The Circle of Ancient Iranian Studies
By: Boris I. Marshak
(Sogd. Pancyknδ), a Sogdian city, the ruins of which are located in the southern periphery of the present-day city of Panjakent (in western Tajikistan).
Location and early history
At the beginning of the 8th century Panjikant was the main settlement of the Panch district, a fact reflected in its name. Some scholars hold that Panjikant was also known as Bo-xi-te (Ma 1987; Yoshida 1993, p. 254), and consider it to be identical with the capital of the principality of Mâymurgh, mentioned by that name in Chinese histories from the Tang period, and situated to the south or south-east of Samarqand. It should be noted, however, that Panjikant is located about sixty km to the east of Samarqand, while the capital of Mâymurgh was about 100 li to the south-east (or south) of it. Since one li is equivalent to 500 m, the distance seems to be almost the same. Nevertheless, in speaking of Central Asia, the Chinese authors might have translated as li a local measurement, equal roughly to 300 m. Thus, Xuan Zang, who visited Samarqand (Afrâsiâb) writes that the perimeter of the city walls was twenty li, but the actual perimeter is six rather than ten km. The same traveler describes Mâymurgh of his time (7th c.) as being extensive from north to south, and rather narrow from east to west. Mâymurgh of this description could not possibly have encompassed Panjikant (Staviskiî 1959, p. 85). In fact the only mention of Bo-xi-te dates from the Tang period (618-906 C.E.), i.e. the time of Xuan Zang or later. Nor do similarities in the names of a few prominent people from the two cities provide sufficient grounds to consider Panjikant identical with the capital of Mâymurgh: Thus the father of Ùekin Ùur Bilgä, who ruled Panjikant at the end of the 7th and the beginning of the 8th centuries, appears to be a namesake of one (or perhaps even two) Mâymurgh princes known as Pychwtt, who held the reigns of power at the beginning of the 7th century and around 658 C.E. Numismatic evidence shows, however, that the man who ruled Panjikant before Bil'ge was not Pycwtt, but a Sogdian named Amogyan or Gamaukyan. Therefore even if Pycwtt, the father of Bilgä, resided in Bo-xi-te, it was only his son who ended up holding sway over Panjikant. Judging by Document V-8 from Mt. Mug, this son of Pycwtt's ruled not only Panch, where he was hailed as "lord" (MR'Y), but also £©t (£št), a hard-to-identify locality, where he was titled "king" (MLK'). It must also be noted that during the Tang era the first consonant of the place-name Bo-xi-te would have been pronounced p, which makes it less likely that the kingdom of Bilgä is identical with this toponym. Yoshida thinks that Panjikant was a part of Mâymurgh because the Chinese named another Panjikant – probably a colony of its namesake in Semirech'e - Mi-guo (State Mâymurgh). However this fact could reflect some kind of a possible Mâymurgh's domination over Panch in the period before the 7th century C.E. E. de la Vaissiére compares Pat-sik-tek (the Tang pronunciation of modern Bo-xi-te) with Panchikant (de la Vaissiére 2002, p.125), but many Sogdian toponyms were transliterated much better with the Chinese characters among which there were rather good equivalents of the last syllable –kan.
Panjikant was the easternmost city of Sogdia. Further to the east, starting from the valley of the river Kshtut, lay Pârghar, which—at least in the 9th-10th centuries and perhaps even earlier—was part of Osrušana, although in the early 8th century it was within the domain of prince Dêwâštich (708?-22; q.v.), ruler of Panjikant. Panjikant of the 5th-8th centuries is known primarily from extensive archaeological excavations, while the scant information about the relatively short period of Dêwâštich's rule is derived mostly from documents found at Mt. Mug (Mogh). The only reference to Dêwâštich and his Panjikant supporters in Tabarî pertains to the year 103/722. In 102/721-2 Dêwâštich, who was still a ruler of Panjikant, claimed the title "king of Sogdia and lord of Samarqand." The Arabs initially recognized his new title, but soon forced him to flee to Pârghar, and later to the castle on Mt. Mug, where he was finally captured (Grenet, de la Vaissiére 2003). After holding him prisoner for a few months, the Arabs executed him. After Dêwâštich, Panjikant had no more native rulers.
Although the site of ancient Panjikant has been known since the 1870s, excavations there started only in the 1930s, when the name of Dêwâštich and his main title "lord of Panch" were identified in documents found at Mt. Mug. Between 1937 and 1940 V. R. Cheîlytko conducted limited excavations in Panjikant, without publishing the results. These excavations uncovered remnants of pre-Islamic structures of sun-dried brick in good condition, and located a necropolis with nauses and ossuaries on the outskirts of the ancient town.
The site was investigated by Aleksandr Yu in 1946. Yakubovskiî, who noted its great importance and put forth a proposal for its systematic and methodical excavation, since in his view Panjikant could yield valuable data about the character of Central Asian cities before the Arab conquest. Yakubovskiî put together a strong team of specialists, including Mikhail M. D'yakonov, historian and art historian; archaeologist Alekseî I. Terenozhkin, Iranist (and later numismatist) Ol'ga I. Smirnova, Arabist Aleksandr M. Belenitskiî, and architect Veronika L. Voronina. Along with these established scholars the expedition included the young archaeologists and Orientalists Boris Ya. Staviskiî, Vladimir A. Livshits, and Oleg G. Bol'shakov, who later became renowned specialists as well. The Panjikant division of the Sogdian-Tajik (later Tajik) expedition, led by Yakubovskiî, began excavating Panjikant in 1947, and since then the excavations have been conducted without a season's interruption. After the death of Yakubovskiî in 1953 the expedition was headed briefly by D'yakonov, who passed away in 1954. From 1954 till the end of the 1970s it was led by Belenitskiî. Since then, the excavations have been conducted under the leadership of Valentina I. Raspopova and the author of this article.
After more than half a century of systematic archaeological investigations, Panjikant has become one of the most thoroughly studied early medieval cities not only in Sogdia, but in Asia as a whole. The excavations show that this city, situated on the rim of a high terrace overlooking a fertile, well-irrigated valley, was founded in the 5th century C.E. and was inhabited until the 770s.
Description of the site
Its citadel (see Isakov 1979) is separated by a ravine from the š or city proper, which lies to the east of the citadel and is surrounded by fortified walls of its own. Two additional walls cross the ravine, linking the š with the citadel, and creating a unified defensive system around the entire city. The central structure of the citadel is a square fort built close to the northern part of a mountain ridge, which runs from south to north. In the end of the 7th or the early 8th century C.E., a square keep was erected in the south-east corner of the fort. At the foot of the fort and to the north of it lies the lower fortification, watered by the abundant Qaynar spring. It shows traces of habitation from the 2nd century B.C.E. to the 1st century C.E. This cultural layer contains remnants of ceramics, but none of buildings. To the south of the fort stood a fortified wall, which defended the citadel against attacks from the top of the ridge. There were no buildings between the wall and the fort. On a hilly site to the east of the fort once rose the richly decorated palace of Dêwâštich, which apparently burned down in 103/722. It was an expansion and an extensive reconstruction of an earlier building, dating from the 6th century C.E. Another palace from the 6th century was located in the lower fortifications.
In the 5th century the area of the city proper (without the citadel) measured about eight hectares. Straight fortified walls defended the settlement: the northern wall running along the rim of the terrace, and the eastern wall perpendicular to it. The southern wall ran straight only where the terrain permitted, and the western wall followed the irregular edge of the hill, departing from the overall regular design. The city walls of Panjikant in the 5th century were ten to eleven meters high, bristling with numerous towers, and punctured by embrasures in a chessboard pattern. Later the walls were made thicker, with fewer towers, a sloping façade, and no embrasures close to the foundations. The residential buildings of the city consisted of several small rooms with low wooden ceilings. All walls were made of sun-dried brick and clay. The streets and alleys intersected at right angles. The land at the city center, where two temples stood, has apparently been dedicated to sacral purposes since the founding of the settlement.
The architectural style of the temples, which by the beginning of the 8th century C.E. had undergone many reconstructions, can be traced back to the traditions of Hellenistic Bactria. The two temples are very much alike: each consisted of a central building facing east and surrounded by a yard, which was adjacent to yet another yard to the east, with an exit to the street. A visitor walking from the street towards the main building would have seen the sacred spaces of the two yards open before his eyes one after the other, until, standing in the inner yard, he would have seen not only the portico, but also the interior of the central hall, which—not enclosed by a wall—opened directly onto the portico of the main building. At the far end of the hall there was a door leading to the cella, and on each side of it two niches with clay statues of divinities. A characteristic feature of the Sogdian temple was its openness to the rays of the rising sun and to the eyes of the laity. The passageways to the corridor, which circumvented the hall and the cella behind it opened onto the portico to the sides of the central hall. Both temples were dedicated to the cult of the gods. A space for the sacred fire was added to Temple 1 only in the late 5th and the 6th century C.E.
The earliest nauses of the necropolis at the edge of the ancient city, with Zoroastrian ossuary burials, date back to the 5th and the beginning of the 6th centuries C.E. At the end of the 5th century the area of the city had grown to 13.5 hectares. New fortifications were built to the south and east, so part of the old walls were enclosed within the perimeter of the new ones, dividing the city into inner and outer quarters. The walls of the inner city were repaired and reinforced in the 6th and the 7th centuries. They were pulled down only at the beginning of the 8th century.
The two temples contained statues and murals from the very beginning (5th c.). The earliest murals in the palaces of the citadel date from the 6th century. Some of the houses built during the 6th century were two stories high, with vaulted ceilings on the lower floor, and murals on the walls of some rooms. However, during the 5th-6th centuries, no building in Panjikant could rival the magnificence of the two temples, and even the houses of the most prosperous residents seemed rather humble in comparison. In the 7th-8th centuries, though, it was the houses of the rich that set the tone of urban architecture in the city. The end of the 7th century and especially the first quarter of the 8th century marked the heyday of early medieval Panjikant. At the beginning of the 8th century the spaces between the houses became passageways (covered with vaults in places), over which towered the walls of the neighboring buildings. Not only the dwellings of the rich, but also those of the poor were more often than not two-story buildings with vaults over the rooms on the first floor. All residential houses from that period—not only those of the rich, but also of the merely well-to-do citizens—were decorated with murals and woodcarvings. Such reception halls were found in more than a third of all houses in the city. The streets were lined with small shops and artisans' workshops, often clustered around the bigger residential buildings. The workshops of the blacksmiths and of other metalworkers are most easily identified among the ruins (Raspopova 1980). The prosperous houses often had a few shops or even a small market built on the master's property. These commercial spaces were wide open to the street, but had no link to the living quarters of the buildings: apparently the tradesmen and small shopkeepers leased the premises from the landowners.
Society and art
A historical-sociological characterization of the residential quarters of Panjikant was published by Raspopova (1990). Each house had its own individual characteristics, along with common traits it shared with other homes belonging to representatives of the same social group. Almost every dwelling in the city had dark vaulted rooms for storage and other household purposes, and a spiral staircase leading to the living quarters on the second floor. The well-to-do houses had a special room outfitted with a fire-altar, and sometimes a ceremonial hall decorated with murals and woodcarvings. In the biggest buildings there were courtyards and a whole system of ceremonial rooms on the first and the second floors. More often than not a vaulted corridor led to the ceremonial hall, which was two stories high and was illuminated by a skylight in the center of the ceiling. The entrance to such a house was often designed as a portico or an arched ayvân. All ceremonial rooms were decorated with wall paintings. In the main hall, across from the entrance, there was an architecturally designed or painted niche up to 4.5 m wide, with giant images of the tutelary gods and small-scale depictions of praying members of the household. Stepping into the hall, a visitor could look up toward the ceiling and its centric skylight, which each architect designed differently. Very often, the central part of the hall was marked with four wooden columns, which supported complex wooden structures overhead, topped with a dome on a square foundation. Sometimes there were no columns, and the hall was covered by a truncated pyramid made of wood. Everything was decorated with high-relief woodcarvings, and even with small statues of caryatids and atlantes. The motifs of the ceiling reliefs varied, but the most common ones featured arched niches sheltering figures of the gods, one of whom was the sun-god on his chariot. The palatial interiors of these houses, which belonged to small-town merchants and landowners, attest to the confidence and self-esteem of Sogdian citizens, who apparently felt as much masters of their own lives as the princes ruling the land.
Next, the ancient visitor would probably have looked at the murals around the niche and along the other three walls. These figures were much smaller than the gods and goddesses of the cult scene facing the main entrance, and formed two or three friezes, which encompassed the walls. Their most common themes were royal feasts featuring several kings with their subjects (i.e., scenes of universal celebration), hunting scenes and, often, illustrations of the exploits of Rostam, of local heroes, of amazons, and of personages from the Indian epic Mahabharata. Sometimes, there was also a row of small figurines or a low ornamental frieze, situated over a wide clay bench (suffa). It depicted animals chasing one another, pairs of lovers, or scenes from fairy tales, parables, and fables (among them themes from the fables of Aesop, from the Panchatantra and the Sendebâd-nâma). Thus, all pictorial representations in the central hall fall into three categories, reflecting the significance they held for the Sogdians. The place of honor was held in the central niche (first category) by the gods, whose figures were larger than all other images. A less important place was accorded to the side friezes, where the figures were smaller (second category). Last in importance and smallest in size were the murals from the lower frieze, which depicted scenes from folk tales and other popular genres (third category). On the whole, the layout of the Sogdian central hall is unique, and its decorations show the familiarity of Sogdians with the artistic and literary traditions of many different cultures, including those of Persia, Greece, and India. Among the other vestiges of city life during the 7th-8th centuries are the nauses strung along the roads leading to the city gates, and the modest suburban farmsteads perched on the rims of terraced fields.
The recurrent Arab campaigns against Samarqand apparently did not reach Panjikant at least until 103/ 722, which allowed the brief florescence of the city, and made possible Dêwâštich's claim to the title "king of Sogdia." However, Dêwâštich's authority in Panjikant itself was not that of an absolute monarch. The documents from Mt. Mug show that the citizens of Panjikant had their own revenues and magistrates. The houses of the prosperous landowners and merchants were similar to those of the rulers, which suggests that the latter were perceived by the city elite as "first among equals."
Bronze coins yield the names of two "lords of Panch": Amoygan (or Gamaukyan) (2nd half of 7th c.), and Bilgä (end of 7th-beginning of 8th cc.)—and also a "lady of Panch," Nana (first decades of 8th c.). Bilgä, whose name appears not only on the coins but also in Document V-8 from Mt. Mug, was neither the son of the previous ruler of Panjikant nor the father of Dêwâštich, whose rule coincided with the time when the coins of Nana, the mistress of Panch, were minted. Two explanations have been offered in this regard: 1. The monetary issues during the reign of Dêwâštich did not bear his name, but the name of the goddess Nana, patron of Panch, referring to her as "the lady of Panch" (Henning). 2. Nana, named so in honor of the goddess Nana, was a historical personality like all other rulers mentioned on Sogdian coinage during the 7th-8th centuries (Livshits). She could have been a daughter of Ùekin Ùur Bilgä, whom she succeeded to the throne, and the wife of Dêwâštich. If so, then Dêwâštich's progeny could indeed claim that Bilgä was their ancestor, as the father (actually, father-in-law) of Dêwâštich. The princess might have two names: the Sogdian name of Queen Nana (or "Lady Nana") and some Turkic one (cf. Documents Nos. 3 and 4 from Mt. Mug). With regard to the first hypothesis, long years of archaeological excavations at the temples of Panjikant have brought no conclusive evidence that they were involved in the minting of coins. Temple II contains a few images of Nana, while, for now, none have been found in Temple I. We can speculate about the political significance of each temple only on the strength of circumstantial evidence. Thus, after the fall of Dêwâštich in 722, fire consumed his palace, as well as many (but not all) of the houses belonging to the city elite, and Temple I, but not Temple II. If one assumes that the burning of these buildings represents a repressive measure of the Arabs against Dêwâštich and his supporters, then Temple II might have been spared because it belonged to a part of the population that sided with the conquerors. One might speculate that Dêwâštich allowed the temple authorities to mint coins dedicated to Nana, the mistress of Panch, as a highly unusual gesture of compromise to some influential and potentially hostile faction of citizens associated with Temple II. Finally, even if Nana the mistress of Panch were a real person, the fact that she was so named, and that her name appeared on the coinage, also suggests that the cult of the goddess was of special significance in the city.
The question of Panjikant's religious affiliation is rather complicated. There is some evidence of the presence of Christianity and Buddhism, and perhaps even of the cult of Shiva, but—judging from the wide distribution of ossuary funerals and fire-altars—the majority of the population observed some local variation of Zoroastrianism in combination with cults of additional deities, not all of whom were of Iranian origin (e.g., the cult of the Mesopotamian goddess Nana/ Nanaya). In the social environment of early medieval Sogdia, marked by a high degree of particularism and by the influential role of the nobility and the merchants, the veneration of the divine patrons of a city, of a dynasty, or even of a family was an important feature of religious life. Evidently in every well-to-do house there were two rooms associated with religious observances: one with a fire-altar, devoted to fire-worship; and the ceremonial hall, where the household prayed to their own divine patron. A fire was lighted before these images too, but in smaller, portable altars, as seen in the murals depicting scenes of worshiping the gods. On a city-wide scale, these ceremonial halls were replicated in Temple I and II.
The iconography of the gods can be traced partially to the Hellenistic period (e.g., the image of the seated goddess), but it seems to have taken its final shape in the 5th-6th centuries, influenced by Sasanian notions of the royal attributes of the deity, and acquiring some Hinduistic features. Although each household had its own divine protector, all gods were part of a single pantheon, as may be seen from murals depicting several divinities together. Easily recognizable are the three-headed god of the wind, Vešparkar, who resembles Shiva, and the four-handed Nana, riding a lion or seated on a lion-shaped throne. The identities of other Panjikant gods, defined on the bases of iconography, remain debatable. There are more than twenty different deities, depicted on small sacred images of stamped terracotta from the 6th century, on murals, woodcarvings, and in clay figurines. Most common are the images of Nana, a god sitting on a camel-shaped throne, and of a god standing over a fallen demon.
Toward the end of the 720s and 730s, during a period of persistent anti-Arab insurrections and punitive expeditions by the Arabs, Panjikant was almost entirely abandoned. Its reconstruction is dated, primarily, from the coinage of the Sogdian king Turgar, who acceded to the throne about 120/738. By that time Panjikant had no native rulers, and the palace of Dêwâštich had been turned into a barracks, apparently by the Arabs. The temples were not restored, but many homes were outfitted with magnificent new murals, some on religious themes. The return of the local inhabitants most probably followed in the wake of the 123/741 treaty between Nasár b. Sayyâr, the Omayyad viceroy in Marv, and those Sogdians who had left the country to fight against the Arabs with the army of the khof the Türgeš. In accordance with this treaty, arrears of taxes were remitted, and apostasy from Islam pardoned. As a result, the Arabs got their taxpayers back, and the Sogdians tried to resume their former way of life. However, soon after the victory of Abu Moslem over the Omayyads in 749, there was another cardinal change. The fire-altars were destroyed or turned into kitchen fireplaces, and the murals were vandalized, apparently from religious motives, with most damage done to the faces and the eyes of the images. The ceremonial halls fell into disrepair, and in many houses the inhabitants continued to use only part of the rooms; other large residential buildings were subdivided into two or three dwellings for the regular citizenry. Nevertheless, the city continued to exist till the 770s. The most common coinage in the 760s became the Arab fels. The latest coins found on the site of the city proper date from 153/770, and those in its southern suburb from 160/ 776-7.
The abandonment of the site inhabited between the 5th and the 8th centuries did not mean the end of Panjikant as a settlement. The inhabitants only moved from the sparsely irrigated terrace, whose main advantage was its easy defense, and took up residence in the fertile valley below, watered by several springs. The consolidation of Abbasid authority led to the disintegration and agrarianization of small but stable and structurally complex urban communities like that of 6th-8th century Panjikant, and facilitated the growth of large cities like Samarqand, which became the new power-centers. To these cities flowed revenues from vast tracts of land, and there they were further distributed. Nevertheless, Panjikant of the 9th-10th centuries continued to exist as a small town. Excavations to the northwest of the citadel uncovered a large winery from the first half of the 9th century, while digs in the gardens to the northeast of the ancient ruins unearthed superb ceramics from the 9th-11th centuries. According to Arab geographers, 10th-century Panjikant boasted a Friday mosque—a formal feature that distinguished a town from a village. This was the easternmost city of Soghd, famous for its walnuts. The existence of a pre-Islamic settlement near Panjikant was noted for the first time in a document from the 16th century.
Thus the city of Panjikant, located on the upper terrace, ceased to exist in the third quarter of the 8th century. It lost its pre-Islamic character around 750. The florescence of the pre-Islamic city dates back to an earlier period—the end of the 7th and the first two decades of the 8th centuries, and the grounds for this florescence were prepared in stages, between the 5th and the middle of the 7th centuries.
Copyright © 1998-2015 The Circle of Ancient Iranian Studies (CAIS)