Thu Aug 22 09:23:36 1996

From owner-tc-list  Thu Aug 22 09:23:36 1996
Return-Path: 
Received: by scholar.cc.emory.edu (SMI-8.6/SMI-SVR4)
	id JAA00147; Thu, 22 Aug 1996 09:22:36 -0400
X-Sender: waltzmn@popmail.skypoint.com
Message-Id: 
In-Reply-To: 
 
References: 
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 22 Aug 1996 08:07:14 -0700
To: tc-list@scholar.cc.emory.edu
From: "Robert B. Waltz" 
Subject: Re: Colwell's 70%
Sender: owner-tc-list@scholar.cc.emory.edu
Precedence: bulk
Reply-To: tc-list@scholar.cc.emory.edu
content-length: 3137

I suppose I need to clarify something here....

On  Wed, 21 Aug 1996, Maurice Robinson  wrote:

>On Wed, 21 Aug 1996, Robert B. Waltz wrote:
>
>> Frankly, I consider the Colwell criterion to be all but worthless.
>> The relationship between manuscripts should not be pursued based
>> on their *percentage* of agreement (unless one is seeking sister
>> manuscripts). After all, *all* manuscripts actually agree on about
>> 95% of the text. Rather, one needs to look at the *nature* of the
>> agreements.
>
>> My humble opinion, of course. Not that I'll brook any argument. :-)
>
>My humble opinion differs as usual (the agreement between us did not last
>long *;-), since I think the criterion has validity, even though the
>precise percentage point of 70% may be subject to some adjustment up or
>down.
>
>Certainly all MSS agree on about 95% of the text in most cases, but the
>establishment of the Colwell-style percentages reflect a 70% agreement as
>found within the 5% or so of the text where the variants actually occur,
>and this is plainly stated in Colwell's "Methodology" article.

True enough; my point is merely that very few scholars (except for
Hurtado in his work on Mark and Richards in his sadly flawed work
on the Johannine Epistles) have tried to study *all* readings. So
the cutoff percentage will depend -- significantly -- on the
particular sample.

>I do not think that the "nature" of the agreements is the significant
>item; rather the fact that the agreements occur in a definite and regular
>pattern -- this pattern-agreement is precisely what establishes a
>"texttype", even if MSS with a definite pattern may find that some
>portions of that pattern otherwise appear in MSS of another texttype.

I am willing to agree with Maurice Robinson that, if we choose the
number carefully, based on the sample, we can find *some* percentage
that defines text-types -- as long as we are working with unmixed
or minimally mixed manuscripts.

That restriction ("unmixed or minimally mixed manuscripts") is, however,
very important. Other than the Byzantine manuscripts, almost all texts
are mixed. Taking Paul as an example, of the roughly 600 manuscripts,

* About 520, or 87%, are Byzantine
* About 70, or 12%, are mixed.
* As best I can tell, exactly ten (p46 Aleph A B C D F G 33 1739) are
  non-Byzantine and unmixed or minimally mixed manuscripts. That's 1.7%.
  (Before someone asks -- yes, C is mixed in the Gospels, but in Paul
  it's almost purely Alexandrian. 33 is Byzantine in Romans, where it
  comes from a second hand, but in the rest of Paul, it is the closest
  relative of Aleph.)

So, of the non-Byzantine manuscripts, seven-eighths are mixed (usually
mixed Byzantine and Alexandrian or mixed Byzantine and family 1739).
Such a mixed manuscript may not agree with *either* the Alexandrian
or the Byzantine text 70% of the time. But they have a text-type --
indeed, they are blessed enough to have two. :-)

That's why I say one must examine the *nature* of agreements -- in order
to figure out what the components of the mixture are.

Bob Waltz
waltzmn@skypoint.com

Back