Terracotta Angel, c.1896
Watts Chapel, England

 Photo ©: Jeff Saward/Labyrinthos

Please note, the contents

of this website are

© 2016 Labyrinthos

unless stated otherwise.
Please contact us for permission to reproduce
any text or images

A New Labyrinth at Knidos

Staffan Lundén

Reprinted from Caerdroia 33 - 2003 - pp.6-12


In Caerdroia 31, (2000), p. 54, the discovery of a labyrinth inscription at Knidos in Turkey was reported, hitherto unknown to labyrinth researchers. The purpose of this brief article is to add some more information on this interesting example.

Fig.1: The labyrinth inscribed
on a block of masonry
 at Knidos, Turkey.
Photograph courtesy of
Prof. Christine Özgen.

Location and context: The labyrinth (fig. 1) is carved on a block of black marble (height 60 cm, width 76 cm, depth 23 cm) broken on the right and left ends, which lies on the ground to the West of the Corinthian temple at Knidos (1). Above the labyrinth is an inscription in Greek: KYRIE BOETHEI “Lord help” (2). On both sides of the inscription the Christian cross is carved. A third cross is found to the left and below the labyrinth. A fourth, larger, cross is to the right of the labyrinth.

Beneath the arms of the large cross are the letters Alpha and Omega, connected to the arms by a vertical stroke. Further to the right of the large cross is a tendril emerging from a vase. The tendril has clusters of grapes and a heart-shaped ivy leaf. Between the labyrinth and the tendril are two palm trees and what appears to be a tree or a bush. The branches of the tree or bush terminate in small dots, which perhaps represent fruit.

The block is clearly a piece of masonry from a building, but probably the carvings postdate the removal of the block from this unidentified building. The rather symmetrical placement of the carvings on the block suggests that they were executed after the right and left ends of the block were broken off and this damage is less likely to have occurred when the block was still in its original position. (The damage in the upper left corner on the block, obliterating the upper parts of the first letters of the inscription, of course postdates the carving.) The block lies on the border of a main road; at Knidos, crosses and other carvings are also found at street junctions, which may tentatively suggest that both these locations were some sort of public meeting or waiting place (3).

Description: Shape: Round. Size: Judging from the dimensions of the block and the size of the letters in the inscription (max. height 4.5 cm according to Bluemel) the size of the labyrinth can be estimated to measure c. 21-23.5 cm high and 18-21 cm wide. At the centre the labyrinth is a small hole that indicates the labyrinth was constructed with the help of an engraving tool, probably a compass or divider.

Date: The carvings on the block appear to be uniform in style and were probably executed at the same time. Crosses of the type found on the block begin to appear at the end of the 4th century CE. Crosses with Alpha and Omega as pendants typically belong to the 6th and 7th centuries CE (4).

Comment: KYRIE BOETHEI “Lord help” is a very common Christian inscription. It may be followed by the words “your servant” and/or the name of the person(s) for whom aid is being sought. Sometimes “Lord” is exchanged for “Lord Christ”, “Jesus” and “Jesus Christ”, “Mother of god” or the names of various Saints (5).  It is interesting to note that on a vertical rock face in the marble quarry at Dokimia (Docimium) modern Iscehisar, Turkey it is found (written KĒ BOETHI) (6) above a number of broken, concentric circles (diameter c. 40 cm) that have been interpreted as a labyrinth (fig. 2) (7).  

Back to Caerdroia Archive

Fig. 2: The inscription at Dokimia, Turkey.
 Difficult to interpret, it would seem likely that this may
 indeed be a weathered representation of a labyrinth.
Illustration by Jeff Saward, after Röder.

Numerous representations of the cross (dating 6th-8th centuries CE) (8) and other Christian motifs are found in the vicinity. Due to the weathering of the rock, it is impossible to establish with certainty whether the Dokimia design actually is the remains of a labyrinth. Earlier I was sceptical, (9) but now when the Knidos labyrinth is associated with the same formula, it is tempting to suggest that the Dokimia carving represent a labyrinth or an attempt to draw a labyrinth.

The Knidos labyrinth, and the possible labyrinth at Dokimia, are among the earliest known examples recorded in Christian contexts. To the present author's knowledge, the only earlier securely dated (10) and securely attested (11) labyrinth is a Roman mosaic labyrinth found in the Reparatus basilica in Orléansville/El-Asnam in Algeria, which was founded in 324 CE (12). All the other motifs on the Knidos block – cross, tendril with grapes emerging from vase, palm tree, fruit tree - are all well established in the Christian repertoire of pictorial elements (13).  

It is beyond the scope of this presentation to try to enquiry into all possible explanations for the presence of the labyrinth in this milieu, but some tentative suggestions may be made. Perhaps the most straightforward explanation for the appearance of the labyrinth in a Christian setting is that the labyrinth design is based on a cross. Mention should also be made that, among the many variants of the cross employed in Early Byzantine times is the staurogram or crux monogrammatica, i.e. a Latin cross with the upright arm in the form of Greek letter Rho, (similar to a Latin “P”) (14). As the centre of the Knidos labyrinth is circular it is possible that it was seen as incorporating this type of cross, with the upper part of the letter Rho corresponding to the right half of the labyrinth’s circular centre.

It may be noted that the cross is found also in the design of the Orléansville labyrinth. The spokes of this labyrinth may be seen as corresponding to the arms of the cross. Moreover, in what looks like a jumble of letters in the centre of the labyrinth, one can read, beginning from the letter S in the middle, the word SANCTA “holy” in all four directions in the form of a cross. It is followed by the word ECLESIA “church” along the sides of the quadrant (Fig. 3). Possibly, the Knidos labyrinth contains a related idea. At the very first glance there are only inextricable lines, but a closer look reveals the cross.

Fig. 3: The mosaic pavemnt labyrinth at
Orléansville (El-Asnam), Algieria.
Illustration by Jeff Saward.

In a previous issue of Caerdroia (29, 1998, p. 38-42) it was suggested that the labyrinth might have a protective function. This conclusion was based on the context of the Conimbriga and Quanawat labyrinths. For the sake of completeness, it should be asked whether there is any evidence that the labyrinth on the Knidos block was executed for protective purposes (15). Looking at how the other motifs on the Knidos block are used in other contexts, such evidence could be argued to exist. The formula KYRIE BOETHEI is found together with protective symbols. For example, on a copper plaque excavated at Anemurion, Turkey, the formula appears above a representation of a the “much-suffering eye”, i. e. the Evil Eye of Envy being threatened by various weapons and animals (16).  

Representations of the cross may be accompanied with inscriptions that show that the cross was considered to give protection against different sorts of evil influences. For example, an inscription in Tepedschi, Turkey, reads: "Where the cross is present, Envy has no power" and on an inscription in Anasartha, Syria we read: "When we engrave your cross, O Christ... we escape every form of wickedness” (17). Still, notwithstanding the fact that the cross and the formula KYRIE BOETHEI were rendered for protective purposes, this does not mean they were executed for this purpose only. The cross, especially, is a symbol that has been invested with a multitude of meanings (18). Thus, although it is certainly possible the carvings on the Knidos block were rendered for protective purposes, substantial evidence to support such an assumption is lacking. Given the polyvalence of the different motifs on the block, other interpretations are also possible (19).

Staffan Lundén; Rome, December 2002.


1.  Knidos is situated on the tip of the peninsula Resadiye Yarimadasi (SE of the Greek island Kos and NW of the Greek island Rhodes) on the West coast of Turkey.

2. The inscription is published in W. Blümel, Die Inschriften von Knidos (Inschriften griechischer Städte aus Kleinasien 41), Bonn 1992, 146, no.244. Blümel gives a brief description of the block and its carvings. To the present authors knowledge this is the only previous mention of this labyrinth in a scholarly publication.

3.  For this information I owe thanks to Prof. Christine Özgen, who also kindly supplied the photograph of the block as well as permission to use the photograph for this publication.

4.  It should be noted that a 5th century AC date has been proposed for a cross with holes for pendants in a private collection in Munich. A. Effenberger in eds. M. Brandt & A. Effenberger, Byzanz. Die Macht der Bilder, Hildesheim 1998, 35-39, no.3, pl.22-25, idem in eds. L. Wamser and G. Zahlhaas, Rom und Byzanz. Archäologische Kostbarkeiten aus Bayern, München 1998, 72-76, no.64. However, Effenberger provides no evidence for this early date. For information on the dating of crosses, as well as the reference to Effenbergs publications, I am indebted to Prof. Josef Engemann (e-mail 31 October 2001).

5.  W. Prentice, Greek and Latin Inscriptions. Part 3 of the publication of an American archaeological expedition to Syria in 1899-1900. New York 1908, 22.

6.  The horizontal stroke over the E indicates that KE is a contraction (i e the nomen sacrum) of KYRIE. A. Paap, Nomina sacra in the Greek papyri of the first five centuries A.D., (Papyrologica lugduno-batava 8), Lugdunum Batavorum 1959, 1f, 101f. Cf. J. Diethart "Kyrie boethei in byzantinischen Notarsunterschriften" Zeitschrift für Papyriologie und Epigrafik 49, 1982, 79-82.

7.  J. Röder, “Marmor Phrygium. Die antiken Marmorbrüche von Iscehisar in Westanatolien“  Jahrbuch des deutschen archäologischen Instituts 86, 1971, 252-312, 288, 291, 293, fig.37. To the left and above the labyrinth and inscription is a further inscription (probably to be read as BOETHI) beneath a cross.

8.  Röder 1971, 294.

9.  "The Labyrinth in the Mediterranean," Caerdroia 27, 1996, 54.

10.  The labyrinth on the Hollywood stone, found beside a pilgrims trackway leading to the monastery of Glendalough, Wicklow, Ireland is dated to the 6th-14th centuries CE. J. Saward, The Caerdroia Field Guide 1987, 13f, idem, Ancient Labyrinths of the World, Thundersley 1997, 3. Accepting a low date for this labyrinth would make it roughly contemporary with the Knidos and Dokimia labyrinths, but as I have noted before (“The Labyrinth in the Mediterranean,” Caerdroia 27, 1996, 32f, n.16) the absence of evidence for labyrinths in the British Isles before the Medieval period calls for caution against such an early dating. Cf. also A. Saward, "The Rocky Valley Labyrinths," Caerdroia 32, 2001, 23.

11.  On the Tigzirt “labyrinth”, see next note.

12.  W. Daszewski, La mosaique de Thésée. Etudes sur les mosaiques avec représentations du labyrinthe de Thésée et du Minotaure. (Nea Paphos 2), Warsaw 1977, 102f, no. 4, H. Kern, Through the labyrinth. Designs and meanings over 5,000 years, Munich, London, New York 2000, 88, no. 117f. This labyrinth stands at the very end of the long tradition of designing labyrinths in mosaic floors. As no securely dated mosaic labyrinth postdates the beginning of the 4th century CE, the "labyrinth" in the basilica in Tigzirt erected in the second quarter of the 5th century CE, (most likely built in the period between 429 and 455 CE) stands in marked isolation.

It is included in both Daszewski's (no.6) and Kern's (101, no.171+) catalogues of mosaic labyrinths without any reservations, but ought to be treated with some scepticism. It is no longer extant and no photograph or drawing of it seems to exist. It is only known from the description by Gavault who writes “on retrouve les traces d'une partie du pavage représentant un labyrinthe de la forme la plus simple, c'est-ŕ-dire une sorte de volute carrée ou de grecque ŕ enroulements multiples, dont les lignes sont blanches en noires.” (P. Gavault, Étude sur les ruines romaines de Tigzirt, (Bibliothčque d'archéologie africaine 2) Paris 1897, 55.) Gavault continues with noting that representations of labyrinths are not rare on Christian monuments, referring to the pavement labyrinths of mediaeval churches. Clearly, Gavault describes a pattern with a complicated design, but can we be sure it is a labyrinth, although he uses this word to describe it?

The term “labyrinth” has been employed by modern scholars for many different sorts of patterns (for example, different swastika or meander patterns), which do not share the basic characteristics of the labyrinth (i e having a unicursal “path”, which must be covered in its entirety to reach from the entrance of the design to its centre) and which were not necessarily conceived as labyrinths by their ancient designers and viewers. (Cf. Kern’s critique of Daszewski using the term “labyrinth” for a meander/swastika pattern on a mosaic close to the Cestius pyramid in Rome Kern 98, no.163). The issue is further complicated by that Gavault adds that the labyrinth is of “the simplest form”. What does this mean?

He may have observed that the mediaeval pavements labyrinths (which are of the Chartres type) have a different lay-out than the Roman mosaic labyrinths, and thus a Roman mosaic labyrinth may be said to be “simpler” than a Medieval pavement labyrinth. But his way of phrasing may also signal that he describes a pattern which is somehow different from that of a Roman mosaic labyrinth. Unfortunately, from Gavault's account of the mosaic it cannot be determined for certain whether he describes a true labyrinth, as opposed to some other sort of geometric pattern.

Above the Tigzirt “labyrinth” is a fragmentary inscription, beginning with the letters CERNI, probably to be read as the latin word CERNIS “you see”. Gavault, 56, fig.13, P. Monceaux, “Enquęte sur l’épigraphie chrétienne d’Afrique,” Revue archaeologique 1906, 126-142, 138 no.210. It has been suggested (by W. Batschelet-Massini, “Labyrinthzeichnungen in Handschriften,” Codices Manuscripti 4, 1978, 33-65, 41) that the inscription quoted a passage from the work Contra Symmachum by the Christian author Prudentius. In this passage (2,882-890), beginning with the word “cernis”, Prudentius speaks of the path, which through many windings leads to error and death. In Prudentius' text this is compared with the simple path that leads to God (2,843-857).

Thus, in Batschelet-Massini’s interpretation the labyrinth becomes a graphic illustration of the crooked path of heresy. The problem with Batschelet-Massini’s theory is that it relies on just one word, which is not uncommon in the Latin language. Latin inscriptions written on mosaics, statues, grave monuments etc., frequently invoke the reader to take a look at the work of art in question. There is a least one (pagan) inscription that begins with the same word “cernis”. F. Buecheler, Carmina latina epigraphica 1, Lipsiae 1895, 173, no.369. Although the Tigzirt inscription begins with “cernis” it is far from certain that it quotes Prudentius.

An oft-repeated dogma in the literature on labyrinths is that the pavement labyrinths of Medieval churches originated in the Roman mosaic labyrinth, but the total absence of evidence for labyrinths in the church floors and in all forms of mosaic art of the intervening period of c. 700 years between the latest mosaic labyrinths in the 4th (or possibly the 5th century CE, if one counts the Tigzirt “labyrinth”) to the appearance of pavement mosaics in North Italian churches in the 12th century CE, makes this assumption untenable. Neither is the composition of the Chartres labyrinth type, which is used for the Mediaeval pavement labyrinths, hereditary to that of the Roman mosaic labyrinth. The Chartres type is instead a development from the cross labyrinth, probably via the Otfrid type. It might be worthwhile to remind that the pavement labyrinth in the church of San Vitale in Ravenna dates to the 16th, not to the 6th century as sometimes stated in older literature. (Kern 160, no.280f.)

13.  The significance of motifs drawn from the natural world in Early Byzantine art are discussed in H. Maguire, Earth and ocean. The terrestrial world in Early Byzantine art, (Monographs in the fine arts 43), University Park and London 1987. Early Byzantine interpretations of such motifs as vegetation bearing fruit may range from the literal to the symbolic. In the former case the motifs illustrate the richness and abundance of nature, and thus show the magnitude of Gods creation. In the latter case they signify different concepts beyond this world, like, for example, the Paradise. A representation of a vine with grapes may be understood literally, i.e. as a further example of the richness of nature, or symbolically. If the latter, the vine may signify a whole range of concepts including the Eucharist, Christ, his people and the Kingdom of God. Maguire 9f. Cf. E. Maguire, H. Maguire, M. Duncan-Flowers, Art and holy powers in the Early Christian house, (Illinois Byzantine studies 2), Urbana and Chicago 1989, 23f. Maguire rightly notes, that often only an inscription explaining the imagery, or a visual pointer within the imagery itself, can reveal which symbolism was intended. On the Knidos block, the interpretation of the carvings becomes difficult as any such key to their understanding is lacking,

14.  The staurogram derives from the Chi Rho monogram, formed of the first two letters of the name of Christ.

15.  Many protective symbols of pagan origin, continued to be used well into he Early Byzantine period, and are found in Christian contexts (like churches) and together with Christian symbols (like the cross). Cf. C. Bonner, Studies in magical amulets, chiefly Graeco-Egyptian. Ann Arbor 1950, 208-221. H. Maguire, “Magic and geometry in Early Christian floor mosaics and textiles,” in eds. W. Hörandner, J. Koder & O. Kresten, Andrias. Herbert Hunger zum 80. Geburtstag, Jahrbuch der Österreichischen Byzantinistik 44, 1994, 265-274. (reprinted in H. Maguire, Rhetoric, nature and magic in Byzantine art, Aldershot 1998.)

16.  J. Russell, “The evil eye in Early Byzantine society. Archaeological evidence from Anemurium in Isauria” Jahrbuch der Österreichischen Byzantinistik 1982, 32:3, 539-548, 544f, pl.4f, idem “The archaeological context of magic in the Early Byzantine period” in ed. H. Maguire, Byzantine magic, Dumbarton Oaks 1995, 35-50, 39f, pl.5f. The protective sense of BOETHEI is discussed in Prentice 1908, 22 and Bonner 1950, 46, 180.

17.  J. Engemann, “Zur verbreitung magischer Übelabwehr in der nichtchristlichen und christlichen Spätantike” Jahrbuch für Antike und Christentum 18, 1975, 22-48, 42f with further examples. When discussing a possible protective and luck-bringing symbolism of the carvings on the Knidos block it should be noted that designs showing, the riches of nature – such the fruit tree and vines with grapes - may be intended to invoke abundance and prosperity. E. Maguire, H. Maguire & M. Duncan-Flowers 1989, 9-13.

18.  On the various meanings and uses of the cross: s.v. “Kreuz 1”, in Reallexikon zur Byzantinischen Kunst 5, 1995, 1-202, A. Felle, s.v. “Croce” in ed. F. Bisconti, Temi di iconografia paleocristiana, (Sussidi allo studio delle antichitŕ cristiane 13) Cittŕ del Vaticano 2000, 158-162.

19.  When discussing the Cominbriga labyrinth, it was suggested that the labyrinth protected through bewildering or confusing the Evil Eye. It might be noted that a very late (9th or 10th century CE) Latin inscription shows that the cross was considered to protect by confusing the Devil. The inscription reads: CRUX XRI CONFUSIO DIABOLI, “The cross of Christ (is) the confusion of the Devil,” eds. R. Hodges & J. Mitchell, San Vincenzo al Volturno. The archaeology, art and territory of an Early Medieval Monastery, (BAR International series 252), Oxford 1985, 158, fig. 6:34f.