Computer assisted collation of New Testament manuscripts

Abstract
From their composition in the first century AD
until 1516 when the first printed edition appeared,
the various books of the Greek New Testament
were hand-copied. Over 5000 Greek New Testament manuscripts still survive, some dating back
to the second century.
A major aid to constructing a critical edition of the
Greek New Testament is a collation of significant
MSS. To reduce the number of MSS included to a
manageable level, only those which are significant
for the reconstruction of the history of the text
should be chosen. Here, this has been achieved by
applying the simple criterion of selecting uncial
MSS alone. This is a rough method, but is adequate provided that the chosen collation technique can
readily admit newly identified significant MSS.
Fortunately, Peter Robinson has developed a flexible collation program called Collate. This is
capable of including up to 100 MSS for collation
against any chosen base text, and inclusion of
additional MSS is simple.
The New Testament Epistle to the Hebrews was
selected as a test case for computer assisted collation. There are 31 known MSS of Hebrews in
uncial script, of which 28 are accessible. They
have variable spelling, sparse accenting, punctuation and intra-word spacing, and utilise various
short-hand devices such as nomina sacra contractions. Among them are highly significant MSS
such as P46, 01, 03 and 06, along with representatives of a number of textual families, including the Byzantine. It is not unreasonable to expect these 28 to contain virtually the whole range
of non-trivial variation found among all MSS of
the Greek New Testament in Hebrews, although
this expectation is yet to be tested.
Frequently, the MSS have been corrected by subsequent scribes. Most have one or two correctors,
while some have none and a few have more than
two. P46 appears to have five! Corrector information is important because its usual effect is to move
a MS from one textual family toward another. As
corrections are often the result of a scribe’s quest
to make a MS conform to an independent exemplar, a corrected MS deserves to be treated as a
distinct entity. If each corrector’s version is separated out, the 28 MSS give rise to over seventy
different witnesses.
Transcription aims to encode information important to mapping trajectories in the history of the
text, as well as the text itself. Which jot or tittle
will eventually point to a relationship between
MSS is not known before analysis of completed
collations, so it is prudent to include as much
information as might reasonably be thought important. In the main, this has been achieved by the
use of tags, as recommended by Peter Robinson in
the Collate user guide. In the interests of standardisation, it would have been better to use the TEI
P3 guidelines for marking up the text. Some ramifications of this deficiency and possible solutions
will be discussed.
My raw transcriptions produce a messy output
from Collate. By contrast, the output is easier to
understand when preprocessed transcriptions are
fed in. A separate program was written to perform
this preprocessing. Now, a verse can be specified
and a clear collation prints out before our eyes!
The preceding outline shows that it is now possible for a personal computer to produce an elegant
collation of significant Greek New Testament
MSS using the Collate program with preprocessed
transcriptions. This could prove useful to those
who would like to have an understandable survey
of variation within a particular section of text at
their finger-tips. It may also provide a new path
towards the goal of a comprehensive critical edition.