« IE7 FAQ & Libronix Update | Main | What's Ugaritic Got to Do with Anything? »

October 25, 2006

Greek Syntax: Searching for Granville Sharp

If you've studied NT Greek, you've likely heard of something called the "Granville Sharp Rule".

If you've been around Bible software, you know that many folks use "finding Granville Sharp" as a sort of litmus test for the capabilities of their Bible software.

The OpenText.org Syntactically Analyzed Greek New Testament gives us an opportunity to examine what the Granville Sharp rule really is and to think about new ways to find instances of it.

Awhile back I wrote a paper for internal use here at Logos examining what "Granville Sharp" is and how to find it using the traditional "morphology+proximity+agreement" approach. This approach has problems because one must approximate relationships between words using morphological criteria (i.e. part-of-speech data), morphological agreement (i.e. terms 'agree' in their specified case), and word proximity (i.e. words are within N words of each other).

Then I examined finding Granville Sharp using the OpenText.org SAGNT. With the syntax annotation, you're freed from approximating relationships with morphology+proximity+agreement and empowered to actually specify relationships that the syntax annotation encodes.

The 17-page PDF document linked below is that paper. It has explanation and screen shots of the queries, graphs and whatnot so it should help in thinking about how to go about isolating syntactic structures via searching the OpenText.org SAGNT. It might even help get the juices flowing for those considering the Logos/SBL Technology Paper Awards.

I've also included the two syntax queries discussed in the paper. I just tested them on 3.0b Beta 2, so if you have that version installed, you should be fine. I would think it would work on any flavor of 3.0, but why not upgrade if you're not up to date?

Copy the queries to your My Documents\Libronix DLS\Syntax Queries folder and then load them as you would any other syntax search, from the Load ... button in the Syntax Search dialogue.

Posted by Rick Brannan at October 25, 2006 6:00 AM

Comments

Thanks for this good post and paper! I suspected that syntactical searching would become the gold standard of precision in answering these types of questions, and the Granville Sharp rule is an obvious example.

I found it interesting that, paradoxically, precision is why you would not restrict articles within your graphical query after ???. Isn’t that what allowed for the majority of false hits (because the query permitted article-governed second substantives)? I wonder how many true TSKS occurrences have an article between the first substantive and the anarthrous one.

But I see your point; this highlights the limitations of purely morphological solutions.

Posted by: Mike at October 26, 2006 1:25 AM

Oops, sorry. I see that you addressed that problem sufficiently in your paper. I should have read it better before commenting. :-)

Posted by: Mike at October 26, 2006 1:45 AM

I was finally able to get the two queries above to work. First I tired to create my own folder and that is where I saved the two files above to. This what I could not get to load, because it would not show up. Finally, I went back and created a simple query and saved it. That created a folder that I could read and then I moved the two queries above into this newly created folder. Then I could load them from the syntax search area.

Posted by: Ken Neighoff at October 26, 2006 11:22 AM

Thanks for your paper. I have just learned about the Grandville Sharp rule and was trying to figure out just how to setup a Syntax search. This is giving me a clearer understanding how to design a search.

Thanks again!

Kevin

Posted by: Kevin Nicholas at November 8, 2006 8:13 PM

Post a comment




Remember Me?

(you may use HTML tags for style)