Rare Text in a Manuscript from Pontigny, ca. 1170, Copied from a Sixth-Century Exemplar


Genre: ,

PRICE: $6500


Didymus the Blind, De spiritu sancto in the Latin Translation of St. Jerome. Single folio on vellum. From the Cistercian abbey of Pontigny, ca. 1170: 462 mm x 337 mm (justification, 325 mm x 230 mm). Written above top line in a confident Gothic minuscule. Prickings survive on the inner, outer and upper margins. Double column, 50 lines. Greek terms have been transliterated and lines have been added above them. Punctus flexus abbreviation used throughout. A few errors in the text have been expunged or corrected in coeval hands. At least one of these corrections seems related to an inability to construe the exemplar.


Rare nowadays is any opportunity to own a manuscript relic from Romanesque Pontigny—or from any of the four great Cistercian houses settled from Cîteaux—at the foundational moment of Cistercian monasticism. Pontigny was established in 1114 by a retinue of twelve monks from Cîteaux, which itself had only been founded sixteen years before (1098). Its reputation grew quickly, and the abbey came to launch some twenty-two daughter houses. From 1168 to 1170 Pontigny hosted Thomas Becket, the exiled archbishop of Canterbury, and Becket’s exile conceivably overlapped with the production of our manuscript. Becket was an obsessive bibliophile whose library was catalogued at Christ Church around 1330 (Montague Rhodes James, The Ancient Libraries of Canterbury and Dover (Cambridge, 1903), pp. 82-85 [nos. 783-853]). Prior Eastry’s inventory reveals a copy of Didymus at Christ Church, but none in Becket’s possession (ibid. 18, no. 32). Perhaps our manuscript had yet to be transcribed when Becket left Pontigny. Other notable men of learning might have encountered it at Pontigny, however. Stephen Langton, a theologian, bible scholar and future archbishop of Canterbury, lectured at the University of Paris. He followed in Becket’s footsteps to Pontigny after being banished by King John, residing there from late 1209 to the summer of 1213. In this instance, it seems entirely feasible that a figure like Stephen found the parent manuscript of our leaf on the library shelves at Pontigny.

The source-text of our manuscript seems to have surfaced in Burgundy in a Merovingian or Carolingian manuscript that descends from a sixth-century Italian Uncial version written in scriptura continua. It has lately been theorized on the basis of stemmatic analysis that the ancient hyparchetype of this exemplar was the key witness in the survival of the text to modern times. Copies of it drifted into monastic libraries across Europe. The textual evidence provided by our manuscript documents this precise international link between a late antique copy in Italy—possibly kept at Vivarium—and one just emerging in France.

Our Didymus text and Florus of Lyon seem like unusual monastic texts, but they had content that made them ideal for monastic study in the twelfth century. Cistercianism was a growing movement in need of books (see Anne Bondéelle-Souchier and André Vernet, Bibliothèques Cisterciennes dans la France Médiévale (Paris, 1991), pp. 260-62). Naturally, the bible was a central to monastic reading, which entailed long hours of rumination on short excerpts that called related passages to mind. The key theological adjuncts to such study were the church “Fathers,” including Jerome, Augustine, Gregory and Ambrose. Isidore of Seville and the Venerable Bede are sometimes also included. The Florus of Lyon text, which was often attributed to Bede, is little more than a florilegium or selection of passages from the works of St. Augustine focusing on the Pauline Epistles. By any measure it qualifies as mainstream work of high monasticism, since it functions as collation of Patristic reflections on Scripture antecedent to the Glossa Ordinaria. Because it circulated in the Latin translation of St. Jerome, De spiritu sancto by Didymus boasted an authoritative sanction. If Jerome thought it worthy of translation, then a Pontigny monk had every reason to read it closely. Yet it hardly needed Jerome’s endorsement, for De spiritu sancto served as a key source for theology on the subject of the Holy Spirit and was, on its own terms, a major statement plagiarized by St. Ambrose.


There is a good modern edition by Louis Doutreleau (Didyme L’Aveugle: Traité du Saint-Esprit, Sources Chrétiennes 386 [Paris, 1992]), in which our text appears on p. 296 (§166.1) to p. 318 (§195,2). A more widely available edition is printed in the Patrologia Latina, vol. 23 (works of Jerome). In the PL edition our fragment comprises §37 (PL 23: 135A) propterea autem ista … §42 (PL 23:139A) eius mortificaverit. A third edition referenced below appears in Migne’s Patrologia Graeca, vol. 39.


Thirty fragments of this manuscript are known, almost all of them complete leaves. Ten currently belong to the cathedral of Auxerre (see below). As for the other twenty, Peter Kidd has reconstructed the engrossing story of their dispersal over a period of years (The McCarthy Collection, vol. III: French Miniatures (London, 2021), pp. 39-43 [no. 9]). In 1995 Les Enluminures offered a group of nineteen complete folios and one partial leaf from a commentary on the Pauline Epistles by Florus of Lyons and the De spiritu sancto by Didymus (A Selection of Manuscripts Through the Ages, 1995, pp. 35-40). It is not precisely known how many constituents belonged to each text, but Kidd’s table of extant leaves suggests that they were more or less equally divided. Two leaves in this group bore four large initials: a leaf now in the McCarthy collection had three, while a second had but one. Kidd further observed that Les Enluminures apparently conveyed its leaves in 1997 to Bruce Ferrini, who then retailed the most decorative one to Bob McCarthy. The remaining nineteen pieces then passed to Charles E. Puckett, an associate of Ferrini’s and a long-time dealer in Akron, OH. He dispersed all but one of them, keeping the leaf of II Corinthian-Galatians with the fourth decorative initial.

In addition to the McCarthy leaf, Kidd has documented eleven others, tracing eight of them to seven current owners (McCarthy Collection 39). Only one of these is institutional. Our folio is the twelfth leaf, to which Kidd drew attention on his blogpost (posted 5 June 2021:, and number thirteen belongs to Puckett. This chain of ownership from Les Enluminures to Ferrini to Puckett explains why all but one of the known leaves reside in American collections. The twentieth partial folio in the Les Enluminures catalogue seems to have been cut into small strips, each sold individually. Two are known, and the one I have seen has six lines of text cut from a single column. In consideration of this evidence, seven complete fragments remain unaccounted for.

In vol. 3 of his McCarthy catalogue, Kidd identified ten sister leaves of our manuscript at Auxerre (Bibliothèque de la Cathedrale Saint-Étienne, Trésor de la Cathédrale ms 3). These had been studied in 1987 and again in 2001, primarily to deduce an origin for the characteristic Cistercian decoration (M. Peyrafort, “Un Nouveau Fragment du Commentaire Augustinien de Florus de Lyon sur les Épîtres de Saint Paul,” Revue des Études Augustiniennes 33 (1987), 132-36 (chiefly on the Pontigny provenance); M. Peyrafort-Huin and P. Stirnemann, La Bibliothèque Médiévale de l’Abbaye de Pontigny, XIIe-XIXe Siècles: Histoire, Inventaires Anciens, Manuscrits [Paris, 2001]). In 1987 Monique Peyrafort proposed an origin at Pontigny, which she and Patricia Stirnemann later affirmed—although an alternative origin at the Premonstratensian abbey of Saint-Marien at Auxerre was also suggested. The twelfth-century Pontigny library catalogue mentions a two-volume compilation of works by Florus followed in volume two by De spiritu sancto of Didymus (Peyrafort-Huin and Stirnemann, Pontigny 251-52). Two decades later Kidd observed that the McCarthy leaf overlaps the Florus and Didymus components and proves beyond doubt that the thirty extant folios of this manuscript belonged to Pontigny. Since all the known leaves come exclusively from the second volume of the two-volume set, it has naturally been assumed that the first volume has been lost and that only these thirty leaves remain.

Some significant deductions have been made about the transmissional history of our manuscript on the basis of extant copies. As Kidd stated in 1968, Louis Doutreleau noticed that the text of De spiritu sancto in BnF ms lat. 2364 had been copied out of sequence (“Incohérence textuelle du De spiritu sancto de Didyme dans le Parisinus, lat. 2364,” Sacris Erudiri 18 (1968), 372-84). The manuscript dates to the same period as our copy and was prepared at Fontenay, a Cistercian abbey proximate to Pontigny. Doutreleau deduced that, because the dislocations and lacunae unfailingly cover about 33 lines or multiples thereof in Migne’s edition (as printed in PG 39), the errors must have arisen through a misbound exemplar with 16 lines of text per page. As Doutreleau remarked, this is a very small number of lines per page and a significant clue to the sixth-century Uncial format of the exemplar.

Now, Kidd further noticed that the leaf of our Didymus at the University of Louisville boasts similarly discoherent text, and he provided a roadmap of the dislocations (p. 43, note 10) as well as a URL to the digitized Louisville fragment. A collation of it demonstrates the that the discontinuity in the Louisville fragment corresponds exactly to sequence VI in the Fontenay copy (Doutreleau, “Incohérence textuelle” 374). In other words, BnF ms lat. 2364 and our manuscript both descend from the same disordered exemplar, either one from the other or both from the same source.

Having observed the strange coincidence that the text is disordered in segments of about 33 lines, Doutreleau proved that this corruption arose from disarranged quires. His evidence proves beyond doubt that the exemplar had two columns not exceeding 16 lines each (including contractions and abbreviations), with 320 to 340 letters per page and dimensions of 190 mm x 180 mm. The best parallels for such a format are Uncial copies of the fourth, fifth and sixth centuries. Five are identified in E. A. Lowe’s Codices Antiquiores, all of them Italian. One of these is dated fourth to fifth century, while a second is dated fifth to sixth. Furthermore, Doutreleau noticed that the scriptura continua of the Uncial writing along with a failure to recognize the nasal suspension abbreviations (of words terminating in -m) explain significant textual revisions.

Our Pontigny codex and its Fontenay congener are exceptionally important in detecting the early transmission of the text—i.e., in proving that it pre-dates 600 AD. In his 1992 edition Doutreleau reasoned that the γ hyparchetype gave rise to all the important witnesses of this recension, which includes ms lat. 2364 and five other copies, as follows:

  1. BnF ms lat. 1688. Moissac.
  2. BnF ms lat. 1689. Tulle.
  3. Leiden, MS Scaliger 2.
  4. Durham, MS B III 2 (now B IV 16). Durham.
  5. Cambridge, Gonville and Caius MS 131/71.

Of the sixty manuscripts of the De spiritu sancto, only those belonging to hyparchetypes α (7 manuscripts) and β (ten manuscripts) did not stem from γ. In fact, it was stated that the γ hyparchetype was equally important as a source as α and β, all three of them deriving from a single version. But where did γ arise? This reconstructed dissemination does not mean that the great Cistercian abbeys of Burgundy had access to an Uncial manuscript. Doutreleau himself spoke of a Merovingian copy that preserves these peculiar features. However, one possible route of dissemination flows through Cluny. A tenth-century copy of De spiritu sancto from Cluny (BnF ms nouv. acq. 1460) belongs to the α family (with twelve lacunae). Cluny, we know, provisioned its library with manuscripts from Rome, so it makes sense that the abbey could have acquired a very old copy of Didymus. As noted above, furthermore, Cîteaux was settled from Cluny. It seems entirely plausible that the Fontenay and Pontigny manuscripts derived from a manuscript at Cluny, and Doutreleau specifically stated that it is just conceivable that nouv. acq. 1460 is descended from a copy antecedent to the disordered version represented in the parent codex of our leaf.


The upper margin, but especially the upper corner of the inner margin bears damp-staining on both recto and verso. This soiling does not affect the text, which is completely legible.

Back to All Manuscripts

Back to Top