Searching for the Subject (Hebrew Construct Search)

[miketisdell]

I am still doing some testing, but not getting the results I would expect. For example if I do the following search:

 

 

 

 

[image attached below]

 

 

 

 

I get the following error:

 

 

 

[image attached below]

 

However, there are a number of places where I would expect hits:

 וַיָּבֹ֙אוּ֙ בְּנֵ֣י יִשְׂרָאֵ֔ל (Ge. 42:5)

 

 

וְנִקְבְּצוּ בְּנֵֽי־יְהוּדָ֤ה וּבְנֵֽי־יִשְׂרָאֵל֙ יַחְדָּ֔ו (Hos. 2:2)

 

 

etc...

 

What am I missing?

Edited May 1, 2018 by miketisdell

[Robert Holmstedt]

I completely understand what you've done. It's natural to think that the whole phrase בני ישראל is the subject. But actually, only בני is the subject, narrowly speaking; ישראל is an adjunct to it and thus within the subject phrase, but not the narrow subject itself.

 

I hope this helps.

[mbcvida]

I believe if you mark "Israel' as an adjunct rather than subject it should work for you

[miketisdell]

I completely understand what you've done. It's natural to think that the whole phrase בני ישראל is the subject. But actually, only בני is the subject, narrowly speaking; ישראל is an adjunct to it and thus within the subject phrase, but not the narrow subject itself.

 

I hope this helps.

I would see it as a compound subject; however, I did also try just tagging בן as part of a compound subject and leaving ישראל without the subject tag. Same results.

 

NOTE: I selected "any" on the subject tag.

Edited May 1, 2018 by miketisdell

[miketisdell]

One of the issues was the text I had selected; the other was the adjunct.

When I was using the BHS-T text, it did not work; but when I switched to the HMT-W4 it began to work better. Note: if I run the original search (with both marked as "subject") it returns Josh. 19:49 which doesn't tag Israel as adjunct. If I remove the subject tag I get everything I would expect.

 

It might be helpful for them to make a change so that the entire construct chain is seen as the subject. However, the primary issue here is that the BHS-T text isn't correctly tagged.

[Λύχνις Δαν]

Hi Mike,

 

  The BHS text is not syntax tagged at all. The HMT is one of the two the syntax tagged texts, and it is tagged with the Holmstedt syntax. The other one is the ETCBC text which is tagged with the ETCBC, WIVU as was, syntax. I'm a little surprised that the search ran against the BHS at all. I'd be inclined to suggest that a construct with syntax elements in its search raising a warning when linked to an incompatible text.

 

Thx

D

[Robert Holmstedt]

A compound subject is one in which two or more conjoined items altogether fulfill the syntactic role of subject. Israel is not the subject in the phrase in question. It is the clitic host (nomen rectum) of the bound (construct) phrase. That is, it is embedded within the subject, but not the subject. This is the linguistic analysis behind my syntax database.

[miketisdell]

A compound subject is one in which two or more conjoined items altogether fulfill the syntactic role of subject. Israel is not the subject in the phrase in question. It is the clitic host (nomen rectum) of the bound (construct) phrase. That is, it is embedded within the subject, but not the subject. This is the linguistic analysis behind my syntax database.

I didn't think "Israel" was the subject but I do think that grammatically it is part of a compound subject. In the example from Hosea 2:2 "וְנִקְבְּצוּ בְּנֵֽי־יְהוּדָ֤ה וּבְנֵֽי־יִשְׂרָאֵל֙ יַחְדָּ֔ו" I would define the subject as "בְּנֵֽי־יְהוּדָ֤ה וּבְנֵֽי־יִשְׂרָאֵל֙". The syntax module seems to agree with that understanding. See attached screen shot. However, that being said, it is only a minor issue to me; the biggest issue I was having was related to using the wrong text i.e. BHS-T. When I did that nothing worked.

 

 

Also, whether a mistake or not, Josh. 19:49 appears to be tagged the way I would expect.

 

‏ וַיְכַלּ֥וּ לִנְחֹל־אֶת־הָאָ֖רֶץ לִגְבֽוּלֹתֶ֑יהָ וַיִּתְּנ֨וּ בְנֵי־יִשְׂרָאֵ֧ל נַחֲלָ֛ה לִיהוֹשֻׁ֥עַ בִּן־נ֖וּן בְּתוֹכָֽם׃

 

 

 

Overall, I love this feature!

 

 

 

 

 

Edited May 1, 2018 by miketisdell

[Robert Holmstedt]

It's important to have clarity on this, or the syntax database will never turn up the results you expect. Look at the Adjunct A tag above Judah and Israel. They are "within" the subject as adjuncts. They are a part of the subject phrase, but not tagged as subjects. I encourage you to read the background docs I've provided for the database.

Edited May 2, 2018 by Robert Holmstedt

[Peter Bekins]

Mike, it will help you create better searches if you think in terms of hierarchy and phrase structure when syntax is tagged. Commands like 'within' are a bit of a work around for when only the morphology was tagged.

 

To restate what Rob said but slightly differently, tags like SUBJECT and ADJUNCT label the role of a single constituent within its immediate phrase. If the subject phrase is only one word, then SUBJECT will basically get you what you want. What you asked for, however, are two separate SUBJECT words, one beginning with בְּנֵ֣י and the second with יִשְׂרָאֵ֔ל, that are within 1.

 

If you want a more complex phrase like בְּנֵ֣י יִשְׂרָאֵ֔ל as a subject phrase but you do not want to specify the internal phrase structure, then you can just do it like this using subject phrase instead of SUBJECT and setting depth to 1.

 

 

Setting depth = 1 tells the algorithm to ignore 1 level of hierarchy, so it skips over the relationship of בְּנֵ֣י to יִשְׂרָאֵ֔ל.

 

Finally, Josh 19:49 is tagged incorrectly, which I am assuming was a mistake by Rob in the editing process because the tagger would never have made such a simple mistake  .

 

Pete