[B-Greek] Scientific Greek Interinear Project
bunning at myrealbox.com
Wed Feb 4 20:47:25 EST 2004
Based on some of the responses I got to my previous post, I thought I ought
to share what it is I have been working on and where it is headed. I think
you may find it interesting (especially if I get to phase 2) and would
appreciate any comments you have to my questions at the end.
SCIENTIFIC GREEK INTERLINEAR PROJECT - PHASE 1
1. Obtain accurate copies of the major composite texts by comparing
currently available copies found on the Internet and conform them to one
- Stephens (1550) (haven't decided if this is useful or not)
- Scrivener (1894) (and then correct it in a few places to better conform
to KJV because after all, these people claim it is the KJV English that is
inspired not the underlying Greek which was unpublished)
- Byzantine Majority (1991)
- Westcott & Hort (1881) (haven't decided if this is useful or not since
we have NA27)
- Nestle-Aland 26/UBS3 and 27/UBS4 (there are minor difference between
The differences found between the Internet texts are then resolved with
printed copies. So far errors have been found in every electronic Greek text
I have analyzed which means that additive works such as parsings and
Strong's numbers must also be corrected in those places.
2. Acquire parsings for all these texts and standardize them to one format.
Should more than one parsing be stored if there are disagreements for a
given word in a text? The same phrases between texts could also be compared
and checked for differences in the parsings.
3. Create a master word list containing all words/parsing combinations from
all texts. Each word/parsing combination will also be linked to its Greek
root using Strongs numbers with additional numbers created for roots not
contained in Strongs.
4. Assign one unique English word or phrase to each word/parsing
combination, trying not to use transliterated words such as angel or
evangelist. Use the reference links from Strong's to group like word sets
for consistency and Vine's to differentiate shades of meaning from related
concepts. Greek words that carry more than one significant connotation will
have their English meanings separated by slashes, such as "word/reason".
Underscores will be used to keep multiple English words as one unit or for
greater accuracy, such as "[they]_have_been_believing" or "brotherly_love"
6. Assign one unique English word or phrase to certain Greek phrases such as
idioms. Greek phrase substitutions will occur before individual word
substitutions when generating the interlinear. May also establish syntax
rules to apply to the translation.
5. Generate Greek/English interlinear for all the texts showing their
6. Make all resources from the project available for free download on the
web (hope to update things on an ongoing basis).
How phase 1 will move the state of the art forward:
1. Electronic versions of texts that are more accurate than those currently
found on the Internet (this doesn't mean that they are necessarily without
errors, although the odds are improved).
2. An accurate interlinear with consistent word substitutions within a given
text and even consistent word substitutions between texts. This is
automatically done by the computer without any theological bias other than
the choice of words. If you don't like a word choice, you just change it in
one place and then it is automatically updated everywhere.
3. The underlying Greek can be determined from reading the interlinear's
English alone since there is a one to one mapping. This reduces 80% of the
time spent looking up words with the right mood, tense, voice etc. so that
more time can be focused on syntax, structure, and meaning.
4. Various texts can be compared and contrasted in a consistent manner (even
in English). This can sort of be done now, but the roots and parsings are
usually not tied together across different texts (this is sort of a
precursor to phase 2).
5. All data will be available in an SQL compliant database such as MySQL
(free), and will also work with Oracle, Microsoft SQL Server, or any other
ODBC compliant database. The project will be released as open source under
the GNU Public License free of charge so that others can correct it, improve
it, or change it as desired.
6. Someone else (not me) could build a front end to this database and make
an awesome Bible study tool with searching, pop up roots, etc. For now, I
just write SQL queries by hand to find out what I am interested in.
What has been done so far:
1. Finished NA26/UBS3 and NA27/UBS4 texts by comparing the differences
between the Internet texts with my own printed copies. After I double-check
everything, I plan on submitting about 20 errors to CCAT people so that they
can correct their text.
2. Working on Scrivener 1894 text. I need a printed copy to verify 27
inconsistencies found among the texts available on the Internet.
3. Created a root table (also in Excel format) complete with Strong's and
Thayer's definitions while finding several errors in the electronic versions
of these. At first glance, it looks like there are about 484 roots to add to
Strong's from the NA27 text (some of these could be spelling differences).
4. Created several database tables in MySQL and loaded the Root table and
NA27 table with data to verify proof of concept and begin searching for
5. Generated database design (currently under peer review by other techies).
SCIENTIFIC GREEK INTERLINEAR PROJECT - PHASE 2 (in a nut shell)
1. Begin collecting electronic copies of all known texts and variants and
update the previously established tables as necessary.
2. Write a program to automatically parse and tie to Strong's numbers any
variant based on database of parsings and roots generated from the composite
texts. I am going to guess that this will take care of about 90% of the work
leaving the rest to be done by hand.
2. Create master data structure to hold all variants containing the same
phrases and showing branches for the differences.
3. Create equivalence table to equate differences in spelling for the same
word (may put this in phase 1).
4. Create textual criticism programs to perform analysis of texts. It should
be possible to create a tree showing which texts were likely copies of other
texts. At this point, the computer could quickly perform many types of
analysis that was impractical to do by hand.
5. Modify interlinear program to generate Greek/English interlinear from all
or even any subset of desired texts showing variant apparatus (might also be
moved up to phase 1).
So here is what I would like to know:
1. What faulty assumptions am I making?
2. What of this work has already been done? (It would be great if someone
has already done word substitutions at the word/parsing level instead of the
root word level.)
3. What things should be added?
4. What technical barriers will I likely run up against?
5. What political barriers might I run up against such as copyrights,
stepping on other's turf, lacking access to texts, etc?
6. Does someone want to sponsor this work or is there someone I should join
with already doing this work. (Someday I would like to obtain a graduate
degree and this could be a good basis for that.)
7. Anyone want to help? Although the computer can do an amazing part of the
processing, quite often certain steps have to be checked by hand. I myself
move at a pretty slow speed, yet I can plod.
I would definitely appreciate any input that you have. I have left out some
details, so let me know if you have any questions. Thanks.
More information about the B-Greek