Monday, December 01, 2008

SBL Boston, Cosaert on Clement's Text

My report from the SBL in Boston continues. The following is a summary from the first paper of the New Testament Textual Criticism session on Saturday afternoon. This was the session in which I presented a paper with Jennifer Wright Knust, which probably caused some distraction on my part.

Carl P. Cosaert, “The Text of the Gospels in the Clement of Alexandria”

Since Ehrman’s dissertation on Didymus the Blind, there has been a steady flow of studies on church father’s text especially those located in Alexandria (Origen, Athanasius, Didymus, Cyril). Now the turn has come to Clement of Alexandria. What form of the text existed in Alexandria in the 2th century? Did his text influence later fathers? These questions led Cosaert to research the subject. In this paper he summarizes his study of the text of Clement.

The study is based on testpassages in Matt, Luke, and John. The methodology was developed by Ehrman (including Comprehensive Group Profile Analysis). The first step, however, involves the identification of citation and allusions. These are then collated against control witnesses representing various textual strands, and also including other Alexandrian fathers.

The first table displayed data based on 118 significant variations in Matthew. At the top, closest to Clement were Athanasius and Origen, then Sinaiticus. After that came a puzzling combination of witnesses. It has been suggested that Clement’s text represents an “early stage of Alexandrian text.” The conclusion in Matthew, however, is that Clement has no clearcut agreement with any one group. It is split between the Alexandrian and Byzantine groups. Cosaert had hoped that the group profile analysis would clear things up, but it did not. Clement does not preserve any distinct readings for a goup. The results have been more definitive in Didymus and Athanasius.

In Luke, there were 143 units of variation (more data). Here Origen comes at the top, then D, 33 and Didymus. Cosaert pointed out that Clement’s text were closer to the Western text in Luke. (Clement had 58% agreements with D i Matt and 66,4% in Luke). [TW: Note that the difference is ca. 10 readings.]

The text of Clemens in John has much clearer affiliation. Alexandrian witnesses dominate the top of the list (7 of the first 10 are Alexandrian). According to Cosaert, it is not surprising that Didymus is at the bottom of the list, since Didymus according to Ehrman shifts textual character in John. Hence, Alexandrian is probably the classfication to apply in John.

But in Matthew and Luke the text is mixed (with Byzantine influence in Matthew, Western in Luke). The results are different from other Alexandrian fathers.

Cosaert suggested two possibie explanations:
1. Clement’s textual relationships may not be so clear because the readings are not distinct or exclusive for any one group (shared by more text types). [This is a good reason!] Then he looked at variants. For example in Matt 12:36 it was apparent that the testpassage was not so useful for classification purposes, since the tradition is split. Same in other examples in Luke 10:21 and 10:42. There is some need of adjustment.

2. There were two major text types, Alexandrian and Western, without any one dominating. [TW: Or, Clement used different sources.] There was no clear situation in Alexandria. The text is more mixed (cf. also the Coptic tradition).

Cosaert also looked at how Clement used the text. For example, whenever he is using Paul he seems to clearly follow a manuscript. Furthermore, he often he attempts to harmonize.

Comments and questions:
Hurtado: The suggestion that there are different transmissional dynamics is compatible with analysis of papyri too. Some are diificult to classify, and other like P75 align with later witnesses (continuity).

Jongkind: What do we mean with a “fluid text”. Does not your conclusion undermine your own analysis (there was no clear text types at this time, they develop later, why include them?).

[I did not record a reply to this question, if there was one.]

Read more about Cosaert’s study as reported previously on this blog here.


  1. If I remember correctly, the answer to the last question went something along the lines of this:

    Cosaert knew about the difficulty with text types being anachronistic, but decided at the start of his study to use them as this would then enable him to compare the evidence against work already done by other scholars. No use trying to define new categories all the time; rather describe from what we've got.

    Then again, this might only be the way I remember it. - Similar to what people would "remember" from a church sermon.

  2. My apologies for writing off topic. I am looking for Marguerite Harl's Grinfield Lectures on the Septuagint on the Greek version of the Song of Songs (Oxford, 1989). If anyone has access to them and is willing to let me look at them (or knows how I might get access to them), please contact me at Thank you for your time and assistance. All the best, Jonathan Kaplan

  3. When do our manuscripts of Clement date from? If they're late enough, one might expect copyists to have made some harmonizations to the Byzantine (even subconsciously).

  4. Is this what is meant by "aligning with later witnesses?"

    An analogy would be to find a manuscript copy of the United States Constitution with the appropriate deletions and additions generated by the forthcoming 27 amendments.