A system of invisible punctuation in papyri and medieval Greek codices

How faithfully do medieval Greek manuscripts reproduce their ancient ancestors?  It’s a question that all of us ask ourselves, from time to time, and it can be hard to answer other than subjectively.  In some cases, however, we can compare ancient papyrus copies with much later medieval versions.  The accuracy can be uncanny.

Origen’s Contra Celsum is known to us from Vaticanus gr. 386 (=A), of the 13th century, plus some extracts in the Philocalia, an anthology of Origen’s thought by Basil of Caesarea and Gregory Nazianzen (=P).  A century ago there was a bitter argument among philologists as to which preserved the text better.  The GCS editor, Kotschau, believed that A was to be preferred, while his critics preferred P.  Which was right?

In 1941 a bunch of papyrus codices were discovered by Egyptian workmen in a gallery in the ancient quarries of Tura.  At the time thoughts were on Rommel and the Afrikakorps, and the workmen stole the lot, broke them up and sold them to dealers.  Among them was a papyrus codex of long extracts from books 1 and 2 of Contra Celsum, made by a learned monk who clearly had before him a complete text.  In the papyrus these are followed by extracts from Origen’s Commentary on Romans, and his Homilies on 1 Samuel.  The papyrus can be dated by paleography to the early 7th century.

This meant that the texts could now be compared with an ancient copy of the text.  Quickly it became clear that the papyrus was from a related but not identical family to A.

Now I would like to share with you a passage in the truly excellent volume by Jean Scherer which published the text of the Contra Celsum extracts.1  I will add a comment or two at the end.  As we pick up the discussion, Scherer is talking about the presence of mysterious blanks or gaps between letters in the otherwise continuously written text.  Note that the papyrus has no word division.  

Clearer still: on pages 30-34 and 56-59, the copyist reproduces in full some long passages of Contra Celsum without selection or omission: however, there are many blanks.

These remarks may appear futile, and we ourselves have been inclined to impute these variations to the whim of the copyist, until the day when we examined the Vaticanus gr. 386, which as we said earlier (p.6) belongs to the same family as our papyrus.  Here — a detail which P. Koetschau signalled in a rapid note in his description — the “blanks”, longer or shorter, are an important element of a system of punctuation in use in this manuscript.  They mark the articulations of the thought, separating and distinguishing the different steps in the argumentation.  Short gaps play a role analogous to that which is observed in the Dialektos.  And if one compares, from this point of view, the manuscript and the papyrus, we can say that, if the manuscript has blanks sometimes which do not appear in the papyrus, nevertheless all the blanks in the papyrus which do not mark an interruption are found in the manuscript. 

Sometimes the correspondance is so perfect as to be uncanny.  Thus in p.33, l.20, before μεμνημαι δε (which introduces a new development) the blank, in the papyrus, is extra long.  In the Vaticanus ms. it is is also extra long. 

Such coincidences cannot be accidental.  They show that, in both the Cairo papyrus and the Vatican manuscript, the use of blanks is not down to the initiative of the copyists.  These have done no more than follow their model here, or, better, beyond their model, the archetype, and beyond that, the editio princeps of the library of Caesarea.2 

This peculiarity is transmitted intact down to the 13th century.  But it was fragile all the same: the second copyist 3 of the Vaticanus did not retain it, and in the ms. Parisinus suppl. gr. 616, which is a careful copy of the Vaticanus gr. 386, the blanks have disappeared.

Thus the variations of the papyrus explain themselves quite naturally.  To separate the extracts, the copyist on his own initiative started by using a double oblique stroke //.  But, under the influence of his model, he gradually started using the blanks, which in the complete text before him had the purpose of separating the parts of the discourse.  And finally he used only blanks, for economy of effort, because it was easier to copy the text mechanically than to substitute systematically one sign for another.

Finally let us note that in the extracts of the Commentary on the letter to the Romans we find numerous blanks and no “//” sign.  This is a clear indication that these extracts were copied, by the same scribe, after the extracts of Contra Celsum.

This is quite something, and also new to me.  I wonder how many editors would have recognised that these apparently random gaps in the text had a meaning, and would have tracked them down into the medieval copy?  Not many, I would guess.

But if this was a normal way to write a 7th century papyrus copy of a literary work, I do wonder what other texts, unrecognised, may have contained it.  It looks like a fingerprint feature to me — a way to detect relationships between manuscripts and papyri.  If so, perhaps editors and those working with papyri should be on the lookout for it.

1. Jean Scherer, Extraits des Livres I et II du Contre Celse d’Origène, d’après le papyrus no. 88747 du Musée du Caire, IFAO 18, Cairo, 1956.  See p.12-13.
2. A note in the papyrus indicates that the text was revised by Pamphilus at Caesarea.
3.  The Vaticanus gr. 386 was written by two copyists, who took turns.  The first was both  more elegant and more accurate than the second.

Leave a Reply