To e-Laborate homepage   Login   Huygens Instituut
English > Carolingian scholarship and Martianus Capella > About the manuscripts >

Description of Leiden BPL 88

Show printer-friendly view Print View          

Leiden,Universiteitsbibliotheek, BPL 88

A ninth-century parchment manuscript of De nuptiis Philologiae et Mercurii of Martianus Capella, heavily glossed especially in the books I, II and IX. The last two quires, containing the text of book IX, were originally part of another manuscript or were especially made to complete an already existing, incomplete one.[1] They contain a different corpus of glosses: in the books I-VIII a version is found of the Anonymous corpus, in book IX of John the Scot's commentary. The joining together of these parts must have taken place already in the ninth century. The last two quires may have been written in Reims.


  • J. Préaux, Le commentaire de Martin de Laon
    Préaux dates the manuscript in the second half of the 9th century and identifies its commentary as a version of the Anonymous corpus, which he attributes to Martin of Laon. Speculating on the presence of 11th-century annotations of ownership of the abbey of St. Peter in Ghent, he suggests that the manuscript was originally written in St.-Omer or Laon, i.e. the two places to which the monks of St. Peter fled during hostile invasions in 880.
  • A. Verhulst, L'activité et la calligraphie du ‘scriptorium’ de l'abbaye Saint-Pierre-au-Mont-Blandin de Gand
    Verhulst mentions the ex libris and anathema-formula of St. Peter's abbey in Ghent and describes with great precision the Wichardian hand writing on the added bifolium 1^2 and the last leaf of the manuscript.
  • J. Préaux, Deux manuscrits gantois
    Préaux again studies the 11th-century additions in ‘Wichardian’ hand, which are found in this and the above described Vatican manuscript (Vatican, 1987). He now suggests a different place of origin. Speculating that the monastery asked her sister abbeys, i.e. Corbie or Reims, for books to enhance the educational and cultural level, he identifies one of these two as the probable place of origin. On the basis of a paleographical study Bischoff suggested a likeness to the script of Tours in the beginning of this manuscript, but from book V onwards a localization in Reims. Préaux argues for a break later in the manuscript using codicological as well as internal evidence; he considers book I-VIII as one, book IX as a second part.
  • C. Leonardi, I codici I, pp. 451-453, 457, 462-464; II,pp. 62-63
    Leonardi establishes the same compilation of two parts: book I-VIII and book IX. The joining together of these two parts must have taken place in the ninth century, for a small, Carolingian hand writes glosses throughout the entire encyclopedia. Referring to Bischoff, he locates the second part of the manuscript in Reims.
  • G. Glauche, Schullektüre im Mittelalter, p. 45
    Because of the presence of ‘Martinus of Laon's’ corpus of glosses, this manuscript is also in Glauche's list of sources which point to a special interest of the “Irish colony” in De nuptiis.
  • T.A.M. Bishop, Autographa of John the Scot
    Although Bishop never mentions the signature of Leiden, BPL 88, he is in fact writing about it in this paper, in which he is concerned with the problem of two hands, I1 and I2, both identified as the autograph of John the Scot.[2] I2 appears as one of the glossators in the ninth book of Leiden, BPL 88. His conclusion is that I2 is not the hand of John the Scot, since this hand shows “nothing peculiarly scholarly; nothing at all to show a moment-to-moment contact between the acts of thinking and writing” (p. 93).)
  • C. Leonardi, Glosse eriugeniane a Marziano Capella
    Leonardi adds to his earlier observations that there are three layers of glosses in the ninth book of Leiden, BPL 88:[3] E — a hand of Irish origin, which writes glosses to the narrative part and the last poem; A — a contemporary continental hand, writing glosses to the music-theoretical part of book IX; and R — a slightly later (but still ninth-century) hand which can also be recognized in the other part of this manuscript (books I-VIII). A and E write a version of John the Scot's commentary. He remains unable to identify the glosses of R.
  • J. Préaux, Les manuscrits principaux, pp. 79, 101, 123
    Préaux dates Leiden, BPL 88 in the third of fourth quarter of the 9th century, locates it in Auxerre (?), without further references (p. 79). Further on he mentions either the region Auxerre-Corbie, or Reims as places of origin (p. 101). Leiden, BPL 88 is a source for the Anonymous commentary, here attributed to Martin of Laon, of a second generation, following earlier sources such as Leiden, Vossianus Latinus Folio 48 and Besancon, Ms. 594.
  • J.C. Frakes, Remigius of Auxerre, Eriugena, and the Greco-Leiden, Latin circumstantiae-formula of accessus ad auctores, p. 241 n. 46
    Frakes cites part of the accessus ad auctorem found in both Leiden, BPL 88 and Vatican, 1987, to illustrate an “Irish predilection for the accessus”.
  • The results of these studies are repeated in, among others, C.E. Lutz, CTC II, p. 370; J.D. Reynolds, Texts and Transmission, p. 246; B.S. Eastwood, The chaster path of Venus, pp. 147, 150, 152, 154; and M.-E. Duchez, Jean Scot Erigène premier lecteur, p. 175.

The manuscript consists of 23 quires; 22 of eight leaves, 1 of four (quire 21) and 1 of six (quire 23). In quire 1 a bifolium, which was added to the originally eight leaves, was numbered in the modern foliation: 18 [1-10]. To quire 13 a single, smaller leaf is added after the sixth leaf in the quire, which was also numbered in the modern foliation (138 [99-107]). For the rest the collation of this manuscript is regular.

Formula: 18 [1-10]

2-128 [11-98]

138 [99-107]

14-208 [108-155]

214 [164-167]

228 [168-175]

236 [176-181]

All quires but the last carry an original quire-signature, written in the lower margin of their last versos: I (10v), II (18v), III (26v), IIII (34v), V (42v), VI (50v), VII (58v), VIII (66v), VIIII (74v), X (82v), XI (90v), XII (98v), XIII (107v), XIIII (115v) XV (123v), XVI (131v), XVII (139v), XVIII (147v), XVIIII (155v), XX (163v), XXI (167v). Quire 22 carries (by error?)[4] the number XXI (175v).

The foliation is modern, written in the upper right corner of each recto in pencil.

The leaves measure ca 253 x 235 mm. This must closely resemble the original size of at least the first part of the manuscript (books I-VIII), for the marginal glosses are all complete. Only on two folia in quire 23 are glosses in the lower margin damaged by cutting (ff 179r, 180v), which suggests that this part of the codex came originally from a slightly larger one.

The parchment is in general of fine quality. The leaves are relatively thick and stiff, some of the flesh-sides being perfectly white, some of the hair-sides perfectly smooth (e.g. quires 5 and 6).[5] The added bifolium (ff 1^2) and the last two quires are clearly of a lesser quality, thinner and more worn. The parchment has very few holes[6] and the margins of only a few leaves show the edge of the skin.[7] All quires of 8 leaves are arranged according to the pattern HFHF (2-12, 14-20, 22), and the quires of lesser leaves are arranged in the same manner, taking the number of their leaves into account: quire 21 has HF|HF, and quire 23 HFH|FHF.

Ruling is executed with a hard point per folium on the hair-side. The dominant pattern of furrows and ridges is ><><|><>< (1-20, 22).[8] The binio and ternio show the same pattern in fewer leaves: ><|>< (21), ><>|<>< (23). The pattern for the lay-out of the page is uniform from quire 1 to quire 21: double vertical through lines on either side of the text, pricked in the upper and lower margins; single horizontal through lines as the first and last textline, pricked in the outer margin; 21 rules, above topline, pricked along the edge of the leaves: 2-2/0/1-1/J, 26<9<135>9>55 x 32<170>50 mm; 21 lines, above topline, height per ten ca 83 mm.

There are only two exceptions: on f 55 the lower horizontal through line stops at the third vertical one (2-2/0/1A-1C/J) and on the bifolium 75^82 both horizontal through lines stop at the fourth vertical one (2-2/0/1-1:C/J).

In the last two quires five different lay-outs are found. In 22 the first and fourth bifolium (i.e. 168^175 and 171^172) have two vertical through lines on either side of the text, one horizontal through line at the head of the text, none below; 2-2/0/1-0/J. A double row of prickings can be seen along the edge of the first bifolium, a single row on the fourth one. The measurements are ca 25<8<142>7>53 x 27<196>29 mm, ten lines are ca 68 mm high. The second and third bifolium, however, have 2 horizontal through lines above, one below the text; 2-2/0/2-1/J, 22<7<142>7>54 x 28<195>28 mm. Again a double row of prickings is found on the one (169^174), a single row on the other bifolium (170^173).

In 23 three different types of lay-out occur in three bifolia:

* ff 176^181: 2-1/0/2-1/J, 19-21A,[9] a single row of prickings along the edge of the leaf, 30 lines, atl (ca 69 mm per ten), 23<7<145>52 x 25<198>22 mm.

* ff 177^180: 2(+2)-2(+2)/0/5-1/J,[10] a double row of prickings along the edge of the leaf, 31 lines, atl (ca 68 mm per ten), 25<8<140>7>52 x 21<205>28 mm.

* ff 178^179: 2-2/0/3-2/J, 19-21A, a single row of prickings along the edge of the leaf, 30 lines, atl (ca 69 mm per ten), 20<7<143>8>52 x 25<200>28 mm.

The many different kinds of lay-out in these two quires may suggest that these were newly composed in order to complete an incomplete or damaged manuscript. These bifolia were, perhaps, prepared for other manuscripts, but not yet used.

The text is written by several scribes in Carolingian minuscule.[11] The general appearance, however, is relatively uniform, with a small, upright letter (ca 2 mm high), 3-shaped g's, x's with long descenders to the left, r's which are tilted to the right and sometimes descend below the line, uncial capital M's and D's (besides rustic capital ones), and lightly clubbed ascenders. The letters are somewhat suspended above the lines. Some of the scribes carefully executed high-quality, very regular letters (e.g. ff 3-19r), others more of a scribbling, wavering script (e.g. ff 46v-48r). Striking are the differences in ink-colour, which varies between an almost faded light brown and nearly black.

In the glosses, several hands can be detected. The main glossing-activity of the first part of the manuscript took place contemporarily with or shortly after its making, by two contemporary glossators. The one who wrote the first layer of glosses wrote his additions in a tiny, very regular Carolingian minuscule of only 1 mm high. He used both the margins and the space between the lines and is precise in tying lemma and gloss together with a sign. The ink he used has a light-brown colour and has occasionally almost faded. The second layer of glosses must have been added only a few years later, for the script of this second glossator closely resembles that of the first one. His glosses are executed in darker ink, seemingly with less care, for the script does not look as regular as that of the first layer. This hand can also be found emending and filling in lacunae in the text. Throughout the codex headers are added which identify the number of the book (verso) and its content (recto). In the margins one sometimes finds subtitles written in rustic capitals, which were probably written by the second glossator. The headers, which look very similar, may also have been written by him.

These two glossators seem to have worked their way through the manuscript, including the additional last two quires containing book IX. They are especially active in books I, II, IV, and at the beginning of VIII. In books I, II and VIII the two hands are approximately equally active, whereas in III, V, VI and VII hand 2 is by far the most active.

In book IX, finally, the first hand disappears after the first page and the second hand appears to have been working together with a third one. This third hand shows many Irish characteristics, such as a high, open e, a low and open a, an r with a relatively long descender and a straight T-shaped t. His letters are low, broad, widely spaced and look as if written in haste. The script was once identified as the autograph of John the Scot Eriugene, then as I2, the hand of a student of John.[12]

In addition to the three hands, contemporary with the manuscript, several notes of later hands can be detected. First of all the Wichardian hands can be observed, who copied large schemes on the topic of Dialectica on ff 1v and 2r, a figure representing the planets and the zodiac on f 2r, an accessus ad auctorem, the refrain Scande celi templa with musical notation[13] and a Greek alphabet with numerical values on f 2v. They have also been active in some of the books of the encyclopedia, as can be seen e.g. on ff 63r and 78r. Several later readers of this manuscript wrote their occasional notes in the margins of Leiden, BPL 88, e.g. on ff 18v, 22r-v, 33v, 43r, 84r, 115r.

The punctuation is partly original, partly added by the (second?) glossator. Majuscules are used for explicits/incipits, at the beginning of verses, passages and “sentences”, for titles and subtitles, headers, refrains, etc. They are written in rustic capital, with occasional uncial M's and D's. Initial letters of 1.5-3 lines high are found at the head of books or prose-passages after verses. In book IX, the initial letters of the verses stand out because of their placement between the double vertical through lines. Colours are used to mark beginnings of books, poems and prose-passages. There are several titles and initials in red ink,[14] and passages with capital letters in brown ink, coloured with ochre.[15] Sometimes the two colours are found together, as for example on ff 96r-96v, where red capital letters are marked with ochre.

Again there are few sources which reveal the history of this manuscript. As was noted above, the codex contains parts from two different manuscripts, bound together in the ninth century. Bischoff noted that the script of book I-IV resembled that of Tours, and suggested that the second part could have been written in Reims.[16] Although Préaux argued for a break later in the manuscript — i.e. from book IX onwards —, he accepted the localization in Reims. Like Vatican, 1987, Leiden, BPL 88 was part of the library of St. Peter's abbey in Ghent in the eleventh century, as can be concluded from the presence of ‘Wichardian’ script in the added bifolium and the anathema on f 2v (the same as that found in Vatican, 1987 on f 146v):

Liber s. petri gandensis ecclesiae. Servanti benedictio. Tollenti maledictio. Qui folium ex eo tulerit. vel curtaverit. anathema sit.

(Book of the church of Saint Peter of Ghent. A blessing to the keeper, a curse to the thief. May he who took or cut a folium from it be damned.)

In an early article Préaux voiced the hypothesis that the manuscript could have been written in St.-Omer or Laon, two places to which the monks of St. Peter's fled for hostile Norman invasions in the 880's.[17] In later articles, however, he opted for a second hypothesis, which dates the arrival of the manuscript in St. Peter's abbey in Ghent in the period when Wichard was abbot, i.e. 1034/35-1058, and seeks the place of origin among its sister abbeys, Corbie or Reims. Because the two ways of establishing the origin of the manuscript, i.e. Bischoff's hypothesis based on paleographical evidence and Préaux' hypothesis based on historical evidence, partly concur, Reims is now generally accepted as the place of origin of the second part, that is, the last two quires, of the manuscript.[18]

Again like Vatican, 1987, the codex was in the sixteenth century obtained by the humanist scholar Franciscus Nansius, from whose heirs the university library of Leiden bought the manuscript ca. 1600. There it received the signature Biblioteca Publica Leiden, Latina 88. A schidula with this signature is attached to the spine of the codex and a stamp of the library can be seen in the lower margin of f 3r.

Today, the codex has a parchment spine and cardboard covers, covered with marbled paper. On the inside, paper leaves are glued to the covers. Two paper fly-leaves open the book, the second one watermarked with a bunch of grapes. The codex has five double cords (spaced 55-42-44-48-42-36 mm). The sewing is not original, for several now unused sewing stations can be seen in the gutters of the quires.[19] In the back of the codex, no paper fly-leaves are present, but there is a parchment folium which carries the signature BPL 88 and the number 182 (i.e. the 182nd folium of the codex). On it is a Leiden, Latin text in a fourteenth-century cursiva.[20] Its measurements do not concur with the manuscript: it is part of a bifolium, from which the first leaf is cut to a strip of only 35 mm wide, the second measures 200 (width) x 240 (height) mm. Its prickings are enigmatic: they run along the edge of the strip and form a rectangle on the edge of the preserved folium. Perhaps another piece of parchment was sewn onto it to make it large enough to serve as a cover. The leaf is now fixed to the sewing with one piece of string.

The manuscript contains:

  • two paper fly-leaves;
  • f 1r: empty, offset of a strip of a missal, used to strengthen the binding;[21]
  • f 1v-2r: ninth-century annotations: schemes highlighting the ars dialectica, a figure of the planets and the zodiac, made by use of a pair of compasses;
  • f 2v: eleventh-century annotations in ‘Wichardian’ hands: an accessus ad auctorem from the Anonymous corpus, the refrain Scande celi with neumes, a Greek alphabet with numerical values, the anathema of St. Peter's abbey, Ghent;
  • f 3r-181v: Martianus Capella, De nuptiis Philologiae et Mercurii, with many glosses taken from the Anonymous corpus (books I-VIII) and John the Scot's commentary (book IX).

    I 3r-18v; II 18v-34r; III 34r-58v; IV 59r-80r; V 80r-106r; VI 106v-133r; VII 133v-152v; VIII 152v-167v;

    (in a different script and with different glosses) IX 168r-181v;

  • f 182r-v: bifolium cut to size with a Latin text in fourteenth-century cursiva, perhaps once used as a cover.

[1] In the last two quires a chaotic mixture of lay-outs is found, which may suggest that these were made from left-overs of already pricked and ruled, but otherwise unused bifolia.

[2] L. Traube, preface to E.K. Rand, Iohannes Scottus, in: Quellen und Untersuchungen I (1906), vii-ix; id., Autographa des Iohannes Scottus, in: Paläographische Forschungen V (1912), (non vidi). Rand denied the identification in the article The supposed autographa of John the Scot.

[3] In the short summary that precedes his paper Leonardi states that he envisages it as a preliminary study for a new, critical edition of the B-version of John's commentary to book IX. This project, however, was never realized.

[4] Leonardi sees in this irregularity in the quire-numbering another argument for his thesis that the last two quires of BPL 88 originally belonged to a different manuscript. Although I do agree with his conclusion, I doubt whether the irregular quire-numbering found in 22 can serve as evidence, for the number has exactly the same lay-out as the other ones. This makes me suspect that the quire-numbering was added after the joining together of the two parts, just like two of the layers of glosses and the headers, uniformly added throughout the manuscript, and that the irregularity is simply an error.

[5] There are, however, also folia that show the markings of the skin, e.g., ff 52v, 54v, 83r, 88v, 90v, all hair-sides of quire 12.

[6] Only two in the main part of the manuscript (book I-VIII): ff 119, 127; five in its additions (bifolium 1-2 and book IX): ff 2, 169, 171, 176.

[7] A bad spot in f 89, and possibly the missing edges of ff 158, 159, 160, and 161.

[8] Both the added bifolium in quire 1 and the added folium in quire 13 remained unruled: 00><><|><>< (1), ><><|><0>< (13).

[9] The ruling of lines 19-21 reaches the edge of the page.

[10] I counted four vertical projecting lines on either side of the text, but for only two of these, prickings were made in the upper and lower margins.

[11] The script changes slowly, and it is difficult to point to the exact places where a change of hands occurs, but I suspect breaks in ff 19r-19v, 32r-32v, 46v, 48r, 54r, 59r, 61v, 71r, 71v, 87v, 167r-168v. The last two quires (ff 168r-181v) are clearly written in a different hand, using long ascenders and descenders (without clubbing), closed g's, one-storeyed a's, st-ligatures. The scribe spaces his letters less widely, which, combined with the larger number of writing lines, makes the page look very full.

[12] T.A.M. Bishop, Autographa of John the Scot, pp. 90-92. For a description of this insular hand, cf. C. Leonardi, Glosse eriugeniane a Marziano Capella, pp. 175-178.

[13] In her study of musical notation in Martianus Capella manuscripts, Corbin did not look at Leiden, BPL 88. She did, however, look at the 11th-century neumes on f 1r-v in the Vatican manuscript Vatican, 1987, which are an exact parallel to those in Leiden, BPL 88. The neumes can thus be identified as of Lorrainian origin. S. Corbin, The neumes of the Martianus Capella manuscripts.

[14] Ff 3r, 22v-24r, 24v-26v, 27r-31v, 40r-v, 79v-80r, 96r-v. Some of these red majuscules are written in red ink, others are redrawn with red ink (ff 24v-26v, 27r-31v, 40r-v), perhaps when the second or third layer of glosses was added to the manuscript.

[15] Ff 3r, 8r-v, 9r, 10v, 11r, 13r, 17v, 18r, 58v, 59r, 60v, 61r, 96r-v.

[16] He suggested this in a personal letter (29 December 1957) to J. Préaux, who repeated his view in Deux manuscrits gantois, pp. 18-19.

[17] Le commentaire de Martin de Laon, p. 457: "Or je crois qu'il est permis de supposer avec assez de vraisemblance que cet important codex de 182 folios [...] ne provient sûrement pas de Saint-Pierre de Gand, dont l'histoire intellectuelle n'est guère brillante. D'autre part il n'est pas sans intérêt de noter ici que lors des invasions normandes de 880, Gand servit de quartiers d'hiver aux pirates et que dès 846 les moines de Saint-Bavon et de Saint-Pierre se réfugièrent à Saint-Omer, puis à Laon [...]. On peut raisonnablement penser que le Leidensis 88 arriva à Gand lors du retour des moines après les invasions [...]."

[18] Each of the two localizations is, however, a hypothesis; neither is in itself based on conclusive evidence. The localization should therefore be taken for what it is: a hypothesis.

[19] I have been unable to reconstruct the sewing-method in its original form.

[20] The text has been identified by one of the librarians as "b. Bernardi, Orationem ad cor b. Virginis". The librarian refers to BPL 90 and BPL 94, in which leaves of the same manuscript can be found.

[21] One of the annotations on the fly-leaves identifies fragments from the same missal in BPL 48 and BPL 120, both of which also belonged to Franciscus Nansius. The fragment itself is now filed away in box BPL 2514 (B:78).

Last modified: 21-11-2008 11:49