The Talmud Text Databank

Students of rabbinic literature are well aware that the great classical
compendia of Judaism, the Babylonian Talmud, is being read and studied
today almost exclusively from a late printed edition, which itself is a
direct textual descendent of the first printed edition of the entire
Talmud (Venice, 1520/21).  Much manuscript material has been made
available in various formats during the last century or so, but the
vulgate text is still often referred to as the definitive Talmud text.

The Lieberman Institute proposes to change this state of affairs with
its Talmud Text Databank. The databank is designed to disseminate the
manuscript evidence for the Babylonian Talmud, with the evidence of the
various witnesses to the text arranged line by line, as an on-line
computer variorum edition.  In addition to the development of this
variorum edition, the software controlling the databank will have
sophisticated search and retrieval capability allowing the scholar to
search for combinations and permutations of text through the entire
mass of material which could never be attempted by means of a simple
printed concordance.

The primary focus of the Talmud Text Databank project at present is
input of the texts into the computer.  First priority has been given to
the major manuscripts of the Babylonian Talmud. Of these manuscripts,
approximately one-third are in various stages of input, proofreading
and verification.  The remaining two-thirds, plus most early printed
editions and the Genizah fragments have yet to be treated.

For the purposes of this project, a unit of text to be input is one
manuscript version of one tractate (i.e. the Munich Ms. 95 version of
Berakhot).  Including Genizah fragments, there will be a total of
approximately 250 text units in the databank.  To date, 67 text units
(including at least one unit for every tractate) have been entered and
have either been verified or are in the proofreading stage.

Information about which manuscripts have been input for each tractate
may be obtained from the Lieberman Institute.

A great deal of attention is devoted to ensuring that the texts input
into the computer are error-free.  Text input from manuscripts is much
more complicated than input from printed editions.  Very often the
handwriting is difficult to read and many manuscripts contain additions
and deletions which must be included in the computer-generated text
though clearly distinguished from the original reading. In order to
attain the highest level of accuracy possible, each text undergoes three
levels of proofreading, the final two by talmudic experts.

In addition to text input, the Lieberman Institute has also been
advancing towards its programming goals.  A prototype line comparison
program has been developed and the first version is in operation. A
sophisticated commercial search-engine program has been purchased which
has been applied to the units of manuscript text already input into the
databank, as well as to a "control" text of the printed Vilna edition
which was acquired with the search engine.

The Talmud Text Databank project is coordinated by Dr. Chaim Milikowsky.
The project is based at the Jerusalem offices of The Saul Lieberman
Institute of Talmudic Research on the Jerusalem Campus of The Jewish
Theological Seminary of America.