Common Hebrew Phrases

[Ben Spackman]

In his recent article (http://www.sbl-site.org/Article.aspx?ArticleId=420) "Why Hebrew Textbooks Are Different From Those For Other Languages," Frederick Greenspahn notes that "rather than expecting students to master phrases, the Hebrew books list isolated words and sometimes even individual forms."

I can think of several common or stereotypical phrases and word pairs off the top of my head, such as hb'Þqen>W rk"ïz".

But how would I construct a search to find the most common phrases? I've tried several different things in the GSE. The problem is that when ranking them, hits are measured on a word level, so you can't see which pairs (or more) of words occur frequently together.

Any suggestions?

[Joshua Luna]

Ben, this is an interesting point and question. It would be neat to know if BW could do something as complex as this.

[Philip Brown]

Hi, Ben,

I think the place to start would be a WTM search like this:

'#2@nc* w #1@nc*

You can play with the specific parsing specifications, but the hits with a phrase search will be phrase-specific, rather than word-specific.

Use the ASE/QSE, not the commandline.

Hopefully that will get you going. Still a lot of grunt work involved in finding the "most common" phrases, as I see it.

[Glenn Weaver]

Perhaps the KWIK/Collocation Table Module (BW7) can help limit the search results. Here are some starting settings, though you may want to refine them.

Version: WTM
Range: Gen 1:1-Mal 3:24
Word: *@nc*
Left: 0
Right: 1
Codes: 2

Glenn

[Ben Spackman]

I've attached the search I ended up running (and I felt clever for figuring it out ;) )

It takes a while to run. Essentially, it looks for a phrase and a near-identical phrase using the same two nouns, within 5000 words (and crpossing word boundaries). I'd set it for higher, but it already takes over 10 minutes to run on my P4 2.4, 1.2 Gig Ram, 7200 rpm hard-drive.

There's still manual work to be done after that in culling, but it looks good so far.