An example of how older editions mislead us about patristic citations

Earlier this week, I was speaking to Dirk Jongkind about Luke 1:50. The THGNT adopts the phrase “εἰς γενεὰς καὶ γενεάς“ as the viable reading that cannot be attributed to harmonisation. Some variants include a form of the phrase that uses singulars instead of plurals, “εἰς γενεὰν καὶ γενεάν“ (ℵ01 Ψ044 f¹ f¹³ [including 69] 892 1424) as well as the reading in the majority of manuscripts, “εἰς γενεὰς γενεῶν“ (A[^{nt; the odes have a different reading}] C² D¹ K Δ^c Θ 33 Byz).

One point of reference that Dirk mentioned is that the there might be patristic citations of the phrase that show that the phrase itself, “εἰς γενεὰς καὶ γενεάς“ is not inherently unusual. Sure enough, the TLG turns up two instances that I want to point out here.

The first instance is Origen’s commentary on Psalm 119(118):90. Origen makes an exegetical point that the text is singular, not plural: the two generations are Israel and the Church, so it is specifically “from generation to generation”.

Origen at the TLG

In the second instance, Chrysostom quotes part of Daniel 4:37 (LXX) in one of his Letters to Olympia (10.9), and he uses the exact same phrase:

Chrysostom in the TLG

So, here we have two patristic sources attesting the phrase that is adopted in the THGNT at Luke 1:50. Admittedly, neither of these phrases is a citation of Luke 1:50, but the question we are considering is whether the phrase itself is particularly unusual, and if so, would that mean that it could be subject to variation that would normalise the phrase into something more common. From the look of it, the phrase “εἰς γενεὰς καὶ γενεάς“ is common enough that Origen and Chrysostom use it, which would be a mark against the rationale that the phrase would be normalised and changed away from “εἰς γενεὰς καὶ γενεάς“ to either “εἰς γενεὰν καὶ γενεάν“ or “εἰς γενεὰς γενεῶν“, which are both more commonly attested.

Wait, not so fast.

Trying to do the responsible thing, I went to the CUL to pull the critical editions. Both are in the Sources chrétiennes series (henceforth, SC). The citation of Origen comes from a catena on Psalm 119(118). The edition is Harl, Marguerite, ed. La chaîne palestinienne sur le Psaume 118 (Origène, Eusèbe, Didyme, Apollinaire, Athanase, Théodoret): Tome I. Introduction, texte grec critique et traduction. Sources chrétiennes 189. Paris: Cerf, 1972. When I turned to the relevant pages, I did not find “εἰς γενεὰς καὶ γενεάς“ as the TLG reported for the text here. Instead, the author of the commentary attributed to Origen roughly quotes Psalm 119:90 in the form of the text attested by the Byzantine tradition at Luke 1:50 (there is an added οὖν, but otherwise it is the same):

Origen in the SC edition

Similarly, the TLG also disagrees with the edition of Chrysostom: Malingrey, Anne-Marie, ed. Jean Chrysostome: Lettres à Olympias. Sources chrétiennes 13. Paris: Cerf, 1947. Instead of the plural form of the phrase “εἰς γενεὰς καὶ γενεάς“ that the TLG says is Chrysostom’s text, we find the singular form “εἰς γενεὰν καὶ γενεάν“:

Chrysostom in the SC edition

But wait, there’s more!

This edition was primarily a revision of an older edition rather than being a genuinely new edition, but there is a second edition of this work in the SC series for that. This second edition of Chrysostom, which is the source of the TLG text, does in fact have “εἰς γενεὰς καὶ γενεάς“, and notes two manuscripts (a: Vatican Library, Palatinus gr. 228, and M: Paris, BnF, gr. 657) that have “εἰς γενεὰν καὶ γενεάν“. Malingrey does note that previous editions used these manuscripts for Chrysostom’s text (“a” in Savile, 1612, and “M” in Fronton du Duc, 1614 and Montfaucon, 1821—this is assuming my memory of the relevant page is correct; the Classics library wouldn’t let me borrow the book, so I had to read and take photos of the text itself with my phone):

Chrysostom in SE 2nd edition

_____

What then is the point of this rather pedantic post? There are two points.

1. Pedantic details matter, because [2].

2. When we rely on older editions, we can be misled about the form of the text that a church father quotes. There even appears to be an error in the first edition of the SC edition for Chrysostom that was carried over from previous editions, so sometimes even critical editions don’t quite get it right. And they differ at a phrase for which there are textual variants when it occurs in Luke 1:50. More than that, Origen (or the author of the comment attributed to him) specifically makes an exegetical point about the precise wording of the text, and even there, the SC edition and the TLG disagree on what that precise wording is. TLG’s text of Origen here is based on J.B. Pitra, Analecta sacra spicilegio Solesmensi parata, vols. 2 and 3 (Paris: Tusculum, 1882). This is also an example why it is appropriate to take older literature with a grain of salt when it claims that certain patristic authors supported one reading against another. They might not have.

For more on why we might have problems citing Chrysostom for particular readings, see Peter Montoro’s guest post, here.

Montoro: The Instability of Chrysostom’s Romans Text

The following is a guest post from Peter Montoro who is currently a doctoral student at the University of Birmingham. It gives a sample of his important paper at the Birmingham Colloquium on the Textual Criticism of the New Testament earlier this year, which was, in turn, based on his BD dissertation for the University of London. An expanded version of this material will be published in the proceedings of the Colloquium. —Peter G.

---

The particular value of patristic citations, as opposed to versions and continuous text manuscripts of the New Testament, is, as Gordon Fee put it, “[W]hen properly evaluated...the Church Fathers have the potential of offering datable and geographically certain evidence.”¹

As has now been well established, recovering usable citations from patristic witnesses poses a number of methodological challenges. Yet while much progress has been made on many fronts, it seems sometimes to be forgotten that the task of “proper evaluation” is incomplete without a careful investigation of the manuscript transmission of the work in which a given patristic citation is located.

Saba 20, Folio 145r

It needs to be more clearly recognized, in practice as well as in theory, that the usability of patristic citations is directly dependent upon their stability within the manuscript tradition of the work from which they derive.

Though a father may have used two different forms of a text at two different times, even in the same work, it is rather difficult to use two different forms at the same time. When a single citation in a single location is itself an instance of variation, that citation is not “stable,” in the sense that I am using the term, and this therefore precludes it—until and unless the priority of one of these textual forms has been clearly and decisively demonstrated—from being utilized as the “datable and geographically certain evidence” that patristic citations are intended to provide.

For this study, I examined those portions of Chrysostom’s Romans Homilies that contain his exegesis of Romans 8 (found in homilies 13–15 using Migne’s numbering), in seven of the eight extant and catalogued manuscripts of the work that date to the tenth century or earlier, as well as two additional eleventh century manuscripts and six editions of the printed Greek text. (The published version of my work will include full data from all extant manuscripts dating to the fifteenth century and earlier.)

Chrysostom’s citation of Rom 8:33–35 provides a clear and striking illustration of the results of the analysis of this dataset as a whole and is what I want to focus on in this post. A PDF containing brief info on each of the manuscripts and editions used, as well as a full collation of the initial citations of 8:33–35 in these manuscripts and editions is available here.

Within the eight manuscripts examined, there are no less than five significant variation units in which two or more manuscripts differ from the others. All five of these non-singular variants are also found in the manuscript tradition of the book of Romans itself. In other words, the variants in the manuscript tradition of the Romans Homilies largely offer a selection of the very same variants that are present in the manuscript tradition of Romans itself. This is the case, not only for individual variants, but also for the particular combination of variants found in each of the two manuscript clusters found in the early manuscripts of the Romans Homilies.

The fifth of these variants, found in 8:35, is particularly interesting. After ἀπὸ τῆς ἀγάπης τοῦ, one cluster of manuscripts reads θεοῦ, while the other cluster reads χριστοῦ, followed by the significant comment, Καὶ οὐκ εἶπε· τοῦ θεοῦ· οὕτως ἀδιάφορον αὐτῷ, καὶ χριστὸν καὶ θεὸν ὀνομάζειν (“And he does not say ‘of God,’ so indifferent is it to him whether he mentions the Name of Christ or of God”).² Phrases such as the Καὶ οὐκ εἶπε(ν)... found in all printed editions at this point, have traditionally been considered the “gold standard” for textual stability in patristic citation. Indeed, Tischendorf thought this phrase so significant that he included the entirety of it in his textual apparatus for this verse.

Yet upon examination of the manuscript tradition of the Romans Homilies, not only is this phrase absent in a significant number of early manuscripts, but also the reading which it is clearly intended to support is itself a point of variation, rendering this traditional “gold standard” for textual stability a rather shaky foundation on which to build one’s text-critical house.

In order to properly evaluate patristic citations, it is not enough to determine from a printed edition that a citation is actually a citation—one must also go behind the edition to consider the stability of the manuscript tradition that underlies it. In the case of Chrysostom’s Romans Homilies, an initial survey of the manuscript tradition demonstrates significant textual instability, raising serious questions about the usability of any of the currently available printed editions of this work as a basis for citations of the form of the text used by Chrysostom in the late fourth century. Given that Chrysostom’s Homilies on Acts—for which similar textual instability has been demonstrated—served as the source for more than a quarter of the entire apparatus of patristic citations in the recent ECM of Acts, the answers that we give to these questions will have a significant impact on the shape of future critical editions of the New Testament.

It is important to note that I am not by any means asserting that these printed editions never give us access to the text of Chrysostom—indeed, in most cases they probably do. Yet this is to say very little indeed. The text of Romans is itself sufficiently stable that the same claim could be made, with equal validity, for almost any purely Greek (excluding the bilinguals, which present special challenges) manuscript or edition of the text of Romans, however far removed from Chrysostom.

As a point of comparison, the Textus Receptus and the NA28, normally considered to be at opposite ends of the textual spectrum, differ from each other only eighteen times in the entirety of Romans 8, four of which variations are merely orthographical. In this same chapter, there are no less than twenty­-two differences between the Romans text found in the two textual clusters of Chrysostom’s Romans Homilies, a number that only includes those instances where all three manuscripts identified as belonging to the one manuscript cluster differ from all five of the manuscripts that belong to the other.

In other words, there is actually more variation between the various manuscript clusters of Chrysostom’s Romans Homilies than there is between the Textus Receptus and the NA28. Until and unless the priority and originality of one of these clusters has been decisively established, no citation from this work, no matter how apparently secure in the printed text, ought to be considered as providing “datable and geographically certain evidence.”

Notes

1. Gordon D. Fee and Roderic L. Mullen, “The Use of the Greek Fathers for New Testament Textual Criticism,” in The Text of the New Testament in Contemporary Research: Essays on the Status Quaestionis, ed. by Bart D. Ehrman and Michael W. Holmes (Leiden: Brill, 2013), 351–352. ↩
2. The translation is taken from the Nicene and Post-Nicene Fathers edition of the Romans Homilies. ↩

British Library, Matthew 23 and Chrysostom

Though I don't think that there is much biblical stuff in the latest batch of Greek mss put online by the British Library, there is always something interesting, and this time it is a citation from the Gospel of Matthew. You can find the BL blog post here from which I copied below the image of Add MS 24371, sermons by Chrysostom.



At the top of the page Mat 23:1-3 is cited and these words are marked, as usual, with diples '>>' in the left margin.

There is a nice textual variant here from line 6 on: πάντα οὖν ὅσα ἄν λέγωσιν ὑμῶν ποιεῖν, ποιεῖτε. (All that they tell you to do, do).

The words ποιειν ποιειτε 'to do, do' are rather rare, I found only two manuscripts in Legg's Matthew volume that have these words, Γ(036) (here) and minuscule 544 (here). Now if I had time I'd love to pursue this and trace the relation between Γ(036), 544 and Chrysostom's text further, but alas ...