[sword-devel] search idea

[Mads Kiilerich]

Hello,

I have been following the list and the discussion on "search ideas". A
friend of mine, Nicolai Winther-Nielsen, are doing research on topics
related to the advanced search ideas proposed in the thread. I will leave
it to him to explain exactly what he is doing.

Attached below are some comments from Nicolai, in the form of a response
to a mail from Jerry Hastings which I had forwarded to him.

Use of this "bleeding edge" research in search-engines goes far beyond
the "simple" synonym-search, but parts of it might be worth considering
when a new search-engine is designed.

I will continue to follow the discussion and "gateway" any follow-ups.

Within some months, after I've completed my masters degree in EE/CS, I
might contribute to the project myself! ;-)

Until then, keep up the good work,

Mads Kiilerich          Sys.Adm. Stud.Polyt
Mads@Kiilerich.com      Tel. +45 38 16 26 00  Mob. +45 26 20 07 73
Møntmestervej 12B 1th,  DK-2400 NV,  Denmark


>If we could use that kind of index there are some other enhancements we
>could add. A Bible text could be edited to include unsaid words that are
>implied or indirectly referred to. To verses that contain "Lamb of God" one
>could add the words "Jesus" and "Christ". Then when the text is index into
>the synonym index those verses will become hits for "Jesus" and/or "Christ"
>when a synonym search is done.
>
>Some other indexes that would be great to have are indexes of mood,
>setting, topic, and action. Of course there are already topical indexes.
>What I want to be able to do is search for verses that are of a certain
>type, like thankful in mood, and contain a word, like "Lord." I would like
>to search for verses that have a topic of "law" and contain a word, like
>"sin." Someone would have to compile the information for some of these
>indexes. But it would be good if in concept there was support for them. If
>the support was there I think we could get people to compile the
>information, though it will take some time.


Dear Jerry Hastings <hastings@dancris.com> and associates.

I was introduced to your discussion by a Danish friend in computational
studies who looks at the problem of search programs. I am an Old Testament
scholar specializing in linguistics and discourse-pragmatic studies in that
area, but I have also on the strictly applicational side -- no programming
for me, I am afraid -- been using suphisticated programs for linguistic
analysis. In my dissertation on Joshua from 1995 I analyzed a whole book in
Hebrew from morpheme to text level using programs developed by Eep Talstra,
director of the Werkgroep Informatica, at the Vrieje Universiteit in
Amsterdam.

I am involved in a reserch group that currently looks at the possibility of
using Conceptual Graphs for translation and narrative analysis of the
Hebrew Bible.

Based on this knowledge I think you have to devide your answer into two.
1. You can develop fully automated next generation hyper-text search
machines. You would need to use all kinds of background knowledge
dictionaries, i.e. simulate the encylcodedia that any humnan being using
language exploits the whole time. The task of language use is perhaps best
described in Herbet Clark, USING LANGUAGE, CUP 1996. Simulation of all
these cooperative aspects of interactive knowledge acquirement might be
developed by finding ways to search for input on different versions +
archaeological, cultural and sociological knowledge.
Or in short: refine AUTOMATIC searching into ALL sorts of current know
knowledge on the Bible emulating the cognitive selection processes of human
beeings.
AND/OR
2. Have Bible scholars develop expert systems BASED on the Hebrew Bible and
incorporting versions at at later stage. ANY MANUAL research work should
build on the Hebrew texts and add-ons associeted very closely to the
source. Again I think that only the Amsterdam group today has theory
independent data on a high level, i.e. this linguistic data base is not
predetermined for any Biblical Critical or Grammatical Formal ideology, and
is not only the result of some scholarly opinion encoded but rather of
fairly obective and theory-neutral encoding up to the level of the clause
and its internal relations.
  If any advances are going to be made these days I think that we ought to
build of all this formal grammatical knowledge that is there, and then
develop more context-sensitive and socio-cultal knowledge from there.

My point is: be careful to distinguish
	between
what the non-biblical scholar can produce automatically (search of existing
translations for what they are worth)
	and
hat we as Bible scholars need to develop in order to put it into a general
computational format; these demands however will also influence the way
that we should work.

Thus, Jerry, your call for indexes of speech acts, event lines and thematic
spans will need to be SOLVED ultimately in relation to current advances
within the professional groups and cooperation would need to include Bible
scholars, linguists and tranlators, who are already building platforms and
have acquired much of the morpho-syntactic knowledge  WHICH ANYONE STARTING
A PROJECT WILL NEED TO TIE INTO!

Any further information on the wider computational aspects of the Werkgroep
Informatica and its application to the New Testamtent and to translation
can be put to
	ulrik_petersen@sil.org
Please note that I do NOT monitor the list, nor has the TIME or the EXPERT
knowledge to do so, but I assume that I am kept posted on any interesting
discussion by my friend and discussion partner who IS MONOTORING THE LIST
	Mads Kiilerich <Mads@Kiilerich.com>, who monitors the list.

Should you have specific questions to me as a Bible scholar and linguist
you canaddress me personally, off the list:
	nwn@hum.auc.dk

Yours

Nicolai Winther-Nielsen
Associate Professor of Old Testament
Lutheran School of Theology at Aarhus in Denmark