External Criteria in Textual Criticism

I recently read Reinventing Jesus: How Contemporary Skeptics Miss the Real Jesus and Mislead Popular Culture (Kregal, 2006).  This book was written by J. Ed Komoszewski, M. James Sawyer, and Daniel B. Wallace.  A significant part of the book deals with textual criticism.  There are a large number of manuscripts of New Testament writings, yet we do not have an original text.  How can we know which manuscripts are faithful to what the original text said?  This book offers solutions.

I've blogged about textual criticism in the past.  See here, here, and here for examples.  Yet, I did learn some things in reading Reinventing Jesus, particularly when the book was talking about external criteria in terms of evaluating manuscripts.  Here are three issues:

1.  In debates against King James Version-Only advocates, defenders of the Alexandria manuscripts hold that they are reliable because they are early.  But are earlier manuscripts necessarily better?  The book offers reasons that earlier could be better, on page 84:

"The preferred variant or reading normally is the one found in the earliest manuscripts.  Less time has elapsed between those manuscripts and the originals and fewer intermediary copies have introduced errors.  The more direct pipeline a manuscript has to the original, the better are its chances of getting the wording right."

But the book does not treat "earlier is better" (my words) as an absolute.  For one, a later scribe who is more meticulous may produce a text that's more reliable than one from an earlier scribe who is not particularly meticulous.  Moreover, a later manuscript can have old material.  For example, on page 78, the book says that Vaticanus is similar to a papyrus from a century earlier.  It was argued that Vaticanus copied from the papyrus, but now it's believed that they both used a common ancient source, for "the wording in Vaticanus is certainly more primitive than that of P75 in several places."

2.  Another external criterion that the book discusses is "genealogical solidarity".  A point that I get out of this discussion is that a particular version is more reliable than a manuscript that has a mixture of versions.  For example, the Alexandrian and the Western text-types came from the second century, and church fathers quoted them.  But the Byzantine text-type came later, as one can see from how it draws from Alexandrian and Western text-types.  Consequently, I'm gathering, the book considers the Alexandrian and Western text-types to be more reliable.  Moreover, in trying to determine which Alexandrian manuscript is better, one should prefer the text that does not contain Western and Byzantine readings.  When Alexandrian manuscripts agree on a reading, "then scholars can be relatively assured that the Alexandrian local original had that reading" (page 86).  The Alexandrian manuscripts had to get their common reading from somewhere, after all.

This makes sense to me, but I have to confess that I have not extensively read the KJV-Only side of the debate, which is rather critical of the Alexandrian manuscripts.

3.  The third external criterion is "geographical distribution".  The idea here is that, if a reading is widely distributed geographically, then the reading is probably from an earlier source.  On page 88, the book says: "Collusion of witnesses is much less probable when these witnesses are distributed in Rome and Alexandria and Caesarea than when they are all in Jerusalem or Antioch."

I have some difficulty understanding this criterion, to tell you the truth.  First of all, what is the "Collusion of witnesses" that the book thinks would be so bad (if it existed)?  Second, couldn't a reading have emerged in one place and then spread out to other places?

(UPDATE: Bart Ehrman in Lost Christianities says that the rationale behind the geographical distribution criterion is that, if a reading is found in a variety of places, there's less of a likelihood that it is merely a local variant, and local variants are not particularly reliable.) 

Despite my cloudiness on this criterion, it does make a degree of sense to me: if texts in vastly different places agree on a reading, then they are probably getting their reading from an earlier source.
The book also discusses internal criteria----looking at the content of the manuscripts themselves to determine which reading is earlier and which is later.  The book also discusses an example in which internal criteria conflicted with external criteria!  But I won't get into that in this post.  But feel free to read the book if you are interested!