daniel stoekl

ABSTRACT: Old Caves and Young Caves: Two Qumran Collections?

The scenario that all or most caves served as emergency hiding places for the Qumran collection around 68 CE has to be discarded or fundamentally modified for the following reason. An examination of the average age of the scrolls of each of the Qumran caves shows a huge gap between “old” caves (caves 1 and 4 with an average age between 37 and 44 BCE) and “young” caves (caves (2, 3, 5, 6 and 11 with an average age between 5 and 25 CE). A statistical analysis proves that such a distribution cannot be achieved by random means. Therefore, the manuscripts from caves 1 and 4 cannot come from the same collection as those found in caves 2,3,5,6 and 11 (p<0.0001 for a Kruskal-Wallis test, similar numbers for a series of T-Tests). Old caves include very young manuscripts and vice versa. The library was therefore not arranged according to age.

Devorah Dimant has shown that the caves are intimately connected by genre and “Sectarianism.” Most probably, therefore, the "old" caves 1 and 4 represent the manuscript collection of the same group as the “young” caves 2,3,5,6,11 but at an earlier point in history. Assuming Qumran was destroyed by fire around 4 BCE, caused most probably by an attack (Jodi Magness), I try to address the question how manuscripts older than that fire survived.
The statistical exams based on the discrepancy in average age shows that the origin of the Qumran scrolls cannot be one book collection that was then distributed in haste before an attack to various caves. We have to speak of at least two collections and/or two events when they were hidden.

This article has been published in Dead Sea Discoveries 14:3 (Oct 2007) pp. 313-333.

 (This full version now replaces the abbreviated version, available since Oct 2005 here and at the Nordic Qumran Network website.
The now outdated version will continue to be accessible via a stable URL at the HAL database of the CNRS here .

Discussions in internet fora in 2005 (pdf)

Echos in the press.