noun phrase, etc.

c stirling bartholomew cc.constantine at
Mon Mar 18 23:04:43 EST 2002

on 3/18/02 5:46 PM, Mark Beatty wrote:

> Greek allows breaking economy for discourse/semantic purposes.

I should add that "semantic" is the critical word when talking about writing
specifications  for noun phrases (NP) and verb phrases (VP).

Defining a sufficient set of production rules for automated identification
of noun phrases, etc. in a morphologically tagged GNT is going to be
difficult precisely because identification of a noun phrase involves both
morphology and semantic information.

To illustrate this point, print out just the morphologically tags from a
complex paragraph of NT Greek and see if you can sort out all the
constituents correctly based on morphological information alone. I think you
will find this a challenge.

So the only point is (a somewhat trivial point actually) that you need
semantic information in your GNT database before you can do the kind of
searching Moon was asking for. Moon's request is a legitimate one. I would
also like to be able to search for NP's, VP's, agents, recipients,  etc.
However, as far as I know there isn't a database in any of the commercially
marketed electronic GNT's that will support his request.

As Mark has pointed out, proposed solutions to questions like this tend to
show up in dissertations. And people writing dissertations often design and
build their own databases for doing their research. This gives them the
freedom to include all kinds of things beyond morphology in their database.
But "the rest of us" are pretty much stuck using  the databases marketed as
packaged products for the masses.



Clayton Stirling Bartholomew
Three Tree Point
P.O. Box 255 Seahurst WA 98062

More information about the B-Greek mailing list