Greek Cryptograms in Southern Italy (and Beyond)

Par Linda Safran
Publication en ligne le 17 novembre 2017

Texte intégral

1Like most stories told at Kalamazoo, this one begins a long time ago. The object that you now see isolated, elevated, and framed in a modern setting is the subject of legend in the Salento region of southern Apulia. The story is that a farmer, plowing outside the town of Parábita one day, accidentally discovered the tall painted monolith. In a familiar topos, his cows knelt down before the image of the Virgin. The stone was brought to the nearby parish church of Parabita where, mirabile dictu, it refused to stay, returning on its own power to the place where it had been found. A small chapel was built on the spot, and both it and the image were called the Madonna della Coltura or Cultura. This alleged name of the local field was very likely connected with agricoltura, agriculture, or with a Greek word for bread, κολλύρα, making this Virgin a protector of food or fields. The chapel is first recorded in a pastoral visit by the local bishop in 1452.1 When it was demolished in 1913 because of its poor condition, the old image was revealed at its full height; for several centuries before that it was visible only through a small porthole in a baroque altar.

2The basilica that now houses the monolith was mostly built between 1913 and 1942.2 The Madonna della Coltura is venerated today with a procession on the last Sunday of May, during which a replica of the stela and a statue of the Virgin are carried through the town and outside it to the supposed place of its discovery.3 The real monolith is 170 centimeters (5 feet 6 inches) tall; that’s two inches taller than I am. The front is 55 cm (21 inches) wide. The Virgin, who is usually described as “Byzantine” even though she is painted in the local late Gothic style, can be dated on stylistic grounds to the 1430s.4 Note, however, that an older layer of painting is visible underneath this one at the upper right corner. The painting we see today covered an earlier image, probably damaged, and presumably of Mary. Both the visible image and the mid-fifteenth-century mention of the chapel accord well with the general vogue for Marian imagery in southern Italy in this period, when other such representations were also discovered miraculously and numerous Marian sanctuaries were built. But what is on the other three sides of the stela, and how did those sides fit into early fifteenth-century cultural practice?

3The back and both short sides of the Parabita stela are each adorned with a cross and a series of four- or eight-letter cryptograms in Greek. In the four-letter groups, which are also called tetragrams, each letter represents the beginning of a word; when there are eight letters, they are the first and last letters of a name or word. The words together make up a meaningful phrase, even if we cannot always decipher it.

4On the back of the stela is a cross with vines growing from its base, an image known as a floriated cross, which underscores that the cross is the Tree of Life. All four arms terminate in crown-like finials, and teardrop-shaped projections decorate the square at the crossing. The Parabita cross is surrounded by five rows of Greek letters on a ground divided by color into quadrants. This two-tone medallion is in turn set into a green-and-black fictive marble ground. The painting on this side is very carefully executed. Many comparisons for this elegant cross with cryptograms can be found in Serbia,5 although the crosses there always have at least two horizontal bars.

5The two short sides of the Parabita monolith are only 27 cm across, half the width of the front and back, and of these the left side was painted almost as carefully as the back that we just saw. Here, however, the cross has no living vines, and the five rows of Greek letters are different. The two-toned medallion is not set into fictive marble, but instead divides a budding vine that unfurls against a two-toned red background. The overall effect imitates intarsia, although this side is considerably more damaged than the back.

6The artistic and epigraphic excellence of the back and left side contrast strongly with the remaining right side. The surface here is noticeably rougher, and so is the execution: the cross and medallion are simplified and squeezed into the space, the Greek letters are irregular, the abbreviation marks are simplified, and the flowering vine is crude and monochromatic. It is very clear that the right side was a poor imitation of the left. Indeed, the Greek on the two sides is almost identical, and I will return to it in a moment.

7There is a very close connection between the Parabita cross and the one we see in a Franciscan church dedicated to St. Catherine at Galatina, 14 kilometers to the north. This important church was decorated with frescoes in the first decades of the fifteenth century under the high-level patronage of the Orsini–Del Balzo family, the local rulers appointed by the Angevins. Artists working here introduced the late Gothic style to the region. In the apse of the so-called “Orsini chapel,” perhaps a family worship space in the right aisle, we find both the living cross with crownlike finials on the same two-tone, quadripartite background and the budding vine on its two-color ground. I cannot prove “influence” in either direction: the Parabita artist may have imitated the prestigious Galatina paintings, or the Galatina painters copied an important local model. And there is one striking difference—either the Orsini chapel artists failed to copy the Greek cryptograms at Parabita, or the Parabita artists added Greek letters to the cross and medallion design that they were otherwise imitating. Let’s look more closely at these Greek letters. All of them have an abbreviation mark on top, indicating that each letter represents an abbreviated word. The mark is more elaborate on the front and left sides; it is just a plain bar on the right. All three sides bear letters representing Ἱησούς Χριστός Νικά, Jesus Christ conquers, here outlined in red; the back uses six letters rather than eight. This is one of the most common cryptograms, appearing across many centuries and in many media. But look more closely: the crude right side copies the letters for Christ incorrectly. Instead of chi, the X shape, for the beginning of Χριστός, it duplicates the iota of Ἱησούς instead.

8On the back we find the chi repeated four times,probably standing for Χριστός χριστιανοῖς χαρίζεται χάριν, Christ Grants Grace to Christians. Under that are the initial letters of Φῶς Χριστοῦ Φαίνει Πᾶσιν, the Light of Christ Shines for All. This is a very ancient phrase from the Liturgy of the Presanctified, directed to the worshippers after the altar has been censed in the form of a cross. These explicitly salvific phrases on the back are, so far, well known. But below them, outlined in light blue, is a cryptogram not found anywhere else: Τ Θ Κ Λ [tau, theta, kappa, lambda], perhaps a phrase that begins τόπος Θεοῦ, place of Christ.

9On the two short sides, in addition to Jesus Christ Conquers, are not one but two undeciphered cryptograms:Χ Φ Χ Χ [in violet, chi, phi, and two more chis] and, in yellow, the equally mysterious Φ Π X X [phi, pi, and two chis]. We tend to assume that Greek cryptograms, like the rest of Greek, are supposed to be read from left to right, but in manuscripts they sometimes appear in graphic form. When arranged in a cross shape the letter at the crossing marks the beginning of all four words, and the order of those words can vary. So, on the two sides of the stela, the two-line phrase outlined in yellow might in fact be read horizontally as phi, pi,chi, chi, but it might also be read from top to bottom, Φ Χ [phi, chi], like the beginning of Φῶς Χριστοῦ, Light of Christ. And perhaps it makes sense to look not only for statements about Christ and the cross but also for epithets of the Virgin, who is shown on the front side. Reading horizontally again, Φυλακτήριον πάντων, Guardian of All, and Φοβερὰ προστασία, Terrible or Fearful Protector, both come from the hymnographic tradition of the Theotókos, as do many other terms and phrases that begin with Φ [phi].A poet and monk named John Ǧeométrēs gave a list of these in the tenth century, as Andreas Rhoby and a colleague pointed out in a recent article.6

10Cryptograms juxtaposed with a painted cross are often found in Orthodox churches in such liminal spaces as doorjambs and window frames. By in the late thirteenth century, when they become very common, the enigmatic letters are thought to have had an apotropaic function that reinforces the protective and mysterious power of the cross itself.7 Crosses with cryptograms are also found in the church sanctuary: either in niches used for preparation of the eucharistic elements, on the altar table, or facing the altar on the inside of the templon barrier that divides the clergy from the laity.8 This is where I would place the Parabita stela. Its right side is rough, which suggests that it originally was attached to something on that side: perhaps that something was a templon screen in a rock-cut Orthodox church. The Virgin and Child would have faced the congregation while the well-executed cross on the reverse faced the clergy. An image of the Virgin flanking the entrance to the sanctuary is standard in Orthodox churches, and in her study of the late Byzantine sanctuary screen, Sharon Gerstel showed how the icons that filled the spaces of the screen were often painted to imitate marble on the side facing the sanctuary.

11If I am correct, the left side of the Parabita monolith would have been the right jamb of the so-called Royal Door into the church sanctuary, as reconstructed hypothetically on the lower left. This hypothesis is supported by the holes found only on this side. They indicate that something was inserted here, such as a rod or hooks to hold the curtain that hid the sanctuary from the view of the laypeople. Such a curtain would have partially obscured the cross and cryptograms on that side, but it did nothing to diminish their power.

12Let me recap my proposed scenario. A farmer stumbled upon a ruined rural chapel in which part of the masonry templon screen still stood. He freed the image of the Virgin, still partly attached on its right side, and transported it to Parabita. The Virgin was returned to her original site, and an Orthodox chapel was built or restored there in her honor in the first decades of the fifteenth century, contemporary with the paintings of the Orsini chapel at Galatina. The Virgin’s image was refreshed by overpainting it, and the crosses and cryptograms were added to the back and left side while the right side was placed against a new screen or a wall. Perhaps soon after, the Virgin was framed in a baroque altar that mostly obscured her. The crude right side of the stela may have been painted in imitation of the left at this time, but it seems more likely that it was painted in 1913, when the monolith’s full height was revealed, and it was isolated as the focus of a new basilica in Parabita dedicated to the Madonna della Cultura. Only a study of the paint could resolve this question.

13Of great interest, then, is the early fifteenth-century context for the original cryptograms on the back and left. Their intended audience was presumably literate and capable of reading Greek; if my scenario is correct, this means Orthodox monks. Crosses with cryptograms are most often found in monastic contexts; even today they decorate the monastic Great Schēma, the garment worn by the highest and holiest level of Orthodox monks and nuns. The three cryptograms at Parabita that lack parallels and remain undeciphered suggest that literate Greek-reading monks were supposed to meditate on their meaning along with the image of the cross.

14Everyone could recognize the Virgin and Child on the front of the stela, and “Jesus Christ Conquers” must also have been familiar to many viewers. The abbreviated forms of “Christ Grants Grace to Christians” and “The Light of Christ Shines for All” were probably accessible only to individuals with greater knowledge. The undeciphered cryptograms speak to yet another level of literacy—to reception by a group of readers who had the liturgical or scriptural knowledge to engage in the mental exercise of puzzling them out.9

15Greek cryptograms also appeared in other artistic contexts in southern Italy, and I show a few examples from the thirteenth century through the early fifteenth. On the left, we find the Φῶς Χριστοῦ Φαίνει πᾶσι,the Light of Christ Shines for All, on an orb held by the Archangel Michael at Soleto, but the cryptograms on Michael’s orb in three other local churches are undeciphered: at the top center, at Miggiano, you can see, reading horizontally, Μ ΘΥ Π Τ [mu, theou, pi, and tau], and the same letters are at another church that I’m not showing. At the lower right, a rock-cut church at San Vito dei Normanni, the orb has a different final letter. The first two letters on all three of these orbs probably refer to Mēter Theou, the Mother of God, and not to St. Michael. At least one of these, Miggiano, is a monastic church, as is the one that contains the cross and cryptograms shown in a drawing at the upper right. This church at Palagianello is dedicated to Santi Eremíti, holy hermits.10

16Finally, on the back of a funerary stela dated 1330, cryptograms are incised in the triangles between the rays of a starburst in relief.11 Once again, the individual letters are topped by an abbreviation mark indicating that they stand in for whole words. Inside the triangles are Α Β  Τ ? [alpha, beta, tau, and something] in the top row, and Τ Χ Λ Ρ Ρ (?) [tau, chi, lambda, and possibly rho] in the lower zone; perhaps this line spells τόπος or τάφος Χριστοῦ, place or tomb of Christ, but the intended order may be different or even multiple. There are additional letters on the semicircles and quarter-circles at the ends of the starburst rays, but these are all omicrons or epsilons, or in one case an ēta, and it is hard to arrange these into the usual groups of four. There is no evidence to connect the deceased Nicholas Ferriaci with the monastic profession, but ample cause to imagine that his family wanted to draw attention to his tomb by using learned-looking cryptograms.

17What explains the presence of Greek cryptograms in the late medieval centuries in a region that the Byzantines relinquished in the eleventh century? Greek was still spoken and Orthodoxy still practiced in Parabita and the other sites I have mentioned. The pastoral visit of 1452 records that Orthodox liturgical books were present in three of Parabita’s churches in that year.12 In the parish church dedicated to St. John there was a horologion, a book of hours, which included the fixed portions of services during the day, as well as a stoichos, which contained the texts for special commemorations. The church of St. Antony had a menaion, with offices for every day of the year; a katanychtikon, for nighttime services; a Triodion, for the Lenten season; an októēchos, for the eight tones, what we would call a hymnal; and a sticherarion, a book with musical notation for Orthros (matins) and Vespers throughout the year. Finally, the church of St. Blasius had a typikon, a monastic book with the order of services and eucharistic hymns, and possibly with specific instructions for running a monastery. Its Latin Quadragenále, used during Lent, is glossed in the pastoral visit “in Greek seragosto,” referring to a Tessarakoste, a Greek book of homilies or sermons for Lent. It’s true that none of these Orthodox service books is associated specifically with Sta. Maria di Cultura, but this is not a problem because books are recorded at very few of the churches included in the episcopal inventory, even though they were certainly there. What the list certainly shows is that an impressive range and variety of Orthodox service books was still available in Parabita in the mid-fifteenth century; some may even have been bilingual. This is just before almost all such books were purchased—or pillaged—for the great manuscript collections in Italy and elsewhere.13 Only three fragmentary Greek manuscripts remain in the Salento today.

18Looking again at the Parabita monolith, the lack of familiarity on the right side with the most basic Greek form of Jesus Christ - the letters that indicate Ἱησούς Χριστός - seems telling. The person responsible for copying four Greek letters only inches away on the left side was unable to do so. I cannot say for certain whether this ignorance of Greek and demonstration of sloppy technique occurred in the twentieth century or much earlier. But I can say that at least until the mid-fifteenth century there were literate people, probably male monastics, who were able to compose and contemplate Greek cryptograms at Parabita. They may have thanked the Madonna della Cultura every day, but that particular miracle would not last much longer.

Documents annexes


1  The chapel was called Ecclesiae Sanctae Mariae de Cultura. Sergio Ortese, “La pittura tardogotica nel Salento leccese: I cantieri “minori.” Tesi di dottorato di ricerca, Università del Salento, 2008, 318–22; C. G. Centonze, A. De Lorenzis, and N. Caputo, Visite pastorali in diocesi di Nardò (1452–1501), ed. Benedetto Vetere (Galatina: Congedo, 1988), 133–35, fols. 54r–55r.

2  Paolo Vincenti, “La Madonna della Cultura a Parabita,”, maggio 2006. Pope John Paul II elevated it to a minor basilica in 1999.

3  Also part of the festival: procession of a papier-maché statue of the Madonna della Coltura through Parabita and eventual enthronement; recapitulations of the original entry of the monolith into the city (a race); fireworks from the campanile; a potato festival.

4  Ortese, “La pittura tardogotica.”

5  Branislav Živković, Mileševa: Les dessins des fresques, Les monuments de la peinture serbe médiévale 10 (Belgrade: Institut pour la protection des monuments historiques de la République de Serbie, 1992), p. 53.

6  Emmanuel Moutafov and Andreas Rhoby, “New Ideas about the Deciphering of the Cryptic Inscription in the Narthex of the Panagia Asinou (Phorbiotissa) Church (Cyprus,” Medioevo Greco 12 (2012): 113.

7  Christopher Walter, “The Apotropaic Function of the Victorious Cross,” Revue des Études Byzantines 55 (1997): 193–220 (esp. 211ff.); Gordana Babić, “Les croix è cryptogrammes, peintes dans les églises serbes des XIIIe et XIVe siècles,” in Byzance et les Slavs, Mélanges Ivan Dujčev (Paris: Association des amis des études archéologiques des mondes byzantino-slaves et du Christianisme oriental, 1979), 1–13.

8  Sharon Gerstel, “An Alternate View of the Late Byzantine Sanctuary Screen,” Thresholds of the Sacred: Architectural, Art Historical, Liturgical, and Theological Perspectives on Religious Screens, East and West, ed. Sharon E. J. Gerstel (Washington, DC: Dumbarton Oaks, 2006), 146f.

9  See Rossitza Schroeder, “Looking With Words and Images: Staging Monastic Contemplation in a Late Byzantine Church,” Word & Image 18.2 (2012): 117–34, at 129.

10  Roberto Caprara, L’insediamento rupestre di Palagianello, I. Le chiese. Florence: Il Davide, 1980 (drawing from 1963); The cryptogram is now transcribed as M I B X.

11  André Jacob, “Notes sur quelques inscriptions byzantins,” 82; idem, “Vaste,” 255–56; Guillou, Recueil, 173–74, no. 161. “The servant of God Nicholas, son of Vitalius Ferriaci, slept at the seventh hour on Tuesday, September 25, 6839 (=1330), fifteenth indiction. And you who read this, pray for him.”

12  Vetere, “S. Maria de Nerito,” 327, 330–31; Visite pastorali, 39, 130, 132.

13  In 1473 another bishop records that a different local church is bene ornatum cum aucona ante altare—well adorned with an icon before the altar: Visite pastorali, 135 (Santi Viti in Parabite).

Pour citer ce document

Par Linda Safran, «Greek Cryptograms in Southern Italy (and Beyond)», In-Scription: revue en ligne d'études épigraphiques [En ligne], Livraisons, Première livraison, mis à jour le : 25/03/2019, URL :