Concerning Three Manuscripts of the Divine Comedy in Private Collections in New York and Milan, Part 1 of 3

Elisabetta Tonello (E Campus University)

Dante Notes / October 22, 2016

For the past fifteen years a group of researchers in Ferrara (Italy) has been working to produce a critical edition of Dante’s Commedia. I have been a member of this group, directed by Paolo Trovato, for the past six years.[1] Recently we accomplished the first phase of a neo-Lachmannian process of critical edition: the recensio and collatio of all the MSS witnesses.[2] We have examined and collationated, on a canon of about 600 loci critici, the 600 manuscripts which are held in libraries around the world; the number rises to 800, if fragments are included. For the great volume of evidence that constitutes the tradition of this work, the Commedia represents one of the most exciting challenges in the field of philology. In addition to being applied to a model case of “overabundant tradition” within romance literature, the classification and creation of a genealogy for its codices (stemma codicum) is further complicated by the contamination which afflicts most of the manuscripts.

Neo-Lachmannian procedure sets as first step in the costitutio textus the examination of all existing codices. A few months ago, when we thought that the first phase of work could be considered completed, we happened upon and entered into a world quite apart from that of academic research: private collecting. The existence of a rather large number of copies of the Commedia held in private hands does not come as a total surprise.[3] Some were already known in the past, and the owners of the manuscripts have given permission to examine and include them in catalogues; these were in any case only a small portion, the traces of which had never been lost. There also exist, however, manuscripts which have never been catalogued, and which at some point were considered to be lost. Three of these witnesses –Kraus 1, Kraus 2 (both in New York) and Vernon (in Milan)– will be the object of this and the following two notes.

Kraus 1

This codex is one of two copies of the Commedia owned by a generous collector, that I was able to inspect in New York.

Dating to the beginning of the 15th century, Kraus 1 MS is among the well-known MSS which was considered lost: it is, in fact, the Phillips 247 MS. It is a vellum manuscript,[4] the binding in red velvet, in common with the Brux MS[5] confirms its having been part of the collection of Charles James Fox, as noted by Sir Thomas Phillips, with the diction «olim celebris C.J. Fox».

Hence the first known owner was Charles James Fox, who owned the manuscript to the first quarter of the XIXth century. Between 1824 and 1837 it was in the possession of Sir Thomas Phillips (one of his earliest acquisitions being included in the series 1-312: 247). Subsequently it was in the Library of Martin Bodmer in Cologny (Genève), where it remained until 1971. Eventually it was sold to H. P. Kraus, and therefore is referred to as Kraus 1 in this paper.

Remarkable features of this manuscript include an illuminated page at the opening of each cantica. The miniatures, formed by leafy and floral elements with birds, surround the initial letters of the canticas. The initial letters of every canto are decorated with red and blue ink. Initial letters of every tercet are touched in yellow and so are the characters and place names appearing in the text. According to Brieger-Meiss-Singleton the miniatures can be considered Tuscan; in particular, the first one could be Sienese. The iconographic project is common to many manuscripts produced by the Tuscan workshops,[6] but the treatment of the subjects is quite remarkable, given the great refinement and the wealth of details. In Inferno Dante stands in front of Virgilio, who is seated at a desk with an open book before him. Beside them there are the three wild beasts (“tre fiere”). In Purgatorio Dante and Virgilio are represented inside a little ship sailing in the wind (“la navicella del mio ingegno”). The third illumination, in Paradiso, represents God in a mandorla, with angels. Marisa Boschi Rotiroti argues that the full page miniature at the beginning of the text could be attributed to Battista di Biagio Sanguigni.[7]

The MS contains the whole Commedia, written in single columns (30 lines), without corrections or notes. Despite what Batines, Moore and Contini have argued, inspection has confirmed that the text was written by two different copyists (in a large and clear littera textualis). The first copyist (Paolo di Duccio Tosi) wrote the “Rubriche di Dante” in red (2r-4v); then a sort of “Prohemio” (5v) and another introduction to the poem (“Prolago”) at 6r. At 10r begins his copy of the Inferno. He stops abruptly at 17v (corresponding to If IV 45). At the end of this sheet there is a lead sign: “Dimmi”. From this point another hand continued copying the Commedia, carrying on to sheet 156v (corresponding to Pg XXVI 87). Here too is given an indication of the next line to copy: “Or sai nostri”. At 157r the first hand returns. It was Paolo di Duccio Tosi, thus, who completed the copy of the poem and added the Credo of Dante and the Capitoli of Bosone and Jacopo. He declares his identity in the colophon of the Paradiso: “Scripto p(er) mano di me Paulo di Duccio tosi / di Pisa negli anni domini MCCCCXII a dì / XI di gen(n)aio. Deo gr(ati)as”. Both copyists were clearly Tuscan but there are no evident dialectal peculiarities.

Paolo di Duccio Tosi was probably a member of a workshop because his name appears in a number of MSS. For instance, in the lost Ricc. 1046, in which there was a colophon very similar to the one of this MS: “Scripto per mano di me Paolo di Duccio Tosi da Pisa negli anni dni MCCCXXVIIII adi VIII di septembre. Deo gratias amen” (In fact the date 1329 must be erroneous; Paolo di Duccio Tosi lived in the XV century, so that one C must be missing: 1429). However, his signature is also in the MS Par. 73 (n. 7255, De Batines 431, dated 1403) and in the Triv. 2263 (n. 4, De Batines 261, dated 1405). We own the collation of both Par. 73 and Triv. 2263, so that we, from a genealogical point of view, can affirm that the first one pertains to a small subgroup related to Boccaccio’s group in If-Pg and is related to the well-known Ash MS in Pd; whereas the second one pertains to the Cento group.[8] It is remarkable that neither Par. 73 nor Triv. 2263 are related to the MS in exam.

Another notation concerns the modus operandi of Paolo di Duccio Tosi da Pisa. In a recent essay Boschi Rotiroti presents the various MSS produced by this copyist (Br. Castiglioni 12, Par. 73 Triv. 2263, Marc. XI 36 and the lost Ricc. 1046), observing that in many of them an alternation of hands occurs. More precisely, both the Par. 73 and the Triv. 2263, containing the text of the Commedia, as well as Marc. XI 36 reveal the presence of two hands. In the latter, the Canzone alla Vergine by Petrarch and the Vita di Dante by Boccaccio were written by Paolo di Duccio, whereas the Epistola consolatoria a Pino de’ Rossi by Boccaccio was copied by another scribe. In Triv. 2263 there is an evident fracture in the copy process. Paolo di Duccio writes through 22v, but by 23r another hand takes over and eventually Paolo di Duccio is writing again at the end of Inferno.[9]

Indeed, the MS Kraus 1 is very close to another manuscript preserved at the National Library in Florence, Fior. II I 43, a link which is particularly evident since both show a marked change of ancestor in the central cantica. Both Kraus 1 and Fior. II I 43 belong to bol family in Purgatorio.[10] Instead, what differentiates them is that, in Inferno and in Paradiso, Fior. II I 43 is quite adherent to the model fi& (despite having a series of vatbocc innovations), while Kraus 1 appears to be closest to vatbocc without the mediation of fi&.[11] This fact confirms a practice of contamination on the part of the copyist, who mixed various ancestors (editio variorum) or worked on a single ancestor belonging to the workshop, which had already been altered by various acts of copying and changing of exemplares. The alternation between the two copyists involved in the copy process is, typically, a further confirmation of the state of contamination but, in this case, represents perhaps an idiosyncrasy of the workshop.

The two MSS share a list of innovations, more or less spread across the largest Florentine families or even, specifically in Paradiso, across the whole of α (a branch of the genealogical tree which contains the most part – and the most complicated – of the tradition of the Commedia).

Innovations by Kraus 1 + Fior. II I 43’ in Inferno Paradiso

Furthermore, in the first cantica the two manuscripts present some typical vatbocc variant readings, but never all of them, nor even those most significant.

Innovations vatbocc in Kraus 1 + Fior. II I 43’ in Inferno

As previously noted, Kraus 1 has only a few typical variant readings of fi& family, while in Fior. II I 43 one can find more meaningful traces of a filiation from fi&.[12]

Innovations fi& (or fi& et alii) in Kraus 1 + Fior. II I 43’ in Inferno Paradiso

As noted above, for the text of Paradiso it is more difficult to place the codex. In the last cantica we are missing significant common errors, at least among the canon of Barbi’s loci,[13] which allow to link this specific manuscript to a precise genealogical area. Only a few innovations can be registered and most of them can be referred to the whole of a instead of to one of these subgroups. The remaining innovations are lectiones singulares.[14] From a certain point onwards (around Pd XXIII) the innovations, even if still widespread, appear slightly more confined. These innovations can be read in the three main Florentine families: cento, parm, vatbocc (which I would collect in a single sigla = cpv). It is noticeable that Fior. II I 43 MS does not share the innovations with Kraus 1 (it presents the vulgata, i.e. common reading or the fi& typical reading), a detail that confirms the generally closer proximity of Kraus 1 to the area which includes vatbocc, if compared with Fior. II I 43.

Innovations cpv in Kraus 1 in Paradiso

The manuscript, as anticipated, radically shifts in its stemmatic position in Purgatorio, where it follows the behaviour of the bol family.

Innovations bol in Kraus 1 in Purgatorio

All data points to the conclusion that this manuscript belongs to the α branch of the tradition, complicated by serious contamination.

Even if the manuscript in question will not be directly useful for a restitutio textus, this case reminds us that every critical edition comes down to a “working hypothesis”, as Contini used to say. Research can never be considered to be concluded, especially in the search for primary evidence materials, that could always be the source of new surprises.


[1] See Nuove prospettive sulla tradizione della Commedia. Una guida filologico linguistica al poema dantesco, edited by Paolo Trovato, Firenze, Cesati, 2007 and Nuove prospettive sulla tradizione della Commedia. Seconda serie. Studi 2008-2013, edited by Elisabetta Tonello, Paolo Trovato, Padova, libreriauniversitaria.it, 2013.

[2] About the neo-Lachmann method see Paolo Trovato, Everything you always wanted to know about Lachmann's method. A non-standard handbook of genealogical textual criticism in the age of post-structuralism, cladistics, and copy-text, with a foreword by Michael D. Reeve, translation by Federico Poole, Padova, libreriauniversitaria.it, 2014.

[3] I wish to thank Prof. Trovato and Dr. Adolfo Tura, who informed me of these manuscripts and contributed to arrange opportunities to study them.

[4] It is composed of 290 ff. (307 x 217 mm [212 x 95 mm]): 1-118, 128+4, 13-218, 2212, 23-338, 34-366. There are 25 blank ff.: 89, 90, 94, 176, 180, 269-286, 289-291.

[5] Bibliothèque Royal de Belgique, Bruxelles, n. 14614-14616.

[6] In the case of the Commedia, Tuscan production is the most numerous and serial, opposed to the northern one, slower and more “personalized”.

[7] Boschi Rotiroti (Marisa Boschi Rotiroti, Censimento dei manoscritti della “Commedia”. Firenze, Biblioteche Riccardiana e Moreniana, Società Dantesca Italiana, Roma, Viella, 2008) reports the opinion of the famous historian of miniatures Mirella Levi D’Ancona (Levi D’Ancona 1970). In fact Boschi Rotiroti was mistaken in her indication of the MS, confusing Bodm. 247 with Ph. 247. The reason is probably that when Levi D’Ancona formulated her judgment the codex was still in the hands of Martin Bodmer. Battista di Biagio Sanguigni, in the opinion of Levi D’Ancona, is responsible for the decoration of other three MSS: Ashb. 827, Laur. C.S. 204, Ricc. 1008.

[8] Both Boccaccio and the Cento group are two stemmatic families. While the Cento group is critically well-known, the Boccaccio’s group has recently been classified by our team. See Nuove prospettive sulla tradizione della Commedia. Seconda serie. Studi 2008-2013, edited by Elisabetta Tonello, Paolo Trovato, Padova, libreriauniversitaria.it, 2013 and Elisabetta Tonello, La famiglia vaticana e la tradizione Boccaccio, in «Filologia Italiana», 11, 2014, pp. 85-110.

[9] Marisa Boschi Rotiroti, Paolo di Duccio Tosi, un copista dantesco e non solo, on line documents for the exhibitions at the Biblioteca Trivulziana, Milano, 4 agosto – 18 ottobre 2015, http://graficheincomune.comune.milano.it/GraficheInComune/bacheca/danteincasatrivulzio/approfondimenti_ita.html.

[10] The northern family called bol is a recent discovery by our team.

[11] Both fi& and vatbocc are new sigla introducted by our team in order to indicate respectively the group of MSS around Fi MS and the group connected to the Vatican and Boccaccio tradition: Elisabetta Tonello, La famiglia vaticana e la tradizione Boccaccio, in «Filologia Italiana», 11, 2014, pp. 85-110 and Elisabetta Tonello, Per l’edizione della Commedia di Dante. Ricerche sulla tradizione tosco-fiorentina (1330-1501), Padova, libreriauniversitaria.it, forthcoming.

[12] For instance: 3.17.78 che notabili fier l’opere sue P] mirabili fi& + Fior. II I 43’ or 3.23.68 (B) quel che fendendo va l’ardita prora P] secando fi& + Fior. II I 43’

[13] In 1891 the young philologist Michele Barbi estabilished a canon of loci critici: a list of points of variation selected with great ability and for this reason still adopted by Dante scholars.

[14] For instance: 3.1.78 con l’armonia che temperi e discerni] tu tempri; 3.7.125 l’aere (aire Kraus 1) la terra e tutte lor misture] veggio; 3.31.142 che’ miei di rimirar fé più ardenti] furon più attenti.