Çapmalı

The game board with the best-recorded archaeological context comes from the Gobustan National Reserve, near the western shore of the Caspian Sea to the southwest of Baku. A covered shelter on the lower terrace of the Böyükdaş Mountain, between rocks 108, 109, 110, and 111, with petroglyphs from the Bronze Age and later periods, was recorded by Ishag Jafarzade as Shelter VII (Figure 3; Rustamov & Muradova, Reference Rustamov and Muradova1974: 10). Later, it was named Çapmalı (çapma in Azerbaijani means percussion on rock and is a word used locally for the abundant cupules in the Gobustan area).

Figure 3.

Aerial view of the Gobustan area. The rectangle indicates the location of the immediate vicinity of Çapmalı, shown in the inset. © CNES/Airbus, Google Earth.

Shelter VII is situated at the apex of a 100 × 200 m enclosed depression surrounded by rocks on all sides (Figure 4). The surface of this area is scattered with cultural remains and artefacts such as river stones and pottery fragments, mostly dating to the Middle Ages, but occasionally to the Bronze Age. Jafarzade documented many Bronze Age and medieval petroglyphs in this area (Jafarzade, Reference Jafarzade1973). Most of the images are of bezoar goats, typical of the Bronze Age (Farajova, Reference Farajova2018: 81; Sigari et al., Reference Sigari, Shirinli, Abdullayev and Smith2019: 5), but images of horses, bows, footprints, and cupules are also present.

Figure 4.

View of Çapmalı, showing stone 111 in the centre, where the fifty-eight holes game was found.

Archaeological excavations in the Çapmalı camp were conducted outside the shelter in front of its northern entrance in 1972 and inside it in 1973 by Rustamov and Muradova (Reference Rustamov and Muradova1972, Reference Rustamov and Muradova1973). Medieval structures and ceramics, as well as second-millennium bc material (Middle Bronze Age, Late Bronze Age/Early Iron Age) were found in the Çapmalı shelter.

The first two strata show intensive use during the Middle Ages, while the third is devoid of cultural material. In the fourth stratum, the shelter was divided by a wall. This 0.9 m layer of greyish ashy soil contained a greater concentration of archaeological material than the others, with ceramics attributed to the Middle and Late Bronze Ages. This stratum can be divided into two substrata. The artefacts in the upper substratum are more numerous than in the lower, but all date to the second millennium bc, with the assemblages of both substrata being similar in type, composition, and shape. They include various pottery sherds, obsidian, flint fragments, broken river stones, and grinding stones. Rustamov and Muradova classified the pottery in this layer into two categories. A grey, wheel-thrown type of finer fabric, which is often decorated, is comparable to that known from the Khojaly-Gadabay culture (thirteenth–seventh century bc). An underfired, rougher type with brittle yellow and pink fabric mixed with powdered limestone and plastered with liquid clay echoes Middle Bronze Age types. This type exhibits yellow and pink spots, and discernible black streaks in the sherds’ section resulting from insufficient firing. Most of these sherds have embossed geometric patterns on the shoulder. A small quantity of pottery reminiscent of clay lamps known in antiquity (500 bc–ad 300) was also recovered. The Çapmalı shelter is interpreted as a settlement occupied during the Middle Bronze Age, Late Bronze–Early Iron Age, antiquity, and the Middle Ages (Rustamov & Muradova, Reference Rustamov and Muradova1973: 9), probably as a shelter used by seasonal cattle herders during the winter, as still practised locally.

The game of fifty-eight holes consists of a pattern pecked into a stone (stone 111) located in front of the shelter's southern entrance (Figure 5). It was discovered by chance in 2015, beneath a layer of earth covering stone 111. The narrow space between stones 108 and 111 on the shelter's south-eastern side had long been enclosed with stones and covered with earth. This space is covered with unhewn stones that make up a height of more than 2 m measured from the inside. According to the shelter's investigators, ‘There is no doubt that the wall is ancient. Because this wall is filled with soil from the shelter's outer side, it is difficult to detect’ (Rustamov & Muradova, Reference Rustamov and Muradova1973: 5). It is likely that these stones were sandwiched between stones 108 and 111 and covered with soil during the first period of use of the shelter, in the Middle Bronze Age, to improve the living environment inside the shelter; indeed, the semi-desert conditions of Gobustan, especially the strong northerly winds and dust brought by the wind and the rainwater flowing through the rocks, make it an unhospitable environment. This soil, which also covered the stone with the game pattern, is only 0.5 m away from the stones that block the space between the stones 108 and 111. After the rocks were covered with soil, it would be difficult or impossible to carve a pattern of holes in that location. Taking these factors into account, the game pattern on stone 111 can be associated with the shelter's earliest, Middle Bronze Age, layer.

Figure 5.

Orthophotograph of Çapmalı, view toward the south. The arrow indicates the location of the fifty-eight holes game. Reproduced by permission of Elgun Pirverdiyev.

Rendered as a series of shallow depressions, with narrow channels connecting certain holes, the pattern closely resembles boards found in south-western Asia and Egypt (Figure 6). The holes and lines are pecked into the stone surface, as are all the boards discussed here. In the centre are two parallel lines of eleven cupules roughly 12–14 mm in diameter (Figure 7). The uppermost cupule in each line is slightly larger than the others (21 mm (right) and 19 mm (left)), and the second from the top in the right-hand row is also larger (20 mm). A channel connects the seventh cupules in each line, as does another connecting the ninth cupules (the numbering of the cupules and holes follows the convention given in Carnarvon & Carter (Reference Carter1912: 57) which proposes separate tracks for each player). Around this is an arc of thirty-four cupules. The twentieth cupule, at the apex, is larger than the others (25 mm). Curved channels connect the fifth cupule to the tenth, and the tenth to the fifteenth on the left side of the arc. This is mirrored on the right side, though the fifth cupule on the right is damaged, and the bottom portion of the arc is no longer extant. Assuming the right side of the arc was originally symmetrical, four cupules are missing. In addition, there is a larger, probably extraneous, depression to the left of the left-hand row of holes, and a zig-zag channel. While similar lines do not appear on other known game boards, the position and shape of this channel evokes the shape of another board, now in the British Museum (2003, 1201.1; Finkel, Reference Finkel, Finkel and Simpson2020: 46–47).

Figure 6.

The fifty-eight holes board from Çapmalı.

Figure 7.

Drawing of fifty-eight holes boards from the Abşeron Peninsula and Gobustan. Grey indicates cupules or lines which may be part of the game pattern, dotted lines indicate features which would have existed in damaged parts of the stones. Drawings of Ağdaşdüzü 3 after Aliyev and Abdullayev (Reference Aliyev and Abdullayev2011: fig. 55), of Yenı Türkan from a photograph by Ronnie Gallagher, and of Dübəndi after Aliyev and Aslanov (Reference Aliyev and Aslanov1988: pl. 15). Drawings not to scale.

Ağdaşdüzü

Three boards were found at Ağdaşdüzü, a second-millennium bc site in the Abşeron Peninsula (Figure 8; Aliyev & Abdullayev, Reference Aliyev and Abdullayev2011: 67, 123, fig. 55) located 1 km south of the Mardakan railway station. Large stone slabs, some with petroglyphs, surround the settlement. Ceramic models of wheels, cylindrical stands, flat braziers with straight walls, and butter churn fragments have been recovered from this site (Aliyev, Reference Aliyev1993: 62–70). The layout of the settlement, which consists of stone structures enclosed by a large stone fence, is characteristic of the period (Aliyev & Abdullayev, Reference Aliyev and Abdullayev2011: 16–17).

Figure 8.

Two fifty-eight holes boards from Ağdaşdüzü found in the centre of the settlement. Reproduced by permission of Ronnie Gallagher.

Two examples of the game were recorded by Abbas Islamov and Ronnie Gallagher on a large stone in the centre of the settlement (Figure 9). The site has since been built over in the urban development of the Abşeron Peninsula, but the stone itself was moved to the Gala Museum. Settlements dating to the late third to early second millennium bc consist of large or small buildings surrounded by large upright stones, often with petroglyphs on those stones or on the bedrock itself (Achundov & Narimanov, Reference Achundov and Narimanov1996: 46–49), as is the case of our game patterns. The location of two games near the settlement's centre suggests that people gathered there to play.

Figure 9.

Abbas Islamov and Ronnie Gallagher indicating the location of the two Ağdaşdüzü boards in the centre of the settlement. Reproduced by permission of Ronnie Gallagher.

The games were located roughly 1.5 m apart. One of the patterns is nearly identical to the array found at Çapmalı (see Figure 7), but the two holes at the top of the central parallel lines are not larger than the others. The second pattern is not as well preserved but follows a similar layout. Ten cupules are preserved in the right-hand parallel row, without connecting channels. Seven survive in the left-hand row. The outer arc is nearly complete, except for the lower four holes on the right-hand side.

A third example, with an unreported find location, also exhibits the same basic pattern of cupules (Aliyev & Abdullayev, Reference Aliyev and Abdullayev2011: pl. 55; see Figure 7). However, in both rows there appear to be curved lines connecting the ninth depression with the eleventh in its row, such that the lines touch one another. It is possible that one of these is extraneous. There are five or six depressions on the left-hand side, and one on the right, below the pattern, which has been cut off from the rest by an unevenly curved line, perhaps to demarcate the gaming pattern and avoid confusion. This line, like that on the Çapmalı board, demarcates the entire array of impressions known from other games, such as that from Kültepe (Özgüç, Reference Ozgüç1986: pl. 132.9).

Yeni Türkan

A fifth instance of a game of fifty-eight holes was discovered in the ‘Kurgan desert’ in the south-eastern Abşeron Peninsula. The Kurgan desert is connected to the settlement of Yeni Türkan, situated between the modern villages of Türkan and Hövsan. Kurgan 40, which had a diameter of 22 m, was excavated in 2004. The mound is dated from the second half of the third millennium bc to the middle of the second millennium bc, based on the ochre-coloured, polished, relatively well-fired ceramics and other finds it contained (Aliyev, Reference Aliyev2005b: 110). Ceramics exhibiting similar characteristics were also recovered from Kurgan 5 at Dübəndi (Aliyev & Aslanov, Reference Aliyev and Aslanov1988: 8–9; see below). Black polished ceramics adorned with a hanging triangle motif were discovered in Yenı Türkan's adjacent Kurgan 41, which yielded the same cultural assemblage. Instances of such ceramics have been documented in sites associated with the Uzarliktepe culture, which flourished during the Middle Bronze Age (Aliyev, Reference Aliyev2005b: 109). A 2 × 3 m rectangular stone structure in the south-western sector of the kurgan's stone platform consisted of large pieces of rock. Based on ceramic fragments, animal bones, hearth, and ash remains, Idris Aliyev interpreted this structure as an altar. This is where the game, as well as a schematic image of a goat, was discovered (Aliyev, Reference Aliyev2005b: 108–09).

The game board is less well preserved than our other examples. The roughly rectangular stone is irregularly shaped around the edges, with the outer arc of depressions following these edges. The two parallel lines in the centre are visible but are much eroded or worn, such that they form a single channel, making it difficult to distinguish them (see Figure 7). On the outer arc, six holes are missing from the bottom of the right-hand side because the stone is damaged, and the left-hand side of the arc is very worn. The apex and top holes of the parallel lines are not enlarged; lines connecting holes in the outer arc are visible on the right, with possible traces on the left. The remnants of one line may have connected cupules in the inner rows. The degree and type of wear on this stone is reminiscent of that seen on other limestone games in Bronze Age Cyprus (Crist, Reference Crist, Chovanec and Crist2021: 81–83), suggesting that it was used many times for play before it was deposited in the kurgan.

Dübəndi

A final instance of the game, on a moveable stone plaque measuring 360 × 300 × 60 mm, was found in Kurgan 5 at Dübəndi (Aliyev & Aslanov, Reference Aliyev and Aslanov1988). The board was found incorporated into the eastern side of the kurgan's stone platform, which was covered with sand (Figure 10). Its use as a game must therefore predate the construction of the kurgan. It was more crudely made than our other examples, and the apex and part of the left half of the arc of cupules are obscured by surface damage. Nevertheless, twenty-nine holes in the outer arc remain (see Figure 7). The cupules are oval (c. 35 × 16 mm), unlike the more circular holes on the other stones. There are no connecting lines or enlarged holes. An extra depression is centred above the parallel lines, also seen on the third example from Ağdaşdüzü. The excavators assign Kurgan 5 to the earlier part of the Middle Bronze Age, specifically the end of the third millennium to the beginning of the second millennium bc, based on the kurgan type and red wares found within such mounds (Aliyev & Aslanov, Reference Aliyev and Aslanov1988: 11–12).

Figure 10.

Plan of Kurgan 5 at Dübəndi, with the location of the fifty-eight holes board indicated by an arrow at the bottom (after Aliyev & Aslanov, Reference Aliyev and Aslanov1988: pl. 3). Reproduced by permission of Idris Aliyev.

Connections with Anatolia, Mesopotamia, and Iran

Much has been written about the connections between the Caucasus region and Mesopotamia, Anatolia, and the Levant, particularly with respect to the Maikop and Kura-Araxes cultures of the Chalcolithic and Early Bronze Age (see Sagona, Reference Sagona2018: 132–280 for a discussion of these periods and broader connections). The Middle Bronze Age in the Caucasus is not as well understood, largely due to the transition from a sedentary lifestyle to one characterized by mobile subsistence strategies leading to a decrease in settlements. Thus, archaeological evidence for the Middle Bronze Age mostly comes from tombs, which generally took the form of kurgans, some of which attest to considerable wealth and a socially stratified society.

Such is the case for the Abşeron Peninsula and the Gobustan Reserve. On the Abşeron Peninsula, a regional style of kurgans exists, as well as evidence of seasonal settlements (Achundov & Narimanov, Reference Achundov and Narimanov1996: 46–49; Bertram, Reference Bertram2003: 86–87; Sagona, Reference Sagona2018: 371–72). Rock art was also a prevalent feature of the Middle Bronze Age in the peninsula, and its thematic similarities to Mesopotamian art point to cultural connections with the wider region in which the game of fifty-eight holes was played (Aliyev, Reference Aliyev2005a). At Gobustan, some of the Bronze Age artefacts are similar to those found at sites in south-western Asia (Muradova, Reference Muradova2011: 109), and some Middle Bronze Age petroglyphs also reflect motifs of Mesopotamian mythology (Abdullayev, Reference Abdullayev2018: 763).

Middle Bronze Age connections to other cultures of south-western Asia are well-attested in other parts of the Caucasus. Recent work has shown that the Caucasus region was included in a network of connections that facilitated exchange—in part, due to the seasonal movement of pastoralists—in Anatolia, Iran, and further south (Rova, Reference Rova, Coppini and Simi2020: 12–16). Even though the Middle Bronze Age saw a transition from the sedentary lifestyle of the Kura-Araxes culture to a more mobile, pastoral mode of subsistence, the successor cultures of the region remained in contact with the wider south-western Asian world. Beginning in the second half of the third millennium bc, the presence of tin bronzes, amongst which several examples were found in Gobustan, points to exchange with Mesopotamian trade networks that acquired tin from Iran, Anatolia, and Afghanistan (Hasanova, Reference Hasanova2016: 222–25).

Further to the west, in Georgia and part of Armenia, the Trialeti culture is noted for its jewellery and other precious metalwork with close parallels to Mesopotamia, Syria, and even Egypt. Silver goblets bear artistic motifs that ‘derive directly from the glyptic tradition of the Old Assyrian colony period, as reflected on the seal impressions from Kültepe Kanesh’ (Sagona, Reference Sagona2018: 348; Rubinson, Reference Rubinson, Smith and Rubinson2003). Basket-handled metal vessels from Trialeti also point to connections with Kültepe Kanesh, where the greatest number of vessels with this distinct handle have been found (Rubinson, Reference Rubinson, Maisuradze and Rusishvili2001). Goldwork also connects the Trialeti culture with the south, including jewellery, cups, and weapons, with clear Mesopotamian inspiration, as well as a leg-shaped agate amulet of a type popular in Egypt from the Sixth Dynasty until the First Intermediate Period (Abramishvili, Reference Abramishvili, Hansen, Hauptmann, Motzenbäcker and Pernicka2010: 169–71).

Blue faience beads originating in Egypt and Mesopotamia also indicate Middle Bronze Age relations with south-western Asia. Turquoise faience beads found in Kültepe II, Qizilburun, Shahtakht, Chalkhangala, and Galajik (all in western Azerbaijan) indicate that this type of item was present there in the first half of the second millennium bc (Bakhshaliyev, Reference Bakhshaliyev2007: 179). Turquoise glazed faience beads were also discovered in large quantities in the Middle Bronze Age monuments of Gobustan and Abşeron, including at Ağdaşdüzü, Dübəndi, and Türkan (Aliyev, Reference Aliyev1993; Muradova, Reference Muradova2011: 83).

Cultural transmission of games from Anatolia to the Caspian coast

The Abşeron Peninsula's connection to exchange networks suggests that the game of fifty-eight holes spread through trade rather than conquest (de Voogt et al., Reference de Voogt, Dunn-Vaturi and Eerkens2013: 1728). Moreover, it appears to have been transmitted with fewer innovations (i.e. changes to the board) than the game of twenty squares over the course of its existence (de Voogt et al., Reference de Voogt, Dunn-Vaturi and Eerkens2013: 1727–28). The boards from Azerbaijan do, however, show innovations that have not been observed previously, which may be relevant since these boards lie at the beginning of the history of the game of fifty-eight holes. De Voogt et al. (Reference de Voogt, Dunn-Vaturi and Eerkens2013: 1727) noted that their sample of fifty-eight holes boards was skewed toward later periods, and that more early examples were needed.

The boards from the Abşeron Peninsula are the only known examples that were made on locally available limestone. Moreover, they were made in such a way that they are unlikely to have been imported: they are the only known immobile versions of the game of fifty-eight holes, thus unequivocally made where they were found. Portable game boards found in other regions, especially those made of precious materials, could have been imported, thus complicating our ability to track innovations in board layout on a case-by-case basis. Comparing our boards to those found elsewhere can therefore provide indications about how the game reached the region, and how it changed as a result of cultural interactions.

While the boards from the Caucasus have no exact parallels with any other known boards used for the game of fifty-eight holes, they all have features that exist in at least one of the others. An example from Kültepe in central Anatolia (Özgüç, Reference Ozgüç1986: pl. 132.9; Dunn-Vaturi, Reference Dunn-Vaturi2019: fig. 10) most closely resembles our boards. It contains the two distinct holes at the top of the two parallel lines in the centre of the board (Özgüç, Reference Ozgüç1986; Gülalp, Reference Gülalp2008, pl. 20b) as well as the lines in the outer arc and a larger hole at the apex. It has a total of sixty-one holes. This board is dated to between 1950 and 1835 bc by Finkel (Reference Finkel, Finkel and Simpson2020: 46), while Manning et al. (Reference Manning, Barjamovic and Lorentzen2017: 75) date Kültepe's Stratum II, where the board was found, to 1885–1836 bc. Regardless, it is one of the earliest boards known. The fact that the examples from Azerbaijan most closely resemble the board from Kültepe is not surprising considering the amount of material from Trialeti—particularly with regard to metal objects—that also indicates connections with Kültepe.

A second example from Kültepe is morphologically different and is slightly later, dating to c. 1800–1700 bc. The board itself has the shape of an arrowhead, with holes six, eight, ten, fifteen, twenty, twenty-five, and the apex outlined in lead. Another example from Karahöyük is nearly identical, and has the same date (Gülalp, Reference Gülalp2008: 46, 48, pls 19b, 20a). Contemporary examples from Acemhöyük are fragmentary, one showing a portion of the outer arc of holes with a curved line consistent with those that connect holes fifteen to twenty or twenty to twenty-five (Öztan, Reference Öztan2005: 397, pl. 6a). Yet another example of this arrowhead-shaped type of board was found at Boğazköy and dated to the Hittite period (Bittel, Reference Bittel1937: 22); since its archaeological context was not recorded, it may belong to the same period as the other Anatolian boards (Gülalp, Reference Gülalp2008: 54; Anne Dunn-Vaturi, pers. comm). Another object that seems to be a fifty-eight holes board was found at Haftavan Tepe in Iranian Azerbaijan (Edwards, Reference Edwards1983: fig. 146.s), in Early Stratum VIB, broadly dated to c. 1900–1550 bc (Edwards, Reference Edwards1986: 70). It is a copper fragment, with depressions arrayed in the remnants of a pattern of two parallel lines surrounded by an arc. Six remain in the left-hand line, four or five in the right-hand line, and seven in the outer arc. There is no indication of lines connecting the holes. In this respect, let us note that only one of the boards from Azerbaijan (that from Kurgan 5 at Dübəndi) is without lines connecting holes.

Boards classified by Finkel (Reference Finkel, Finkel and Simpson2020) as a ‘crossover’ type, which include those from Azerbaijan and Kültepe, indicate a kind of gameplay that differs from that shown on the early Egyptian boards with two parallel tracks. Because of the regularity in some of the lines—such as those that connect the holes in the outer arc, which always connect the same holes—it is likely that these share some common origin. Innovations appear in some of the boards’ appearance: whether the apex of the arc is accentuated, whether there are ‘starting holes’ or any other lines connecting holes, and whether they cross between the left and right side of the board. It may be possible to trace a sequence of innovations from the Kültepe board to the Ağdaşdüzü board (Ağdaşdüzü 3 in Figure 7), which adds two crossover lines to the board, to the other examples from Azerbaijan that eliminate some of the crossover lines. This sequence of a single change from board to board could indicate a process of cultural transmission, as the game was adopted into the society living on the Caspian coast. More finely grained dating of these sites than is currently possible would be necessary to confirm this, but conceptually the process is more visible here than, for example, between Anatolia and Egypt.

In Egypt, there are never crossover lines on fifty-eight holes boards. While the south-western Asian boards exhibit a considerable degree of variability, the Egyptian examples are remarkably regular in design (see Figure 1), with lines connecting the eighth and tenth holes in the inner rows, the sixth hole in each inner row to the twentieth in the outer arc on each side, and the fifteenth and twenty-fifth holes marked with the hieroglyph nfr (‘good’). This Egyptian pattern is never exactly replicated in south-western Asia. Here, where it can be determined, boards with lines connecting holes always display a crossover pattern. It is not easy to hypothesize a sequence of transmission and innovation for the Egyptian board type, since there are several differences between the Egyptian games and the Anatolian ones, without morphological intermediates in the Levant. A board dated to the Middle Bronze Age is said to have been found at Tyre (Dunn-Vaturi, Reference Dunn-Vaturi2019: 83). It is a flat slab of limestone, without lines or marked holes, except for the apex and two ‘starting holes’ that are larger than the others. There is also a single small hole at the bottom of the board, but it is unclear whether it is meant to be used for play. It is the only known contemporary example which is purportedly from the Levant. Lebanon or Syria would be expected to be the connection between Egypt and Anatolia (Dunn-Vaturi, Reference Dunn-Vaturi2019: 83), but this board is several innovations away from either the Egyptian or the Anatolian boards.

Finkel (Reference Finkel, Finkel and Simpson2020: 53–55) suggests that the development of the crossover board configuration is an adaptation of the game of fifty-eight holes to make it more interesting than the Egyptian version by imposing increased interaction between each player's pieces. This is based on the possibility that the games from Azerbaijan could date to the late second or first millennia bc, but this is not supported by the archaeological evidence from Çapmalı, Dübəndi, Yenı Türkan, and indeed Kültepe, which clearly demonstrate that the crossover game was played during the Middle Bronze Age. The more complex board from Tepe Sialk in central Iran, which dates from the eleventh to ninth century bc and features a more elaborate pattern of crossover lines (Ghirshman, Reference Ghirshman1939: 44), may indicate such a development toward more complex rules, but it is clear that crossover boards existed in the early history of the game of fifty-eight holes.