The previous post showed how an analysis of the variant data from the ABS scroll versus the BLS, HUC, SMU, and VS scroll showed the the ABS scroll has from more in common with scrolls in the Masoretic tradition that the post 1642 flood Kaifeng scrolls. Exodus in ABS shows this as well. Here is a summary of the evidence:
Main result
The five scrolls do not form one homogeneous group. There is a very strong structural pattern:
Pair Similarity
ABS – MS 93.7%
ABS – HUC 86.4%
HUC – MS 83.1%
BLS – HUC 53.8%
BLS – VS 51.6%
ABS – BLS 46.2%
BLS – MS 46.2%
HUC – VS 23.1%
ABS – VS 7.6%
VS – MS 7.6%
Here “similarity” means: when both scrolls are explicitly represented in an entry, what percentage of the time are they on the same side of the variant—both giving the variant or both giving the Masoretic spelling?
The data strongly separates into two tendencies:
Cluster 1: ABS / HUC / MS
• ABS–MS: 93.7%
• ABS–HUC: 86.4%
• HUC–MS: 83.1%
These three are extraordinarily close relative to the other relationships.
Cluster 2: BLS / VS
• BLS–VS: 51.6%
And VS is very unlike the first cluster:
• VS–ABS: 7.6%
• VS–MS: 7.6%
• VS–HUC: 23.1%
BLS appears to be much more intermediate: it agrees with VS about half the time and with the ABS/HUC/MS group roughly half the time.
An especially interesting finding
Looking only at entries where all five scrolls are explicitly assigned, there are 59 such entries.
The two dominant configurations are:
• MVMVM — 25/59 (42.4%)
• MMMVM — 22/59 (37.3%)
Using the order ABS / BLS / HUC / VS / MS, this means:
MVMVM
ABS = Masoretic
BLS = Variant
HUC = Masoretic
VS = Variant
MS = Masoretic
while:
MMMVM
ABS = Masoretic
BLS = Masoretic
HUC = Masoretic
VS = Variant
MS = Masoretic
So in the majority of the fully specified cases, VS is the outlier, while ABS, HUC, and MS move together. That's a considerably more interesting result than simply saying that the scrolls have different numbers of variants.
Variant type matters
The 74 Plene cases dominate the dataset. Among those:
• ABS–MS: 98.4%
• ABS–HUC: 88.0%
• HUC–MS: 87.0%
• ABS–VS: 3.2%
• VS–MS: 2.7%
• HUC–VS: 14.8%
So the basic clustering isn't being driven merely by the 16 Error cases. It is especially pronounced in the Plene material.
The 16 Error cases show more complicated relationships—for example, BLS–HUC rises to 80%—so I would analyze Plene, Defective, and Error separately rather than treating them as one population.
ABS, MS, and HUC form a statistically robust similarity cluster, while VS is strongly differentiated from that cluster. BLS occupies an intermediate position, and its apparent association with VS is substantially less stable.
As I wrote in the other entry, the ABS scroll most likely is not the Scroll of Moses. It does not have enough characteristics in common with BLS, SMU and VS. It is strongly paired with HUC square books. Whatever source or sources the Kaifeng Jews used to produce the BLS and VS scroll (we will need to wait for my reading of the remainder of the SMU, the rest of CUS, then OUS, in other words, most of four other scrolls) to reach firmer conclusions.
Comments
Post a Comment