PDA

View Full Version : Homonyms sensitive search in BHS



yugu
06-16-2007, 11:23 AM
Is it possible to arrange homonyms sensitive search in WTM? If yes - how should it be done?
Also once i proposed to the BW staff to add the possibility to search on lemma, that is equal to the very homonym which was Right-Button clicked - is there anyone else who thinks it would be helpful?
Thanks,
Yuri.

Ben Spackman
06-16-2007, 01:40 PM
It is indeed possible. Make sure the pop-up morphology help in the command line is on.

The homonyms are coded by letter. Homonym 1=a, b=2 and so on. In the morphology search (ie. when you're searching in the WTM), there is a space to indicate homonyms at the end of the string. (You can see a chart for this in section 44d of the help.)

By using the pop-up morphology help in the command line, you can see where this comes in. Type .*@ and then just keep typing ? until you see where the option is for the Homonym code. For any noun, homonym 1, it would look like this. .*@n????+??????Ha

Of course, you don't need all those ?'s. You could also type *@n*a, or specify as much of the morphology as you need to.

Ben Spackman
06-17-2007, 01:47 PM
Forgot to add- I would also really like to see lemma searching implemented, but both in the way you propose, but also differently way.

I'd like to be able to run root searches, ie. something that would find zbx as a verb, mzbx the noun, etc.

It would make finding cognate accusatives much easier, for example.

yugu
06-17-2007, 03:33 PM
Thank you very much Ben,
Yuri.

Philip Brown
06-18-2007, 09:49 PM
I'd like to chime in here.

There are at least two places in BW where, IMO, homonym sensitivity should be default behavior:
1) All lemma searches should by default be homonym sensitive. If a user doesn't want such sensitivity, it could be turned off, but since many (most?) first year users of Hebrew don't know what homonyms are, their ignorance combined with a lemma search that is not homonym sensitive produces less than exegetically satisfactory results. I've received a few such results in student papers.

2) Vocabulary counts in the Vocab module should be redone with homonyms distinguished.

Cognate searches such as you're proposing, Ben, would be nice, but so far as I know the cognates would have to be manually identified and cataloged.

Seems like a person could get a fairly decent idea of cognates simply by doing a xbz* WTM search. If a person knows the typical morphological patterns of Hebrew nouns, they could relatively easily generate a set of cognates. And there's always BDB's root-based approach. Start with a verb and scan through the subsequent entries. 'Course, I know you already know all this stuff, Ben, but others perhaps don't.