Homoeoteleuton and Genealogy

The large number of permutations in layout possible that could generate a master-copy vulnerable to identical homoeoteleuton errors, means the possibility of multiple manuscripts with the same error by coincidence, as James Snapp Jr. has warned:

'You [Naz] mentioned that “When then, for instance, Aleph/B share some 70 such homoioteleuton errors, we know that coincidence is effectively ruled out.” I don’t deny that Aleph and B show signs of a historical relationship and common descent. The thing to watch out for, when an analytical approach is used which assigns special favor to the reading distributed among the consensus of text-types (i.e., Sturz’s approach) is the occurrence of erroneous readings in witnesses which usually disagree. Sturz’s approach seems very vulnerable to this phenomenon.'

- James Snapp Jr., KJV Only Debate Blog, "Theol.Illusions...", Pt 4

From this complication, we can deduce two important cautions:

a) One or two such instances of "agreement in error" between manuscripts does not necessarily imply a common ancestor.

b) Omissions need not be singular to be 'first hand' by the copyist of the manuscript. You could have multiple copies, unrelated genealogically, with the same error.

1) Together these cautions make shared homoeoteleuton errors (h.t.) much less strong in isolation as evidence of a genealogical relationship. It is only in significant numbers that what was flimsy becomes in the end clinching for establishing the source of the errors, and a shared ancestor.

2) At the same time, we become less certain of a prior source for a shared h.t. in the individual case, while maintaining strong certainty of a prior source in the case of multiple instances of shared h.t.

In terms of method and confidence, we can never be very sure of any individual case of a shared h.t. (i.e. single Variation Unit) between MSS originating in a prior ancestor, although that confidence increases with the number of MSS sharing the error.

But if by method (i.e., fiat) we assign all such cases to the reconstructed ancestor (when we have a significant number of shared h.t.), we can nonetheless be confident that the majority of them indeed originated in a prior document, and this probability also increases in strength with the number of shared h.t. cases between the MSS under consideration.

Finally, in spite of these cautions, homoeoteleuton remains one of the strongest types of evidence of shared ancestry, i.e., "agreement in error".

This is because it is far more definite and unique than mere itacisms (which can easily and commonly be generated independently by accident or policy without genealogical dependence), and far more common than other errors (like w.o.r.), and far less likely to be caught in copying (like dittography).

mr. scrivener