Thanks go to the Danish National Research Foundation (grant DNRF119) for supporting the research undertaken within the framework of Centre for Urban Network Evolutions and to the Carlsberg Foundation for funding the Palmyra Portrait Project (2012-2020).
1Edessa, modern Urfa (Şanlıurfa) in the south of Turkey, was the centre of an independent kingdom from c. 135 BC to AD 213, at which point it was annexed by the Roman emperor Caracalla and fell under direct Roman rule (though the royal house of Edessa did not disappear completely until c. AD 248) (brief history in Healey, 2009, pp. 13-16). Between AD 73 and 253/83 (for newly revised dates see Caillou & Brelaud, 2016) the city gives us the first evidence of the local form of Aramaic, which later came to be called ‘Syriac’. This appears in the form of inscriptions, many of them carved on stone and set in mosaic and associated with tombs. These have been known and studied since the late 19th century (see Drijvers & Healey, 1999 for the corpus of inscriptions to that date; for earlier publication note, e.g., Pognon, 1907; Sachau, 1882).
2However, rather little attention has been given to the actual monuments on which the inscriptions appear and with which they were associated (for exceptions, see Colledge, 1994; Parlasca, 1984; Segal, 1970, pp. 29-41 and plates). The most striking of these monuments are the mosaics, most of which come from multi-burial family tombs carved into the hillsides outside the city walls (see Çetin, Demir, Desreumaux, Healey, & Liddel, 2020 for recently discovered tombs just outside the walls of Edessa). Some tombs have no surviving mosaics but have inscriptions and images carved into stone surfaces. Similar reliefs with funerary inscriptions have been found also in the area around Edessa, notably at Sumatar, a village c. 60 km to the south-east of the city in the Tektek mountains. Although the tombs are of greater importance for the authors of this paper, Sumatar was also a significant religious centre and the “seat of the governors of ‘Arab” (Segal, 1970, p. 23, pp. 56-61; Drijvers & Healey 1999, pp. 40-41).
3Over the last decade, extensive research on funerary traditions in the Near East has been conducted and published (i.e., Blömer & Raja, 2019a; de Jong, 2017; Raja, 2022; Bobou & Raja, 2023). This new research has underlined that local and regional traditions encountered in the funerary spheres were strong and persisted across centuries. These traditions, which included the depiction of the deceased and their family members, living or dead, in the form of sculpted, painted, or mosaic portraits, are often highly recognizable in their styles and can be ascribed to a city or a region just on the basis of their looks. However, it is also possible to disentangle various elements pointing to outside influences, such as Roman styles or styles imported from the East, reflected in the representation of hair, clothing fashions, jewellery, or other attributes. Such mixtures have also been observed in the funerary portraits from Edessa, but these have never been studied as a group in their own right. Despite the fact that they are not as numerous as the several thousand found in Palmyra, primarily made of limestone, the funerary portraits from Edessa, make up an important group of material stemming from one place in the Near East. What follows is an attempt to catalogue and present a systematic study of the inscribed funerary portraits of the tombs from Edessa. These have been selected out of the corpus of funerary portraits from the city since they carry information such as dates, that can help establish a timeline of Edessene art and iconographic themes, and genealogies that help elucidate the identities and relationships of the people portrayed. It can also aid in the study of the uninscribed reliefs and statues that may be associated with the funerary sphere and can add more information on the commemorative practices of the Edessenes. The article aims to draw attention to common themes and similar materials from elsewhere in the region and embed the funerary portrait habit of Edessa within its local setting and that of the wider region.
4A wealthy person wishing to construct a tomb in Edessa and its environs had three different options available to him: rock-cut chamber tombs, tumuli, and tower tombs. The most common funerary monument was the rock-cut chamber tomb: almost a hundred were documented in the necropoleis of Edessa (Segal, 1970, p. 27). They consist of one or two connected chambers of relatively small size, sometimes with columns or pilasters or pediments with figural decoration. Inside the chamber, the burials took place inside arcosolia, sometimes in the shape of a kline and sometimes with architectural decoration (Segal, 1970, p. 28; Rumscheid, 2013, p. 110). Other decorations included wall paintings, mythological figures, vegetal and floral motifs, as well as the statues of tomb owners (Önal, 2017, p. 22). The reliefs cat. nos. 8-10 were located on the walls of the chamber above the arcosolia. One tomb had three uninscribed busts inside the chamber (Segal, 1970, p. 28). Some of the rock-cut tombs had mosaics on the floors, and these were placed in the centre of the chamber. From the evidence so far, the tombs either had sculptural, painted, and architectural decoration or mosaics (Segal, 1970, p. 28; Blömer, 2019a, p. 210; see also the publication of recently excavated tombs in Çetin et al., 2020, where the tombs either have paintings or sculptures or mosaics).
5The truly wealthy Edessenes seem to have preferred the tumulus and the tower tomb types for their burials, judging by the expense necessary for both types of monuments. No tumulus burial is known from the city of Edessa, but they were common in the region, especially in the Tektek mountains. They remain largely unexplored and understudied, and range from plain “heaps of rubble” to tumuli with a built base (Blömer, 2019a, pp. 213-214). The tower tombs were not as common and seem to have been limited to the wealthiest members of Edessene society. None has survived from the city of Edessa, but several were recorded in the city’s surrounding area (Segal, 1970, p. 29). No certain funerary portrait has been associated either with a tumulus or a tower tomb. The only possible portrait is a relief showing a reclining figure, from a tower at Dayr Ya‘qub (Deyr Yakup), incorporated in the later monastery of that name, that has been identified as the portrait of a nobleman (Segal, 1970, p. 29). The same relief, however, has been identified as a depiction of the god Tammuz (Hvidberg-Hansen, 2012, with previous bibliography) so it cannot be used as evidence for the presence of funerary portraiture in the tower tombs.
6The evidence from the texts recording the foundation of the tombs reveals some of the beliefs of the Edessenes regarding the afterlife. The tomb is often described as a “house of eternity”, especially in the mosaics (Drijvers & Healey, 1999, cat. nos. As7, As9, Am2, Am3, Am5, Am6, Am7, Am10, Add6), while curses on whoever disrespects the tomb are attested in some of the inscriptions (Drijvers & Healey, 1999, cat. nos. As60, Am5, Bs2; see also Drijvers & Healey, 1999, p. 40, and Salman, 2008, pp. 110-113). The inscriptions also reveal how the Edessenes emphasized their local identity through language: most of the texts are in their local dialect. Only one tomb, at Deyr Yakup mentioned above, has a bilingual, Greek and Aramaic inscription (Drijvers & Healey, 1999, cat. no. As62, with previous bibliography).
7Typologically, the reliefs can be divided into two groups: free-standing stelae and wall-carved reliefs, while the mosaics were all placed on the floor of rock-cut tomb chambers. The survey of their find locations shows that each type had a different function in the funerary context. The stelae cat. nos. 1-5 are associated with Sumatar: cat. nos. 1-4 were found there, while cat. no. 5 may also be from the area. The stele cat. no. 1 was found near the road leading to Sumatar (Drijvers, 1973, p. 2) and the stele cat. no. 2 was found either in situ or very close to its original location in Sumatar village, placed “in front of a stairway cut in the rock which led to a cave-tomb” (Drijvers, 1973, p. 6). Stelae cat. nos. 3 and 4 were also found in Sumatar, but cannot be associated with a specific tomb (stele cat. no. 3: Drijvers 1993, pp. 148-152; Laflı, 2016, p. 444; stele cat. no. 7: Güler, 2014, p. 33). Stele cat. no. 5 was probably also found in Sumatar (Drijvers, 1980, p. 20). The other two stelae are associated with Urfa: stele cat. no. 7 was recorded in Urfa, but its provenance is unknown (Segal, 1959b, pp. 39-40, pl. 6; Laflı, 2016, p. 446) and stele cat. no. 8 was also recorded in Urfa before being taken to the museum at Diyarbakir. It too is of unknown provenance (Segal, 1959b, p. 40, pl. 7; Drijvers & Healey, 1999, p. 208, cat. no. Cs2).
8Stele cat. no. 2 gives insight as to the placement and function of this group of monuments: it served as a commemorative marker. However, as the cave-tomb near which it was found had no inscribed texts, it is not possible to confirm the tempting hypothesis that the stele was associated with that particular grave monument (for discussion of tomb markers in Roman Syria, see de Jong, 2017, passim). This single stele found near a tomb indicates that the other stelae were most likely placed outside tombs, perhaps marking the location of a family grave rather than an individual burial, although it was common for individual burials in simple pits to be marked with a single stele in the wider region (de Jong, 2017, p. 41). The burials, both from the same area and villages around Edessa, but also from the city itself, point to the predominance of family tombs for the elite of the city, either in the form of rock-cut cave-tombs, or in the form of tumuli or towers (for discussion of burial customs in Edessa, see Segal, 1970, pp. 27-29; Blömer, 2019a, pp. 207-214).
9The reliefs that were carved onto the walls of tombs were observed and studied in situ first by Henri Pognon (cat. nos. 9-11), while cat. no. 6 is now free-standing, but scholars propose that it formed part of the carved decoration of a cave-tomb (Drijvers & Healey, 1999, p. 57, cat. no. As6). Cat. no. 9 was inside a cave-tomb at Kırk Mağara south-west of the citadel of Edessa (Segal, 1970, p. 27) and it was placed in a carved niche (Segal, 1970, pl. 25b). Cat. nos. 10 and 11 were found in the Kara Köprü cave tomb a few kilometres north of Edessa. Both reliefs were inside the niches carved within the almost square tomb chamber, one on the left wall and one on the right (Pognon, 1907, pp. 179-183).
10The majority of the funerary portraits were on floor mosaics inside chamber tombs that were cut into the soft limestone that forms the bedrock of Edessa (Rumscheid, 2013; Önal, 2017, pp. 19-22; Blömer, 2019a, p. 210; Demir, 2019; Çetin et al., 2020, p. 121). These had one or two chambers with burials in klinai under arcosolia along the walls, and the mosaic was placed in the centre of the floor (Önal, 2017, pp. 19-22). Most of the mosaics were not excavated properly or were recorded in the art market, so with their original context lost, but from the evidence of those that were found in situ during proper archaeological excavations, it is clear that they were placed in an axial relationship with the corridor leading to the tomb and with the faces looking towards the entrance as if staring at the visitors. This was done regardless of the type of representation: with busts, with banqueting scenes, or with standing figures, indicating that the most important thing was the direct visual engagement between the viewer and the depicted person (see Rumscheid, 2013, fig. 1; Önal, 2017, figs 11-14).
Fig. 1 Mosaic fragments (cat. no. 34)
S. P. Brock
- 1 Compared to Palmyra, where there are 1576 inscribed portraits in 1123 reliefs, statues, and wall-pa (...)
11The funerary portraits can be divided into two large groups according to their typology: (1) sculptural portraits, and (2) mosaic portraits. The majority of the portraits were in mosaics (cat. nos. 12-38), while only eleven sculptures are securely connected to the funerary sphere (cat. nos. 1-11). In total, there are thirty-eight reliefs and mosaics with 122 funerary portraits, and with 118 inscriptions naming the figures represented, without taking into account the inscriptions recording the building of the tomb.1 In one mosaic the inscriptions indicate that the number of portraits was greater than the number which have been preserved (cat. no. 34) (Fig. 1), while in others, there are more portraits than inscribed names (cat. nos. 6, 11, 15).
12The eleven funerary sculptures can be divided into two groups: (1) free-standing stelae (cat. nos. 1-5, 7-8), and (2) rock-cut reliefs (cat. nos. 6, 9-11). The stelae are similar in shape: they are rectangular and divided into two sections. In the three stelae where the top is preserved, the top is flat and straight (cat. nos. 2, 4, 8). Their height varies: stelae cat. nos. 3 and 7 are 111 and 157 cm in height respectively (the height is recorded in Laflı 2016, pp. 444, 446; Segal, 1959b does not give the dimensions) but stelae cat. nos. 1 and 2 have a height of 390 cm and 295 cm respectively (Drijvers, 1973, pp. 6, 2) (Figs. 2-3). The upper section had a recessed, carved niche with a straight (cat. no. 2) or an arched (cat. nos. 7, 8) top, while the lower section was inscribed with a text giving the name(s) of the deceased and the name of their relative who had commissioned the funerary monument. The portraits of the deceased were placed inside the niche. Their number varied from one (cat. nos. 1, 3, 5, 8), to two (cat. no. 7), to three (cat. no. 2). The portraits were in the bust form: the figure is shown frontally from the height of the upper torso, and with the arms bent in front of or next to the chest.
Fig. 2a - 2b Stele with male bust (cat. no. 1)
H. J. W. Drijvers
13This type differs from the traditional Roman bust format, where only the head, the neck, and part of the torso around and sometimes below the collarbone are shown (for these traditional Roman busts, see Motz, 1993; Stewart, 2003, p. 46) While the traditional Roman bust is known from Italy, Greece and Asia Minor, the full torso form is known mainly from the provinces: it was typical of Palmyra, Zeugma, Hierapolis and other cities of Syria and its neighbouring areas, but also of provinces such as Pannonia. In one stele (cat. no. 7), the second figure is shown in full, and at a much smaller scale than the bust of the woman that occupies the centre of the relief (Fig. 4). This schema is known from Palmyra, where it is commonly used for the depiction of children next to their parents or guardians (Krag & Raja, 2016). The relief cat. no. 4 from Sumatar cannot be placed in one of the above-mentioned groups. The stele was found broken in two parts and is now lost (personal communication with S. E. Güler).
Fig. 3 Stele with three female busts (cat. no. 2)
H. J. W. Drijvers
Fig. 4 Stele with female bust and standing woman (cat. no. 7)
J. F. Healey, Şanlıurfa Museum
14The second group comprises reliefs that were carved onto the rock (cat. nos. 6, 9-11). The reliefs cat. nos. 9-11 show a reclining figure with one or two other figures either standing or seated. Finally, detached from a cave wall, relief cat. no. 6 shows two portrait busts, a male and a female. The placement of the reliefs cat. nos. 9-11 that are in their original location makes clear that they functioned as grave markers, as they were carved on the walls above the burials.
15The mosaics cat. no. 13 and cat. no. 24 were found in the same tomb (Rumscheid, 2013), while all the mosaics that have a documented provenance come from areas in the city of Urfa, and thus indicate that this was a very localized phenomenon, even though the tombs themselves resemble those of Zeugma and Doliche in shape (Önal, 2017, p. 19; Blömer, 2019a, p. 210).
Fig. 5 Mosaic fragment (cat. no. 35)
K. Parlasca, Art Gallery of S. Australia, Adelaide
- 2 A fourth category of mosaics, also from tomb floors, with symbolic and legendary scenes (Orpheus, t (...)
16The depictions can be divided into three groups: (1) banqueting scenes (cat. nos. 12-16), (2) figures in bust form (cat. nos. 17-23), and (3) standing figures (cat. nos. 24-29). In addition to these, several fragments of mosaics have made their way to various museums outside Turkey and in the international art market (cat. nos. 30-39) (Fig. 5). Their fragmentary nature does not allow their assignment to one of the groups mentioned above, but they too can offer useful information about the iconography of the mosaics, even if their way of extraction has meant that their context is lost and the contribution that they can make to the history of Edessa reduced.2
17The banqueting scenes are multifigural compositions. In all the mosaics, the centre and the greatest part of the field is occupied by a reclining man – no mosaic has been found with a reclining woman so far. The most common composition has a seated woman to the left of the reclining figure (cat. nos. 12, 14-16), and the other figures depicted standing around the reclining man. In one mosaic (cat. no. 13), two men are shown sharing the banqueting kline, while the woman is shown behind the kline. In all the mosaics where the woman is seated, she is shown on a high-backed chair. The standing figures can be distinguished between active and static attendants: the active ones bring cups or other objects associated with the banquet to the reclining man (cat. nos. 14-16), while the static figures look on (cat. nos. 12-16). In cat. no. 15, one of the standing figures, a woman, is on a small pedestal or stool.
18Bust portraits (cat. nos. 17-23) are similar in all the mosaics: the figure is shown frontally from the waist or the chest up. Only in one mosaic is the hand of the figure shown, and that occurs in two of the figures in cat. no. 17, which also show the bust from the waist up. In the other figures, the arm is shown from the shoulder down to just above the elbow. The busts on the field of the mosaic are usually divided by simple lines that create a square or rectangular frame around the figure, but in two cases, the figures are shown one next to and above the others without being separated (cat. nos. 17, 19). In many cases, individual bust portraits have been removed from mosaics (no doubt as part of looting and for reasons of portability and saleability), but in examples which survive intact, it is clear that the individual busts were often part of a bigger scheme, often a series of portraits of members of a family, with the individual busts fitted into a very distinctive grid pattern (as in cat. nos. 18, 19, 20 and, less clearly, 17).
19The last group is that of mosaics with standing figures (cat. nos. 24-29). In this group, the figures are all shown one next to the other, fully frontal. In cat. nos. 25 and 27, a younger, sub-adult figure is shown behind the older, adult ones, while in cat. nos. 26 and 28, the younger figures are shown in front of the adults.
20Lastly, cat. nos. 30-38 represent fragments that could have belonged to any of the above-mentioned groups. Most of these show heads of figures, men or women (cat. nos. 30, 31, 35, 37, 38), but there is one fragment (cat. no. 32) that shows the upright, upper part of a sub-adult located next to a hand holding a flower, and two showing young children, one shown reclining on a bed (cat. no. 33), and one standing (cat. no. 36). Most of these were recorded in the art market (only cat. nos. 31, 32, 35, 38 are currently in museums), and even though they originally must have belonged to multifigural groups, it is impossible to say anything about the original compositions.
21Of the thirty-eight reliefs and mosaics, only three are securely dated by an inscription: cat. no. 9 to AD 201/2, cat. no. 14 to AD 218 (or 228), and cat. no. 28 to either AD 253 or AD 283 (Caillou & Brelaud, 2016). In addition to these, two mosaics depicting Orpheus can be dated with precision to AD 194 and AD 228, and a mosaic with a Phoenix was made in AD 236. The others are dated either by stylistic or palaeographic criteria (see Güler & Çelik, 2002; Önal, 2017, in passim).
22The single relief and the five mosaics dated by inscription give us fixed stylistic and iconographic points on which to anchor the chronology of the other reliefs and mosaics. Cat. no. 9 is not well-preserved but it shows that the banqueting motif was already fully formulated by AD 201/202 (Fig. 6). Best-preserved for the consideration of style are the clothes of the figures: the body is not emphasized under the voluminous fabrics (for example, the woman’s breasts are implied only by the presence of deep, V-shaped groves in the middle of the torso), and the folds are indicated by shallow, mostly linear grooves.
Fig. 6 Relief with banqueting scene (cat. no. 9)
J. B. Segal
23The mosaic cat. no. 14 is not preserved and survives only through the drawings made by Mrs. Seton Lloyd (Fig. 7). As such, it is possible to make only a few observations about the style and colour palette used in the mosaic. The pupils are rendered touching the upper edge of the eye only, and only a few folds are indicated. The main colours used for the garments are pink, yellow/ochre and silver/grey, with darker red, red/brown and dark grey/black used to render the folds. Cat. no. 28 is another lost mosaic, preserved only through a drawing from 1881 (Fig. 8). The only possible observation regarding the style that one can make is that the eyes of the figures appear to be large.
Fig. 7 ‘Funerary Couch’ mosaic (cat. no. 14)
After Segal, 1970, pl. 2
Fig. 8 Sketch of mosaic with female figure and children (cat. no. 28)
After Renan, 1883, 250
24The first of the two mosaics with Orpheus, dating from AD 194, shows the figure of the singer wearing clothes that, while not revealing, allow for a sense of the body underneath. The folds, however, are mostly rendered by linear and slightly curving grooves (see Önal, 2017, pp. 30-31). The second mosaic is much more schematic in the rendering of the clothes and the human form underneath (see Önal, 2017, p. 32). While the Phoenix mosaic cannot be used for comparisons between the figures, the geometric patterns on the frame can be used for comparisons with other such frames, especially that of the mosaics cat. nos. 25 and 27.
25The securely dated relief and mosaics give very few fixed points for establishing the chronology of the other monuments. In addition to their small number, the difference between media and workshops means that one must be cautious when using the evidence from the reliefs to date the mosaics, and vice-versa. This becomes clear when one compares the very different ways that the human form is depicted in the relief cat. no. 9 and the earlier of the two Orpheus mosaics. Even though they are separated by six or so years only, the figure of Orpheus is inspired by classical Greco-Roman models. This is clear in the rendering of the folds under the right thigh and over the left knee. In contrast to that, the clothes of the figures of the relief cat. no. 8 do not emphasize the body, while the folds that fall over the right knee of the reclining figure are more in common with the folds on the clothes of reclining men and women in Palmyrene art rather than that of Imperial Rome.
26The second Orpheus mosaic, however, has a stylized way of presenting the clothed human body through a combination of geometric motifs that has more in common with the way that the body is depicted in the other funerary mosaics. The frame of the Orpheus mosaic with the geometric motifs is also comparable to that of mosaics cat. no. 17, 23, and 25. Thus, while the comparison between mosaics and reliefs must be done with caution, it is possible to see evidence of common stylistic and iconographic choices in the mosaics regardless of their subject matter (mythological versus portraits).
27The other monuments have been dated primarily on the basis of association with dated inscriptions (many of which have no connection with tombs) and, less certainly, of palaeography to the period from the middle of the 2nd century AD to the 3rd century AD (see Drijvers & Healey, 1999, pp. 16-19 for the form of letters and problems of dating).
28The earliest of the monuments are the reliefs cat. nos. 1-3, that were probably created around the middle of the 2nd century AD. The earliest of the mosaics are cat. nos. 12-13, dating at around the end of the 2nd and the beginning of the 3rd century (though one of the Orpheus mosaics, also from a tomb, dates to AD 194, indicating that the practice of using mosaics inside the tombs had started before the use of mosaics with portraits). Most of the funerary monuments, however, date from the 3rd century AD (Drijvers & Healey, 1999, p. 17; Blömer, 2019a, p. 211).
29Regardless of whether the figures were depicted in reliefs or mosaics, alone or with their families, in banqueting scenes or standing against an empty background, their iconography, as far as it can be ascertained from the well-preserved reliefs and mosaics, was the same. They have large eyes and serious expressions that can perhaps be associated with trends in Severan art (Önal, 2017, pp. 22-23). Their clothes follow local, rather than Greco-Roman fashions, as shown by the comparison between the clothes of the figures in the funerary and in the mythological mosaics (Önal, 2017, p. 23).
30Men are depicted wearing long-sleeved tunics and trousers that are fairly loose. A mantle or chlamys may be worn over the tunic (cat. nos. 1, 3, 10, 11, 17, 18, 26, 29) that, in two cases, is shown fastened over the right shoulder with a round brooch (cat. nos. 3, 29) (Fig. 9). In one mosaic, cat. no. 25, the male figure is shown wearing a long overcoat, while the other three men seem to be wearing either short overcoats (Segal, 1953, p. 117) or tunics with varied and colourful decoration in their upper part (Leroy, 1957, p. 316). When the waist and the lower body are shown, the tunic is almost belted, except in the two figures in cat. no. 15. Two types of shoes are worn: the first is a boot that covers the ankle and the calf (cat. nos. 10, 14, 15, 29), and the second is a closed type of shoe that is tied around the ankle (cat. nos. 25, 26) (Salman, 2008, p. 108).
Fig. 9 Stele with male bust (cat. no. 3)
J. F. Healey, Şanlıurfa Museum
31There is a difference, however, in the attributes and headdresses of the men in the reliefs and in the mosaics: in two of the reliefs, the man is shown holding a sword (cat. nos. 3, 5), while the reclining men in reliefs carry a dagger attached to their belt (cat. nos. 4, 9, 10), and hold a bowl (cat. nos. 4, 10) with the fingertips of the upturned left hand (Fig. 10). None of the standing men in the mosaics have swords or daggers. In cat. no. 14, one of them holds a piece of cloth, identified as a napkin (Segal, 1959a, p. 157), and another a small container, perhaps a spice box (Segal, 1959a, p. 157). In cat. no. 16, the standing man holds a bowl, in cat. no. 15 an object that is perhaps a piece of fruit, and in cat. no. 26 he holds a flower.
Fig. 10 Relief with banqueting scene (cat. no. 10)
H. J. W. Drijvers
32The reclining men are shown holding either a footless cup with two handles (cat. no. 14), or a bowl (cat. no. 15). In the reliefs cat. nos. 4, 9, 10, where the head survives, although it is not very well-preserved, the male figures wear no headdress, while in the mosaics, the reclining men are shown wearing a headdress, usually a turban (cat. no. 14) or a conical cap whose end falls backward, a ‘Phrygian’ cap (cat. no. 16) (Segal, 1953, p. 117). Standing men or men shown in bust wear the ‘Phrygian’ cap (cat. nos. 17, 18, 20, 24, 25) or the turban (cat. nos. 20, 24, 25, 37) (Fig. 11), or they have their hair uncovered (see Rumscheid, 2009; 2013, pp. 126-127; Salman, 2008, pp. 108-109; Önal, 2017, p. 24 for male iconography).
Fig. 11 Mosaic fragment (cat. no. 37)
K. Parlasca
Fig. 12 Mosaic fragment with woman (cat. no. 38)
© Musée du Louvre, dist. RMN-Grand Palais / Paul Veysseyre
33Women wear a tall, conical headdress (Fig. 12), over which falls a long veil that reaches to the ground, a long-sleeved tunic, and a long himation. The himation leaves the arms free and is fastened over the left shoulder, falling diagonally over the chest, covering the area of both breasts and is wrapped around the body and reaches to the ground. Only in three examples, one relief (cat. no. 7) and two mosaics (cat. nos. 26, 30), is the headdress under the veil cylindrical with a flat top, while in one relief (cat. no. 11) and one mosaic (cat no. 18) the women wear a soft cap under the veil. In several of the reliefs and the mosaics, the women wear necklaces (cat. nos. 4, 9, 11, 13, 15, 38), but brooches (cat. nos. 4, 16, 17, 21) and bracelets are rarer (cat. nos. 7, 9, 10, 25). Where their feet are shown protruding from under the himation, they wear closed shoes. Most of the women hold no attributes: only in cat. no. 26 does the woman hold an object in her hand, identified as a mirror (Segal, 1959a, p. 154), a spindle (Segal, 1970, pp. 38-39), or a distaff (for female iconography, see Rumscheid, 2009; 2013, pp. 127-128; Salman, 2008, p. 108; Önal, 2017, p. 24).
- 3 It is tempting to suggest that this group and the next comprise children up to seven years of age, (...)
- 4 Also tempting is to identify in this group and the next children around eleven, twelve, or fourteen (...)
34The iconography of sub-adults is varied. None of the inscriptions give the ages of the children depicted, but the images show that children could be differentiated from each other based on height, facial features, and clothes. Thus, it is possible to identify the following groups in the mosaics according to the relative age and gender of the children: (1) infants and very young boys: they are shown wearing a tunic with a broad belt and ‘Phrygian’ cap and they are small in size compared to other figures in the scene; two of these (cat. nos. 32, 33) are shown next to adults, although their relation to the adult figure is not obvious: in cat. no. 32, a hand with a flower is located behind the child, while in cat. no. 33, there is a hand placed over the child’s left shoulder (Figs. 13-14); (2) young girls who are able to walk but are not yet approaching puberty:3 they are shown wearing a tunic with broad belt, ‘Phrygian’ cap and closed shoes: cat. no. 26; (3) young boys who are able to walk but are not yet approaching puberty: they are shown wearing a belted tunic, trousers, boots and a ‘Phrygian’ cap: cat. no. 26; (4) young boys approaching puberty:4 they wear the same costume as adult men but without a headdress: cat. no. 14; (5) young girls, approaching puberty: they wear a tunic with a broad belt, and have their hair arranged in three buns, one on top, one to the left and one to the right of the forehead: cat. nos. 25, 27, 31 (Fig. 15).
Fig. 13 Mosaic fragment with boy (cat. no. 32)
© Musée du Louvre, dist. RMN Grand Palais / Paul Veysseyre
Fig. 14 Mosaic fragment with child (cat. no. 33)
K. Parlasca
Fig. 15 Mosaic fragment with girl (cat. no. 31)
© Musée du Louvre, dist. RMN-Grand Palais / Paul Veysseyre
35One mosaic fragment showing a girl with the same type of costume but without the three buns may also represent a girl of this age-group: cat. no. 36 (Fig. 16); (6) boys who have reached puberty. This group can be subdivided into two: (a) boys who are beardless, and shown wearing a tunic (cat. nos. 18, 21b, 22a, 27), and (b) boys with facial hair but not fully grown beards. In cat. no. 15, it is possible to see clearly how facial hair was used for distinguishing between these older, yet not fully mature men: the son on the far right is shown with a moustache, he wears a belted tunic, trousers, and boots, and seems to be holding a piece of fruit; the son in the middle wears an unbelted tunic, a turban, and has a moustache and partial beard; the son on the left wears an unbelted tunic and has a small moustache. It may be significant that the man with the most facial hair is also the one with the darkest, ruddiest complexion, while the face of the son with the moustache is rendered mostly with white and a few red tesserae, while the one with the sparest facial hair and the small moustache has the palest complexion of the three. In cat. nos. 20 and 29, the facial hair is again used as a marker for separating older and younger men, in cat. no. 20 from two different generations (that is, father and son), in cat. no. 29 of the same generation (two brothers) (Fig. 17). Finally, the last group (7), is that of girls who have reached puberty, possibly near the age of marriage: they are shown wearing the headdress and costume of adult women: cat. nos. 7, 8?, 14, 15, 25 (see also Salman, 2008; Önal, 2017, p. 25).
Fig. 16 Mosaic fragment with girl (cat. no. 36)
K. Parlasca
Fig. 17 Mosaic with sons of Bar‘ata (cat. no. 29)
K. Parlasca
36There is great variation regarding the placement of the figures in group compositions. In cat. no. 7, Šalmat, who is an adult, is shown in a bust portrait, while the daughter is shown as a full figure but of much smaller proportions: she is as tall as the bust of her mother. In banqueting scenes, the reclining man takes pride of place. In the reliefs, he is the only figure to be shown consistently in full, while there is variation in the depiction of the other figures: in cat. no. 9, all the figures are in full, but in cat. nos. 10 and 11, the woman is shown seated but the lower half of her body is condensed and appears squat (that is especially visible in cat. no. 11; in cat. no. 10, it appears as if the body is emerging out of a rounded pillow) (Fig. 18). The greatest variation appears in the mosaic representations. In the banqueting scenes, the couch with the reclining man takes up the centre of the mosaic. In cat. no. 12, the tomb’s founder is the son of Barsimya, however, the three figures whose relationships are clearly indicated by the surviving inscriptions are Severus and Barnabu, sons of Aptuḥa, and Šalum, Barnabu’s daughter. Thus, there are two generations of the same family depicted, whose connection to the founder is no longer identifiable. The clearest depictions are that of Barnabu and his daughter: he is depicted reclining on the couch, while his daughter is seated on a high-back chair to his left. It is not clear how the other figures were positioned around, or on, the banqueting couch. In cat. no. 13, all three family members are shown on the couch, the founder’s wife seated, with her son and her husband reclining. Both father and son have full beards, but the father is differentiated further through the wearing of a turban. In cat. no. 14, the tomb’s founder, Zaydallat, reclines on a couch and takes centre stage; his wife sits on a high-back chair to his left. Behind and in front of the couch, their children are attending to their father. One brings a cloth, another carries what is most likely an incense burner, and the daughter brings a small object, perhaps a fruit or a flower. One of the boys, Baršalma, places one hand behind his father, in a supportive or affectionate gesture. A similar scene is played out in cat. no. 16: the patriarch of the family, depicted with a grey beard, is shown reclining in the centre of the composition, while a woman is seated on a high-back chair to his left and two younger men are attending to him, one bringing a bowl. Even though the relationships of the figures are not specified, it is likely that the reclining Garmu is the husband of Atu, identified only as the “daughter of Waʾel”, who is seated next to him, and the two young men are their sons. A similar relationship must have existed between the figures in cat. no. 15: the surviving inscriptions do not allow us to identify the relationships of the figures, but it is likely that Barnabu was the founder and was depicted next to his wife and children. His ‘wife’ is seated on a stool, while the younger members of the family are shown standing and, with the exception of the youngest person in the scene, they are placed close to each other, with Qami standing on a low footstool. The youngest person is shown in front of the couch, holding a small vessel in his hand, perhaps a drinking cup.
Fig. 18 Relief with banqueting scene (cat. no. 11)
H. J. W. Drijvers
37In the mosaics with busts, the figures are almost always inserted within a frame composed of black tesserae, except for cat. no. 17. In cat. nos. 17-19, the tomb’s founder is placed on the upper register, as testified by the additional inscription recording the tomb’s construction. The person next to him, however, may be the king (certainly a figure of authority), with the founder’s father placed next to that central figure (cat. no. 17), or his daughter (cat no. 19). In cat. no. 17, the figures are placed next to each other in two rows (Fig. 19). The portrait of Abgar occupies the central position; he may be identified with Abgar the Great, although an identification with Abgar X has also been proposed (Drijvers & Healey, 1999, pp. 187-188), or not a king at all. The tomb’s founder, Barsimya, is represented next to and slightly behind Abgar, thus visibly showing his support to the king, his “lord and benefactor” (as the inscription says). On the other side of Abgar is the portrait of Barsimya’s father. Unlike the other figures, his bust is shown from the upper chest upwards, in order to accommodate the full bust portrait of ‘Azil, Barsimya’s mother (and presumably Ašadu’s wife, even though she is not named as such in the inscriptions), who is shown wearing a tall headdress. Her position within the mosaic is extraordinary, as she is taller than all the other men in the relief, and perhaps indicates her status as matriarch of the family. In cat. no. 18, the relationship of Šumu to the founder, whose portraits are placed next to each other, is not identified. Below them, are two young men and, in the third register, two young women. Their relationship to the founder Aptuḥa is also not made explicit in the inscription. One of the young men, however, is called Garmu, and has the same name as Aptuḥa’s father, so it is likely that he is the founder’s son. If we accept this identification, then the second young person may also be a son of Aptuḥa. Because both men in the middle register are shown as beardless, the women in the lower register are most likely their sisters, since the evidence from the inscriptions points to maturity, reflected by the presence of a full beard, as a marker of married men. In cat. no. 19, the founder Balay is placed in the upper register (Fig. 20); according to the foundation inscription, his tomb was made for his children and for his heirs, but the inscribed portraits reveal that the tomb was used for the extended family, and not just his direct descendants. According to these inscriptions, Balay also has his sister Šalmat depicted, his two children, Barnabas and Ani, the mother of Ani, Šalmat, as well as Arḥemta with her son Abšay and his wife Qiṣat and several individuals whose relationship is not specified. In cat. no. 21, where the relationships of the figures are clear, the parents are placed in the upper register, while the son and his wife are in the lower. Their placement also seems to be divided according to gender and family lines: the son is placed under his father, while the two wives are both placed on the left side of the mosaic. Cat. nos. 21 and 22 are detached panels from mosaics with busts.
Fig. 19 ‘Abgar Mosaic’ (cat. no. 17)
H. J. W. Drijvers
Fig. 20 Mosaic with busts (cat. no. 19)
J. Euting Diary, Tübingen University Library
38Only two tomb foundation inscriptions are associated with mosaics with standing figures (cat. nos. 26 and 28), with a third possible foundation implied in cat. no. 27. Cat. no. 26 shows two standing men, of which one is named as Gabbay, while the inscription accompanying the other is not preserved (Fig. 21). The tomb’s founder, however, is named Adona, son of Gabbay, indicating that the second standing man must be Adona. The two men are standing next to each other and at almost the same level, with the foot of Adona placed slightly more forward than that of Gabbay. The women that appear behind them are not named, and the one behind Adona seems to be closer to him and perhaps even embracing him. Adona’s two children are placed in the foreground. This may indicate that the tomb was constructed by Adona on the occasion of the death of Gabbay, Adona’s father, for the use of Adona’s family. Such circumstances surrounding the tomb’s foundation may also explain the epigram’s focus on mourning one’s forefathers. Cat. no. 27 shows two men with full beards, with three younger persons next to and between them. The inscription places particular emphasis on Ruma, by the use of the formula “may he be remembered”. He is also shown as slightly forward, compared to the other adult male figure in the mosaic, Barhadad. What is perhaps significant is that Barhadad places his hand behind Ruma; in the other mosaics, this supportive gesture only appears in the mosaic cat. no. 17, where Barsimya places his hand behind Abgar. This may be another subtle indication that Ruma is the founder and the more important man of the two. Cat. no. 28 survives only in a drawing; the founder’s figure was not preserved even at the time that the drawing was made, and even though the female figure’s iconography conforms to that of other women in Edessene mosaics, that of the two children cannot fit any of the iconographic groups for sub-adults. Either the scene is unique, or the person who drew it took liberties with the scene (either because the mosaic was not well-preserved, or because he did not decipher the garments well or for some other reason).
Fig. 21 ‘Tripod Mosaic’ (cat. no. 26)
After Segal, 1970, pl. 3
- 5 Similar branches can be seen in the hands of men in Palmyrene reliefs, though their meaning there i (...)
39The relationships of the figures in cat. no. 24, cannot be separated from those in cat. no. 13, since both mosaics were located in the same family tomb. There, the reclining father was named as Ma‘na. In cat. no. 24, two sons of Ma‘na are named, with two women, most likely their wives, and another young relative, Gadya. In cat. no. 25, Moqimu is shown surrounded by his children, his wife and his granddaughter Šalmat (Fig. 22). He has a full, mostly grey beard, and wears a distinctive costume and headdress. He also holds a (so far) unique attribute in the mosaics, a small, green branch with two leaves.5 His wife is also shown with grey hair. Of his sons, one is shown in a ‘Phrygian’ cap, and the others are bare-headed, with grey-black hair and beards. His daughter and granddaughter have black hair. Cat. no. 29 is a fragment of a larger mosaic, showing three brothers. Two are well-preserved: one completely beardless and one with a slight beard. A third person, most likely an older brother, was located in front of him, as indicated by the remains of a mainly pink tunic.
Fig. 22 Mosaic with family of Moqimu (cat. no. 25)
After Segal, 1970, pl. 1
40A similar gesture to the supportive one is that of the ‘embrace’ that appears mostly in the mosaics with standing figures (cat. nos. 15, 25, 26, 29): one figure either has one hand behind the figure to his left, and with the other hand holds a small object (flower or fruit) in front of the figure to his left (cat. nos. 15, 26), or holds the arm of the figure to his left (cat. nos. 25, 29). The gesture differs from the ‘support’ gesture in that the figure’s hand is placed in front of the figure next to him, instead of in front of his own chest, and is used, perhaps, to indicate affection between equals, since, based on the inscriptions, it appears between siblings (cat. nos. 25, 29). The inscriptions do not allow us to identify the relationship between the figures in the ‘embrace’ gesture in cat. nos. 15 and 26.
41The personal names of the early Syriac inscriptions (including both the texts mentioned in this article and the other texts not accompanied by images, notably the early Syriac parchments and papyri) have recently been subject to a detailed study (al-Jadir, 2021). It may be noted that theophoric names formed as genitive compounds (such as “Son-of-DIVINE-NAME”, “Slave-of-DIVINE-NAME” and “Handmaid-of-DIVINE-NAME”) are common. Less common are names formed as sentences (“DIVINE-NAME has given”). The following are attested in the items catalogued in this article (grouped here by divine name):
42ʿAbdallat, Zaydallat
43ʿAbdnaḥay, Amatnaḥay
44ʿAbdšamaš, Baršamaš, Baršama (hypocoristicon), Lišamaš, Šamašyahb
45Baralaha, Bartalaha
46Barʿata, Šelamʿata
47Barsimya, Batsimya
48Other theophoric names in the corpus are ʿAlbel, Barbaʿšamin, Barhadad, Barkalba, Barnabu and Matʿuzzat.
49The difficulty in interpreting this evidence lies in the fact that names can become traditional in families and communities and remain popular even when little religious significance is attached to them. Cat. no. 3 provides a case in point. It contains two Šamaš names (Lišamaš and Šamašyahb), which might indicate devotion to the Sun God, but the inscription goes on to mention that the stele is under the protection of Sin, the Mood God. Bel and Sin were popular in name-formation, though no Sin-names appear in the catalogue here (for these see al-Jadir, 2021). We can, at least, say that personal names alluding to Allat, Bel, Naḥay, Nabu, ʿAta, Simya, Sin and Šamaš were popular.
50In theory, names might also provide us with the opportunity to make family links between inscriptions — by identifying individuals mentioned in different texts. In fact, the dataset is too small to allow this, though this aspect comes to the fore in a tomb containing two chambers which has been documented. Cat. nos. 13 and 24 (separated here because of genre criteria) are in fact linked both because they are from different chambers within a tomb-complex and through the texts in each chamber. The main figure in cat. no. 13 in the inner chamber is Maʿna, son of Baršamaš, while the main figure in cat. no. 24 in the outer chamber is Barbaʿšamin son of Maʿna (depicted alongside his brother Barkalba). It is reasonable to assume that Maʿna is to be identified as the founder of the family tomb consisting of two chambers. In due course, the outer chamber was taken over by Maʿna’s son, Barbaʿšamin. (See catalogue entry 24 for reasons for departing from Voigt’s interpretation [2013]).
51Most of the monuments, regardless of their material, have multifigural compositions. Only the stelae cat. nos. 1, 3, 4, 5, 8, show a single individual. Thus, the imagery emphasizes the bond between the figures, while the inscriptions accompanying the portraits clarify the relationships between them. Even in the reliefs with single figures, however, the inscriptions highlight how the deceased is not alone, but he or she is placed within a family network, identified as a son or daughter or wife. The inscription, where fully preserved, also records the name of the person who commissioned the monument and their relationship to the depicted figure. This information is preserved fully in cat. nos. 1, 3, 6, 9, 10, 12, 14, 15, 17-19, 26, 28 and partially preserved in cat. nos. 4, 8. Thus, even when the image seems to isolate the deceased, the text re-inserts him into the family unit and into the family “house of eternity”, as the tomb is called in the inscriptions of cat. nos. 12, 15, 17-19, 26, 28.
52Of direct interest in relation to the funerary portraits is the fact that reference to the builder of the tomb is one of the most common features of the inscriptions. This information appears in several mosaics with banqueting scenes or busts in a dedicatory panel which is separate from the ‘labels’ that attach names to individuals pictured in the scene (see cat. nos. 12, 14, 18, 19). These texts have certain typical elements: a date (in some cases) and a statement of who made the tomb (called a byt ʿlmʾ, literally “house of eternity”). Thus, we find this formula in inscriptions: Drijvers & Healey, 1999, cat. nos. As7-9, As20, As21-22, As55c, As56, As59, Am6, Am7c, Am9, Bs2, Cs1, Add1, and in more recent finds, such as those in Healey, 2006; Apamaea inscription 9: Desreumaux, 1999, pp. 99-100); Tell el-Magāra (Teixidor, 1998); and Healey & Gioia, 2023). Typical phraseology, using cat. no. 18 (Drijvers & Healey, 1999, cat. no. Am2) as an example, is “I, Aptuḥa son of Garmu, made for myself this house of eternity, for myself and for my sons/children and for my heirs for ever”. In the mosaics with standing figures, however, these inscriptions are almost completely absent; only in cat. nos. 26 and 33 is there an explicit reference to the tomb founder, and the inscriptions are not placed within frames but rather ‘float’ on the surface of the mosaic like the other inscriptions. The evidence from the mosaics that were found in recent excavations gives different possible explanations for the lack of the foundation inscription. One possibility is that the mosaics came from tombs with multiple chambers, where the founder’s name would have been inscribed in a different mosaic, as was the case with mosaics 13 and 24, where the foundation inscription was in the mosaic with the banqueting scene (cat. no. 13). Another is that the foundation was placed in a different location inside or outside the tomb. The re-use of the cave tombs as storage spaces or houses would explain the disappearance of a foundation inscription (e.g., cat. no. 27, that was found in a tomb used as a storage room: Drijvers, 1993, pp. 154-155; Önal, 2017, p. 51). A third possibility is that the formula “may he be remembered” which appears in cat. no. 27 next to the portrait of the man named Ruma, may also imply the foundation of the tomb by him.
53Even though the patriarch of the family was the one usually responsible for the construction of the tomb or monument (identified with certainty in cat. nos. 1, 4, 5, 8, 12, 14, 15, 17, 18, 19, 26, 28; likely in cat. nos. 13, 23, 24, 25), one monument was commissioned by the brother of the deceased (cat. no. 3: Drijvers & Healey, 1999, cat. no. As60), while cat. no. 11 (Drijvers & Healey, 1999, cat. no. As6) was commissioned by the husband of the deceased. The relationship of Dardu, the sculptor, with most of the figures mentioned in the inscription, cannot be reconstructed with certainty in the case of cat. nos. 10 and 11, except for Matʿuzzat, who is identified as his aunt by the inscription (see Drijvers & Healey, 1999, cat. nos. As12-14). The evidence from the corpus of Edessene inscriptions shows that other commissioners were possible: for example, one of the tombs was commissioned by a woman (Drijvers & Healey 1999, cat. no. As20, for which there is no evidence of any portraits). In some inscriptions, the commissioner was not named: this is the case in cat. no. 7, as well as cat. no. 5, where he may be identified as the father of the deceased, but his name is not given.
54The inscriptions may also hint at complicated family structures. So, while cat. no. 7 commemorates a mother and daughter, the commissioner is not named and the mother, Šalmat, is identified as the “daughter of Maruna”, rather than the ‘wife of X’. In addition to this epigraphic puzzle, a small carved protrusion to Šalmat’s right may be indicative of yet another diminutive figure that had been placed beside her. In cat. no. 16 also, the woman is only identified as “daughter of Waʾel”. These inscriptions prioritize the woman’s relationship with her paternal home rather than that of her husband. In cat. nos. 17 and 19, the woman is identified as a mother rather than a wife. Cat. no. 6 commemorates Qaymi, daughter of Arku, while ‘Abdallat’s name is placed next to hers. Another inscription, either in a non-surviving part of this monument or in the same tomb, would have clarified the relationship between Qaymi and ‘Abdallat.
55This difficulty in identifying relationships appears in mosaics as well. In cat. no. 16, while all the figures are identified by an inscription, there are no mentions of their relationship, and the woman is identified as the “daughter of Waʾel”, whose name does not appear in connection with any of the figures. In cat. nos. 18 and 19, several of the figures are named but their relationship to the other figures is not elucidated by any inscription. This is also the case for cat. no. 27: the two adults are named but they are both identified as sons of different fathers, while only one of the children is identified as the daughter of one of the two men, Ḥata daughter of Ruma.
56The inscriptions indicate that trusted household servants could also be commemorated in the mosaics: in cat. no. 12, one of the figures is identified as a “maidservant”, and in cat. no. 23 as a “steward”. Thus, the inscriptions on the mosaics and reliefs of Edessa hint at the complexity of the urban elites that created them in terms of family structures and members of the household. They also demonstrate how individual families prioritized different relationships, and how these varied from tomb to tomb.
- 6 A fourth sculptor known from inscriptions may be Man’u, according to Drijvers’ restoration of the i (...)
57The relief cat. nos. 10 and 11 preserve the name of one of the few sculptors known from the Edessene epigraphic record: Dardu (?). In cat. no. 10, he identifies himself as the nephew of Matʿuzzat, for whom he has also made the monument. The Syriac word used for her is ḥaltā, which means “maternal aunt” (Drivers & Healey, 1999, cat. no. As13). The reference to their family relation, might separate Dardu (?) from the other two sculptors mentioned in the inscriptions, Haššay (Drivers & Healey, 1999, cat. no. As55) and Šila (Drivers & Healey, 1999, cat. nos. As47 and Add3), in that he may also be involved in the commissioning process of the monument. Unlike him, Haššay and Šila are mentioned as carvers or sculptors of monuments “made by” others, that is, commissioned and paid for by Zarbiyan, Wa’el and …Malik (Drivers & Healey, 1999, cat. nos. As55, As47, and Add3 respectively).6
- 7 Abadie-Reynal, 2002, p. 769, proposes that they are two different individuals because of the stylis (...)
58Unlike sculptors, none of the mosaicists who worked on the funerary mosaics of Edessa have left their signatures, and the monuments are all attributed to the people who paid for them. Perhaps this reflects the higher social standing that sculptors may have had in Edessene society and their own awareness of their importance as creators of sculpted images. Another explanation may be that, while sculptors could work as individuals, mosaicists worked as members of a team, or workshop, and so it was more difficult for an individual to take credit for their collective work. A third reason may have to do with the signature habit of artists in general: while sculptors tended to sign their works, mosaicists rarely signed their creations in antiquity. In the Levant, in particular, the mosaicists’ signatures are more often found in synagogue and church floors from the 5th to the 8th centuries AD rather than in private houses (see Dunbabin, 1999, pp. 271-273). Only one signature of a mosaicist is known from Edessa, that of Barsaged (or Barsagar), who calls himself a “mosaicist”, rṣwpʾ in Syriac (Healey, 2006, especially pp. 323-324). Another signature is in Greek, that of Zosimos: it appears on the Synaristosai mosaic from a house in Zeugma (Önal, 2002, p. 61) while a Zosimos of Samosata, probably the same individual, signs another mosaic at Zeugma (house of Pasiphae) with a depiction of Aphrodite (Aylward, 2013, p. 16; Lepinski & Rousseau, 2022, p. 234).7 Another possible mosaicist active at Zeugma is Theodotos, whose name is recorded in a fragmentary inscription (Dunbabin, 2013, p. 162). A now lost bilingual inscription from Tell el-Masʿūdiyyeh carries the signature of a mosaicist in the Greek part of the text: Eutyches states that “he made it” (Parlasca, 1983; Drijvers & Healey, 1999, cat. no. Bm1).
59While itinerant workshops have been proposed as the propagators of classical Greek and Roman myths and motifs in mosaics in the East (see Dunbabin, 1999, pp. 273-274), the craftsmen of the Edessa funerary mosaics were most likely from the region, and if they travelled, they probably did so within the greater area. The difference between border motifs – widespread across the Roman Empire – and the local iconography of the depicted figures, has made scholars suggest that the craftsmen were familiar with Greco-Roman mosaic art but adapted it for a local clientele (Dunbabin, 1999, p. 173). The discovery of a villa in Shioukh al-Tehtani with mosaics with mythological scenes made in a style reminiscent of the funerary mosaics indicates the presence of, at least, one local workshop active in the 3rd century in the city of Edessa, and shows that the Edessene craftsmen were familiar with the themes and motifs popular in the Greco-Roman East, and especially in places such as Antioch on the Orontes, but preferred to work in their own style for their local customers (for the house, see Abdallah et al., 2020; see also Blömer, 2019a, pp. 212-213). The alternative hypothesis would suggest the co-existence of two different groups of artists active in the region: the mosaics were created over a period of over sixty years and there are significant gaps in our knowledge (see Healey, 2007, pp. 10-11).
60None of the surviving sculptures retains any traces of pigment, although it is likely that they were painted, and so the mosaics and frescoes are our best source for the use of colour in Edessene art. In the mosaics, which are the focus of this paper, the background is white, so that the figures in their colourful garments can be better viewed. Perhaps the most popular colour is yellow; it appears on garments in cat. nos. 12, 14-19, 20-21, 24-27, 29. Blue appears almost as often in cat. nos. 12, 15-19, 20, 24, 26. A variation of blue seems to be the blue-grey, where the folds are in a darker colour, but the rest of the fabric is rendered by greyish tesserae (cat. nos. 14, 15, 18, 23, 25, 27, 29). Close to blue is purple, appearing in cat. nos. 15-19, 20, 26. Red appears in cat. nos. 12, 15, 18, 20, 22. A lighter variation is pink, appearing in cat nos. 12, 14, 15, 17, 25, and a warmer variation is orange, appearing in cat. nos. 17, 19, 25, 26. Green is not common, appearing only in cat. nos. 19, 25, 26, while white is used for a garment only in cat. no. 25. Thus, the Edessene artists seem to have preferred bright or deep colours to make the figures stand out (Önal, 2017, p. 23).
61The mosaic scenes are all placed within frames. First, there is a plain, linear frame, rectangular in shape that is created by single-coloured tesserae (black in almost all the mosaics, except for cat. no. 14, which has red tesserae). Following that, there can be a sequence of frames containing geometric patterns; these are sometimes divided by frames containing single-coloured, white or black tesserae (cat. nos. 13, 14, 17, 20, 27), or only by the black tesserae delineating each frame (cat. nos. 19, 24, 25). The repertory of geometric patterns is fairly limited: (i) guilloche (cat. nos. 14, 19, 21, 24, 25, 27), (ii) single row of triangles (cat. nos. 17, 19, 20, 23, 25, 27), (iii) single row of waves (cat. no. 23, 24, 26), and (iv) three-strand guilloche (cat. nos. 13, 17). Most of the mosaics have only one frame with geometric patterns, though some mosaics combine patterns that create a more varied effect (cat. nos. 19, 20, 25, 27).
- 8 The tombs were also decorated with statues: see Blömer, 2019a, p. 219, no. 34-35.
62The present material, which is characterized by a small number of objects and the paucity of dated reliefs and mosaics, does not allow for the refinement of the chronology of Edessene funerary monuments. It is possible, however, to trace broadly the practice of tomb embellishment with markers and commemorative monuments. The evidence shows that there is a shift in the funerary practices of Edessa from stelae to reliefs and mosaics, which is essentially a change from the use of exterior tomb markers to that of interior commemorative reliefs and mosaics, that takes place near the end of the 2nd century AD. The habit of using stelae and reliefs predated that of mosaics, practiced from the middle of the 2nd century AD to perhaps the early 3rd century AD. Mosaics appeared at the end of the 2nd century AD and dominated the funerary art of the 3rd century AD. The earliest mosaic is that of Orpheus, from AD 194; mosaics with portraits, however, were much more popular compared to those with mythological scenes alluding to beliefs about the afterlife, of which, only four have survived, two with Orpheus, one with Prometheus, and one with a phoenix (see Önal, 2017, pp. 30-35, with previous bibliography). In addition to these, two tombs had geometric mosaics: one dated in AD 238 and one in the 3rd century AD (for these, see Önal, 2017, pp. 56-57, with previous bibliography).8 Mosaics with busts and banqueting scenes are most likely the first to appear, judging by their higher numbers in the period between the end of the 2nd century AD and the early 3rd century AD, while mosaics with standing figures were more popular in the 3rd century AD. The two mosaics from the second half of the 3rd century AD and the 4th century AD respectively, show the combination of an inscribed emblem surrounded by geometric motifs (see Önal, 2017, pp. 119-121; Desremaux & Önal, 2017, pp. 136-137). No mosaic with portraits or mythological figures can be dated in the second half of the 3rd century. AD, indicating that this particular trend is associated with the aristocracy that was allied to the royal house of the Abgarids.
- 9 See also Eristov, Vibert-Guigue, al-As’ad, & Sarkis, 2019, for the tomb of the Three Brothers in Pa (...)
63The funerary art of Edessa has been compared often with the art of Palmyra (see recently Healey, 2019, with further bibliography) and the material shows how motifs and compositions in the art of the city were connected to that of the wider region. The portraits from each site, however, have distinctive features that differentiate them from those from neighbouring towns. The tall, conical headdress under a veil that is typical of the Edessene female attire, differs from that of Palmyra (a headband, often decorated, under a turban and a veil, see Krag, 2018, pp. 106-109), that of Zeugma (a high turban usually worn under a veil), and that of Hierapolis in Syria (a high turban whose upper part appears triangular because of the way that the veil is fastened over it) (see Rumscheid, 2019). Thus, each city seems to have its own particular type of headdress, which is the one most commonly depicted in the art of the city. However, there are exceptions to this rule. For example, the headdresses of the women in cat. nos. 2 and 4 are cylindrical with a flat top, rather than conical.9
64The placement of one or more portrait busts inside a recess (cat. nos. 1-3), and the depiction of an older woman in bust format together with the full figure of a standing child (cat. no. 7) have their closest parallels in the art of Palmyra, Zeugma, and Hierapolis in Syria, and can be said to be part of this wider, Syrian artistic tradition. It is striking that, while in the reliefs from Edessa cat. nos. 7 and 8 the child is dressed in the same manner as the adult mother, in the art of Palmyra, only adult (perhaps unmarried) women share the same costume as their mothers; when the girls are pre-adolescents, they are shown with tunics and holding bunches of grapes and/or birds, typical attributes of children (Krag & Raja, 2016, pp. 156-157). The compositional schema of the paratactic arrangement of multiple busts is also part of a wider Syrian tradition (see, for example, Blömer & Raja, 2019b, p. 12, fig. 2.11; Blömer, 2019b, p. 57, fig. 4.15). But not every motif has parallels in neighbouring regions in the same type of monument: the seated figures in cat. no. 8 can be compared only with the singular relief of a seated woman from Hierapolis (Blömer & Raja, 2019b, p. 11, fig. 2.9). The motif of the seated woman, however, appears in stelae from Roman Britain: the famous stele of Regina from Arbeia (South Shields), and a stele in Carlisle (Carroll, 2012), and is common in the Palmyrene banqueting reliefs and sarcophagi (Krag & Raja, 2017; Krag, 2018, pp. 32-34, 62-64). The standing male figure in cat. no 6 has no parallels in Palmyrene stelae but appears in sarcophagus boxes from Palmyra (see, for example, the sarcophagus boxes from the exedra of Julius Aurelius Maqqai within the hypogeum of ‘Atenatan: Ingholt, 1935, pp. 58-75), and statues from Hatra (for the statues, usually depicting kings, see Anderson, 2017). The most striking difference is, however, in the use of the banqueting motif in the art of Edessa and that of other sites, most notably Palmyra. The most obvious difference is that the reclining man in the reliefs and mosaics of Edessa was not depicted as larger than the other adults in the mosaic, as was the case with reclining men in Palmyrene funerary reliefs with banquet scenes, but there are other, subtler differences as well.
Fig. 23 Palmyrene banqueting relief. National Museum of Damascus, inv. no. 10.941. Ingholt Archive, PS 1094
© Palmyra Portrait Project and Rubina Raja, courtesy of the Ny Carlsberg Glyptotek
- 10 Unless the artist intended to have her depicted as reclining on her stomach and raising her torso u (...)
65The reliefs cat. nos. 9-11 have their closest parallels in Palmyrene sarcophagi and banqueting reliefs: the main (male) figure is shown reclining on a bed and resting on a stack of cushions. With the fingertips of his upturned left hand, he holds a bowl, while his right hand rests on his right leg (for the gesture, see Miller, 2011, p. 114; Heyn, 2021, p. 69). In two of the reliefs, a woman is shown seated on the bed to his right (cat. nos. 9 and 11), and one figure is depicted standing behind the reclining man. All these motifs are found in Palmyrene sarcophagi and banqueting reliefs in exactly the same order and placement on the relief (see Bobou & Raja, 2023) (Fig. 23). However, the reliefs of Edessa were not free-standing or carved on a plaque covering the sarcophagus, but rather they are carved on the wall of the arcosolium inside the chamber (Segal, 1970, p. 28). If the composition and iconographic details such as the holding of the cup are identical, there are also iconographic differences: in cat. no. 9, a second figure is depicted standing to the right of the reclining man; in cat. no. 9, the woman is shown in the form of a bust portrait to the right of the reclining man,10 while there is a kind of ‘halo’ behind the man: the object was described as a “crown” or “aureole” by Pognon (Pognon, 1907, p. 179) and as a hat with peacock or ostrich feathers (Segal, 1970, p. 41) and has no known parallels (Segal, 1970, p. 41 observes that the closest parallel is the turban worn by figures on mosaics, but unlike the turban, it seems to be placed behind the head of the figure and its beginning is located near the base of the neck). Finally, in cat. no. 11, a winged figure is placed between the reclining man and the seated woman, another iconographic peculiarity that has no clear parallels.
66The differences between the banquet scenes of Palmyra and Edessa become even more pronounced in the 3rd century AD when the imagery of the banquet shifts from the reliefs to the floor mosaics. The composition changes: instead of the figures shown on the same mattress, regardless of whether they are seated or reclining, the reclining figure is the only one on the kline, while the other figures are arranged behind him or to his left or right, or even in front of the kline (cat. nos. 14, 15). The woman is seated on a chair that is adjacent to the kline (cat. nos. 12, 14-16) and is similar to the image of the wife of the reclining deceased common in the Roman banquet reliefs (see Dentzer, 1978; Draycott & Stamatopoulou, 2016). The most obvious differences in the iconography can be seen in the costume choices of the participants in the scene (briefly mentioned above), which reflect the attire of the Edessene elites. There are other differences as well: while the recumbent man continues to hold a bowl or cup with the same elegant gesture (that was not always expertly rendered: see cat. no. 14), he no longer carries a dagger, and he sometimes wears a turban (cat. no. 14) or a ‘Phrygian’ cap (cat. no. 15, 16). Furthermore, in one of the banquet mosaics (cat. no. 15), there is an object in front of the kline that is perhaps a small table or an incense burner, an object that does not appear in Palmyrene banquet scenes except in one single relief (Colledge, 1976, p. 63, 74, 77, 132, 155, 157, 240, pl. 105). Despite the differences in iconographic motifs and details, the banquet scene appears to have a similar function in Edessa and Palmyra: in both cities, it is used as a symbol of wealth and status, with the reclining figure(s) at the centre to emphasize his position of authority as head of the family (Önal et al., 2013, p. 14; Bobou & Raja, 2023).
- 11 For a good example of this Palmyrene practice, see the — partly hypothetical, but evocative — recon (...)
67The mosaics with bust portraits follow the tradition of the bust portrait that first appeared in the city through the reliefs, and the frequent use of framing devices around the figures reinforces the sense of continuity, interconnection, and iconographic transmission between the two types of monuments. It is even likely that the placement of the busts inside a grid pattern in the mosaics reflects the Palmyrene practice of covering loculus burials with relief plaques that created a gallery of portraits within the tombs (Raja, 2019).11
68The arrangement of the family in the mosaics with standing figures is another motif that finds iconographic parallels in the art of Palmyra (see, for example, the stele with three standing figures in the Palmyra Museum, inv. no. A215, from tower tomb no. 68: Henning, 2013, p. 211, pl. 60b; Krag & Raja, 2016, pp. 155-157, p. 172, cat. no. 84, fig. 31; Krag, 2018, p. 232, cat. no. 249), Perrhe (Blömer, 2019b, p. 51, fig. 4. 6) and the Savur plain (Blömer, 2019b, p. 57, fig. 4. 16). Compared to the reliefs, the mosaics of Edessa offer more dynamic compositions, with the family members shown not only one partially behind the other (cat. nos. 24-27, 29), but also with the younger members of the family shown either at a lower register in the field (cat. nos. 26, 27) or behind the adult figures (cat. nos. 25, 27). In mosaic cat. no. 28, a female figure with two children standing in front of her, is completely surrounded by the garments of the protective female figure, in a motif that seems to prefigure that of the Madonna of Mercy, although a connection between the two types is highly unlikely (see Brown, 2017 for the Madonna type).
69Another common feature is the close contact between the family members (cat. nos. 15, 25, 26, 29). In the art of Palmyra, affectionate gestures such as holding the arm or the hand are reserved for depictions of children, mothers with young children (Fig. 24), or mothers mourning the loss of an adult child (see Krag & Raja, 2016; 2017). In the mosaics of Edessa, such gestures are seen between siblings (cat. no. 25, 29), or between other close relatives (cat. nos. 15, 26). In addition to the ‘embrace’ gesture, in cat. no. 15, two of the figures are holding hands.
70Even though the inscriptions identify the figures as portraits in all the reliefs, the inclusion of the winged figure in cat. no. 11, and the peculiar ‘halo’ (if not a hat) in cat. no. 10 suggests that the banquet scene may take place in the afterlife. That the Edessenes believed in life after death for them and their families is made clear by several inscriptions and objects (Salman, 2008, pp. 110-113; Healey, 2018, p. 62). It is not possible to say, however, how the winged figure related to the deceased. The figure itself, described only as a “winged deity” by Pognon (Pognon, 1907, p. 182), is not well-preserved. Therefore, it is not possible to say whether it was a similar figure to the Genius in the Apotheosis relief from the column of Antoninus Pius, the Mithraic and Orphic Phanes, who was closely connected to Aion (Vogel, 1973, pp. 33-38), or a siren-like winged, common in Greek funerary art (Estrin, 2021), but also found in the art of Palmyra (Abdul-Hak, 1952, p. 236, fig. 14), or a winged deity reminiscent of victory such as found in Parthian and Sasanian art (see Mazloumi & Nasrollahzadeh, 2017).
Fig. 24 Stele with woman and two children. Palmyra Museum, inv. no. A 130. Ingholt Archive, PS 1114
© Palmyra Portrait Project and Rubina Raja, courtesy of the Ny Carlsberg Glyptotek
71The colour palette used in the mosaics is not different from that used in Zeugma or in the mythological mosaics from Edessa itself (see Önal, 2017, pp. 12-18). The simple frames and geometric motifs appear in the houses of Zeugma (see Görkay, 2021, pp. 160-259, especially fig. 4 and fig. 6 for the running dog motif, similar to the one appearing in cat. no. 26; fig. 10 and 22 for the combination of bands with triangles enclosing a trellis, similar to the frame of cat. nos. 21, 25 and 27, and fig. 26 for the plain band frame similar to those on cat. nos. 21 and 23). They appear also in the Greco-Roman mythological mosaics found in the houses of the region and their presence has been used as an indicator of the knowledge of Edessene artists of the Greco-Roman artistic koine (Dunbabin, 1999, p. 173; Güler & Çelik, 2002, p. 182, Önal, 2017, pp. 12-18).
- 12 A mosaic was found used as a sarcophagus cover in a tomb in Sidon, but it has been suggested that t (...)
72The brief survey above shows the great variety in funerary commemoration that existed in Edessa, and the differences in the use of the monuments. Reliefs were used both outside the tombs as markers for the location of the burial monument proper, and inside as commemorative markers for the deceased and his family. Mosaics were only used inside the chamber tombs. One of the reasons for this divergence must have been the choice of material: the stone was not as delicate and fragile as the tesserae used in the mosaics. This does not explain, however, why some of the inhabitants of Edessa chose to decorate their tombs with mosaics: the practice is so far unattested in other neighbouring areas or the Roman Empire, and is a unique local tradition.12
73The corpus of funerary portraits shows two distinct groups and phases. The first is the group of sculptures: these were created first and demonstrate the closest connections between the art of Edessa and that of the wider region. The use of the bust format for the portraits and their placement within a recess, which is seen in the free-standing stelae, which are the earliest of the sculptures, can be compared to how bust portraits were placed in funerary reliefs that covered loculus niches in the funerary art of Zeugma, Palmyra and Hierapolis. The Edessene stele form, however, was suitable for a monument that served as an external tomb marker, as can be seen by its large dimensions – especially the height. The wall-reliefs that were carved inside tombs also share similarities with the funerary art of Palmyra, especially in the choice of the banqueting motif for the depiction of the deceased. The similarities do not signify imitation, as the deliberate use of attributes, costume and even the mode of depicting the seated and reclining figures shows the grounding of the imagery in Edessene social practices and habits, even when the overall schema of multiple figures shown on a banquet together seems to have been borrowed from other regions, especially Palmyra.
74The second group is that of mosaics, mostly dating to the 3rd century AD. This group reveals best the emergence of a local artistic tradition that still had points of contact with the art of greater Syria, but privileged local habits and practices. The banqueting motif retains its importance, but it is only one of the ways chosen to commemorate members of the same family. The biggest difference between the sculpted banquet and the one in the mosaics is that there is only one banqueter in the mosaics, and his family is shown around him. While this is similar to the form of banquet seen in the Palmyrene sarcophagi, in the Edessene mosaics there is variation as to the placement of the figures, as well as their actions and gestures, in addition to the depiction of local costumes and attributes. The mosaics with busts of the various family and household members placed within frames can be related to the earlier sculpted busts placed in recesses, as well as the practice best seen in Palmyra, but also in Zeugma and other cities, of covering rows of loculus niches with limestone plaques with portrait busts. Finally, the mosaics with standing figures can also be compared with reliefs from neighouring cities, but in comparison to them, they offer more dynamic and complex compositions that emphasize family bonds through the location of the figures and gestures denoting affection.
75The common iconographic and compositional traits shared by sculptures and mosaics show that all the monuments served the needs of a local elite that, already from the middle of the 2nd century AD, distinguished itself from that of other cities while being aware of them, in contact with them, and perhaps even in competition with them.
Flux RSS
