tc-list What groups are being named?
- Jean Valentin has begun a lively discussion in relation to the
nomenclature we use.
It seems to me that we are dealing primarily with a classification
problem. The names we give to the groups we discover are a secondary,
albeit thorny, problem.
Powerful classification methods already exist for the kinds of data we
deal with. I tried out a few of these methods on my Hebrews data and found
groups that correspond to the Alexandrian and Byzantine texts we are
accustomed to speaking of. In addition, there was another group that we
are not accustomed to speaking of. (I think that it is a Palestinian or
Syrian text. It includes 025, 044, 0150, and 0278 but not D 06.)
Here is my point. Before arguing about labels, argue about the
classification methods that lead to groups that require the labels. I
suggest that the following have potential to lead us forward:
(1) Multivariate techniques such as classical scaling. These end up
producing maps of manuscript dispositions relative to each other. (Such
maps for Hebrews seem to me to correspond to geographical maps).
(2) Cladistic techniques such as used on the Canterbury Tales corpus by
Peter Robinson and others (see Nature 394, 27 August 1998, 839).
These techniques have fundamental advantages over the ad hoc methods
currently used by New Testament researchers (e.g., quantitative analysis,
profile methods). They are based on mathematical or iterative techniques
that are not tied to preconceived notions of the data. Many of them are
optimal. That is, they produce the best possible classification of the