "Learn to Use Biblical Greek and Hebrew with Logos" product. Dr. Heiser covers all sorts of tips and tricks on how to best search for Hebrew word forms, and explains how to avoid many of the common pitfalls. He also discusses extensively the differences you'll encounter between searching a Hebrew Bible directly and searching for Hebrew words through the mediating influence of an English reverse interlinear (ie, searching NASB).

I hope that helps!

Posts 2589
Lee | Forum Activity | Replied: Mon, Dec 9 2013 12:22 PM

Strong suggestion: can threads like these be "stickied" in the forums?

An alternative (but not as good, in my opinion) would be to mention it in the wiki.

Posts 9947
George Somsel | Forum Activity | Replied: Mon, Dec 9 2013 1:31 PM

Brad Berglind:

Before the update, I searched for the manuscript form of Yom from Gen 1:5 in the NASB95 Old Testament and got 495 hits.  After the update, the same search got 1,399 hits.  It seems that the search is returning hits like יּ֖וֹם and where י֔וֹם is in a word - likeכָּל־הַיּֽוֹם׃.  

What's up with this??  It seems like the search before the update was more correct than after the update.

From your question I would assume that you don't know Hebrew.  You can nevertheless search the Hebrew morphology for the word you wish (whether יוֹם or something else) by using the morphology search.  Note the screenshot below which has two searches for יוֹם, one a search of BHW4.18 and the other of the NASB.  While they do differ slightly, they are relatively close in their results (Don't ask me why they differ since I haven't checked and don't wish to take the time).  I would recommend the BDB lexicon or HALOT (newer), but if you don't have one of them, you can download the FREE ABDB (Abridged Brown-Driver-Briggs) which should give you some idea of the various ways in which the word is used.

george
gfsomsel

יְמֵי־שְׁנוֹתֵינוּ בָהֶם שִׁבְעִים שָׁנָה וְאִם בִּגְבוּרֹת שְׁמוֹנִים שָׁנָה וְרָהְבָּם עָמָל וָאָוֶן

Posts 11
Brad Berglind | Forum Activity | Replied: Mon, Dec 9 2013 2:04 PM

Great help - Thank You very much.  I do have HALOT and will dig into it.  And you are right, I do not know Hebrew.  So far it looks like I have to identify each location of YOM and research the verse to see if it means 24 hours or something longer.  Lots of work if there are over 1000 locations.  I was hoping to whittle the task down substantially.

Again, thanks for the advice.

Brad

Brad Berglind

MacBookPro Retina, OSX 10.8, 16GB RAM,  Solid State Drive

Atlanta Logos User Group

Page 1 of 1 (8 items) | RSS