114 Maria Evangelatou

In general, the subjects treated in this Christian the corpus of miniatures hosted in the pages of this codex
compilation are of theological, moral, and social import: is exceptionally rich: according to Weitzmann’s count,
they focus on the relationship of God and humankind, the extant folios contain 1,658 miniatures,5 but originally
the virtues and vices of human nature, and the proper the whole manuscript would have contained around 1,830
conduct of Christians toward their fellow human beings miniatures.6 Although the majority are authors’ por-
and the state and religious authorities. Other, similar traits, narrative illustrations are also numerous,7 and the
florilegia also existed in Byzantium and survive in vari- subject matter of some is unique in the surviving corpus
ous recensions.2 Our codex is not only the earliest sur- of Byzantine art. On top of that, this large and richly
viving Sacra Parallela text, it is also the only one that illustrated codex makes lavish and extensive use of gold
is illustrated; indeed it is the only illustrated Byzantine leaf, for most parts of the miniatures as well as most of
florilegium of any kind that has been preserved.3 In addi- the titles (chapter headings) and excerpt identifications
tion, at 35.6 × 26.5 cm it is among the largest illustrated within titles.8 This technical characteristic is very rare
Byzantine manuscripts known today, but originally it was among surviving Byzantine manuscripts.9 In addition,
even larger: its pages are now trimmed, and of the more
than 420 original folios only 394 still survive.4 Moreover, Vision and Meaning in Ninth-Century Byzantium: Image as Exegesis in
the Homilies of Gregory of Nazianzus (Cambridge, 1999), 2.
Damascenus, TU 16.1 (Leipzig, 1897), esp. 42ff., 81–88, 161. Also M. 5  Weitzmann, Sacra Parallela, 11.
Richard, “Florilèges spirituels grecs,” Dictionnaire de Spiritualité, fasc. 6  Weitzmann (ibid., 6) counted “more than forty scenes and more than
XXXIII–XXXIV (Paris, 1962), 476–80, with references to previous one hundred thirty busts and medallions” in the first 33 illustrated folios
literature (the recension to which cod. Paris. gr. 923 belongs, the “Flo- of the manuscript, and he estimated that the 31 missing folios would
rilegium PML,” is mentioned in columns 482–83). The main points of have contained at least as many miniatures. He also noted: “The study
Richard’s analysis, still valid today, are summarized by P. Odorico, “La of illustrated books in general teaches us that the more richly a page is
cultura della ΣΥΛΛΟΓΗ: (1) Il cosiddetto enciclopedismo Bizantino; illustrated the more likely it is to be cut out by an admirer. Thus the esti-
(2) Le tavole del sapere di Giovanni Damasceno,” BZ 83 (1990): 1–21, mate here seems rather conservative; numbers aside, we might venture to
esp. 13–20. I thank Dr. G. Papadogiannakis for drawing my attention guess that some of the best and most interesting illustrations are lost.” It
to this article. The above issues are also mentioned to some extent by should be noted that this codex is densely illustrated throughout, so that
Weitzmann, Sacra Parallela, 8–10. the miniatures are not more numerous at the beginning and drastically
2  These were called by M. Richard spiritual florilegia, in distinction fewer toward the end (as happens in other Byzantine manuscripts); thus
to those he called dogmatic florilegia. See Richard, “Florilèges spirituels Weitzmann’s use of the first 33 folios to estimate the number of missing
grecs,” 475–511 and ODB 2:793–94. miniatures (and not, for example, the last 33 folios) should not be consid-
3  Weitzmann, Sacra Parallela, 10. ered inappropriate—although it is not clear why he used 33 folios instead
of 31, that is, a number equal to the number of missing folios.
4  Weitzmann (ibid., 3) mentioned that “a note at the end of the man-
uscript from about the thirteenth or fourteenth century states that at 7  Weitzmann (ibid., 11) mentioned “402 scenic illustrations and 1,256
that time it had ‘υκδ’ (= 424) folios.” For information on the missing portraits” and added that “this division must be considered approxi-
pages (random losses throughout the codex), see ibid., 4–5. Weitzmann mate.”
said also (ibid., 10) that cod. Paris. gr. 923 “is the largest illustrated Greek 8  The impressive effect produced by the extensive use of gold can be
manuscript in existence.” He was probably referring to the unparalleled glimpsed in ibid., color plate of folio 208r (unnumbered plate, before
number of miniatures this codex contains (see below) rather than to the black and white plates). The only parts of the miniatures that are
the dimensions of the manuscript. Although the pages of this codex are not gilded are faces, hands, and other naked parts of bodies, and rarely
fairly large in comparison to the majority of surviving Byzantine illus- some other elements like headgear and other attributes (e.g., monastic
trated manuscripts, other codices are much larger in size, the largest of koukoulion and cruciform staff on fol. 208r mentioned above); natural
all (according to John Lowden) being a pair of Prophet Books with cat- elements like fire, rays of light, or the sea (e.g., fols. 68v, 69r, 207v); and
enae, codices Vat. gr. 1153 and 1154, which are 50.8 × 37.7 cm and 49.5 animals (e.g., storks and swallows on fol. 200v). See ibid., 14–15, on the
× 35.5 cm respectively (the latter is trimmed). Several other illustrated use of gold. See also Figs. 25–27 below.
manuscripts with catenae (for example, Prophet Books, Octateuchs, and 9  Well known to scholars are the ninth-century codices Ambros.
Psalters) have dimensions similar to or larger than those of cod. Paris. gr. 49–50 (Homilies of Gregory Nazianzenus) and Vat. gr. 749 (Book of
gr. 923. For some examples see J. Lowden, Illuminated Prophet Books: A Job). The use of gold is very extensive in the Milan Gregory and rather
Study of Byzantine Manuscripts of the Major and Minor Prophets (Uni- limited in the Vatican Job; see A. Grabar, Les manuscrits grecs enluminés
versity Park and London, 1988), 9, 14, 22, 32, 42; idem, The Octateuchs: de provenance italienne (IXe–XIe siècle) (Paris, 1972), 15–16. Grabar men-
A Study in Byzantine Manuscript Illustration (University Park and Lon- tioned two other Byzantine manuscripts with the same technical char-
don, 1992), 11, 22, 26; idem, “Observations on Illustrated Byzantine Psal- acteristic (ibid., 25, 16): the ninth-century cod. Paris. gr. 2179 (Diosko-
ters,” ArtB 70 (1988): 245–48, esp. figs. 3–4 with valuable comparative rides), where gold leaf is used for the clothes of a very few figures at
material. A number of large illustrated manuscripts have survived from the beginning of the manuscript; and the eleventh-century cod. Paris.
ninth-century Byzantium. One example which surpasses the dimensions gr. 74, which is illustrated by minute figures, sometimes dressed in
of Paris. 923 is cod. Paris. gr. 510, 41 × 30 cm, 464 folios; see L. Brubaker, gold. Gold leaf is also used for the clothes of the figures which form the

DOP 62
 Word and Image in the Sacra Parallela (Codex Parisinus Graecus 923) 115

the exceptionally high number of miniatures and titles however, no systematic and comprehensive study on the
in Parisinus 923 means that no other manuscript with sources, method, and purpose of its production has ever
the same technical characteristic contains as much gold been written.13 The only monograph on this manuscript
leaf. In sum, the Sacra Parallela of Paris not only is excep- is the 1979 publication by Weitzmann,14 but as I shall
tional in the preserved corpus of Byzantine art, but was soon explain, his work is primarily a study, not of the
intended to be exceptional in its time as well. Sacra Parallela itself, but of pre-iconoclast manuscript
Parisinus 923 is dated to the ninth century,10 a period illumination as Weitzmann understood it on the basis
when the dispute over the use of religious imagery was an of his assumptions about the Paris codex.
important issue for much of the intellectual and ecclesi- Weitzmann was interested not in the Sacra Paral-
astical community of Byzantium, even after the triumph lela per se but in the glimpses of the past he thought he
of the iconophiles in 843.11 Since few illustrated manu- could catch through it.15 The following quotations from
scripts survive from that period, and since the Paris Sacra his monograph are characteristic of his approach. He
Parallela is of such an exceptional character, it is quite wrote: “It is no exaggeration to say that no other known
frequently discussed by Byzantinists, and certain aspects manuscript can contribute so much to our knowledge
of its illustration have been studied to some extent;12 of pre-iconoclastic book illumination, so little of which
has survived. . . . The pictures are of only secondary
interest as illustrations of a florilegium. Their primary
illustrated initials in a little-known Byzantine manuscript: cod. Paris.
interest lies in the fact that they are epitomes of vast lost
gr. 41, a small Psalter (9.8 × 9.7 cm), 12th c., laconically described by
H. Omont, Inventaire sommaire des manuscrits grecs de la Bibliothèque miniature cycles whose existence can be proved by their
Nationale et des autres bibliothèques de Paris et des départements (Paris, migration into the Sacra Parallela.”16 “Only a handful of
1898), 7. See also M. Evangelatou, “The Illustrated Initials of Codex Parisi- Early Christian illustrated manuscripts has come down
nus graecus 41,” Word and Image 24, no. 2 (2008): 199–218.
to us, but clearly many lacunae in our knowledge can be
10  Obviously, a dating before or after 843 is also related to the issue of
provenance. See Weitzmann, Sacra Parallela, 20–25. For an overview of
filled by the excerpted miniature cycles in the Sacra Par-
the literature (concerning also dating and provenance) see L. Brubaker, allela. This constitutes perhaps the primary importance
Byzantium in the Iconoclast Era (ca. 680–850): The Sources; An Anno- of our manuscript, more significant even than its value
tated Survey (Ashgate, 2001), 49–50.
11  This is evident, for example, in the writings of Patriarch Photios,
as J. Lowden and L. Brubaker, whose observations will be mentioned
who often referred to arguments of the iconophile literature about the
on various occasions in this article.
status and function of images and was actively involved in the celebra-
tion for the restoration of holy images, as his speech on the inauguration 13  Its exceptional nature is perhaps what makes this codex so difficult
of the mosaic in the apse of St. Sophia clearly shows. See, for example, to approach and understand. The lack of a facsimile or other illustrated
the perceptive comments on this homily by L. James, “Senses and Sen- edition of the manuscript makes its study particularly challenging, espe-
sibility in Byzantium,” Art History 29 (2004): 522–37, esp. 529–32. The cially for the evaluation of the relationship between word and image. In
illustration of cod. Paris. gr. 510, which was commissioned by Photios, is my experience, the script of cod. Paris. gr. 923 is discouragingly difficult
characterized by various elements which betray his interest in iconophile to read from the microfilms available (especially because many of the
theology. See Brubaker, Vision and Meaning (n. 4 above), for example, lines in the letters of the text were made exceedingly thin by the scribe
51–52, 216, 220, 225, 281–307, 401. For an overview of the function of and are hardly visible). As access to this manuscript is very restricted for
images according to iconophile theology and the possible links of the reasons of conservation, it would be of great service to Byzantinologists
Sacra Parallela illustration to iconophile arguments, see L. Brubaker, if the Bibliothèque Nationale were to produce a photographic record of
“Byzantine Art in the Ninth Century: Theory, Practice, and Culture,” the entire manuscript in digital form of high resolution.
BMGS 13 (1989): 56–81; eadem, Vision and Meaning, 19–58. 14  Sacra Parallela (n. 1 above).
12  For example, the style of the miniatures has been discussed by Grabar, 15  In his book review of the Sacra Parallela monograph, Cyril Mango
Manuscrits grecs de provenance italienne (n. 9 above), 22–24, and by rightly observed that what interested Weitzmann most was “the Paris.
O. Oretskaia, “A Stylistic Tendency in Ninth-Century Art of the Byz- gr. 923 not in itself, but rather as a witness for the reconstruction of Early
antine World,” Zograf 29 (2002–3): 5–18. Certain iconographic themes Christian picture cycles.” See The Antiquaries Journal 62 (1982): 162.
have been discussed, for example, by J. R. Martin, “An Early Illustration 16  Weitzmann, Sacra Parallela, vii. C. Walter, “Liturgy and the Illustra-
of the Sayings of the Fathers,” ArtB 32 (1950): 291–95; E. Revel-Neher, tion of Gregory of Nazianzen’s (sic) Homilies: An Essay in Iconographi-
“Problèmes iconographiques dans les Sacra Parallela,” in L’art et les révo- cal Methodology,” REB 29 (1971): 188–89, has acutely criticized the use
lutions: Actes [du] XXVIIe Congrès international d’ histoire de l’art, Stras- of the term “migration” for the copying of miniatures from one codex to
bourg, 1–7 septembre, 1989, sect. 4, Les Iconoclasmes (Strasbourg, 1992), the other (a term introduced by Weitzmann), because it undermines the
7–12; M. Bernabò, “L’illustrazione del Salmo 105 (106) a Bisanzio ed una active role of the painters in deciding which miniatures to adopt from
nota sui Sacra Parallela di Parigi,” Medioevo e Rinascimento 14 (2000): their models (“miniatures do not move themselves”). I thank Prof. M.
85–109, esp. 108–9; and in works of more general scope by scholars such Papadaki for drawing my attention to this article.

DOP 62